Environmental Protection and Disaster Risks: Proceeding of the 2nd International Conference on Environmental Protection and Disaster Risks and 10th Annual CMDR COE Conference on Crisis Management and Disaster Response 3031267532, 9783031267536

Table of contents :
Preface
Organization
Contents
Disaster Management, Natural Hazards, Risk reduction and Building Resilience
Assessment of the Effects of Strong Earthquakes on the City of Ruse
1 Introduction
2 Creating a Database for the Building Stock
3 Methodology for Seismic Risk Assessment
4 Identification and Classification of the Building Stock
5 Seismic Vulnerability Assessment of the Building Stock
6 Deterministic Earthquake Scenario for the City of Ruse in Terms of Intensity
7 Assessment of Direct Damages and Destruction
8 Assessment of Social Losses - Injured People and Victims
9 Estimation of the Economic Losses Due to Direct Physical Damage and Destruction
10 Conclusion
References
Earthquake Risk Scenarios for the City of Veliko Tarnovo
1 Introduction
2 Tectonic Settings
3 Soil Properties
4 Ground Motion Attenuation (GMPE’s)
5 Deterministic Scenarios for the City of Veliko Tarnovo
6 Probabilistic Seismic Hazard Assessment
6.1 Seismic Source Model
6.2 Seismic Source Parameters
6.3 Treatment of Uncertainties
7 Probabilistic Earthquake Scenarios for the City of Veliko Tarnovo
8 Risk Scenarios and Population Exposure
8.1 Population and Buildings of the City of Veliko Tarnovo
8.2 Methodology for Calculation of the Population Exposure Index (PEI)
9 Results and Discussion
References
Recent Trends in Hail Precipitation for Administrative Districts of Bulgaria
1 Introduction
2 Data and Methods
3 Climate Analysis of Distribution of Hail Days During the Period 1991–2020
4 Comparative Analysis of Annual Distribution of Hail Days During the Period 1961–2020
5 Conclusions
References
Torrential Catchments from Belasitsa Mountain (SW Bulgaria) - Geological and Geomorphological Characteristics and Related Hazards
1 Introduction
2 Study Area
2.1 Geology and Geomorphology Settings
2.2 Climate, Vegetation, and Anthropogenic Influence
3 Data and Methods
3.1 Data
3.2 Methodology
4 Results
4.1 December 2021 Interacting Hazardous Processes
4.2 Morphometric Analysis – Linear, Aerial and Relief Parameters of the Catchments
5 Conclusions
References
Identification of Coastal Flooding Hotspots in a Large Bay Using an Index-Based Risk Assessment Approach
1 Introduction
2 Study Site
3 Data and Methods
3.1 Flooding Hazard
3.2 Exposure Vulnerability
3.3 Coastal Index
4 Results and Discussion
4.1 Flooding Hazard Evaluation
4.2 Exposure Evaluation
4.3 Coastal Index Evaluation
5 Conclusions
References
Evaluation of the Nakamura Vulnerability Index of a Cast-in-Situ Reinforced-Concrete Building from Ambient Noise Records
1 Introduction
2 Methodology and Measurements
3 Results
4 Conclusions
References
Historical Earthquakes and Tsunami Waves in the Sea of Marmara: Review and Modelling
1 Introduction
2 Historical Earthquakes and Tsunami Waves
3 Tsunami Simulation Procedure
4 Results of Tsunami Simulations
4.1 Șarköy-Mürefte Tsunami Simulation
4.2 1999 Kocaeli Tsunami Simulation
5 Conclusions
References
Climate Change and Security Implications
Wind Speed and Temperature Variations in Burgas Region Since 1836
1 Introduction
2 Data and Methods
3 Results
3.1 Time Series Spectra
3.2 TSI Influence on Wind Velocity and Air Temperature
4 Conclusions
References
Climate Change Implications on the Condition of the Road Surface in Bulgaria
1 Introduction
2 Materials and Methods
3 Research and Results
4 Conclusion
References
Climate Change Challenges and Security Implications on National Security in North Macedonia
1 Introduction
2 Security Policy and Climate Change
3 Climate Change Challenges and Security Implications on National Security in North Macedonia
4 Conclusion
References
The Frequency of Freeze-Thaw Cycles Across Balkan Peninsula in the Period 1991 – 2020
1 Introduction
2 Data and Methods
3 Results and Discussion
3.1 Annual Frequencies
3.2 Seasonal Distribution
3.3 Monthly Distribution
3.4 Trend Analyzes
4 Conclusions
References
Assessment of Contemporary Climate Change in Bulgaria Using the Köppen-Geiger Climate Classification
1 Introduction
2 Data and Methods
3 Results and Discussion
4 Conclusions
References
An Artificial Intelligence and Simulation Approach to Climate-Conflict-Migration Driven Security Issues
1 Background
1.1 Climate Change, Conflict and Migration Driven Challenges
1.2 Climate Change Impact in Nigeria Camp Sites
2 Aim, Objectives and Methodology
2.1 A Novel Approach to Response and Humanitarian Security
2.2 MIP Medical and Security Threats Analysis
2.3 MASA SYNERGY Simulation in Humanitarian Crisis
2.4 SYNERGY Simulation Decisional Process Workflow
3 Results and Discussion
4 Conclusions
References
Estimation of the Historical and Future Renewable Energy Potential with RegCM4 over the Region of Southeastern Europe
1 Introduction
2 Methods
3 Results
4 Conclusion
References
Variations of Daily Danube Discharge at 16 Gauging Stations
1 Introduction
2 Data and Methods
3 Time Series Spectra
4 Seasonal Variations and Floods
5 Solar Signals in Maximal Discharge
6 Conclusions
References
Investigation of the Dependence of Ultraviolet Radiation on the day
1 Introduction
2 Data and Instruments
3 Results
4 Conclusion
References
Air Pollution and Health
Time Series Analysis of Asthma Hospital Admissions and Air Quality in Sofia – A Pilot Study
1 Introduction
2 Methods
2.1 Hospital Admission Data
2.2 Air Pollution and Meteorological Data
2.3 Statistical Analysis
3 Results
3.1 Description of the Data
3.2 Time Series Analysis
4 Discussion
4.1 Overall Findings
4.2 Strengths and Limitations
5 Conclusions
References
Impact of Regulatory Measures on Pollutants Concentration in Urban Street Canyon – A Pilot Study
1 Introduction
2 Methods and Input Data
2.1 Methods
2.2 Meteorological Input
2.3 Emissions Input
3 Models Set-Up and Flows
3.1 Weather Research and Forecasting Model Set-Up
3.2 ADMS-Urban System Set-Up
4 Results and Discussion
5 Conclusions
References
Multivariate Statistical Modelling of Urban Air-Quality
1 Introduction
2 Materials and Methods
2.1 Data Harvesting
2.2 Principal Component Analysis
3 Results and Discussion
4 Conclusions
References
Transport Emissions from Sofia’s  Streets - Inventory, Scenarios, and Exposure Setting
1 Introduction
2 Research Framework, Materials and Methods
2.1 Materials and Methods for Rapid Traffic Modeling and Emission Inventory
2.2 Materials and Methods for Fleet Inventory and Scenario Development
2.3 Materials and Methods for Urban Morphology Modeling
2.4 Materials and Methods for Activities Modeling
3 Results
3.1 Rapid Traffic Modeling, Fleet and Emission Inventory, Scenarios and Urban Street Canyons
3.2 Activities Modeling and Exposure Setting
4 Discussion
5 Conclusions
References
Temporal Variations of Black Carbon in the Urban Air Particulate Matter of Sofia–Observed and Modelled
1 Introduction
2 Methodology
2.1 Sampling Site and Equipment
2.2 Black Carbon Analysis
2.3 Description of Modelling Systems
3 Results and Discussions
3.1 Temporal Variation of PM2.5 and BC in Sofia
3.2 Modelled Results
3.3 Specific Episodes
4 Conclusions
References
Evaluation of the Effects of the National Emission Reduction Strategies for Years 2020–2029 and After 2030 on the Sulphur and Nitrogen Wet and Dry Depositions on the Territory of Bulgaria
1 Introduction
2 Materials and Method
3 Emission Scenarios and Numerical Experiments
3.1 Results for Nitrogen Depositions for Different Scenarios
3.2 Results for Sulphur Depositions for Different Scenarios
3.3 NMB Nitrogen Deposition for Different Scenarios
3.4 NMB Sulphur Deposition for Different Scenarios
4 Conclusion
References
PAHs and Black Carbon in Urban Air Particulate Matter in Bulgaria
1 Introduction
2 Methodology
2.1 Sampling Sites and Equipment Description
2.2 Analytical Techniques
2.3 Source Identification of BC Mass Concentration and PAHs
2.4 Estimation of Cancer Risk of PAHs
3 Results and Discussions
3.1 PM2.5 Mass Concentrations in Sofia and Burgas
3.2 PAHs, BC and BrC Concentration
3.3 Diagnostic Ratio for Source Identification of PAHs in PM2.5
3.4 Estimation of Cancer Risk
4 Conclusions
References
Forecasting Hourly NO2 and O3 Concentrations Using Data Analytics Models at Pavlovo Station in Sofia*-12pt
1 Introduction to Data Analytics Air Pollution Models
2 Data Description and Exploratory Data Analysis
3 Mixtures-of-Experts Time Series Regression Models
4 Model Predictors
5 Data Splitting for Time Series Validation
6 Accuracy Measures
7 Graphical Models Assessment
8 Conclusions
References
Solar and Climate Impacts on Groundwater Level Variations in Two Wells in England
1 Introduction
2 Study Sites
3 Data and Methods
4 Results and Discussion
4.1 Comparison of the Time Series Spectra
4.2 Precipitation Influence on Groundwater Level Variations
4.3 Common Cycles of Solar and Groundwater Level Variations
5 Conclusions
References
Water Resources and Management
Plastic Ingestion by Phocoena Phocoena and Tursiops Truncatus from the Black Sea
1 Introduction
2 Materials and Methods
3 Results
4 Discussion
5 Conclusions
References
The Role of Precipitation Variability in Water Content at Four Reservoirs in Central Western Bulgaria for the Period 2016–2019
1 Introduction
2 Data Sets
3 Results
4 Conclusion
References
Informatics, Remote Sensing, GIS, High Performance Computing
Application of Drones in Crises Management Supported Mobile Applications and C4IRS Systems
1 Introduction
2 Choosing UAV with the Fuzzy AHP (Analytical Hierarchical Process) and the VIKOR (Multi-criteria-Compromise Ranking) Method
3 C4IRS
4 Conclusion
References
Verification of 2020 Geomagnetic Models Over the Bulgarian Territory
1 Introduction
1.1 Geomagnetic Models
1.2 Repeat Station Networks
1.3 Geomagnetic Observatories
2 Bulgarian Infrastructure for Geomagnetic Observations
3 Bulgarian Geomagnetic Models
4 Data Used for Model Verification
5 Results and Discussion
References
Atmospheric Boundary-Layer Height at Marine and Land Air Masses Based on Sodar Data
1 Introduction
2 Measuring Site and Equipment
3 Data and Analysis Overview
3.1 Data Availability
3.2 Analysis
4 Results
4.1 Nocturnal Stable Boundary Layer (SBL) of Air Masses from the Land
4.2 Nocturnal CBL of Marine Air Masses
4.3 Daytime CBL of Land Air Masses
5 Conclusions
References
Bulgarian Platform for Natural Hazards Data Collection and Decision Support in Field Operations
1 Introduction
1.1 Digital Systems About Hazards in Bulgaria
2 Forest Fires in Bulgaria
3 Flood Events in Bulgaria
3.1 Small Floods
3.2 Dangerous Floods
3.3 Very Dangerous Floods
3.4 Devastating Floods
3.5 Calamitous Floods
4 RiskMap Platform and Test Areas
5 Conclusion
References
Author Index

Citation preview

Lecture Notes in Networks and Systems 638

Nina Dobrinkova Orlin Nikolov   Editors

Environmental Protection and Disaster Risks Proceeding of the 2nd International Conference on Environmental Protection and Disaster Risks and 10th Annual CMDR COE Conference on Crisis Management and Disaster Response

Lecture Notes in Networks and Systems

638

Series Editor Janusz Kacprzyk, Systems Research Institute, Polish Academy of Sciences, Warsaw, Poland

Advisory Editors Fernando Gomide, Department of Computer Engineering and Automation—DCA, School of Electrical and Computer Engineering—FEEC, University of Campinas—UNICAMP, São Paulo, Brazil Okyay Kaynak, Department of Electrical and Electronic Engineering, Bogazici University, Istanbul, Turkey Derong Liu, Department of Electrical and Computer Engineering, University of Illinois at Chicago, Chicago, USA Institute of Automation, Chinese Academy of Sciences, Beijing, China Witold Pedrycz, Department of Electrical and Computer Engineering, University of Alberta, Alberta, Canada Systems Research Institute, Polish Academy of Sciences, Warsaw, Poland Marios M. Polycarpou, Department of Electrical and Computer Engineering, KIOS Research Center for Intelligent Systems and Networks, University of Cyprus, Nicosia, Cyprus Imre J. Rudas, Óbuda University, Budapest, Hungary Jun Wang, Department of Computer Science, City University of Hong Kong, Kowloon, Hong Kong

The series “Lecture Notes in Networks and Systems” publishes the latest developments in Networks and Systems—quickly, informally and with high quality. Original research reported in proceedings and post-proceedings represents the core of LNNS. Volumes published in LNNS embrace all aspects and subfields of, as well as new challenges in, Networks and Systems. The series contains proceedings and edited volumes in systems and networks, spanning the areas of Cyber-Physical Systems, Autonomous Systems, Sensor Networks, Control Systems, Energy Systems, Automotive Systems, Biological Systems, Vehicular Networking and Connected Vehicles, Aerospace Systems, Automation, Manufacturing, Smart Grids, Nonlinear Systems, Power Systems, Robotics, Social Systems, Economic Systems and other. Of particular value to both the contributors and the readership are the short publication timeframe and the world-wide distribution and exposure which enable both a wide and rapid dissemination of research output. The series covers the theory, applications, and perspectives on the state of the art and future developments relevant to systems and networks, decision making, control, complex processes and related areas, as embedded in the fields of interdisciplinary and applied sciences, engineering, computer science, physics, economics, social, and life sciences, as well as the paradigms and methodologies behind them. Indexed by SCOPUS, INSPEC, WTI Frankfurt eG, zbMATH, SCImago. All books published in the series are submitted for consideration in Web of Science. For proposals from Asia please contact Aninda Bose ([email protected]).

Nina Dobrinkova · Orlin Nikolov Editors

Environmental Protection and Disaster Risks Proceeding of the 2nd International Conference on Environmental Protection and Disaster Risks and 10th Annual CMDR COE Conference on Crisis Management and Disaster Response

Editors Nina Dobrinkova Institute of Information and Communication Technology Bulgarian Academy of Sciences Sofia, Bulgaria

Orlin Nikolov Crisis Management and Disaster Response Centre of Excellence Sofia, Bulgaria

ISSN 2367-3370 ISSN 2367-3389 (electronic) Lecture Notes in Networks and Systems ISBN 978-3-031-26753-6 ISBN 978-3-031-26754-3 (eBook) https://doi.org/10.1007/978-3-031-26754-3 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

Environmental Protection and Disaster Risk topics are challenging fields that scientific world is trying to address as much as it can. Earthquakes, floods, fires, droughts, blizzards, dust storms, natural releases of toxic gases and liquids, diseases, and other environmental variations affect hundreds of millions of people each year. Many disaster events are triggered by human activities. Examples that affect the environment and natural biodiversity are activities such as adding contaminants to air and water, changing land use, reducing and fragmenting the habitat of some species, introducing non-native species, and changing natural fluxes and cycles of energy and materials. The challenges associated with environmental protection today are multifaceted and affected by many interacting factors. Usually, they cover various, often large, spatial scales, unfold on long temporal scales, and have global implications (e.g., carbon dynamics, nutrient cycles, and ocean acidification). Dealing with these problems will require systems thinking and integrating multidisciplinary science. Actions in these directions are taken more and more in the recent years by political bodies, NGOs, and scientific groups trying to find sustainable solutions for the future generations. Every point of view matters when it comes to our global home—The Planet Earth. This volume is a result of discussions done during the 2nd International Conference on “Environmental protection and disaster Risks,” Sofia, Bulgaria, 2022, held together with the 10th Annual CMDR COE Conference on Crisis Management and Disaster Response. It was a hybrid participation event in the period June 6–9, 2022. The participants have agreed that the relevance of the conference topic and quality of the contributions have clearly suggested that a more comprehensive collection of extended contributions devoted to the area would be very welcome and would certainly bring value to a wider public in the field of environmental protection and disaster risks. The topics covered by this volume are: disaster management, natural hazards, risk reduction and building resilience, climate change challenges and security implications, air pollution and health, water resources and management and informatics, remote sensing, GIS, and high-performance computing. November 2022

Nina Dobrinkova Orlin Nikolov Editors of Environmental Protection and Disaster Risks ENVIRORISKs 2022

Organization

The 2nd International Conference on “Environmental protection and disaster Risks,” Sofia, Bulgaria, 2022, held together with the 10th Annual CMDR COE Conference on Crisis Management and Disaster Response as a hybrid participation event in the period June 6–9, 2022.

Conference Co-chairs Kostadin Ganev Nina Dobrinkova Dimitar Dimitrov Orlin Nikolov Georgi Gadzhev Nikolay Miloshev Petya Trifonova Stelian Dimitrov

Bulgarian Academy of Sciences, Bulgaria IICT-BAS, Bulgaria CMDR COE, Bulgaria CMDR COE, Bulgaria NIGGG-BAS, Bulgaria NIGGG-BAS, Bulgaria NIGGG-BAS, Bulgaria Sofia University, Bulgaria

Program Committee Alexander Arakelyan Artemi Cerdà Anna Ganeva Aleksandar Petrovski Bojko Berov Chuck Bushey Christos Dimopoulos Constantin Ionescu Dimitrios Melas Dimcho Solakov Evangelos Katsaros George Boustras George Drakatos Georgios Eftychidis Geert Seynaeve Hrachya Astsatryan

American University in Armenia, Armenia University of Valencia, Spain IBER-BAS, Bulgaria Uni “Goce Delcev”-Stip and MA “Gen.Mihailo Apostolski”-Skopje, N.Macedonia GI-BAS, Bulgaria International Association of Wildland Fire, USA European University Cyprus, Cyprus NIRD for Earth Physics, Romania Aristotle University of Thessaloniki, Greece NIGGG-BAS, Bulgaria European University Cyprus, Cyprus European University Cyprus, Cyprus National Observatory of Athens, Greece Center for Security Studies-KEMEA, Greece EUSDEM, Belgium IIAP, NAS of Armenia

viii

Organization

Harald Pauli Horst Schwichtenberg Ilias Gkotsis Ivan Georgiev Jesús Rodrigo-Comino Kristalina Stoykova Nikolai Dobrev Reneta Dimitrova Roberto San Jose Snejana Moncheva Stefan Florin Balan Tania Marinova Todor Gurov Velichka Milousheva Viacheslav Berman Yancho Todorov

Global Observation Research Initiative in Alpine Environments, Austria Fraunhofer SCAI, Germany Center for Security Studies-KEMEA, Greece NIGGG-BAS, Bulgaria Uni-Valencia, Spain and Uni-Trier, Germany GI-BAS, Bulgaria GI-BAS, Bulgaria Sofia University and NIGGG-BAS, Bulgaria Technical University of Madrid, Spain IO-BAS, Bulgaria NIRD for Earth Physics, Romania NIMH, Bulgaria IICT-BAS, Bulgaria IMI-BAS, Bulgaria IH-NAS of Ukraine, Ukraine VTT Technical Research Centre of Finland, Finland

Contents

Disaster Management, Natural Hazards, Risk reduction and Building Resilience Assessment of the Effects of Strong Earthquakes on the City of Ruse . . . . . . . . . Dimitar Stefanov, Dimcho Solakov, and Jordan Milkov

3

Earthquake Risk Scenarios for the City of Veliko Tarnovo . . . . . . . . . . . . . . . . . . . Dimcho Solakov, Stela Simeonova, Petya Trifonova, Plamena Raykova, and Metodi Metodiev

15

Recent Trends in Hail Precipitation for Administrative Districts of Bulgaria . . . . Lilia Bocheva and Vulcho Pophristov

28

Torrential Catchments from Belasitsa Mountain (SW Bulgaria) Geological and Geomorphological Characteristics and Related Hazards . . . . . . . Zornitsa Dotseva, Ianko Gerdjikov, and Dian Vangelov Identification of Coastal Flooding Hotspots in a Large Bay Using an Index-Based Risk Assessment Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nataliya Andreeva, Petya Eftimova, Nikolay Valchev, Bogdan Prodanov, Todor Lambev, and Lyubomir Dimitrov Evaluation of the Nakamura Vulnerability Index of a Cast-in-Situ Reinforced-Concrete Building from Ambient Noise Records . . . . . . . . . . . . . . . . . Emil Oynakov, Radan Ivanov, Irena Aleksandrova, Jordan Milkov, and Mariya Popova Historical Earthquakes and Tsunami Waves in the Sea of Marmara: Review and Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lyuba Dimova and Reneta Raykova

40

51

66

77

Climate Change and Security Implications Wind Speed and Temperature Variations in Burgas Region Since 1836 . . . . . . . . Yavor Chapanov

91

Climate Change Implications on the Condition of the Road Surface in Bulgaria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Georgi Belev, Petja Ivanova-Radovanova, Vladimir Ivanov, and Hristo Chervenkcov

x

Contents

Climate Change Challenges and Security Implications on National Security in North Macedonia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Aleksandar Petrovski, Nenad Taneski, Andrej Iliev, and Nikola Spasov The Frequency of Freeze-Thaw Cycles Across Balkan Peninsula in the Period 1991 – 2020 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 Vulcho Pophristov, Hristo Chervenkov, Radoslav Evgeniev, Lilia Bocheva, and Dimitrina Todorova Assessment of Contemporary Climate Change in Bulgaria Using the Köppen-Geiger Climate Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 Krastina Malcheva and Lilia Bocheva An Artificial Intelligence and Simulation Approach to Climate-Conflict-Migration Driven Security Issues . . . . . . . . . . . . . . . . . . . . . . . 149 Walter David, Michelle King-Okoye, Irene Mugambwa, and Beatriz Garmendia Doval Estimation of the Historical and Future Renewable Energy Potential with RegCM4 over the Region of Southeastern Europe . . . . . . . . . . . . . . . . . . . . . . 160 Vladimir Ivanov, Georgi Gadzhev, Kostadin Ganev, and Ivelina Georgieva Variations of Daily Danube Discharge at 16 Gauging Stations . . . . . . . . . . . . . . . . 170 Yavor Chapanov and Emil Bournaski Investigation of the Dependence of Ultraviolet Radiation on the day . . . . . . . . . . 177 Rumiana Bojilova, Plamen Mukhtarov, and Nikolay Miloshev Air Pollution and Health Time Series Analysis of Asthma Hospital Admissions and Air Quality in Sofia – A Pilot Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 Angel M. Dzhambov, Krasimira Dikova, Tzveta Georgieva, Plamen Mukhtarov, and Reneta Dimitrova Impact of Regulatory Measures on Pollutants Concentration in Urban Street Canyon – A Pilot Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 Margret Velizarova and Reneta Dimitrova Multivariate Statistical Modelling of Urban Air-Quality . . . . . . . . . . . . . . . . . . . . . 216 Stefan Tsakovski, Ventsislav Danchovski, Reneta Dimitrova, and Plamen Mukhtarov

Contents

xi

Transport Emissions from Sofia’s Streets - Inventory, Scenarios, and Exposure Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 Angel Burov and Danail Brezov Temporal Variations of Black Carbon in the Urban Air Particulate Matter of Sofia–Observed and Modelled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 Elena Hristova, Emilia Georgieva, and Blagorodka Veleva Evaluation of the Effects of the National Emission Reduction Strategies for Years 2020–2029 and After 2030 on the Sulphur and Nitrogen Wet and Dry Depositions on the Territory of Bulgaria . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 Ivelina Georgieva, Georgi Gadzhev, Kostadin Ganev, and Vladimir Ivanov PAHs and Black Carbon in Urban Air Particulate Matter in Bulgaria . . . . . . . . . . 260 Elena Hristova, Blagorodka Veleva, Stela Naydenova, Anife Veli, Zilya Mustafa, and Lenia Gonsalvesh-Musakova Forecasting Hourly NO2 and O3 Concentrations Using Data Analytics Models at Pavlovo Station in Sofia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272 Neyko Neykov, Nadya Neykova, Anton Petrov, Tatiana Spassova, Hristomir Branzov, and Valeri Nikolov Solar and Climate Impacts on Groundwater Level Variations in Two Wells in England . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284 Tatiana Orehova, Yavor Chapanov, and Emil Bournaski Water Resources and Management Plastic Ingestion by Phocoena Phocoena and Tursiops Truncatus from the Black Sea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 Svetlana Mihova, Valentina Doncheva, Kremena Stefanova, Elitsa Stefanova, Dimitar Popov, and Marina Panayotova The Role of Precipitation Variability in Water Content at Four Reservoirs in Central Western Bulgaria for the Period 2016–2019 . . . . . . . . . . . . . . . . . . . . . . 308 Nikolay Rachev Informatics, Remote Sensing, GIS, High Performance Computing Application of Drones in Crises Management Supported Mobile Applications and C4IRS Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321 Aleksandar Petrovski, Dimitar Bogatinov, Marko Radovanovic, and Marko Radovanovic

xii

Contents

Verification of 2020 Geomagnetic Models Over the Bulgarian Territory . . . . . . . 335 Petya Trifonova, Metodi Metodiev, and Ivaylo Radev Atmospheric Boundary-Layer Height at Marine and Land Air Masses Based on Sodar Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346 Damyan Barantiev and Ekaterina Batchvarova Bulgarian Platform for Natural Hazards Data Collection and Decision Support in Field Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361 Nina Dobrinkova Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371

Disaster Management, Natural Hazards, Risk reduction and Building Resilience

Assessment of the Effects of Strong Earthquakes on the City of Ruse Dimitar Stefanov(B) , Dimcho Solakov, and Jordan Milkov National Institute of Geophysics, Geodesy and Geography (NIGGG), Bulgarian Academy of Sciences, Acad. G. Bonchev Str. Block 3, Sofia, Bulgaria [email protected]

Abstract. In order to determine the effects of strong earthquakes for a given settlement, it is necessary to solve several tasks in a few stages. The first task is identification of the building stock for the surveyed area. The assessment of the effects of strong earthquakes is based on procedures for integration of seismic hazard and seismic vulnerability of the bearing structures of the building stock. The analyses are made for seismic excitation defined from seismic hazard analysis - a deterministic earthquake scenario (expressed in macroseismic intensity). The scenario map accounts soil amplification effects using the geotechnical zonation of the considered urban area. Based on the collected information about the building stock and the performed analysis, an assessment of the direct damages and destructions is obtained. The distribution of the buildings, respectively by number and by total floor area, at the different damage levels is presented in graphical form. Economic losses are calculated based on the results obtained for damages and destruction in the buildings. The main task of this research is a quantitative evaluation of the effects of strong earthquakes for the city of Ruse, based on an accurate assessment of the seismic hazard. This evaluation is of practical importance for the municipal administration, as it allows for the development of a well-grounded prevention program, priority planning of rescue teams, equipment, resources and ability to plan the need for hospital beds and shelter. Keywords: Seismic vulnerability · Building stock · Damages · Losses

1 Introduction In the last few years, a number of earthquakes have occurred on the Balkan Peninsula, some of which (Albania, Croatia) have caused significant damage, social and economic losses. The efforts focusing on Disaster Management and Disaster Risk Reduction have become increasingly important worldwide. Scientific knowledge about the sources of earthquake ground motion as well as the vulnerability of the building stock to seismic excitations make it possible to more accurately determine the adverse effects of strong earthquakes. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 3–14, 2023. https://doi.org/10.1007/978-3-031-26754-3_1

4

D. Stefanov et al.

The seismic hazard for Bulgaria is significant, which is a prerequisite for continuous updating of seismological information and subsequent seismic risk assessments. The seismic hazard for the city of Ruse is controlled by the Vrancea intermediate-depth source in Romania that is located at about 150 km to the N-NW from the city. The Vrancea seismogenic zone of Romania is a very peculiar seismic source, often described as unique in the world, and it represents a major concern for most of the northern part of Bulgaria. The events generated in this seismogenic zone are characterized by relatively deep hypocentres and wide area of macroseismic impact. Strong intermediate earthquakes with a depth of 90–230 km are realized in this region. The scenario earthquake ground motion fields in Ruse are derived postulating strong intermediate depth earthquakes in Vrancea [1].

2 Creating a Database for the Building Stock The process of gathering input data in a format that is appropriate for processing and analysis is delicate and responsible. Over the years, several public databases have undergone numerous updates, changes, and even deformations anywhere, as the subject matter of their application varies from one institution to another. Some databases lack information on the year of construction of buildings, while others lack the type of construction. Databases that are up to date a few years ago are logically missing new buildings, but contain information about the year of construction of buildings for which this information is not available in modern databases, created for another purpose. After extensive comparative analyzes and restructuring of a large part of the data, filtering and combining, the database is combined and harmonized in one general form. The graphic base is connected in two directions with the textual basis for the building stock through the cadastral identifiers of the individual buildings and properties, thus providing the possibility for the future visualization of different thematic map materials. To help this, an orthogonal coordinate division (zoning) of the graphic parts is constructed - by means of quadrants 450m x 450m. The network is individualized by alphanumeric identifiers – Fig. 1. Figure 2 shows a quadrant of the grid illustrating the location of the buildings in the city center.

3 Methodology for Seismic Risk Assessment The present study is based on the “Methodology for Analysis, Evaluation and Mapping of the Seismic Risk of the Republic of Bulgaria”, developed by the NIGGG-BAS for Ministry of Regional Development and Public Works, approved by the Minister and having the force of a normative document [2]. The first step is to identify the regions with equal macroseismic intensity referring to the seismic intensity map. Then each building is assigned to a relevant vulnerability class following a procedure based on EMS-98 [3]. The buildings are grouped according to their vulnerability class: from A to F. To each group of buildings, the number of those that reach certain damage grade is calculated directly using the values of the DPM (damage probability matrix) for the vulnerability class and depending on the seismic intensity for the location of the building.

Assessment of the Effects of Strong Earthquakes on the City of Ruse

5

The buildings with equal damage grade are selected among all the buildings in the region with equal intensity and belonging to the same vulnerability class. In this way the number of the damaged building is determined. Subsequently the area of the damaged buildings according the damage grades is also calculated from the GIS database for the region. That area is used to assess the monetary loss and the human casualties. The approach is applied to each elementary cell of the GIS representation of the region.

Fig. 1. Quadrant network numbering for the city of Ruse

Fig. 2. Illustration of the data for quadrants - the central part of the city

6

D. Stefanov et al.

4 Identification and Classification of the Building Stock Ruse is the fifth-largest city in Bulgaria, with a population of near 140 000. Ruse is situated in the northeastern part of the country, on the right bank of the Danube in the mouth of Rusenski Lom River. It is the most significant Bulgarian river port. The city is an administrative, transport and tourism center of Bulgaria, Ruse is known for its 19th and 20th century Neo-Baroque and Neo-Rococo architecture – Fig. 3.

Fig. 3. Historic buildings in the center of Ruse

The total number of buildings in the city of Ruse is 35 831. The total built-up area is 10 798 557 m2 . Table 1 illustrates the height distribution of buildings by grouping them into four groups. Most of the buildings (33354 units) can be classified as low-rise up to 3 floors, 1530 buildings are in the range of 4 to 6 floors, 846 buildings are in the range of 7 to 9 floors and only 101 buildings are above 10 floors. Table 2 illustrates the height distribution of buildings by total floor area (TFA). Low-rise buildings are extremely numerous - 93% of all buildings, but their area is only 46% of the total area. Mediumheight buildings (between 4 and 9 floors) are about 48%, that is approximately as many as those of the first group. High-rise buildings have a relatively small contribution - about 6%. Table 1. Distribution of buildings (pieces) by number of floors Floors

1 to 3

4 to 6

7 to 9

Above 10

Pieces

33354

1530

846

101

%

93,1

4,3

2,4

0,3

Table 2. Distribution of buildings (total floor area – m2 ) by number of floors Floors

1 to 3

4 to 6

7 to 9

Above 10

TFA (m2 )

4963665

2688544

2507156

639193

%

46,0

24,9

23,2

5,9

One of the important parameters of the building stock is the year of design (construction). It is directly related to the current state of design seismic codes for the year.

Assessment of the Effects of Strong Earthquakes on the City of Ruse

7

Bulgarian national seismic codes have been periodically updated and modernized. The timing of these changes is also the boundaries of the respective time periods. In this case 7 time periods are considered. Tables 3 and 4 show the distribution of buildings by construction period. The largest number of buildings (10584 units) were built in the period 1966–1977. After that period, 20 years follow, at which rates are relatively constant. There is a significant decline in the last decade. In terms of TFA, the maximum is in the period 1978–87. Table 3. Distribution of number of buildings by construction period Period

Until 1929

1930–57

1958–65

1966–77

1978–87

1988–2007

After 2007

Pieces

307

3806

4321

10584

7257

6462

3094

%

0,9

10,6

12,1

29,5

20,3

18,0

8,6

Table 4. Distribution of buildings (total floor area – m2 ) by construction period Period

Until 1929

1930-57

1958-65

1966-77

1978-87

1988–2007

After 2007

TFA m2

114129

331907

472686

2612188

3209080

2623237

1435331

%

1,1

3,1

4,4

24,2

29,7

24,3

13,3

Based on the information about the cadaster design system, each building is classified in a separate type according to the matrix of typologies of the Bulgarian buildings. The total number of typologies is 10. The distribution of buildings by typologies (number of buildings) is given in Table 5. Table 5. Distribution of buildings by typologies (number of buildings) Type

M1

M2

A1

St1

RC1

RC2

RC4

RC5

RCp6

RCp7

Pcs

18067

11974

836

36

3357

371

116

74

10

990

%

50,4

33,4

2,3

0,1

9,4

1,0

0,3

0,2

0,03

2,8

Half of the building stock falls within the typology M1 (masonry structures, unreinforced masonry with reinforced concrete slabs, beams and belts, not framed or framed by columns). There is also a significant number in M2 (masonry structures, unreinforced masonry with wooden beams without reinforced concrete belts, not framed by columns). Third is the concrete frame structures - RC1 type buildings. In the fourth place are the buildings of type RCp7 (large-panel buildings). The distribution of the buildings by typologies (total floor area) is given in Fig. 4. The area of the RC1 typology dominates all others, but the area of the large-panel buildings RCp7 is larger than that of the masonry structures M1.

8

D. Stefanov et al.

Fig. 4. Distribution of buildings by typologies (total floor area m2 )

5 Seismic Vulnerability Assessment of the Building Stock The seismic vulnerability assessment is performed for each building individually, based on information on its load-bearing structure (typology), construction time (design), number of floors and other technical parameters. Each building is assigned a corresponding vulnerability class. The results obtained are illustrated as follows. Figures 5 and 6 show the distribution of the buildings by number and by TFA - total floor area (m2 ).

Fig. 5. Number of buildings (%) in vulnerability classes

The largest number of buildings fall into class “C” - 35%. In second and third place are the buildings of class “B” - 33% and class “D” - 27%. However, when considering the total built-up area, class “D” dominates with 42,7%. In second place is class “C” with 39%. In third place is class “E” with 9.6% (compared to 1,32% in number of buildings).

Assessment of the Effects of Strong Earthquakes on the City of Ruse

9

Fig. 6. Total floor area (%) in vulnerability classes

The most drastic difference is in class “B” - only 7% (33% in terms of number of buildings).

6 Deterministic Earthquake Scenario for the City of Ruse in Terms of Intensity Seismic history of Ruse shows that the hazard for the city is mainly influenced by the intermediate depth quakes occurred in the region of Vrancea (Romania). The scenario is based on the information about Vs30 for the built-up part of the city (Fig. 7a). The Vs30 values are derived from [1] and USGS slope-based global map (https://earthquake. usgs.gov/) [4]. The scenario earthquake is MW 7,5 Vrancea quake at depth 94 km and hypocentral distance of 240 km [1]. A deterministic earthquake scenario in terms of intensity is developed in three steps: – Prediction of peak ground acceleration (PGAgm - the geometric mean of two horizontal components) using the ground motion prediction equation presented in [5] – 88.3 cm/s2 for EC8 soil class B and 112.8cm/s2 for soil class C – Conversion of PGA in intensity using relation [6]:   I = 3.058 log PGAgm + 0.731 (1) where I is the macroseismic intensity (MSK) and PGAgm is in cm/s2 – I(88.3) = 6.7, I(112.8) = 7.0 – Correction of the intensity due to high groundwater level – increasing of intensity with 0.5 for territories close to Danube river characterized with water-saturated sands and cohesionless gravels up to 10 m thick [1]. The deterministic earthquake scenario in terms of intensity is presented in Fig. 7b.

10

D. Stefanov et al.

Fig. 7. Vs30 distribution (left) and deterministic earthquake scenario for the city of Ruse in terms of intensity (right)

7 Assessment of Direct Damages and Destruction The building stock of Ruse is divided into three main groups, taking into account the specifics of the urban development and the available information on the function and habitation of the buildings - residential buildings; industrial, commercial and administrative buildings and special buildings. The first group includes residential buildings in different parts of the city. The second group includes all industrial buildings in the industrial areas, farm buildings and administrative buildings in the rest of the city. The third group is conventionally called “special buildings”. It includes buildings with a high concentration of people and those of particular importance in a post-earthquake period. This group includes: hospitals, schools, universities and kindergartens. The distribution of buildings, by number and total floor area - TFA (m2 ), in the different levels of damage is summarized in Table 6. Table 6. Distribution of buildings (by pieces and TFA) in different levels of damage Damage grade/

No damage Slight

Medium Heavy

Very heavy Destruction

Number of buildings

18769

11204

4331

1247

255

25

Number of build. (%) 52,3

31,3

12,1

3,5

0,7

0,1

Total floor area (m2 )

7002477

2844204 740471

167853 37571

5960

Total floor area (%)

64,8

26,3

1,6

0,1

6,9

0,3

The built-up area of buildings without and with slight damages is 64,8% and 26,3% respectively. Following are buildings with moderate damage – 6,9%. High levels of damage are relatively low – 1,6% for heavy and 0,3% for very heavy damages. The area

Assessment of the Effects of Strong Earthquakes on the City of Ruse

11

of the destroyed buildings is minimal – 0,1%. Figures 8 and 9 illustrate some of the results obtained. The most significant damages are observed in the central part of the city, where the oldest buildings are located and the seismic intensity is the highest (I = 7,5).

Fig. 8. Distribution of buildings with heavy damages - TFA (m2 )

Fig. 9. Distribution of buildings with very heavy damages - TFA (m2 )

In assessing the consequences, it is very important to know the distribution of unusable and demolished buildings, so that the volume of restoration work can be planned in advance. Table 7 summarizes the results obtained by detailing the respective groups of buildings - residential, industrial, administrative and special. The results clearly show that residential buildings are the most vulnerable and with seismic excitation will suffer more damages than the other two groups of buildings.

12

D. Stefanov et al. Table 7. Distribution of unused and demolished buildings (by pieces and TFA) Residential

Unused buildings (pcs) Unused buildings – TFA(m2 ) Demolished buildings (pcs) Demolished build. – TFA(m2 )

Industrial

Special

Total

488

284

5

777

78631

25147

5209

108987

15

10

1

26

5073

503

295

5871

8 Assessment of Social Losses - Injured People and Victims The socio-economic vulnerability of the urban system also needs to be assessed in terms of casualties, social disruption and economic loss for a comprehensive earthquake damage and loss scenario. Casualties in earthquakes arise mostly from structural collapses and from collateral hazards. Lethality per collapsed building for a given class of buildings can be estimated by the combination of factors representing the population per building, occupancy at the time of the earthquake, occupants trapped by collapse, mortality at collapse and mortality post-collapse. The population of Ruse is 141231 people, according to data from the National Statistical Institute (31.12.2019). The results are available in two variants, depending on when the earthquake occurs - day or night. Table 8 summarizes the social losses - the distribution of the affected people - injured and victims. Table 9 presents a detailed distribution of the injured by category. The number of people who remain without shelter (need for accommodation) is in the range 1326–1473 people. If the earthquake is in the summer, these people could be temporarily housed in tents. If the earthquake is in winter, they must be housed indoors. At an average rate of 6 m2 per person, the required indoor and heated area is 7955–8839 m2 , respectively. On the basis of detailed data on injured people, the need for hospital beds was determined immediately after the seismic event: • Day 110–124 people; • 197–220 people at night.

Table 8. Distribution of the affected people - injured and victims (day and night) Injured - by day (pcs.)

Victims - by day (pcs.)

Injured - by night (pcs.)

Victims - by night (pcs.)

169–190

47–53

302–337

88–98

Assessment of the Effects of Strong Earthquakes on the City of Ruse

13

Table 9. Distribution of the injured by category (day and night) S1-day (pcs.) S2-day (pcs.) S3-day (pcs.) S1-night (pcs.) S2-night (pcs.) S3-night (pcs.) 59–66

55–62

54–61

105–117

97–108

100–112

Designations: S1 - Minor injuries; S2 - In need of hospital treatment; S3 - Severely injured.

9 Estimation of the Economic Losses Due to Direct Physical Damage and Destruction The economic losses are calculated on the basis of the results obtained for the damages and destructions of the three main groups of buildings. The results are summarized in Table 10. Table 10. Economic losses in the main groups of buildings

Economic losses (thousands of BGN)

Residential

Industrial

Special

Total

82651

34028

5650

122329

Comparison of results for individual groups shows the largest share of buildings in the first group (residential) - BGN 82 million. The losses in the second group of buildings (industrial and administrative) are smaller - approximately 40% than those for the first group. The losses in the third group (special buildings) are much smaller, which is favorable in terms of their importance, immediately after the earthquake.

10 Conclusion The assessment of the consequences of catastrophic earthquakes is based on procedures for integrating seismic hazard with the seismic vulnerability of building structures. A detailed quantitative assessment of the consequences in the form of losses is made as follows: • Assessment of direct damages and destructions of the building stock. The results clearly show that residential buildings are the most vulnerable and during an earthquake will suffer more damages than the other two groups of buildings. This fact is important and should be taken into account when planning the necessary resources (building materials and labor) to recover the damaged building stock. • Assessment of social losses - injured people and victims. The assessment of casualties and injuries is based on parameters that take into account the specific conditions in the study area, such as type of construction system, building density, time of the event, etc. The number of people who remain without shelter (need for accommodation) is determined. On the basis of the detailed data on injured people, the need for hospital beds is determined immediately after the earthquake. In general, social losses are

14

D. Stefanov et al.

greater at night. The main reason for this is that then people are at home and residential buildings are the most vulnerable and with the greatest damages and destruction. • Estimation of economic losses due to direct physical damage and destruction of the building stock of Ruse. These amounts can be used to develop disaster response plans and plan the need for financial resources to rebuild the building. Relevance of management decision-making results and their benefits to society: • Opportunity to develop a well-grounded prevention program; • Opportunity for priority planning of rescue teams, equipment, resources by the Ruse municipal administration; • Ability to plan the need for hospital beds and shelter.

Acknowledgements. This work has been carried out in the framework of the National Science Program “Environmental Protection and Reduction of Risks of Adverse Events and Natural Disasters”, approved by the Resolution of the Council of Ministers No 577/17.08.2018 and supported by the Ministry of Education and Science (MES) of Bulgaria (Agreement No D01– 363/17.12.2020).; and also in the framework of the Contract No D01–404/18.12.2020 (project “National Geoinformation Center (NGIC)” financed by the National Roadmap for Scientific Infrastructure 2020–2027.

References 1. Solakov, D., Simeonova, S., Raykova, P., Rangelov, B., Ionescu, C.: Earthquake scenarios for the city of Ruse. In: Proceeding of 1st International Conference on Environmental Protection and Disaster RISKs, Sofia, October 2020 2. Methodology for analysis, evaluation and mapping of seismic risk of the Republic of Bulgaria, Ministry of Regional Development and Public Works, Construction and Architecture, Issue no. 5 (2018). ISSN 0324-0711. (in Bulgarian) 3. Grunthal, G., Musson, R., Schwarz, J., Stecci, M. (eds.).: European Macroseismic Scale 1998-EMS98; (Cahiers du Centre Europeen de Geodynamique et de Seismologies), vol. 15, Luxembourg (1998) 4. Heath, D., Wald, D.J., Worden, C.B., Thompson, E.M., Scmocyk, G.: A global hybrid VS30 map with a topographic-slope-based default and regional map insets”. Earthq. Spectra 36(3), 1570–1584 (2020) 5. Vacareanu, R., Radulian, M., Iancovici, M., Pavel, F., Neagu, C.: Fore-arc and back-arc ground motion prediction model for Vrancea intermediate depth seismic source. J. of Earthq. Eng. 19(3), 535–562 (2015) 6. Ardeleanu, L., Neagoe, C., Ionescu, C.: Empirical relationships between macroseimic intensity and instrumental ground motion parameters for the intermediate-depth earthquakes of Vrancea region, Romania. Nat. Hazards 103(2), 2021–2043 (2020). https://doi.org/10.1007/s11069020-04070-0

Earthquake Risk Scenarios for the City of Veliko Tarnovo Dimcho Solakov , Stela Simeonova , Petya Trifonova , Plamena Raykova(B) and Metodi Metodiev

,

National Institute of Geophysics, Geodesy and Geography, Bulgarian Academy of Sciences Acad., Georgi Bonchev Str., Bl. 3, 1113 Sofia, Bulgaria [email protected]

Abstract. Earthquakes adversely affect large parts of the Earth. Global seismic risk and vulnerability to earthquakes are increasing steadily as urbanization occupy more areas that a prone to effects of strong earthquakes. In this study, the earthquake risk posed to the city of Veliko Tarnovo (that is known as the historical capital of the Second Bulgarian Empire) is quantified by considering the seismic context of the city that contributes to its hazard and the population exposure. The population, is directly associated with the amount and special distribution of buildings in the city. Risk assessment and its associated management is a most effective approach to estimate the impact of natural hazards on the city of Veliko Tarnovo that exhibits high seismic activity. Earthquake scenarios and social vulnerability metrics are combined in a geographic information system (GIS) to identify the vulnerability of exposed population to the seismic risk, and the locations of areas with high exposure and vulnerability level. The study focuses on earthquake risk identification and assessment while the findings provide some basis for local government to review their susceptibility and preparedness. Spatial distribution of the obtained results is available in GIS format and can be used not only for scientific purposes but also for practical measures to reduce the risk and limit the consequences of a future strong earthquake. The consideration of the earthquake scenarios into the policies for seismic risk reduction will allow focusing on the prevention of earthquake effects rather than on the activities following the disasters. Keywords: Seismic hazard · Risk scenario · City of Veliko Tarnovo

1 Introduction Seismic risk scenarios may have different forms and outputs but generally, they pre-sent a combination of the probability of an event and its negative consequences. Nowadays they are tools that use the capabilities of science to protect human life and health, as well as the economy and the environment. Recently, studies on seismic hazard were performed for some of the largest cities in Bulgaria applying the approach adopted for developing ground motion hazard maps © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 15–27, 2023. https://doi.org/10.1007/978-3-031-26754-3_2

16

D. Solakov et al.

in the Risk EU Project (2001–2004). In Bulgaria, such scenarios are generated for the cities: Blagoevgrad [1], Ruse [2] and Plovdiv [3]. In addition, in some of the studies a base level of the seismic risk is also estimated using the population exposure in a region [4] or in a city [5]. The present study focuses on the seismic hazard and population exposure for another large city of Bulgaria, which has experienced strong earthquake nearly hundred years ago - the city of Veliko Tarnovo. The city is situated in the central part of Northern Bulgaria and it is one of the oldest settlements in Bulgaria, with a history of more than five millennia. Veliko Tarnovo is known as the historical capital of the Second Bulgarian Kingdom. Nowadays, the city of Veliko Tarnovo is an important administrative, economic, educational, and cultural point of Northern Bulgaria. In this regard, there is a concentration of business and administrative centres, as well as many tourist sites.

2 Tectonic Settings Tectonics of the Eastern Mediterranean is dominated by the collision of the Arabian and African plates with the Eurasian (among others [6]). The recent tectonics of Bulgaria is determined by the geotectonic of the region in which dominate the processes of extension with general direction north south. Analysis of the horizontal velocities (based on the permanent GNSS stations and geodynamic network) shows that the horizontal velocities in Northern Bulgaria are negligible (about 1 mm/y), which confirms the hypothesis that this region belongs to the Eurasian plate [7]. Figure 1 represents epicenter map of historical and instrumental earthquakes occurred near the city of Veliko Tarnovo (in seismogenic zone Gorna Oryahovitsa).

Fig. 1. Spatial pattern of seismicity (historical and instrumental earthquakes with MW > 3.0) and active faults map for the city of Veliko Tarnovo and surroundings.

In Fig. 1 is also shown the SHARE model [8] of the active faults in the area of the city of Veliko Tarnovo. Seismic history of the city shows that the 1913 MW 6.8 earthquake caused the strongest on city seismic effects in the city of Veliko Tarnovo. The consequences of the earthquake are described in the work of Spas Watzof [9]. About 95% of Gorna Oryahovitsa and 80% of Veliko Tarnovo were destroyed [9]. Some of damages of the 1913 MW 6.8 earthquake are presented in Fig. 2.

Earthquake Risk Scenarios for the City of Veliko Tarnovo

17

Fig. 2. Damages in the cities of Veliko Tarnovo and Gorna Oryahovitsa caused by the 1913 MW 6.8 earthquake (modified from [9]).

3 Soil Properties Representation of the soil properties of the city is defined by the parameter Vs30 - average shear-wave velocity in the upper 30 m of the soil/rock profile. The Vs30 values for the city of Veliko Tarnovo, (Fig. 3) are based on the results presented in [10] and the USGS slope-based Vs30 model (available at https://earthquake.usgs.gov/data/vs30). Figure 3 illustrates the varying values of Vs30 (320–470 m/s) throughout the urban area. Most of the territory is characterized with Vs30 higher than or equal to 360 m/s.

Fig. 3. Average shear-wave velocities in the upper 30 m of the soil/rock profile for the city of Veliko Tarnovo.

4 Ground Motion Attenuation (GMPE’s) In our study, six GMPE’s models for shallow and three for intermediate depth earthquakes, presented in [11] were used for seismic hazard assessment. The method presented in [12] is chosen for testing the selected models against the actually recorded ground motion data from Balkan countries, Italy and Middle East. The data used are from Engineering Strong Motion (ESM) presented in [13]. The approach presented in [14] is applied for ranking and weighting the selected GMPE’s.

18

D. Solakov et al.

5 Deterministic Scenarios for the City of Veliko Tarnovo In the deterministic scenario evaluations, specific values of magnitude can be associated to individual sources in order to compute, via attenuation relations for magnitude and ground motion parameters, the severity of ground shaking in the urban area of interest. The work on scenarios was guided by the perception that usable and realistic ground motion maps had to be produced for urban areas. In the present study, the deterministic scenarios were defined on the base of the seismogenic potential of the causative faults defined in the SHARE project. The epistemic uncertainty in the ground motion attenuation is represented by the application of the selected six GMPE’s for active shallow crustal tectonic regime. The calculation procedure is described in [1]. The deterministic scenario in PGA for MW 6.8 earthquake is mapped in Fig. 4. The estimated PGA values vary between 0.29 g and 0.41 g throughout the city. The highest PGA values (0.36 g–0.41 g) are in the north-eastern part of the city.

Fig. 4. Deterministic Earthquake Scenario for the city of Veliko Tarnovo in PGA [g].

6 Probabilistic Seismic Hazard Assessment The probabilistic seismic hazard is the probability that various levels of strong ground motion will be exceeded during a specified time period at a site. The ground motion levels may be expressed in terms of peak ground acceleration (velocity, displacement) and/or peak response spectral amplitudes for a range of frequencies. The resultant hazard at a specified site is obtained by integrating the effects of ground motion from earthquakes of different size occurring at different locations within different seismic source regions and with different frequencies of occurrence. The probabilistic approach (PSHA) was developed in the late 1960s and early 1970s to provide a systematic method to deal with the “uncertainty in the number, sizes, and

Earthquake Risk Scenarios for the City of Veliko Tarnovo

19

locations of future earthquakes” [15]. In our study, the PSHA was performed by using a version of machine code EQRISK [16] that was developed and used in practice for probabilistic hazard assessment in Bulgaria (presented in [17]). 6.1 Seismic Source Model A seismic source model, which demands translating seismotectonic information in to a spatial approximation of earthquake location and recurrence, is a key component of seismic hazard assessment. Seismicity in the 200 km region surrounding the city of Veliko Tarnovo is associated within 15 seismic sources. The regional seismic source model used in seismic hazard analysis for the city is illustrated in Fig. 5 and specified in Table 1. The source model includes all seismic sources that substantially influence the seismic hazard of the city of Veliko Tarnovo.

Fig. 5. Seismic source model (modified from [17]).

6.2 Seismic Source Parameters The seismicity statistics are defined by specifying the Gutenberg - Richter (GR) magnitude-frequency relation [18]. Parameters a and b of cumulative Gutenberg – Richter relation and Mmax (maximum observed and maximum expected magnitude) for each seismic source in the region (within a radius of 200 km) surrounding the city are presented in Table 1 (details are presented in [17]). 6.3 Treatment of Uncertainties The epistemic variability (uncertainties) comes from statistical or modelled variations. The large uncertainties in seismic hazard result from lack of knowledge about earthquake

20

D. Solakov et al.

cause, characteristics, ground motions, i.e. from uncertainties in seismic input. Modern methods of seismic hazard analysis incorporate uncertainties into the analysis to assess their impact on the estimate of the expected level of seismic hazard as well as the uncertainty in that estimate. In the present study for accounting the epistemic uncertainties in the seismic input, a logic tree approach ([19, 20]) is used. For the PSHA analysis of the city, the 4 logic tree levels have been used (details are presented in [17]): Level 1 - Seismic source modelthe seismic source model presented in Fig. 5 is considered; Level 2 - Seismic sources a logical tree is generated for each seismic source; Level 3 - Ground Motion Prediction Equations (GMPE’s) – the selected 6 GMPE’s for shallow earthquakes; Level 4 - Soil properties - the VS30 map for the city presented in Fig. 3 is applied. Table 1. Seismic source parameters. Source

Mmax

Mmax

N

Name

a

GR b

Observed

Expected

S1

G. Orjahovica

2.18

0.8

6.8

7.1

S2

East Balkan

2.25

0.8

5.0

7.1

S3

Dobrich

1.84

0.8

5.7

6.2

S4

Marica East

2.63

0.8

7.1

7.5

S5

Sr. Gora

1.83

0.8

5.2

7.0

S6

Sliven Jambol

2.52

0.8

6.0

6.9

S7

Marica West

2.03

0.8

6.3

7.0

S8

Sofia East

2.39

0.8

6.5

7.2

S9

Moesian

1.84

0.8

4.4

7.0

S10

Dulovo

2.03

0.8

6.3

7.2

S11

S-E Bulgaria

2.02

0.8

4.9

6.2

S12

Rhodope West

2.91

0.9

5.7

6.4

S13

Rhodope East

2.41

0.9

4.7

6.2

S14

West Turkey

2.68

0.9

7.1

7.7

S15

Romania-BS

2.9

0.94

5.4

6.5

GR—estimated parameters of the GR relationship; Mmax —maximum magnitude

7 Probabilistic Earthquake Scenarios for the City of Veliko Tarnovo Probabilistic seismic analysis (PSHA) for the city of Veliko Tarnovo was performed using the model of seismic sources presented in Fig. 5 and specified in Table 1. Values of PGA (in [g]) are calculated using the selected six GMPE’s for shallow and three for

Earthquake Risk Scenarios for the City of Veliko Tarnovo

21

intermediate earthquakes. Four level logic trees for each seismic source are developed to account for uncertain-ties in seismic input. The probabilistic scenario has been generated in terms of Peak Ground Acceleration (PGA) for 475 and 1000 years return periods. Figure 6 shows that the estimated PGAs for the city of Veliko Tarnovo slightly vary along the city (from 0.21 g to 0.24 g) for 475 years return period. The highest PGA values are in the north-eastern part of the city.

Fig. 6. Probabilistic earthquake scenario in PGA [g] for the city of Veliko Tarnovo for 475 years return period.

8 Risk Scenarios and Population Exposure The earthquake risk scenario for the city of Veliko Tarnovo is assessed by taking into account the seismic hazard and the population (human) exposure as an element at risk. If we add the vulnerability of that element to earthquake ground shaking, we would obtain the three factors used [21] to explain the risk. Following the multiplicative rule, it means that if the hazard is Ø (null), then the risk is also Ø (null): R = H . Pop.Vul

(1)

where R is the risk, i.e. the expected human impacts, H is the hazard, e.g. earthquake hazard, Pop is the population living in a given exposed area, Vul is the vulnerability, which depends on some additional factors (economic, social, etc.). 8.1 Population and Buildings of the City of Veliko Tarnovo According to the data of the National Statistical Institute of Bulgaria (https://nsi.bg/ nrnm/), 66 104 inhabitants live in Veliko Tarnovo by 2021. Usually that number increases

22

D. Solakov et al.

due to the students educated in the local universities and the high number of tourists but our methodology is based on the resident population only (i.e. night-time) and do not account for its spatio-temporal variation. Veliko Tarnovo is relatively large city - fifteenth in terms of population and seventeenth in terms of territory for Bulgaria, with almost 6200 residential buildings. The distribution of buildings, by number of floors and total floor area - TFA (m2 ) is given in Table 2. Almost three quarters of the buildings have low-raise constructions up to 3 floors, but they represent only 22 percent of the living space in the city. Nearly equal number of people lives in buildings with floors between 4–6, and 7–10, around 25 000 per-sons in each category. The highest density of population is observed in the central and southern part of the city. Table 2. Distribution of buildings (by number of floors and TFA) in Veliko Tarnovo Residential Buildings

1–3 floors

4–6 floors

7–10 Floors

Above 10 floors

Number of buildings

4615

1058

480

18

Number of build. (%)

74,8

17,1

7,8

0,3

TFA [m2 ]

758 413

1 324 962

1 269 350

98 614

TFA (%)

22,0

38,4

36,8

2,8

8.2 Methodology for Calculation of the Population Exposure Index (PEI) Seismic risk evaluation, which is determined by Eq. 1, might be simplified by a parameter, which we call (similar to [21]), Population Exposure (PopExp) which represents the hazard multiplied by the population: PopExp = H . Pop

(2)

Following the methodology described in [5] and using GIS we overlapped the seismic hazard and population living in the potentially affected area of the city of Veliko Tarnovo. The analysis is performed on a grid with single spatial element E, which here is the site (polygon) of every separate residential building. The input data, which are imported in GIS and used for the analysis are: 1) seismic hazard (deterministic with reference earthquake MW 6.8 and probabilistic with return periods 475 displayed in Figs. 4 and 6, for the territory of Veliko Tarnovo and 2) distribution of population according to the total number of citizens published by the National Statistical Institute (NSI) for 2020 and the data of built-up land properties of the city given by the Geodesy, Cartography and Cadastre Agency of Bulgaria (GCCA). The population density for each element of the grid was calculated using the information on the total floor area of the residential buildings in the city of Veliko Tarnovo and the total number of residents living there.

Earthquake Risk Scenarios for the City of Veliko Tarnovo

23

Table 3. Seismic hazard levels Seismic hazard

Value

Seismic hazard PGA[g]

≤ 0.10

0.1–0.14

0.141–0.18

0.181–0.25

> 0.25

Seismic hazard levels

1

2

3

4

5

Table 4. Population density levels Population density

Value

Population density [pers./100m2 ]

12

Population density levels

1

2

3

4

5

According to the intervals given in Table 3 and Table 4 respectively, we prescribe a level for the seismic hazard H E and population PopE (in terms of population density per 100 m2 ) for each spatial element of the grid. Equation 2 obtained the form: PopExpE = H E . PopE

(3)

The number of levels and the parameter variation intervals are predefined for consistency according to [4].

9 Results and Discussion Equation 3 is applied to obtain the population exposure as a base level earthquake risk scenario for all single spatial elements E of the grid. The results are categorized in the same five classes giving more readable index form of the parameter. For this reason, the five classes of a standard risk matrix (see e.g. http://jasonpope.co.uk/risk-management/) are chosen: Minor, Low, Moderate, High and Major and displayed in Fig. 7. Risk Scenario 1: Magnitude MW 6.8 earthquake is associated with the closest fault to the city of Veliko Tarnovo (displayed in Fig. 1) and deterministic hazard is assessed in terms of PGA (Fig. 4). The estimated PGA values vary between 0.29 g and 0.4g with maximum in the north-eastern part of the city that is the closest to the fault. According to the classification in Table 3 to the whole studied territory is assigned the highest seismic hazard level - 5. Concerning the population, the value for each element of the grid is calculated using the average population density of 1.915 pers./100m2 obtained from the total floor area of the residential buildings and the total number of citizens in Veliko Tarnovo, and the size of each building site. These values are classified using the levels of Table 4. Having parameters, H E and PopE we have applied Eq. 3 to obtain the population exposure for all 6172 elements of the grid. Values were classified using the model matrix of Fig. 7 to find the Population exposure index (PEI) as a base level of the seismic risk. Results are shown in Fig. 8. In this scenario, we have three categories of the Population

24

D. Solakov et al.

exposure index – Moderate, High and Major. The most endangered sites with Major value of the PEI are 1137. They are unevenly spaced over the territory of the city, predominating in its central and southern parts.

Fig. 7. Model of a standard risk matrix used for PEI classification.

Fig. 8. Deterministic earthquake risk scenario for the city of Veliko Tarnovo.

Risk Scenario 2: A seismic source model is used for probabilistic hazard assessment in terms of PGA. In Fig. 9 are presented the results for 475 years return period. These values are in a narrow interval from 0.21 g to 0.24 g and could be categorized to the penultimate seismic hazard level - 4. The grid of the population density is the same as in Scenario 1 - with assigned value for every single element according to Table. 4. In

Earthquake Risk Scenarios for the City of Veliko Tarnovo

25

the probabilistic seismic scenario for 475 years return period we have four categories of the risk – starting from Low and ending to the Major. The highest category has similar spatial distribution as in the deterministic scenario but only for 498 grid elements. About 1875 elements are classified as High-risk places and more than half of the sites are in the favourable Low and Moderate risk classes. Most of the lower level elements are observed in the western and east/north-eastern part of the city.

Fig. 9. Probabilistic earthquake risk scenario for the city of Veliko Tarnovo (475 years).

In conclusion, it should be noted that our results present a base-level earthquake risk scenario, which might be upgraded to fully operational tool if the infrastructure of the city and building vulnerability analysis is added to it. Acknowledgements. The present study has been carried out in the framework of the National Science Program “Environmental Protection and Reduction of Risks of Adverse Events and Natural Disasters”, approved by the Resolution of the Council of Ministers № 577/17.08.2018 and supported by the Ministry of Education and Science (MES) of Bulgaria (Agreement № D01-279/03.12.2021); and the project “National Geoinformation Center (NGIC)”Contract № D01-404/18.12.2020 financed by the National Roadmap for Scientific Infrastructure 2020–2027.

References 1. Solakov, D, Simeonova, S., Raykova, P., Rangelov, B.: Deterministic seismic scenarios for the city of Blagoevgrad. In: 20th Internatinal Multidisciplinary Scientific Geoconference: Science and Technologies in Geology, Exploration and Mining, SGEM 2020, vol. 2020, Issue 1.2, pp. 527–534 (2020). Albena 18–24 August 2020 (2020a)

26

D. Solakov et al.

2. Solakov, D., Simeonova, S., Raykova, P., Rangelov, B., Ionescu, C.: Earthquake scenarios for the city of ruse. In: Proceeding of 1st International Conference on Environmental Protection and Disaster RISKs. In: Gadzhev, G., Dobrinkova, N. (eds.) Part 2 Natural Hazards and Risks, pp. 371–380, Az-buki National Publishing House, Sofia (2020b) 3. Solakov, D., Simeonova, S., Raykova, P.: Earthquake scenarios for the city of Plovdiv. In: 21th International Multidisciplinary Scientific GeoConference: Science and Technologies in Geology, Exploration and Mining, SGEM 2021, vol. 2021, pp. 701–708, Albena (2021a). ISBN: 978-619-7603-20-0. ISSN: 1314-2704 4. Solakov, D., Metodiev, M., Simeonova, S., Trifonova, P.: Population exposure index – an element of seismic risk assessment. In: 10th Congress of the Balkan Geophysical Society, ALBENA 2019 (2019a). https://doi.org/10.3997/2214-4609.201902659 5. Trifonova, P., Solakov, D., Simeonova, S., Metodiev, M., Balan, S.F.: Seismic scenario and people exposure for Blagoevgrad Region, Bulgaria. In: Dobrinkova, N., Gadzhev, G. (eds.) EnviroRISK 2020. SSDC, vol. 361, pp. 293–305. Springer, Cham (2021). https://doi.org/10. 1007/978-3-030-70190-1_20 6. Jackson, J., McKenzie, D.P.: The relationship between plate motions and seismic moment tensors, and the rates of active deformation in the mediterranean and middle east. Geophys. J. 93, 45–73 (1988) 7. Georgiev, I., Dimitrov, D., Briole, P., Botev, E.: Velocity field in Bulgaria and Northern Greece from GPS campaigns spanning 1993–2008. In: 2nd INQUA-IGCP 567 International Workshop on Active 2nd INQUA-IGCP 567 International Workshop on Active Tectonics, Earthquake Geology, Archaeology and Engineering, 19–24 Corinth, Greece, pp.54–56 (2011) 8. Basili, R., Kastelic, V., et al: The European Database of Seismogenic Faults (EDSF) compiled in the frame of the Pr. SHARE (2013). http://diss.rm.ingv.it/share-edsf/. https://doi.org/10. 6092/INGV.IT-SHARE-EDSF 9. Watzof, S.: The earthquakes in Bulgaria. Report on the earthquakes felt in 1913–1916, Central Institute for Meteorology, Sofia, p. 195 (in Bulgarian, abstract in French) (1923) 10. Rangelov, B., Solakov, D., Dimovsky, St., Kisyov, A., Georgieva, B.: Mapping and digitalization of the ground conditions for the seismic hazard assessment. In: Proceedings 19th “Days of Physics 2020”, pp. 91–97, Technical University, Sofia (2020) 11. Solakov, D., Simeonova, S., Raykova, P., Oynakov, E., Aleksandrova, I.: GMPEs used in seismic hazard assessment for Bulgaria-selection and testing in Bulgaria. In; Conference Proceeding of 10th Congress of the Balkan Geophysical Society, 2019-SM2.5-N17P6, ALBENA 2019 (2019b) 12. Scherbaum, F., Delavaud, E., Riggelsen, C.: Model selection in seismic hazard analysis: an information-theoretic perspective. BSSA 99(6), 3234–3247 (2009) 13. Luzi, L., Puglia, R., et al.: The engineering strong-motion database: a platform to access pan-European accelerometric data, Seism. Res. Lett. 87(4) (2016). https://www.orfeus-eu. org/data/strong/ 14. Delavaud, E., Scherbaum, F., Kühn, N., Allen, T.: Testing the global applicability of ground-motion prediction equations for active shallow crustal regions. Bull. Seism Soc. Am. 102(2),707 (2012). https://doi.org/10.1785/0120110113 15. Cornell, C.: Engineering seismic risk analysis. BSSA 5, p.1583 (1968) 16. McGuire, R.: FORTRAN Computer Program for Seismic Risk Analysis, U.S. Geol. Surv. 491 Open-File Rep. 76–67, U.S., Rep. 1–90 (1976) 17. Solakov, D., et al.: Building Seismic Risk Management, Part 2: Probabilistic Seismic Hazard Assessment BAS Publ. House, Sofia, pp. 13–110 (in Bulgarian) (2019c) 18. Gutenberg, B., Richter, C.F.: Frequency of earthquakes in California. Bull. Seism. Soc. Am. 497(34), 185–188 (1944)

Earthquake Risk Scenarios for the City of Veliko Tarnovo

27

19. Kulkarni, R., Youngs, R., Coppersmith, K.: Assessment of confidence intervals for results of 501 seismic hazard analysis. In: Proceedings of the 8th World Conference on Earthquake Engineering, vol. 1, p. 263, San Francisco, California (1984) 20. Coppersmith, K., Youngs, R.: Capturing uncertainty in probabilistic seismic hazard assessments 504 within intraplate environments. In: Proceedings of the 3rd National Conference on Earthquake Engineering, vol. I, p, 301, Charleston, August 24–28 (1986) 21. Dao, H., Peduzzi, P.: Global Risk And Vulnerability Index Trends Per Year (GRAVITY), Phase IV: Technical annex and multiple risk integration, UNDP/BCPR, Geneva, Technical report, 31 p. (2003)

Recent Trends in Hail Precipitation for Administrative Districts of Bulgaria Lilia Bocheva(B)

and Vulcho Pophristov

National Institute of Meteorology and Hydrology, Tsarigradsko Shose Blvd. 66, 1784 Sofia, Bulgaria {lilia.bocheva,vulcho.pophristov}@meteo.bg

Abstract. Convective storms, which produce hail, cause significant damage to agriculture, buildings, vehicles, ecosystems and even cause human casualties in many countries around the world, as well as in Bulgaria. The relief and geographical location of our country characterize it as one of the countries with the highest frequency of hail event in Europe. This study summarized the monthly and annual variations of hail days for each administrative district in Bulgaria for the period 1991–2020. The data used in the analysis include daily, monthly and annual number of hail days from the meteorological network of the National Institute of Meteorology and Hydrology for the period considered. A comparison of the frequency of hail for one meteorological station, representative for each district, for two periods (1961–1990, 1991–2020) was made. In the last 30 years, there was an increase in hail days in May, as well as in the relatively cold month of March for almost all regions. The obtained trend is statistically significant only for few districts. During the cold half of the year (October-March) there is an increase in hail events in almost the whole territory of Bulgaria, more significant for districts from South Bulgaria. Keywords: Bulgaria · Hail variations · Climatology

1 Introduction Hail events are frequent in many parts of the world, especially in mid-latitudes most often during the spring and summer. The areas with the highest frequency of hail in the Northern Hemisphere are located in North America (along the Rocky Mountains from Alberta in Canada in the north to New Mexico in the south) and in Europe (from the northern Iberian Peninsula, through Central Europe and the Balkans). In the Southern Hemisphere, the highest incidence of these extreme events is in Mendoza (Argentina), South Africa and parts of Australia. For many countries in these areas, the damage caused by hail precipitation is a significant part of the damage from natural disasters in general. For the United States, these damages average about $ 1.433 billion U.S. dollars (USD) annually [1]. For Australia over the last 100 years, 11% of building damage has been caused by hail [2]. Damage from hail to buildings, vehicles, the electricity grid and crops in some Central European countries such as Germany, Switzerland and France © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 28–39, 2023. https://doi.org/10.1007/978-3-031-26754-3_3

Recent Trends in Hail Precipitation for Administrative Districts of Bulgaria

29

is estimated at several million euros per year [3]. In Cyprus, it was found that between 1996 and 2005, 91% of agricultural losses were due to hail [4]. Even the individual event can cause damages for more than $ 1billion USD [5, 6]. The studied spatial-temporal distributions and the frequency of hail precipitation are characterized by great variability due to both the peculiarities of the formation of these precipitation and the limited information in the meteorological databases about them. The conventional network of meteorological stations is not suitable for measuring such phenomena that are too local in time and place, as hail often falls in areas outside the stations and remains unregistered. Therefore, many studies use data from other sources such as insurance companies, hail-prone areas, media and eyewitness data [3, 7]. For a more detailed presentation of the characteristics of hail processes, data from a local specialized measurement network, characterized by a high density of measurements and located in hail-prone areas in the respective countries are used [8–10]. Some recent works presented the climatology of hail precipitation, based on radar or satellite information [11, 12], but they analyzed the relatively short periods of 10–15 years. However, long-term climatic studies of the spatio-temporal characteristics of hail in different countries are based mainly on data from standard meteorological observations [7, 10, 13–15]. The areas with the most frequent hail events in Europe are positioned between 39.N and 50.N [16]. Moreover, the simulation for future climate over Europe [17] show that for central and eastern parts of the continent the strong increase in frequency of severe environmental conditions is expected till end of 21st century. Not only the geographical location, but also the complex orography of Bulgaria defined it as one of the most hail-stormy countries in Europe. During the last 20 years the frequency and the severity of hail storms, mainly accompanied by heavy rain, which caused flood over Bulgaria, increase. The damages produced by individual hail events also increased recently. [18, 19]. The present study summarizes and analyzes the spatial and temporal variations of hail precipitation in all 28 administrative districts in Bulgaria during the period 1991–2020. Also long-term changes in the frequency of hail for a 60-year period (1961–2020) for individual meteorological stations, representative for each of the districts in the country, are examined. The aim of the study is to present the current changes in the frequency and distribution of hail precipitation, summarized for each administrative district of the country. The study was conducted on this principle, as the summarized results are crucial in assisting state and local authorities in developing contingency plans and policies for adaptation to climate change at the regional level.

2 Data and Methods In this study all data for existence of hail precipitation events from the meteorological records of the National Institute of Meteorology and Hydrology (NIMH) of Bulgaria for the period 1991–2020 are used, taking into account all synoptic, climatological and raingauge stations, in which regular observation are completed. Days with hail are reported for each station in which at least one such phenomenon is registered between 00:00 and 24:00 local time. The hail precipitation is recorded only when they occur in limited site

30

L. Bocheva and V. Pophristov

where the weather station is located and so their frequency seems to be under-evaluated in the recent times. On Fig. 1 the position of all meteorological stations in each of 28 administrative districts in Bulgaria are presented. The current study estimates of the monthly, seasonal and annual number of days with hail precipitation for each district in Bulgaria for the period of interest, using specialized procedures in the SQL environment. The data obtained from each station for the 30year period were analyzed, the necessary calculations were performed and the results were summarized for each district. For seasonal analysis we divided the year on 2 parts: warm half from April to September and cold half from October to March. The statistical significance of all tendencies (at the 5% significance level) for the period 1991–2020 are estimated by means of the Mann-Kendall test [20] with Sen’s slope estimator [21], which present the magnitude of the slope in number of days/10 years [16]. All annual, seasonal and monthly days with hail precipitation for each administrative district are statistically calculated.

Fig. 1. Locations of meteorological stations used in the study

The annual and seasonal hail-storm frequency for single event for two 30-years periods (1961–1990, 1991–2020), recommended by World Meteorological Organization (WMO) for calculation of previous and current climate normal, is determined for representative station from each district (S1 ÷ S28 on Fig. 1). The stations are selected in accordance to continuity and quality of their time-series. For statistical evaluation of the comparative analysis of the two periods 1961–1990 and 1991–2020, ANOVA with Poisson distribution and Chi square estimation is used, which are suitable for such discrete samples of hail days.

Recent Trends in Hail Precipitation for Administrative Districts of Bulgaria

31

3 Climate Analysis of Distribution of Hail Days During the Period 1991–2020 Analyzes for the period 1991–2020 show that the average annual number of hail days in single meteorological station for all districts vary from 0.67 in Dobrich to 1.69 in Sofia-city. These estimates do not include the average annual values for high mountain peaks because there an annual number of days with hail events is much higher (between 4.1 days on Cherni vrah and 6.73 days on peak Musala). The values obtained for the individual stations in practice represent the climatic normal for the number of days with hail precipitation and are calculated in accordance with the requirements of the World Meteorological Organization [26]. Summarizing the results for the individual stations, the average values for each district are calculated. For the last 30 years the most frequent hailfalls (over 1.2 days per year) are for the region of Sofia-city (1.69) and Sofia-District (1.3), as well as the central and northeastern parts of the Danube plain (up to 1.32 days/year in districts Targovishte and Lovech), the eastern part of Upper Thracian lowlands (up to 1.54 in district Yambol) and the Rhodopes (up to 1.38 in district Haskovo). The least hail precipitation is registered in Northeastern Bulgaria (district Dobrich – 0.69) and along the Danube river (districts Vidin – 0.76; Ruse – 0.81 etc.). Atypically small are number of days with hail precipitation during the period 1991– 2020 in part of southwestern Bulgaria – about 0.85 days/year for districts Blagoevgrad and Kyustendil. According to climate studies concerning the frequency of hail in Bulgaria in older periods of time, for the regions of Southwestern Bulgaria the frequency of these processes should be one of the largest in the country and comparable to that in the Sofia region [24, 27]. The reasons for this can be found in the changes in the precipitation regime and the atmospheric circulation in the region, as well as other changes that are not the subject of this study. The registered days with hail in at least one meteorological station in Bulgaria are 2510, and their annual variations in the period 1991–2020 have a negative, statistically insignificant trend (Fig. 2, left), which coincides with the results of earlier climate studies for the Balkan Peninsula [9, 23, 24]. In different districts number of days with hail precipitation varies between 130 days (Gabrovo) and 743 days (Sofia-District) during the 30-years period and on annual basis the trends in the distribution of hail days for 28 administrative districts in Bulgaria are quite different. There are both positive and negative trends, but only a small part are statistically significant. The results are presented on Table 1 and statistically significant trends are bolded. It can be summarized that the decreasing tendency in distribution of hail precipitation days is observed for 11 administrative districts and half of them are situated in West Bulgaria (Vidin, Vratsa, Kyustendil, Sofia-District and Sofia-city), which confirms the findings for earlier years presented in [24]. All negative trends are statistically insignificant with exception of only one – in district Pazardzhik. In the other 17 districts of the country the number of hail days’ increase during the period 1991–2020, and this positive trend is statistically significant for 6 districts – 3 in northern Bulgaria (Pleven, Veliko Tarnovo and Targovishte) and 3 in southern Bulgaria (Haskovo, Kardzhali and Blagoevgrad). Statistical estimates show that the increasing trend for hail days is most significant in Blagoevgrad district (0.36 days/10 years), followed by Targovishte (0.25 days/10 years). Blagoevgrad is the only district in Bulgaria in which there is a statistically significant increase in the

32

L. Bocheva and V. Pophristov

frequency of hail days/per decade not only on an annual basis, but also in the individual half year periods recently.

Fig. 2. Distribution of all days with hail precipitation in Bulgaria during the period 1991–2020: annual (left), cold half of the year (middle) and warm half of the year (right)

The analyses show that annual peak of hail events occurred during the period AprilJuly – about 65–70% of all cases. The maximum number of days with hail is registered in May and June, which is also typical for other countries on Balkan Peninsula – North Greece [9] and Croatia [25]. The percentage distribution of hail days for the warm and cold half of each year for each district is presented on Fig. 3. The main part of hail events for the territory of all administrative districts in Bulgaria is registered during the warm half of the year (between 76% and 96% of all cases). However, the significant share (12–24%) of hail precipitation during the cold part of the year in areas with a Continental-Mediterranean climate should be noted. The most common are autumn and winter days with hail in the eastern Rhodopes (Kardzhali – 24.1%), Strandzha and the Black Sea coast (Burgas – 18.8%; Yambol – 15.6%; Varna – 14.7% and Dobrich – 13%), as well as in the south westernmost region (Blagoevgrad – 12.7%).

Fig. 3. Distribution of hail days in percent (%) for warm half (left) and cold half (right) of the year

Recent Trends in Hail Precipitation for Administrative Districts of Bulgaria

33

Table 1. Mann-Kendall trend test results (Z and p-value) and the slope magnitude for annual and seasonal distribution of hail days in each district of Bulgaria (statistically significant trends are bolded) № District

Annual z-stat

p-value Slope

Warm half year

Cold half year

z-stat

z-stat

p-value Slope

1

Vidin

−0.13 0.90

0.00 −0.41 0.68

2

Montana

0.72

0.00 0.54

3

Vratsa

−0.04 0.97

4

Pleven

2.06

5

Lovech

−1.06 0.29

6

Gabrovo

1.39

0.16

0.06 1.14

7

V.Tarnovo

2.06

0.04

8

Ruse

0.29

0.77

9

Targovishte

2.77

10 Razgrad 11 Silistra

0.47 0.04

0.59

0.00 −0.07 0.94

0.00

1.76 0.08

0.00

0.00

0.77 0.44

0.00

0.00 −0.27 0.79

0.00

0.03

0.14

0.04 0.97

0.00

−0.12 −1.36 0.17

−0.12

0.00 1.00

0.00

0.25

0.05

1.96 0.05

0.00

0.19 1.58

0.11

0.13

1.99 0.05

0.00

0.00 0.18

0.86

0.00

0.27 0.79

0.00

0.01

0.25 3.07

0.00

0.24

0.62 0.54

0.00

0.99

0.32

0.07 0.69

0.49

0.00

1.80 0.07

0.00

0.50

0.61

0.00 0.05

0.96

0.00

1.99 0.05

0.00

12 Shumen

0.29

0.77

0.00 0.16

0.87

0.00

0.40 0.69

0.00

13 Dobrich

−1.03 0.30

−0.05 −1.05 0.29

−0.04

0.71 0.48

0.00

14 Varna

0.65

0.04

0.00 1.00

0.00

15 Burgas

−1.26 0.21

−0.13 −1.36 0.17

−0.10 −0.48 0.63

0.00

16 Yambol

−0.27 0.79

0.00 −0.88 0.38

−0.06

0.90 0.37

0.00

17 Sliven

1.77

0.08

0.20 1.85

0.06

0.18

1.05 0.29

0.00

18 St.Zagora

0.93

0.35

0.08 0.48

0.63

0.05

1.98 0.05

0.00

0.52

0.13 2.12

p-value Slope

0.04 0.61

0.54

19 H askovo

2.13

0.03

0.18 1.94

0.05

0.14

0.86 0.39

0.00

20 Kardzhali

2.05

0.04

0.14 1.17

0.24

0.07

1.63 0.10

0.00

21 Smolyan

−1.93 0.05

−0.25 −0.29 0.77

0.00

22 Plovdiv

0.75

23 Pazardzhik

−2.09 0.04

24 Blagoevgrad 2.85

0.45 0.00

−0.24 −1.95 0.05 0.06 0.39

0.69

−0.16 −2.38 0.02

2.21 0.03

0.00

−0.15 −0.10 0.92

0.00

0.01

0.29

1.98 0.05

0.06

25 Kyustendil

−0.32 0.75

0.00 −0.38 0.71

0.00

0.22 0.82

0.00

26 Pernik

0.56

0.04 0.23

0.81

0.00

0.23 0.81

0.00

27 Sofia Distr

−0.90 0.37

−0.07 −1.38 0.17

−0.13

0.96 0.34

0.00

28 Sofia City

−0.18 0.86

0.00 −0.14 0.89

0.00

0.02 0.98

0.00

0.58

0.36 2.56

0.04

The seasonal trend analysis for each district is presented also in Table 1. The distribution and direction of the trends for warm half of the year completely coincides with those already described for the whole year. The difference is that during the warmer months

34

L. Bocheva and V. Pophristov

the statistically significant changes are in fewer administrative areas. Positive trend is statistically significant in 3 districts – in northern Bulgaria (Pleven and Targovishte) and 1 in southern Bulgaria (Blagoevgrad). Again only in district Pazardzhik the obtained seasonal negative trend is statistically significant (Table 1). During the cold half of the year for most of the country the trends are towards increasing the number of registered hailstorms, although the calculated trends are mostly statistically insignificant. According to applied Mann-Kendal test the observed trends are statistically significant for 6 districts (underlined or bolded in the Table 1), but the slope of trend can be estimated only for one of them due to the extreme rarity of events and the many zeros in the time series. So, only for Blagoevgrad district (bolded in Table 1) a statistically significant positive trend is evaluated (0.06 days/10 years).

Fig. 4. Mann-Kendal trend statistic (p-value) in color codes: positive trends - significant in red, insignificant in pink; negative trends – significant in dark blue, insignificant in light blue; without any trends – white.

The statistical estimate of the monthly variations of hail days for the period 1991– 2020 for each administrative district is even more difficult due to the reasons mentioned above. Due to the impossibility to adequately assess the slope of the trend, here we will comment only on the tendency for change in distribution of the days with hail precipitation in different month, which is presented on Fig. 4. During the cold half of

Recent Trends in Hail Precipitation for Administrative Districts of Bulgaria

35

the year the number of hail days generally increased after 1991. Significant positive tendencies are well pronounced for almost all districts in May and March as in May the increase of number of hail days is greater. On the other hand, in April and in the period June - September, the number of days with hail precipitation decreases in a larger number of districts, and the most significant is the decrease in August. It is significant for 3 districts from northern and 3 from southern part of the country.

4 Comparative Analysis of Annual Distribution of Hail Days During the Period 1961–2020 The annual regime of hail-storm days for 28 representative stations (one from each district) is compared for two periods: 1961–1990 and 1991–2020. The results of statistical estimations are presented in Table 2. Obviously for the majority of observed meteorological stations there are no statistically significant differences between two periods. However, for seven stations (bold in Table 2), a statistically significant increase in the frequency of hail days was recently observed. Five of them are in the regions of central (Pleven, Gabrovo and Veliko Tarnovo) and eastern (Targovishte and Razgrad) parts of Northern Bulgaria, and another two are in southeastern Bulgaria (Yambol and Haskovo). The largest increase is in Haskovo (about 2 times more days with hail compared to the previous period); then come the stations from Northeastern Bulgaria.

Fig. 5. Deviation of mean seasonal number of hailstorm days during the 1991–2020 period towards 1961–1990 period.

36

L. Bocheva and V. Pophristov

Table 2. Statistical comparison between two samples of average number of hail days in different districts of Bulgaria, using the Poisson distribution for the 1961–1990 (1) and 1991–2020 (2) data set (statistically significant trends are bolded). № of samples

1

1

2

2

Hail-storm days

Mean

Max

Mean

Max

№

Station

μ1

S1

Vidin

0.5

3

1.0

S2

Lom

0.7

2

0.6

S3

Vratsa

1.5

6

1.7

4

0.667

0.414

0.18

S4

Pleven

1.1

4

2.1

10

8.804

0.003

0.85

S5

Lovech

0.9

3

1.4

4

3.690

0.055

0.59

S6

Gabrovo

0.8

3

1.6

7

7.371

0.007

0.92

S7

Veliko Tarnovo

1.0

4

1.7

5

5.371

0.020

0.68

S8

Ruse

1.0

4

1.3

3

0.916

0.338

0.26

S9

Targovishte

0.9

5

1.8

5

9.175

0.002

1.00

S10

Razgrad

0.4

2

1.1

3

9.433

0.002

1.67

S11

Silistra

0.6

3

0.7

5

0.401

0.527

0.22

S12

Shumen

1.4

5

1.5

5

0.186

0.666

0.10

S13

Generl Toshevo

1.0

4

1.0

3

0.016

0.898

0.03

S14

Varna

0.6

3

1.0

4

2.908

0.088

0.63

S15

Burgas

0.6

2

0.9

3

1.734

0.188

0.47

S16

Yambol

1.5

6

2.4

6

5.383

0.020

0.54

S17

Sliven

1.9

4

2.1

6

0.412

0.521

0.13

S18

Stara Zagora

1.2

4

1.4

6

0.621

0.431

0.19

S19

Haskovo

0.8

3

2.6

6

28.625

0.000

2.12

S20

Kardzhali

1.5

4

2.0

9

2.192

0.139

0.34

S21

Smolyan

1.1

4

1.4

4

1.083

0.298

0.27

S22

Plovdiv

0.7

3

1.0

3

1.662

0.197

0.45

S23

Pazardzhik

0.8

3

0.7

3

0.091

0.763

−0.09

S24

Blagoevgrad

1.4

3

1.5

5

0.011

0.915

0.02

S25

Kyustendil

1.7

4

1.4

3

1.089

0.297

−0.20

S26

Pernik

1.3

5

1.3

5

0.013

0.909

0.03

S27

Dragoman

1.5

5

1.5

5

0.000

1.000

0.00

S28

Sofia

2.0

6

2.2

5

0.281

0.596

0.10

μ2

1,2

Tail

(µ2 − µ1 )/µ1

Probability χ2

p

%

4

3.810

0.051

0.81

3

0.401

0.527

−0.18

Recent Trends in Hail Precipitation for Administrative Districts of Bulgaria

37

Only in 3 stations there is a statistically insignificant decrease – Lom (about 18%), Pazardzhik (6%) and Kyustendil (20%). In all other districts in the second period (1991– 2020) there was a statistically insignificant increase in the average number of days with hail, but in Southwestern Bulgaria this increase is much smaller, almost negligible, than in other districts. These trends are well outlined in the seasonal differences in the average number of hail days presented in Figs. 5. The percentage deviation of the average seasonal number of days with hail for the period 1991–2020 compared to the period 1961–1990 (Fig. 5) increases faster in Northern Bulgaria (stations S1 to S14), especially during the cold half of the year. Convective storms, which are not typical for winter, are becoming more frequent at the resent years. The reason for this is not only the rise in winter temperatures, but also changes in atmospheric circulation over the country.

5 Conclusions For the period 1991–2020 the highest annual frequency of hail days is calculated for districts Sofia-city (1.69) and Yambol (1.54), followed by Haskovo, Targovishte, Lovech and Sofia-District where there are more than 1.3 days/year. The smallest number of days with hail precipitation are registered in Northeastern Bulgaria (Dobrich – 0.69) and along the Danube River (Vidin −0.76; Ruse – 0.81). Atypically small mean annual number of hail days during the last 30 years is observed in part of southwestern Bulgaria – about 0.85 days/year for districts Blagoevgrad and Kyustendil. The main part of hail events for the territory of all administrative districts in Bulgaria is registered during the warm half of the year (76%–96% of all cases). The share of hail days in the cold half of the year is over 10% for the regions with a ContinentalMediterranean climate as eastern Rhodopes (district Kardzhali – 24.1% of all), Strandzha and the Black Sea coast (Burgas – 18.8%; Yambol – 15.6%; Varna – 14.7% and Dobrich – 13%), as well as in the south westernmost region (Blagoevgrad – 12.7%). Statistical analysis show that the increase of number of days with hail precipitation is most significant in Blagoevgrad district (0.36 days/year), followed by Targovishte (0.25 days/year). The statistically significant increase is observed also in four more districts – two from North central Bulgaria (Pleven and VelikoTarnovo) and two from South Bulgaria, region of East Rodopes (Kardzhali and Haskovo). Statistically significand decreasing trend is calculated only for district Pazardzhik – on annual basis as well as for warm half of the year. During the cold half of the year for most of the country the trends are towards increasing the number of registered hail, although the calculated trends are mostly statistically insignificant. In the warm half of the year the trend is rather the opposite, especially for April and between June and September, where in more areas the number of hail days decreases, with the most significant decrease in August. The annual regime of hail-storm days for 28 representative stations (one from each district) is compared for two periods: 1961–1990 and 1991–2020. The results of statistical estimations show that for greater part of stations there are no statistically significant differences between two periods. Only in 3 stations there is a statistically insignificant decrease. In all other districts in the second period (1991–2020) there are a statistically insignificant increase in the average number of days with hail, but in Southwestern Bulgaria this increase is much smaller. The average seasonal number of days with hail for the

38

L. Bocheva and V. Pophristov

period 1991–2020 compared to the period 1961–1990 increases faster in stations from North Bulgaria, especially during the cold half of the year, which is mainly due to faster increase of winter maximal temperatures and connected with this increase of a number of winter convective storms. In seven stations a statistically significant increase in the frequency of hail days was observed and 5 of them are in central and east regions of North Bulgaria, while another 2 are in southeastern Bulgaria. The largest is increase in Haskovo (212%), followed by the stations from Northeastern Bulgaria (Targovishte – 167% and Razgrad – 100%).

References 1. Changnon S., Changnon D., Hilberg S.M.: Hailstorms across the nation. An atlas about hail and its damages. Illinois. State Water Survey (2009) 2. Leigh R.: Hail storm – one of the costliest natural hazards. Coastal Cities Natural Disaster Conference, AU, Sydney, 20–21 February 2007 (2007) 3. Mohr, S., Kunz, M.: Recent trends and variabilities of convective parameters relevant for hail events in Germany and Europe. Atmos. Res. 123, 211–228 (2013) 4. Nikolaides, K.A., et al.: The impact of hail storms in agricultural economy of Cyprus and their characteristics. Adv. Geosci. 17, 99–103 (2009) 5. Kunz, M., et al.: Characteristics, impacts and meteorological conditions. Quart. J. Roy. Meteor. Soc. 144, 231–250 (2013). https://doi.org/10.1002/qj.3197(2018) 6. Brown, T.M., Pogorzelski, W.H., Giammanco, I.M.: Evaluating hail damage using property insurance claims data. Wea. Climate Soc. 7, 197–210 (2015). https://doi.org/10.1175/WCASD-15-0011.1 7. Tuovinen, J.-P., Punkka, A.-J., Rauhala, J., Hohti, H., Schultz, D.: Climatology of severe hail in Finland: 1930–2006. Mon. Weat. Rev. 137, 2238–2249 (2009) 8. Dessens, J., Berthet, C., Sanchez, J.L.: A point haillfall classification based on hailpad measurements: the ANELFA scale. Atmos. Res. 83, 132–139 (2007) 9. Sioutas, M., Meaden, T., Webb, J.: Hail frequency, distribution and intensity in Northern Greece. Atmos. Res. 93, 526–533 (2009) 10. Berthet, C., Dessens, J., Sanchez, J.L.: Regional and yearly variations of hail frequency and intensity in France. Atmos. Res. 100, 391–400 (2011) 11. Fluck, E., Kunz, M., Geissbuehler, P., Ritz, S.P.: Radar-based assessment of hail frequency in Europe. Nat. Hazards Earth Syst. Sci. 21, 683–701 (2021). https://doi.org/10.5194/nhess21-683-2021 12. Punge, H.J., Bedka, K.M., Kunz, M., Reinbold, A.: Hail frequency estimation across Europe based on a combination of overshooting top detections and the ERA-INTERIM reanalysis. Atmos. Res. 198, 34–43 (2017). https://doi.org/10.1016/j.atmosres.2017.07.025 13. Mezher, R., Doyle, M., Barros, V.: Climatology of hail in Argentina. Atmos. Res. 114–115, 70–82 (2012) 14. Suwala, K., Bednorz, E.: Climatology of hail in Central Europe. Quaestiones Geographicae 32(3), 99–110 (2013) 15. Jelicґ, D., Megyeri, O. A., Malec.icґ, B., Belusicґ Vozila, A., Strelec Mahovicґ, N., Telisman Prtenjak, M.: Hail climatology along the northeastern Adriatic. Journal of Geophysical Research: Atmospheres 125, JD032749 (2020). https://doi.org/10.1029/2020JD032749 16. Punge, H.J., Bedka, K.M., Kunz, M., Werner, A.: A new physically based stochastic event catalog for hail in Europe. Nat. Hazards 73(3), 1625–1645 (2014). https://doi.org/10.1007/ s11069-014-1161-0

Recent Trends in Hail Precipitation for Administrative Districts of Bulgaria

39

17. Pusic, T., et al.: Future changes in European severe convection environments in a regional climate model ensemble. J. Climate 30, 6771–6794 (2017) 18. Bocheva, L., Dimitrova, Ts., Penchev, R., Gospodinov, I., Simeonov, P.: Severe convective supercells outbreak over western Bulgaria on July 8, 2014. IDӦJÁRÁS 122(2,) 177–202 (2018). ISSN: 03246329 19. Chipilski, H.G., Tsonevsky, I., Georgiev, S., Dimitrova, T., Bocheva, L., Wang, X.: Analysis of a case of supercellular convection over Bulgaria: observations and numerical simulations. Atmosphere 10(9), 486 (2019). (ISSN: 2073-4433) 20. Mann, B.: Nonparametric tests against trend. Econometrica 13, 245–259 (1945) 21. Sen, K.: Estimates of the regression coefficient based on Kendall’s tau. J. Am. Stat. Assoc. 63, 1379–1389 (1968) 22. StatSoft, Inc., STATISTICA (data analysis software system), version 6 (2004). www.statsoft. com 23. Gavrilov, M., Lazic, L., Milotinovic, M., Gavrilov, M.: Influence of hail suppression on the hail trend in Vojvodina. Serbia. Geographica Pannonica 15(2), 36–41 (2011) 24. Bocheva L., Simeonov P.: Spatio-temporal variability of hailstorms for Bulgaria during the period 1961–2010. In: 15th International Multidisciplinary Scientific GeoConference SGEM 2015, www.sgem.org, SGEM2015 Conference Proceedings, ISBN 978-619-7105-38-4 / ISSN 1314-2704, June 18–24, 2015, Book4, 1065–1072 (2015) 25. Pocakal, D., Vecenaj, Z., Stalec, J.: Hail characteristics of different regions in continental part of Croatia based on influence of orography. Atmos. Res. 93, 516–525 (2009) 26. WMO: Guidelines on the Calculation of Climate Normals. WMO 1203. Chairperson, Publications Board, ISBN 978-92-63-11203-3 (2017) 27. Stanchev, K.: On the frequency and territorial distribution of hail events in Bulgaria. Hydrol. Meteorol. 2, 15–23 (1964). (in Bulgarian)

Torrential Catchments from Belasitsa Mountain (SW Bulgaria) - Geological and Geomorphological Characteristics and Related Hazards Zornitsa Dotseva(B)

, Ianko Gerdjikov , and Dian Vangelov

Department of Geology, Paleontology and Fossil Fuels, Sofia University St. Kliment Ohridski, Sofia, Bulgaria [email protected]

Abstract. Floods, debris flows, and landslides are hazardous processes that often occurred in mountainous terrains over the world and could be highly destructive. In the last few decades, these events occur more often global, causing disasters as a result of climate changes and poor hazard and risk mitigation measures in the countries. On a regional scale, the territory of Bulgaria is not an isolated case and in the last decades, there is a higher intensity of hydro-meteorological and geological hazardous events. The main object of this paper is the study of the geological and geomorphological characteristics of torrential catchments located on the northern slopes of Belasitsa Mountain (SW Bulgaria), in the area of Petrich town. The area was affected by an intense rainfall event in December 2021 that triggered hazard processes like floods, landslides, debris floods, and mudflows that caused major damage to the road network, properties, and buildings, showing the urgent need to take hazard and risk assessment and mitigation measures. For the studied area, we made a susceptibility analysis that includes a description of geological and geomorphological conditions in the area, and GIS-based morphometric analysis of the torrential catchments. For further development of hazard assessment and mitigation decisions, Melton’s ruggedness ratio was calculated in order to distinguish the sediment transport type in the catchments. Keywords: Torrential catchments · GIS · Morphometric analysis

1 Introduction Steep torrential catchments in the mountainous areas could be impacted by fast-moving flows like clear water flows, debris flows, debris floods, and intensive erosion and landslides and their distribution poses a high risk to the population, buildings, and infrastructure in these areas. The trend is these events increase in feature due to climate change and urbanization of low-prone areas. Based on field investigations, geological, geomorphological, and hydrological characteristics of the drainage network, the assessment of susceptibility at the catchment scale is one of the main steps for prioritizing zones that can be affected and a key part of land management plans. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 40–50, 2023. https://doi.org/10.1007/978-3-031-26754-3_4

Torrential Catchments from Belasitsa Mountain

41

The main objective of the study is to provide data for geological and geomorphological conditions on 11 selected catchments in the northeastern part of the Belasitsa Mountain, near Petrich town and the villages of Belasitsa and Kolarovo which are susceptible to natural hazard processes. Short but intense rainfall on 11th December 2021 serves as a triggering source for floods, and activation of slope failures such as shallow landslides, as well as the formation of debris floods and debris flows that affect and damage road infrastructure, buildings, and other structures. GIS and remote sensing techniques are useful tools for evaluating the hydrological process in catchments [1]. The use of morphometric parameters to analyze drainage networks and catchments is critical in catchment management and planning. Important drainage network parameters like geometry, drainage texture and relief characteristics are all connected to the hydrological process in the catchments [2]. The Melton ratio [3] is one of the relief parameters that have high predictive power and is used to differentiate flood processes in catchments combined with field data for sediment deposition [4–6]. In this work, we use machine learning approaches in combination with morphometric characteristics and field surveys of torrential catchments and fans to assess susceptibility and hazards in the area.

2 Study Area The study catchments cover an area of 65.31 km2 and are situated on the northern slopes of the Belasitsa Mountain in southwestern Bulgaria, near the town of Petrich and Belasitsa and Kolarovo villages (Fig. 1). 2.1 Geology and Geomorphology Settings In the Neogene-Quaternary time, Southwestern Bulgaria experienced intense neotectonic movements that were associated with the tectonic events that occurred in Southeast Europe, including the collapse of the Late Alpine orogen, the Aegean back-arc extension and the crustal motions in the Pannonian and the Black Sea regions [7]. The Belasitsa Mountain is a linear horst that has been uplifted and extends in an east-west direction between two parallel faults - the Kerkini normal fault in Greece and the Podgorie normal fault in Bulgaria. The Podgorie normal fault is well geomorphologically expressed in the steep Mountain’s northern slopes with a displacement amplitude of up to 2.5 km [8–11] and is associated with the formation of the Strumeshnitsa graben also. Horst uplift has caused extensive erosion and a large amount of sediments were deposited, forming alluvial fans with a width of more than 250 m in some places, composed mainly of blocks of weathered metamorphic rocks combined with coarse gravel and sand in the highest parts of the fans, and smaller and finer particles at the lowest parts of the fans, from which many springs emerge [11–13]. The study area is mainly composed of Paleozoic metamorphic rocks from the Ograzhden metamorphic complex, including migmatites, metagranites, and meta-basic rocks, as well as Ng-Q sediments that fill the Strumeshnitsa graben [11] (Fig. 1B). Some river valleys are filled with Quaternary alluvial deposits.

42

Z. Dotseva et al.

Fig. 1. Location of the study area. A. Google Earth imagery of the area; B. Geological characteristics of the area (based on data by [11])

The area is characterized by mountainous relief, cut by deeply incised river valleys, steep slopes, and summits (Fig. 2A). The slope angle plays a major role in water flow which controls the water infiltration, erosion rate and surface run-off. The mean slope angle in the studied area is 17.46° and reaches up to 60°, with a sudden change in the gradient in places (Fig. 2 B). The slope aspect is with dominating NE orientation. 2.2 Climate, Vegetation, and Anthropogenic Influence The Mediterranean climate has a strong influence over Belasitsa Mountain [14]. Climate conditions are characterized by hot summer and short winter and are influenced by the warm Mediterranean air masses from the south, and the cooler continental air masses from the north. The mountain serves as a barrier to moist air masses, allowing the creation of conversion and heavy rainfall [15]. The annual mean temperature and precipitation are 13.9 Co and 676 mm, respectively. The precipitation rate varies with height and reaches 900 mm in the highest part of the Mountain, with summer minimums (nearly null) and autumn-winter maximums [15, 16]. From November through March, the water flow

Torrential Catchments from Belasitsa Mountain

43

Fig. 2. The research area’s elevation and slope characteristics. A. ASTER GDEM coupled with hill shaded relief; B. Distribution of slope angles

rises, with the lowest flow occurring in August. Extreme flood events are uncommon, but when they do occur, they have a huge impact. The vegetation in the area is represented by native broadleaf forests and forest plantations affected by anthropogenic activity [17, 18], such as tourism and logging.

3 Data and Methods 3.1 Data To better understand the process activated by the 11th December event, we collect field data and analyze crowdsourcing data like photo and video materials from the social

44

Z. Dotseva et al.

network, news sources, locals, and data supplied by Petrich Municipality and Petrich Forestry. Data about the geology, catchments closing sections, relief characteristics and deposition zones were obtained by geological maps in scale 1:50 000, topographic maps in scale 1:5000, satellite images (Google Earth) and images obtained by Unmanned Aerial Vehicle (UAV). Rainfall data was accessed from the Meteoblue platform. For the purposes of the study, we use freely available Digital Elevation Model (DEM) with a spatial resolution of 30 m cell size [19]. 3.2 Methodology Following the flood event that took place in December 2021, we carried out field surveys between late December 2021 and March 2022 to determine the geological and geomorphological conditions and to document the damages and characteristics of event-related sedimentary deposits. Although field identification of hydrogeomorphic processes is required, assessment procedures based on remote sensing data are often useful for regional studies. Remote sensing and GIS techniques are effective tools for morphological and hydrological analysis [1]. Furthermore, flood hazard evaluations are performed using great effectiveness by remote sensing and GIS approaches [20, 21, etc.]. ASTER DEM and the Spatial Analyst tools in ESRI ArcGIS 10.2 software were used to create maps for slope angle, flow direction and accumulation and automatic generation of a stream network and catchments for a thorough investigation and analysis. The morphometric parameters were evaluated with the help of established mathematical equations [22–28]. Morphometry is a quantitative assessment of the geometry of a river basin and relief. In most research morphometric parameters are divided into three classes: linear, aerial, and relief [29]. The linear features include characteristics such as stream order, stream number, and stream channel length, as well as their interactions with morphometric laws. The examination of perimeter, shape, and basin area are all examples of aerial characteristics. Relief of the catchments plays a vital role in drainage development, rate of surface runoff, infiltration into the sub-surface and channel flow. Under the relief aspects, other criteria such as basin relief, roughness number, and relief ratio are handled. We include in the analysis the Melton ratio (MR) [3] which is a slope parameter that provides a spatialized representation of relief roughness within the catchment and often is applied to distinguish different types of flows (flood, debris flood, debris flow) that perform in the catchments as a complemented method to field analysis of the sedimentary deposits [4, 6, 30, 31].

4 Results 4.1 December 2021 Interacting Hazardous Processes The events in December 2021 were preceded by heavy rainfall for five days in a row on the northern slopes of Belasitsa Mountain, which moistens the slopes and makes them unstable. The most intense rainfall occurred on the 11th of December, with precipitation rates exceeding 110 mm and acting as a triggering factor for landslides, erosion of the

Torrential Catchments from Belasitsa Mountain

45

river valleys, flooding, formation of debris floods and debris flows that caused great damage in the area of Petrich Town, Belasitsa and Kolarovo villages (Fig. 3) [32]. The most affected was the area of Buhoto District (the westernmost part of Petrich town) where several surges carrying mud, large boulders and trees destroyed? streets and damaged buildings. The main road between Petrich town and Belasitsa village was partially destroyed and covered by mud and clasts of different sizes. After a field survey and analysis of images obtained by UAV, we documented the reactivation of already registered and formation of a large number of new landslides, some of which disrupt the road Petrich to Belasitsa hut and a large number of forestry roads. Active sediment sources, such as landslides, produced a significant volume of sediment that enter the river valleys and affected the character of the flows. Thus, December 2021 case is a good example of a cascade of processes triggered by heavy rainfall. Various types of flows (debris flows, debris floods, mudflows, etc.) transported and deposited a huge amount (hundreds of m3 ) of clasts. The locations of the terminal deposits are mainly at places where channel gradient decrease and in several cases they coincide with the road-stream crossings. Sedimentary deposits in the study area are represented by very poorly to poorly sorted material, with a pebble to boulder size of the angular to sub-rounded clasts that spread on the fans and margins of the active river channels and fill the river valleys. Floods and debris floods are differentiated by sediment amount and clast orientation. Flood sediment concentrations are less than 20% by volume and are usually confined or localized in the channels. Debris floods contain sediment concentrations of 20 to 50% by volume and spread sediments on the fans, generating deposition forms like levees and lobes. 4.2 Morphometric Analysis – Linear, Aerial and Relief Parameters of the Catchments We analyzed 11 catchments listed as C-1 to C-11 (east to west). Catchments’ area varies from 1.04 to 19.40 km2 , where the smaller catchments didn’t reach the main watershed divide. The results from the morphometric analysis are shown in Tables 1 and 2 and discussed in the text below. In general, the catchments had a dendritic drainage pattern which complements the area’s relatively homogenous lithology [33]. The parallel drainage pattern shows a tiny shift, which might be connected to fault distribution. Linear Parameters The hierarchical relationship between stream segments is expressed by stream order (u) [26, 27]. The stream order has the highest value of 4th . The total number of streams (Nu) is 263 and their distribution in different orders shows a decrease in the number of streams with increasing order (Table 1). The stream length (Lu) was determined using Horton’s law [22] and defines the basin’s surface runoff characteristics. The total stream length is about 113 km, with decreasing length with increasing stream order which could indicate the catchments undergo homogeneous weathering and erosion [34]. The Mean stream length (Lsm) [22] is a dimensional parameter and reveals the characteristic size of drainage network components and its contributing basin surface [27]. The Lsm ranges between 0.39 and 0.51 and the total Lsm of the catchments is 2 km. The Stream length ratio (RI) has an important relationship with the surface flow and discharge [22].

46

Z. Dotseva et al.

Fig. 3. Damage to the infrastructure and buildings after the December 2021 flood event near Petrich Town and Belasitsa Village.

The values show decreasing in the RI with increasing stream order. The changes of RI from one order to another could indicate the youth stage of geomorphic development, or changes in slope and topography which control the discharge and different erosion stages [35, 36]. Bifurcation ratio (Rb) [24] ranges between 1.08 and 18.18 with a mean of 9.54 which indicates the fast increase and decrease of the floodwater surface and the variations from one order to the next order are attributed to the geological and lithological development of a catchment [27]. Table 1. Linear parameters of the drainage network Stream order u

Number of streams Nu

Stream length Lu (km)

Mean stream length Lsm (km)

Stream length ratio RI

Bifurcation ratio Rb

1

137

53.98

0.39

-

18.18

2

74

37.73

0.51

9.81

9.36

3

40

16.16

0.40

3.35

1.08

4

12

4.69

0.39

0.92

-

263

113

2

Mean 4.69

Mean 9.54

Total

Torrential Catchments from Belasitsa Mountain

47

The sinuosity index (SI) [25] is a parameter that reflects the geomorphic history and depends also on geological control like active tectonics [37]. The obtained values range from 0.68 to 0.98 and indicate straight to low sinuosity channels in the studied catchments (Table 2). Rho coefficient indicates the storage capacity during the flood defined as the ratio of RI and Rb [22]. The Rho coefficient of the catchments varies between 0.22 and 1.41. The catchments C-3, C-5, C-7, and C-10 have a value lower than 0.50 that indicating lower hydrologic storage during the flood. Relief Parameters Basin relief (R) [38] value ranges between 676 m and 1709 m, with a mean of 1202 m. The high value of R could indicate low infiltration and faster run-off conditions which leads to intensive erosion processes in the catchments. Relief ratio (Rr) [24] could be used as an indicator of the intensity of erosional processes and sedimentation rate [27]. The computed Rr values in the studied catchments range from 0.22 to 0.51 with a mean of 0.34. The ruggedness number (Rn) [39] value ranges between 1.06 and 3.28 with a mean of 2.11 where the high Rn values suggest an area susceptible to erosion, and structural complexity of the terrain with high relief and drainage density. The Melton ratio is a slope parameter that provides a specialized representation of relief ruggedness within the catchment [3]. High values correspond to basins with rough relief possibly affected by tectonic uplift. Furthermore, the Melton ratio and catchment length are often used for differentiation of the hydrogeomorphic processes – floods (bedload), debris floods and debris flows [4, 6]. The values range from 0.39 to 0.60 indicating debris flood development in 9 of the catchments. In C-9 and C-11, the higher values of the MR indicate the formation of debris flows. The obtained results indicate active and intensive erosional processes and high susceptibility to debris floods and debris flows instead of bedload. The predictive capability of MR was very good and corresponds well to the characteristics of the depositions from the December 2021 event and the geological and geomorphological conditions of the catchments. Aerial Parameters Drainage density (Dd) [22, 27, 40] values range from 1.45 to 2.07 km/km2 , indicating that the stream network in the catchments has moderate to high drainage density. If the Dd is less than 2 km/km2 , water may enter a channel fast, increasing the discharge [42]. The low drainage density is attributed to the presence of massive, resistant, and impermeable rocks with coarser drainage texture, low infiltration, and faster run-off. The high Dd values could be due to the density of fractures in the host rocks. After the Smith classification [41], the values of Dt [22] in the studied catchments indicate a very coarse to coarse drainage texture, which reflects a massive and resistant rock basement. Drainage intensity (DI) is defined as the ratio of Fs to Dd [28]. The high DI values for studied catchments show that Dd and Fs have a significant impact on the degree to which the surface has been affected by denudation. The Stream frequency index (Fs) [22] might be used to identify catchments where flooding could be expected [42]. The Fs value is high, ranging from 2.89 to 4.81 with an average of 3.88. These values correspond to higher surface run-off, impermeable subsurface material, steep gradient and medium to high relief. Values of the Dd, Fs and DI indicate that surface run-off could be quickly removed from the catchments.

48

Z. Dotseva et al.

The form factor (Ff), Elongation ratio (Re) and Circularity ratio (Rc) [22–24] are used to determine the catchment form and assess the flow intensity of a particular basin. In the analyzed catchments, the values indicate an elongated form which leads to low surface run-off, but the area is suitable for a large amount of run-off if significant rainfall is received, strong structural control on the drainage development and an area susceptible to high erosion and sediment yield [27, 35]. Table 2. Calculated linear, relief and aerial parameters of the catchments C.no/Index

C-1

C-2

C-3

C-4

C-5

C-6

C-7

C-8

C-9

C-10

C-11

SI

0.97

0.97

0.94

0.82

0.68

0.92

0.91

0.98

0.74

1.01

0.68

Rho

0.61

0.81

0.45

1.41

0.44

0.61

0.47

0.22

0.68

0.36

0.77

R

676

714

1709

734

1367

1065

1614

1581

1379

1598

785

Rr

0.26

0.24

0.22

0.32

0.40

0.32

0.31

0.32

0.49

0.31

0.51

Rn

1.40

1.29

2.75

1.06

2.16

2.12

2.61

2.75

2.21

3.28

1.52

MR

0.51

0.46

0.39

0.51

0.54

0.60

0.51

0.57

0.80

0.53

0.77

Dd

2.07

1.81

1.61

1.45

1.58

1.99

1.62

1.74

1.60

2.05

1.93

DI

1.94

2.11

2.69

2.38

2.81

1.45

2.05

1.97

2.11

2.34

2.49

Dt

1.06

1.22

2.97

1.08

2.31

1.09

2.25

1.91

1.15

3.03

0.95

Fs

4.02

3.81

4.33

3.45

4.45

2.89

3.32

3.43

3.38

4.80

4.81

Ff

0.25

0.25

0.29

0.26

0.24

0.24

0.29

0.32

0.20

0.35

0.20

Re

0.57

0.56

0.61

0.58

0.56

0.55

0.61

0.64

0.51

0.67

0.51

Rc

0.50

0.54

0.54

0.61

0.54

0.57

0.58

0.53

0.49

0.55

0.47

5 Conclusions Using GIS and remote sensing help to better understand topography features such as infiltration capacity, run-off, lithology, and relief, all of which impact the basin’s hydrological process. The obtained results are useful in understanding the studied catchment’s geology, geomorphology, hydrology, and structural properties. The northern slopes of Belasitsa Mountain are highly impacted by different natural hazards. In general, this is a result of the specific geological and geomorphological characteristics in the area – young relief and active tectonics in a combination with climate conditions. Because the bedrock is represented mainly of tectonically reworked metamorphic rocks landslides are frequent. As shown by the described events, heavy rainfall triggers a hazard process chain that ultimately causes severe infrastructure damage. Furthermore, the area is urbanized, and the population is exposed to high risk, and this requires more detailed research of the catchments that pose a risk, estimation of the rainfall thresholds that can trigger the processes, complete risk assessment and urgent mitigation measures.

Torrential Catchments from Belasitsa Mountain

49

Acknowledgements. This work has been carried out in the framework of the National Science Program “Environmental Protection and Reduction of Risks of Adverse Events and Natural Disasters”, approved by the Resolution of the Council of Ministers № 577/17.08.2018 and supported by the Ministry of Education and Science (MES) of Bulgaria (Agreement № D01-363/17.12.2020).

References 1. Waikar, M.L., Nilawar, A.P.: Morphometric analysis of a drainage basin using geographical information complex: a case study. Int. J. Multi. Current Res. 2, 179–184 (2014) 2. Rastogi, R.A., Sharma, T.C.: Quantitative analysis of drainage basin characteristics. Jour. Soil Water Conserv. India 26(1–4), 18–25 (1976) 3. Melton, M.A.: The geomorphic and paleoclimatic significance of alluvial deposits in southern Arizona. J. Geol. 73, 1–38 (1965) 4. Marchi, L., Pasuto, A., Tecca, P.R.: Flow processes on alluvial fans in the Eastern Italian Alps. Zeitschrift fur Geomorphologie 37, 447 (1993) 5. Marchi, L., Tecca, P.R.: Alluvial fans of the Eastern Italian Alps: morphometry and depositional processes. Geodin. Acta 8(1), 20–27 (1995) 6. Wilford, D.J., Sakals, M.E., Innes, J.L., Sidle, R.C., Bergerud, W.A.: Recognition of debris flow, debris flood and flood hazard through watershed morphometrics. Landslides 1, 61–66 (2004) 7. Burchfiel, B.C., Nakov, R., Tzankov, T.Z., Royden, L.H.: Cenozoic extension in Bulgaria and Northern Greece: the northern part of the Aegean extensional regime. In: Bozkurt, E., Winchester, J.A., Piper, J.D.A. (eds.) Tectonics and Magmatism in Turkey and the Surrounding Area. Geophysical Society. Special Publication, vol. 173, pp. 325–352 (2000) 8. Jaranoff, D.: La tectonique de la Bulgarie. Technica, Sofia, 283 p. (1960) 9. Jaranoff, D.: La neotectonique de la Bulgarie. Revue de Geographie Physique et de Geologie Dynamique 5(2), 75–83 (1963) 10. Zagorchev, I.: On the neotectonic movements in a part of SW Bulgaria. Bull. Geol. Inst., ser. Geotectonics 19, 141–152 (1970) 11. Klimov, I., Marinova, A., Marinova, R., Petrov, I., Milovanov, P., Ilieva, P.: Explanatory note for the geological map of the Republic of Bulgaria in scale 1:50 000, K-34-94-G (Makrievo), K-34-95-V-(Petrich), K-34-106-B, K-34-107-A (Belasitsa) map sheets. Ministry of environment and waters, Bulgarian National Geological Survey, 63 pp. (2009) 12. Galabov, Z.H., Ivanov, I., Penchev, P., Mishev, K., Nedelcheva, V.: Physical geography of Bulgaria. Narodna Prosveta, Sofia (1977) 13. Nikolov, V., Yordanova, M.: The mountains in Bulgaria. Academic Publishing House “Prof. Marin Drinov”. Sofia (2002) 14. Topliyski, D.: Climate of Bulgaria. Publishing house “Amstels”, Sofia (2006) 15. Koleva, E., Peneva, R.: Climate Directory - Rainfall in Bulgaria. BAS, Sofia (1990) 16. Velev, S.: The Climate of Bulgaria. Heron Press, Sofia (2010) 17. Bondev, I.: The Vegetation of Bulgaria. Map 1:600000 with explanatory text. St. Kliment Ohridski University Press, Sofia (1991) 18. Tonkov, S.: Sedimentation and local vegetation development of a reference site in SW Bulgaria. In: Lang, G., Schluchter, C.H. (eds.) Lake, Mire and River Environments, pp. 99–101. Balkema. Rotterdam (1988) 19. NASA/METI/AIST/Japan Spacesystems and U.S./Japan ASTER Science Team: ASTER Global Digital Elevation Model V003, distributed by NASA EOSDIS Land Processes DAAC (2019)

50

Z. Dotseva et al.

20. Yan, K., Di Baldassarre, G., Solomatine, D.P., Schumann, G.: A review of low-cost spaceborne data for flood modelling: topography, flood extent and water level. Hydrol. Process. 29, 3368–3387 (2015) 21. Samanta, S., Pal, D.K., Palsamanta, B.: Flood susceptibility analysis through remote sensing, GIS and frequency ratio model. Appl. Water Sci. 8(2), 1–14 (2018). https://doi.org/10.1007/ s13201-018-0710-1 22. Horton, R.E.: Erosional development of streams and their drainage basins; hydrophysical approach to quantitative morphology. J. Jpn. For. Soc. 56, 275–370 (1945) 23. Miller, V.C.: A Quantitative Geomorphic Study of Drainage Basin Characteristics in the Clinch Mountain Area, Virginia and Tennessee, Project NR 389-042. Technical report 3, Columbia University, Department of Geology, ONR, Geography Branch, New York (1953) 24. Schumm, S.A.: Evolution of drainage systems and slopes in Badlands at Perth Amboy. New Jersey. Geol. Soc. Am. Bull. 67, 597–646 (1956) 25. Schumm, S.A.: The Fluvial System, p. 338. John Wiley and Sons, New York (1977) 26. Strahler, A.N.: Quantitative analysis of watershed geomorphology. Trans. Am. Geophys. Union 38, 913–920 (1957) 27. Strahler, A.N.: Quantitative geomorphology of drainage basins and channel networks. In: Chow, V.T. (ed.) Handbook of Applied Hydrology, pp. 439–476. McGraw Hill, New York (1964) 28. Faniran, A.: The index of drainage intensity—a provisional new drainage factor. Aust. J. Sci. 31, 328–330 (1968) 29. Panda, S.P.: Quantitative analysis of Baitarani drainage basin using geographical information system. Vistas Geolog. Res. Spec. Publ. Geol. 14(1), 165–176 (2016) 30. Marchi, L., D’Agostino, V.: Estimation of debris-flow magnitude in the Eastern Italian Alps Earth Surf. Proc. Land. 29, 207–220 (2004) 31. Marchi, L., Fontana, G.D.: GIS morphometric indicators for the analysis of sediment dynamics in mountain basins. Environ. Geol. 48(2), 218–228 (2005) 32. Gerdjikov, I., Dotseva, Z., Dimitrov, S., Vangelov, D.: The natural disaster of December 2021 in the area of Petrich. SW Bulgaria. Geol. Min. Resour. 1, 2–7 (2022) 33. Mesa, L.M.: Morphometric analysis of a subtropical Andean basin (Tucumán, Argentina). Environ. Geol. 50, 1235–1242 (2006) 34. Nag, S.K., Chakraborty, S.: Influences of rock types and structures in the development of drainage network in hard rock area. J. Indian Soc. Remote Sens. 31(1), 25–35 (2003) 35. Singh, S., Singh, M.C.: Morphometric analysis of Kanhar river basin. Nat. Geogr. J. India 43, 31–34 (1997) 36. Sreedevi, P.D., Subrahmanyam, K., Ahmed, S.: Integrated approach for delineating potential zones to explore for groundwater in the Pageru River basin, Cuddapah District, Andhra Pradesh. India. Hydrogeol. J. 13, 534–545 (2005) 37. Burnett, A.W., Schumm, S.A.: Alluvial-river response to neotectonic deformation in Louisiana and Mississippi. Science 222, 49–50 (1983) 38. Hadley, R.F., Schumm, S.A.: Sediment sources and drainage basin characteristics in upper Cheyenne river basin. US Geolog. Surv. Water-Supply Paper 1531-B, 198 (1961) 39. Schumm, A.: Quaternary palaeohydrology. In: Wright, N.E., Frey, D.G. (eds.) The Quaternary of The United States, pp. 783–794. Princeton, Princeton University Press (1965) 40. Horton, R.E.: Drainage-basin characteristics. Eos Trans. Am. Geophys. Union 13, 350–361 (1932) 41. Smith, K.G.: Standards for grading texture of erosional topography. Am. J. Sci. 248, 655–668 (1954) 42. Howard, A.D.: Drainage analysis in geologic interpretation: a summation. Am. Assoc. Petroleum Geol. Bull. 51, 2246–2259 (1967)

Identification of Coastal Flooding Hotspots in a Large Bay Using an Index-Based Risk Assessment Approach Nataliya Andreeva(B) , Petya Eftimova , Nikolay Valchev , Bogdan Prodanov , Todor Lambev , and Lyubomir Dimitrov Institute of Oceanology, Bulgarian Academy of Sciences, 40 Parvi May Blvd., Varna, Bulgaria [email protected]

Abstract. The study aims to assess in probabilistic terms the magnitude of storminduced flooding hazard and vulnerability along Burgas regional coast (Bulgaria, western Black Sea) and to identify susceptible coastal sectors (hotspots). It employs the Coastal Risk Assessment Framework – a screening process that allows estimation of hazard intensities, extents and potential receptors’ exposure vulnerability within predefined sectors. Total water level, consisting of maximum wave run-up estimated using empirical models relevant for sandy beaches, artificial or rocky coastal slopes and storm surge level, is considered for calculation of coastal flooding hazard. Furthermore, hazard extents (flooded areas) are determined by two approaches to form the hazard indicator, while land use, social vulnerability, transport systems, utilities and business settings are considered as exposure indicators. Finally, the potential risk is assessed by combining the hazard and exposure indicators into a single index, thereby allowing rapid comparison of coastal sectors’ vulnerability. The study found that the main concentration of hotspots is in the town of Pomorie and its surroundings, while other candidates are scattered across the innermost Burgas bay. Keywords: Storms · Coastal flooding · Receptors · Hazard exposure · Vulnerability

1 Introduction Coastal areas are among the most valuable and highly dynamic environments that unite rich natural resources, large concentration of human population and a wide variety of socio-economic activities like urbanization, tourism & recreation, industrial production, port activities & shipping, agriculture, etc. At the same time, coastal zones are highly vulnerable to different natural phenomena, among which the storm-induced waves and surges can be considered as the main causes for coastal flooding and erosion capable of causing loss of human lives, damage to infrastructure, disruption of livelihood and enormous economic consequences. Historically and in the recent years, heavy storms that caused significant damages along the Bulgarian Black Sea coast have occurred in the late 1970s, early 1980s and during 2010–2015 [1–4]. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 51–65, 2023. https://doi.org/10.1007/978-3-031-26754-3_5

52

N. Andreeva et al.

The increase of occurrence and intensity of coastal flooding due to climate change [5] poses a necessity to assess contemporary and future coastal risks in support of coastal management, disaster risk reduction, decision-making and relevant policy & strategy implementation. According to the requirements of the EU Floods Directive 2007/60 [6], adopted in the national 2000 Water Act [7] by the Ministry of Environment and Water through its four River Basin Directorates, Bulgaria delivered a report on Preliminary Flood Hazard and Risk Assessment (PFHRA), flood hazard risk maps of Areas of Potential Significant Flood Risk (APSFR) and Flood Risk Management Plans (FRMP) concerning river basins and the Black Sea region for the period 2016–2021 [8]. Since each of these is subjected to revision and update every six years, for the next period 2022–2027 revised and updated so far are the methodology and report on PFHRA and flood hazard risk maps of APSFR, while updates on the FRMPs is still in progress [9]. The methodology applied at the national level includes risk assessment due to wave-induced hazard to designated coastal stretches highly susceptible to flooding. However, due to the large scale of assessment (along the entire Black Sea coast) the outcome attributes same level of risk for coastal stretches with different lengths and diverse geomorphic settings that in some cases extend for several tens of kilometers. Therefore, the present study delivers a more detailed estimation of flooding hazard, vulnerability and associated risks along a part of the Bulgarian Black Sea coast. An approach that offers such support is the Coastal Risk Assessment Framework (CRAF) that was developed within the RISC-KIT project [10]. The methodology was elaborated for identification of critical coastal areas (hotspots) considering the probability of occurrence of extreme storm events and has been applied and validated on ten European case studies [11] and worldwide [12]. The initial phase of the framework represents a screening process that delimits susceptible coastal sectors (≈1 km) on a regional scale (≈100 km) by assessing relevant hazard intensities, hazard extents and receptors’ exposure vulnerability for each sector producing assessment of the potential risk posed by coastal flooding [13]. The CRAF follows the index-based approach [14] and combines hazard and exposure indicators for different receptors (land use, population, transport, utilities and business) into a single index in order to compare coastal sectors. Hence, the aim of the present study is application of the CRAF and identification of ‘hotspots’ along Burgas regional coast in support of coastal managers, local & regional stakeholders, decision- and policy-makers. Successful application of the framework for Varna regional coast has been reported in [15], assessment of flooding & erosion hazards in [16] and exposure vulnerability of coastal receptors in [17]. Similar studies for Varna regional coast, which include indicator-based approach, were performed by [18, 19].

2 Study Site The study site with an approximate length of 90 km is located on the western Black Sea within the regional coast of Burgas and comprises the coastal areas of Pomorie and Burgas municipalities. To the north, it starts from Ravda, continues along the coast of the Pomorie peninsula and further on comprises almost the entire Burgas Bay down to the Naval Base Atiya at the south (Fig. 1).

Identification of Coastal Flooding Hotspots

53

Fig. 1. A schematic map of the study site showing: 1) types of slope at the shoreline within each predefined sector: yellow sectors – beaches backed or not by cliffs, orange sectors – rocky slopes (capes, etc.), grey sectors – technogenic type at the shoreline and green sectors - mixed type; 2) location of grid points for wave climate time series: star - wave point for peak-over-threshold (POT) analysis and red circles - wave forcing points.

Along Ravda and Aheloy the coast represents the alternation of low cliffs and narrow beaches locked between linear groynes. Further south, the coastal area is low and occupied by the Pomorie Lake, which is an ultra-saline lagoon separated from the sea by a natural sand spit reinforced with a dike along its entire length of 6.9 km. The lake is a natural deposit of healing mud rich in minerals and a source of sea salt extraction. Town of Pomorie occupies the entire peninsula, as a long urbanized beach with a large number of linear groynes is located at its eastern coast. The peninsula’s forehead is heavily man-modified area protected by several coastal structures, including local marina, fishing harbor and boat wharf. The southern shores from Pomorie to Sarafovo, being prone to erosion, abrasion and landslides are partially protected by dam. The city of Burgas is located in the innermost part of the Burgas Bay. To the west, the largest complex of coastal lakes (wetlands) with various salinities surrounds the city: Burgas Lake, Atanasovsko Lake, Mandra Lake and two smaller lakes Poda and Uzungeren, as the latter two are protected areas. Together with the Pomorie Lake they are included in the network of NATURA 2000 protected zones for their extraordinary biodiversity and substantial natural resources. The Atanasovsko Lake is separated from the sea by a narrow sand spit and having extremely high salinity is used for extraction of sea salt. The urbanized coastal area includes a long city beach backed by a promenade and a public park, as well as the premises of Port Burgas and related commerce, transport and

54

N. Andreeva et al.

industry zones. The coast from the Foros Bay to the Naval Base Atiya is presented by relatively narrow beaches and low-laying marshlands, which alternate with protruding high cliff rocky capes (Fig. 1). Municipality of Burgas is distinguished by its high level of industrialization, as among the main industries within the Burgas city (the fourth largest city in Bulgaria) could be pointed chemical & oil refining industry, shipbuilding, metalworking, fishfarming & fish-processing, tourism, etc., contributing significantly to local, regional and national economy [20]. For Pomorie municipality along with manufacturing and agriculture, the defining industry is tourism, as it successfully develops sea, spa, cultural and eco-tourism with national and international importance. Moreover, the town of Pomorie is classified as a secondary support center of micro-regional importance in the territory of Burgas region [21].

3 Data and Methods 3.1 Flooding Hazard By definition, coastal flooding is a combination of water levels presented by storm surges & high tides, and wave action producing wave-induced run-up. Due to small tidal variations in the Black Sea only storm surge plus wave run-up was considered as Total Water Level (TWL) parameter used for estimation of the flooding hazard intensity. To implement the methodology the study site’s coast was divided into 71 sectors with coastline lengths ranging from 0.5–1 km due to complexity of the coastline and variety of the geomorphic setting. Subsequently, for comparability the results for each sector were recalculated to correspond to a length of 1 km. Four main types of coastal sectors were recognized according to the morphological features of the coast near the shoreline and the degree of human influence– beach (31 sectors), rocky shore (11 sectors), manmodified shore (21 sectors) and seven sectors of mixed type (Fig. 1). The available very high-resolution topography data collected during photogrammetric surveys with an unmanned aerial vehicle system [22] consisted of digital surface models (resolution less than 20 cm/pix) and orthophoto mosaics (2.5 cm/pix). To determine the hazard intensities at least one cross-shore profile was chosen for each coastal sector and in case of more complex coastline a larger number of profiles was considered. Thus, the total set of 121 representative cross-shore profiles comprised 73 profiles in accumulative coastal stretches and 48 profiles for sectors with rocky shores and man-modified coast. The study followed an approach that takes into account the response of a system to floods from a group of selected individual extreme storm events [23], using longterm wave and sea level data to assess flood hazard parameters – maximum wave runup, water level and overflow over dunes, dikes, low shores, etc. Extreme storms were defined using wave hindcast data from WAM-SWAN wave model train covering a 69year period (1949–2018) [24]. Identification of individual storms was done by the peakover-threshold (POT) analysis of off-shore grid point times series using threshold values for significant weave height of 2 m and its exceedance for more than 18 h [15] – Fig. 1. This resulted in the selection of 200 storms, each represented by maxima of surge level, significant wave height, peak wave period, mean wave direction and storm duration for representative model grid points (Fig. 1). Storm surge data for the same period were

Identification of Coastal Flooding Hotspots

55

derived from tide gauge daily measurements at bay conditions (Burgas Bay) and at open coast. For sandy beach profiles, the maximum wave run-up was calculated by the empirical model of Holman [25] and on profiles representing rocky and man-modified sectors by the EurOTop formulation [26]. Both models have been tested and validated at the Bulgarian Black Sea coast in [16]. The resulting TWLs were subjected to the Extreme Value Analysis and best fitted to the Generalized Extreme Value distribution to produce flood hazard intensities associated with return period of 50 years. The areas of inundation (hazard extents) were determined using two approaches depending on the slope of the hinterland. For continuously rising slopes the bathtub approach was used [27], while for low-laying lands an overwash extent was calculated following [28]. As a flooding hazard indicator I FH was used the parameter Weighted TWL developed in [15], which considers the TWL value and the relevant inundated areas for each sector. It combines the contribution of each of these quantities to the overall hazard. 3.2 Exposure Vulnerability To evaluate the relative exposure of coastal receptors five types of exposure indicators were considered: Land Use, Population, Transport systems, Utilities and Business settings [13]. To reflect the exposure degree of each receptor’s type the resultant values of the indicators were ranked from 1 to 5: non-existent or very low (1), low (2), moderate (3), high (4) and very high (5) exposure to flooding hazard. The land use indicator I LU considers only the presence of receptors and not the vulnerability of land use classes. For each coastal sector within the limits of the inundated area, it takes into account the exposed surface and a weight (importance) value assigned to each land use class: n Si ∗ wi (1) ILU = i=1

where n - number of land use classes, S i - surface in m2 of each land use class and wi is weight (importance) value assigned to each land use class (from 1 to 10). Herein, as land use data was used the national Corine Land Cover (CLC) dataset version 2018 [29] with an initial scale of geo-data 1:100,000 and complemented with a refined coastline of the Burgas regional coast updated to the year 2020. For population the indicator I POP measures the relative exposure of various communities along the coast. It assesses the relative vulnerability of different areas to long-term health and financial recovery from an event, considering only the socio-economic capabilities of the population living in areas exposed to certain hazards [30]. To this end, six variables falling into four categories were selected: Financial deprivation (Unemployment, Non-car ownership, Overcrowding of households), Health (Long-term sick), Household structure (Single parent) and Age (Elderly 70+) given the present socioeconomic conditions and development in Bulgaria [31]. As input, the 2011 Census data on municipal level [32] were acquired and after statistical processing the indicator was calculated for each municipality as: n Ci ∗ wi (2) IPOP = i=1

56

N. Andreeva et al.

where n - number of variables, C i - average of the variables and wi is a weight of each variable. The next three indicators: I TS for Transport (roads and railroads), I UT for Utilities (e.g. water, electricity, telecom, emergency services, etc.) and I BS for Business were assessed by methodologies proposed in [30], which regard these receptors as networks, i.e. a set of elements interconnected to provide functions & outputs. For each indicator the ranks were assigned according to the importance value of the receptor within its network (Table 1). For the transport, geo-referenced data on terrestrial networks and the relative importance (road class) were obtained from Burgas and Pomorie municipal Master plans (scale 1:25,000) [33, 34] and online platform OpenStreetMap [35]. Evaluation for each coastal sector within the flooded area was done according to the national road classification [36]. Information on Utilities was compiled from the relevant Master plans, online platform Wikimapia [37], and field surveys, while data on assets’ relative importance – from desktop research. The indicator I UT was evaluated for each coastal sector within the flooded areas. Business assets’ location, their type, number and economic significance were acquired through field surveys, desktop research and Wikimapia & World Imagery [37, 38]. For each sector the indicator was evaluated based on businesses presence and significance of economic activities. Table 1. Ranking values and description of ranking rules for evaluation of transport, utilities and business indicators Rank

Description

(1) None or very low

No significant networks

(2) Low

Mainly local small networks

(3) Moderate

Presence of network with local or regional importance

(4) High

High density & multiple networks of local or regional importance

(5) Very high

High density & multiple networks of national or international importance

The overall exposure indicator I EXP unites the ranks of all exposure indicators as their geometric mean [13]: IEXP = [ILU ∗ IPOP ∗ ITS ∗ IUT ∗ IBS ]1/5

(3)

Ranking from 1 to 5 shows the degree of receptors’ exposure to flooding hazard. 3.3 Coastal Index Finally, in order to evaluate the potential flood risk for each coastal sector the Coastal Index (CI) combines hazard indicator and exposure vulnerability as [13]: CI = [IFH ∗ IEXP ]1/2

(4)

Identification of Coastal Flooding Hotspots

57

where I FH – hazard indicator and I EXP is the overall exposure indicator. A ranking of the results from 1 to 5 helps identification of the hotspots and comparison of the sectors along the coast.

4 Results and Discussion 4.1 Flooding Hazard Evaluation Using neither the TWL nor the inundated area allows for understanding the occurrence and correct evaluation of flooding hazard. The main reason lies in the fact that the former is strongly bound to the morphological conditions near the shoreline, while the latter is predetermined by the hinterland topography. Therefore, in this section, both elements are described in order to justify the hazard ranking and its contribution to the overall coastal index. The values of the estimated flooding hazard indicator I FH were ranked from 1 to 5 by the natural breaks data classification method [39] provided in the ArcGIS software. The assessment reveals that the most threatened by coastal flooding are 6 sectors: 18, 21, 26, 28, 29 and 34, while 15 sectors: 2, 6, 10, 12–13, 16–17, 20, 27, 30, 33, 35–36, 57 and 61 have a “high” susceptibility to flooding hazard (Fig. 2, Table 2). According to the coastal type, the above commented 21 sectors are mostly man-modified and natural beaches (15 sectors), five sectors are classified as technogenic coasts and one sector is of mixed type represented by small narrow beaches and coastal defense structures. In

Fig. 2. Flooding hazard indicator along the coastal areas of Burgas & Pomorie municipalities for 50-year return period

58

N. Andreeva et al.

terms of spatial distribution, they form two main clusters of flood-prone areas, while several sectors are scattered across the study site. The first cluster comprises the southern part of low-laying Pomorie sand spit, the urbanized beach of Pomorie town (sectors 10, 12, 13), the southern coast of Pomorie town fortified by a dam (sectors 16, 17), Pomorie south beach (sector 18) and sectors located to the west of cape Krotiriya (sectors 20, 21). The shore is exposed to storms from the eastern half and the beaches are dissipative to intermediate with relatively mild beach face slopes (0.04 ÷ 0.20). The estimated TWLs ranged between 2.9 m and 4.1 m (average value of 3.4 m) with flooded areas varying from 35,000 m2 to 82,200 m2 (average value of 51,100 m2 ). The second cluster is located in the inner Burgas Bay. It includes the beaches along Sarafovo (sectors 26–28), the coastal strip protected by a dam (sectors 29–30), the southern end of the sand spit at the Atanasovsko Lake (sector 33) and the urbanized beaches of Burgas city (sectors 34–36). The shore is moderately exposed to eastern storms and the beaches are dissipative with mild beach face slopes (0.03 ÷ 0.11). The estimated TWLs are between 2.6 m and 4.4 m (3.5 m being an average value), as the flooded areas are calculated to be between 39,600 m2 and 70,600 m2 with an average value of 52,500 m2 . For both spatial clusters, the considered hazard intensities were able to cause significant inundations reaching beyond the beach rear lines and overtopping the coastal protection structures to flood the low-laying portions of the coast. The flooded areas calculated for the first cluster have wider range in comparison to the ones relevant to the second cluster, since the type of coast is more diverse, but according to the average values, the second cluster is slightly more threaten by the flooding hazard. The rest of the sectors are located across the site and include sectors 2, 6, 57 and 61 (Fig. 2). The calculated TWL for sector 2, located near Aheloy, is quite high – 6.3 m, but since the sector is presented by a narrow beach (beach face slope of 0.22) backed by a cliff the flooded area is limited to 24,450 m2 . Sector 6 occupies the most northern part of Pomorie sand spit and is represented by a mild-sloped beach (beach face slope of 0.12). The total water level for this sector is 4.3 m and the flooded area equals 30,600 m2 . Sector 57 is located within Kraymorie residential district and its coast is moderately exposed to NE waves. Kraymorie beach is dissipative with very mild slopes (0.06 ÷ 0.07). It is a man-modified beach due to the presence of a fishing port at the southeastern sector’s end. Herein, the calculated TWLs are the lowest (2.70 ÷ 2.89 m) among the considered group of 21 sectors, but the inundated area is still quite large. It can be attributed to the assessment method that does not account for the protective role of the marina, so flooded area is being overestimated (49,600 m2 ). The last sector (sector 61) is part of the Chengene Skele protected site and is located in a more sheltered area, with northeastern exposure. It is represented by a dissipative beach (slope 0.07) backed by low-laying wetlands. In this case, the estimated TWL of 3.4 m results in a large inundation area (56,700 m2 ). 4.2 Exposure Evaluation The CLC classes [40] for the indicator I LU were identified within the 500 m hinterland strip. Some of the classes were merged to facilitate the evaluation resulting in 12 land use classes. Higher importance (weight) values were assigned to classes related to human occupation and activity. The land use classes and their weight values are: 1)

Identification of Coastal Flooding Hotspots

59

Discontinuous urban fabric (10), 2) Port areas (9), 3) Industrial or Commercial units (8), 4) Harbors & Marinas (7), 5) Green urban areas (6), 6) Sport & Leisure facilities (5), 7) Agriculture (4), 8) Nature conservation & Wetlands (4), 9) Coastal protection (4), 10) Beaches & Dunes & Sands (3), 11) Forest & Vegetated areas (2), 12) Bare rocks (1). After calculation, the indicator’s values were ranked from 1 to 5 by the natural breaks method. Results on the relative proportion (%) of each class affected by inundation state that the most affected are “Beaches” (37.99%), “Nature conservation & Wetlands”(17.10%), “Discontinuous urban fabric” (7.08), “Port areas” (6.71%), “Coastal protection” (6.45%) and “Industrial or commercial Units” (4.85%). To much lower extent are affected “Harbors & Marinas” (1.87%), “Forest & Vegetated areas” (0.71%) and “Sport & Leisure facilities” (0.16%), while “Agriculture” and “Green urban” areas suffer no inundation. For statistical reliability, the population indicator was assessed for all 32 municipalities of the coastal districts of Dobrich, Varna and Burgas. Ranks from 1 to 5 were evaluated by the method of natural breaks. This approach does not permit the assignment of a particular value for each coastal sector, so a single value was attributed for all sectors within a given municipality. The results show that the population of Pomorie has low (rank 2) social vulnerability, while for the municipality of Burgas the exposure is very low (rank 1). Since, the census information dated back to 2011 the results were supported by the most recent information on the socio-economic development of Bulgarian districts [41], classifying the Burgas district to have “Good” socio-economic conditions at present. To estimate the Transport exposure the national road classes were ranked according to the rules in Table 1 and values varied between 1 (local roads) to 5 (national highways and railroads). Results show predominance of exposure levels corresponding to ‘very low’ and ‘low’ ranks. Higher ranks (3 & 4) were attributed to sectors holding the presence of Burgas Salt works (sector 32) and port activities (sectors 40–44, 66). The highest rank (5) was given to a single sector (21) holding a stretch of the European route E 87, which is part of the United Nations international E-road network. It performs an integrating function for all the municipalities along the Black Sea coast and provides interregional connections between the northern and southern parts of the coast. The assessment for Utilities indicator shows the same tendency, but with lower ranks, i.e. predominance of low ranks (1 & 2) across the study site, moderate exposure for sectors related to salt extraction (sectors 12, 31), port premises (sectors 38, 40–43, 67) and residential areas of town of Pomorie (sectors 16, 18) – Fig. 1. “High” utilities’ exposure is detected only for sector 17 in the town of Pomorie and sector 66 located at port Rosenets. Three common types of coastal businesses were identified for the study site: 1) beach frontage urban area – Pomorie town and Burgas city, 2) port and related commercial & industrial zones – ports Burgas & Rosenets and their surroundings, and 3) coastal harbor (w/o marina) and related urban area – Aheloy, Pomorie, Sarafovo, Kraymorie, Chengene Skele. Moderate to very high ranks were given to sectors, representing businesses with regional to international importance. Finally, all five exposure indicators were combined into the Overall exposure indicator I EXP by Eq. 3, as the ranks were distributed by the method of natural breaks. According to results (Fig. 3, Table 2) moderately exposed are 9 sectors, as ‘high’ are ranked 16 sectors and only 5 sectors have the highest vulnerability to hazard. All these

60

N. Andreeva et al.

Fig. 3. Overall exposure indicator along the coastal areas of Burgas & Pomorie municipalities for 50-year return period flooding hazard

sectors are associated to: 1) seaside tourism at Aheloy, Pomorie, Sarafovo, Burgas and Kraymorie (sectors 3, 11–14, 20, 21, 32, 33, 57–58, 64), 2) port & industry & commerce of Burgas Salt works, port Burgas, shipbuilding & ship repair activities and port Rosenets (sectors 31, 38, 40–46, 48, 49, 66, 67) and residential quarters of Pomorie (sectors 15–18). Analysis suggests that in sectors relevant to tourism the indicator I EXP is predominantly controlled by the presence of various business receptors and affected land use. On the other hand, for sectors with presence of port & industry facilities in the Foros bay and the Chengene Skele bay, the influence of transport & utility assets increases. For the residential areas of Pomorie the exposure depends primarily on the land use, followed by business and utility assets. 4.3 Coastal Index Evaluation Coastal index values obtained by Eq. 4 were ranked again by the method of natural breaks. Results show that 16 sectors could be considered as hotspot having ‘high’ (13 sectors) and ‘very high’ ranks (3 sectors) – (Fig. 4, Table 2). Similar to the flood hazard indicator, these sectors are related to seaside recreation areas at Pomorie town beaches (sectors 12–14, 18, 20, 21), Burgas city beaches (sector 33) and Kraymorie (sector 57), as well as the residential areas in Pomorie town (sectors 16, 17) and industry & commercial areas at Burgas Salt works (sectors 31, 32), Burgas Shipbuilding & Ship repair factories (sector 44) and port Rosenets oil tanking premises (sectors 66, 67).

Identification of Coastal Flooding Hotspots

61

Fig. 4. Coastal indices along the coastal areas of Burgas & Pomorie municipalities for 50-year return period flooding hazard Table 2. Ranking of sectors for the flood hazard indicator, the overall exposure indicator and the coastal index, 50 year return period flood hazard Rank

Hazard indicator I FH (sectors)

Exposure indicator I EXP (sectors)

Coastal index CI (sectors)

(1) None-existent or very low

1, 15, 38, 40–43, 46–49, 52, 54,63, 68, 71

2, 7–9, 19, 22–25, 30, 39, 47, 50–52, 54–56, 59, 60, 62, 63, 65, 68, 70, 71

1, 19, 22–25, 50–52, 54–56, 59, 60, 63, 65, 68, 70, 71

(2) Low

19, 22–25, 37, 39, 45, 50, 51, 53, 55, 56, 59, 60, 64, 65, 70

1, 4–6, 10, 26, 27, 29, 34–37, 53, 61, 69

2, 7–9, 15, 30, 37–43, 46–49, 53, 62

(3) Moderate

3–5, 7–9, 11, 14, 31, 32, 44, 58, 62, 66, 67, 69

3, 11, 15, 28, 33, 45, 46, 58, 64

3–6, 10, 11, 26, 27, 29, 34–36, 45, 58, 61, 64,69

(4) High

2, 6, 10, 12, 13, 16, 17, 20, 27, 30, 33, 35, 36, 57, 61

12–14, 16, 20–21, 31, 32, 38, 40–43, 48, 49, 57

12–14, 16, 20, 28, 31–33, 44, 57, 66, 67

(5) Very high

18, 21, 26, 28, 29, 34

17, 18, 44, 66, 67

17, 18, 21

62

N. Andreeva et al.

The indicated connection would suggest that CI is mainly hazard controlled. However, the more detailed analysis reveals that for hotspot allocated as sea recreation areas and residential quarters the CI is almost equally driven by flooding hazard and assets’ exposure, while for sectors, representing port & industry & commerce the exposure in terms of business presence, affected land use, and to lower extent utilities and transport, have stronger influence. Thereby, the most vulnerable to storm-induced flood hazard are the coastal areas of the entire town of Pomorie and those located within the limits of Burgas city and its surroundings, experiencing substantial anthropogenic pressure and supporting significantly local, regional and national economy.

5 Conclusions The presented study demonstrated the application of a screening, but comprehensive index-based methodology that combines estimation of flooding hazard and exposure vulnerability of socio-economic elements to assess the potential risk of coastal storminduced inundation over the regional scale of the municipalities of Pomorie and Burgas. The estimated flooding hazard indicator integrated hazard intensities and relevant flooding extents corresponding to a return period of 50 years by means of the weighted TWL parameter. The assessment revealed that most vulnerable to flood hazard coastal areas are concentrated into two main clusters. The first one is located along the coast of Pomorie peninsula, enclosing the town beaches, parts of the dam-protected coast at Pomorie south and the coast to the west of cape Krotiriya. The second cluster includes the beaches and the adjacent coastal areas from Sarafovo to the north of Burgas city. The rest of the threatened areas are scattered across the study site including beaches located along Aheloy, Kraymorie, Pomorie sand spit, and one being part of the protected site “Chengene Skele”. Evaluation of the relative exposure of different impact categories: land use, population, transport, utilities and businesses, merged into the overall exposure indicator, showed that it is mainly controlled by different business settings and land use, and to a lower extent by transport and utility assets located along the entire coast of Pomorie town, its southern surroundings and in the innermost Burgas bay. The ultimate product of these indicators – the Coastal Index enabled identification of hotspots, as well as their clustering along the study site. The outcome of the index estimation indicated that hotspots are allocated at seaside recreational and residential areas where the CI is driven equally by flood hazard and assets exposure, and port & industry & commerce zones with increased exposure with regard to business presence and affected land use. On a larger scale, the hotspots are spatially concentrated along the coastal stretches of the entire Pomorie town and the innermost Burgas Bay. Thus, the aforementioned coastal areas along the Burgas region are considered the most vulnerable and at risk to coastal flooding. According to the available flood hazard risk maps of Areas of Potential Significant Flood Risk along the Burgas regional coast for both 2016–2021 and 2022–2027 [6, 7] the coast from Ravda to the innermost Burgas bay is susceptible to the highest level of risk to storm-induced flooding. Thus, the risk assessment methodology applied in the present study has demonstrated very good ability to determine the coastal stretches vulnerable to flooding hazard and at the same time, it gives a more detailed evaluation on the level of the coastal risk on a regional scale.

Identification of Coastal Flooding Hotspots

63

Acknowledgements. This work has been carried out in the framework of the National Science Program “Environmental Protection and Reduction of Risks of Adverse Events and Natural Disasters”, approved by the Resolution of the Council of Ministers № 577/17.08.2018 and supported by the Ministry of Education and Science (MES) of Bulgaria (Agreement № D01-279/03.12.2021). We also would like to thank Mr. Louis Haquette, a student from ENSTA-Bretagne, France, for his contribution to the geospatial data processing in the frame of his internship at the Institute of Oceanology-BAS, Varna, Bulgaria.

References 1. Andreeva, N., Valchev, N., Trifonova, E., Eftimova, P., Kirilova, D., Georgieva, M.: Literary review of historical storm events in the western Black Sea. In: Proceedings of Union of Scientists – Varna, “Marine sciences”, pp. 105–112 (2011). ISSN 1314-3379. (in Bulgarian) 2. Trifonova, E., Valchev, N., Andreeva, N., Eftimova, P.: Reconstruction of severe storms in the Western Black Sea and assessment of their impact on the coast. In: Proceedings of the 10th International Conference on Marine Sciences and Technologies – Black Sea 2010, pp. 254– 260, Varna technical and Scientific Unions, Varna (2010). ISSN 1314-0957 3. Andreeva, N.K., Kiresiewa, Z.K., Valchev, N.N., Eftimova, P.T.: Cultural insights into coastal risks and climate change resilience of a society ‘in transition.’ In: Martinez, G. (ed.) Culture and Climate Resilience. PSCRS, pp. 15–43. Springer, Cham (2021). https://doi.org/10.1007/ 978-3-030-58403-0_3 4. Galabov, V., Chervenkov, H.: Study of the western Black Sea storms with a focus on the storms caused by cyclones of North African origin. Pure Appl. Geophys. 175(11), 3779–3799 (2018). https://doi.org/10.1007/s00024-018-1844-7 5. Vousdoukas, M.I., Voukouvalas, E., Annunziato, A., Giardino, A., Feyen, L.: Projections of extreme storm surge levels along Europe. Clim. Dyn. 47(9–10), 3171–3190 (2016). https:// doi.org/10.1007/s00382-016-3019-5 6. Directive 2007/60/EC of the European Parliament and of the Council on the assessment and management of flood risks. http://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:320 07L0060. Accessed 28 Aug 2022 7. The 2000 Water Act. https://www.mrrb.bg/en/act-on-waters/. Accessed 28 Aug 2022 8. Flood Hazard and Risk Assessment, 2016-2021, Black Sea Basin Directorate Homepage. https://www.bsbd.org/bg/ndex_bg_2934486.html. Accessed 29 Aug 2022 9. Flood Hazard and Risk Assessment, 2022–2027, Black Sea Basin Directorate Homepage. https://www.bsbd.org/bg/index_bg_965885.html. Accessed 29 Aug 2022 10. Van Dongeren, A., et al.: Introduction to RISC-KIT: resilience-increasing strategies for coasts. J. Coast. Eng. 134, 2–9 (2018). https://doi.org/10.1016/j.coastaleng.2017.10.007 11. Viavattene, C., Jiménez, J.A., Ferreira, O., Priest, S., Owen, D., McCall, R.: Selecting coastal hotspots to storm impacts at the regional scale: a coastal risk assessment framework. Coast. Eng. 134, 33–47 (2018). https://doi.org/10.1016/j.coastaleng.2017.09.002 12. Pazini, K.C., Bonetti, J., da Silva, P.G., Klein, A.H.F.: Spotting areas critical to storm waves and surge impacts on coasts with data scarcity: a case study in Santa Catarina. Brazil. Natural Hazards (2022). https://doi.org/10.1007/s11069-022-05275-1 13. Ferreira, O., et al.: CRAF Phase 1, a framework to identify coastal hotspots to storm impacts. In: 3rd European Conference on Flood Risk Management FLOODrisk 2016, E3S Web of Conferences, vol. 7, p. 10002 (2016). https://doi.org/10.1051/e3sconf/20160711008 14. Balica, S.F., Wright, N.G., van der Meulen, F.: A flood vulnerability index for coastal cities and its use in assessing climate change impacts. Nat. Hazards 64, 73–105 (2012). https://doi. org/10.1007/s11069-012-0234-1

64

N. Andreeva et al.

15. Valchev, N., Andreeva, N., Eftimova, P., Prodanov, B., Kotsev, I.: Assessment of vulnerability to storm induced flood hazard along diverse coastline settings. In: 3rd European Conference on Flood Risk Management FLOODrisk 2016, 17–21 October 2016, Lyon, France. E3S Web of Conferences, vol. 7, p. 10002 (2016). https://doi.org/10.1051/e3sconf/20160710002 16. Eftimova, P., Valchev, N., Andreeva, N., Prodanov, B., Dimitrov, L.: Calculation of maximum wave run-up and erosion at Varna regional coast (Western Black Sea) using empirical models. Coast. Eng. Proc. 1(35), Management 17 (2017). https://doi.org/10.9753/icce.v35.manage ment.17 17. Andreeva, N., Valchev, N., Prodanov, B., Eftimova, P., Kotsev, I., Dimitrov, L.: Assessment of coastal receptors’ vulnerability to flood hazard along Varna regional coast. Coast. Eng. Proc. 1(35), Management 8 (2017). https://doi.org/10.9753/icce.v35.management.8 18. Valchev, N., Andreeva, N., Eftimova, P., Trifonova, E.: Prototype of early warning system for coastal storm hazard (Bulgarian Black Sea coast). Comptes rendus de l’Academie bulgare des Sciences 67(7), 971–978 (2014) 19. Villatoro, M., et al.: An approach to assess flooding and erosion risk for open beaches in a changing climate. Coast. Eng. 87, 50–76 (2014). https://doi.org/10.1016/j.coastaleng.2013. 11.009 20. Development Plan of Burgas municipality 2014–2020. https://www.burgas.bg/bg/obshtinskiplan-za-razvitie-2014-2020/. Accessed 14 June 2022 21. Development Plan of Pomorie municipality 2014–2020. https://www.pomorie.bg/48364/ obshtinski-plan-za-razvitie-na-obshtina-pomorie-za-perioda-2014-2020g/. Accessed 14 June 2021 22. Prodanov, B., Kotsev, I., Lambev, T., Bekova, R.: Unmanned aerial vehicles for surveying the Bulgarian black sea coast. Comptes rendus de l’Academie Bulgarie des Sciences 73(5), 666–672 (2020). https://doi.org/10.7546/CRABS.2020.05.09 23. Garrity, N.J., Battalio, R., Hawkes, P.J., Roupe, D.: Evaluation of the event and response approaches to estimate the 100-year coastal flood for Pacific coast sheltered waters. Coast. Eng. Proc., 1651–1663 (2006). https://doi.org/10.1142/9789812709554_0140 24. Valchev, N., Trifonova, E., Andreeva, N.: Past and recent trends in the western Black Sea storminess. Nat. Hazard. 12, 1–17 (2012). https://doi.org/10.5194/nhess-12-961-2012 25. Holman, R.A.: Extreme value statistics for wave run-up on a natural beach. Coast. Eng. 9(6), 527–544 (1986). https://doi.org/10.1016/0378-3839(86)90002-5 26. Pullen, T., Alsop, N.W.H., Bruce, T., Kortenhaus, A., Schüttrumpf, H., van der Meer, J.W.: EurOtop. Wave overtopping of sea defenses and related structures: assessment manual. In: Samuels, P. (ed.) Flood Risk Management: Research and Practice, 193 p. Taylor & Francis Group, London (2007) 27. Orton, P., Vinogradov, S., Blumberg, A., Georgas, N.: Hydrodynamic mapping of future coastal flood hazards for New York City, Revised final project report, Stevens Institute of Technology, 36 p. (2014) 28. Donnelly, C.: Coastal Overwash: Processes and Modelling. Ph.D. thesis, University of Lund, 53 p. (2008) 29. Executive Environment Agency (EEA) of Bulgaria Homepage. http://eea.government.bg/bg/ projects/korine-14/kzp-danni-clc-data. Accessed 14 June 2022 30. Viavattene, C., Micou, P., Owen, D., Priest, S.J., Parker, D.J.: Library of coastal vulnerability indicators: guidance document, EU FP7 603458 research project RISC-KIT - Deliverable No: D.2.2, 136 p. (2015) 31. Bulgaria 2030: Analysis of the socio-economic development of the country after its accession in the EU, p. 105 (2019). Ministry of Finances Homepage. https://www.minfin.bg/upload/ 41549/Bulgaria+2030+analiz.pdf. (in Bulgarian). Accessed 29 May 2022 32. National Statistical Institute of Bulgaria (NSI) Homepage. https://nsi.bg. Accessed 14 June 2022

Identification of Coastal Flooding Hotspots

65

33. Burgas municipality Homepage. https://www.burgas.bg/bg/obsht-ustroystven-plan-na-gr-bur gas-s-negovite-kvartali-i-tehnite-zemlishta/. Accessed 29 Aug 2022 34. Pomorie municipality Homepage. https://www.pomorie.bg/62506/obsht-ustroystven-planna-obshtina-pomorie-v-predvaritelen-proekt/. Accessed 29 Aug 2022 35. OpenStreetMap (OSM) Homepage. https://www.openstreetmap.org. Accessed 26 Aug 2022 36. Ordinance No 2 for planning and design of transport and communication systems in urbanized territories. Official Gazette 86 (2004) (in Bulgarian). https://www.lex.bg. Accessed 29 May 2022 37. Wikimapia Homepage. https://wikimapia.org. Accessed 28 Aug 2022 38. ESRI, “World imagery” [Layer], Scale Not Given, “World imagery”, 15 Aug 2022. https:// www.arcgis.com/home/item.html?id=10df2279f9684e4a9f6a7f08febac2a9. Accessed 26 Aug 2022 39. Jenks, G.F.: The Data Model Concept in Statistical Mapping, pp. 186–190. International Yearbook of Cartography VII. Bertelsmann Verlag, Germany (1967) 40. European Environment Agency (EEA) Homepage. https://www.eea.europa.eu/publications/ COR0-landcover. Accessed 14 June 2022 41. Slavova, Z., Nikolov, A., Ganev, P.: Regional profiles – indicators of development, Institute for market economics, Report on the project “Regional Profiles: Indicators of Development” financed by the America for Bulgaria Foundation, 152 p. (2021). https://www.regionalprofile s.bg. Accessed 14 June 2022

Evaluation of the Nakamura Vulnerability Index of a Cast-in-Situ Reinforced-Concrete Building from Ambient Noise Records Emil Oynakov , Radan Ivanov(B) , Irena Aleksandrova, Jordan Milkov, and Mariya Popova National Institute of Geophysics, Geodesy and Geography - Bulgarian Academy of Sciences, Acad. G. Bonchev Street, Bl. 3, 1113 Sofia, Bulgaria [email protected]

Abstract. The deformation characteristics and the dynamic amplification of the response of a building subjected to an earthquake can be estimated from the dynamic characteristics of the structure. In the present study, the fundamental natural frequency of a structure, a key factor influencing the damage after earthquakes, is examined. The investigated structure is the building of the National Institute of Geophysics, Geodesy and Geography (NIGGG), to the Bulgarian Academy of Sciences, which is a five-floor, cast-in-situ reinforced concrete (RC) building. The campus area adjacent to the building was also studied. Two indexes, related to the performance during earthquake were calculated using ambient vibration data; (i) the vulnerability index of the studied structure (K b ), and (ii) the coefficient of vulnerability of ground of the adjacent free field (K g ). The vulnerability index was calculated at ten locations on each floor. The maximum value computed in the present study was 81, indicating the location where the structure experiences the highest dynamic amplification and is hence the weakest and most likely to fail first during earthquake. The maximum possible interstorey drift ratio due to a design earthquake peak ground acceleration (PGA) was also determined. All measurements were performed using four ETNA 2 mobile accelerometers with a recording frequency of 100 samples/second, and recording time at each point 20 min. Keywords: Vulnerability index · Ambient noise · H/V spectral ratio

1 Introduction Bulgaria is located in the central part of the Balkan Peninsula, which is the most active seismic node in the Alpo-Himalayan earthquake zone in Europe. The capital Sofia is located in the Sofia basin, surrounded by mountains, within the territorial boundaries of the Sofia seismic zone. To the North, the zone stretches to the Trans-Balkan deep fault and the ridge of Stara Planina; to the east and southeast–to the fault junction near the village of Sturgel and the line between the Etropole Balkans and Ihtiman Sredna Gora; to the south–to the Pernik fault, the Iskar dam and the southern © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 66–76, 2023. https://doi.org/10.1007/978-3-031-26754-3_6

Evaluation of the Nakamura Vulnerability Index

67

slopes of Vitosha and Plana planina; to the west and northwest–to the border with Serbia. According to Bonchev, [1], this territory overlaps with the Sofia Srednogorie. Historical data on seismic events in the Sofia Valley, although inaccurate, have existed since the 15th and 16th centuries. In 1450 there was a devastating earthquake with an intensity (I0 ) of about VIII on the scale of Medvedev-Sponhoyer-Karnik-64 (MSK), during which occurred the first destruction of the church of St. Sofia and also of the small church of St. Marina, located in the yard of the former diocese (today’s Independence Square). Cursory data for an earthquake of 1557 with an intensity of about VII, [2, 3] are also available. The data from 1818, when two very strong earthquakes of intensity IX and VII occurred, are more reliable. The first happened on March 23 (April 4) at 12 o’clock. It was a devastating earthquake. The second one was on September 7/19 and was also quite strong. Its intensity is estimated at about VII. It is obvious that this was a strong aftershock of the earthquake of March 23/April 4, [2]. The earthquake of 1858 is considered to be the strongest event in the Sofia area. Its intensity is estimated as IX and its magnitude as 6.6. The aftershock series of this earthquake lasted about 4 months (felt), [2]. The strongest earthquake in the 20th century, realized in the vicinity of Sofia, was the event of 1917 with magnitude Mw = 5.7 (I0 = VII–VIII MSK). The earthquake caused significant damage to the buildings in the city, [3]. Almost a century later, on May 22, 2012, an earthquake with Mw = 5.6 shook the Sofia area, with an epicenter 25 km southwest of Sofia, between the cities of Pernik and Radomir. Moderate to severe damage was observed in the city of Sofia and its surroundings. The elements of risk today for a city like Sofia are far more significant than in earlier times and losses due to earthquakes may increase. The extent of the damage is largely due to the proximity of the epicenters to the big city. Modern earthquakes can be more damaging to the city than in the past, even if their magnitudes do not exceed the highest observed so far, due to increased population growth and building stock, the concentration of interconnected administrative, commercial, industrial and residential buildings infrastructure networks (underground and above ground), etc. Probable side effects are also important, especially in densely populated areas. All of this requires measures to reduce the risk to reasonable limits. The investigation of buildings of social or historical significance by non-destructive geophysical method is considered largely beneficial as a preparation for renovation and strengthening works. Some of the widely used geophysical methods are those relying on the analysis and interpretation of ambient noise, and more specifically methods based on the horizontal-to-vertical spectral ratios (HVSR), initially introduced by Nakamura, [4]. The underlying idea is that any change in the stiffness of the structure would influence dynamic response parameters (modal frequencies, amplitudes) of the system, [5]. In fact, the variations in the dynamic response may be due to variations in the actual boundary conditions between ground and structure, alterations in the initial design assumptions such as refurbishment, changes in the mechanical properties of the material due to a history of seismic actions, or other environmental impacts. The method has been used successfully for assessing the vulnerability of architectural heritage (the Colosseum, [4] the leaning tower of Pisa), as well as multiple administrative buildings and other

68

E. Oynakov et al.

structures (viaducts, embankments) in Japan and a number of countries of the European Union. For the needs of the method, the observation and registration of natural and anthropogenic microseisms, which usually requires complex and often expensive organization, is accomplished in the form of a set of triaxial measurements at reference points with the measuring equipment being moved across the points, without the need of temporal synchronization of the records. In the present application of the method, the vulnerability indexes of each point of a number of vertical profiles of the building are computed (the variations of the indexes with height, while identifying the reasons behind the variations). In this way it is possible to identify the weakest, i.e. the most vulnerable locations in the structure and inform all interested parties for taking further action for strengthening the specific locations. The method usually produces about 10% to 15% standard deviation which is typical for seismological studies, but with increasing the number of investigated points this value is reduced. The cost of implementing the method is several times smaller than the cost of standard methods, such as the forced-vibration test, [6] or the free-vibration test [7]. The Nakamura technique for buildings as a medium for wave propagation (which differs substantially from ground) considers the investigated building as an equivalent linear system, in which the records of the horizontal channels are assumed to be the output signal, whereas the records of the vertical channel are the input ones. A “black-box” problem is hence solved–the characteristics of the vibrating object are extracted from the input and output signals. The HVSR provides the transfer function or the spectral response of the “black box”. The Nakamura technique provides as an approximation a direct relation between the actual characteristic of the transfer function and the amplitudes of the likely shear deformations of the building. It follows that the resonance peaks of a spectral characteristic are due to shear deformations occurring at the particular frequencies, and the building is most likely to be damaged through vibrating at these frequencies, [8] while subjected to external dynamic actions.

2 Methodology and Measurements The investigated structure is a five-story, cast-in-situ RC building raised in 1980, which houses the NIGGG, to the Bulgarian Academy of Sciences. The main building, where the measurements were performed, is rectangular in plan with dimensions 67.4 m by 14.8 m, and story height of 3.05 m. The vertical loads are transferred by 0.18 m thick slab, girders and columns, while the lateral stability is provided by shear walls in both directions. The building is supported on strip foundations. The building belongs to one of the most common structural types for residential and office buildings in Bulgaria. Therefore, the obtained results can be useful for general insights for this structural type, as well as for justification of further research. For the purposes of this study a total of 60 measurements were made at 10 points situated on each story of the building, Fig. 1 (top), with each group of 6 points with the same location in plan, but situated on different stories constituting one vertical profile. In this way it was possible to analyze “rays” of the seismic wave propagating vertically. By measuring at 10 points relatively uniformly distributed at each story, we assure adequate horizontal coverage. Deviations of the location of points on the same vertical

Evaluation of the Nakamura Vulnerability Index

69

profile were sometimes necessary due non-uniform distribution of the walls on different stories; these deviations were kept to a minimum and never exceeded 2–3 m, which enables the changes in the nature of the seismic signal from the bottom to the top stories to be properly observed.

Fig. 1. a). (top). Map of the measurement points; the projections of the vertical profiles are designated by 1, 2, …, 10 (the stories are designated, bottom to top as follows: −1, 0, 1, 2, 3, 4). The orientation and deviation of the instrument X-axis from North (40°) is shown at the top-right, and the instrument Y-axis is parallel to the long side of the building. (bottom) 3D model of the structure with a possible deformed shape of the vertical axes of the columns during a seismic event (grey–at rest, red–deformed during shaking).

All measurements are made by four mobile accelerometers ETNA 2, with recording frequency of 100 samples/second, and recording time at each point 20 min. The Xaxis of the instruments was always kept perpendicular to the long side of the building, while the Y-axis was parallel to the long side. The processing and analysis of the obtained records was done by the program GEOPSY v.3.4.1 (http://www.geopsy.org/). Taking into account the frequency characteristics of the recording equipment and the recommendation for the practical implementation of the HVSR method by the SESAME project, [9] the signals were band-pass filtered in the frequency range 0.1–20 Hz, which enhances the performance of the method. For the investigated points, the graphs of the H/V spectral ratio were plotted against the frequency, after which the Nakamura vulnerability index, which is a dimensionless coefficient, was computed for each of them, [10].

70

E. Oynakov et al.

The vulnerability index is one of the most important parameters for evaluation of potential damage in urban regions caused by earthquake, [12]. It can be used to describe the robustness of a building against earthquake, [4, 13]. It is generally accepted that the vulnerability of a structure during an earthquake can be evaluated by the drift angle, which in turn is related to the effective acceleration of the input ground motion at base level (in cm/s2 ). In the following equation α is the part of the acceleration which contributes to setting the building in motion given input acceleration a: α = e.a

(1)

where e is the coefficient of effectiveness of the ground motion for the particular structure. The deformation characteristics and the degree of amplification of the motion of the structure can be assessed by the dynamic characteristics of the structure. The fundamental natural frequency which strongly influences the damage during earthquake is considered. The displacement δ i of the ith story is computed from the fundamental natural frequency F and the amplitude Ai of the ith story, which in turn is taken from the graph of the spectral ratio of the horizontal component of the record to its vertical component, Fig. 2, as follows: δi =

Ai α

(2)

(2π F)2

Hence, the drift angle γi of the ith story is computed as: γi =

δi+1 − δi (Ai+1 − Ai )hi (Ai+1 − Ai )hi α = = eKbi a; Kbi = 10000 2 hi (2π F) (2π F)2

(3)

where (Ai+1 –Ai ) is the difference in the amplifications of the two adjacent stories, hi is the story height in meters, and F is the fundamental frequency of the structure. As Eq. 3 suggests, the drift angle γi for each story is computed as the vulnerability index K bi , multiplied by the free-field peak ground acceleration a in cm/s2 and the coefficient of effectiveness e of the earthquake motion. The average value of the vulnerability index av K b is computed for each structure as follows: avKb =

A H (2π F)2

10000

(4)

where, A is the amplification of the top story, and H is the height of the building in meters. Analyzing the wave field by the Nakamura method the resonant frequencies can be determined experimentally and compared to those computed during the design of the building. Furthermore, by analyzing the phase differences of the wave fronts, the torsional vibrations, as well as the stress and strain developing in the structural members can be determined. Summarizing the obtained dynamic parameters and their variations in time it is possible to draw conclusions about the overall physical condition of the investigated object, [5].

Evaluation of the Nakamura Vulnerability Index

71

Fig. 2. Model of n-story structure. δ i– displacement; hi –height of the ith story; α1 , αi , αi+1 – effective horizontal acceleration of the 1st , ith and i + 1th story; A1 , Ai , Ai+1 –the amplitude of the spectral H/V ratio of the 1st , ith and i + 1th story; H–height of the building.

The coefficient of vulnerability of ground, K g is computed by data obtained from a previous investigation, [14] by the relation, [5]: Kg =

A2 F

(5)

where A is the maximum value of the amplification; F is the frequency corresponding to it, both obtained from the graphs of the HVSR.

3 Results The analysis of the obtained data was performed by the methodology proposed by Okada, [15] and Peterson, [16], which allows a distinction to be made between microseismic motions generated by natural or anthropogenic sources in the vicinity of the investigated building. From the spectral curves obtained for the area surrounding the building, Fig. 3, the coefficient of vulnerability of ground is computed with average value from all 19 measurement points of K g = 1.12.

Fig. 3. Graphs of the HVSR for the 19 measurement points in the vicinity of the building.

72

E. Oynakov et al.

Fig. 4. Spectral characteristics of the HVSR for all vertical profiles. Each curve represents a single story.

3.05

6.1

9.15

12.2

15.25

Story 0

Story 1

Story 2

Story 3

Story 4

2.25

2.12

2.76

2.05

0.91

1.22

0

3.05

6.1

9.15

12.2

15.25

Profile number

Story −1

Story 0

Story 1

Story 2

Story 3

Story 4

16.74

1/700

av K b

γ

2.25

2.91

2.25

2.25

2.10

1.25

F (Hz)

1/1000

6

H (m)

γ

13.03

0

Story −1

1

F (Hz)

av K b

H (m)

Profile number

5.18

3.81

3.41

2.14

1.62

1.42

Ai

4.42

2.23

2.05

1.88

1.53

1.65

Ai

2

74.15

1/1300

10.33

2.89

2.26

2.20

−5.71

11.70

2.15

1.60

5.49

F (Hz)

7

1/700

17.49

2.21

2.21

2.12

2.21

2.26

1.98

F (Hz)

18.18

8.65

9.30

K b6 (m/s2 )−1

63.31

35.41

1.70

2.89

30.54

5.52

K b1 (m/s2 )−1

5.28

4.74

3.47

3.43

1.63

0.78

Ai

5.20

4.08

2.96

2.43

1.57

1.28

Ai

45.73

7.61

18.99

0.63

50.99

2.03

K b7 (m/s2 )−1

77.45

16.66

18.01

7.94

12.28

5.40

K b2 (m/s2 )−1

1/800

16.55

2.25

2.12

2.21

2.25

2.12

1.10

F (Hz)

8

1/1700

7.59

2.91

2.85

2.85

2.08

1.60

1.13

F (Hz)

3

5.12

4.33

3.50

2.93

1.71

1.58

Ai

3.92

3.73

2.08

1.91

1.46

1.36

Ai

73.28

12.76

12.29

8.15

19.72

7.61

K b8 (m/s2 )−1

33.59

1.69

14.72

2.79

12.98

5.44

K b3 (m/s2 )−1

Table 1. The computed vulnerability indexes. 4

1/1300

10.43

2.91

2.97

2.91

2.91

2.08

1.28

F (Hz)

9

1/700

18.31

2.25

2.97

2.97

1.88

1.28

1.13

F (Hz)

5.39

4.45

2.97

2.24

1.51

1.36

Ai

5.66

2.69

2.00

1.60

1.45

1.16

Ai

46.18

7.76

12.64

6.27

12.30

6.51

K b9 (m/s2 )−1

81.08

24.43

5.73

8.10

6.60

16.88

K b4 (m/s2 )−1

5

1/1300

10.03

2.85

2.91

2.91

2.85

2.69

1.34

F (Hz)

10

1/2300

5.68

2.97

2.26

2.25

2.21

2.20

1.48

F (Hz)

4.98

4.47

3.51

2.44

1.33

1.35

Ai

3.05

4.25

4.32

2.29

1.66

1.45

Ai

44.41

4.39

8.20

9.62

11.12

0.80

K b10 (m/s2 )−1

25.13

17.07

1.00

30.29

9.33

7.08

K b5 (m/s2 )−1

Evaluation of the Nakamura Vulnerability Index 73

74

E. Oynakov et al.

In all vertical profiles the HVSR values increase as the story level increases, which confirms the validity of the method, i.e. the clear dependence of the computed spectral characteristics on the structure of the investigated building. In Fig. 4 frequencies smaller than 1 Hz correspond to the vibrations of the underlying soil layer, and frequencies larger than 1 Hz to the actual structural vibrations. The graphs of the HVSR for each point of all vertical profiles are shown in Fig. 4. The computed vulnerability indexes are shown in Table 1. On the fourth story of vertical profile 4 (Fig. 4; Table 1) the amplification factor reaches its maximum of 5.66, and this is probably the most vulnerable location in the building. The values of the vulnerability index are put on the plan of the building, Fig. 5, and interpolated using the method „Natural Neighbor”, which assigns the largest weight to the neighboring points. The maximum value of the vulnerability index is 81.08 exactly at the above mentioned point with maximum amplification factor (vertical profile 4; 4th story). It is within the range (25 to 100) reported for rigid frame viaducts, [10], and the range (40 to 90) for the low heights of the 200 m tall Folkart towers in Turkey, [11]. Again, we suppose this is the weakest location in the building. For each vertical profile the average vulnerability indexes, av K b are computed as well. Using their values, the maximum drift angles during a seismic event have been determined.

Fig. 5. Distribution of the vulnerability index along the stories of the building.

Evaluation of the Nakamura Vulnerability Index

75

If we assume a maximum PGA of about 0.32g (~314 cm/s2 ), which is close the current maximum design PGA for Bulgaria, and further assume that the effective seismic motion is about 1/3 of the maximum (~105 cm/s2 ), the maximum drift angle would be between 1/700 (vertical profiles 3 and 4) and 1/2300 (vertical profile 5). On the other hand, for most structures significant damage or collapse occurs at drift angles of around 1/100–1/200. Comparing these values indicates that the structure would withstand an earthquake of the considered size. For the investigated region, the fundamental frequency of the underlying sediment soil layers (average depth 160 m to 240 m), is evaluated in a previous study, [14] using the HVSR method. The computed values are hereby compared to the fundamental period of the analyzed building. The results show that the building is raised on ground with fundamental frequency of about 1.4–1.6 Hz (period 0.63–0.71 s.), Fig. 3, which is double the fundamental frequency of the building 2.2–2.9 Hz (period 0.34–0.45 s.), Table 1. According to this comparison we can conclude that the resonance of ground and structure is avoided for their fundamental periods, and it is the fundamental mode which usually has the largest contribution to the dynamic response.

4 Conclusions Since recording of microseismic noise in buildings is easy, non-destructive and fast, it would be useful to do it on a regular basis in buildings situated in seismically active regions. Processing of the accumulated data would then facilitate the assessment of the dynamic behavior and the design of strengthening and other mitigation measures for each monitored building. The data can also help to reveal structural flaws undetectable by the usual visual surveys, and their growth in time. The data can also aid the design of regular maintenance activities and structural alterations, assuring they will not be detrimental to the structural performance during future earthquakes. In the present investigation the vulnerability indexes, K b and the coefficient of vulnerability K g of the building and the surrounding area respectively were assessed. Their values were computed from recorded microseismic noise, allowing to evaluate the vulnerability of the building structure throughout its volume. The maximum value of the vulnerability index was 81.08 at the location where the HVSR also reached its maximum. This location is identified as possibly the most vulnerable one in the whole building. The distribution of the vulnerability index along the stories indicates that the top story as a whole is much more susceptible to damage than the rest of the building. The maximum interstory drift angles were determined for a scenario of maximum plausible PGA, and the building is not likely to be damaged by this scenario. Acknowledgement. This work has been carried out in the framework of the National Science Program “Environmental Protection and Reduction of Risks of Adverse Events and Natural Disasters”, approved by the Resolution of the Council of Ministers № 577/17.08.2018 and supported by the Ministry of Education and Science (MES) of Bulgaria (Agreement № D01–279/03.12.2021); Presentation of this work is supported by Contract No D01–404/18.12.2020 (Project “National Geoinformation Center (NGIC)” financed by the National Roadmap for Scientific Infrastructure 2017–2023.

76

E. Oynakov et al.

References 1. Bonchev, E.: Problems of Bulgarian Geotectonics, 203 p. Tehnika, Sofia (1971) 2. Watzof, S.: Earthquakes in Bulgaria during XIX century, 93 p. Central Meteorological Station, Imprimerie de I’Etat, Sofia (1902). (in Bulgarian and French) 3. Kirov, K.: Contribution to study of earthquakes in Sofia region. Ann. Main Dep. Geol. Mining Res. 5, 407–440 (1952) 4. Nakamura, Y., Gurler, E.D., Saita, J., Rovelli, A., Donati, S.: Vulnerability investigation of Roman Colosseum using microtremor. In: Proceeding of 12th WCEE, pp. 1–8, Auckland, New Zealand (2000) 5. Nakamura, Y.: A method for dynamic characteristic estimation of subsurface using microtremor on the ground surface. Q. Rep. Railway Tech. Res. Inst. 30(1), 25–33 (1989) 6. Shabbir, F., Omenzetter, P.: Forced vibration testing of a thirteen storey concrete building. In: Proceedings of New Zealand Society for Earthquake Engineering Annual Conference, pp. 1–8. Wairakei, New Zealand (2008) 7. Butterworth, J., Lee, J.H., Davidson, B.: Experimental determination of modal damping from full scale testing. In: Proceedings of 13 WCEE, paper No. 310, Vancouver, Canada (2004) 8. Okamoto, S.: Seismic Stability of Engineering Structures, 342 p. Strojizdat, Moscow (1980) 9. Bard, P.Y., et al: SESAME, guidelines for the implementation of the H/V spectral ratio technique on ambient vibrations measurements, processing and interpretation. 62 p. SESAME European research project WP12–Deliverable D23.12 European Commission–Research General Directorate Project No. EVG1-CT-2000-00026 (2004) 10. Nakamura, Y.: Seismic vulnerability indices for ground and structures using microtremor. In: Proceedings of World Congress on Railway Research, Florence, Italy (1997) 11. Timur, E., Ozicer, S., Sari, C., Uyanik, O.: Determination of buildings period and vulnerability index using microtremor measurements. In: Proceedings of 8th Congress of the Balkan Geophysical Society, Chania, Greece (2015) 12. Hosseini, K., Hosseini, M., Jafari, M.K., Hosseinioon, S.: Recognition of vulnerable urban fabrics in earthquake zones: a case study of the Tehran metropolitan area. J Seismol. Earthq. Eng. (JSEE) 10(4), 175–187 (2009) 13. Sato, T., Nakamura, Y., Saita, J.: The change of dynamic characteristics using microtremor. In: Proceedings of the 14th WCEE, Beijing, China (2008) 14. Popova, M., Oynakov, E., Solakov, D., Aleksandrova, I., Dragomirov, D., Ivanov, R.: Experimental Evaluation of the Dynamical Parameters f0 and A0 Using the Method HVSR. In: Proceedings of 11th Congress of the Balkan Geophysical Society. European Association of Geoscientists & Engineers, Bucharest, Romania (2021) 15. Okada, H.: The microtremor survey method. Geophysical Monograph Series No. 12. Society of Exploration Geophysicists, USA (2003) 16. Peterson, J.: Observations and modeling of seismic background noise. USGS Open-File Report 93-322. USGS, USA (1993)

Historical Earthquakes and Tsunami Waves in the Sea of Marmara: Review and Modelling Lyuba Dimova(B)

and Reneta Raykova

Department of Meteorology and Geophysics, Faculty of Physics, Sofia University “St. Kliment Ohridski”, 5 James Bourchier, 1164 Sofia, Bulgaria [email protected]

Abstract. Marmara Sea is located near the collision border between the Eurasia and the Arab plates, where a major continental strike slip fault called the North Anatolian Fault (NAF) runs south of the Black Sea and pass through the Anatolian blocks with a dextral motion. Seismic ruptures along the NAF are one of the main causes of strong earthquakes and along with submarine landslides, tsunami waves generated. Nevertheless, the seismic activity is concentrated along the Izmit Fault in the east and the Ganos Fault in the west of Marmara Sea. In this study we review the historical earthquakes generated tsunamis and model some of them numerically. One of the largest earthquakes of the XXth century in the Balkans, occurred on the active Ganos Fault on August 9th , 1912 (Ms 7.4). The so-called S, arköy-Mürefte earthquake produced tsunami waves which were observed all the way to the eastern coast of Marmara [1]. Another strong earthquake caused tsunami near Istanbul and Izmit Bay happened on August 17th , 1999 (Mw 7.4). The model UBO-TSUFD [2], used in the computations, is based on the nonlinear shallow-water equations of Navier-Stocks coupled with bathymetry and topography data. The initial tsunami conditions are calculated via the Okada model [3], based on known focal mechanisms. Results of the simulations are presented as maximum water elevation field, tsunami travel time maps and synthetic mareograms in certain points. Keywords: Earthquakes · Tsunami · Numerical modelling · Marmara Sea

1 Introduction The Marmara Sea region is characterized by high seismicity [4]. Historical earthquakes are documented in many papers starting with the BC period up to now [5–8]. An assessment of the XXth century earthquakes is given in many different studies [9–11]. The main tectonic features and geodynamics are discussed in brief. Marmara Sea is located in a complex area where different stress regimes occur. The Anatolian Peninsula is subjected to three large fault systems: the Hellenic Arc, the North Anatolian Fault (NAF) and the Eastern Anatolian Fault. The boundary between the Eurasia and Arab plates is formed by the dextral NAF, which splay into three different branches: the northern part crosses the Marmara Sea and the Gulf of Saros; the southern part elongates © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 77–87, 2023. https://doi.org/10.1007/978-3-031-26754-3_7

78

L. Dimova and R. Raykova

to the city of Bursa; the middle part follows the coastline of the Marmara Sea from Gemlik Bay to Bandirma and ends into the Aegean Sea [12]. The seismicity, together with GPS data and geological surveys suggest that the northern branch of the fault zone is more active than the other two southern parts [13, 14].

2 Historical Earthquakes and Tsunami Waves The Table 1 gives a summary of the strongest earthquakes between 1800 and 2022 using several open access catalogues and papers [15, 16]. It must be noted that in XIXth century the seismicity is unusual inactive, with only two earthquakes of M > 7.0, while in XVIIIth and XXth centuries there have been four and five respectively [16]. Table 1. Strong earthquakes in the broad Marmara Sea region 1800–2022. Date

Magnitude

Lat. [°]

Lon. [°]

6 Oct. 1841

6.1

40.8

29.0

28 Feb. 1855

7.1

40.1

28.6

11 Apr. 1855

6.3

40.2

28.9

21 Aug. 1859

6.8

40.3

26.1

19 Apr. 1878

5.9

40.7

30.2

9 Feb. 1893

6.9

40.5

26.2

27 Jul. 1893

5.9

40.7

27.0

10 Jul. 1894

7.3

40.7

29.6

9 Aug. 1912

7.4

40.6

27.0

4 Jan. 1935

6.4

40.7

27.5

26 May 1957

7.4

40.7

31.1

18 Sep. 1963

6.2

40.7

29.2

17 Aug. 1999

7.4

40.8

30.0

24 May 2014

6.3

40.3

25.4

20 Feb. 2022

7.0

40.8

27.5

About 35 tsunami waves occurred in the area of the Marmara Sea since 120 AC. The map in Fig. 1 presents the location of the tsunami sources according the National Geophysical Data Center [17]. The size of the stars corresponds to the tsunami event validity and ranges from 0 to 4. Some of the notable tsunamis are described below. After an earthquake on 18 October 1343 [17]/14 October 1344 [12, 18] a tsunami stepped in for 2000 m [19]. In 1509 a tsunami was generated and the wave run-up was about 6.0 m [20]. An earthquake back in 1894 induced tsunami in the region of Istanbul and the Prince Islands, where the Golden Horn was inundated. The sea first receded and not long after the coast was inundated with run-ups up to 2.7 m [18]. The

Historical Earthquakes and Tsunami Waves

79

events from 1912 and 1999 are investigated in more details in this study and simulated numerically. The S, arköy-Mürefte earthquake occurred on August 9, 1912 with epicenter located in land and estimated magnitude M7.4. The sea withdrew near Tekirdag, the coasts of Çanakkale were inundated. Near Istanbul (Ye¸silköy, Kadıköy and Bosphorus) an eyewitness accounted retreat and inundation and several strong sea waves before and after the earthquake [21]. The earthquake from August 17, 1999 happened on the northern branch of NAF with an epicenter location in the province of Kocaeli. The generated tsunami waves were observed in the whole Izmit Bay, from Darica to Yalova. In Halidere the sea receded for 15 m and then penetrated inland for 50 m. Near the port of Izmit the sea receded about 40 m and several boats which were anchored fell down by 2 m [18].

Fig. 1. Tsunami sources in the Marmara Sea denoted with blue stars [17]. The size of the stars corresponds to the tsunami event validity. The brown lines show the fault sources.

3 Tsunami Simulation Procedure The tsunami simulation procedure is based on two main steps: the tsunami generation process and the propagation and inundation processes. The initial tsunami conditions are based on the simplified method proposed by Okada [3]. For a rectangular fault placed on a fixed depth together with the geometry of the source and the focal mechanism, Okada presents a set of solutions which give the sea floor deformation. The calculated seafloor displacements are the initial tsunami conditions. In Table 2 we summarize the known focal mechanisms for the two events. For the S, arköy-Mürefte earthquake there isn‘t so much information about the focal mechanism, so we retrieve this information from a paper. Table 3 presents the calculated geometry of the faults according the regressions proposed in the papers [22, 23]. For Kocaeli earthquake the simulations are based on the three types of the geometry in combination with different focal mechanisms.

80

L. Dimova and R. Raykova Table 2. Focal mechanisms (FM) of the simulated earthquakes.

Event

Strike [°]

Dip [°]

Rake [°]

Author

9 Aug. 1912

68

75

213

[24]

17 Aug. 1999

255

84

178

ISC

75

89

179

NEIC

95

81

180

NEIC

90

72

196

ZUR_RMT

91

87

164

GCMT

81

88

183

CSEM

271

78

192

MOS

FM

[24] – see references; ISC – International Seismological Centre (United Kingdom), NEIC – National Earthquake Information Center (U.S.A.), ZUR_RMT – Swiss Seismological Service (Switzerland), GCMT – The Global CMT Project (U.S.A.), CSEM – The European-Mediterranean Seismological Centre (France), MOS – Geophysical Survey of Russian Academy of Sciences (Russia). Table 3. Geometry of the selected seismic sources.

Geometry

Length [km]

Width [km]

Displacement [m]

84

23

2.5

56

26

3.5

74

38

1.7

The numerical model UBO-TSUFD [2] is used for the tsunami propagation and interaction with the coast. The procedure combines the outputs from Okada’s model, the built grid and the equations of Navier-Stokes both in linear and non-linear approximation. As a result, we have the propagation field, tsunami travel times, the fields of maximum elevation and when fine nested grids are applied, we compute the inundation line on

Historical Earthquakes and Tsunami Waves

81

the coastline. This procedure is validated in several cases in eastern Mediterranean and Aegean Sea [25–28] and the results show that the model describe very well the processes related to tsunami propagation. Tsunami simulations are computed numerically using rectangular grid built on available data for bathymetry and topography (see Fig. 2). GEBCO [29] provides 15 arc seconds freely distributed data for global coverage thus for the Marmara Sea we built 250 m spatial resolution grid. The time step for each cell is 1 s.

Fig. 2. Topography and bathymetry map. Black and red lines show the seismogenic faults according the European Database of Seismogenic Faults [30]. M1 and M2 show the sources related to the simulated events.

Fig. 3. The position of virtual tide gauges is shown as black triangles. Synthetic mareograms of the virtual mareographs (red triangles) are presented in Sect. 4.

82

L. Dimova and R. Raykova

4 Results of Tsunami Simulations 4.1 S, arköy-Mürefte Tsunami Simulation For this event we accounted the first and the second nodal plane of the focal mechanism and the three types of geometry, thus we have 6 different scenarios. In this section we present only one parametrization. For the geometry we chose L = 84 km, W = 23 km, D = 2.5 m, the upper border of the depth of the fault is 2 km, and the first plane solution is taken into account. Figure 4 (left) illustrates the computed initial tsunami elevations. The heights vary in the range −0.7 up to 0.4 m, as the negative values are closer to the sea. Figure 4 (right) presents the maximum tsunami elevations. Higher values are concentrated northeastern from the epicenter location. It is well seen that the waves propagate through Dardanelles Strait and central part of Marmara Sea. The computed heights reach 1.4 m near Tekirdag.

Fig. 4. Initial tsunami elevations (left) and maximum tsunami heights (right) for S, arköy-Mürefte simulation.

The tsunami travel time maps are quite useful in tsunami early warning. The computed time of the first arrival wave is presented in Fig. 5 top left panel. The colored scale is divided in 5 min intervals. The map shows the first 25 min after the earthquake onset. It is clearly seen that the whole western part of Marmara, from Gelibolu to Tekirdag, is affected by the tsunami in less than 5 min. The sea level oscillations at certain points are calculated. These virtual mareographs are located close to the shore (Fig. 3). The model computes the tsunami elevation over the simulation time. Figure 5 upper right and bottom panels present the tsunami heights near S, arköy, Tekirdag and Çanakkale. In the latter one the mareogram shows very short period oscillations. For Tekirdag the estimated elevation reaches more than 0.4 m.

Historical Earthquakes and Tsunami Waves

83

Fig. 5. Travel time map (upper left) for S, arköy-Mürefte simulation. Synthetic mareograms (upper right and bottom panels). The position of the virtual tide gauges is shown in Fig. 3 as red triangles.

4.2 1999 Kocaeli Tsunami Simulation Reviewing the International Seismological Center‘s catalogue we saw that the epicenter‘s location of the earthquake from 1999 is not defined unambiguous. Many authors estimated the location of the epicenter on land, rather than offshore. NEIC data center evaluate the location offshore Izmit Bay. We selected their solution and the one proposed by ISC. In this section we present results based on the first nodal plane solutions (Table 2) and for the geometry we chose L = 84 km, W = 23 km and D = 2.5 m. Figure 6 show the initial tsunami elevations for both scenarios. They are quite similar if one considers the focal mechanism solution. Nevertheless, the epicenter‘s location is very important. The model in this case underestimates the heights reported in the paper. The tsunami initial elevations range between −0.22 and 0.2 m. The maximum elevation field presented in Fig. 7 indicate the low tsunami amplitudes, compared to the observed one in 1999. It is clearly seen that the tsunami did not propagate toward the western part of the Marmara Sea. The energy of the tsunami is concentrated in the Izmit Bay.

84

L. Dimova and R. Raykova

The tsunami time travel maps in Fig. 8 show that potentially threatened zones in just 5 min are the localities from the coastline between Darica and Yalova. Istanbul and adjacent areas are not affected due to these scenarios. The synthetic mareograms presented in Fig. 9 show the first two hours of the tsunami simulation. For Kocaeli virtual gauge (upper and bottom right) we show the first 30 min from the simulation in order to vie more detailed the first negative waves, thus the receding of the see. In Istanbul the wave period is about 1 h and the amplitudes are insignificant. In Yalova is the opposite, very short period oscillations.

Fig. 6. Initial tsunami elevations for Kocaeli simulation: ISC left panel, NEIC right panel.

Fig. 7. Maximum elevation fields for Kocaeli simulation: ISC left panel; NEIC right panel

Historical Earthquakes and Tsunami Waves

85

Fig. 8. Travel time maps for Kocaeli simulation: ISC left panel; NEIC right panel.

Fig. 9. Synthetic mareograms for Kocaeli (NEIC) simulation. The position of the virtual tide gauges is shown in Fig. 3 as red triangles. The left panels present mareographs near Istanbul and Yalova. The right panels show virtual tide gauge offshore Kocaeli: upper – first 2 h; bottom: first 30 min.

5 Conclusions The region of Marmara Sea suffered strong earthquakes and tsunamis in its historical record. In many cases the information about certain event is not sufficient, therefore numerical models are used to reproduce the process of generation or propagation of tsunami waves. In this study we did an attempt to review important earthquakes generated tsunami and describe them in brief. Tsunami simulations for two significant earthquakes (1912 S, arköy-Mürefte and 1999 Kocaeli) are performed using the numerical model UBO-TSUFD, based on the Navier-Stokes equations in approximation of shallow water theory, coupled with bathymetry and topography data (250 m). A total

86

L. Dimova and R. Raykova

number of 6 scenarios for 1912 S, arköy-Mürefte earthquake and 18 scenarios for 1999 Kocaeli earthquake are built. Numerical models of the seismically generated tsunami often underestimate the elevations of the observed tsunami waves. Nevertheless, underwater landslides are not excluded in the generation of tsunamis. Most probably this is the case in both events, an earthquake that activate an underwater slide mass movement and consecutive tsunami waves. Such complex sequence of natural hazards may increase significantly expected tsunami heights. The travel time maps show difficulties related to the short evacuation times for certain places. In order to solve this problem proper evacuation plans and population training must be in place. Acknowledgements. The first author would like to thank the Tsunami Research Team from University of Bologna for the opportunity to work with UBO-TSUFD. This paper was supported by the project CP-06-COST-7/24.09.2020 “Tsunami Hazard Assessment in the Southeastern European region”, funded by BNSF. The first author (LD) contributed to the European Cooperation in Science and Technology COST project “AGITHAR-Accelerating Global science In Tsunami HAzard and Risk analysis”.

References 1. Altinok, Y., Alpar, B., Oezer, N., Aykurt, H.: Revision of the tsunami catalogue affecting Turkish coasts and surrounding regions. Nat. Hazards Earth Syst. Sci. 11(2), 273–291 (2011) 2. Tinti, S., Tonini, R.: The UBO-TSUFD tsunami inundation model: validation and application to a tsunami case study focused on the city of Catania, Italy. Nat. Hazards Earth Syst. Sci. 13, 1759–1816 (2013) 3. Okada, Y.: Surface deformation due to shear and tensile faults in a half-space. BSSA 75(4), 1135–1154 (1985) 4. Ambraseys, N.: Value of historical records of earthquakes. Nature 232, 375–379 (1971) 5. Ambraseys, N., White, D.: The seismicity of the Eastern Mediterranean region 550-1 BC: a reappraisal. J. Earthquake Eng. 1, 603–632 (1997) 6. Guidoboni, E., Comastri, A., Traina, G.: Rom Istituto Nazionale di Geofisica. Catalogue of Ancient Earthquakes in the Mediterranean Area up to the 10th Century, p. 504. Istituto nazionale di geofisica, Rome (1994) 7. Evangelatou-Notara, F.: Earthquakes in Byzantium from the 13th to 15th Century. Historical Research, Athens (1993) 8. Ambraseys, N., Finkel, C.: The Seismicity of Turkey and Adjacent Areas: A historical review, 1500–1800. Muhittin Salih EREN, Istanbul (1995) 9. Ambraseys, N., Finkel, C.: Seismicity of Turkey and neighbouring regions, 1899–1915. Ann. Geophys. Series B. Terr. Planet. Phys. 5(6), 701–725 (1987) 10. Ambraseys, N., Finkel, C.: The Saros–Marmara earthquake of 9 August 1912. Earthquake Eng. Struct. Dynam. 15(2), 189–211 (1987) 11. Ambraseys, N.: Engineering seismology. J. Earth. Eng. Struct. Dyna. 17, 1–105 (1988) 12. Yalçıner, A.C., Alpar, B., Altınok, Y., Özbay, ˙I, Imamura, F.: Tsunamis in the Sea of Marmara: historical documents for the past, models for the future. Mar. Geol. 190(1–2), 445–463 (2002) 13. Barka, A.: Neotectonics of the Marmara region. In: Active Tectonics of Northwestern Anatolia—The MARMARA Poly-Project, pp. 55–87 (1997) 14. Barka, A., Reilinger, R.: Active tectonics of the Eastern Mediterranean region: deduced from GPS, neotectonic and seismicity data. Annali di Geofisica, XL 3, 587–610 (1997)

Historical Earthquakes and Tsunami Waves

87

15. International Seismological Centre: ISC-GEM Earthquake Catalogue (2022). https://doi.org/ 10.31905/d808b825 16. Ambraseys, N.: The seismicity of the Marmara Sea area 1800–1899. J. Earthquake Eng. 4(3), 377–401 (2000) 17. National Geophysical Data Center/World Data Service: NCEI/WDS Global Historical Tsunami Database. NOAA National Centers for Environmental Information (2022). https:// doi.org/10.7289/V5PN93H7 18. Altinok, Y., Ersoy, S., Yalciner, A.C., Alpar, B., Kuran, U.: Historical Tsunamis in the Sea of Marmara. In Int. Tsunami Symp. ITS 4–2, 527–534 (2001) 19. Ambraseys, N.: Data for the investigation of the seismic sea-waves in the Eastern Mediterranean. Bull. Seismol. Soc. Am. 52(4), 895–913 (1962) 20. Oztin, F., Bayulke, N.: Historical earthquakes in Istanbul, Kayseri, Elazig. In: Proceedings of the worckshop on Historical Seismicity and Seismotectonics of the Mediterranean Region, pp. 150–173. 10–12 Oct 1990. Istanbul, Turkish Atomic Energy Authority, Ankara (1991) 21. Altınok, Y., Alpar, B., Yaltırak, C.: Sarköy ¸ – Mürefte 1912 Earthquake’s Tsunami, extension of the associated faulting in the Marmara Sea. Turkey J Seismol 7, 329–346 (2003). https:// doi.org/10.1023/A:1024581022222 22. Mai, P., Beroza, G.: Source scaling properties from finite-fault-rupture models. Bull. Seismol. Soc. Am. 90(3), 604–615 (2000) 23. Wells, D., Coppersmith, K.: New empirical relationships among magnitude, rupture length, rupture width, rupture area, and surface displacement. Bull. Seismol. Soc. Am. 84(4), 974– 1002 (1994) 24. Aksoy, M.E.: The 9 August 1912 Mürefte-Sarköy ¸ earthquake of the North Anatolian fault. Med. Geosc. Rev. 3(1), 95–114 (2021). https://doi.org/10.1007/s42990-021-00050-z 25. Dimova, L., Raykova, R.: Tsunami radiation pattern in the Eastern Mediterranean. J. Phys. Technol. 1(2), 22–27 (2017) 26. Dimova, L.: Tsunami radiation pattern in the southern Aegean Sea Annual of Sofia University “St. Kliment Ohridski.” Fac. Phys. 111, 23–40 (2018) 27. Dimova, L., Raykova, R., Armigliato, A., Pagnoni, G., Tinti, S.: Modelling of earthquakeinduced tsunami in the Eastern Mediterranean region. AIP Conf. Proc. 2075(120024), 2019 (2019). https://doi.org/10.1063/1.5091282 28. Dimova, L., Raykova, R., Armigliato, A., Pagnoni, G. Tinti, S.: Aggregated tsunami scenario for Karpathos Island. In: Proceedings of: 1st International Conference on ENVIROnmental protection and disaster RISKs, 29–30 Sep 2020, pp. 443–451 (2020). https://doi.org/10.48365/ envr-2020.1.40 29. GEBCO: https://www.gebco.net/data_and_products/gridded_bathymetry_data/ 30. Basili R., et al.: The European Database of Seismogenic Faults (EDSF) compiled in the framework of the Project SHARE. http://diss.rm.ingv.it/share-edsf/ (2013). https://doi.org/ 10.6092/INGV.IT-SHARE-EDSF

Climate Change and Security Implications

Wind Speed and Temperature Variations in Burgas Region Since 1836 Yavor Chapanov(B) Climate, Atmosphere and Water Research Institute, Bulgarian Academy of Sciences (CAWRI-BAS), Sofia, Bulgaria [email protected]

Abstract. The global wind speeds vary in time. They had been decreasing for several decades, starting in the 1970s and they are getting faster during the last decades. The Intergovernmental Panel on Climate Change states that “there is evidence for long-term changes in the large-scale atmospheric circulation, such as a poleward shift and strengthening of the westerly winds” and that these observed changes likely will continue. The wind speed variations are important in the field of renewable energy sources. Other application of wind variation study is the influence of winds on air quality in urban areas. Recently new time series of wind data have been created in 20th Century Reanalysis Version 3 data provided by the National Oceanic and Atmospheric Administration / Oceanic and Atmospheric Research / Earth System Research Laboratories - Physical Sciences Laboratory, Boulder, Colorado, USA. The 20th Century Reanalysis Version 3 data set cover large time period (1836–2016) globally gridded data set that represents several Earth’s atmosphere parameters. Time series of air temperature, U- and V- components of winds over Burgas for the trapezoid 43-44N, 27-28E are extracted from the gridded data, and next, time series of wind speed is calculated. The variations of air temperature and wind speed are analyzed and compared with the cycles of solar activity, represented by the Index of Total Solar Irradiance (TSI). The spectra of wind oscillations are determined by Fast Fourier Transform (FFT). The common solar-climate cycles are separated from time series by the Method of Partial Fourier Approximation (PFA). These cycles are from 16 narrow frequency bands, whose periods are between 5 and 180 years. Their superposition may help long-term forecasts and better understanding of solar influence on atmospheric circulation over South-Eastern part of Bulgaria. Keywords: Climate change · Wind · Temperature · TSI

1 Introduction The Earth’s mean temperature rises during the last century and much more during the last decades. This rising affects all climate cycles in the atmosphere and ocean, in particular it affects the variations of winds speed and directions. Wind power has been threatened by reductions in global average surface wind speed, which have been occurring over land since the 1980s, a phenomenon known as global terrestrial stilling. The decadal © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 91–99, 2023. https://doi.org/10.1007/978-3-031-26754-3_8

92

Y. Chapanov

variations of near surface wind are probably determined by internal decadal ocean– atmosphere oscillations, where the significant part of energy excitation comes from solar-terrestrial influences and variations of TSI. The wind speed variations change the amount of available renewable energy. A great interest exists in authors about the topics of wind speed variations and renewable energy generation. The most of publications in the field of wind study are devoted to this problem. Another important application of wind properties is connected with the study of ecosystems, air quality and air pollution in urban areas. The extreme wind events could be dangerous and can cause significant infrastructure damages, economic losses and even loss of human life. Therefore, the results of research in this field are of great importance for decision-making purposes related to better management and reduction of risks (Barantiev et al. 2021; Kirova et al. 2015; 2017; 2018).

2 Data and Methods The time series of temperature, U- and V-components of wind at geopotential height 1000 Hpa (110.8 m above sea level) for the period 1836.0–2016.0 are taken from the NCEP/NCAR Reanalysis data set (20th Century Reanalysis Version 3 Dataset) for the trapezoid between 43-44N and 27-28E. (Figs. 1 and 2). The wind speed over Burgas region is calculated from its components. The solar data are presented by the Total Solar Irradiance (TSI) variations (Fig. 3). The daily reconstruction of TSI data since 1850 is a composite of SATIRE-T reconstruction from (Krivova et al. 2010) for 1850 to 22 August 1974; and SATIRE-S reconstruction from (Yeo et al. 2014) for 23 August 1974 onwards. The 0.1-year values of TSI are calculated by means of robust Danish method

Fig. 1. Map of Burgas region in Bulgaria.

Wind Speed and Temperature Variations in Burgas Region Since 1836

93

(Juhl 1984; Kegel 1987; Kubik 1982). This method allows to detect and isolate outliers and to obtain accurate and reliable solution for the mean values. The spectra of wind component variations are determined by the Fast Fourier Transform. The periodical terms are determined by the Method of Partial Fourier Approximation (Chapanov et al. 2015). After estimating the Fourier coefficients by the Method of Least Squares, it is possible to identify a narrow frequency zone presenting significant amplitude, and defining a given cycle. Then this cycle can be reconstructed in time domain as the partial sum limited to the corresponding frequency bandwidth. Doing this for terrestrial and solar time series, we shall identify their respective cycles, isolate and compare the common ones. This method provides Least Squares estimate of 150 harmonically coefficients of wind speed variations with average accuracy of about 6 cm/s and coefficients of TSI variations with accuracy 5 mW/m2 .

TSI [W/m**2]

Fig. 2. Time series of U- and V- coordinates of the wind, wind velocity Vel and temperature T over Burgas region.

1840 1363 1362 1361 1360 Epoch 1840

1860

1880

1900

1920

1940

1960

1980

2000

1860

1880

1900

1920

1940

1960

1980

2000

2020 1363 1362 1361 1360 2020

Fig. 3. Time series of daily values of TSI (in yellow color) and their normal points at 0.1-year intervals (in red color).

94

Y. Chapanov

3 Results 3.1 Time Series Spectra The time series spectra, determined by the Fast Fourier Transform, are shown in Fig. 4. The spectra of wind components are not coherent and they have different peaks for some frequencies. So, they have different sources of excitation of interannual and decadal oscillations. The TSI variations are the main source of temperature variations, but only limited number of oscillations have common spectral peaks in Fig. 4.

Fig. 4. Time series spectra of U- and V- wind components, wind velocity Vel, temperature T over Burgas region and TSI.

3.2 TSI Influence on Wind Velocity and Air Temperature The solar influence on wind and temperature variations over Burgas region are analyzed by comparison of common interannual, decadal and centennial cycles in Figs. 5–7.

Wind Speed and Temperature Variations in Burgas Region Since 1836

95

Fig. 5. Common decadal and centennial cycles of TSI, wind velocity and temperature over Burgas region.

96

Y. Chapanov

Fig. 6. Common decadal and interannual cycles of TSI, wind velocity and temperature over Burgas region with periodicity below 22 years.

Wind Speed and Temperature Variations in Burgas Region Since 1836

97

Fig. 7. Common interannual cycles of TSI, wind velocity and temperature over Burgas region.

The long-term oscillations of TSI, wind velocity and air temperature are compared in 6 narrow frequency bands with periods between 22.5 and 179.8 years in Fig. 5. The amplitudes of long-term wind velocity vary between 0.05 and 0.4 m/s, where their cumulative effect is up to 1 m/s. The temperature amplitudes vary between 0.8 and 6 °C in 5 frequency bands. Their possible total effect may reach 13 °C. These bands

98

Y. Chapanov

cover several cycles of solar activity – solar magnetic cycle with 22-year period; 45-year cycles of North-South solar asymmetry; Gleisberg cycles (70–120 years) and 178.7-year planetary term of solar cycles. The decadal oscillations of TSI, wind velocity and air temperature with periodicity below 22.5 years are compared in 6 narrow frequency bands in Fig. 6. The periods of these bands are between Schwabe (11 years) and Hale (22 years) cycles. Relatively good agreement exists between solar and climatic cycles with some small phase and amplitude deviations and 90-year anti phase between wind velocity and TSI for periodicity band 16.3–18.0 years. The amplitude of wind velocity from these bands vary between 0.05 and 0.10 m/s, while the amplitudes of temperature are between 0.4 and 0.8 °C. The total amplitude of wind velocity is up to 0.4 m/s, and the total amplitude of temperature – up to 3.5 °C. The subdecadal oscillations of TSI, wind velocity and air temperature with periodicity below 9 years are compared in 5 narrow frequency bands in Fig. 7. Relatively good agreement exists between solar and climatic cycles with small amplitude deviations. The amplitude of wind velocity from these bands vary between 0.01 and 0.13 m/s, while the amplitudes of temperature are between 0.1 and 0.5 °C. The total amplitude of wind velocity is up to 0.3 m/s, and the total amplitude of temperature – up to 1.5 °C.

4 Conclusions The climatic conditions at Black Sea coast near Burgas strongly depend on the solar activity cycles and their harmonics. The wind velocity and air temperature have common oscillations with TSI in 16 narrow frequency bands, whose periodicity are centennial, decadal and interannual. The models of periodical oscillations of climatic indices, based on these common frequency bands, may help to create long-term forecasts in this century. Acknowledgements. The study is supported by the National Science Fund of Bulgaria, Contract KP-06-N34/1/30-09-2020 “Natural and anthropogenic factors of climate change – analyzes of global and local periodical components and long-term forecasts”.

References Barantiev, D., Batchvarova, E., Kirova, H., Gueorguiev, O.: Numerical modeling of extreme wind profiles measured with SODAR in a coastal area. In: Dimov, I., Fidanova, S. (eds.) HPC 2019. SCI, vol. 902, pp. 171–183. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-553470_15 Chapanov, Y., Ron, C., Vondrák, J.: Millennial cycles of mean sea level excited by Earth´s orbital variations. Acta Geodyn. Geomater. 12(3), 259–266 (2015) Juhl, J.: The “Danish Method” of weight reduction for gross error detection. In: XV ISP Congress proc., Comm. III, Rio de Janeiro (1984) Kegel, J.: Zur Lokalizierung grober Datenfehler mit Hilfe robuster Ausgleichungsfervahren. Vermessungstechnik 35, 348–350 (1987) Kirova, H., Barantiev, D., Nikolov, V., Batchvarova, E.: Wind field in a closed breeze cell in Ahtopol – modelling and observations. International scientific on-line journal. Science &Technologies, V, 3, Union of Scientists – Stara Zagora, pp. 25–29 (2015). ISSN: 1314-4111

Wind Speed and Temperature Variations in Burgas Region Since 1836

99

Kirova, H., Batchvarova, E., Barantiev, D.: Horizontal scale of closed breeze cells at the southern Bulgarian black sea coast. In: Proceedings of the 18th International Conference on Harmonisation within Atmospheric Dispersion Modelling for Regulatory Purposes (harmo18), pp. 401–405, 9–12 Oct 2017, Bologna, Italy (2017) Kirova, H., Barantiev, D., Batchvarova, E.: Evaluation of mesoscale modelling of a closed breeze cell against sodar data. In: Mensink, C., Kallos, G. (eds.) ITM 2016. SPC, pp. 151–155. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-57645-9_24 Krivova, N.A., Vieira, L.E.A., Solanki, S.K.: Reconstruction of solar spectral irradiance since the Maunder minimum. J. Geophys. Res. 115(A12112), 1–11 (2010) Kubik, K.: An error theory for the Danish method. In: ISP Symposium, Comm. III, Helsinki (1982) Yeo, K.L., Krivova, N.A., Solanki, S.K., Glassmeier, K.H.: Reconstruction of total and spectral solar irradiance from 1974 to 2013 based on KPVT, SoHO/MDI and SDO/HMI observations. Astron. Astrophys. 570(A85), 1–18 (2014)

Climate Change Implications on the Condition of the Road Surface in Bulgaria Georgi Belev1(B)

, Petja Ivanova-Radovanova1 and Hristo Chervenkcov2

, Vladimir Ivanov3

,

1 Climate, Atmosphere, and Water Research Institute, Bulgarian Academy of Sciences

(CAWRI-BAS), Bul. Tsarigradsko Chausse, 66, Sofia, Bulgaria [email protected], [email protected] 2 National Institute of Meteorology and Hydrology (NIMH), Bul. Aleksandar Malinov, 1, Sofia, Bulgaria [email protected] 3 National Institute of Geophysics, Geodesy, and Geography at the Bulgarian Academy of Sciences (NIGGG-BAS), Ul. Akad. G. Bonchev, Bl. 3, Sofia, Bulgaria [email protected]

Abstract. The passage of the temperature through 0 °C creates conditions for freezing of the surface layer of the road surfaces, due to the accumulation of moisture from the condensation of water vapor. The risk of freezing of the road surface in Bulgaria is assessed by observing data on air temperature and humidity in the period 2017–2020. By applying a well-established methodology and creating raster images in a geoinformation environment, the processes favorable to the conditions for freezing of roads are investigated. The analysis of the data and images reveals interrelations and regularities between regions with a high, medium, and low degree of risk of frost and the climatic regions in Bulgaria. Keywords: Road frozen · Climate change · Seasonal conditions

1 Introduction Freezing of the road surfaces is a dangerous meteorological phenomenon that can cause serious traffic accidents causing great damage, injuries, and death. Frost processes often occur in clear and calm weather in anticyclonic conditions, hence the civic name – “quiet death” (Mahura et al. 2008). There are studies investigating the factors for road freezing and the role of different factors. For example, a study (Lim and Kim 2020) analyzed the scenario by applying the road freezing evaluation algorithm and derived evaluation through the change of the road surface condition and water film thickness of the freezing section. Other studies (Fujimoto et al. 2014), (Sangyoup et al. 2015) consider the problem of optimization of the application of deicing agents on the winter road surface and testing how water and deicing agents disperse due to passing vehicles as well as for calculating the dissolution rates of salt on the road surface. It is also recognized that the winter season is both very challenging for the process of road freezing period and for providing © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 100–109, 2023. https://doi.org/10.1007/978-3-031-26754-3_9

Climate Change Implications

101

essential transportation services in northern regions (Sladen et al. 2019). The type and structure of the road surface is another important factor, thus several types of pavement damage were measured by long-term observations (Baocun et al. 2020). Research on the course of temperature in Bulgaria is carried out by various scientific and research units – Bulgarian Academy of Sciences (BAS), Sofia University “St. Kl. Ohridski”, National Institute for Meteorology and Hydrology (NIMH). There was a great scientific contribution to climate change research published by (Tishkov 1976; Velev 1990; Velev 2010; Tishkov et al., 2002; Vekilska 2012; Nojarov 2014, 2019; Chervenkov et al. 2019; Chervenkov and Slavov 2019, 2020a, b, 2021) and especially for the cold period of the year (Malcheva 2017). The main focus of the present study is the freezing of the road surface mainly in the spring and autumn period, in connection with the behavior of owners and drivers of motor vehicles in the time outside the “seasonal” winter when vehicles are not equipped with winter tires and the “setting” of the drivers is not for driving in winter conditions (increased attention, lower speed, longer driving distance). The climate of Bulgaria suggests that there are specific circumstances for freezing and icing of roads in certain areas and outside the “winter season” (Octobe–March), which contributes to unfavorable and dangerous conditions with preconditions for traffic accidents with severe consequences – material damage, injuries, and deads. The study needs to analyze climatic factors and conditions in combination with an analysis of the features of the terrain and land cover. The main factors are the processes of temperature passing through 0 °C, humidity, sunshine, clouds, local atmospheric circulation, exposure, altitude, afforestation, and anthropogenic factors and conditions – width and composition of the pavement (asphalt, paving, macadam), the slope and curvature of the road, the environment of the road (urban and intercity road networks) (Kenderova et al. 2015). The problem’s urgency is related to investigating the relationships and patterns between dangerous and unfavorable processes and climatic features in Bulgaria. The location of the country in the south-southeastern parts of Europe and the characteristic of the region, a complex mosaic relief suggest peculiarities in the manifestations of climate elements. The main objectives of the present study are to differentiate the territory of the country by defining areas with high and low risk of road freezing and to establish the overlap of the defined areas with climate regions. For the study, the climate zoning of Bulgaria according to Velev (2010) is used. We hypothesize that there is a contradiction between climate change conditions when the warming scenarios decrease the risk of road freezing and on the other side the danger of more often occurrences of extreme temperatures remains high.

2 Materials and Methods The analysis of the genetic features of the freezing mechanism of the road surface is essential for the present research. Widespread among drivers is the term “black ice”, which is a thin transparent ice cover on the asphalt, “invisible” to drivers. It is the patterns associated with the formation of this cover that is the focus of this study. The genesis of the processes leading to the formation of “black ice” is associated with certain meteorological conditions in combination with the influence of the terrain. The main factors

102

G. Belev et al.

are the temperature conditions and mainly the environment in which the temperatures pass through 0 °C. The genesis of the “black ice” is analogous to the genesis of frost formation. It is known that during the spring the advective frosts predominate – related to the intrusion of cold air, while in the autumn the radiation frosts predominate, whose genesis is related to the radiative cooling of the earth’s surface in clear and calm weather and the concave forms of the relief (Velev 2010). The utilized meteorological data include surface temperature, 2 m. air temperature and the dew point are from the fifth-generation reanalysis of the European Centre for Medium-Range Forecasts ERA5 (Hersbach et al. 2020) downloaded from the Copernicus Data Store for a period of three years (2016–2018). Based on these data, the potential conditions leading to the development of icing, classified into five categories, are computed applying the methodology, described in (Mahura et al. 2008). The results are digital maps, stored in the standard form of NetCDF files. In the next stage, these files are processed in a geoinformation environment, raster images are generated of the temperature transition through 0, 1 and 2 °C, (respectively – CW0, RA1, and RA2), and the probability of temperature transition is expressed as a percentage. Situations of real freezing are presented through risk models, marked as ‘emergencies’ – Red Alert (RA), RA1, and RA2 (Mahura et al. 2008). The basic data is temperature with added layers of the different images of isometric indicators, road networks, and the location of settlements. The aim is to compile a generalized image representing the areas with conditions most likely to freeze. The cartographic models, created by the authors, are in a geo-informative environment. This process is done by using the software Arc GIS owned by ESRI-Bulgaria and licensed for usage by the Climate, Atmosphere, and Water Research Institute to the Bulgarian Academy of Sciences (CAWRI-BAS). The process of freezing the road surface is carried out in several stages: initially, a thin layer of snow accumulates on a wet surface; a stage of crystallization followed in which the crystals grew in the form of a column or also called congelation ice (Barrette 2015). The next stage is provoked by the dynamics of atmospheric processes, among which the most significant is the wave current of the wind. As a result of that, the ice is layered and due to lower ambient temperature, the ice cover begins to rise rapidly. Thus the process of freezing is completed by the increase of the temperature. In the presence of a layer of snow on the ice, the decomposition of the ice decreases. Several types of ice can be formed on the road surface: 1. Columnar – in which the axis of the ice crystal is in a horizontal plane, while the grains extend vertically in a columnar shape; 2. Pure (or black) – its genesis is similar to the columnar ones, but its invisible appearance creare difficulties to distinguish them when driving and that creates the most dangerous situations; 3. White ice – also called snow, originates from a layer of snow on an ice sheet, cracks are filled with snow and do not allow dissolved air to come out, therefore white ice is less dense than pure (Barrette 2015).

Climate Change Implications

103

The main indicators in the analysis of the study are the number of days of freezing, which is the annual sum of days in which the daily minimum temperature is below 0 °C, and the number of days of icing – the annual sum of days with maximum daily temperature below 0 °C, the extreme values of the minimum temperatures for a certain period, the monthly minimum values of the average daily temperatures and the percentage of days with minimum temperatures (www.moew.government.bg).

3 Research and Results The location of Bulgaria in the south-southeastern parts of Europe determines the peculiarities of the climate, among which are the presence of four seasons, and the predominant impact of northwestern transport of Atlantic air masses. The complex dissected relief (Fig. 1) of the country determined azonal climatic factors. The differences in the individual seasons are well expressed and have a significant impact on the conditions for freezing the road surface. The most important for the studied processes is the winter season with the greatest frequency of the days which transited the 0 °C temperature (Andreev et al. 2010). During this season, the conditions for freezing presuppose the implementation of active measures to reduce the dangerous impact of freezing on the road surface on the driving of motor vehicles to protect the lives and health of drivers and passengers. During the winter season, there is a special scenario on the part of the authorities to combat icing by cleaning and sanding the roads, as well as tight control over the speed of vehicles. At the same time, the special adjustment in the behavior of drivers during the winter is of great importance, which for the territory of Bulgaria covers the months of December, January, and February (unlike the astronomical December 21–March 20). The start and end dates of the first autumn and last spring frosts are essential for the analysis of the freezing processes (Chervenkov and Slavov 2022). The scope of the study does not consider the high mountain areas – like peaks Musala (2925 m), Vihren (2914 m), Botev (2376 m), etc., which are not used actively for transportation purposes. The attention of the study is focused on certain areas with special physical and geographical locations: high valley fields – valleys with an average altitude of 400 m such as Sofia (550 m), Pernik (700 m), Dupnitsa (450 m), Samokov (950 m), etc., open to the northwestern transmission of air masses and located in the northern parts of the mountains – Etropole region, Beglika (1552), low valley sections in the Fore-Balkans and the Danube plain – Kneja, Danube, open to the north and northeast – Dobrudzha plateau, Burgas lowland (Tishkov et al. 2002). Many of the given above sample regions are characterized by the location of large cities, economic activity, and related highways with heavy traffic during all seasons of the year. A peculiarity in terms of climate is the section Beglika in the Western Rhodopes. This region is characterized by one of the lowest average minimum temperatures (−12 °C) and the earliest date of the first frost – 20.08 and the latest date of the last – 21.07 (Velev 1990). An important road artery connecting the southern parts of the Western Rhodope Mountain with the Gorna Trakia Lowland passes through this section. The regions with early frosts are the most vulnerable to creating ice on the road surface, which for Bulgaria

104

G. Belev et al.

Fig. 1. Hypsometry and road transport network of Bulgaria (data based on JICA).

are in August for the region of Beglika, October – the Danube region, the northern parts of the Rhodopes, and the high valleys of Western Bulgaria, with the latest manifestation of the first frost, are the Black Sea region, the eastern parts of the Upper Thracian lowlands and the southern parts of the valleys of the rivers Struma and Mesta – the end of December. The onset of the last frost is important given the length of the period creating a risk of freezing the roads. At the latest, the frost subsides in June in the high parts of the Western Rhodopes and some high valleys of Western Bulgaria such as Dragoman, Breznik, Tran, and Ihtiman. Several maps had been created for the research, through which the processes of transitioning 0 °C are visualized, taking into account three autumn-winter periods – 2017/18, 2018/19, and 2019/20 (Fig. 2, 3, and 4) and hypsometric map of Bulgaria with a superimposed layer for the road network (Fig. 1 and Table 1). The input variables within the accepted approach required are air temperature (Ta), De, w/sublimation point (Td), and surface temperature (Ts). Situations of real freezing are presented through risk models, marked as ‘emergencies’ – Red Alert (RA), RA1 and RA2 represented in percentage according to a cartographic method for quantitative background. This method is used for the visualization in Fig. 2 of the risk of complicated situations on the road and the 1st and 2nd-degree risks of freezing. The higher degree of risk is represented by the higher density of the red color. On the maps (Fig. 2, 3 and 4) the image of the Black Sea is with maximum

Climate Change Implications

105

Table 1. Summary of the occurrence of potential conditioning leading to the development of icing on Bulgarian roads (based on road weather seasons 2017–2020). (The factors in Table 1 are adopted by Fig. 1 in Intercept of the report of Mahura, A., Petersen, C., and Sass, B.-H. (2008) Road Icing Conditions in Denmark, Scientific Report 08-03 (available at: www.dmi.dk/dmi/sr0803.pdf). Situation

Definition

Level

# cases

% cases

RA1

Ts < 0 & Ts ≈ Td

Red alert

4 216 233

9,1

RA2

Ts < 0 & Td ≈ Ta

Red alert

3 803 553

8,21

CW0

Ts ≤ 0

Warning

11 282 410

21,39

dense red color and is simply due to the consequence of “program reading” during the transformation of the digital data into raster data and it is not the subject of analysis in the present study. The map in Fig. 5, was created by the authors, to synthesize generated images represented in Figs. 2, 3, and 4 into one common layer of the risks for road freezing for the three basic time frame periods (2017–2018, 2018–2019, and 2019–2020) combined with the climate regions of Velev (2020).

Fig. 2. a) risk of freezing of the road surface in percentages for the period 2017/18; b) ‘emergency of the first degree in percentages for the period 2017/18; c) “emergency” of II degree for the period 2017/18

Fig. 3. a) risk of freezing of the road surface in percentages for the period 2017/18; b) ‘emergency of the first degree in percentages for the period 2018/19; c) “emergency” of II degree for the period 2018/19

106

G. Belev et al.

Fig. 4. a) risk of freezing of the road surface in percentages for the period 2019/20; b) ‘emergency of the first degree in percentages for the period 2019/20; c) “emergency” of II degree for the period 2019/20

4 Conclusion In the research of the scenarios related to the genesis and the interrelations between the dangerous natural processes and the processes of global climate change, it is necessary to pay attention to the special processes taking place in the atmosphere. The atmosphere itself can be seen as an open active thermodynamic system where the laws of Clapeyron and Carnot operate. When additional heat is applied, the system is activated. Therefore, when studying dangerous and unfavorable processes, one should start from the extreme manifestations in the atmosphere – big global warming and related climate change will not reduce the risk, but on the other side it will cause the earlier occurrence of frost on the road surface. At the same time, the additional warming of the ground layer will increase the vulnerability to the front of the road surface. The analysis of the temperature indicators in combination with the data for slopes, exposure and density, and distribution of the road network on the territory of Bulgaria shows that the roads in the mountainous parts, the high valley fields of Western Bulgaria have the highest risk of frost and the open spaces of Northeastern Bulgaria. In addition, given the projected global warming scenarios, the potential risk of road frosts will also decrease, but the danger of occurrence of extreme temperatures during the transition periods from autumn towards winter (October–November) and spring towards summer (March–April) remain. Under the influence of the General Atmospheric Circulation, the intrusion of cold air masses from the northeast in their collision with warm humid air in Mediterranean cyclones from the southwest can create a favorable environment for freezing roads, especially when “sucking” cold air masses from the Arctic or Siberia and establishing of an anticyclone lasting several days. Based on the analysis, several highly vulnerable regions are identified, which in certain extreme weather events will create a high risk of freezing of roads. Thus it leads to icing and a high risk of accidents while under the influence of frost weathering can lead to the destruction of the road surface, especially in the narrow ones where there is a risk of landslides and rock falls that can cause the closure of the road. The riskiest regions are the high valleys of Western Bulgaria, the northern parts of the Western Rhodopes, the Danube lowlands, the Dobrudzha coat, and the Burgas lowland (Fig. 5).

Climate Change Implications

107

The Gorno-Trakiiska nizina (Upper Thracian lowland) is defined with a medium degree of risk, the Black Sea coast, the southeastern parts of the country – Strandzha and Sakar and the valleys of the rivers Struma and Mesta are defined with low risk. The overlapping coverage of such defined risk areas with climatic regions has is partial character. Most of the high-risk areas fall into the temperate-continental climate regions – the Danube plain with the Pred-Balkans Mountain and the Danube lowlands and the Dobrudzha plateau. Respectively low-risk areas fall into the continental Mediterranean climatic region – the valleys of the rivers Struma and Mesta, the mountains – Strandzha and Sakar Mountains, and the Black Sea coast. The transitional climatic region includes areas of high and medium risk – the first is the Western Rhodopes Mountain, Krajina (Pernik, Kyustendil), and the Western parts of Sredna Gora Mountain. Prevention to combat road frost does not offer many solutions – the main measure is sanding, accompanied by the use of salt, which in turn when dissolved in water during the thawing process damages the structure of road vehicles, so control remains essential on road traffic and the personal responsibility of drivers.

Fig. 5. Climatic regions and areas at risk of freezing (based on data by Velev 2010, Chervenkov and Slavov 2020a, b and JICA)

Acknowledgments. The authors would like to express their deep gratitude to the organizations and institutions (Max Planck Institute for Meteorology, the European Centre for Medium-Range Forecasts, and the Copernicus Data Store), which provide free-of-charge software and data.

108

G. Belev et al.

This work has been carried out in the framework of the National Science Program “Environmental Protection and Reduction of Risks of Adverse Events and Natural Disasters”, approved by the Resolution of the Council of Ministers № 577/17.08.2018 and supported by the Ministry of Education and Science (MES) of Bulgaria (Agreement № D01–363/17.12.2020).

References Andreev, V., Alexandrov, V., Bachvarova, E.: Current risk phenomena in the atmosphere. In: Demetra EOOD (ed.) (2010). ISBN 978–954–9526–76–9, C (BG) Baocun, Y., Zipeng, Q., Qingping, Z., Hongwei, L., Liang, L., Xiaosong, Y.: Pavement damage behavior of urban roads in seasonally frozen saline ground regions. J. Cold Reg. Sci. Technol. 174, 103035 (2020). https://doi.org/10.1016/j.coldregions.2020.103035 Barrette, P.D.: Overview of ice roads in Canada: Design, usage and climate adaptation, Technical Report OCRE-TR-2015–011, 1200 Montreal Rd, Ottawa, ON KIA 0R6, October 18, 2015, National Research Council Canada (ENG) (2015) Chervenkov, H., Slavov, K., Ivanov, V.: STARDEX and ETCCDI climate indices based on E-OBS and CARPATCLIM. In: Nikolov, G., Kolkovska, N., Georgiev, K. (eds.) NMA 2018. LNCS, vol. 11189, pp. 360–367. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-106928_40 Chervenkov, H., Slavov, K.: STARDEX and ETCCDI climate indices based on E-OBS and CARPATCLIM. In: Nikolov, G., Kolkovska, N., Georgiev, K. (eds.) NMA 2018. LNCS, vol. 11189, pp. 368–374. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-10692-8_41 Chervenkov, H., Slavov, K.: Historical climate assessment of temperature-based ETCCDI climate indices derived from CMIP5 simulations. C. R. Acad. Bulg. Sci. 73(6), 784–790 (2020). https:// doi.org/10.7546/CRABS.2020.06.05 Chervenkov, H., Slavov, K.: Historical climate assessment of precipitation-based ETCCDI climate indices derived from CMIP5 simulations. C. R. Acad. Bulg. Sci. 73(7), 942–948 (2020). https:// doi.org/10.7546/CRABS.2020.07.06 Chervenkov, H., Slavov, K.: ETCCDI climate indices for assessment of the recent climate over southeast Europe. In: Dimov, I., Fidanova, S. (eds.) HPC 2019. SCI, vol. 902, pp. 398–412. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-55347-0_34 Chervenkov, H., Slavov, K.: Inter-annual variability and trends of the frost-free season characteristics over Central and Southeast Europe in 1950–2019. J. Central Eur. Agric. (JCEA) 23(1), 154–164 (2022). https://doi.org/10.5513/JCEA01/23.1.3394 Fujimoto, A., et al.: A road surface freezing model using heat, water and salt balance and its validation by field experiments. J. Cold Reg. Sci. Technol. 106 – 107, 1–10, (2014). https:// doi.org/10.1016/j.coldregions.2014.06.001. Elsevier Hersbach, H., et al.: The ERA5 global reanalysis. Q. J. R. Meteorol. Soc. 146, 1999–2049 (2020). https://doi.org/10.1002/qj.3803 Kenderova, R., Rachev, G., Baltakova, A., Nikolova, N., Krenchev, D.: Variations in soil surface temperature in the pirin high mountain area and their relation with slope processes activity. Comptes rendus de lAcademie Bulgare des Sciences 68(8), 1027–1034 (2015). http://www. proceedings.bas.bg/ Lim, H.-S., Kim, S.-T.: A study on road ice prediction by applying the road freezing evaluation model. J. Korean Appl. Sci. Technol. 37(6), 1507–1516 (2020). https://doi.org/10.12925/ JKOCS.2020.37.6.1507 Mahura, A., Petersen, C., Sass, B.H.: Road icing Conditions in Denmark, DMI Scientifik Report 08–03, pp. 1–25 (2008). ISSN 1399–1949

Climate Change Implications

109

Malcheva, K.: Cold waves on the territory of Bulgaria in the period 1952–2011. Bul. J. Meteo. Hydro. 22(3–4), 16–31 (2017) Nojarov, P.: Atmospheric circulation as a factor for air temperatures in Bulgaria, Springer, Meteorology and Atmospheric Physics, vol. 125, Combined 3–4, pp. 145–158 (2014). ISSN 0177–7971 Nojarov, P.: Factors affecting air temperature in Bulgaria. Theor. Appl. Climatol. 137(1–2), 571– 586 (2019). https://doi.org/10.1007/s00704-018-2622-2 Sangyoup, K., Youngsoo, J., Sungkyu, K., Dongchan, M., Hohyuk, N., Jaisung, C.: A study on the effects of factors of traffic accidents caused by frozen urban road surfaces in the winter. Int. J. Highw. Eng. 17(2), 79–87 (2015) Sladen, W.E., Wolfe, S.A., Morse, P.D.: Evaluation of threshold freezing conditions for winter road construction over discontinuous permafrost peatlands, subarctic Canada. Cold Reg. Sci. Technol. 170, 102930 (2020). https://doi.org/10.1016/j.coldregions.2019.102930 Tishkov, H.: The Climate of the Mountain Regions in Bulgaria – Structure, and Genesis. In: BAS, S (BG) (ed.) (1976) Tishkov, H., Nikolova, M., Mateeva, Z., Velev, S.: Climate. In: Geography of Bulgaria: Physical geography, Socio-economic, Chapter 2, pp. 141–183. ForCom (BG) (2002). ISBN 954–464– 123-B Vekilska, B.: Genaral Climatology. St. Kliment Ohridski (2012). ISBN: 978–954–073 -383 -8 Velev, S.: The Climate of Bulgaria. Heron Press Ltd. (2010). ISBN: 978–954–580–283–6, C https://www.moew.government.bg/bg/analiz-i-ocenka-na-riska-i-uyazvimostta-na-sectorsite-vbulgarskata-ikonomika-ot-klimatichni-promeni/ – Analysis and assessment of the risk and vulnerability of the sectors in the Bulgarian economy to climate change, Operational Program 2007–2013 Copernicus Climate Change: https://climate.copernicus.eu/ JICA: http://www.wcoomd.org/en/topics/capacity-building/activities-and-programmes/cooper ation-programmes/the-wco-jica-joint-project.aspx

Climate Change Challenges and Security Implications on National Security in North Macedonia Aleksandar Petrovski1(B) , Nenad Taneski2 , Andrej Iliev2 , and Nikola Spasov2 1 University “Goce Delchev” Shtip, Military Academy “General Mihailo Apostolski”, Vasko

Karangjeleski Bb, Skopje, North Macedonia [email protected] 2 Army of North Macedonia, Shtip, North Macedonia

Abstract. It is the right of each of us to exist and live safely in our environment, in our country, but also globally on this planet. The large number of natural disasters that have occurred in the past in certain countries and which have caused serious damage to the national security of those countries as just a warning to all other countries that natural disasters can happen at any moment and that no country is immune to these phenomena. Some of these disasters are caused by normal natural phenomena, but most of them are unfortunately caused by the activities of people who are increasingly contributing to global warming as a major cause of climate change. A large number of human casualties and the destruction of the natural environment in which we live and depend on it and the endangerment of the vital interests of countries as a result of these climate changes are reason enough that in the future we all commit to reducing the causes that led to these consequences. Climate change is indeed a threat to the national security of nations, and global warming is a “multiplier of threats” of instability in some regions of the world, and such instability threatens to affect stable regions, given the tensions caused by climate change. Developing countries, many of which are strategically located around the world, are the least able to adapt and are likely to experience political instability in the face of extreme weather. The purpose of this research is to understand the impact of climate change on the future security environment, disruption of citizens’ safety, forms of disruption of national security and security environment, as well as the necessary measures and activities for prevention and timely prevention of threats, risks, and challenges of this kind. The paper will also consider the impacts of the human factor on the natural environment and general safety, through its activities and actions on it. Keywords: Climate change · National security · Natural environment · Prevention · Security implications

1 Introduction National security has no predefined content, and we cannot accurately measure the danger that threatens the population and the state, whether that danger comes from inside or from © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 110–124, 2023. https://doi.org/10.1007/978-3-031-26754-3_10

Climate Change Challenges and Security Implications

111

the outside, and today in the 21st century we are witnessing new security challenges, in which certainly include climate change and their impact on security. Nutrition systems, economies, settlements and societies are adapted to the climate in which we are currently living. But what if the climate changes too fast and we can’t follow that step? Climate change characterized the overall history of human society and consistently forced man and his activities to adapt to these climate changes. During the whole development of humanity, the devastating forces of nature and the repair of the consequences of natural disasters pose a huge challenge, as victims and material losses in these cases are always huge. Climate change is indeed a threat to the national security of nations, and global warming is a “multiplier of threats” of instability in some regions of the world, and such instability threatens to affect stable regions, given the tensions caused by climate change. Developing countries, many of which are strategically located around the world, are the least able to adapt and are likely to experience political instability in the face of extreme weather. Since 1988, scientists have warned that stroking the planet’s climate will cause unintentional, uncontrolled, globally sensitive consequences, which may be the second largest after the eventual global nuclear war. Since then, hundreds of scientific studies have documented more and more evidence that human activities have changed the climate in the world. The number of world leaders who are concerned about climate change is the biggest problem we face today, even a bigger problem than the threat of terrorism. Climate change is likely to have severe consequences for local, regional and global security. Drought, floods, hunger and natural time-related disasters can take thousands or even millions of lives and worsen existing tensions within states and between states and encourage diplomatic and trade disputes. In the worst case, further warming will reduce the capacities of Earth’s natural systems and raise the already rising seas, which could endanger island states, destabilize the global economy and geopolitical balance and cause violent conflicts. Existing security threats will be enhanced as climate change will have an increased impact on regional water supply, agricultural productivity, human health and ecosystems, infrastructure, the flow of finance and the economy, as well as the patterns of international migration.

2 Security Policy and Climate Change Security is a complex concept. In all communications, national security is what dominates. Large-scale violent conflict is a concern that receives the most attention from policymakers and the development of military capabilities to respond to it consumes the most national resources, but alternative security threats such as environmental change are increasingly being considered.1 Climate change is a security problem for some countries, cultures, and societies, through its effects on ecosystems and their populations, and through its indirect effects on development and political stability. The European Union and the world, in general, are focused on upgrading security policy and better identifying the long-term challenges of its strategic interests. The 1 Worldwide average 2.9% of GDP is spent on defense (UNDP).

112

A. Petrovski et al.

European Union is one of the first organizations to identify climate change as a multiplier of threats and has significantly increased the initiatives to integrate climate change into the foreign and security policies of the member states. In addition to this security policy development, the EU is increasingly making changes to its climate and energy policies. In January 2014, the European Commission proposed energy policy guidelines be achieved by 2030, focusing on reducing carbon emissions by 40%.2 Security and climate change are linked, and that connection has led to an agenda called “climate security”. This concept can be defined as a variety of foreign policy actions aimed at addressing the strategic and political consequences of climate change.3 Climate security was created with the intention of showing that security is not just a military practice or that environmental change is just a local problem, and to raise environmental issues from a “low policy” to a “high policy” level that states will commit to energy and resources to deal with environmental problems as well as other security problems. Policymakers need to have a good understanding of the likelihood of the consequences of climate change, the security threats of climate change to be treated in the necessary context, and defense mechanisms, including security and defense systems, to be properly organized to deal with potential climate shocks. Climate change has now become a relevant factor in strategic security policy thinking, moving away from its status of irrelevant environmental issues. The preliminary results are stark: about 70% of nations in the world explicitly state that climate change is a national security concern.4 Almost all nations that have official military planning have stated that their government considers missions like humanitarian assistance and disaster relief as critical responsibilities of their armed forces (Fig. 1).

Fig. 1. World map of countries that recognize climate changes as a security threat 5 2 European Commission, Communication from the Commission to the European Parliament, the

Council, the European Economic and Social Committee and the Committee of the Regions: A policy framework for climate and energy in the period from 2020 to 2030, COM(2014)15 final (Brussels: January 22, 2014). 3 EU, Climate change and EU security policy: An Unmet Challenge,Richard Youngs, May 2014. 4 The Global Security Defense Index on Climate Change sets out how governments around the world view climate change as a matter of national security and how their security agencies have begun to plan for the consequences of climate change. 5 https://www.americansecurityproject.org/climate-energy-and-security/climate-change/gsdicc/

Climate Change Challenges and Security Implications

113

In June 2013, the European Union’s Foreign Affairs Council adopted a conclusion on climate diplomacy and security, together with a new European External Action Service (EEAS) document that obliges ministers on an annual report on Progress in Implementing Foreign and Security Policy Parameters in Climate Change Management Strategies.6

3 Climate Change Challenges and Security Implications on National Security in North Macedonia North Macedonia it’s National Concept of Security and Defense as potential and real risks and dangers that can lead to crises and conflicts have recognized the degradation and destruction of the environment as one of the causes of climate change7 , while the Ministry of Environment and Physical Planning of North Macedonia has adopted a National Climate Change Plan which is a guiding document for policymakers in northern Macedonia in creating strategies for reducing greenhouse gas emissions and adapting to climate change and should strengthen dialogue, cooperation and exchange of information between all relevant factors in the country including the governmental, nongovernmental, scientific and private sector, which shows that North Macedonia has taken climate change seriously in its policies and strategies. The most important climatic factors for determining the climate in North Macedonia are geographical position, relief, proximity to the surrounding seas, atmospheric currents, and more recently human activities. The influences of these climatic factors on the territory of our country enabled the presence of 4 climatic types: altered Mediterranean, temperate-continental, continental and mountainous. From here, two seasons stand out: cold, wet winters and dry, hot summers, associated with the transition seasons, spring and autumn. Spring starts earlier and autumn later. Autumn (especially October) is warmer than spring (April). Additionally, in the high mountainous areas there is a mountainous climate characterized by short and cold summers and significantly cold and moderately humid winters, where the precipitation is mostly snowy. The average annual temperature is high (above 14–15 ºC), and the average annual fluctuation is low (below 20 ºC). Absolute temperature fluctuation is lower (usually below 50 ºC). In the mountainous climates, the average annual temperatures are: 4.7 ºC in Popova Sapka (1750 m), 6.8 ºC in Lazaropole (1330 m), and 8.2 ºC in Krushevo (1230 m). Climate is influenced by several climatic factors such as: • • • • •

Latitude, Winds and pressure fields, Ocean currents, Oceans and air currents, The relief,

6 Foreign Affairs Council Meeting, “Council Conclusions,” Luxembourg, June 24,2013; External

Action Service, EU Climate Diplomacy for 2015 and Beyond: Reflection Paper, 2013. 7 Assembly of the Republic of Macedonia, National Concept for Security and Defense, 2003.

114

A. Petrovski et al.

• Altitude, • Clouds and • other factors. Climate and climatic factors are an element that significantly affects the exposure of the state to the risks of natural disasters and catastrophes that can cause more serious consequences for human life, as well as the material and natural resources of the state. According to scientists, the effects of climate change so far can be seen in the following:8 • • • • • • • • • • •

Increased surface temperatures, Sea level rise, Retreat of glaciers and melting of icy sea surfaces, Change in precipitation, More intense weather extreme events such as heat waves, tornadoes, hurricanes and heavy rains, Long, extreme droughts, Expansion of subtropical deserts, Endangered species and extinction and loss of biodiversity, Reduced yields from agriculture, Spread of vector-borne diseases and Increasing the acidity of the oceans.

Regarding whether climate change is a threat to the national security of the Republic of N. Macedonia, according to the results obtained from the survey conducted for the needs of this paper (pie 1), the opinion of 128 respondents, or 62% of respondents is that they pose a threat. National security and that in planning the country’s security policy, climate change should be considered as one of the most serious threats to the national security of the country, and appropriate measures should be taken to prevent the causes leading to climate change both locally and globally.

PIE 1.

Climate change is a threat to the national security of the Republic of N. Macedonia? NO 80 (38%) YES 128(62%)

Out of 208 respondents, 80 or 38% believe that climate change does not pose any threat to the national security of the country and that the Republic of N. Macedonia as 8 www.ecolife.com/define/climate-change.html#sthash.gaYTryBr.dpuf.

Climate Change Challenges and Security Implications

115

a small country is not a factor that greatly contributes to global warming and cannot greatly contribute to preventing the same. Rising temperatures in the last century have led to a package of climate change, including some that have led to greater flood risks. The increase in heavy rainfall is one of the clearest trends observed over the past few decades and projected by climate models for the next century. Global warming is contributing to increasing rainfall as warm air can absorb more water. For each heating of 0.5 ºC, the atmospheric water vapor increases by 3–4% (Table 1). Satellite observations over the past 20 years combined with studies of climate models have confirmed that such increases are occurring worldwide. As the earth warms, the atmosphere will be able to absorb more water, and with more water in the air, the trend of increasing heavy rainfall will continue.9 In the mountains there is melting snow and rising river levels earlier in the spring, leading to water shortages in summer and autumn, but also floods in winter and early spring, and at the same time due to rising temperatures, winter rains are more rain than snow, which also leads to more floods because of a combination of rain and melting snow at the same time10 . Table 1. Average precipitation in mm in R. N. Macedonia, 2008–2020

Rains

Meteo station

2008–2013

2013–2019

2020

Berovo

658.1

700.5

1040.0

Bitola

618.8

666.6

850.7

Demir Kapija

520.6

593.4

813.1

Kriva Palanka

624.3

693.8

966.5

Ohrid

659.0

828.1

693.2

Prilep

471.9

587.2

801.4

Skopje

456.1

502.8

782.9

Stip

431.2

480.5

803.1

From Table 2, it can be seen that in the Republic of N. Macedonia in the period from 2008 to 2020 we have an increase in the days during the year which are followed by rain and snow, and in contrast we have in some places a decrease in the days that were followed by fog. The frequency and intensity of floods in the last few years in N. Macedonia are increasing. According to statistics, the floods were caused by the outflow of the large rivers Vardar, Crna Reka, Strumica, Pchinja, Lepenec and Bregalnica. In the period 1989–2006, 44% of all disasters were floods or flood-related occurrences. 9 CCSP(Climate Change Science Program), 2008. 10 Knowles, N., M.D.Dettinger and D.R.Cayan, 2006. Trends in Snow fall versus Rainfall. Journal

of Climate.

116

A. Petrovski et al. Table 2. Average rainy, snowy and foggy days in R. North Macedonia, 2008–2020

City Berovo

Bitola

Demir Kapija

Kriva Palanka

Ohrid

Prilep

Skopje

Stip

2008–2013

2014–2019

2020

Rain

88.4

71.2

107

Snow

20.6

18.4

23

Fog

7

5.2

10

Rain

99.8

90

134

Snow

20.8

20.8

29

Fog

20.8

17.6

13

Rain

82.6

69

74

Snow

10.6

9.6

30

Fog

28.2

24.4

3

Rain

112.2

99.6

162

Snow

29

24.4

16

Fog

13.8

12.0

11

Rain

115

93.8

145

Snow

14

14.6

10

Fog

3

1.6

10

Rain

101.8

80.4

124

Snow

22

20.2

25

Fog

12.4

6.8

16

Rain

99.6

92.8

155

Snow

12.8

18.4

19

Fog

15.6

12

9

Rain

89.4

82

127

Snow

11.8

12.8

43

Fog

15.8

10.4

9

Global climate change is affected by various drought-related factors. There is a high certainty that rising temperatures will lead to more rainfall than snow, earlier melting of snow, and increased evaporation and transpiration. Because of this, the risk of hydrological and agricultural droughts increases as temperatures rise. In the Mediterranean and Balkan countries, prolonged droughts and extremely high air temperatures are common. In such ecologically unfavorable conditions for sustainable development of vegetation, the amount of active water in plants decreases rapidly, which increases the risk of occurrence and rapid spread of fires. Their geographical location, which conditions hot and dry climates, plays a crucial role in the occurrence of large-scale forest fires in the Balkans and the Mediterranean.

Climate Change Challenges and Security Implications

117

In the Republic of N. Macedonia over 3000 forest fires were recorded in the period between 1999–2021 in which almost 150,000 hectares of forest and forest land were burned, which caused direct or indirect damage estimated at about 100 million euros. Ozone and airborne particles are strongly influenced by weather changes (eg heat waves or droughts). Concentrations of greenhouse gases such as carbon monoxide (CO), methane (CH4), nitrogen oxides (NOx), total suspended particles/dust (TSP), and sulfur dioxide (SO2) prevent solar energy from returning to the atmosphere, which upsets the balance. of received and returned energy from the sun. The greenhouse effect contributes to increased atmospheric pollution, and thus to global warming (Fig. 2).

Fig. 2. Total annual emissions of polluting substances in the air in R.N. Macedonia, 2002–2011

Based on future climate scenarios and in the absence of a reduction in additional greenhouse gas emissions, the International Panel on Climate Change attributes declining air quality and pollution as a result of climate change. Pollution is one of the biggest disruptors of human health. There is almost no organ in the human body that does not suffer from the consequences of pollution, but the lungs and the cardiovascular system are most affected, and the most vulnerable category of people are children up to 5 years, the elderly, the chronically ill, and people with asthma. PIE 2.

Which consequence of climate change mostly reflects on the national security of the Republic of N. Macedonia? No one Tempreature 8 (4%)

Air polution 64 (31%) Forest fires 34 (16%)

10 (5%) Floods 34 (16%) Droughts 58 (28%)

According to the opinion of the citizens of the Republic of N. Macedonia, from the conducted survey for the needs of this paper (pie 2), pollution in the Republic of N.

118

A. Petrovski et al.

Macedonia is the biggest consequence of climate change which reflects on the national security of the country. Out of a total of 208 respondents, 64 or 31% of the respondents think that pollution in their opinion is the biggest danger for them at the moment, 58 respondents, or 28% answered that the drought most affects national security, 34 or 16% think that floods and fires are equally a security threat. Many of the respondents (31%) believe that in the Republic of N. Macedonia there is too much pollution in the larger cities where the majority of the population is concentrated daily, whether for living or performing work responsibilities, and that more specific measures must be taken for 44% believe that droughts and floods in the country are a major threat to security and that they have greatly contributed to the displacement of the population from the agricultural regions of the country to larger cities, in search of work. In the Republic of N. Macedonia annually, approximately 1,350 people die as a result of air pollution, and the national economy suffered a loss of about 253 million euros as a result of premature death, health costs, reduced productivity of the population, and absenteeism of citizens11 . Climate change creates the preconditions for bloody conflicts that will jeopardize the well-being of society and the ability to counteract the effects of global warming12 .

PIE 3. Can climate change cause social danger and crisis in the Republic of N. Macedonia? NO 50 (24%) YES 104 (50%) I don't know 54 (26%) According to the survey conducted for this paper (pie 3), and in terms of whether climate change can cause social danger and crisis in the country, 104 people or 50% of respondents believe that their impact through different types of consequences (pollution, floods, droughts,…) can cause huge material damage, danger, and crisis in the country, 54 or 26% do not know or are not sure, and 50 respondents 24% are sure that climate change does not have the opportunity to cause social danger and crisis and that the country is not largely vulnerable to the impact of climate Change Water supply is one of the basic challenges of our time. By the middle of this century, over 40% of the global population will be living under increased stress from water scarcity. As the population grows, tensions between different water needs will increase. 11 Institute of Public Health of the Republic of N. Macedonia, December 2020. 12 Dan Smith, Secretary General of the independent peacebuilding organization International Alert

and is currently the Director of the Stockholm International Peace Research Institute (SIPRI).

Climate Change Challenges and Security Implications

119

Many analyzes predict that there will be an increase in air temperature of 1–2 °C by 2050 as a result of global warming. In arid regions, this can result in 10% less rainfall and a 40–70% reduction in available water from rivers and lakes. In colder regions at higher altitudes, winter melts can be more intense, causing floods and lower river levels during the summer (Fig. 3). From Fig. 4, it can be seen that the Republic of N. Macedonia in the coming period until 2040 will face a high risk of water shortage. According to the results of the analysis of the organization “Institute of World Resources (WRI)”, based on 12 indicators, such as the availability, extraction, and consumption of water, but also based on specific hydrological components, warns that high risk of running out of water, or with little water in 2040 are 33 countries in the world, including Kuwait, Qatar, Bahrain, Saudi Arabia, Libya and Jordan, and North Macedonia is positioned on the 17th place on the list of countries that have the opportunity by 2040 to face the risk of water shortages.

Low Low to middle Low to high High Extremely High

Fig. 3. Water stress by country: 2040

HIGH level of water scarcity MEDIUM level of water scarcity LOW level of water scarcity

Fig. 4. Water scarcity in the world

120

A. Petrovski et al.

PIE 4. World water usage

Water supply 8% Electrical Power and GAS 7% Production 4%

Households 9%

Agriculture 67%

Minning 2% Others 3% Pie 4 - shows that most of the available water for use in the world or 67% is used for agricultural purposes, i.e., for irrigation of agricultural areas for food production for both humans and livestock.

4 Conclusion The Republic of North Macedonia is not a country that has large quantities of water, i.e., the water surfaces are represented by only 2% of its total area, i.e. 25,713 km2 , 488 km2 of water surfaces. According to an analysis by the Ministry of Environment, water resources in Macedonia are very vulnerable to the effects of climate change, and the consequences of water scarcity for human health and safety will be very serious. The same analysis for the Republic of Macedonia envisages an education of precipitation by 15% by 2050 and a drastic reduction of leakage in all river basins, and the general availability of water is expected to decrease by 18% by 2100, while the dry periods and sudden floods will become more frequent and more intense.13 By the end of the century, the eastern parts of the country will face the most severe water shortage, the average runoff in Vardar will be reduced by 20% compared to 2000, and the projected water losses for the Bregalnica watershed will be 24%. Climate scenarios have also been prepared for the Republic of N. Macedonia using the models of the Intergovernmental Panel on Climate Change, namely A1B-AIM, A1FI-MI, A1T-MES, A2-ASF, B1-IMA and B2-MES whose characteristics describe the current level and the future technological, demographic, economic and sociological development of Macedonia. Estimation of changes in temperature and precipitation is made compared to the period 1961–1990 which is taken as a reference and initial. 13 The analysis of the Ministry of Environment entitled "Water resources and the challenge of

climate change", 2014.

0.8

Low

Summer

1.1

1.5

2.2

3.0

3.8

1.1

1.7

2.7

3.9

5.0

0.7

0.9

1.0

1.1

1.4

0.9

1.3

1.8

2.4

3.0

1.4

1.9

2.7

3.6

4.6

1.4

2.1

3.3

4.8

6.2

1.2

1.6

1.7

1.9

2.4

1.5

2.1

3.0

3.8

4.8

2.4

3.4

4.6

6.2

7.9

Source: Hydro-meteorological service. Climate change scenarios for N. Macedonia, 2012.

0.5

1.5

1.0

Medium

0.8

1.9

Medium High 0.9

Medium Low 0.7

2.4

1.1

High

Spring

Autumn

Yearly

2.7

3.9

5.8

8.2

106

0.7

1.0

1.1

1.2

1.5

1.0

1.3

1.9

2.4

3.0

1.6

2.2

3.0

3.9

5.0

1.8

2.5

3.7

5.2

6.7

0.8

1.1

1.2

1.3

1.6

1.0

1.4

2.0

2.6

3.3

1.6

2.2

3.1

4.2

5.3

1.7

2.5

3.9

5.5

7.1

2025 2050 2075 2100 2025 2050 2075 2100 2025 2050 2075 2100 2025 2050 2075 2100 2025 2050 2075 2100

Winter

Table 3. Projected changes in air temperature for R.N. Macedonia for 2025, 2050, 2075, and 2100, the four seasons (winter, spring, summer, autumn) and annual (yearly)

Climate Change Challenges and Security Implications 121

122

A. Petrovski et al.

Based on the modeling results, the following can be concluded: • It is likely that there will be a steady increase in temperature between 2025 and 2100. • Compared to the period between 1961 and 1990, the projected changes for the period between 2025 and 2100 will be the most intense in the warmest period of the year. • It is possible for the average monthly temperatures during the transition between winter and spring to equalize in this period. • For the period between 2025 and 2100, a decrease in precipitation is forecast, in all seasons and annually, and the largest decrease will be during the summer. • The intensity of the changes is greatest in the warmest part of the year (in July and August, there may be no precipitation at all); and • In the cold period of the year, a decrease in precipitation of up to 40% of the average monthly quantities is predicted. According to the modeling data shown in Table 3, in the period 2025–2100, a continuous increase in air temperature is overlooked. Compared to the period 1961–1990, the projected changes are most intense in the warm part of the year. Summers would be warmer, and the temperature rise more pronounced. An increase in air temperatures in the cold part of the year is also forecast, but with a lower intensity. At the transition from winter to spring, the average monthly temperatures in this period are probably approaching and equalizing. From Table 4, it is likely that in the period 2025 – 2100, there will be a continuous decrease in the amount of precipitation. The forecasted changes will be the most intense in the warm part of the year, so the summers will be drier, and it is likely that some summer months (July, and August) will be without precipitation. Different climatic regions in N. Macedonia will react differently to climate change, and it is predicted that the region in the southeastern part of North Macedonia with a continental climate, which is near Lake Ohrid and Lake Prespa, is predicted to have the weakest response to climate change on a large scale in the context of changes in absolute temperature and precipitation, and the northwestern part which is under the dominant influence of the mountain-alpine climate to have the strongest reaction. Due to the reduction of effective rainfall in N. Macedonia in the coming period, which can be seen from the data in Table 2, and the fact that 84% of available water resources are generated on the territory of Macedonia, mainly from rainfall and snowmelt, a significant reduction of effective rainfall will cause a reduction in available water resources in 2050 and 2100. According to these forecasts, water resources for water supply would be reduced, and due to rising temperatures due to climate change, the demand for drinking water will increase by up to 30%, and the available water resources for the needs of agriculture where is currently used 40 will be reduced. % of the total consumed water in the Republic of N. Macedonia, and in general, the availability of water in the Republic of N. Macedonia is expected to decrease by 18% by 2100. In the end, this means that climate change will have a serious impact on National Security in North Macedonia (Pie 5).

−3

−4

−6

−8

− 10

−1

Medium High −1

−3

Medium Low −4

−5

Summer

− 14

− 11

−7

−3

−2

− 20

− 16

−9

−2

−1

−5

−4

−3

−2

−2

− 12

−9

−8

−6

−5

− 21

− 17

− 13

− 10

−7

− 29

− 23

− 17

− 12

−9

− 25

− 20

− 13

−6

−4

− 48

− 38

− 25

− 15

− 12

− 68

− 54

− 46

− 38

− 29

Source: Hydro-meteorological service. Climate change scenarios for N. Macedonia, 2012.

Low

Medium

High

Spring

Autumn

Yearly

− 80

− 66

− 57

− 47

− 36

−5

−4

−2

−1

−1

− 14

− 11

−9

−7

−5

− 25

− 21

− 14

− 10

−8

− 34

− 27

− 20

− 13

−9

−6

−5

−4

−3

−2

− 14

− 11

− 10

−8

−6

− 25

− 21

− 15

− 10

−8

− 33

− 27

− 19

− 12

−8

2025 2050 2075 2100 2025 2050 2075 2100 2025 2050 2075 2100 2025 2050 2075 2100 2025 2050 2075 2100

Winter

Table 4. Anticipated changes in precipitation amounts (%) for R.N. Macedonia for 2025, 2050, 2075, and 2100, the four seasons (winter, spring, summer, autumn), and annual (yearly). Climate Change Challenges and Security Implications 123

124

A. Petrovski et al.

PIE 5 Which sector will be most affected by climate change in N. Economy Agriculture 22 Macedonia? 4 (2%)

(11%)

Energy system 10 (5%)

Watersupply 4 (2%) Enviorement 49 (23%)

NATIONAL security 47 (23%)

PUBLIC Health 72 (34%)

References 1. Worldwide average of 2.9% of GDP is spent on defense (UNDP) 2. European Commission, Communication from the Commission to the European Parliament, the Council, the European Economic, and Social Committee, and the Committee of the Regions: A policy framework for climate and energy in the period from 2020 to 2030, COM (2014)15 final, Brussels (January 22, 2014) 3. EU, Climate change and EU security policy: An Unmet Challenge, Richard Youngs (May 2014) 4. The Global Security Defense Index on Climate Change sets out how governments around the world view climate change as a matter of national security and how their security agencies have begun to plan for the consequences of climate change 5. https://www.americansecurityproject.org/climate-energy-and-security/climate-change/gsd icc/ 6. Foreign Affairs Council Meeting, “Council Conclusions,” Luxembourg, June 24, 2013; External Action Service, EU Climate Diplomacy for 2015 and Beyond: Reflection Paper, (2013) 7. Assembly of the Republic of Macedonia, National Concept for Security and Defense, (2003) 8. www.ecolife.com/define/climate-change.html#sthash.gaYTryBr.dpuf 9. CCSP(Climate Change Science Program) (2008) 10. Knowles, N., Dettinger, M.D., Cayan, D.R.: Trends in Snow fall versus Rainfall. Journal of Climate (2006) 11. Institute of Public Health of the Republic of N. Macedonia, (December 2020) 12. Smith, D.: Secretary General of the independent peacebuilding organization International Alert and is currently the Director of the Stockholm International Peace Research Institute (SIPRI)

The Frequency of Freeze-Thaw Cycles Across Balkan Peninsula in the Period 1991 – 2020 Vulcho Pophristov(B) , Hristo Chervenkov , Radoslav Evgeniev , Lilia Bocheva , and Dimitrina Todorova National Institute of Meteorology and Hydrology, 66 Tsarigradsko Shousse, Sofia, Bulgaria {vulcho.pophristov,hristo.tchervenkov}@meteo.bg

Abstract. High and low near-surface air temperatures are normal part of the climate of the mid-latitudes and in particular the bigger part of Southeastern Europe. The variations of temperature below and above the freezing point represent socalled freeze-thaw phenomenon, which has essential and important role in many spheres of the science, public life and the infrastructure. Several ways for calculation of freeze-thaw cycles exist. In present study a day with transition through 0 °C is defined as minimum daily temperature below 0 °C and maximum temperature over 0 °C, measured at 2 m above the ground surface in the meteorological stations, according to World Meteorological Organization (WMO) recommendations. Annual and monthly frequencies of the parameter are examined for 103 meteorological stations on the Balkan Peninsula for 30 years period 1991 – 2020. The highest number of days with transition through 0 °C exceed 100 on average annual basis and is observed in stations with continental characteristics in negative terrain forms, favorable for radiation cooling and large daily temperature amplitudes. The values decreases in places with lower latitudes and near the larger water basins, especially close to Adriatic coast, where the average annual values approaching zero. In about 58% of the stations the maximum number of days with transition through 0 °C is observed in January. Around 17% of the stations have its maximum in December and the same also in February. Maximum in April and May have just mountain regions above 1800 m. Performed trend analyses reveals statistical significant annual decreasing tendency in 46% of the meteorological stations in the examined period. Keywords: Freeze-thaw phenomenon · 0 °C transition · Spatial distribution

1 Introduction Freeze-thaw phenomenon or variation of the temperature above and below freezing point is widespread on large areas in mid-latitudes of the continents of the North Hemisphere. The water has a feature to expand its volume by 7 – 9% when freezing and passing into a solid state. When the water in a liquid form infiltrates in the pores of different materials and consequently freezes, it causes concentrating tensile stresses and generates cracks in areas adjacent to these pores [1]. When this ice melts, the water fills the new enlarged volume of the crack and in the next freezing cycle this volume is increasing again. Various © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 125–136, 2023. https://doi.org/10.1007/978-3-031-26754-3_11

126

V. Pophristov et al.

studies examine and evaluate different environmental factors on mechanical, physical, chemical and mineralogical properties of different materials putting them through variety of tests. One of these environmental factors is crossing the freezing point event or socalled freeze-thaw cycles. In [2] is considered, for example the effect of freeze-thaw cycles on the gloss of polished stone surfaces, used as the pavement of buildings in open areas and affected by temperature variations, especially in cold regions. Understanding of damaging mechanisms of salt and ice crystallization processes in pore spaces in natural building stones are reviewed in [3]. The freeze-thaw cycles show to have a considerable influence on the physical and mechanical properties of granites [4]. Another study in the field of the mining industry shows the important impact of freeze-thaw cycles on the rock engineering structure of open-pit mine slopes in high-altitude and cold areas [5]. Freeze-thaw cycles can also damage the mechanical properties of rollercompacted concrete surfaces used in airport runways and thus have serious effect on the flight safety [6]. The problem of degradation of structural concrete over many phase transitions is reviewed in [7]. Other laboratory research reveals sudden degradation of asphalt road mixes causing major and sudden swelling effects when crossing the freezing point [8]. Freezing of the soils can lead to a reduction in infiltration and an increase in runoff possibility, resulting in a greater potential for soil erosion [9]. Results of another experimental study show that greater number of cases with transition through 0 °C are associated with a decrease in the adsorption capacity of phosphorus in brown soils [10]. The global warming affects permafrost zones in northern regions, where periods of thaw increase carbon decomposition rates which leads to the release of carbon dioxide and methane into the atmosphere creating potential for climate feedback [11]. Relatively old Canadian research [12], finds freezing and thawing of water in rock crevices and soil material is an important factor in mechanical weathering. A study [13], shows connection between soil erosion primarily at high latitudes and altitudes and freezethaw cycles. Several studies display freeze-thaw cycles in meteorological point of view researching in detail its annual and monthly frequencies. Most of them refer to North America and except measuring air temperature in meteorological shelter at 2m above the ground, including soil temperature measurements in different depths. Nine methods for calculating frequencies of transition through 0 °C are propounded in [14]. A freezethaw day is defined as a weather observation day with a maximum temperature of 0 °C or above and a minimum temperature of −2.2 °C or below in [15]. The study includes 30 winters 1950 – 1980 at 228 stations in nine northeastern American states. One of the most large-scale works of freeze-thaw cycles in the USA [16] contains data from 1300 meteorological stations. It calculates the days with transition through 0 °C as maximum temperature above 0 °C and minimum temperature below 0 °C in the observation day. Two other studies depict freeze-thaw cycles in North America. The first and more recent study is devoted to the region of Toronto (Canada) and shows dependency between monthly average temperature and freeze-thaw cycles [17] and the second concerns American state Kentucky [18]. The summary of the researches for North American continent show the highest number of days yearly are observed in mountainous regions of the West. Also, the freeze-thaw phenomenon has its highest values in the months with mean temperature close to 0 °C and have two maximums in autumn and in spring at the northern stations, while in warmer climates at south this

The Frequency of Freeze-Thaw Cycles Across Balkan Peninsula

127

bimodal characteristic is replaced with single-peak maximum, most often in January. The Climate Service Center Germany (GERICS) [19] and its research project “IMPACT2C” [20] quantifying projected climate impacts under 2 °C of warming, concerns the freezethaw cycles in European continent [21] with the method defined by [15] with minimum temperature below −2.2 °C and maximum above 0 °C. A few bulgarian researches also affect transition of the temperature through 0 °C or freeze-thaw cycles on the territory of Bulgaria. [22] Shows yearly and monthly frequencies of about 50 meteorological stations across Bulgaria in the period 1930 – 1970, using the method mentioned in [16]. Additionally, the cases with freeze-thaw cycles are divided by its average daily temperature on the days with < 0 °C and days with average daily temperature > 0 °C. Zoning of the territory of Bulgaria and characteristics of some climatic parameters, including crossing of 0 °C, for products for the industry, are reviewed in [23]. The current work aims to reveal annual and intra-annual frequencies and tendency of freeze-thaw cycles across Balkan Peninsula in the last climatological period 1991 – 2020.

2 Data and Methods In present study the definition in [16] is used, as freeze-thaw cycles, or at least one transition of the temperature trough 0 °C in twenty-four hours period is defined as minimum temperature below freezing point (min t °C < 0 °C) and maximum temperature above it (max t °C > 0 °C) in a single observation day. For example, in [21] the other definition is chosen with minimum temperature below (–2.2 °C) and maximum above 0 °C and their argument for this approach is that temperatures between zero and (–2 °C) do not constitute a hard freeze and that impacts on vegetation and infrastructure are modest. The reason to choose the approach in [16] in our research is the fact that temperatures are measured in meteorological shelter approximately 1.5 – 2 m above the ground surface. But depending on the wind, the amount of cloudiness, humidity and other factors, the minimum temperature on the ground surface can be 5 – 6 °C or even lower than in meteorological shelter at 2 m. For example, 0 °C, or even higher temperature in a meteorological shelter could be −5 °C at the ground surface. The data of minimum and maximum daily temperatures of 103 stations across Balkan Peninsula are considered in the study for the period 1991 – 2020. The data for minimum and maximum daily temperature are mainly from Global Surface Summary of the day (GSOD), derived from National Centers for Environmental Information (NCEI), a part of National Ocean and Atmospheric Administration (NOAA) [24]. The other sources of meteorological data used in the present study are Ogimet [25], MeteoManz [26], Rp5 [27] and Daily Climate Weather Data Statistics [28]. The data gaps are filled when possible with data from at least two other sources after verification for the current station and exact date. If there is no data for the current date in the other sources, such a day is excluded from statistics. When more than six days in a month are missing data for minimum and maximum temperature, this month is excluded too. A year with excluded mount is out of the annual statistics too and don’t attend in average annual frequency calculations. The year or winter season with freeze-thaw cycles is counted from July and ends in June like it is shown in [14]. Thus we count a real winter season with freeze-thaw cycles,

128

V. Pophristov et al.

instead of a calendar year, where we could count cases in two different winter seasons – one at the beginning of the year and the other one at its end. The records suspected for errors are checked also in two other different sources when such exist. Data records of neighboring stations are also verified as an additional checkup. Meteorological stations in Bulgaria performed revisions in their metadata for changed locations during the time period of the study and the data from them are used with attention and trend analysis is not done on them. All the stations outside Bulgaria are synoptic meteorological stations with 8 observations in a day. In Bulgaria, there are a couple of climatological stations included in the research, where there are 3 observations in 24 h, but this does not affect the readings of minimum and maximum temperatures. Trend significance and steepness are analyzed with performed Mann-Kendall test and Theil-Sen slope estimator [29–31] on a yearly and monthly basis on the rows of freeze-thaw cycles with a significance level of 5%. Mann-Kendall test is suitable for non-normally distributed data, data containing outliers and is robust [32]. It is recommended by the World Meteorological Organization [33]. The trend test is not applied for stations with more than one consecutive missing year.

3 Results and Discussion 3.1 Annual Frequencies The first assessment of freeze-thaw cycles, or at least one transition of the temperature trough 0 °C in twenty-four hours period is its average annual value. The highest annual number of days with 0 °C transition in the Balkan Peninsula are registered in meteorological stations situated at the bottom of hollow terrains, where there are appropriate conditions for rapid radiation cooling at night, during the synoptical situations with a clear sky and lack of wind in general. In such orographically closed hollow fields with altitude 600 – 800 m a.s.l, large daily temperature amplitudes are realized and hereby, a possibility of frosts not just in the winter, but also in all spring and autumn months as even in the summer months there are days with frost. During the winter such places have more days with positive maximum temperatures, compared for example to mountainous regions, thus increasing additionally the annual number of days with freeze-thaw cycles. In the examined thirty years period in more than 100 meteorological stations across the Balkans, Romanian Miercurea ciuc is the station with the highest average annual number of days – 123. Along with the North Macedonian Berovo with 111 and Bulgarian station Tran with 110 days, these are the only three meteorological stations on the Balkan Peninsula with days with transition through 0 °C more than 100 on average, on annual basis. Miercurea ciuc as the northernmost of these three is the only station on the peninsula below 1000 m. a.s.l with recorded frosts and thus with days with 0 °C transition in all months of the year in the period 1991 – 2020. In the summer months, Tran measured frosts in July. Normally the average annual number of days decreases in lower altitudes near the coasts, especially near the Adriatic coast where these areas are protected from cold air masses by high Dinaric and Pindus mountains. The lowest number of days of all stations has Dubrovnik with less than 2 days yearly on average. For comparison Thessaloniki on the Aegean coast has 24 days and Constantsa on the Black sea coast, not sheltered

The Frequency of Freeze-Thaw Cycles Across Balkan Peninsula

129

by any mountain ridges from cold air intrusions – 37 days. The days with freeze-thaw cycles commonly increase with the altitude, but over 1000 – 1500 m a.s.l. they start decreasing because in the winter months the temperatures at the higher altitudes are constantly below the freezing point and there is no 0 °C transition for long periods. On the highest peaks of the Peninsula over 2500 m, average annual values are increasing again due to more cases in summer and in May and September in general (Omu – 72, Musala – 78, Kredarica – 85). It can be said that the considered parameter strongly depends on the complex mix of factors, as form of the terrain, proximity of great water basins, altitude and latitude, possibility of advection of cold, warm or humid air masses and radiation cooling on every specific location. In concave terrain forms, where the temperature amplitudes are higher, the number of days increases and the opposite, at the convex terrain forms like the peaks, temperature amplitudes are much less and the number of days decreases. For example, Bjelasnica in Bosnia (2070 m. a.s.l.) – 56 days per year and Botev peak (2376 m a.s.l.) – 58, have relatively low average annual values (Fig. 1). In addition to averages, the range of transition through 0 °C parameter or socalled freeze-thaw cycles could be shown also by its minimum and maximum values. Minimum annual values below ten days in a year appeared near Aegean and Adriatic coasts, as they are zero only in narrow line across Croatian, Montenegro, Albanian and Greek Adriatic seaside (Dubrovnik, Tirana, Podgoritsa), where there are winters without frosts. Over 60 days have places favorable for radiation cooling with closed, hollow relief and latitude 600 – 800 m. above sea level, in the inner parts of the peninsula, far from the seas like stations Tran – 84 days, in the western part of Bulgaria, Miercurea Ciuc – 88 in Romania and Berovo – 75 in North Macedonia. Maximum annual values of freeze thaw-cycles vary from 7 in Dubrovnik (2001/2002) and 11 in Split (2004/2005), to 127 in Berovo (2018/2019) and Sevlievo (1991/1992, Bulgaria), 134 in Tran (2018/2019) and 158 (2019/2020) in Miercurea Ciuc. Just six stations have maximum annual values below 50 days. The average maximum annual value for all examined stations is 85, while the average minimum annual is 37 days. It is obvious the distribution of maximum annual values is similar to minimum and average ones, so almost the same stations are at the bottom and at the top of the minimum, maximum and average rows (Table 1). The average for the area of the Balkan Peninsula in the period 1991 – 2020, there are about 60 days yearly with transition through 0 °C. 3.2 Seasonal Distribution Concerning seasons, winter (December, January and February) is with the most cases of freeze-thaw cycles, about 37 days on average for the Balkan Peninsula region. The highest value of almost 64 days is registered in Berovo, several stations in the western part of Bulgaria, south Romania in the Danube valley and Liubliana and Maribor in Slovenia have more than 50 days. The lowest number of days except for the stations near the Adriatic coast (Split – 4, Dubrovnik – 2) have also highest mountainous meteorological stations on the peninsula (Musala, Omu – 6, Botev peak – 11, Bjelasnica – 12, Kredarica – 18 (Fig. 2).

130

V. Pophristov et al.

Fig. 1. Average annual number of days with 0 °C transition on Balkan Peninsula (1991–2020)

In spring, 13 days on average are with transition through 0 °C among 103 meteorological stations in this study. Over 25 days have Miercurea Ciuc – 42 and mountainous stations Ceahlau Toaca (Romania) and Kredarica (Slovenia) – 30 days, Omu – 28 and Musala – 25, also Tran – 28 and Berovo - 26 days. Below 5 days have stations near Adriatic, Aegean and Black sea coasts (Fig. 2). In summer, frosts are observed only in meteorological stations over 1000 m in the mountains and at these situated at the bottom of hollow fields like above mentioned Tran and Miercurea ciuc. The cases in this season strongly increase with altitude in the mountains, Bjelasnica and Ceahlau Toaca – 2, Botev peak – 3, Kredarica – 9, Omu – 10 and Musala – 15 days. All other stations with frosts in summer have average season values of less than 1 day (Fig. 3). In autumn, average values vary from over 20 to almost 40 days in the mountain stations above 2000 m a.s.l. (Musala – 31, Omu – 29, Kredarica – 27) and stations like Berovo, Tran and Sibiu with 21 – 23 days. The highest seasonal value is in Miercurea ciuc – 38 days. Normally the lowest number of days with freeze-thaw cycles below 4 days have stations with low altitudes near the coasts, where in this season still warm sea waters have a great influence on minimizing the days with frosts. About 10 are the average number of days with transition through 0 °C for the autumn season in studied area (Fig. 3). For more than 94% of the stations, maximum is in the winter. Only six stations at the altitude above 1800 m. have maximums in other seasons. In two of them above 2500 m. – Musala and Omu in eastern part of the peninsula, the maximum is in autumn and these two stations are the only ones, where there are more cases with transition through 0 °C in autumn than in spring in examined period. Furthermore, in these two stations only, winter is the season with lowest values. The rest four have its highest values in spring. (Fig. 4-first row). Seasonal distribution of the three studied stations in high hollow fields and three stations near the coasts of Adriatic, Aegean and Black Sea are also presented in the second row of (Fig. 4). On the third row of the figure are depicted stations in the most populated areas in lowlands. In all of them winter is the season with most cases.

The Frequency of Freeze-Thaw Cycles Across Balkan Peninsula

131

Fig. 2. Number of days with transition through 0 °C in winter (left) and in spring (right) in the period (1991 – 2020) on Balkan Peninsula

Fig. 3. Number of days with transition through 0 °C in summer (left) and in autumn (right) in the period (1991 – 2020) on Balkan Peninsula

Fig. 4. Seasonal distribution of average number of days with transition through 0 °C in mountainous stations (first row), stations in high hollow fields and stations near the sea (second row), meteorological stations in lowlands and in most populated areas (third row), in the period (1991 – 2020). (0.0) – indicates seasons with average values less than 0.05. (0) – indicates absence of cases in current season.

132

V. Pophristov et al.

3.3 Monthly Distribution January is the month with most cases – about 13 days average, little less in December – 12 days, February – also almost 12 and 9 days in March. In 58% of the stations maximum of freeze-thaw cycles is in January. These are stations with more southern location or such with lower altitudes in northern part of the peninsula, where in common in January the average monthly temperature fluctuates around 0 °C. 17% of the stations have maximum in December in the area of Croatia, 2 stations about 1000 m a.s.l. in Serbia (Zlatibor and Crni vrah), some stations in north-western part of Bulgaria, Sarajevo and Bania Luka in Bosnia and several stations in eastern Romania. Maximum in February have also other 17% of the stations, situated in Hungary, western part of Romania, Ljubliana and Maribor (Slovenia) and Split in Croatia. Only two stations have maximum in March, Murgash peak (1687 m.) in the western part of the Balkan ridge in Bulgaria and the station with the most annual cases – Miercurea Ciuc. In April maximum have three stations in mountain regions in the range 1800 – 2400 m a.s.l. – Ceahlau Toaca (1858 m a.s.l.) in Carpatian mountains in Romania, Bjelasnica (2070 m.) in Bosnia and Botev peak (2376 m) in Bulgaria. Maximum in May have also three meteorological stations, all above 2500 m a.s.l. and are the highest meteorological stations on Balkan Peninsula – Omu (2511 m) in Romania, Kredarica (2515 m) in Slovenia and the highest peak on the Peninsula – Musala (2925 m) in Bulgaria. Monthly distribution of average number of days with transition through 0 °C in some meteorological stations in the period (1991 – 2020) are presented below at (Fig. 5). In the mountains it’s common that intra-annual distribution have two maximums, one in autumn and one in spring. The most significant difference at highest meteorological stations is most cases in the winter from November to March in Kredaritsa (Slovenia). This distinction is probably due to proximity of Adriatic Sea and relatively easy reaching of warm Mediterranean air masses to these most western parts of Balkans. In altitude range 1800 – 2400 m, the cases decrease with increasing the altitude from November to April. In May and June the cases with transition through 0 °C increase with altitude and are higher at Botev peak. In September and October, values at Bjelasnica are slightly smaller due to its western location, respectively relative proximity to Adriatic Sea and that the transport of air masses is mainly from west to east. Musala as the highest peak on the Peninsula has maximum values from all stations during the summer months. In May – 15, June – 9, July – 4, August – 2 and September – 10 days. Seasonally, Musala and Omu have the most cases in autumn, while the month with highest values in both stations is May. Concerning the stations situated near the bottom of hollow fields, Miercurea ciuc has two maximums (in November and March) like mountain regions, but compared to them, with more cases in the winter. In Tran station these two maximums are still visible (December and March), but values are much higher especially in December and January, due to its southern location and more positive temperatures during the days in the winter. In southernmost of these three stations – Berovo, just one maximum is observed. In the months December – 22, January – 22 and February – 20 days, Berovo has maximum average values from all stations. Miercurea ciuc has the highest average values from all stations in the months March – 23, April – 15, October – 15 and November – 19 days. Observing meteorological stations representative for the coasts of Adriatic (Split), Aegean (Tessaloniki) and Black Sea (Varna), it is obvious the eastern coast of Peninsula have more average number of

The Frequency of Freeze-Thaw Cycles Across Balkan Peninsula

133

days with transition through 0 °C than western coast. Split have cases of frosts only from December to March. In Tessaloniki cases of frosts exist from November to April and in Varna – from October to April. Six meteorological stations in lowland areas below 600 m are also presented at (Fig. 5) from west to east, as in all of them one maximum is observed.

Fig. 5. Monthly distribution of freeze-thaw cycles: above 2500 m; in the range 1800 – 2400 m; at in hollow fields; near the coasts; in valleys and most populated areas on Balkan Peninsula.

3.4 Trend Analyzes The tendency of at least one daily transition of the temperature through 0 °C, or so-called freeze-thaw cycles on annual basis is presented at (Fig. 6). In 46% of the examined stations situated mainly in Bulgaria, Romania, the eastern part of Hungary and northern Serbia, the annual trend is negative and statistical significant. Decreasing of the freezethaw cycles in the stations with statistical significant negative trend is in the interval between 4 and 10 days on every 10 years. In other stations it is around 42%, it is also negative, but statistical insignificant. Just 10% have positive, but insignificant trend and only in two meteorological stations, Miercurea ciuc in Romania – (8.7 days / 10 years) and Sopron in Hungary – (4.9 days / 10 years), the tendency is statistical significant positive. Examining tendency by months, December and January are the only with more stations with positive than negative trends. The reason of this fact is probably the warming of these two winter months and as consequence, many of the stations with average temperatures close to 0 °C have more days with positive maximum temperatures and respectively, more days with freeze-thaw cycles. In transitional season months October, November, March, April and also February almost 90% of stations have decreasing tendency. The warming affects this months rising the temperatures and respectively reducing the days with frosts, but in the mountains the effect is opposite and days with transition through freezing point are rising. Musala in (April, November, December), Kredarica (April) and Botev peak (March) are with statistical significant positive tendency. Statistical significant negative is the trend in May and September in the mountains. In summer months almost only mountainous stations have cases with temperatures below freezing point and the tendency is negative almost everywhere, as an exlusion is westernmost Kredarica in August, with positive, but statistical insignificant tendency.

134

V. Pophristov et al.

Fig. 6. Tendency of annual values of the number of days with transition through 0 °C in the period 1991 – 2020. Stations with: dark blue color – statistical significant negative, light blue – insignificant negative, light orange – insignificant positive, red – significant positive trend. Meteorological stations with changed location and with missing data in more than one consecutive years, or with more than 10% of missing data are not included Table 1. Minumum, maximum, average annual number of days with freeze-thaw cycles, P-value of Mann-Kendall test (P-v) with 5% significance level, Sen-Slope estimator Qm . Bolded stations are with statistical significant trend. na – trend is not calculable Station

Min Max Ave P-v

Musala

53

113

Qm

78 0.986 0

Station

Min Max Ave P-v

Varna

17

77

41

0.001 −1 0.362 −0.18

Omu

43

111

72 0.083 −0.50 Tessalon

5

50

24

Kredarica

59

117

85 0.765 −0.18 Split

0

11

4

Botev

33

84

52

113

91

Bjelasnica 40

56

C. Toaca

55

74

Tran

84

M. Ciuc Berovo

58 0.510 0.16

Liubliana

79 0.720 −0.35 Zagreb

Qm

0.706 0 0.601 0.11

33

80

59

0.297 −0.39

109 0.164 −0.48 Budapest 28

74

55

0.042 −0.5

134

110 na

77

52

0.033 −0.44

88

158

75

127

na

Beograd

18

123 0.015 1.00

Sofia

49

96

70

0.055 −0.50

111 na

Bucuresti

63

103

85

0.474 −0.13

na

4 Conclusions The considered parameter depends of the terrain, proximity of great water basins, altitude, latitude, possibility of advection of cold, warm or humid air masses and radiation cooling on every specific location. The highest annual number of over 100 days with freeze-thaw cycles in Balkan Peninsula are registered in meteorological stations situated at the bottom of closed hollow terrains at 600 – 800 m a.s.l. The lowest annual number of days are registered near the coasts and commonly rise with the altitude, but over 1500 m a.s.l. they start decreasing, because in winter months the temperatures in the higher altitudes are constantly below 0 °C. In the highest peaks over 2500 m, average annual values

The Frequency of Freeze-Thaw Cycles Across Balkan Peninsula

135

are increasing again due to more cases in summer and also in May and September. In 94% of the stations maximum is in winter season, over 2500 m in autumn and in the range 1800 – 2400 m in spring. Between 1800 and 2400 m maximum is in April, over 2500 – in May. In transitional season months October, November, March, April and also February almost 90% of stations have decreasing tendency. The warming affects this months rising the temperatures and respectively reducing the days with frosts, but in the mountains the effect is opposite and days with transition through freezing point are rising. In 46% of the examined stations situated mainly in Bulgaria, Romania, eastern part of Hungary and north Serbia, the annual trend is negative and statistical significant. Acknowledgements. We want to appreciate (NCEI) for the submitted freely available datasets (GSOD) and to Kamelia Nikolova for the assistance provided.

References 1. Wang, S., Chen, Y., Ni, J., Zhang, M., Zhang, H.: Influence of freeze-thaw cycles on engineering properties of tonalite: examples from China. Adva. Civil Eng. 2019, 12 (2019). Article ID 3418134. https://doi.org/10.1155/2019/3418134 2. Ozcelik, Y., Careddu, N., Yilmazkaya, E.: The effects of freeze–thaw cycles on the gloss values of polished stone surfaces. Cold Regions Science and Technology 82, 49–55 (2012). ISSN 0165-232X, https://doi.org/10.1016/j.coldregions.2012.05.007 3. Ruedrich, J., Siegesmund, S.: Salt and ice crystallisation in porous sandstones. Environ. Geol. 52(2), 225–249 (2007). https://doi.org/10.1007/s00254-006-0585-6 4. Martins, L., Vasconcelos, G., Lourenco, P., Palha, C.: Influence of the freeze-thaw cycles on the physical and mechanical properties of granites. J. Mater. Civil Eng. 28(5), (2016). https:// doi.org/10.1061/(ASCE)MT.1943-5533.0001488 5. Xiao, Y., Li, C., Cao, J., Wang, Y., Hou, Z., Hu, N.: Investigation of the effects of freeze-thaw cycles on geomechanical and acoustic characteristics of tuff specimens under different stress paths. Advances in Civil Engineering (2020). https://doi.org/10.1155/2020/6689181 6. Zhang, W., Zhang, J., Chen, S., Gong, S.: Degradation of roller-compacted concrete subjected to freeze-thaw cycles and immersion in potassium acetate solution. Adv. Materi. Sci. Eng. 8 (2018). https://doi.org/10.1155/2018/4282181 7. Norvell, C., Dusicka, P., Sailor, D.J.: The effect of microencapsulated phase-change material on the compressive strength of structural concrete. In Press: J. Green Building 8(3), 116–124 (2013) 8. Mauduit, V., et al.: Sudden degradation of asphalt road mixes during periods of freeze-thaw events: Analysis of field cases and exploratory laboratory research. Bulletin des Laboratoires des Ponts et Chaussées 279, 47–63 (2013) 9. Sinha, T., Cherkauer, K.: Time Series Analysis of Soil Freeze and Thaw Processes in Indiana. Department of Agricultural and Biological Engineering, Purdue University, West Lafayette, Indiana, Journal of Hydrometeorology 9(5), 936–950 (2008). https://doi.org/10.1175/2008JH M934.1 10. Wang, Q., Liu, J., Wang, L.: An experimental study on the effects of freeze–thaw cycles on phosphorus adsorption–desorption processes in brown soil. Royal society of chemistry RSC Adv. 7, 37441 (2017). https://doi.org/10.1039/c7ra05220k 11. Wilson, R., et al.: Greenhouse gas balance over thaw-freeze cycles in discontinuous zone permafrost. J. Geophy. Res. Biogeosci. 122(2), 387–404 (2017). https://doi.org/10.1002/201 6JG003600

136

V. Pophristov et al.

12. Fraser, K.: Freeze-thaw frequencies and mechanical weathering in Canada. Arctic 12(1), 40–53 (1959). http://www.jstor.org/stable/40506957 13. Sun, B., et al.: Effects of freeze-thaw on soil properties and water erosion. Soil & Water Res. 16, 205–216 (2021). https://doi.org/10.17221/143/2020-SWR 14. Baker, G., Ruschy, L.: Calculated and measured air and soil freeze-thaw frequencies. J. Appl. Meteorol. 34, 2197–2205 (1995). https://doi.org/10.1175/1520-0450(1995)034%3c2 197:CAMAAS%3e2.0.CO;2 15. Schmidlin, W., Deither, K., Eggleston, K.: Freeze-thaw days in the northestern United States. J. Climate Appl. Meteorol. 26, 142–155 (1987) 16. Hershfield, M.: The frequency of freeze-thaw cycles. J. Appl. Meteorol. Climatol. 13(3), 348– 354 (1974). https://doi.org/10.1175/1520-0450(1974)013%3c0348:TFOFTC%3e2.0.CO;2 17. Ho, E., Gough, W.: Freeze thaw cycles in Toronto, Canada in a changing climate Theor. Appl. Climatol. 83, 203–210 (2006). https://doi.org/10.1007/s00704-005-0167-7 18. Conner, G.: Freeze thaw events in Kentucky 1948 – 1973. Kentucky climate center Publ. 22 (1979) 19. GERICS Homepage https://www.gerics.de/index.php.en. 26 May 2022 20. “IMPACT2C” project final report Homepage https://www.gerics.de/imperia/md/content/csc/ projekte/impact2c_final_report.pdf. 26 May 2022 21. “IMPACT2C” Freeze-thaw days atlas Homepage - https://www.atlas.impact2c.eu/en/climate/ freeze-thaw-days/?parent_id=22. 26 May 2022 22. Monthly and annual number of days with transition of the temperature through 0C. (minimum t < 0 °C; maximum t > 0 °C). Climate guide of Bulgaria 3, 38–41 (1974). (in Bulgarian) 23. Characteristics and zoning of climatic parameters for industrial products on the territory of Bulgaria (1974). (in Russian) 24. GSOD Homepage. https://www.ncei.noaa.gov/acess/metadata/landingpage/bin/iso?id=gov. noaa.ncdc:C00516 20 March 2022 25. OGIMET Homepage https://www.ogimet.com/gsynres.phtml.en. 27 January 2022 26. MeteoManz Homepage http://www.meteomanz.com/index?l=1. 04 February 2022 27. RP5 homepage https://rp5.ru/. 29 November 2021 28. Daily Climate Weather Data Statistics https://www.geodata.us/weather/. 16 December 2021 29. Mann, B.: Nonparametric tests against trend. Econometrica 13, 245–259 (1945) 30. Gocic, M., Trajkovic, S.: Analysis of changes in meteorological variables using Mann-Kendall and Sen’s slope estimator statistical tests in Serbia. Global Planet. Change 100, 172–182 (2013). https://doi.org/10.1016/j.gloplacha.2012.10.014 31. Sen, K.: Estimates of the regression coefficient based on Kendall’s tau. J. American Statis. Asso. 63, 1379–1389 (1968) 32. Chervenkov, H., Slavov, K.: Theil–Sen Estimator vs. Ordinary Least Squares—Trend Analysis for Selected ETCCDI Climate Indices. Comptes Rendus Acad. Bulg. Sci. 72, 47–54 (2019). https://doi.org/10.7546/CRABS.2019.01.06 33. WMO: Detecting Trend and Other Changes in Hydrological Data; WCDMP-45, WMO/TD 1013. WMO, Geneva, Switzerland (2000)

Assessment of Contemporary Climate Change in Bulgaria Using the Köppen-Geiger Climate Classification Krastina Malcheva(B)

and Lilia Bocheva

National Institute of Meteorology and Hydrology, Tsarigradsko Shose Blvd. 66, 1784 Sofia, Bulgaria {krastina.malcheva,lilia.bocheva}@meteo.bg

Abstract. Understanding climate system shifts has a crucial role in assessing the ecosystems’ stability and resilience of human society to climate change. The territory of Bulgaria belongs to the temperate climate zone of Europe, which has experienced fast warming in recent decades and expected ongoing warming in the future. The Köppen-Geiger climate classification turned out to be an efficient diagnostic tool to evaluate spatial-temporal climate transitions under the changing climate. Our study aims to establish not only the shift between the two reference periods, 1961–1990 and 1991–2020 but also the extent and the tendency of changes at different timescales using observational data and the Köppen-Geiger classification criteria. The sliding window technique was applied on the annual cycles of precipitation and mean air temperature, calculated for all stations from the meteorological network of the National Institute of Meteorology and Hydrology with dense time series in the period 1961–2020. The results show that hotter and dryer subtypes of temperate climates expand their coverage in Bulgaria, corresponding to the reduction of the cooler and cold climate subtypes. Our findings are consistent with other regional studies and illustrate the potential of the presented climate classification-based approach for climate variability assessment and monitoring. Keywords: Climate change · Köppen-Geiger climate classification · Spatial-temporal climate shifts

1 Introduction Recent climate studies projected unprecedented and irreversible climate changes to the end of the 21st century under the worst-case scenarios of anthropogenic forcing of the climate system [1]. The poleward warming shift will affect not only polar climates but today’s hot and arid climates are expected to expand worldwide and to reach climate states with no current analog [2]. Although there is an internationally agreed system of climate change indicators, the climate classification-based approaches receive increasing attention because of their capability to identify the spatial-temporal patterns of the potential biodiversity changes across the world [3].

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 137–148, 2023. https://doi.org/10.1007/978-3-031-26754-3_12

138

K. Malcheva and L. Bocheva

Among the widely used global climate classifications, Köppen’s classification system has become a standard, especially as its modern implementations allow convenient mapping using climate data from the worldwide network of weather stations [4]. Despite some shortcomings and the elaboration of modern alternative climate classification approaches, Köppen’s system remains one of the most widely used in climate change studies [4, 5]. It is often referred to as the Köppen-Geiger classification to designate the final version, published by Rudolf Geiger [6] after the Köppen’s death. Köppen’s climate classification was proposed at the end of the 19th century [7, 8] and further developed over the following decades [9]. Köppen introduced the concept of climate types by looking at bioclimatic similarities between distant areas worldwide. He has established simple meteorological indicators to distinguish vegetation zones and used the same indicators to delineate the boundaries of the respective climate zones through a straightforward classification scheme. The good correlation of the Köppen’s climate types to major vegetation types has been confirmed by many studies [e.g., 10, 11], which makes this climate classification system useful both for reconstructing past climates based on plant fossils and forecasting changes in vegetation induced by shifts in climate types [e.g., 12, 13]. The territory of Bulgaria belongs to the temperate climate zone of Europe, which has experienced fast warming in recent decades and expected ongoing warming in the future [1, 14]. The latitudinal position of the country determines the significant amplitude in the annual course of insolation, resulting in a clearly expressed difference between winter and summer temperatures. Because of the large distance to the Atlantic Ocean, maritime air masses reach the country quite transformed, and the influence of the Black Sea is limited to a narrow area of around 30 km along the coast. Thus, the annual temperature amplitude turns out large enough to characterize the climate as temperate-continental in most of the country [15]. The precipitation regime is determined by seasonal atmospheric circulation patterns, significant diversity of landforms and graduate transition to a more typical Mediterranean climate in the south. As a result, the precipitation is not only unevenly distributed by territory, but also its annual course in the northern and southern regions of Bulgaria is substantially different (with a summer maximum in the north and a winter maximum in the south). All this makes the delineation of climate zones a non-trivial task, especially under the expected climate changes by the end of the century. Although the first classification of the climate in Bulgaria according to Köppen’s system was made as early as 1929, it was used during the following decades in research mostly related to drought conditions. A brief summary, mentioning a map from the climate atlas for the period 1921–1950, is presented in [15]. We have not found newer studies dedicated to a more thorough analysis of the country’s climate through the Köppen-Geiger classification. Therefore, the present study aims to fill this gap, using modern approaches for spatial-temporal analysis to establish the shift between the two reference periods, 1961–1990 and 1991–2020, as well as the extent and the tendency of changes at different timescales.

2 Data and Methods Meteorological information (including metadata) from all stations of the National Institute of Meteorology and Hydrology (NIMH) operating in the period 1961–2020 had

Assessment of Contemporary Climate Change

139

undergone primary expert control. Next, the monthly precipitation and average air temperature values for stations with at least 80% density of time series were calculated according to the WMO requirements [16] using specialized procedures in the SQL environment [17]. The missing monthly values were imputed using the R-package “missMDA” by applying an iterative principal component analysis (PCA) imputation technique, which considers the internal structure of data and the links between individual time series [18]. The calculated monthly and yearly temperature and precipitation normals for the reference periods 1961–1990 and 1991–2020 were interpolated on a regular 30 × 30 arcsec grid. The spatial interpolation method used is regression kriging, which combines a linear regression model with kriging of regression residuals [19]. The altitude and its derivatives at grid points were obtained from the digital terrain model AW3D30 of the Japan Aerospace Exploration Agency [20] using the freely available GIS software QGIS 3.4.9-Madeira [21]. The R-packages “caret” [22] and “gstat” [23] were used for regression analysis and spatial modeling. After that, all grid points were processed under the Köppen-Geiger classification scheme using the own-developed R-script. Additionally, a sliding centered window (with a time step of 1, 11 and 31 years) was applied on the calculated annual cycles of precipitation and average air temperature from 108 meteorological stations, relatively evenly distributed on the territory of the country (Fig. 1), to obtain the respective “sliding” Köppen-Geiger classification in the period 1961–2020.

Fig. 1. Locations of the meteorological stations used in the spatial-temporal analysis on the background of the hypsometric map of Bulgaria, prepared by using the digital terrain model AW3D30 of the Japan Aerospace Exploration Agency [20]

All calculations and scripts were made in the free software environment R, version 3.6.2 [24] and RStudio 1.2 [25]. The maps shown in Figs. 1, 2 and 3 were prepared using QGIS 3.4.9-Madeira.

140

K. Malcheva and L. Bocheva

Table 1. Three-level scheme of Köppen–Geiger climate classification (L1, L2 and L3). MAT: mean annual air temperature (°C); Tcold : the air temperature of the coldest month (°C); Thot: the air temperature of the warmest month (°C); Tmon10 : the number of months with air temperature > 10 °C; MAP: mean annual precipitation (mm); Pdry : precipitation in the driest month (mm per month); Psdry : precipitation in the driest month in summer (mm); Pwdry : precipitation in the driest month in winter (mm); Pswet : precipitation in the wettest month in summer (mm); Pwwet : precipitation in the wettest month in winter (mm); Pthreshold = 2 × MAT if > 70% of precipitation falls in winter, Pthreshold = 2 × MAT + 28 if > 70% of precipitation falls in summer, otherwise Pthreshold = 2 × MAT + 14. Summer (winter) is defined as the warmer (cooler) six-month period of October–March and April–September L1

L2

L3

Description

Criterion

Tropical

Tcold ≥ 18 °C

f

– Rainforest

Pdry ≥ 60 mm

m

– Monsoon

Not (Af) & Pdry ≥ 100 – MAP/25

w

– Savannah

Not (Af) & Pdry < 100 – MAP/5

Arid

MAP < 10 × Pthreshold

W

– Desert

MAP < 5 × Pthreshold

S

– Steppe

MAP ≥ 5 × Pthreshold

h

– Hot

MAT ≥ 18 °C

k

– Cold

MAT < 18 °C

Temperate

Thot > 10 °C & Tcold > − 3 °C

w

– Dry winter

Pwdry < Pswet /10

s

– Dry summer

Not (w) & Psdry < 40 mm & Psdry < Pwwet /3

f

– Without dry season

Not (s) or (w)

A

B

C

a

– Hot summer

Thot ≥ 22 °C

b

– Warm summer

Not (a) & Tmon10 ≥ 4

c

– Cold summer

Not (a or b) & 1 ≤ Tmon10 < 4

Boreal

Thot > 10 °C & Tcold ≤ − 3 °C

w

– Dry winter

Pwdry < Pswet /10

s

– Dry summer

Not (w) & Psdry < 40 mm & Psdry < Pwwet /3

f

– Without dry season

Not (s) or (w)

D

a

– Hot summer

Thot ≥ 22 °C

b

– Warm summer

Not (a) & Tmon10 ≥ 4

c

– Cold summer

Not (a), (b) or (d)

d

– Very cold winter

Not (a) or (b) & Tcold < − 38 °C (continued)

Assessment of Contemporary Climate Change

141

Table 1. (continued) L1

L2

L3

Description

Criterion

Polar

Thot ≤ 10 °C

T

– Tundra

Thot > 0 °C

F

– Frost

Thot ≤ 0 °C

E

The Köppen–Geiger classification scheme has a hierarchical structure generating a convenient two- or three-letter code. The description of the classification scheme (following [26] and [27]) is presented in Table 1. The first-level criteria (L1) define the five major climate types: A (tropical), B (arid), C (temperate), D (boreal) and E (polar). The temperature thresholds are based on empirical observations: the 10 °C summer isotherm defining the boundary of E climates approximately correlates with the poleward tree growth limit. The 18 °C isotherm in the coldest month corresponds to the physiological cold tolerance for a broad range of tropical plants and defines the boundary of A climates. The –3 °C isotherm delineates the boundary of persistent snow cover during the cold season at midlatitude, splitting C and D climates. The threshold criteria for B climates are based on the empirical relationships between annual precipitation and average annual temperature, considering the seasonality of precipitation. The second-level criteria (L2) account for the precipitation regime peculiarities. Especially for climates C and D, the second letter indicates whether there is a dry season: “f” for climates without a dry season (Cf and Df), “s” for dry summer (Cs and Ds) and “w” for dry winter (Cw and Dw). The third-level criteria (L3) account for the thermal regime for climates B, C and D. Similarly, the climate types in the mountains are presented as altitudinal belts [28].

3 Results and Discussion Our analysis is focused primarily on evaluating the climatic shift between the two reference periods, 1961–1990 and 1991–2020. The higher spatial resolution used (30 × 30 arcsec) is not an end in itself but a proven approach in contemporary research on climate change for areas with complex orography [13, 28, 29]. We classified all grid points inside the country’s borders. Three major climate types are distinguished in both periods (Fig. 2): 1. Temperate (C), which can be without a dry season, with hot (Cfa) or warm (Cfb) summer, or Mediterranean-type – with dry and hot (Csa) or warm summer (Csb); 2. Boreal (D) – without dry season, with warm (Dfb) or cool (Dfc) summer; 3. Polar (E) – alpine tundra (ET). In the period 1961–1990, temperate climates occupied 92.6% of the country’s territory; Cfb (61%) is the predominant subtype, followed by Cfa (28.7%). The Mediterranean influence is most pronounced in the southern border regions – the Csa subtype

142

K. Malcheva and L. Bocheva

(1%) occurs along the river valleys, and Csb (2%) is spread mainly in the Eastern Rhodopes and Strandzha, including the southernmost coast. Regarding the boreal climates, Dfb (4.9%) has the largest share in the low mountain belt and some closed valleys, while Dfc (1.8%) is typical for the middle mountain belt. The alpine type ET (0.7%) dominated above 2200 m. In the period 1991–2020, climate types C, D and E covered 94.3%, 5.5% and 0.2% of the total area, respectively. The absolute differences between periods seem insignificant, but a substantial reduction of 69% for E climates and 17% for D climates is obtained when accounting for their relative share. Another noticeable change is the inversion in the ratio between the Cfa and Cfb subtypes, namely from 28.7% for Cfa and 61.0% for Cfb in the first period to 59.1% and 30.9% in the second period, respectively.

Fig. 2. Köppen-Geiger climate classification for the territory of Bulgaria for two reference periods 1961–1990 (left panel) and 1991–2020 (right panel).

Differences in the Köppen-Geiger climate classification between reference periods were mapped and summarized (Table 2 and Fig. 3). L3 transitions from warm to hot summer in the temperate climates have the greatest contribution to the climatic shift (31.8%). The share of L2 transitions is 1.7% comprising shifts from the temperate humid subtypes Cfa and Cfb to the drier subtypes Csa and Csb but also the opposite shift from dry to more humid climates (0.2%). Further analysis of transitions from Cfb to Dfb and from Dfb to Dfc subtypes (mostly in southern and southwestern mountain areas) is required to distinguish real shift from interpolation uncertainty [27]. The other L1 transitions, shown in Table 2, consist of shifts from boreal to temperate (2%) and from alpine to boreal (0.5%) subtypes, but as mentioned above, the actual area reduction for the boreal and alpine climates is 17% and 69% relative to 1961–1990. The improved spatial resolution disclosed more details about the spatial patterns of transitions between climate subtypes. Comparing our results with global or regional scale studies showed that the last century pattern of climate zones is largely preserved in Bulgaria, as well as in Europe [28, 30, 31]. As stated by Jylhä et al. [31], during the latter half of the twentieth century, the European climate shifted from colder to a warmer or warmer and drier climate. These shifts have affected much larger areas (12.1%) than changes toward cooler or wetter climates (2.2%). The alpine/tundra climates indicated

Assessment of Contemporary Climate Change

143

Table 2. Transitions in the Köppen-Geiger classification between reference periods 1961–1990 and 1991–2020 and respective area as a percentage of the total area; # denotes both “a” and “b” C (Temperate) Transition L1

D (Boreal) Area

Transition

0.4% Cfb → Dfb

L2

0.4%

Cs# → Cf# L3

Transition

2.0% Dfb → Cfb

1.7% Cf# → Cs#

E (Polar) Area 2.0%

Area 0.5%

ET → Dfc

0.5%

0.0%

0.0%

0.5%

0.0%

1.5% 0.2% 31.8%

Cfb → Cfa

30.6%

Dfc → Dfb

0.4%

Csb → Csa

1.2%

Dfb → Dfc

0.1%

Fig. 3. Differences in the Köppen-Geiger climate classification between reference periods 1991– 2020 and 1961–1990 derived in a 30 × 30 arcsec regular grid; changes consist mainly of L1 transitions to different climate types (E → D, D → C), L2 transitions to drier subtypes (Cf → Cs) and L3 transitions to warmer subtypes (Dfc → Dfb, Cfb → Cfa, Csb → Csa); transitions to cooler/wetter subtypes are colored in blue.

the most considerable percentage change (–30%). In absolute terms, C climates showed the largest increase (9%). Because of Bulgaria’s narrow latitudinal range and southern location, the changes are distributed among fewer climate types and are generally larger:

144

K. Malcheva and L. Bocheva

the affected areas are 35.8% and 0.7%, respectively; the percentage change of alpine climates is –69%; and C climates increased only by 1.7% in absolute terms. The dynamics of long-term shifts play a key role since the variations could impact ecosystems’ resilience. The climate types and subtypes for a given location may shift from one to another at different timescales, and their stability decrease towards shorter timescales [5]. The evolution of the main climate subtypes, calculated by stations’ data at 31-year and 11-year timescales (i.e., approximately at a reference and decade timescale), is shown in Fig. 4 and Fig. 5. The script applying the sliding window technique to the Köppen-Geiger classification scheme generates time series of calculated Köppen’s subtypes for each station, attributed to the central years of the sliding window. The computed frequency of individual subtypes for each central year by all stations is divided by the number of stations to obtain the relative frequency of Köppen’s subtypes.

Fig. 4. Long-term variation of the temperate climate subtypes, obtained by stations’ data in a centered 31-year sliding window; the years on the abscissa indicate the middle of the temporal window, e.g., 1990 corresponds to the period 1975–2005.

The dominating subtypes Cfa and Cfb turn out to be the most variable ones but in opposite directions, which results in a small total change in C climates (Fig. 4). The alpine climates presented by the three high-altitude meteorological stations on peak Musala (2925 m), peak Botev (2384 m) and peak Cherni Vrah (2286 m) show a stable transition to boreal climates at the 11-year timescale for peak Cherni Vrah since the beginning of the 21st century (Fig. 5). While the Köppen-Geiger classification, applied to stations’ data by a sliding 31-year window in the period 1961–2020, revealed the tipping point in the late 1980s for Cfa and Cfb subtypes, the sliding 11-year window finds out climate fluctuations. Decadal transitions between climate subtypes confirm the long-term reduction of boreal climates and the 1970s’ cooling but also revealed cyclicity in the Mediterranean (Csa and Csb) and steppe (Bsk) climates. The year-to-year variations in the Köppen-Geiger subtypes (or so-called “annual climate types”) are used in classification approaches accounting for both average climate types and interannual climate variability [32]. We found 13 key climate types (Bsk, Csa, Csb, Cwa, Cwb, Cfa, Cfb, Dsa, Dsb, Dwa, Dwb, Dfb and ET), which corresponds to the mentioned above significant climate diversity in Bulgaria. The boreal climate subtypes dominate in 1963, 1969, 1980, 1985, 2012 and 2017, distinguished by very cold winters. Dry periods and events, such as the drought in 2000, are well reflected by the enlargement of the Bsk subtype. The Cwa and Cwb subtypes

Assessment of Contemporary Climate Change

145

Fig. 5. Temporal changes in the Köppen-Geiger subtypes, obtained by stations’ data in a centered 11-year sliding window; the chart type is percentage stacked area; the years on the abscissa indicate the middle of the window, e.g., 1990 corresponds to the period 1985–1995.

Fig. 6. Interannual changes in the Köppen-Geiger climate subtypes; the chart type is percentage stacked columns.

prevailed in the rainy 1972, 1989, 2014 and 2015 (Fig. 6). The general tendency in the period 1961–2020 is expressed by the reduction of cold subtypes and expansion of temperate subtypes. The natural precipitation seasonality is represented by the balanced proportion of “w” and “s” subtypes. The appearance of many different subtypes in some years indicated larger regional differences.

4 Conclusions In this study, we attempted to show the potential of climate classification-based approaches for climate variability assessment using the Köppen-Geiger climate classification and modern techniques for spatial-temporal analysis. The climatic shift between

146

K. Malcheva and L. Bocheva

reference periods, 1961–1990 and 1991–2020, as well as the extent and the tendency of changes at different timescales, were analyzed by observational data from the meteorological network of the National Institute of Meteorology and Hydrology. The calculated monthly and yearly temperature and precipitation normals for the reference periods were interpolated on a regular 30 × 30 arcsec grid using the regression kriging method. The improved spatial resolution disclosed more details about the spatial patterns of transitions between climate subtypes. Generally, seven Köppen-Geiger subtypes (Cfa, Cfb, Csa, Csb, Dfb, Dfc and ET) characterized the climate in Bulgaria. The shifts from colder to warmer or drier climates have affected much larger areas (35.8%) than changes toward cooler or wetter climates (0.7%). Alpine climates indicated the most considerable percentage change (–69%), while in absolute terms, temperate climates showed the largest increase (1.7%). Our findings are consistent with other regional-scale studies. A sliding centered window (with a time step of 1, 11 and 31 years) was applied to the annual cycles of precipitation and mean air temperature, calculated for all stations with dense time series, to obtain the respective “sliding” Köppen-Geiger classification. The dominating subtypes Cfa and Cfb turn out to be the most variable ones on the 31-year timescale but in opposite directions, which results in a small total change. Decadal transitions between climate subtypes confirm the long-term reduction of boreal climates and the 1970s’ cooling but also revealed cyclicity in the Mediterranean subtypes (Csa and Csb). The interannual variability of the Köppen-Geiger subtypes corresponds to the significant climate diversity in Bulgaria and confirms the general tendency to the reduction of cold subtypes and expansion of temperate subtypes in the period 1961–2020.

References 1. IPCC: Summary for Policymakers. In: Masson-Delmotte, V., et al. (eds.) Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [in press] (2021) 2. Garcia, R.A., Cabeza, M., Rahbek, C., Araújo, M.B.: Multiple dimensions of climate change and their implications for biodiversity. Science 344(6183), 1247579 (2014) 3. Guan, Y., et al.: Changes in global climate heterogeneity under the 21st century global warming. Ecol. Ind. 130, 108075 (2021) 4. Netzel, P., Stepinski, T.: On using a clustering approach for global climate classification. J. Clim. 29(9), 3387–3401 (2016) 5. Chen, D., Chen, H.W.: Using the Köppen classification to quantify climate variation and change: An example for 1901–2010. Environmental Development 6, 69–79 (2013) 6. Geiger, R.: Klassifikation der Klimate nach W. Köppen. In: Landolt-Börnstein– Zahlenwerte und Funktionen aus Physik, Chemie, Astronomie, Geophysik und Technik, alte Serie, vol. 3, pp. 603–607. Springer, Berlin (1954) 7. Köppen, W.: Die Wärmezonen der Erde, nach der Dauer der heissen, gemässigten und kalten Zeit und nach der Wirkung der Wärme auf die organische Welt betrachtet. Meteorol. Z. 1, 215–226 (1884) 8. Köppen, W.: Versuch einer Klassifikation der Klimate, vorzugsweise nach ihren Beziehungen zur Pflanzenwelt. Geogr. Zeitschrift 6(593–611), 657–679 (1900) 9. Köppen, W.: Das geographische System der Klimate. In: Köppen, W., Geiger, R. (eds.) Handbuch der Klimatologie, Band 1, Teil C., pp. 1–44. Gebr. Borntraeger, Berlin (1936)

Assessment of Contemporary Climate Change

147

10. Poulter, B., et al.: Plant functional type mapping for earth system models. Geoscientific Model Development 4, 993–1010 (2011) 11. Rohli, R.V., Joyner, T.A., Reynolds, S.J., Ballinger, T.J.: Overlap of global Köppen-Geiger climates, biomes, and soil orders. Phys. Geogr. 36, 158–175 (2015) 12. Denk, T., Grimm, G.W., Grímsson, F., Zetter, R.: Evidence from “Köppen signatures” of fossil plant assemblages for effective heat transport of Gulf Stream to subarctic North Atlantic during Miocene cooling. Biogeosciences 10, 7927–7942 (2013) 13. Beck, H.E., Zimmermann, N.E., McVicar, T.R., Vergopolan, N., Berg, A., Wood, E.F.: Present and future Köppen-Geiger climate classification maps at 1-km resolution. Scientific Data 5, 180214 (2018) 14. Twardosz, R., Walanus, A., Guzik, I.: Warming in Europe: recent trends in annual and seasonal temperatures. Pure Appl. Geophys. 178(10), 4021–4032 (2021). https://doi.org/10.1007/s00 024-021-02860-6 15. Stanev, S., Kyuchukova, M., Lingova, S.: The Climate of Bulgaria. BAS Publishing house, Sofia (1991). [in Bulgarian] 16. WMO: Guidelines on the Calculation of Climate Normals. WMO 1203. Publications Board, Chairperson (2017). ISBN 978-92-63-11203-3 17. Brimhall, J., Gennick, J., Sheffield, W.: SQL Server T-SQL Recipes, Fourth Edition, Book Series: Professional and Applied Computing. Apress Berkeley, CA (2015). e-ISBN 978-14842-0061-2 18. Josse, J., Husson, F.: missMDA: a package for handling missing values in multivariate data analysis. J. Stat. Softw. 70(1), 1–31 (2016) 19. Hengl, T., Heuvelink, G.B.M., Rossiter, D.G.: About regression-kriging: From equations to case studies. Comput. Geosci. 33(10), 1301–1315 (2007) 20. JAXA Earth Observation Research Center: ALOS Global Digital Surface Model “ALOS World 3D-30m (AW3D30)”, https://www.eorc.jaxa.jp. Last accessed 31 January 2018 21. QGIS Development Team: QGIS Geographic Information System. Open Source Geospatial Foundation Project, http://qgis.osgeo.org. Last accessed 12 May 2018 22. Kuhn, M.: Building predictive models in r using the caret package. J. Stat. Softw. 28(5), 1–26 (2008) 23. Gräler, B., Pebesma, E.J., Heuvelink, G.: Spatio-Temporal Interpolation using gstat. The R Journal 8(1), 204–218 (2016) 24. R Core Team: R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, https://www.R-project.org. Last accessed 21 November 2021 25. RStudio Team: RStudio: Integrated Development for R. RStudio, Inc., Boston, MA, http:// www.rstudio.com. Last accessed 21 November 2021 26. Kottek, M., Grieser, J., Beck, C., Rudolf, B., Rubel, F.: World Map of the Köppen-Geiger climate classification updated. Meteorol. Z. 15, 259–263 (2006) 27. Cui, D., Liang, S., Wang, D., Liu, Z.: A 1 km global dataset of historical (1979–2013) and future (2020–2100) Köppen-Geiger climate classification and bioclimatic variables. Earth System Science Data 13, 5087–5114 (2021) 28. Rubel, F., Brugger, K., Haslinger, K., Auer, I.: The climate of the European Alps: Shift of very high resolution Köppen-Geiger climate zones 1800–2100. Meteorol. Z. 26, 115–125 (2017) 29. Fridley, J.D.: Downscaling climate over complex terrain: high fine scale (14 years [10]. 2.2 Air Pollution and Meteorological Data Air pollutants of interest in this study were particulate matter with a diameter less than 2.5 (PM2.5 ) and 10 microns (PM10 ), nitrogen dioxide (NO2 ), sulfur dioxide (SO2 ), carbon monoxide (CO), and ozone (O3 ). In addition to air pollutants, data was extracted on temperature and relative humidity. These exposures were measured at one background air quality monitoring station located in the “Hippodrome” Park (xy coordinates: 23.296338, 42.680806). That station is part of the official network of monitoring sites of the Ministry of Environment and Water of Bulgaria and offers the most complete coverage of the 10year period of interest. It is also the only station in Sofia where PM2.5 was measured continuously. Hourly concentrations of air pollutants, temperature, and relative humidity were obtained from the Executive Environment Agency of Bulgaria. After purging the hourly data by removing implausible values and outliers, daily average values were calculated and matched with the health data.

Time Series Analysis of Asthma Hospital Admissions

193

2.3 Statistical Analysis The data characteristics were inspected and summarized using descriptive statistics. Annual and seasonal changes in air pollution and admission counts were plotted to investigate any apparent patterns. Spearman correlations were calculated to analyze associations between air pollutants and meteorological factors. The associations between air pollution and daily hospital admission counts were analyzed using time series regressions separately for the three age groups (0–4, 5–14, > 14 years). Optimal model selection followed an iterative algorithm of re-specification and refitting. First, asthma counts were regressed on each pollutant using a generalized linear model with Poisson distribution. This model was gradually refined to a log-link negative binomial model with Newey-West corrected standard errors. This decision was informed by model diagnostics – overdispersion and zero-inflation tests and examination of autocorrelation (correlogram and partial correlogram plots, deviance residuals). Comparison of zero-inflated vs. standard negative binomial models was made on Akaike information criterion (AIC) and Bayesian information criterion (BIC) values. The incidence rate ratio (IRR) of hospital admission for asthma was estimated in relation to different exposure scenarios. First, models were fitted with same-day pollutant concentrations (lag0). Second, potential delayed effects were investigated by constructing a lag template of 1 to 7 days for each air pollutant, temperature, and humidity (lags1– 7). Different lags were modeled together in an unconstrained distributed lag model to disentangle specific lagged effects. For CO, however, implausible IRR estimates were obtained, therefore some lagged effects were constrained to be equal (i.e., IRR at lag1 = lag2 and lag3 through lag7). The net effect of lags 0–7 was also calculated. Third, the series was smoothed by consolidating the daily exposure data points into lag 3- and 7-day moving averages of pollutant concentrations. In all these models, risk estimates were scaled per 10 µg/m3 increase in pollutant concentrations, except for CO, where the unit increase was 1 mg/m3 . All models were adjusted for flexible splines of the date variable with 69 knots to account for time trend and seasonality, day of the week to account for weekly patterns in air quality and asthma incidence, temperature and relative humidity as categorical variables in deciles to allow for nonlinear effects. Air pollutants were tested one-at-a-time (single-exposure models). Finally, models were fitted with pollutant concentrations at lag0 dichotomized according to the respective short-term guideline values (45 µg/m3 for PM10 , 15 µg/m3 for PM2.5 , 40 µg/m3 for SO2 , 25 µg/m3 for NO2 , 60 µg/m3 for O3 , and 4 mg/m3 for CO) [11]. The risk estimates from these models were used to estimate air pollution-attributable fractions of asthma admissions compared with a counterfactual scenario, in which pollution level was lower than the respective threshold. A population attributable fraction (PAF) was only calculated for pollutants associated with an increased risk according to the formula: P*(RR-1)/1 + P*(RR-1), where “P” is the prevalence of exposure (in this case, proportion of days with mean concentration exceeding the threshold) and “RR” is the risk of asthma hospitalization (i.e., IRR). Results were considered statically significant if the 95% confidence interval (CI) around the point estimate did not include the null value (1 for IRRs and 0 for PAF). All analyses were conducted using StataMP v. 17. For estimating population attributable fractions (PAF), the punaf module in Stata was used [12].

194

A. M. Dzhambov et al.

3 Results 3.1 Description of the Data Table 1 shows descriptive statistics for the air pollutants and meteorological variables. Overall, 3651 days were covered in the time series and the proportion of missing exposure data was negligible. Correlations between air pollutants, temperature, and humidity were high and in expected directions (Table 2). In addition to clear seasonality, a general trend toward decreasing air pollution over time was observed (Fig. 1, lower three panels). Winter crests in the trends of PM2.5 , PM10 , NO2 , SO2 , and CO, and spring crests for O3 notably shifted downwards over the years. Table 1. Descriptive statistics for the exposure variables in the study. Variable

Missing data (N, %)

Mean

SD

Min

Max

PM10

45 (1.23)

44.50

48.82

3.60

601.04

PM2.5

204 (5.59)

26.70

33.61

0.49

485.77

NO2

23 (0.63)

34.68

20.91

0.01

184.89

SO2

25 (0.68)

8.61

6.91

0.01

123.31

O3

26 (0.71)

38.32

20.23

0.39

97.27

CO

45 (1.23)

0.80

0.74

0.00

7.83

Temperature

20 (0.55)

11.81

8.83

−14.39

Relative humidity

20 (0.55)

66.40

12.79

30.50

31.375 98.26

Table 2. Spearman correlations between the exposure variables in the study. Variable

PM2.5

PM10

NO2

SO2

O3

CO

Temp

PM2.5

1.00

PM10

0.88

1.00

NO2

0.69

0.77

1.00

SO2

0.55

0.50

0.42

1.00

O3

−0.43

−0.48

−0.62

−0.27

1.00

CO

0.89

0.92

0.80

0.54

−0.48

1.00

−0.34

−0.20

−0.21

−0.53

0.49

−0.38

1.00

0.15

0.08

0.07

0.12

−0.50

0.28

−0.53

Temperature Relative humidity

Humidity

1.00

Time Series Analysis of Asthma Hospital Admissions

195

Fig. 1. Time trend of hospital admission counts of asthma (top panel) and air pollutant concentrations (lower three panels) in Sofia in the period 2009 – 2018. Note. For better visualization, the daily data were smoothed by plotting the monthly sum of hospital admissions and average pollutant concentrations.

196

A. M. Dzhambov et al.

Over the entire period, there were 2676 hospital admissions in the age group 0– 4 years, 2976 in the group 5–14 years, and 7619 in the group >14 years. Plotting the exposure and outcome data against time, there was a gradual decrease in hospital admissions in young children (Fig. 1, top panel). The admission counts generally spiked in winter months. Overall, seasonal variations in air pollution somewhat resembled the trend for hospital admissions. 3.2 Time Series Analysis Initial diagnostics of model performance showed that a Poisson regression was inappropriate due to overdispersion in the data, according to Deviance and Pearson goodnessof-fit statistics >1.00. Therefore, a negative binomial model was used instead, which improved the dispersion statistics to 14 years. These autocorrelations were 0.65, where H is the average height of the buildings, W – canyon width. Those segments are described, within the model, as road sources, with the emission height adjusted to account for the range of heights of vehicle emissions, and the horizontal turbulence describes the additional traffic-produced turbulence (CERC 2020). In order to describe the road sources more accurately, the Advanced Street Canyon module option in network mode was used. This module can model the channeling of

208

M. Velizarova and R. Dimitrova

flow along a street, can represent asymmetric canyons and the effect of pavements within the canyon, and can calculate the effect of a street canyon on the surrounding area. The complex terrain option applies 3D flow and turbulence field to the dispersion model, which modifies the plume trajectory and its dispersion, in order to account for the disturbances in the air flow, originating from the topography of the area. Dry and wet deposition processes are included.

Fig. 2. Area of simulations with the WRF model – all domains (a) and the innermost domain D3 with 1 km (b); domain used for the ADMS-Urban model with 5 m grid and all simulated roads included.

4 Results and Discussion Model verification was made using monthly mean concentration values of NO2 at the measurement site G. S. Rakovski str. №193. The measurements were made at 2.5 m height. Greenpeace volunteers, with help from the organization Deutsche Umwelthilfe, have installed and assembled small diffusion tubes containing a chemical substance (triethanolamine), which is absorbing the measuring component (NO2 ). After measuring 31 days, the tubes were sent to the Passam AG Laboratory in Switzerland to provide results (LANUV 2015). Full description of the analyses is available in the report of nonprofit civil society organization “For the Earth” (2020). Figure 3a shows the monthly mean concentration of NO2 at the measurement site with location marked on the map (Fig. 3b) near the crossroad between bul. Vasil Levski and G. S. Rakovski str. Numerical simulations were performed for the same period as for the small diffusion tube collection – from 3rd January to 3rd February. The obtained modelled values of NO2 for the baseline and the four different scenarios are presented in Table 1. The concentration for the baseline scenario corresponds to the

Impact of Regulatory Measures on Pollutants Concentration

209

Fig. 3. Monthly mean concentration at the measurement site G. S. Rakovski str. №193 (at 2.5 m height) (a) and location of the site (b).

Table 1. Description of different cases used in simulations. Scenario

NO2 µg/m3

Case 2018 Baseline

61.52

Case 2026 - unaffected traffic

56.99

Case 2026 - 20% reduced traffic

53.39

Case 2030 - unaffected traffic

47.49

Case 2030 – 20%reduced traffic

44.49

real emission inventory and show 61.52 µg/m3 or slight overestimation of the measured value of 57.50 µg/m3 . The impact of regulatory measures on the pollutants’ concentration fields in a deep street canyon can be seen in Table 1. The monthly mean concentration decreases for both years (2026 and 2030), with more considerable reduction of 14 to 17 µg/m3 for the cases with 20% traffic reduction compared to the cases with unaffected traffic, where the reduction is approximately between 4 to 8 µg/m3 for each case, respectively.

Fig. 4. Monthly mean vertical profiles of NO2 concentration at the measurement site without traffic reduction for Case 1 (2026) and Case 2 (2030) (a) and with 20% traffic reduction for Case 1 (2026) and Case 2 (2030) (b).

210

M. Velizarova and R. Dimitrova

The monthly mean vertical profiles (all data for the period are averaged at the given level) of the NO2 concentration at the measurement site for the baseline and the four emission scenarios are presented on Fig. 4. For the cases with 20% traffic reduction (Fig. 4b) the concentration is considerably lower than the baseline (2018) and the two cases with unaffected traffic (Fig. 4a). The differences between the baseline and the two cases for 2026 at the ground-level are approximately 5 µg/m3 with the unaffected traffic and 17 µg/m3 with 20% traffic reduction. For the cases of 2030 the decrease is approximately 10 µg/m3 with the unaffected traffic and 20 µg/m3 with 20% traffic reduction. A slight increase of the concentration at approximately 12 m height might be due to the formation of recirculation zone inside the street canyon. G. S. Rakovski str. is a typical narrow street canyon with averaged ratio of 2.21 that is favorable for vortex formation inside the street (Oke et al. 2017). Hourly monthly mean profiles of NO2 concentration averaged over the entire domain (Fig. 5) can represent the general traffic dynamics for the central part of Sofia city. Different scenarios for 2026 (Fig. 5a) and 2030 (Fig. 5b) are compared with the baseline scenario. The NO2 concentration for the cases with 20% traffic reduction are lower than the concentration for the baseline and the two cases with the unaffected traffic. There are also two strongly pronounced peaks in the concentration at 8 a.m. and between 5 p.m. to 6 p.m., which corresponds to the rush hours related to travel to and from work.

Fig. 5. Hourly monthly mean profiles of NO2 concentration averaged over the entire domain with unaffected traffic for Case 1 (2026) and Case 2 (2030) (a) and with 20% traffic reduction for Case 1 (2026) and Case 2 (2030) (b).

Comparison between the hourly monthly mean profiles of NO2 concentration averaged over the entire domain by weekday for the baseline (2018) and differences between the baseline (2018) and Case 1 (2026) as an example is shown in Fig. 6. The concentrations are lowest during the weekend, which significantly affects the concentration (two times less at the rush hours) and are highest at 8 a.m. for Fridays and between 5 p.m. and 6 p.m. for Wednesdays (Fig. 6a). The difference in concentration is almost 3 times greater using emissions with 20% traffic reduction (Fig. 6c) than the unaffected traffic cases (Fig. 6b). The values are 1.9 µg/m3 with unaffected traffic and about 6.5 µg/m3 with 20% traffic reduction for Wednesdays, where the greatest differences occur.

Impact of Regulatory Measures on Pollutants Concentration

211

Fig. 6. Hourly monthly mean profiles of NO2 concentration averaged over the entire domain by weekday for the baseline (2018) (a), differences between the hourly monthly mean profiles of NO2 concentration by weekday for the baseline (2018) and Case 1 (2026) without traffic reduction (b) and with 20% traffic reduction (c).

Monthly mean horizontal surface concentration field of NO2 for the baseline and differences between concentration for the baseline and both cases (2026 and 2030) without and with 20% traffic reduction are shown in Fig. 7. The highest concentrations of NO2 are along G. S. Rakovski str., bul. Vasil Levski and bul. Patriarh Evtimy and range from 50 µg/m3 to 200 µg/m3 , except at the crossroads, where the values exceeded 200 µg/m3 (Fig. 7a). Wind from NWW and E directions covers more than a half of the cases, and the wind is approximately perpendicular to the middle part of G. S. Rakovski str., between bul. V. Levski and W.Gladstone str.. The heavy traffic on V. Levski and Patriarh Evimiy boulevards and the predominant wind direction are most likely the reason for the very high concentration. The greatest differences for all of the considered cases are located along the same roads and crossroads. For unaffected traffic the differences in the concentration values between baseline and 2026 emission scenario reach 10 µg/m3 (Fig. 7b), for the 2030 emission scenario reach approximately 25 µg/m3 at the crossroads and along the segment between Racho Dimchev str. And Graf Ignatiev str. (Fig. 7c). The differences between the baseline and the scenarios with 20% traffic reduction reach values between 15 (for 2026) to 35 µg/m3 along the roads and above 60 µg/m3 at the two crossroads. We can also conclude that the measures with additional traffic reduction affects more significantly the pollution reduction inside the street canyon.

212

M. Velizarova and R. Dimitrova

Fig. 7. Monthly mean horizontal surface concentration field of NO2 for the baseline (2018) (a), differences between the monthly mean concentration field of NO2 for the baseline (2018) and Case 1 (2026) with unaffected traffic (b) and with 20% traffic reduction (d) and the baseline (2018) and Case 2 (2030) with unaffected traffic (c) and with 20% traffic reduction (e).

5 Conclusions This work is a pilot study and presents the first steps in developing the basis of new methodology. The results show the numerical modelling capabilities in the process of selecting potential measures and giving recommendations for the decision-making institutions. Some general conclusions are difficult to be made based on only one month of simulations, but the main findings in this study are summarized below. • Model verification shows slight overestimation by the model of monthly mean concentration values of NO2 by approximately 7% at the measurement site, located at G. S. Rakovski str. №193 (at 2.5 m height).

Impact of Regulatory Measures on Pollutants Concentration

213

• Additional traffic reduction with 20% leads to two times lower pollution compared to the cases only with restriction on entry into LEZ of most polluting vehicle types. • Two prominent peaks in the NO2 concentration around 8 a.m. and 5 to 6 p.m. correspond to the rush hours related to travel to and back from the offices. • The hourly dynamics of different days of the week is different, with the lowest traffic during the weekend, which significantly affects concentration (two times less at the rush hours). • The greatest effect from the LEZ implementation in the simulated domain occurs along the G. S. Rakovski str., bul. Vasil Levski and bul. Patriarh Evtimiy, and this effect is most prominent at the crossroads between those roads. • In spite of the described measures, expected concentrations remain high at this locations, more than 150 µg/m3 and require further investigation and probably application of more strict measures. Presented study demonstrates the promising abilities of the new methodology, which includes high resolution air quality dispersion modelling (with 5 m) for limited domain with street canyons, coupled with the numerical results from a regional meteorological model and new developed sector emission inventories for road transportation. It is a useful tool that allows for simulation of different scenarios for future emission reduction and can help authorities with decision making. Some deficits of the study also have to be pointed. The lack of ground air quality data to assess the air quality at street level makes more extensive model verification a difficult task. Data from AQS Pavlovo is used to estimate the ratio between NO2 /NOx which is very high – 0.67, and it can be a possible reason for the average monthly value overestimation by the model of 4 µg/m3 . Measurements at more traffic sites are needed to estimate the real traffic structure and ratio between NO2 /NOx . More simulations are necessary to cover the entire 2020 year in order to compare the modelling results with the annual limit values for the protection of human health of 40 µg/m3 . And finally, the extended modelling domain, covering Sofia municipality, will ensure more reliable results for air quality related to the road traffic in the entire highly urbanized area. All these points will be investigated in our future studies. Acknowledgments. This work has been carried out in the framework of the grant № KP-06H54/ (Development of a methodology for air quality and human health risk assessment in urban areas) supported by the Research Fund at the Bulgarian Ministry of Education and Science. We acknowledge the provided access to the e-infrastructure of the NCDSC - part of the Bulgarian National Roadmap on RIs, with the financial support by the Grant No D01-221/03.12.2018.

References AMS. Glossary of meteorology (2020). https://glossary.ametsoc.org/wiki/Rain Biggart, M., et al.: Street-scale air quality modelling for Beijing during a winter 2016 measurement campaign. Atmos. Chem. Phys. 20, 2755–2780 (2020) Builtjes, P.J.H., van Loon, M., Schaap, M., Teeu wisse, S., Visschedijk, A.J.H., Bloos, J.P.: ‘Project on the modelling and verification of ozone reduction strategies: contribution of TNO-MEP’, TNO-report, MEP-R2003/166, Apeldoorn, The Netherlands (2003)

214

M. Velizarova and R. Dimitrova

Burov A. and Brezov D. (2022). Transport Emissions from Sofia’s Streets - Inventory, Scenarios, Exposure Setting, Studies in Systems, Decision and Control, XXX (in the same issue) CERC. EMIT Atmospheric EMissions Inventory Toolkit user guide version 3.4 (2015). https:// www.cerc.co.uk/environmental-software/assets/data/doc_userguides/CERC_EMIT3.4_U ser_Guide.pdf CERC. WRF to Met utility user guide Version 1.4 (2016). https://www.cerc.co.uk/environmentalsoftware/as-ets/data/doc_userguides/WRFtoMet_User_Guide.pdf CERC. ADMS-Urban (Urban Air Quality Management System Version 5.0) (2020). http://www. cerc.co.uk/environmental-software/assets/data/doc_userguides/CERC_ADMS-Urban5.0_U ser_Guide.pdf Chen, F., Dudhia, J.: Coupling an advanced land surface–hydrology model with the penn state– NCAR MM5 modeling system. Part I: model implementation and sensitivity. Mon. Wea. Rev. 129, 569–585 (2001) Dimitrova, R., Velizarova, M.: Assessment of the contribution of different particulate matter sources on pollution in Sofia City. Atmosphere 12, 423 (2021). https://doi.org/10.3390/atmos1 2040423 Dimitrova, R., et al.: Modeling the impact of urbanization on local meteorological conditions in Sofia. Atmosphere 10, 366 (2019). https://doi.org/10.3390/atmos10070366 Egova, E., Dimitrova, R., Danchovski, V.: Numerical study of meso-scale circulation specifics in the Sofia region under different large-scale conditions. Bul. J. Meteol. Hydrol. 22, 54–72 (2017) “For the Earth” non-profit civil society organization. Analysis of data on nitrogen dioxide levels in Sofia (2020) German Report by LANUV. Measurement of nitrogen dioxide in ambient air with passive collectors in NRW. Demonstration of equivalence with the reference method of the European Directive 2008/50/EC and the 39th BlmSchV (2015). https://www.lanuv.nrw.de/fileadmin/lan uvpubl/3_fachberichte/30059.pdf HEI - The Health Effects Institute (2020). https://www.stateofglobalair.org/health/global. Assessed 2 June Hood, C., et al.: Air quality simulations for London using a coupled regional-to-local modelling system. Atmos. Chem. Phys. 18, 11221–11245 (2018) Iacono, M.J., Delamere, J.S., Mlawer, E.J., Shephard, M.W., Clough, S.A., Collins, W.D.: Radiative forcing by long-lived greenhouse gases: calculations with the AER radiative transfer models. J. Geophys. Res. Atmos. 113, 2–9 (2008) Jimenez, P., Dudhia, J., Rouco, J.F.G., Navarro, J., Montávez, J., Garcia Bustamante, E.: A revised scheme for the WRF surface layer formulation. Monthly Weather Rev. 140 (2012). https://doi. org/10.1175/MWR-D-11-00056.1 Kain, J.S.: The Kain-Fritsch convective parameterization: an update. J. Appl. Meteorol. 43, 170– 181 (2004) Kirova, H., Batchvarova, E., Dimitrova, R., Vladimirov, E.: Validation of WRF with detailed topography over urban area in complex terrain. In: Mensink, C., Matthias, V. (eds.) ITM 2019. SPC, pp. 353–357. Springer, Heidelberg (2021). https://doi.org/10.1007/978-3-662-637609_51 Krzyzanowski, M.: Health effects of transport-related air pollution: summary for policy-makers, WHO Regional Office for Europe (2005). ISBN 92-890-1375-3. https://www.euro.who.int/__ data/assets/pdf_file/0007/74716/e86650sum.pdf Mansell, E.R., Ziegler, C.L., Bruning, E.C.: Simulated electrification of a small thunderstorm with twomoment bulk microphysics. J. Atmos. Sci. 67(1), 171–194 (2010) Oke, T.R., Mills, G., Christen, A., Voogt, J.A.: Urban Climates, 1st edn. Cambridge University Press, Cambridge (2017)

Impact of Regulatory Measures on Pollutants Concentration

215

Shin, H., Hong, S.-Y.: Representation of the subgrid-scale turbulent transport in convective boundary layers at gray-zone resolutions. Mon. Weather Rev. 143(1), 250–271 (2015). https://doi. org/10.1175/MWR-D-14-00116.1 Skamarock, W.C., et al.: A description of the advanced research WRF version 3. NCAR Tech. Note NCAR/TN-475+STR, p. 113 (2008) https://doi.org/10.5065/D68S4MVH Stocker, J., Hood, C., Carruthers, D., McHugh, C.: ADMS-urban: developments in modelling dispersion from the city scale to the local scale. Int. J. Environ. Pollut 50, 308–316 (2012) Vladimirov, E., Dimitrova, R., Danchovski, V.: Sensitivity of the WRF model results to topography and land cover: study for the Sofia region; Annuaire de l’Université de Sofia “St. Kliment Ohridski.” Faculté de Physique: Sofia, Bulgaria 111, 87–106 (2018)

Multivariate Statistical Modelling of Urban Air-Quality Stefan Tsakovski1(B)

, Ventsislav Danchovski2 , Reneta Dimitrova2,3 and Plamen Mukhtarov3

,

1 Faculty of Chemistry and Pharmacy, Department of Analytical Chemistry, Sofia University

“St. K. Ohridski”, 1 James Bourchier Blvd, 1164 Sofia, Bulgaria [email protected] 2 Faculty of Physics, Department of Meteorology and Geophysics, Sofia University “St. K. Ohridski”, 5 James Bourchier Blvd, 1164 Sofia, Bulgaria [email protected] 3 National Institute of Geophysics, Geodesy and Geography, Bulgarian Academy of Sciences, Acad. G. Bonchev, Bl. 3, 1113 Sofia, Bulgaria [email protected]

Abstract. The air quality data for the city of Sofia were retrieved for the period 2009 to 2018 and includes daily observations from four automatic monitoring stations (Hypodruma, Pavlovo, Nadejda, Drujba) for the following variables: PM10 , NO, NO2 , SO2 , O3 , air temperature, humidity, wind and solar radiation. The multivariate statistical modelling of pre-treated data set was performed by the use of Principal component analysis (PCA). The PCA reveals three principal components, which are responsible for urban air quality in all monitoring stations explaining between 80 and 84% of total data variance. The first principal component (PC1), explaining between 45 and 51% from data variance is positively associated with PM10 , NOx and SO2 , and negatively correlated with ozone. PC1 describes the elevated concentrations of the primary pollutants during winter and elevated ozone levels during summer. Together with the expressed seasonality, factor scores of PC1 reveal a decreasing trend of all associated pollutants during the investigated period. The PC2 (19 to 21% of explained data variance) reflects the high concentrations of PM10 and NOx during calm summer days. The PC3 (10 to 13% of explained data variance) reflects the elevated concentrations of SO2 during windy days. A significant seasonality and decrease trend for PM10 , NOx and SO2 is observed at all monitoring stations. There is no significant difference in urban air quality between working days and weekends. Keywords: Urban air pollution · Multivariate statistics

1 Introduction Most of the major challenges related to air pollution and its impact on human health unfortunately will be relevant over the next decades. It is noteworthy that more than half of the worlds’ population lives in urban (55%) rather than rural, areas. This distribution is © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 216–222, 2023. https://doi.org/10.1007/978-3-031-26754-3_19

Multivariate Statistical Modelling of Urban Air-Quality

217

expected to reach 68% by 2050, with one in three people living in cities with at least half a million inhabitants (UN 2018). This trend is also noticeable in Bulgaria, note that about 70% of population lives in urban areas, and more than one third lives in the four largest cities: Sofia, Plovdiv, Varna, and Burgas (BNSI 2017). Cities are becoming centres of human activity because people expect to improve the “quality of human life”, but large urbanized regions, on the other hand, are the largest sources of greenhouse gases and air pollutants, and cause changes in the land-use (Romero-Lankao and Doodman 2011). Air pollution is associated with both acute adverse health effects (e.g., higher risk of hospital admissions in people with chronic diseases on days when air quality is poor), as well as with chronic health impacts, even at low exposure levels (HEI 2020). The World Health Organization (WHO) identifies Particulate Matter (PM), nitrogen dioxide (NO2 ), sulphur dioxide (SO2 ), and ground-level ozone (O3 ) as the air pollutants that are most harmful to human health (WHO 2015). A special report by the European Court of Auditors (EU Special Report 2018) which regularly scrutinizes the effectiveness of European Union policies and programs, concludes that action, taken so far to improve air quality, is not sufficiently protecting citizens from pollution. Cities that auditors visited for the report, have made little or no progress since 2009 in reducing particulate matter pollution (Kraków and Sofia). Unfortunately, Bulgaria is leading in number of lost years of healthy life from ambient air pollution per hundred inhabitants in European Union, according to WHO (2012) and European Environment Agency. The complex and multivariate nature of urban air quality data required the appropriate data treatment. The principal component analysis is often used as an approach that can analyze and model air pollution while in parallel extract knowledge in terms of similarities, differences and interdependencies of the studied air quality parameters in cities like Thessaloniki, Helsinki and Hong Kong (Voukantsis 2011, Wei-Zhen 2011). Air pollution is strongly related to meteorological conditions and when the atmospheric conditions are stable, the turbulence is missing and emitted contaminants are accumulated near the surface increasing concentration. The objective of this study is to investigate the association between air quality and meteorological conditions in Sofia city using principal component analysis. The capital of Bulgaria as a study object is a challenging complex urban system, because of its geographical setting (complex terrain and favourable meteorological conditions for high pollution) and presence of several automatic monitoring stations which can provide long time series of measured concentrations of the main pollutants. These type of analysis enrich our understanding and quantify the intricate interplay between processes underlying the lifecycles of pollution.

2 Materials and Methods 2.1 Data Harvesting The air quality data for the city of Sofia were retrieved from the National System for Environmental Monitoring managed by the Minister of Environment and Water through the Executive Environment Agency. The period from 2009 until 2018 years is investigated in this study, and it includes daily observations from four automatic monitoring stations (AMS) Hypodruma, Pavlovo, Nadejda, Drujba for the following variables: PM10 , NO, NO2 , SO2 , O3 , air temperature, humidity, wind and solar radiation. The AMS Pavlovo is

218

S. Tsakovski et al.

located in the intensive traffic area while the other monitoring stations are typical background urban ones. The row data underwent cleaning procedure including removing of negative values (excluding temperature) and removing of outliers using the 3 sigma rule. 2.2 Principal Component Analysis The Principal Component Analysis (PCA) is a multivariate approach to data reduction. The aim is to find and interpret the latent interdependencies between the variables (air pollutants and meteorological conditions) in the data set (Jolliffe 2002). Such variables form new ones, called latent factors or principal components. In addition to discovering the data structure, the PCA data set can be modelled, compressed, classified and visualized on a plane. The main task in PCA is decomposition of the data matrix into two parts—a matrix of factor results and a matrix of factor weights. The factor weights show the participation of each of the original variables in the formation of the main components while the factor results are the coordinates of the objects (sampling days) in the newly formed variables. The determination of the number of significant principal components is based on their eigenvalues and the percentage of explained variation in the data. Before the analysis the variables in pretreated data matrix undergo autoscaling transformation. All PCA calculations were performed in MATLAB R2018b using PLS Toolbox 8.7 (Eigenvector Research Inc, Manson, WA, USA).

3 Results and Discussion The PCA of pretreated data of AMS Hypodruma station reveals three principal components (PCs) explaining above 80% of data variance. Their factor loadings are presented in Fig. 1. The PC1, explaining 47.37% of data variance, is positively associated with PM10 , NOx and SO2 , and negatively correlated with ozone.

Fig. 1. The PCA factor loadings for AMS Hypodruma.

PC1 describes the elevated concentrations of the primary pollutants during winter and elevated ozone levels during summer. The second PC (20.78%) is positively correlated

Multivariate Statistical Modelling of Urban Air-Quality

219

with PM10, NOx and temperature and negatively with the wind. The PC2 reflects the high concentrations of PM10 and NOx during calm summer days. The PC3 (12.22%) is dominated by the SO2 and wind. The third principal component represents the elevated concentrations of SO2 during windy days. The performed PCA of the other three monitoring stations resembles extraction of the principal components with quite similar origins to those discussed above. The results presented in Table 1 resemble the similar data structure of all monitoring stations which is a proof that the presented principal components are “responsible” for urban air quality at all monitoring stations. Table 1. The PCA results of all monitoring stations. Station

PC1

PC2

PC3

Explained variance

Hipodruma

47.73%

20.78%

12.22%

80.36%

Pavlovo

51.19%

19.46%

13.21%

83.86%

Nadejda

47.92%

21.02%

11.90%

80.84%

Drujba

50.62%

20.71%

10.99%

82.31%

The plots of the factor scores derived by PCA give an opportunity to estimate the impact of each principal component to each sampling day. The PC1 scores for Hypodruma station reveals pronounced seasonality characterized with the elevated concentrations of the primary pollutants during winter and elevated ozone levels during summer (see Fig. 2). 15

Scores on PC 1 (47.37%)

10

5

0

2009 2010 2011 2012 2013 2014

-5

2015 2016 2017 2018 95% Confidence Level

-10 500

1000

1500

2000 Sample

2500

3000

Fig. 2. The PC1 scores for AMS Hypodruma station.

3500

220

S. Tsakovski et al.

The scores show a decreasing trend of all associated pollutants during the investigated period. The same results are observed in the PCA analysis of the other three monitoring stations. The PC2 scores plots for all monitoring stations are similar to the presented for AMS Hipodruma (see Fig. 3). In addition to the reflected by PC1 high concentrations of PM10 and NOx the second principal component represents also their elevated concentrations during calm summer days. The PC3 scores plot for AMS Hipodruma do not show pronounced seasonality behavior of SO2 (see Fig. 4). It should be pointed out that there are elevated concentrations of SO2 during the last winter (2017/2018) in the investigated period. This fact is observed at all monitoring stations. 8 2009 2010 2011 2012 2013 2014

Scores on PC 2 (20.78%)

6

2015 2016 2017 2018 95% Confidence Level

4

2

0

-2

-4 500

1000

1500

2000 Sample

2500

3000

3500

Fig. 3. The PC2 scores for AMS Hypodruma.

10 2009 2010 2011 2012 2013 2014

8

2015 2016 2017 2018 95% Confidence Level

Scores on PC 3 (12.22%)

6

4

2

0

-2

-4 500

1000

1500

2000 Sample

2500

3000

Fig. 4. The PC3 scores for AMS Hypodruma.

3500

Multivariate Statistical Modelling of Urban Air-Quality

221

The bivariate PC1 vs PC2 scores and loadings plots are used to identify potential significant difference between urban air quality during working days and weekends. The absence of well-defined groups of similarity (see Fig. 5) is an indication that there is no significant difference in urban air quality between working days and weekends based on data from a typical background urban station.

Fig. 5. The PC1 vs PC2 scores and loadings plots for AMS Hypodruma.

4 Conclusions The PCA reveals three principal components responsible for urban air quality in all monitoring stations (Hypodruma, Pavlovo, Nadejda, Drujba). The PC1 (explaining between 45 and 51% from data variance) describes the elevated concentrations of the primary pollutants during winter and elevated ozone levels during summer. Together with the expressed seasonality, factor scores of PC1 reveal a decreasing trend of all associated pollutants during the investigated period. The PC2 (19 to 21% of explained data variance) reflects the high concentrations of PM10 and NOx during calm summer days. The PC3 (10 to 13% of explained data variance) reflects the elevated concentrations of SO2 during windy days. There is no significant difference in urban air quality between working days and weekends for a typical background urban environment. Acknowledgements. This work has been carried out in the framework of the National Science Program “Environmental Protection and Reduction of Risks of Adverse Events and Natural Disasters”, approved by the Resolution of the Council of Ministers №577/17.08.2018 and supported by the Ministry of Education and Science (MES) of Bulgaria (Agreement № D01-279/03.12.2021).

References BNSI (Bulgarian National Statistical Institute), Population and Demographic Processes 2017 (2018). http://www.nsi.bg/sites/default/files/files/publications/DMGR2017.pdf. Accessed 10 Jun 2022

222

S. Tsakovski et al.

EU Special Report, European Court of Auditors, Air pollution: Our health still insufficiently protected, report N° 23 (2018) HEI (2020). https://srzi.bg/bg/spravki. Accessed 10 Jun 2022 Jolliffe, I.: Principal Component Analysis, 2nd edn. Springer, New York, NY, USA (2002) Romero-Lankao, P., Dodman, D.: Cities in transition: transforming urban centres from hotbeds of ghg emissions and vulnerability to seedbeds of sustainability and resilience: introduction and editorial overview. Curr. Opin. Environ. Sustain. 3, 113–120 (2011). https://doi.org/10.1016/j. cosust.2011.02.002 UN (United Nations), DESA/Population Division World Urbanization Prospects: The 2018 revision. (2018). https://www.un.org/development/desa/publications/2018-revision-of-world-urb anization-prospects.html. Accessed 10 Jun 2022 Voukantsis, D., Karatzas, K., Kukkonen, J., Räsänen, T., Karppinen, A., Kolehmainen, M.: Intercomparison of air quality data using principal component analysis, and forecasting of PM10 and PM2.5 concentrations using artificial neural networks, in Thessaloniki and Helsinki. Sci. Total Environ. 409, 1266–1276 (2011). https://doi.org/10.1016/j.scitotenv.2010.12.039 Wei-Zhen, L., Hong-Di, H., Li-yun, D.: Performance assessment of air quality monitoring networks using principal component analysis and cluster analysis. Build. Environ. 46, 577–583 (2011). https://doi.org/10.1016/j.buildenv.2010.09.004 WHO (World Health Organization), Public Health and Environment: ambient air pollution DALYs attributable to ambient air pollution, (2012) WHO (World Health Organization), Regional Office for Europe, OECD Economic cost of the health impact of air pollution in Europe: Clean air, health and wealth. Copenhagen: WHO Regional Office for Europe (2015)

Transport Emissions from Sofia’s Streets Inventory, Scenarios, and Exposure Setting Angel Burov(B)

and Danail Brezov

University of Architecture Civil Engineering and Geodesy, Sofia 1164, Bulgaria [email protected]

Abstract. The present study aims to demonstrate high quality traffic emissions inventory in Sofia city serving as an input for dispersion modeling and simulations through baseline and scenario development juxtaposed to exposure characteristics of the urban environment. The spatial scope of the study is the compact city of Sofia. The baseline year is 2018 and the scenario development is up to 2030 with several reference years in between. The methods applied include a wide array of data gathering and processing as well as rapid traffic, activity and morphology mapping and modeling steps and techniques. The traffic distribution model takes into account diverse characteristics of the street network and the spatial development of Sofia. The utilized traffic data is from various sources with poor integrity. In order to obtain a relatively good estimate of the average annual daily traffic for the entire city street and road network, data interpolation and machine learning regression models are tested through QGIS and Python. The relatively large portion of missing data requires pre-processing (filtering) and phased imputation using a well-trained multivariable regression (we found the Random Forest Algorithm to be an excellent choice), preferably with optimally selected parameters. As the latter may differ in the analysis of the main traffic arteries, the primary and the secondary street network, we study each of the clusters separately, allowing given predictions to propagate down the hierarchy. Further on the space-time activity model stems from land use and functional analysis of points of interest with differentiation of the presence of mobility modes and people throughout time. The urban morphology and surface modeling makes use of the Street Canyon Tool from CERC implemented in ArcMap environment as well as the UMEP plugin and other native tools in QGIS. The specific emission scenarios falling under predefined general assumptions are calculated through CERC EMIT and rely on both fleet composition changes and urban plan provisions. The paper raises an array of issues for discussion over the results through the link between emissions generation and exposure settings changing under different conditions as a preliminary step towards spatially and temporally differentiated health impact evaluation. A major conclusion is that the specific configuration of urban street canyons and traffic load in the city impose varying degrees of impacts and risks from the transport sources which need to be addressed by more specific scope and design of measures that can prevent public health in an efficient and equitable manner. Keywords: Traffic · Inventory · Scenarios

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 N. Dobrinkova and O. Nikolov (Eds.): EnviroRISKs 2022, LNNS 638, pp. 223–233, 2023. https://doi.org/10.1007/978-3-031-26754-3_20

224

A. Burov and D. Brezov

1 Introduction Most urban dwellers of cities today in global terms are exposed to dangerous levels of air pollution from a cocktail of sources and chemical species, fine and ultrafine particles. This continues to be one of the major challenges for the urban environment in the light of public health [1].The Sustainable Development Goals address directly the challenge through two goals, three targets and related indicators [2], namely: a) Goal 3 “Ensure healthy lives and promote well-being for all at all ages” with Target 3.4 “By 2030, reduce by one third premature mortality from non-communicable diseases through prevention and treatment and promote mental health and well-being” and the Indicator 3.4.1 “Mortality rate attributed to cardiovascular disease, cancer, diabetes or chronic respiratory disease”, as well as with Target 3.9 “By 2030, substantially reduce the number of deaths and illnesses from hazardous chemicals and air, water and soil pollution and contamination” and the Indicator 3.9.1 “Mortality rate attributed to household and ambient air pollution”; b) Goal 11 “Make cities and human settlements inclusive, safe, resilient and sustainable” with Target 11.6 “By 2030, reduce the adverse per capita environmental impact of cities, including by paying special attention to air quality and municipal and other waste management”, and the Indicator 11.6.2 “Annual mean levels of fine particulate matter (e.g. PM2.5 and PM10) in cities (population weighted)”. The indicators are monitored in general through various statistical and environmental screening tools, but the more fine resolution picture for every particular city is only available if ambitious scientific methods and applied techniques are in place to link statistics, inventories, models, measurements, surveys, etc. into integrative reliable instruments for support of planning and decisions on the path of achieving the global, national and local goals. The study of the background levels of emission concentrations has advanced a lot in the last decades but the fine grained picture for many cities isn’t that close to become available in due time because of lack of reliable measurement, inventory and modeling systems in place. Different tiers of inventories as well as physical and health modeling techniques exist that can predict relatively well at a reasonable cost the dispersal and exposure to pollutants [3]. At the base of such approaches is an array of geospatially referenced data sets. These can vary depending on the tools and techniques but it can be said that more comprehensive databases can be much more precise in revealing the patterns and how they are overlaying and interplaying. South-East European cities are the most affected by the air pollution problem in Europe and are among the World hotspots [4, 5]. Many of the Bulgarian cities are at the top of the list of polluted cities in the EU and it is important to say that this is not due to heavy industry as it was in the past but to domestic heating and transportation [6]. Sofia, the capital and biggest city with its almost 1.5 million citizens and visiting commuters and tourists is the most polluted capital in the EU. There are many underlying cumulative reasons both natural and anthropogenic which overlap as various sources and make its situation complicated and challenging for understanding and adequate policy making. The amalgam of building morphology and street network patterns with varying degrees of roughness and porosity of the urban tissue shape the many types of urban street canyons and other situations of poor ventilation. The growing polycentricity, the densification of secondary mixed use centers and parallel contrasting advantaged and disadvantaged new peripheral suburban development which is often called ‘muddy’ with missing public works, altogether add

Transport Emissions from Sofia’s Streets

225

to the factors behind pollution and its distribution in the cityscape. The lagging behind and low quality public rail and road works and maintenance provoke the deepening car dependence and the vicious cycle of growing motorization rate along with many critiques to the development of the public transport and neglected shared and light mobility and access modes of travel. The mixture of sources leads to high concentration of air pollution and numerous exceedances of the limits throughout the last decade that provoked debates on where to start from in resolving the issue, which sources of pollution to blame for, to regulate and invest into them to comply, hopefully following the results of reliable and focused research. Various modeling [7] and apportionment measurement [8] studies have shown different picture of the contribution from the domestic heating, transportation and other sources depending on the methods, data and spatial scope applied. Several generations of official inventories of the traffic emissions fall in the first tier as they do not step on the basis of actual transport model and rely on very rough estimates [9]. Bias is unavoidable when there are many data and knowledge gaps and lack of cross validation or shared baselines used. The question remains open in terms of more comprehensive studies and verification investigating the changing pollution patterns in the city space. This piece of research which is linked to other experiments and future steps in development provides first insights in an integrated approach and methods starting from steps of provision of a baseline towards more in-depth modeling, verification and evaluation of impacts.

2 Research Framework, Materials and Methods The more general aim of the research is the link between air pollution, health and urban development while here in the focus is the improved inventory of transport emissions with the help of rapid traffic modeling, the elaboration of scenarios for the fleet in the context of implementation of measures and spatial development, and the exposure setting relying on estimates from a model of activities in Sofia city. Further on the research framework includes measurements; numerical experiments in the field of dispersion modeling; epidemiological survey and longer term spatial development scenarios with their health related aspects. One of the biggest challenges at the base is the quality of the underlying traffic and fleet data. It comes from various sources with poor access to the more complete sets, no integrity and many methodological mismatches in the data gathering approaches. It is a typical practice led fragmentation stemming from the sectorized institutional silos. The scarce public access to traffic count data and the lack of actual municipal level transport model are the results from this fragmentation. There are rich data sets made available through the open data access from a municipal enterprise [10]. With the help of own data gathering and various processing steps for matching of diverse geometry, attribute data and methods the research tries to bridge the available resources. Thus the need for time and cost efficient approach is pressing and the answer is through performing the most steps by using public accessible data and open source software which has the ambition to demonstrate a pathway which out beats the public spending boulevard to the urban knowledge arena for the air pollution discourse and contemporary choices.

226

A. Burov and D. Brezov

2.1 Materials and Methods for Rapid Traffic Modeling and Emission Inventory As key input for traffic modeling we use data for key boulevards and junctions, various types of streets in Sofia from several sources at our disposal: measurements provided by the municipality and its planning enterprise from single days in 2017 and 2018 for almost 40 junctions, car count geolocation sample along the street network for the entire 2018 and 2019 (part of the street network) [11, 12], previous and next traffic count campaigns from consultancies [13, 14] covering sets of primary or secondary streets, and the Open Transport Model [15]. For the visualization and analysis of the available data we use both QGIS tools and Python modules. Several additional factors are considered, including space syntax (choice and integration) performed with depthMapX, functions distribution (points but also buildings and areas of interest) and population density with coefficients of the motorized share of users’ and dwellers’ mobility mode. After applying properly tuned heat maps in order to ‘smear’ the point data across the grid, we select an optimal set of features based on correlation. The data is then conveniently normalized, calibrated and clustered in preparation for the ML techniques we resort on in the following stages of our work. 2.2 Materials and Methods for Fleet Inventory and Scenario Development The fleet baseline inventory steps on one dimensional fleet stratification available as input data, distributed by general vehicle typology, type of fuel, engine size, EURO category, year of production and availability of filters [16, 17]. This available data is then juxtaposed through an Analytical Hierarchical Process to the CERC EMIT most complete and differentiated database with the help of Excel for intermediate calculations. Thus the matching is performed with minimum expected deviation due to the association and uncertain distribution for some sub-categories. The NAEI2014 Urban 2014 version 2 is used as the closest possible database to the emission performance of the fleet in Sofia and Bulgaria at the baseline year 2018. We are assuming that the older fleet in Sofia corresponds to that previous year in the database. It consists of almost 550 sub-categories based on multi-dimensional stratification including the fuel types (Zero Emissions Vehicle, Petrol, Diesel and LPG), Vehicle types (Passenger vehicles, LGV N1 (I, II, III), HGV (rigid, artic) (diesel), Bus, Coach (diesel), Moped and Motorcycle (petrol)), engine volume and vehicle weight types (2.5 tonnes ( HMW (43%) > LMW (11%) for Sofia and MMW (57%) > HMW (39%) > LMW (4%). A similar pattern of distributions were obtained for both cities for July 2021: MMW (46%) > HMW (31%) > LMW (23%) for Sofia and MMW (50%) > HMW (30%) > LMW (20%) for Burgas.

Fig. 3. Mean PAHs concentration: (a) LMW, (b) MMW, (c) HMW and (d) PAHs

PAHs and Black Carbon in Urban Air Particulate Matter in Bulgaria

267

To find the correlation between PM2.5 , BC, BrC and PAHs the regression analysis was performed for the whole data set (Table 1). A strong correlation was obtained for PM2.5 and BC, PM2.5 and PAHs for Sofia (0.7), and PM2.5 and BC for Burgas (0.8). A positive weak correlation is observed for PAHs and BrC (0.3) for both sites. Despite the strong correlation between PM2.5 and BC for Burgas, a weak correlation between PAHs and PM2.5 (0.33) is observed. This is probably due to the influence of other sources around the sampling region for which we have no information. Table 1. Correlation coefficients between PM2.5 , PM associated BC, BrC and PAHs

3.3 Diagnostic Ratio for Source Identification of PAHs in PM2.5 Diagnostic ratios (DR) between PAHs are commonly used as a tool for identifying pollution emission sources [15, 20, 26]. The obtained data for six DR are listed in Table 2. In the present study, the values obtained for the Burgas and Sofia suggest that vehicle exhaust emissions, fossil fuel combustion and biomass combustion source were the major contributors depending on the season. Higher ratios for most DR in summer than in winter indicate enhanced photochemistry [21]. An exception is BaA/(BaA+Chr) ratio demonstrating lower values for summer period. In general both PAHs, i.e. BaA and Chr, are persuasive markers of coal combustion [27], which is atypical for both regions during this season, and respectively their concentration are extremely low, especially that of BaA being at the edge of the limit of detection and quantitation. The BaP/(BaP+Chr) ratio has been reported to be successful in discriminating vehicle emissions with values ~0.73 and ~0.5, respectively, for gasoline and diesel exhausts [26]. In this study this ration vary from 0.29 (Oct 20_Burgas) to 0.42 (Jul_21 Burgas and Sofia) indicating diesel vehicles as main source. This is confirmed from the IndP/BghiP ratio for Oct 20 in both cities as well. The contribution of grass/wood/coal combustion to the PAH is more pronounced in BaP/BghiP and IndP/BghiP ratios for both cities. The BaP/BghiP and IndP/BghiP ratios are >1 for Burgas and Sofia in cold period.

268

E. Hristova et al.

Table 2. Diagnostic ratio (the sources indicating values are taken from [21, 22] and the presented values for Burgas and Sofia are calculated in this study) Diagnostic Ratio

Indicator Source

0.5 grass/wood/coal combustion 0.42-0.53 road dust unburned petroleum 0.5 grass/wood/coal combustion 0.35-0.7 diesel vehicles IndP/BghiP