Advances In Geosciences (A 4-volume Set) - Volume 30: Planetary Science (Ps) And Solar & Terrestrial Science (St) : Planetary Science (PS) and Solar and Terrestrial Science (ST 9789814405744, 9789814405737

Citation preview

A d v a n c e s

i n

Geosciences Volume 30: Planetary Science (PS) and Solar & Terrestrial Science (ST)

8474hc.v30.9789814405737-tp.indd 1

24/7/12 4:23 PM

ADVANCES IN GEOSCIENCES Editor-in-Chief: Kenji Satake (University of Tokyo, Japan) A 5-Volume Set Volume 1: Solid Earth (SE) ISBN-10 981-256-985-5

A 6-Volume Set

Volume 2: Solar Terrestrial (ST) ISBN-10 981-256-984-7

Volume 17: Hydrological Science (HS) ISBN 978-981-283-811-7

Volume 3: Planetary Science (PS) ISBN-10 981-256-983-9

Volume 18: Ocean Science (OS) ISBN 978-981-283-813-1

Volume 4: Hydrological Science (HS) ISBN-10 981-256-982-0

Volume 19: Planetary Science (PS) ISBN 978-981-283-815-5

Volume 5: Oceans and Atmospheres (OA) ISBN-10 981-256-981-2

Volume 20: Solid Earth (SE) ISBN 978-981-283-817-9

A 4-Volume Set

Volume 21: Solar Terrestrial (ST) ISBN 978-981-283-819-3

Volume 6: Hydrological Science (HS) ISBN 978-981-270-985-1

Volume 16: Atmospheric Science (AS) ISBN 978-981-283-809-4

Volume 7: Planetary Science (PS) ISBN 978-981-270-986-8

A 6-Volume Set Volume 22: Atmospheric Science (AS) ISBN 978-981-4355-30-8

Volume 8: Solar Terrestrial (ST) ISBN 978-981-270-987-5

Volume 23: Hydrological Science (HS) ISBN 978-981-4355-32-2

Volume 9: Solid Earth (SE), Ocean Science (OS) Volume 24: & Atmospheric Science (AS) ISBN 978-981-270-988-2 Volume 25: A 6-Volume Set Volume 10: Atmospheric Science (AS) Volume 26: ISBN 978-981-283-611-3

Ocean Science (OS) ISBN 978-981-4355-34-6 Planetary Science (PS) ISBN 978-981-4355-36-0 Solid Earth (SE) ISBN 978-981-4355-38-4

Volume 11: Hydrological Science (HS) ISBN 978-981-283-613-7

Volume 27: Solar Terrestrial (ST) ISBN 978-981-4355-40-7

Volume 12: Ocean Science (OS) ISBN 978-981-283-615-1

A 4-Volume Set

Volume 13: Solid Earth (SE) ISBN 978-981-283-617-5

Volume 28: Atmospheric Science (AS) & Ocean Science (OS) ISBN 978-981-4405-67-6

Volume 14: Solar Terrestrial (ST) ISBN 978-981-283-619-9

Volume 29: Hydrological Science (HS) ISBN 978-981-4405-70-6

Volume 15: Planetary Science (PS) ISBN 978-981-283-621-2

Volume 30: Planetary Science (PS) and Solar & Terrestrial Science (ST) ISBN 978-981-4405-73-7 Volume 31: Solid Earth Science (SE) ISBN 978-981-4405-76-8

A d v a n c e s

i n

Geosciences Volume 30: Planetary Science (PS) and Solar & Terrestrial Science (ST)

Editor-in-Chief

Kenji Satake

University of Tokyo, Japan

Volume Editor-in-Chief

Anil Bhardwaj

Space Physics Laboratory, Vikram Sarabhai Space Centre, India

Andrew Yau

University of Calgary, Canada

World Scientific NEW JERSEY

•

LONDON

8474hc.v30.9789814405737-tp.indd 2

•

SINGAPORE

•

BEIJING

•

SHANGHAI

•

HONG KONG

•

TA I P E I

•

CHENNAI

12/7/12 5:26 PM

Published by World Scientific Publishing Co. Pte. Ltd. 5 Toh Tuck Link, Singapore 596224 USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601 UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE

British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library.

ADVANCES IN GEOSCIENCES A 4-Volume Set Volume 30: Planetary Science (PS) and Solar & Terrestrial Science (ST) Copyright © 2012 by World Scientific Publishing Co. Pte. Ltd. All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage and retrieval system now known or to be invented, without written permission from the Publisher.

For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is not required from the publisher.

ISBN 978-981-4405-66-9 ISBN 978-981-4405-73-7

(Set) (Vol. 30)

Typeset by Stallion Press Email: [email protected]

Printed in Singapore.

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-fm

EDITORS

Editor-in-Chief:

Kenji Satake

Volume 28: Atmospheric Science (AS) and Ocean Science (OS) Editor-in-Chief: (AS) Chun-Chieh Wu Editor-in-Chief: (OS) Jianping Gan Editors: (AS) Kevin K. W. Cheung Hyun Mee Kim Tieh-Yong Koh Mong-Ming Lu Seon-Ki Park Editor: (OS) Minhan Dai Volume 29: Hydrological Science (HS) Editor-in-Chief: Gwo-Fong Lin Editors: Kwan Tun Lee Sanjay Patil Srivatsan Vijayaraghavan Volume 30: Planetary Science (PS) and Solar & Terrestrial Science (ST) Editors-in-Chief: (PS) Anil Bhardwaj Editor-in-Chief: (ST) Andrew W. Yau Editors: (PS) Takashi Ito Paul Hartogh Editors: (ST) Yusuke Ebihara Susan Mckenna-Lawlor Gang Lu

v

July 13, 2012 9:4

vi

ADGEO

9in x 6in

Editors

Volume 31: Solid Earth Science (SE) Editor-in-Chief: Ching-Hua Lo Editors: Yih-Min Wu J. Gregory Shellnutt

b1389-fm

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-fm

REVIEWERS

The Editors of Volume 30 would like to acknowledge the following referees who have helped to review the manuscript publish in this volume: P. Senthil Kumar Ciska Kemper Munetaka Ueno Janet Luhmann Wing-Huen Ip

Markus Fraenz Murray Dryer Susan McKenna-Lawlor Duggirala Pallamraju Hui Tian

vii

July 13, 2012 9:4

ADGEO

9in x 6in

This page intentionally left blank

b1389-fm

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-fm

PREFACE

The present volume set (volumes 28 to 31) of Advances in Geosciences (ADGEO ) is the sixth round of ADGEO series edited by the Asia Oceania Geosciences Society (AOGS), and contains papers presented at the eighth annual meeting held in Taipei in 2011. The AOGS is an international society legally registered in Singapore, aiming to cooperate and promote discussion on studies of the Earth and its environment, as well as the planetary and space sciences. To achieve this objective, the AOGS has held its annual meetings since 2004. The AOGS has six sections, Atmospheric Sciences (AS), Hydrological Sciences (HS), Ocean Sciences (OS), Planetary Sciences (PS), Solar and Terrestrial Sciences (ST), and Solid Earth Sciences (SE). In the current set, papers presented at AS and OS sections are included in volume 28, those at HS section are in volume 29, at PS and ST sections are in volume 30, and at SE section are in volume 31. ADGEO is not a scientific journal, but a monograph series or proceeding volumes of the AOGS meetings. Only papers presented at the AOGS meetings are invited to ADGEO series, and are published after peer reviews. The first (volumes 1 to 5), second (6 to 9), third (10 to 15) sets corresponded to the second, third and fourth AOGS annual meetings. The fourth volume set (16 to 21) included papers presented at the fourth and fifth annual meetings, and the fifth set (22 through 27) included those at the sixth and seventh meetings. As a young scientific society, AOGS needs to develop ways to promote information exchange and interaction among scientists in Asia and Oceania region, in the era of internet. Until we establish a journal or other means of publication, ADGEO is expected to serve as a publication tool among the AOGS members and society at large. Finally, I would like to thank authors, reviewers, volume editors and volume editors-in-chief for their timely efforts to publish the current

ix

July 13, 2012 9:4

x

ADGEO

9in x 6in

b1389-fm

Preface

volume set, Meeting Matters International (MeetMatt) for developing and maintaining a system for submission, review and editorial processes, and World Scientific Publishing Company (WSPC) for the editorial, publication and marketing processes.

Kenji Satake Editor-in-Chief

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-fm

PREFACE TO PS VOLUME

The eighth Annual Meeting of the Asia Oceania Geosciences Society (AOGS) was held in Taipei, Taiwan in 2011. Roughly one-sixth of the total number of papers at this annual meeting was presented in different Planetary Science (PS) sessions. The present PS volume of Advances in Geosciences (ADGEO) contains some of the highlights reported in the PS sessions papers at the Taipei annual meeting. The papers cover different planets in the solar system, satellites, as well as minor bodies and ice. The papers deal with observation, modeling, laboratory measurements, simulations, instrumentation, and missions. A wide range of planetary sciences activities are covered in these papers, including surfaces, atmospheres, ionospheres, exospheres, magnetospheres, and solar wind interaction. This issue will be a quality contribution that would serve the purpose of disseminating science to the planetary science community. On behalf of the Planetary Science volume editorial team, I greatly appreciate the efforts of the referees in providing timely and careful reviews.

Anil Bhardwaj PS Section Editor-in-Chief

xi

July 13, 2012 9:4

ADGEO

9in x 6in

This page intentionally left blank

b1389-fm

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-fm

PREFACE TO ST VOLUME

Solar terrestrial science is concerned with the interactions between the Earth, the Sun, and interplanetary space and the resulting effects these interactions induce at the Earth such as those collectively known as Space Weather. This volume comprises recent research results and invited reviews on solar terrestrial research with an emphasis on work undertaken in the Asia Oceania region. Aspects of Solar Terrestrial research covered include: solar flares and active regions, the solar wind and the interplanetary medium, the forecasting of space weather, and its effects on global navigation satellite system. Observational measurements, modeling and theoretical developments are included.

Andrew W. Yau ST Section Editor-in-Chief

xiii

July 13, 2012 9:4

ADGEO

9in x 6in

This page intentionally left blank

b1389-fm

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-fm

CONTENTS

Editors

v

Reviewers

vii

Preface

ix

Preface to PS Volume

xi

Preface to ST Volume

xiii

Asymmetric Cratering on the Moon: Numerical Result From a New NEA Flux Model Takashi Ito

1

Diffuse Interstellar PAH Emission in the LMC Observed with the AKARI/IRC H. Umehata, I. Sakon, T. Onaka and D. Kato

15

Formation of an Extended Halo of Hot Oxygen Atoms in the Wake Region of Venus Ying Liao and Wing Huen Ip

25

Interaction of Solar Wind with Moon: An Overview on the Results from the SARA Experiment Aboard Chandrayaan-1 Anil Bhardwaj, M. B. Dhanya, R. Sridharan, Stas Barabash, Futaana Yoshifumi, Martin Wieser, Mats Holmstr¨ om, Charles Lue, Peter Wurz, Audrey Schaufelberger and Kazushi Asamura

xv

35

July 13, 2012 9:4

xvi

ADGEO

9in x 6in

b1389-fm

Contents

Plasma Transport Processes in the Topside Martian Ionosphere Tariq Majeed, Stephen W. Bougher and S. A. Haider

57

The 3D Analysis of the Heliosphere Using Interplanetary Scintillation and Thomson-Scattering Observations B. V. Jackson

69

Forecasting Transient Heliospheric Solar Wind Parameters at the Locations of the Inner Planets B. V. Jackson, P. P. Hick, A. Buffington, J. M. Clover and M. Tokumaru Recent Progress of Solar Weather Forecasting at NAOC Han He, Huaning Wang, Zhanle Du, Liyun Zhang, Xin Huang, Yan Yan, Yuliang Fan, Xiaoshuai Zhu, Xiaobo Guo and Xinghua Dai A New Approach for Identifying Ionospheric Gradients in the Context of GAGAN System Ravi Chandra Kudala

93

117

127

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch01

Advances in Geosciences Vol. 30: Planetary Science and Solar & Terrestrial Science (2011) Ed. Anil Bhardwaj c World Scientific Publishing Company


ASYMMETRIC CRATERING ON THE MOON: NUMERICAL RESULT FROM A NEW NEA FLUX MODEL TAKASHI ITO National Astronomical Observatory of Japan [email protected]

The asymmetric cratering on satellites is generally related to the synchronous rotation of satellites. On the Moon, the asymmetric distribution of craters has been ascribed to the impacts of the near-Earth asteroid (NEA) population. However, the observed rayed crater distribution’s asymmetry on the Moon stared from a debiased NEA population is significantly more pronounced than what had been predicted by previous numerical studies. This suggests the existence of an undetected population of slower (low impact velocity) projectiles. In this paper, as an extension of our previous trials, we carried out numerical simulations of the orbital evolution of NEA-like particles generated from a new NEA flux model which contains substantial amount of highinclination component as well as close-Earth component. We tried to determine their impact flux on the Moon and resulting asymmetric distribution of craters. The new model is considered to be closer to “true” distribution of NEAs than the conventional NEA flux model is. As a result we obtained slightly enhanced degree of cratering asymmetry from the new model. But it is not quite different from what the conventional model had yielded: The discrepancy between the observational crater record remains. Existence of more, slower objects is still implied from the current result.

1. Introduction Many planetary satellites are locked in synchronous rotation, and their mean rotational angular speed and mean orbital motion is in a 1:1 commensurability. The synchronous rotation of these satellites leads to asymmetric spatial distribution of impact craters on these satellites: The leading hemisphere tends to have more craters than the trailing hemisphere, as is observed on the Galilean satellites of Jupiter, on Neptune’s moon Triton, and on the Moon around the Earth.1–5 Particularly, the asymmetric cratering on the Moon6, 7 is quite interesting because it reflects the steadystate of modern near-Earth asteroids (NEAs) impact flux recorded on 1

July 13, 2012 9:4

2

ADGEO

9in x 6in

b1389-ch01

T. Ito

morphologically young and fresh craters with bright rays, called rayed craters.8 In the paper by Morota and Furumoto, the observed ratio of crater density (D > 5 km) at the apex to that at the antapex is shown to be ∼1.65.6 In addition, there is a recent report that small seismic events observed by the Apollo mission can be used to obtain information of the current lunar bombardments with small magnitude.9 In the paper by Kawamura et al., the number density ratio of the seismic events of roughly 1.4–1.9 has been reported between leading and trailing sides.9 The degree of the leading/trailing asymmetric crater distribution on a synchronized satellite orbiting its mother planet is a function of satellite’s orbital velocity and the average relative velocity between projectiles and the satellite–planet system. When a satellite with a synchronous rotation has a large orbital velocity around its mother planet, or when the average relative velocity between projectiles and the planet–satellite system is small, the asymmetric distribution of craters becomes the most remarkable. Smaller orbital velocity of the satellite, or larger average relative velocity of projectiles, tends to diminish the asymmetry of crater distribution. For the purpose to quantitatively test the hypothesis that impacts from the NEA population account for the observed asymmetric crater distribution on the Moon, in the past we had simulated numerically the spatial distribution of impacts of NEAs, using a numerical model with a steady-state population of impactors based on current estimates of debiased NEA population.10 Starting from the population of NEAs that had been through a debiased processing process,11, 12 we had compared the results of the simulation with the observed asymmetry of the population of rayed craters on the leading/trailing hemispheres of the Moon. Our numerical simulation had yielded a leading/trailing hemispherical ratio of ∼1.32 for lunar impacts by NEAs, which is only marginally compatible with the observed ratio of ∼1.65 found by the geological observation.6 For a comparison test, we carried out another set of numerical integrations of the raw, not debiased population of NEAs, expecting to contain more slower objects that can produce higher asymmetric cratering than the debiased population.13 However the resulting asymmetry turned out to be ∼1.37, not as high as the observed asymmetry deduced from the rayed crater record. A possible explanation for the discrepancies is that there exists a hitherto undetected population of small objects whose average impact velocities on the Moon are much lower than the average impact velocity of the known NEA population. Other explanations are possible, including the possibility that a more comprehensive study of young lunar craters could reveal a

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch01

Numerical Result from a New NEA Flux Model

3

smaller leading/trailing asymmetry and thereby remove the discrepancy with the dynamical modeling. In this paper, as an extension of our previous studies,10, 13 we carried out yet another set of numerical integrations of an NEA population including a different type of component: Particles with higher inclination and smaller semimajor axis. The population is created through a synthetic NEA model that is based on the most credible basis of NEA dynamics and observation to date. Our numerical experiment in this paper will serve as a check as to how differently debiasing models work on changing the impact velocity distribution and asymmetric impacts of the Earth/Moon colliding projectiles. In Sec. 2 we describe our model, method, and our choice of initial conditions. Our results on NEA encounters and collisions with the Earth– Moon system are given in Sec. 3. This section includes the result about the NEA impact fluxes, impact velocities and their spatial distribution on the Moon. In Sec. 4, we compare our numerical result with the actual observation record. Section 5 goes for some discussion.

2. Initial Conditions and Numerical Model Our numerical model follows that in the paper by Ito and Malhotra, having two stages.10 In the first stage, our numerical integrations include the eight major planets and the Sun, and a large number of test particles with NEA-like orbits (Fig. 1). We numerically integrate their orbital evolution (a)

(b)

(c) 0.15

0.10

0.06

relative fraction

0.08 0.10 0.04 0.06

0.04 0.05

0.02

0.02

0

0 0

1

2

3

semimajor axis [AU]

4

5

0 0

0.2

0.4

0.6

eccentricity

0.8

1

0

30

60

90

orbital inclination [degree]

Fig. 1. Initial distribution of the osculating orbital elements of the NEA population in our numerical model of the first stage. (a) Semimajor axis, (b) eccentricity, and (c) orbital inclination. The solid lines are for the population A particles, and the dashed lines are for the population B particles.

July 13, 2012 9:4

4

ADGEO

9in x 6in

b1389-ch01

T. Ito

for up to 100 million years. Throughout these integrations, we record all close encounters of the particles that reach the Earth’s activity sphere (see Sec. 3 for more detail). We use this record in our second stage of numerical simulation, in which we adopt the restricted N -body model consisting of the Earth, the Moon, the Sun, and cloned test particles within the Earth’s activity sphere. In the second stage, we do not include the effects of any planets except the Earth but we include the Moon’s gravity. Our aim and numerical method are similar to those in what was published in previous literatures as numerical14 or analytical15, 16 work, but we believe our model is more realistic and straightforward. For our first stage numerical simulation in this paper we used two different populations of NEA-like particles. Both from the synthetic, “debiased” NEA population models, but one of them is a conventional model, and the other is a revised one. The conventional NEA model (hereafter called the population A) was devised in the paper by Bottke et al.12 This is also the model that we consulted as standard in our previous studies.10, 13 The NEA population described by this model is assumed to be continuously supplied from five intermediate source regions: the ν6 secular resonance in the main asteroid belt, the 3:1 mean motion resonance at 2.5 AU, the intermediate source Mars-crossers, the outer main belt, and the trans-Neptunian disk. This model is established by taking a linear combination of the (a, e, I) distributions from each of the source regions with best fit parameters based on the Spacewatch observation. The set of the population A particles in this paper has an orbital distribution that obeys the histograms shown in Fig. 12 of the paper by Bottke et al.12 which gives the debiased orbital distribution of the NEA of absolute magnitude H < 18. We produced 18,000 particles along with this distribution and used for the numerical integrations described in the next sections. On the other hand, there is another numerical NEA model which we hereafter call the population B. This model is not yet officially published, but mentioned and described in detail in the paper by Moon et al.,17 referred to as a model by “Morbidelli (2006, personal communication)”. Basically this new model is an updated version of the conventional model with the help of observational bias correction of NEAs,18, 19 adding two more high inclination sources such as Hungaria (1.77 < a < 2.06AU, I > 15◦ ) and Phocaeas (2.1 < a < 2.5AU, above the ν6 resonance) to the five intermediate source regions used in the conventional NEA model in the paper by Bottke et al.12 For this paper 18,000 particles with H < 18 were created along with

July 13, 2012 9:4

ADGEO

9in x 6in

Numerical Result from a New NEA Flux Model

b1389-ch01

5

the revised NEA model by the courtesy of A. Morbidelli for the authors, and these particles were used for the numerical integrations described in the next sections. Note that from the result of our previous study using the raw NEA population,13 we are aware that the orbital distributions of the fainter NEAs, such as H > 18, are different from what the brighter NEA component yields. However, the diameter range of the crater record that we are concerned is rather large, >5–10 km, which roughly corresponds to the brighter population of the current NEAs. Hence in this paper we do not consider the fainter components of NEAs than H = 18 in our numerical calculation. This criterion, however, can be of course changed depending on discovery and detection of more and more NEAs in various size ranges through large survey programs in the near future. The orbital element distribution of the particles (18,000 each) belonging to the NEA populations A and B are shown in Fig. 1. In the panel (c) the excess of high inclination component in the population B is obvious, indicating the evident inclusion of the high inclination particles belonging to Hungaria and Phocaeas. Also, it is clear that the particles with larger semimajor axis (such as a > 2.5AU) are less frequent in the population B than in the population A, leading to a fact that there are more particles around a ∼ 1AU in the population B. This difference is due to the difference in observational debiasing used in the new NEA model,18, 19 and will eventually have an influence on the difference in cratering asymmetry between the two populations (see Sec. 4 for detail). For the numerical orbit integration scheme of these particles, we used the regularized mixed-variable symplectic method.20 The basic framework of our first stage simulation follows the work of Ito and Malhotra21 : When a test particle approaches within the physical radius of the Sun or that of planets, we consider the particle to have collided with that body and lost from the NEA population. Also, when the heliocentric distance of a test particle exceeds 100 AU, the particle is considered lost. Over the 100 Myr length of the simulation, a large fraction of the both populations would be expected to be removed in this way, and if this loss were not compensated, we would not be able to mimic a steady-state NEA flux. We realize the steady-state NEA flux in our numerical simulation as follows: for each “lost” particle, we immediately introduce in our simulation another particle with the original position and velocity of that “lost” particle. For example, when the particle i is removed from the simulation by any of the reasons described above at the position ri with the velocity v i , another particle,

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch01

T. Ito

6

also denoted by the subscript i, at the position r i,0 and with the velocity v i,0 is immediately introduced in the simulation where ri,0 and v i,0 are the initial position and the velocity of the particle i at the beginning of the integration.

3. Particle Encounters with Earth’s Activity Sphere In our first stage numerical simulation described above, we recorded the encounters of particles at the Earth’s activity sphere (∼144 Earth radii) over the 100 Myr integration, and found a large number (several ten millions) of encounters with the Earth’s activity sphere. In our simulation, average encounter velocities of the particles at the Earth’s activity sphere are 22.36 km/s for the population A particles, and 22.25 km/s for the population B particles (see Fig. 2 for the encounter velocity distribution). We think the number of the encounters is large enough to establish an orbital distribution function of the particles that can be used to create “clones” of particles in order to increase the reliability of the collision statistics between the particles and the Earth or the Moon. Using the particle encounters at 0.08

relative fraction

0.06

0.04

0.02

0

0

10

20

30

40

50

encounter velocity [km/s] Fig. 2. Fractional encounter velocity distribution at the Earth’s activity sphere. The solid line is for the population A particles, and the dashed line is for the population B particles.

July 13, 2012 9:4

ADGEO

9in x 6in

Numerical Result from a New NEA Flux Model

b1389-ch01

7

Earth’s activity sphere, we generated cloned particles by perturbing the encounter position r and velocity v of each of the original particles so that their initial trajectories at the activity sphere become slightly different: r clone = (1 + δr )r original and v clone = (1 + δv )v original , where δr and δv are random numbers in the range [−0.1, 0.1]. This procedure produces a large number of particles that obey nearly the same orbital distribution function as the original particles but with somewhat different paths toward the Earth and the Moon. We repeated this cloning procedure five hundred times for all the results of the first stage numerical integrations, generating a large number of particle initial conditions on the Earth’s activity sphere. Using these sets of cloned particles, we performed a second set of numerical integrations, this time with the restricted N -body problem including the Sun, the Earth, the Moon, and the cloned test particles. Here we did not include the effect of other planets than the Earth, but we included the Moon’s gravity. All the cloned particles started near the Earth’s activity sphere, and were integrated until they hit the Earth or the Moon or went out of the sphere. We used the present orbital elements of the Moon with true anomaly randomly chosen from 0 to 360◦ for each of the 500 sets of clones. We employed the regularized mixed-variable symplectic method again with a stepsize of 84.375 seconds (= 2−10 days). For the population A, the second stage calculations within the activity sphere of the Earth yielded 1,509,364 collisions with the Earth and 73,923 collisions with the Moon. For the population B, we have 1,155,955 collisions with the Earth and 64,604 collisions with the Moon. The ratio of the number of collisions with the Earth and those with the Moon is found to be 20.4 for the population A, and 17.9 for the population B. For comparison, we note that it has been reported that the ratio of collisional cross sections of the Earth and the Moon becomes ∼23 by assuming isotropic collisions and average impact velocity of Earth-crossing asteroids to be 16.1 km/s on the Earth.22 Figure 3 shows the distribution of impact velocities of the clones on the Earth and on the Moon. Overall, the average impact velocities of the clones on the lunar surface (22.41 km/s for the population A, and 22.28 km/s for the population B) are almost the same as the average encounter velocity of the original particles at the Earth’s activity sphere. This means that lunar gravity plays only a minor role in accelerating particles to the lunar surface in our numerical model.

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch01

T. Ito

8

(a)

(b) 0.06

relative fraction

0.08

0.04

0.04 0.02

0

0

10

20

30

40

50

0

0

10

20

30

40

50

impact velocity [km/s]

Fig. 3. Distribution of impact velocity of the cloned particles on the Moon (a) and on the Earth (b). The solid lines are for the population A particles, and the dashed lines are for the population B particles.

4. Particle Collisions with the Earth and the Moon In order to compare the distribution of impacts in our numerical model with the actual lunar crater record, first we have to consider a correction to the raw numerical results due to the systematic difference in the impact velocities on the leading and trailing hemispheres, a difference that owes to the orbital motion of the satellite about its mother planet. For a satellite with synchronous rotation, the average impact velocity of projectiles is somewhat larger on the leading side than on the trailing side. This difference means that the apparent crater sizes would be larger on the leading side than on the trailing side (assuming the projectile size-frequency distribution (SFD) is not different on the two sides), resulting in the apparent increase (shift) in crater density on the leading side.23 The magnitude of the shift depends upon the relationship between the impact velocity vimp and the crater size, D. From the results of our second stage simulation, we computed the average impact velocity, vimp  in km/s, of NEAs on the lunar surface as a function of angle from apex, γ (degrees), by a least squares fit (see Fig. 4 for the raw impact velocity data for the lunar collision). We find vimp  = 22.9 − 0.00540γ for the population A, and vimp  = 22.6 − 0.00333γ for the population B. This indicates that difference of vimp  between the γ = 90◦ point and the apex (γ = 0) or antapex (γ = 180◦) is less than 0.486 km/s for the population A, less than 0.300 km/s for the population B. Compared with the average of vimp over the entire range of 0 ≤ γ ≤ 180◦ , these velocity differences amount to
2.12% for the population A, and

July 13, 2012 9:4

ADGEO

9in x 6in

Numerical Result from a New NEA Flux Model

b1389-ch01

9

Fig. 4. Dependence of the projectile impact velocity at the lunar surface (vimp , km/s) on the angular distance from apex (γ, degree). (a) For the population A, and (b) for the population B.


1.35% for the population B, which are quite small in their effect on the crater number density change. This small difference is owed to the fact that the lunar orbital velocity of ∼1 km/s is much lower than the average impact velocity vimp . As a result of this small dependence, apparent change of the crater SFD due to the impact velocity difference between the leading side and the trailing side is quite modest. Including this correction to our second stage simulation, we computed the simulated spatial density of NEA impacts on the Moon. Normalizing to unity at antapex, our simulation results for the crater density as functions of apex angle are shown in Fig. 5, panels (a) and (b). In Fig. 5, we used a simple sinusoid with the function form of α + β cos γ for a fitting curve where α and β are fitting parameters, normalizing α + β cos 180◦ = 1. For comparison, panel (c) shows the distribution found from the analysis of observed lunar rayed craters.6 Note that the number of the lunar rayed craters in the observational data analyzed by Morata and Furumoto6 is only 222, while we have some 60,000 to 70,000 impacts in our simulations. This difference is reflected in the difference of the errorbar magnitudes in Fig. 5. Examining Fig. 5, what we notice is that the apex/antapex asymmetry is less prominent in the numerical results (panels (a) and (b)) compared with the observed lunar rayed crater record (panel (c)). The maximum crater density at apex is about 1.65 (normalized to unity at antapex, and estimated

July 13, 2012 9:4

ADGEO

9in x 6in

T. Ito

10

(a)

2

relative crater density

b1389-ch01

(b)

2

1.8

1.8

1.6

1.6

1.4

1.4

1.2

1.2

1

1

0.8

0.8 0

30

60

90

120

150

180

0

30

60

(c)

2

90

120

150

180

angle from apex [deg]

1.8

1.6

y z

1.4

x

1.2

1

0.8 0

30

60

90

120

150

180

angle from apex [deg]

Fig. 5. Modeled and observed impact crater distribution on the Moon. We normalized the crater density to unity at antapex (γ = 180◦ ) using the best fit sinusoid (solid line curve). (a) Numerical result including the correction due to the difference of average impact velocity as a function of the angular distance (γ) from apex (γ = 0) for the population A particles. (b) Same as (a), but for the population B particles. (c) The observed rayed crater distribution of D > 5 km.6 The inset at the bottom right illustrates the coordinate system in this frame: The Earth always lies along −x direction, the Moon velocity is toward +y direction, and +z is the north of the Earth–Moon system. Apex point is defined as (x, y, z) = (0, RM , 0) where RM is the lunar radius.

from the best-fit sinusoid) in the observed crater record, whereas in our simulations, it is 1.31 ± 0.02 for the population A, and 1.34 ± 0.03 for the population B particles. This also means that the NEA population created by the new model (the population B) yields slightly stronger asymmetry in terms of the lunar impacts compared with what the NEA population created by the conventional model (the population A) does. But obviously the difference is not remarkable.

July 13, 2012 9:4

ADGEO

9in x 6in

Numerical Result from a New NEA Flux Model

b1389-ch01

11

5. Summary and Discussion We carried out a new set of numerical integrations of a population of particles created by a new, revised debiased NEA flux model. The aim of our calculation is to search potentially “slower” objects which can reproduce the stronger cratering asymmetry on the Moon that has been actually observed for the lunar rayed craters; this is because the leading/trailing cratering asymmetry becomes more prominent when the average relative velocity between the Moon and the projectiles is low. In our previous publications, we had performed similar numerical experiments using a debiased population of NEAs created by the conventional NEA flux model10 as well as a “raw” NEA population.13 But either of the results did not reach the observed rayed cratering asymmetry. In this paper using a revised NEA flux model, the resulting relative crater density at apex became slightly larger than when we used the conventional NEA flux model. We suspect that this slight enhancement is caused by the fact that the particle population produced from the new NEA flux model (i.e., the population B) contains relatively more particles around the Earth, a ∼ 1AU, than the population that was produced from the conventional NEA flux model (i.e., the population A) does, as seen in Fig. 1. The closer-Earth NEA component has generally lower relative velocity with respect to the Earth/Moon system, hence creating a higher cratering asymmetry. The difference of the new and the conventional models is also characterized by the inclusion or no inclusion of the high inclination NEA component. But it does not have a significant effect on the final result of cratering asymmetry because the high inclination NEA component does not have a large collision probability on the Earth or the Moon due to their geometric configuration (i.e., inclined orbits). Overall, difference between the results yielded by the new and conventional models is small, and we could not say that we have approached any explanation about the discrepancy between the crater record and what is suggested by the NEA dynamics. Hence, our result still implies the existence of undetected NEA populations with even lower relative velocity with respect to the Earth/Moon system. Deduced from the low relative velocity, this kind of populations could be Earth-coorbiting, some of them perhaps being produced by fragmentation due to Earth’s tidal force when a projectile approaches the Earth–Moon system. Recently one of the very examples, the first Earth’s Trojan object was detected through the survey observation by the WISE mission.24 In the forthcoming publications we will present our numerical result along the same computing scheme including

July 13, 2012 9:4

12

ADGEO

9in x 6in

b1389-ch01

T. Ito

slower NEA populations with several different dynamical characteristics including Earth’s Lagrangian points, and will make an estimate as to what kind of NEA populations could be responsible for the lunar crater asymmetry as high as what is currently observed. Of course, more complete observational surveys of NEAs will test our prediction, such as what is going on with Pan-STARRS. Also, future progress in the reconstruction of the true orbital distribution of NEAs by debiasing techniques, as well as reexamination of the lunar crater data including the latest lunar mission data, would be waited. Also, more careful examination as to how close to a steady-state the lunar impact flux has been is needed, as some large rayed craters are argued to be older than what they had been thought.25

References 1. E. M. Shoemaker, B. K. Lucchitta, D. E. Wilhelms, J. B. Plescia and S. W. Squyres, The geology of ganymede, in Satellites of Jupiter , ed. D. Morrison (The University of Arizona Press, Tucson, 1982), pp. 435–520. 2. G. P. Horedt and G. Neukum, Icarus 60 (1984) 710. 3. P. Schenk and S. Sobieszczyk, Bull. Am. Astron. Soc. 31 (1999) 1182. 4. K. J. Zahnle, L. Dones and H. F. Levison, Icarus 136 (1998) 202. 5. K. J. Zahnle, P. Schenk, S. Sobieszczyk, L. Dones and H. F. Levison, Icarus 153 (2001) 111. 6. T. Morota and M. Furumoto, Earth Planet. Sci. Lett. 206 (2003) 315. 7. S. C. Werner and S. Medvedev, Earth Planet. Sci. Lett. 295 (2010) 147. 8. A. S. McEwen, J. M. Moore and E. M. Shoemaker, J. Geophys. Res. 102 (1997) 9231. 9. T. Kawamura, T. Morota, N. Kobayashi and S. Tanaka, Geophys. Res. Lett. 38 (2011) L15201. 10. T. Ito and R. Malhotra, Astron. Astrophys. 519 (2010) A63. 11. W. F. Bottke, R. Jedicke, A. Morbidelli, D. Vokrouhlick´ y, M. Broˇz, D. Nesvorn´ y, J.-M. Petit and B. Gladman, Science 288 (2000) 2190. 12. W. F. Bottke, A. Morbidelli, R. Jedicke, J.-M. Petit, H. F. Levison, P. Michel and T. S. Metcalfe, Icarus 156 (2002) 399. 13. T. Ito, Asymmetric cratering on the moon: Case of the raw near-earth asteroid population, in Advances in Geosciences, ed. A. Bhardwaj (World Scientific, Singapore, 2010), pp. 109–119. ´ 14. J. Gallant, B. Gladman and M. Cuk, Icarus 202 (2009) 371. 15. M. Le Feuvre and M. A. Wieczorek, Icarus 197 (2008) 291. 16. M. Le Feuvre and M. A. Wieczorek, Icarus 214 (2011) 1. 17. H.-K. Moon, Y.-I. Byun, H.-S. Yim and S. N. Raymond, J. Korean Astron. Soc. 41 (2008) 7. 18. J. S. Stuart, Science 294 (2001) 1691.

July 13, 2012 9:4

ADGEO

9in x 6in

Numerical Result from a New NEA Flux Model

b1389-ch01

13

J. S. Stuart and R. P. Binzel, Icarus 170 (2004) 295. H. F. Levison and M. J. Duncan, Icarus 108 (1994) 18. T. Ito and R. Malhotra, Adv. Space Res. 38 (2006) 817. K. J. Zahnle and N. H. Sleep, Impacts and the early evolution of life, in Comets and the Origin and Evolution of Life, eds. P. J. Thomas, C. F. Chyba and C. P. McKay (Springer–Verlag, New York, 1997), pp. 175–208. 23. Y. Ishizaki and M. Furumoto, Planet. People 6 (1997) 12, text in Japanese. 24. M. Connors, P. Wiegert and C. Veillet, Nature 475 (2011) 481. 25. J. A. Grier, A. S. McEwen, P. G. Lucey, M. Milazzo and R. G. Strom, J. Geophys. Res. 106 (2001) 32847.

19. 20. 21. 22.

July 13, 2012 9:4

ADGEO

9in x 6in

This page intentionally left blank

b1389-ch01

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch02

Advances in Geosciences Vol. 30: Planetary Science and Solar & Terrestrial Science (2011) Ed. Anil Bhardwaj c World Scientific Publishing Company


DIFFUSE INTERSTELLAR PAH EMISSION IN THE LMC OBSERVED WITH THE AKARI/IRC H. UMEHATA∗ Institute of Astronomy, Graduate School of Science, The University of Tokyo, 2-21-1 Osawa, Mitaka, Tokyo 181-0015, Japan ∗ [email protected] I. SAKON and T. ONAKA Department of Astronomy, University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo 113-0033, Japan D. KATO Institute of Space and Astronautical Science, Sagamihara, Kanagawa 229-8510, Japan

We present the results of mid-infrared (MIR) slit spectroscopic observations of the diffuse interstellar medium (ISM) in the Large Magellanic Cloud (LMC) with the Infrared Camera (IRC) onboard AKARI. We have observed seven slit points across the whole galaxy, which have the 12 CO (J = 1 − 0) detection by NANTEN survey, and detect distinct unidentified infrared (UIR) bands (6.2 µm, 7.7 µm, 8.6 µm, 11.2 µm) for six points, while there are no signature of UIR bands in the spectra of one points. The comparison between the intensity ratios of 7.7 µm/11.2 µm and the IRAS 60 µm/100 µm color are carried out. We find that the band ratios peak at positions with an intermediate IRAS 60 µm/100 µm color of I60µm /I100µm ∼ 0.5 and decreases at positions with lower or higher IRAS 60 µm/100 µm colors of I60µm /I100µm ∼ 0.4 and I60µm /I100µm ∼ 0.6–0.7, respectively. This result indicates the transition of the ionization state of PAHs relying on the gas phase and radiation environment.

1. Introduction Gillett et al. discovered the 11.3 µm emission band feature in a planetary nebulae in 1973, which is the first detection of the unidentified infrared (UIR) bands (Gillett et al., 1973), which is a series of emission bands appeared in the mid-infrared (MIR) spectra (3.3, 6.2, 7.7, 8.6, 11.3, 12.7, and 16.4 µm). The UIR-bands are observed in various objects or environments such as photo-dissociation regions (PDRs) (Peeters et al., 2002), the diffuse interstellar medium (ISM),11 reflection nebulae (Peeters et al., 15

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch02

H. Umehata et al.

16

2002), nearby/distant galaxies (Lutz et al., 2005).3 Polycyclic aromatic hydrocarbons (PAHs) or PAH-containing carbonaceous dusts are thought to be the carrier of the UIR bands.1,12 PAHs absorb a single ultra-violet photon, and excited, and then release the energy with a number of photons at the infrared wavelength via various lattice vibration modes of C–H and C–C.1 The molecular structure, ionization state, and temperature of PAHs could affect the appearance of the PAH-features. One of the most remarkable factor to affect the PAHfeatures is the ionization state, which is controlled by U/ne , approximately, where U and ne represent the strength of the radiation field and the number density of electrons.2 Experimentally and theoretically it is argued that the band features at 6–9 µm are much weaker than the features at 11–14 µm for neutral PAHs, while the situation is reversed for ionized PAHs.4,14 In this paper, we present the result of the MIR spectroscopic observations with AKARI for the Large Magellanic Cloud (LMC) and consequently discuss the ionization states of PAHs in the LMC using the diagnostics of the intensity ratios of PAH features. 2. Observations and Data Reduction 2.1. Observations Our target is the LMC, which is a nearby external galaxy (55 kpc), oriented in an almost face on angle (i = 35◦ ; van der Marel et al., 2001), and whose metallicity is a lower than that of the Milky Way galaxy (0.3–0.5Z; Westerlund, 1997). Slit spectroscopic observations for seven positions in the LMC are carried out using the InfraRed Camera (IRC) onboard AKARI with Mid-Infrared Short (MIR-S) channel (Onaka et al., 2007b). The slit positions are summarized in Table 1. The observations are carried out as Table 1.

Information for observations.

Slit position (=Region number)

Observation Date

Object R.A. (degree, J2000)

Object DEC (degree, J2000)

1 2 3 4 5 6 7 zod

2006-10-14 2006-10-16 2006-10-16 2007-01-02 2006-10-19 2006-10-23 2007-05-18 2007-05-15

84.5836 84.9897 85.9747 81.4809 84.9295 84.8055 81.5265 67.9999

−70.1237 −69.7576 −69.3562 −66.1750 −70.0094 −69.5012 −68.6011 −69.9999

July 13, 2012 9:4

ADGEO

9in x 6in

Diffuse PAH Emission in the LMC

b1389-ch02

17

a part of the AKARI mission program “ISM in our Galaxy and Nearby Galaxies” (“ISMGN”, Kaneda et al., 2009). Observations were performed with the slit spectroscopic mode (AOT04; Ohyama et al., 2007). The spectra were taken with two grisms, SG1 (4.6–9.2 µm) and SG2 (7.2–13.4 µm) and the imaging data are also taken with S9W filter (9.0 µm) in the MIR-S channel. 2.2. Data Reduction The data reduction was carried out using the IRC spectroscopy data reduction toolkit (Ohyama et al., 2007). However some different reduction process was applied since a typical signal level of our targets is very faint. In the present observations, a single image set contains one short exposure and three long exposures. The dark image are obtained at the beginning and the end of the spectroscopic observations for each position. The dark current subtraction are made by utilizing the combined dark flame obtained by median-averaged three dark images. Then for each slit positions three long-exposure images in a single frame are also median-averaged to remove-out the cosmic-ray effects. Additionally we employ the datasets obtained at positions off the LMC with AOT04 to estimate and subtract the foreground components like the zodiacal light and diffuse Galactic emissions. 3. Strategy and Results 3.1. Strategy The target selection is very important for this research since our primary objective is to investigate the relationship between the property of PAHs and the surrounding environment, especially on the variation of the radiation field across the whole galaxy. Figure 1 shows the selected slit positions overlaying on the IRAS color map (I60µm /I100µm ) (Boulanger et al., 1988,10), and the superposed contours show the 12 CO(1 − 0) flux.6,9 We consider the following criteria to cover a wide range of the incident radiation field. First we focus on molecular cloud indicated by the detection of CO. the second condition is the dust temperature, which is indicated by the IRAS color ratio (I60µm /I100µm ). A high value suggests that there is active star forming activity and young stars radiate the surrounding dust at the position. Indeed, the positions of Regions 2 and 6, which show high value of (I60µm /I100µm ), are consistent with the positions of known young

July 13, 2012 9:4

18

ADGEO

9in x 6in

b1389-ch02

H. Umehata et al.

Fig. 1. Color map shows 60 µm/100 µm ratio based on IRAS data. Contours shows 12 CO(1 − 0) intensity observed by NANTEN. White points with number are slit posions of our observation.

star cluster N159 and N158, respectively. Third, AKARI/IRC must have the capability to observe target region. Eventually we successfully obtain the spectra for seven slit positions in different radiation environment. 3.2. Results The MIR imaging and the resultant spectra at our target positions are shown in Figs. 2–8. The PAH features are clearly seen at 3.3, 6.2, 7.7, 8.6 and 11.3 µm in six regions except Region 4. To diagnose PAH properties, the intensity of each PAH features are measured. We estimate continuum component using linear interpolation between the edges of the band feature and subtract them in the measurement of the intensity. Table 2 shows the measurement of the intensity ratios of 6.2 µm/ 11.2 µm, 7.7 µm/11.2 µm, 8.6 µm/11.2 µm. As shown in Fig. 9, the band ratios of 6.2 µm/11.2 µm and 7.7 µm/11.2 µm are largest at positions with I60µm /I100µm ∼ 0.5 and intermediate and the lowest values at positions with I60µm /I100µm ∼ 0.6–0.7 and I60µm /I100µm ∼ 0.4, respectively.

July 13, 2012 9:4

ADGEO

9in x 6in

Diffuse PAH Emission in the LMC

b1389-ch02

19

Fig. 2. (Left) Mid-infrared imaging around the slit position. (Right) Obtained midinfrared spectra of Region 1.

Fig. 3.

The MIR imaging and the spectra of Region 2.

Fig. 4.

The MIR imaging and the spectra of Region 3.

July 13, 2012 9:4

ADGEO

9in x 6in

H. Umehata et al.

20

Fig. 5.

The MIR imaging and the spectra of Region 4.

Fig. 6.

The MIR imaging and the spectra of Region 5.

Fig. 7.

The MIR imaging and the spectra of Region 6.

b1389-ch02

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch02

Diffuse PAH Emission in the LMC

Fig. 8.

21

The MIR imaging and the spectra of Region 7.

Table 2.

Intensity Ratios for each regions.

Region number

6.2 µm/11.2 µm

7.7 µm/11.2 µm

8.6 µm/11.2 µm

1 2 3 5 6 7

0.76 ± 0.17 1.05 ± 0.09 1.43 ± 0.34 1.06 ± 0.21 1.25 ± 0.04 1.51 ± 0.20

1.68 ± 0.31 2.56 ± 0.19 3.39 ± 0.65 1.55 ± 0.34 2.84 ± 0.07 3.25 ± 0.38

0.42 ± 0.14 0.50 ± 0.07 1.02 ± 0.23 0.32 ± 0.15 0.34 ± 0.02 0.40 ± 0.10

Fig. 9.

The diagram of IRAS 60/100 µm versus 7.7/11.2 µm.

July 13, 2012 9:4

22

ADGEO

9in x 6in

b1389-ch02

H. Umehata et al.

4. Discussion 4.1. Ionization state of PAHs We can provide reasonable description for Fig. 9 considering theoretical model of the ionization states of PAHs. Studies based on quantum chemical calculations and laboratory experiments show that what affects the intensity ratio of the PAH features are the ionization fraction and the molecular size of PAHs.5,15 The ionization of PAHs enhances the PAH band intensity which appears in the 6–9 µm relative to the PAH band in the 11–14 µm since the former are attributed mainly to stretching modes of C–C bonds, and the latter are attributed to out-of-plane bending modes of C–H bonds.2,4,14 Therefore, 7.7 µm/11.2 µm ratio shown in Fig. 9 is interpreted as the ionizedto-neutral band ratios and useful to investigate the ionization conditions of PAHs. Figure 9 indicates that Regions 3 and 7 might be expected to be dominated by photo dissociation regions (PDRs) rather than ionized (HII) regions and the difference between Region 2, 6 and Region 3, 7 might be caused by the rapid decline of ne at the boundary between HII regions and PDRs. The result implies that PDRs should offer the suitable environment for PAHs to be positively ionized.

4.2. The possibility of PAHs destruction There are no clear PAH features on the spectra of Region 4 while there are emission from very small grains (VSGs). Region 4 is located at the boundary between two supergiant shells (SGSs)8 and therefore it might be implied that PAHs were destroyed by a supernova shock wave.

5. Summary MIR slit spectroscopic observations toward the LMC by AKARI/IRC were carried out, and we obtain continuous spectra from 5 to 14 µm of the diffuse ISM at seven positions to investigate the spatial distribution of the PAH ionization states. The target positions are selected based on IRAS I60µm /I100µm color ratio and 12 CO(1 − 0) observations to cover a wide range of the incident radiation field across the galaxy. We find obvious PAH features in six positions, and derive the band intensity ratio of 6.2 µm/11.2 µm and 7.7 µm/11.2 µm. The PAH intensity ratio peaks at positions with an intermediate I60µm /I100µm ratio and our results suggest

July 13, 2012 9:4

ADGEO

9in x 6in

Diffuse PAH Emission in the LMC

b1389-ch02

23

that PDRs should offer a suitable environment for PAHs to be positively ionized.

Acknowledgments This research is based on observations with AKARI, a JAXA project with the participation of ESA. We would like to thank all members of the AKARI project.

References 1. L. J. Allamandola, A. G. G. M. Tielens and J. R. Barker, ApJS 71 (1989) 733. 2. E. L. O. Bakes, A. G. G. M. Tielens and C. W. Bauschlicher, Jr., ApJ 556 (2001) 501. 3. D. A. Dale et al., ApJ 646 (2006) 161. 4. D. J. DeFrees, M. D. Miller, D. Talbi, F. Pauzat and Y. Ellinger, ApJ 408 (1993) 530. 5. B. T. Draine and A. Li ApJ 551 (2001) 807. 6. Y. Fukui et al., ApJS 178 (2008) 56. 7. F. Galliano, S. C. Madden, A. G. G. M. Tielens, E. Peeters and A. P. Jones, ApJ 679 (2008) 310. 8. J. Meaburn, MNRAS 192 (1980) 365. 9. N. Mizuno et al., PASJ 53 (2001) 971. 10. T. Onaka, D. Tokura, I. Sakon, Y. Y. Tajiri, T. Takagi and H. Shibai, ApJ 654 (2007a) 844. 11. T. Onaka, I. Yamamura, T. Tanabe, T. L. Roellig and L. Yuen, PASJ 48 (1996) L59. 12. A. Sakata, S. Wada, T. Tanabe and T. Onaka, ApJ 287 (1984) L51. 13. I. Sakon, T. Onaka, D. Ishihara, T. Ootsubo, I. Yamamura, T. Tanabe and T. L. Roellig, ApJ 609 (2004) 203. 14. J. Szczepanski and M. Vala, ApJ 414 (1993) 646. 15. A. G. G. M. Tielens, ARA&A 46 (2008) 289.

July 13, 2012 9:4

ADGEO

9in x 6in

This page intentionally left blank

b1389-ch02

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch03

Advances in Geosciences Vol. 30: Planetary Science and Solar & Terrestrial Science (2011) Ed. Anil Bhardwaj c World Scientific Publishing Company


FORMATION OF AN EXTENDED HALO OF HOT OXYGEN ATOMS IN THE WAKE REGION OF VENUS YING LIAO∗ and WING HUEN IP† Institute of Astronomy, National Central University, No. 300, Jhongda Rd., Jhongli City, Taoyuan County, Taiwan ∗ [email protected] † [email protected]

From the detailed measurements in Venusian ionosphere by the Pioneer Venus Orbiter, it was well-known that there is a large day-to-night flow of ionospheric plasma with the horizontal speed reaching a value as high as 5 km/s at 500 km altitude near the terminator. This large-scale anti-sunward convective motion could lead to a significant distortion of the hot oxygen corona maintained by oxygen atoms from O+ 2 dissociation recombination into a tadpole-like structure. A Monte-Carlo model is developed to simulate the two-dimensional configuration of such a hot oxygen corona.

1. Introduction The atmospheres of both Venus and Mars are dominated by CO2 to be followed by N2 and other minor species. Photodissociation of CO2 leads to the production of CO and O that are efficiently recycled back to CO2 . The photodissociative products of water molecules, namely, H and O, could not be recycled, however. The H atoms are mostly lost by Jeans escape. The oxygen atoms are lost via several channels. Most of them have to do with the conversion of the oxygen atoms into O+ and O+ 2 ions by direct photoionization (for O+ ) or photochemical reactions (for O+ 2 ) such as + + + CO and CO + O → O + CO. The oxygen ions can O+ + CO2 → O+ 2 2 2 subsequently be lost because of a cometary-like solar wind interaction.1 Another important loss mechanism first proposed by McElroy.2 for Mars has to do with the exothermic electronic dissociative recombination of the O+ 2 ions. The branching ratios and excess energies of these reactions are shown below: Because of the large values of the excess energies, in particular, of the first two channels, an extended corona of hot oxygen atoms would be built 25

July 13, 2012 9:4

ADGEO

26

9in x 6in

b1389-ch03

Y. Liao and W. Huen Ip

up at Venus and Mars, respectively. Nagy et al.4 were the first to examine the formation of a hot oxygen atom corona around Venus by taking into consideration the O+ 2 dissociative recombination processes. They compared their theoretical results with the measured hot oxygen density profiles on the dayside and nightside from the ultraviolet spectrometer observations of Pioneer Venus Orbiter (PVO) and found good agreement. Along the same line but a little later, Ip5,6 and Nagy and Cravens7 produced hot oxygen corona models for Mars. Since Mars has a lower surface gravity which makes the dissociative recombination of O+ 2 an important escape process of its atmospheric oxygen. For example, Lammers et al.8 compared the oxygen escape rates of various mechanisms and found that 23 −1 for they are 5 × 1024 s−1 for O+ 2 dissociative recombination, 4.3 × 10 s 24 −1 solar wind sputtering, and 3 × 10 s for pickup ions. It is expected that the corresponding oxygen escape rate via dissociative recombination should be less important for Venus because of its lager surface escape velocity (i.e., 10.4 km/s versus 5.0 km/s). But the existence of the hot oxygen corona has nevertheless interesting effect on its solar wind interaction. We would like to point out that in previous studies of the formation of the hot oxygen coronas, the structures of the atmospheres and ionospheres of Mars and Venus have always been considered to be static, namely, there are no significant ionospheric flows in the upper atmospheres. However, measurements by the retarding potential analyzer on PVO from the exobase to about 500 km/s altitude detected a significant day-tonight horizontal flow.9 Cravens et al.10 and Shinagawa11 developed twodimensional ionospheric flow models to simulate such an ionospheric flow. In this work, we will outline the Monte Carlo model calculations to simulate the dynamical effect of the ionospheric plasma flow on the density distribution of the hot oxygen atoms around Venus. In Sec. 2, the basic scheme will be described. In Sec. 3, we will present the preliminary results of a 2D calculation. The summary and discussion will be presented in Sec. 4.

2. Model Description There are three main ingredients in our study: 1. Global distributions of the O+ 2 ions and the neutral atmosphere of Venus. 2. The flow field of the ionospheric plasma.

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch03

Hot Oxygen Halo of Venus

27

3. Initial ejection velocity distribution of the hot oxygen atoms created in dissociative recombination. Figure 1 shows the horizontal ionospheric flow pattern from the dayside to the nightside derived from the expression given by Cravens et al.10 We note that similar flow field has been obtained in the model calculation of Shinagawa.11 It can be seen that at the terminator the flow velocity u ∼ 1 km/s at 200 km high and u ∼ 5 km/s at about 500 km high. Because the peak velocity obtained from the first reaction branch in Table 1 is

Fig. 1. The speed distribution of horizontal ionospheric flow from the dayside to nightside.

Table 1. Major sources of suprathermal oxygen forming the exosphere of Venus.3 Reaction Branch 8 3 O( P) + O(3 P) > > > > 3 1 > > < O( P) + O( D) − O+ 3 2 +e → O( P) + O(1 S) > > > > O(1 D) + O(1 D) > > : 1 O( D) + O(1 S)

Energy (eV)

Probability (%)

6.99

22

5.02 2.80

42 20 amu 9◦ × 180◦ 4.5◦ × 22.5◦ 0.1–5 452 g

The instrument parameters of CENA and SWIM are presented in Table 1.

3. Observations and Results 3.1. Observation pertaining to ENAs 3.1.1. ENAs backscattered from lunar surface SARA has observed energetic neutral hydrogen atoms (hydrogen ENAs) from the Moon surface.1 The source of these ENAs are solar wind protons

July 13, 2012 9:4

ADGEO

9in x 6in

Interaction of Solar Wind with Moon

b1389-ch04

39

Fig. 2. Energy spectrum of solar wind and the ENAs scattered from lunar surface on 5 February 2009 for three orbits which are showed in solid line, dashed line and dotted line with dayside equator crossings at 05:22 UTC, 07.20 UTC, and 09:18 UTC, respectively. The dependence of the flux of scattered ENA on the incident solar wind flux is clearly visible from these spectra.1

which get neutralized in the interaction with the Moon surface and are scattered back to space. This is contrary to the earlier belief that the lunar surface completely absorbs the incident solar wind.16,17 Most of the ENAs are found to have energy less than ∼50% of the incident solar wind energy. The energy spectrum of the hydrogen ENAs shown in Fig. 2 is broader than the incident solar wind protons. One of the mechanisms for the energy loss may be the multiple scattering process due to interaction with the lattice. The observations showed that about 20% of the incident solar wind flux is backscattered as ENAs from the lunar regolith (Fig. 3). The solar zenith angle (SZA) dependence of the flux of scattered ENAs could be seen in Fig. 3. The higher energy of the ENAs is indicative that they are indeed solar wind backscattered ENAs as opposed to the sputtered ENAs whose energies based on model calculations are expected to be much lower (around few eV).18 The IBEX observations of lunar ENAs suggests that the solar wind backscatter efficiency is 10%29,43 and the analysis of the energy spectra indicates that they are indeed backscattered ENAs and not sputtered particles. The solar wind ions which are absorbed or implanted in the lunar regolith can get released to the lunar exosphere due to processes such as

July 13, 2012 9:4

40

ADGEO

9in x 6in

b1389-ch04

A. Bhardwaj et al.

Fig. 3. Map of scattered energetic neutral hydrogen atoms as observed by SARA on 6 February 2009. The direction of Sun is indicated in the figure. The ENA signature was seen for three consecutive orbits and the orbital plane was at an angle of 39◦ with the day–night terminator plane. The gradient in grey-scale and length of the lines indicates the ENA flux. The small flux seen on night-side is due to instrument background. The variation of ENA flux with solar zenith angle is seen in the map. The lunar surface map is from Clementine image data.1

diffusion, solar wind sputtering, and micrometeorite impact vaporization. The significant energy loss of the observed ENAs compared to impinging solar wind is due to elastic and inelastic processes during surface–plasma interactions, and possibly other processes such as retarding of solar wind ions by a positive surface potential. Hodges 44 has modeled the scattering of the solar wind from lunar surface using an inter-atom transport model for fast hydrogen in the surface layers of lunar soil grains and rocks. He had taken into account the rapid charge exchange neutralization of incident solar wind protons and also multiple encounters of free hydrogen with loosely packed soil grains. The modeled energy spectra compare well with CENA/SARA observations,1 but the modeled scattering efficiency is higher by a factor of three compared to the SARA observation. 3.1.2. Mini-magnetosphere observed in backscattered ENAs Although Moon does not possess a global magnetic field,45,46 it is found to have regions of localized magnetic field called magnetic anomaly

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch04

Interaction of Solar Wind with Moon

41

regions.10,14,47–49 There were speculations12 that these small areas of locally strong magnetic field can create mini-magnetospheres that may deflect the solar wind in the same way that Earth’s magnetosphere shields most of the planet from the solar wind. SARA has shown for the first time an existence of a mini-magnetosphere above the magnetic anomaly region using the backscattered ENAs.2 The image of a lunar magnetic anomaly of strength ∼100 nT in the backscattered hydrogen ENAs for the Crisium antipode anomaly near the Gerasimovic crater is shown in Fig. 4. The image shows that a partial void of the solar wind, a mini-magnetosphere, is formed above the strong magnetic anomaly. Above the magnetic anomaly region, the flux of backscattered hydrogen ENAs was significantly lower than the surrounding areas indicating that the region was shielded from solar wind particles by a mini-magnetosphere.2 The spatial variation in the flux of backscattered hydrogen ENA shows a reduction of about 50% within the area of the mini-magnetosphere (inside dotted circle in Fig. 4a) compared to the surrounding ring shaped region of enhanced flux (dashed line) for ENAs in the 150–600 eV energy range. In the map for lower energies shown in Fig. 4b, it is seen that the large scale depletion in the neutral hydrogen flux above the magnetic anomaly is replaced by small scale fluctuations and the region of enhanced flux has become almost a filled circle, which are due in part to a low instrument count rate and may also reflect energy dependent angular scattering properties of regolith surfaces, with higher energy scatter products possibly being more specularly reflected than lower energy scatter products. The minimagnetosphere is 360 km across at the surface and is surrounded by a 300 km thick circular region of enhanced plasma flux that results from the solar wind flowing around the mini-magnetosphere, for a magnetic anomaly of strength 100 nT at lunar surface.2 3.1.3. Distribution of hydrogen ENAs backscattered from the Moon The angular distribution of the scattered ENAs was investigated by making use of the CENA observations.50 Around 290,000 data points were used for the study so that almost full coverage of SZA, polar (0–90◦ ), and azimuthal (0–360◦) angles of scattering is achieved. The observation of hydrogen ENAs in the energy range 19–740 eV were considered. The observed angular scattering depends on the SZA (Fig. 5) such that as the SZA increases, more scattering takes place in the sunward direction than in the anti-sunward direction, which is contrary to what is expected

July 13, 2012 9:4

42

ADGEO

9in x 6in

b1389-ch04

A. Bhardwaj et al.

Fig. 4. The energetic neutral hydrogen flux from the surface over the magnetic anomaly near 22◦ S and 240◦ E on the lunar farside based on the observation on 17 June 2009 from 200 km altitude. The maps show a unit-less reflection coefficient: neutral hydrogen number flux integrated over the specified energy range divided by total solar wind number flux integrated over energy and multiplied by cosine of lunar latitude in the energy ranges (a) 150–600 eV (b) 30–100 eV. Black contours in the center show the magnetic field magnitude at 30 km altitude obtained from Lunar Prospector data, with lines for 5 nT, 15 nT, and 25 nT. The dotted circle represents the region of magnetic anomaly and the dashed circle represents the region just surrounding the anomaly. (c) Context image taken from the Clementine grey scale albedo map where the regions M, E and U indicate three sample regions inside the mini-magnetosphere, the enhanced flux region, and the undisturbed region, respectively.2

July 13, 2012 9:4

ADGEO

9in x 6in

Interaction of Solar Wind with Moon

b1389-ch04

43

Fig. 5. The angular distribution of the scattered hydrogen ENA for 15◦ SZA intervals. Empty squares correspond to (θ, φ) configurations for which we have no measurements. (a) Shows the coordinate system used for generating the maps shown in panels c and d. The arrow labeled “Sun” points in the sunward direction. For the sub-solar point (SZA = 0◦ ) the arrow would be pointing out of the plane of the page along the normal, for a SZA of 90◦ the arrow lies in the page plane (depicted). (b) Contains the grey-scale bar, the first bar belongs to panel c, and the second bar belongs to panel d. (c) The measured angular distributions. (d) The number of observations.50

based on laboratory studies.51 In addition, with the increase in SZA, the polar angle of scattering measured from the direction of surface normal also increases (shallower scattering). A mathematical expression which describes the observed angular scattering was derived as a function of SZA, polar, and azimuth angles of scattering.50 This provides a useful way to represent the

July 13, 2012 9:4

44

ADGEO

9in x 6in

b1389-ch04

A. Bhardwaj et al.

scattering function for solar wind produced ENAs from a regolith covered non-magnetized bodies in the solar system.

3.2. Observations pertaining to ions around the Moon The SWIM sensor of SARA has observed different populations of ions around Moon.5 The energy-time spectra from the SWIM observations for two orbits on 25 January 2009 is shown in Fig. 6. It may be noted that a few viewing direction of SWIM directly faces the Moon surface (Fig. 1). The spectra is divided into three panels depending on the viewing direction in which the ions are observed, such as space, horizon, and surface viewing. The different ion populations are marked by letters A to E. The population A, which is observed in the space viewing directions, when SWIM was on dayside, are the solar wind protons which have energy around 500–600 eV/q. In the same viewing directions, SWIM has observed ion population B during dayside, which have an energy/charge almost twice that of solar wind H+ energy/charge. These are the He++ ions in the solar wind. The population C is seen in the surface viewing pixels of SWIM on the dayside which have lower count rate and broader energy compared to that of solar wind protons. This represents the solar wind protons which are scattered from the lunar surface. The population D which are seen mostly close to the horizon are

Fig. 6. The energy time spectra of ions observed by SWIM for two consecutive orbits on 25 January 2009. The bottom panels represent the observation from the view directions of SWIM which looks at space, the center panel represents data from the view directions towards the Moon limb, and the top panel represents observations from view directions which looks at Moon surface. The population marked A and B are the direct solar wind H+ and He++ ions, respectively. The population C is the solar wind H+ reflected from Moon surface, D is the accelerated protons from population C, and E is the night side ions. The grey boxes on the top represents when Chandrayaan-1 was on lunar night side.5

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch04

Interaction of Solar Wind with Moon

45

the reflected ions which are the scattered solar wind protons picked up by the interplanetary magnetic field (Bsw ) and accelerated by the solar wind convective electric field (Esw ), thereby gaining energy. The population E are the ions seen when Chandrayaan-1 was on the nightside (plasma wake) of Moon. This represents the ions in the near-lunar plasma wake, which are found by mass analysis to be mostly protons. Except the population D, all others are seen for both the orbits. The trajectory of population D depends on the orientation of IMF and the convective electric field (Esw = −Vsw × Bsw , where Vsw is the solar wind velocity). The orientation of Bsw was different for the two orbits. So the trajectories of the reflected ions were such that they were able to enter SWIM FoV in the first orbit and not in the second orbit due to the change in orientation of Bsw .5 3.2.1. Deflection of solar wind protons over the magnetic anomalies As mentioned above, SWIM has measured the solar wind protons reflected from the lunar surface in the surface looking pixels. This has been observed by SWIM in almost all orbits of Chandrayaan-1. The SWIM observations of reflected protons were found to be highly correlated to the crustal magnetic field.4 Figure 7 shows SWIM data where the spacecraft passes by a large region of magnetic anomalies. Reflected protons, with energies similar to the solar wind, are clearly observed when magnetic anomalies are within the instrument field of view. This indicates that the observed protons are actually solar wind protons that have been deflected over the magnetic anomalies.4 A map of these deflected protons for the lunar far-side is shown in Fig. 8. It is notable that deflection can be observed even from small, isolated anomalies. The deflection ratio of solar wind protons integrated over the full far-side hemisphere is 1%, while over the magnetic anomaly regions alone it is around 10%, and over the strongest anomalies it reaches over 50% (see Fig. 9). Kaguya32 has also reported the deflection of solar wind protons by the magnetic anomaly near the South Pole Aitken region with an efficiency of more than 10%. 3.2.2. Ions in the near-lunar wake Significant proton fluxes were detected (cf. Fig. 6) by SWIM in the nearplasma wake region of the Moon. On 25 January 2009, when the protons moved along IMF in to the wake their energy was slightly higher than that of the solar wind protons. The protons were detected close to the lunar equatorial plane at a SZA of 140◦ , i.e. ∼50◦ behind the terminator at a

July 13, 2012 9:4

46

ADGEO

9in x 6in

b1389-ch04

A. Bhardwaj et al.

Fig. 7. SWIM-SARA observation during dayside pass of Chandrayaan-1 between 04:58 and 05:57 on 29 April 2009, over the Imbrium antipode magnetic anomalies. The top two panels represent the energy distribution of the ions with the differential flux summed over (a) five space-pointing directions (104–142◦ from nadir) and (b) five surface-pointing directions (44◦ –80◦ from nadir). (c) The direction distributions in angle from nadir, with the differential flux integrated over the 100 eV–3 keV energy range. (d) The strength of magnetic anomalies near the horizon is shown for reference. The strength is given in field magnitude at 30 km altitude based on model by Purucker (2008),52 from Lunar Prospector data. Also shown is the time of observation as well as the selenographical coordinates.4

July 13, 2012 9:4

ADGEO

9in x 6in

Interaction of Solar Wind with Moon

b1389-ch04

47

Fig. 8. The map of deflected solar wind protons observed in the 200 eV–1.7 keV energy range obtained by SWIM/SARA observations. The peak differential flux of protons is traced linearly to the surface of the Moon and binned to a 1◦ × 1◦ resolution spatial grid. Black contours show 2 nT, 3 nT, and 5 nT magnetic field strength at 30 km altitude in model by Purucker.52 The large anomaly cluster at the Imbrium Antipode (IA), the Serenitatis Antipode (SA) and, the Crisium Antipode (CA) are marked in the figure.4

height of 100 km. The protons came from just above the local horizon, and moved along IMF in to the wake.5 The comparison of the observed proton flux with the predictions of 1-D expansion of plasma into a vacuum showed that the observed velocity is higher than the velocity predicted by analytical models by a factor of two to three while the observed density is lower than the model value.5 As the solar wind is a magnetized plasma, the entry of solar wind ion in to the plasma wake is governed by the orientation of IMF. The protons in

July 13, 2012 9:4

48

ADGEO

9in x 6in

b1389-ch04

A. Bhardwaj et al.

Fig. 9. Map of the ratio of the outflowing proton flux to the incident proton flux. The outgoing proton flux is calculated from SWIM data and inflowing proton flux is calculated from WIND data.4

the near-lunar wake have also been observed by Kaguya33,34 and Chang’E,35 where the protons entered the wake in a direction perpendicular to IMF. There are four possible mechanisms suggested for entry of ions in to the near-lunar plasma wake. Out of these, three mechanisms favor entry in a direction perpendicular to IMF, can be called as perpendicular entry,33–35 and one mechanism favors entry of ions parallel to IMF5 (parallel entry). 3.2.3. The reflected ion trajectory: Modeling The protons reflected from the Moon surface will be influenced by the magnetic field carried by the solar wind (IMF) and the convective electric field (Esw ). For an IMF of 5 nT, the convective electric field is in the range 1.5–5.5 mV/m for a solar wind velocity (Vsw ) in the range 300–700 km s−1 (assuming Vsw and IMF are perpendicular to each other). Thus, the protons could get accelerated by Esw and gyrate around IMF (E × B drift). These are surface reflected ions and seen as population D in Fig. 6. Modeling the trajectory of these reflected ions (under Lorentz force) using a test particle approach, as well as hybrid model,3 showed that the reflected solar wind protons affect the global plasma environment (Fig. 10). The results of the model is compared with observations. The broadening of the energy spectrum of reflected protons seen in the model result is consistent with the SARA observation. Comparing with Nozomi’s ion observation revealed that the probable source of the non-thermal protons observed by Nozomi53 in the lunar vicinity were the reflected protons.

July 13, 2012 9:4

ADGEO

9in x 6in

Interaction of Solar Wind with Moon

b1389-ch04

49

Fig. 10. The results of a run of the test particle simulation. The grey-scale shows the ratio of the magnitude of the yz component of the proton number flux (number density times average velocity) around the Moon, to the magnitude of the solar wind proton number flux. Here, 100% reflection (f = 1) of the precipitating protons is considered but this can be scaled for any other value of f. The left panel, middle panel and the right panel shows the cuts through the planes x = 0, z = 0, and y = 0, respectively.3

4. Discussion The ENA scattering from Moon reduces the implantation of solar wind protons. The implantation of solar wind hydrogen and its subsequent bonding with oxygen in the regolith has been suggested as a potential candidate for the formation of OH/H2 O on Moon surface.54–56 Although laboratory simulations by Burke et al.57 using solid samples of ilmenite (FeTiO3 ) and anorthite (CaAl2 Si2 O8 ) did not show evidence for this, the very recent laboratory simulations by Managadze et al.58 by using olivine and SiO2 powders, which is more representative of the regolith composition,

July 13, 2012 9:4

50

ADGEO

9in x 6in

b1389-ch04

A. Bhardwaj et al.

has indeed shown that the solar proton is a potential source of lunar OH/H2 O molecules. The contradiction in the results are attributed to the use of solid surface by Burke et al.57 than the powdered sample used by Managadze et al.58 Thus, the 20% backscattering of solar wind protons as ENAs will affect the production of OH/H2 O on Moon surface by reducing the implantation rate of hydrogen. The proton scattering from Moon surface as ions also affect the implantation of hydrogen, but the fraction is much lower (0.1–1%) compared to the ENA scattering. On the other hand, the proton scattering has implications on the lunar plasma environment. The presence of the mini-magnetosphere causes deflection of solar wind around the magnetic anomaly region, resulting in net increase in the solar wind incident flux on the area surrounding the magnetic anomaly and decrease in the incident flux over the anomaly region. This correlates well with the Chandrayaan-1 Moon Mineralogy Mapper (M3 ) analysis of the spectral features of swirls and off-swirls at Reiner Gamma, Gerasimovich, and Mare Ingenii regions.59 The analysis indicated that the swirls are depleted in OH relative to their surrounding region, supporting the view that the magnetic anomalies deflect the solar wind away from the swirls and onto off-swirl surfaces. In addition to the effect of reduced OH production in magnetic anomaly regions, the deflection of solar wind protons affects the spectral properties22 of the soil at the anomaly region compared to the neighboring regions. This can be associated with magnetic shielding of the surface and the accumulation of fine charged dust above the swirls owing to the electrostatic interaction between charged fine grained dust and the electric field generated due to the charge separation between electrons and protons in the solar wind at magnetic anomaly.60,61 The maturation of the soil will be higher in the region adjacent to the anomaly due to increased flux of protons deflected from the swirls. Similar spectral studies of immature craters and surface soils both on and adjacent to the lunar swirls at Mare Ingenii region using the Clementine ultraviolet, visible and near infra-red cameras23 had also supported this hypothesis. Blewett et al.60 have done an extensive study of the spectral characteristics of swirls and found them to be similar to those of immature soil and also that their FeO content to be lesser than the surrounding region. Connecting it with the presence of magnetic anomaly, they suggest that the difference in composition could be due to either the magnetic shielding of solar wind or the accumulation of dust moving under the influence of the electric field induced by solar wind interaction with magnetic anomaly. Observations of the enhanced albedo of the Descartes C crater by Mini-RF (radio frequency) synthetic

July 13, 2012 9:4

ADGEO

9in x 6in

Interaction of Solar Wind with Moon

b1389-ch04

51

aperture radar on the Lunar Reconnaissance Orbiter is also found to be related to its location within a magnetic anomaly, and hence supports an origin hypothesis that invokes interaction between the solar wind and the magnetic anomaly.62 The absence of lunar-type swirls on Mercury based on Messenger flybys also suggest that models for the formation of lunar swirls invoking the interaction between solar wind and crustal magnetic anomalies are of prime importance.63 The protons scattered from lunar surface can directly affect the lunar plasma environment.3 Recent observations have shown that they cause whistler waves on the dayside of Moon.64 Their trajectories can lead to their entry to the nightside.34,35 The protons in the near wake cause instabilities and results in wave generation.65 The existence of ions in deep near-wake indicates that not only the processes which happen in the immediate downstream but also the ones on dayside play a role in determining the nature of dynamics in the immediate downstream. Thus, both the fluid nature (plasma expansion into vacuum)5 and particle nature of the solar wind (scattered protons reaching the nightside) play their role in the nearwake plasma features. Observations using Lunar Prospector magnetometer have shown the presence of whistler waves above the magnetic anomaly regions due to the interaction of solar wind with magnetic anomaly.66 All these indicate that the interaction of supersonic plasma flow with a planetary body without an atmosphere and intrinsic magnetic field is yet to be fully understood. The manifestation of the angular distribution of backscattered ENAs to be different from the expectations based on laboratory studies points to the requirement of further investigation of the micro-physics of the solar wind interaction with the regolith-covered planetary body. These results apply not only to Moon but also to any regolith covered planetary body with almost no atmosphere and lack a global magnetic field, but can have localized magnetic fields. Hence, these processes are expected to happen on Mercury, asteroids, and Moons of the giant planets also. The backscattering of solar wind protons have been observed recently on Phobos,67 the Moon of Mars, by the Mars Express.

5. Summary With a neutral particle detector (CENA) and ion-mass analyzer (SWIM), the SARA experiment on Chandrayaan-1 has made several interesting observations. These include (1) scattering of solar wind protons as ENAs

July 13, 2012 9:4

52

ADGEO

9in x 6in

b1389-ch04

A. Bhardwaj et al.

from Moon surface,1 (2) scattering of solar wind protons itself from Moon surface,3 (3) mini-magnetosphere around the magnetic anomaly region using both the scattered ENAs2 as well as solar wind protons deflected by the magnetic anomaly,4 (4) reflected ions which are the scattered solar wind protons influenced by the solar wind convective electric field and interplanetary magnetic field,3 and (5) protons in the near lunar plasma wake.5 Based on SARA observations: (1) the angular distribution of backscattered ENAs has been modeled,50 (2) a model to understand the influence of trajectories of reflected protons on the lunar plasma environment has been developed.3 The micro-physics underlying the interaction between the solar wind particles and the porous lunar regolith which results in the observed angular distribution of the scattered ENAs and the scattering of protons are areas for future investigation. References 1. M. Wieser, S. Barabash, Y. Futaana, M. Holmstr¨ om, A. Bhardwaj, R. Sridharan, M. B. Dhanya, P. Wurz, A. Schaufelberger and K. Asamura, Planet. Space Sci. 57 (2009) 2132. 2. M. Wieser, S. Barabash, Y. Futaana, M. Holmstr¨ om, A. Bhardwaj, R. Sridharan, M. B. Dhanya, P. Wurz, A. Schaufelberger and K. Asamura, Geophys. Res. Lett. 37 (2010) L015103. 3. M. Holmstr¨ om, M. Wieser, S. Barabash, Y. Futaana and A. Bhardwaj, J. Geophys. Res. 115 (2010) A06206. 4. C. Lue, Y. Futaana, S. Barabash, M. Wieser, M. Holmstr¨ om, A. Bhardwaj, M. B. Dhanya and P. Wurz, Geophys. Res. Lett. 38 (2011) L03202. 5. Y. Futaana, S. Barabash, M. Wieser, M. Holmstr¨ om, A. Bhardwaj, M. B. Dhanya, R. Sridharan, P. Wurz, A. Schaufelberger and K. Asamura, J. Geophys. Res. 115 (2010b) A10248. 6. R. M. Killen and W.-H. Ip, Rev. Geophys. 37 (1999) 361. 7. S. A. Stern, Rev. Geophys. 37 (1999) 453. 8. R. Sridharan, S. Ahmed, T. P. Das, P. Sreelatha, P. Pradeepkumar, N. Naik and G. Supriya, Planet. Space Sci. 58 (2010a) 947. 9. R. Sridharan, S. Ahmed, T. P. Das, P. Sreelatha, P. Pradeepkumar, N. Naik and G. Supriya, Planet. Space Sci. 58 (2010b) 1567. 10. P. J. Coleman, B. R. Lichtenstein, C. T. Russell, L. R. Sharp and G. Schubert, Geochem. Cosmochem. Acta 36 (1972) 2271. 11. L. L. Hood and G. Schubert, Science 208 (1980) 49. 12. R. P. Lin, D. L. Mitchell, D. W. Curtis, K. A. Anderson, C. W. Carlson, J. McFadden, M. H. Acu˜ na, L. L. Hood and A. Binder, Science 281 (1998) 1480.

July 13, 2012 9:4

ADGEO

9in x 6in

Interaction of Solar Wind with Moon

b1389-ch04

53

13. L. L. Hood, A. Zakharian, J. Halekas, D. L. Mitchell, R. P. Lin, M. H. Acu˜ na and A. B. Binder, J. Geophys. Res. 106 (2001) 27825. 14. J. S. Halekas, D. L. Mitchell, R. P. Lin, S. Frey, L. L. Hood, M. H. Acu˜ na and A. B. Binder, J. Geophys. Res. 106 (2001) 27841. 15. N. C. Richmond, L. L. Hood, J. S. Halekas, D. L. Mitchell, R. P. Lin, M. Acu˜ na and A. B. Binder, Geophys. Res. Lett. 30 (2003) 1395. 16. W. C. Feldman, D. J. Lawrence, R. C. Elphic, B. L. Barraclough, S. Maurice, I. Genetay and A. B. Binder, J. Geophys. Res. 105 (2000) 4175. 17. D. H. Crider and R. R. Vondrak, Adv. Space Res. 30 (2002) 1869. 18. P. Wurz, U. Rohner, J. A. Whitby, C. Kolbb, H. Lammer, P. Dobnikar and J. A. Mart´ın-Fern´ andez, Icarus 191 (2007) 486. 19. A. Milillo, S. Orsini, K. C. Hsieh, R. Baragiola, M. Fama, R. Johnson, A. Mura, C. Plainaki, M. Sarantos, T. A. Cassidy, E. D. Angelis, M. Desai, R. Goldstein, W. H. Ip, R. Killen and S. Livi, J. Geophys. Res. 116 (2011) A07229. 20. A. Bhardwaj, S. Barabash, Y. Futaana, Y. Kazama, K. Asamura, R. Sridharan, M. Holmstr¨ om, P. Wurz and R. Lundin, J. Earth Syst. Sci. 114 (2005) 749. 21. Y. Futaana, S. Barabash, M. Holmstr¨ om and A. Bhardwaj, Planet Space Sci. 54 (2006) 132. 22. C. M. Pieters, E. M. Fischer, O. Rode and A. Basu, J. Geophys. Res. 98 (1993) 20. 23. G. Y. Kramer, J. P. Combe, E. M. Harnett, B. R. Hawke, S. K. Noble, D. T. Blewett, T. B. McCord and T. A. Giguere, J. Geophys. Res. 116 (2011a) E04008. 24. K. W. Ogilvie, J. T. Steinberg, R. J. Fitzenreiter, C. J. Owen, A. J. Lazarus, W. M. Farrell and R. B. Torbert, Geophys. Res. Lett. 10 (1996) 1255. 25. E. F. Lyon, S. B. I-I and I. B. J, J. Geophys. Res. 72 (1967) 6113. 26. W. M. Farrell, M. L. Kaiser and J. T. Steinberg, Geophys. Res. Lett. 24 (1997) 1135. 27. S. D. Bale, C. J. Owen, J.-L. Bougeret, K. Goetz, P. J. Kellogg, R. P. Lin, R. Manning and S. J. Monson, Geophys. Res. Lett. 24 (1997) 1427. 28. J. S. Halekas, Y. Saito, G. T. Delory and W. M. Farrell, Planet. Space. Sci. (2010). 29. D. J. McComas, F. Allegrini, P. Bochsler, P. Frisch, H. O. Funsten, M. Gruntman, P. H. Janzen, H. Kuchareck, E. M. obius, D. B. Reisenfeld and N. A. Schwadron, Geophys. Res. Lett. 36 (2009) L12104. 30. Y. Saito, S. Yokota, T. Tanaka, K. Asamura, M. N. Nishino, M. Fujimoto, H. Tsunakawa, H. Shibuya, M. Matsushima, H. Shimizu, F. Takahashi, T. Mukai and T. Terasawa, Geophys. Res. Lett. 35 (2008) L24205. 31. S. Yokota, Y. Saito, K. Asamura, T. Tanaka, M. N. Nishino, H. Tsunakawa, H. Shibuya, M. Matsushima, H. Shimizu, F. Takahashi, M. Fujimoto, T. Mukai and T. Terasawa, Geophys. Res. Lett. 36 (2009) L11201. 32. Y. Saito, S. Yokota, K. Asamura, T. Tanaka, Y. M. N. Nishino, Y. Terakawa, M. Fujimoto, H. Hasegawa, H. Hayakawa, M. Hirahara, M. Hoshino,

July 13, 2012 9:4

54

33.

34.

35.

36.

37. 38.

39. 40.

41.

42. 43. 44. 45. 46. 47. 48. 49.

ADGEO

9in x 6in

b1389-ch04

A. Bhardwaj et al.

S. Machida, T. Mukai, T. Nagai, T. Nagatsuma, T. Nakagawa, M. Nakamura, K. Oyama, E. Sagawa, S. Sasaki, K. Seki, I. Shinohara, T. T. Terasawa, H. Shibuya, M. Matsushima, H. Shimizu and F. Takahashi, Space Sci. Rev. 154 (2010) 265. M. N. Nishino, K. Maezawa, M. Fujimoto, Y. Saito, S. Yokota, K. Asamura, T. Tanaka, H. Tsunakawa, M. Matsushima, F. Takahashi, T. Terasawa, H. Shibuya and H. Shimizu, Geophys. Res. Lett. 36 (2009a) L12108. M. N. Nishino, M. Fujimoto, K. Maezawa, Y. Saito, S. Yokota, K. Asamura, T. Tanaka, H. Tsunakawa, M. Matsushima, F. Takahashi, T. Terasawa, H. Shibuya and H. Shimizu, Geophys. Res. Lett. 36 (2009b) L16103. X.-D. Wang, W. Bian, J.-S. Wang, J.-J. Liu, Y.-L. Zou, H.-B. Zhang, C. Lu, J.-Z. Liu, W. Zuo, Y. Su, W.-B. Wen, M. Wang, Z.-Y. Ouyang and C.-L. Li, Geophys. Res. Lett 37 (2010) L07203. S. Wiehle, F. Plaschke, U. Motschmann, K.-H. Glassmeier, H. U. Auster, V. Angelopoulos, J. Mueller, H. Kriegel, E. Georgescu, J. Halekas, D. G. Sibeck and J. P. McFadden, Planet. Space Sci. 59 (2011) 661. D. McCann, S. Barabash, H. Nilsson and A. Bhardwaj, Planet. Space Sci. 55 (2007) 1190. Y. Kazama, S. Barabash, A. Bhardwaj, K. Asamura, Y. Futaana, M. Holmstr¨ om, R. Lundin, R. Sridharan and P. Wurz, Adv. Space Res. 37 (2006) 38. Y. Kazama, S. Barabash, M. Wieser, K. Asamura and P. Wurz, Planet. Space Sci. 55 (2007) 1518. S. Barabash, A. Bhardwaj, M. Wieser, R. Sridharan, T. Kurian, S. Varier, E. Vijayakunar, V. Abhirami, K. V. Raghavendra, S. V. Mohankumar, M. B. Dhanya, S. Thampi, K. Asamura, H. Andersson, Y. Futaana, M. Holmstrm, R. Lundin, J. Svensson, S. Karlsson, R. D. Piazza and P. Wurz, Current Science 96 (2009) 526. A. Bhardwaj, M. Wieser, M. B. Dhanya, S. Barabash, Y. Futaana, M. Holmstrm, R. Sridharan, P. Wurz, A. Schaufelberger and K. Asamura, Advances in Geosciences 19 (2010) 151. M. B. Dhanya, A. Bhardwaj, Y. Futaana, S. Barabash, M. Wieser, R. Sridharan, M. Holmstr¨ om and P. Wurz, private communication. D. Rodriguez, L. Saul, P. Wurz, S. Fuselier, H. Funsten, E. M. obius and D. McComas, Planet. Space Sci. 60 (2012) 297. R. R. Hodges, Geophys. Res. Lett. 38 (2011) L06201. S. S. Dolginov, E. G. Eroshenko, L. N. Zhuzgov and N. V. Pushkov, Geomagn. Aeron. Engl. Transl. 1 (1961) 18. N. S. Ness and K. W. Behannon and C. S. Scearce and S. C. Cantarano, J. Geophys. Res. 72 (1967) 5769. K. A. Anderson, R. P. Lin, R. E. McGuire and J. E. McCoy, Space Sci. Inst. 1 (1975) 439. R. P. Lin, Phys. Earth Planet. Int. 20 (1979) 271. D. L. Mitchell, J. S. Halekas, R. P. Lin, S. Frey, L. L. Hood, M. H. Acu˜ na and A. Binder, Icarus 194 (2008) 401.

July 13, 2012 9:4

ADGEO

9in x 6in

Interaction of Solar Wind with Moon

b1389-ch04

55

50. A. Schaufelberger, P. Wurz, S. Barabash, M. Wieser, Y. Futaana, M. Holmstrm, A. Bhardwaj, M. B. Dhanya, R. Sridharan and K. Asamura, Geophys. Res. Lett. 38 (2011) L22202. 51. H. Niehus, W. Heiland and E. Taglauer, Surf. Sci. Rep. 17 (1993) 213. 52. M. E. Purucker, Icarus 197 (2008) 19. 53. Y. Futaana, S. Machida, Y. Saito, A. Matsuoka and H. Hayakawa, J. Geophys. Res. 108 (2003) 1025. 54. C. M. Pieters, J. N. Goswami, R. N. Clark, M. Annadurai, J. Boardman, B. Buratti, J.-P. Combe, M. D. Dyar, R. Green, J. W. Head, C. Hibbitts, M. Hicks, P. Isaacson, R. Klima, G. Kramer, S. Kumar, E. Livo, S. Lundeen, E. Malaret, T. McCord, J. Mustard, J. Nettles, N. Petro, C. Runyon, M. Staid, J. Sunshine, L. A. Taylor, S. Tompkins and P. Varanasi, Science 326 (2009) 568. 55. J. M. Sunshine, T. L. Farnham, L. M. Feaga, O. Groussin, F. Merlin, R. E. Milliken and M. F. A’Hearn, Science 326 (2009) 565. 56. T. B. McCord, L. A. Taylor, J. P. Combe, G. Kramer, C. M. Pieters, J. M. Sunshine and R. N. Clark, J. Geophys. Res. 116 (2011) E00G05. 57. D. J. Burke, C. A. Dukes, J.-H. Kim, J. Shi, M. Fam´ a and R. A. Baragiola, Icarus 211 (2011) 1082. 58. G. G. Managadze, V. T. Cherepin, Y. G. Shkuratov, V. N. Kolesnik and A. E. Chumikov, Icarus 215 (2011) 449. 59. G. Y. Kramer, S. Besse, D. Dhingra, J. Nettles, R. Klima, I. G. Bethell, R. N. Clark, J. P. Combe, J. W. H. III, L. A. Taylor, C. M. Pieters, J. Boardman and T. B. McCord, J. Geophys. Res. 116 (2011b) E00G18. 60. D. T. Blewett, E. I. Coman, B. R. Hawke, J. J. Gillis-Davis, M. E. Purucker and C. G. Hughes, J. Geophys. Res. 116 (2011) E02002. 61. I. Garrick-Bethell, J. W. H. III and C. M. Pieters, Icarus 212 (2011) 480. 62. C. D. Neish, D. T. Blewett, D. B. J. Bussey, S. J. Lawrence, M. Mechtley, B. J. Thomson and T. M.-R. team, Icarus 215 (2011) 186. 63. D. T. Blewett, B. W. Denevi, M. S. Robinson, C. M. Ernst, M. E. Purucker and S. C. Solomon, Icarus 209 (2010) 239. 64. T. Nakagawa, F. Takahashi, H. Tsunakawa, H. Shibuya, H. Shimizu and M. Matsushima, Earth Planets Space 63 (2011) 37. 65. M. N. Nishino, M. Fujimoto, Y. Saito, S. Yokota, Y. Kasahara, Y. Omura, Y. Goto, K. Hashimoto, A. Kumamoto, T. Ono, H. Tsunakawa, M. Matsushima, F. Takahashi, H. Shibuya, H. Shimizu and T. Terasawa, Geophys. Res. Lett. 37 (2010) L12106. 66. J. S. Halekas, D. A. Brain, D. L. Mitchell and R. P. Lin, Geophys. Res. Lett. 33 (2006). 67. Y. Futaana, S. Barabash, M. Holmstr¨ om, A. Fedorov, H. Nilsson, R. Lundin, E. Dubinin and M. Fr¨ anz, J. Geophys. Res. 115 (2010a) A10213.

July 13, 2012 9:4

ADGEO

9in x 6in

This page intentionally left blank

b1389-ch04

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch05

Advances in Geosciences Vol. 30: Planetary Science and Solar & Terrestrial Science (2011) Ed. Anil Bhardwaj c World Scientific Publishing Company


PLASMA TRANSPORT PROCESSES IN THE TOPSIDE MARTIAN IONOSPHERE TARIQ MAJEED University of Michigan, 2455 Hayward St., Ann Arbor, Michigan, 48109-2143 USA and American University of Sharjah, P.O. Box 26666, Sharjah, UAE STEPHEN W. BOUGHER University of Michigan, 2455 Hayward St., Ann Arbor, Michigan, 48109-2143 USA S. A. HAIDER Physical Research Laboratory, Ahmedabad, India

The altitude profiles of the Martian ionosphere have been measured extensively over the past five decades with orbiters and fly-bys missions. The magnetic field has also been observed with the magnetometer on board MGS spacecraft. These observations provided no evidence of a significant planetary magnetic field at Mars. However, the observations of electron density scale heights indicate quite variable topside ionospheric structure which seems to violate diffusive equilibrium: The condition that would have been imposed by a magnetic field-free ionosphere. Plasma transport has been suggested to be the most viable process for the interpretation of such an ionosphere by using an external magnetic field arising from solar wind interaction with the Martian ionospheric/atmospheric system. We have developed a 1-D chemical diffusive model to describe plasma transport processes for the topside Martian ionosphere. For the case of a purely induced horizontal field, the plasma loss is due to both the downward flow and horizontal divergence of ion velocities. However, for the case of a vertical magnetic field, an upward flow of plasma seems to play an important role in the upper ionosphere. The vertical transport of plasma is described in our model by the upward ion flux (Φi ) and downward drift velocity (WD ). The impact of Φi on the topside ionosphere has been simulated with the values of ion fluxes that the model can allow. WD in our model is varied from −10 ms−1 to −50 ms−1 . While these transport processes reduce the topside electron densities yielding smaller scale heights, the photochemical equilibrium prevails near and below the ionospheric peak.

57

.

July 13, 2012 9:4

58

ADGEO

9in x 6in

b1389-ch05

T. Majeed, S. W. Bougher & S. A. Haider

1. Introduction Since its discovery in 1965 with the Mariner 4 spacecraft,1 the ionosphere of Mars has been observed many times with flybys and orbiters over the past five decades using both the radio occultation technique and in situ probes. Earlier observations [e.g., 2, 3, 4, 5, 6, 7, 8], represent the dayside ionosphere with primary and secondary peaks in the electron density (Ne ) profiles for very low solar activity (F10.7 ∼ 70) to a moderately high solar activity period (F10.7 ∼ 170). The range of solar zenith angle (SZA) for these Ne measurements was ∼45◦ –90◦ . The most recent measurements of the Martian ionosphere have been made with the Mars global surveyor (MGS)9 and Mars express (MEX) orbiters.10,11,12 Radio science subsystem (RSS) experiments on board MGS using the radio occultation technique were performed to observe the high-latitude Martian ionosphere during its mission from December 1998 to March 2005. Analysis of these observations provided a comprehensive set of Ne profiles near the Martian terminator (SZA ∼71◦ –87◦ ). The MEX spacecraft is orbiting Mars since 25 December 2003. The MEX is equipped to perform ionospheric measurements with a couple of experiments onboard; radio science (MaRS)12 and the Mars advanced radar for subsurface and ionospheric sounding (MARSIS).10 Unlike the MaRS instrument, the MARSIS instrument was placed to probe the topside ionosphere and the total electron content of the atmosphere. Because MARSIS operates by vertical sounding, it can also measure the Ne profiles near the subsolar point. The MaRS instrument probed the Martian ionosphere from April to August 2004 (AA2004) and December 2004 to January 2005 (DJ2005) for low solar activity (F10.7 ∼ 80). The observations of AA2004 provided 13 early morning Ne profiles at SZA between 85◦ and 108◦ and 77 evening profiles between 70◦ and 84◦ . The MARSIS instrument probed the topside ionosphere from 5 July 2005 to 10 October 2005 for a period with relatively low solar activity (F10.7 ∼ 100−130).10 Although the details of MaRS Ne profiles are not known they are expected to have two peaks. An upper peak appears in the range ∼125–150 km, and a lower peak observed in the range ∼105–135 km similar to the MGS Ne profiles. The peak magnitude of the observed Ne profiles for the upper region is in the range 1.6 × 105 to 3 × 104 cm−3 for the subsolar point to the region of larger SZA. The magnitude of the lower region Ne peak is shown to be in the range 2.5 × 104 to 1.7 × 104 cm−3 . The major source of the upper region peak Ne is the process of photoionization of the neutral gas by EUV photons with wavelengths of ∼15.0–100 nm while the

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch05

Plasma Transport Processes in the Topside Martian Ionosphere

59

lower region peak Ne is attributed to the process of the electron-impact ionization of the neutral gas.13,14 Soft X-ray photons (λ < 15 µm) also strongly contribute to the magnitudes of the Ne peaks in the lower region of the Martian ionosphere.15 The details of these processes and Cosmic ray impact ionization on the D-region of the Martian ionosphere are given in a recent review by Haider et al.16 The mystery of magnitudes and height variations of the measured Ne peaks has been debated in the current literature with various scenarios. These scenarios are different than what one must expect for changes in SZA. The planetary scale-waves in the lower atmosphere has have used to interpret variations in the peak altitudes as a function of planetocentric longitude.17,18,19,20 The oscillations caused by wave activity may impact atmospheric pressure, temperature, and densities up to altitudes of about 150 to 160 km.21 Wang and Nielson22 adopted a wave model and explained the height variation in the Mariner 9 Ne data with the elevation of the Martian surface. Seasonal changes with this wave model have been predicted to affect the Ne peak altitude with a variation of about 15 km from aphelion to perihelion.23 The Martian dust storm has also been considered as one of the causes of the measured variations in the peak magnitudes and peak altitudes of the Ne profiles.21,24,25,26 The measured ionospheric Ne profiles have also been analyzed by using Chapman layer theory. Assumptions of this theory include: photoionization of a single molecular species, monochromatic ionizing radiation, an isothermal atmosphere characterized by the neutral temperature, and photochemical equilibrium. Although conditions in the Martian ionosphere strongly differ from these assumptions, Chapman theory is applied as a first approximation to describe the characteristics of the main ionospheric peak.10,17,27 Chapman theory has also been used to analyze measured variation of Ne profiles with SZA, solar EUV fluxes, latitude and longitude, neutral scale height and the temperature in the vicinity of the ionization peak by several investigators.28,29,30,31,32 Most recently, Fox and Yeager33 and Morgan et al.10 using the MGS and MEX Ne data derived information on the neutral and plasma scale heights and variations in the peak Ne in response to varying SZA and F10.7 fluxes. The impact of embedded magnetic field due to solar wind interaction with the ionosphere/atmosphere [cf., Ma et al.34 ] on the variability of the observed Martian ionosphere has been discussed by Krymskii et al.35 They used MGS magnetometer-electron reflectometer (MAG-ER) and

July 13, 2012 9:4

60

ADGEO

9in x 6in

b1389-ch05

T. Majeed, S. W. Bougher & S. A. Haider

MGS accelerometer (MGS-ACC) data to investigate the role of magnetic field and neutral atmospheric scale heights as a driving force to explain the ionospheric variability observed in MGS/RS Ne data both for the northern and southern hemispheres. They found that this variability is more pronounced in the southern than in the northern hemisphere, consistent with observed crustal magnetic field and neutral scale height variability. The topside ionosphere has been shown to be sensitive to the dimensions of the crustal magnetic field35,36 and provide an effective source of day-to-to-night plasma transport near the terminators. Assuming the condition of largescale minimagnetospheres between 120◦–250◦ longitude, Krymskii et al.35 determined that the topside electron density scale heights, on average between altitudes of 145–165 km and 165–185 km, decrease more rapidly in the southern hemisphere than in the northern hemisphere particularly for large SZA. Clearly, this suggests that cross-terminator plasma convection at altitudes above 165 km is effective within the large-scale minimagnetosphere and it is more rapid in the southern hemisphere than in the northern hemisphere. Ma et al.34 interpreted the topside ionosphere observed by MGS and Viking 1 with their 3-D multispecies high spatial resolution MHD model. While explaining the inherent characteristics of the measured Ne profiles they concluded that the topside Ne scale heights are modified as a result of temperature gradients, convective plasma transport, and embedded magnetic fields due to the presence of minimagnetospheres in the plasma environment of Mars [see 37 for details]. Thus within the topside ionosphere, variation in the Ne scale heights indicates either horizontal removal of plasma38,39 to establish an effective trans-terminator plasma transport or tailward escape of the ionospheric plasma. These escape fluxes of plasma vary with solar cycle condition as simulated by Ma and Nagy.40 The ionospheric escape makes up a significant fraction of the total escape flux estimated from Phobos 2 measurements.41,42 Fox43 calculated transterminator plasma fluxes consistent with those later simulated by Ma and Nagy40 to match the observed Ne densities. Most recent studies of the Martian ionosphere-thermosphere system provided evidence of magnetized regions. In these regions, transport of photoelectrons near the equator cause plasma loss in the upper ionosphere44 and charged particle precipitation processes at the high-latitudes produce aurora,45 joule heating, and electric fields.46 Thus, the Ne distribution in the upper ionosphere of Mars can be accounted for by including the effects of vertical plasma transport through the magnetic field lines.

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch05

Plasma Transport Processes in the Topside Martian Ionosphere

61

In this paper, we present results of our chemical diffusive model for the southern high-latitude regions which are characterized by strong magnetic fields either due to Martian solar wind interaction or remanent crustal magnetization. In our model the variation of the Martian upper ionosphere is described by the upward ion flux resulting from solar wind dynamic pressure and the vertical ion drift induced by meridional winds in the neutral atmosphere or electric field.

2. Model Description The process of photoionization is assumed to be the major source of ionization of the neutral atmosphere in the dayside upper ionosphere of Mars. In the lower ionosphere, the ionization due to photoelectron and soft X-ray photons is assumed to become significant making a contribution of up to 25% of the total ionization rate. The six ion species included in + + + + + the present version of the model are CO+ 2 , CO , O2 , O , NO , and H . The chemical scheme and corresponding reaction rate coefficients in the model are adopted from Fox and Sung.13 Because our focus for these exploratory calculations is to describe upper ionospheric structure for the Martian region with embedded magnetic field due to solar wind interaction and/or remanent crustal magnetization as observed in the high-latitude southern hemisphere (160◦ E and 250◦ E longitudes), we use ionization rates and background neutral atmosphere from published studies. The profiles of neutral densities of CO2 , CO, N2 , O2 and O for the solar minimum conditions along with ion, neutral and electron temperature profiles were taken from Bougher et al.17,20 The density profiles of NO and H were adopted from Fox and Sung.13 The corresponding profiles for ionization + + + rates of CO+ 2 , CO , O2 , and O (including ionization due to photoelectron impact) for the solar minimum condition for ∼66◦ S latitude at high solar zenith angle were adopted from Bougher et al.17 while the profiles of ionization rates for NO+ and H+ for almost the same conditions were adopted from Fox and Sung.13 The diffusion of ions in the background atmosphere is included in the present version of the model. The ion-neutral diffusion coefficients in the model were calculated using the data and formulae given by Banks and Kockarts.47 The governing equations to be solved were the continuity equation for each ion, ∂φi ∂ni + = Pi − Li ∂t ∂z

July 13, 2012 9:4

62

ADGEO

9in x 6in

b1389-ch05

T. Majeed, S. W. Bougher & S. A. Haider

and flux equation for each ion given as, 
∂ni ni ni ∂ ni Te ∂ne φi = −Din + + (Te + Ti ) + ni WD , + · · · ∂z Hi ne Ti ∂z Ti ∂z where the vertical drift velocity, WD , is given as, WD = −αEy cos I − βUn cos I − Vn sin I cos I + Wn sin 2I. In the above equations, Te and Ti are the electron and ion temperatures, respectively, ni and ne are ion and electron densities, Hi is the ion scale height and I is the magnetic dip angle. We assume charge neutrality condition. Un , Vn , and Wn , are the eastward (zonal), northward (meridional), and vertical winds of the neutral atmosphere defined with respect to the plane of the magnetic field. Ey is the eastward electric field component. Since vertical neutral winds are much smaller than zonal and meridional winds,23 the drift velocity primarily arises due to electric field and meridional component of the neutral winds. The dip angle for given latitude is calculated using a simple dipole model aligned with the rotation axis.47 The terms involving α and β are defined by Banks and Kockarts.47 The lower boundary in the model is located at 100 km and is assumed to be in photochemical steady state (PCSS), although for some cases we use lower boundary as zero flux boundary condition. The upper boundary is located at 400 km and is assumed to be in diffusive equilibrium (φi = 0). We have also performed calculations with non-zero boundary fluxes to simulate outflow of ions resulting from the direct solar wind interaction with the Martian ionosphere. There is an upper limit to the upward flux because the upward flow of ions is diffusion limited when the external electrical driving force is zero [see reference 47 for details].

3. Results and Discussions Figure 1 shows electron density profiles calculated with different scenarios. The model electron density (Ne ) profile with zero magnetic field is compared with the MGS Ne profile at 66◦ S latitude. For this model run, a zero-flux upper boundary condition is used with WD = 0 while the plasma at lower boundary is assumed to be in photochemical steady-state. The diffusion of ions in the background atmosphere is allowed because plasma is not impeded by the magnetic field and thus diffuses purely in response to photochemical imbalances in the ionosphere. The peak in the model Ne profile occurs at

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch05

Plasma Transport Processes in the Topside Martian Ionosphere

63

Fig. 1. A comparison of model Ne profiles with the MGS/RS profile at SZA ∼ 83◦ . Solid curve is the model Ne profile with Φi = 0 and at top boundary. Dotted curve is the model Ne profile with Φi = 1 × 107 cm2 s−1 while dashed curve is the model Ne profile with Φi = 3 × 107 cm2 s−1 .

an altitude of ∼135 km in good agreement with the measured Ne profile. However, the magnitude of the model Ne profile above the peak is found to deviate upward from the measured Ne profile. This clearly suggests that the plasma in the topside ionosphere observed by MGS is not in diffusive equilibrium owing to magnetic field effects. Note that below the main ionospheric peak, the model electron density is underestimated presumably because of lower photoelectron impact ionization rates used in the model.34 It may also be possible that soft X-ray impact ionization rate for this high latitude Ne profile of the southern hemisphere is too small at the time of observation.15,17,23 Because of short recombination time constant, the calculated ionosphere is in photochemical equilibrium from the lower boundary up to an altitude of ∼140 km. Above this height, the ionosphere

July 13, 2012 9:4

64

ADGEO

9in x 6in

b1389-ch05

T. Majeed, S. W. Bougher & S. A. Haider

is predominantly controlled by vertical diffusion of ions and is shown to be in diffusive equilibrium. This type of ionosphere is very similar to those completely shielded from solar wind and in which a strong intrinsic magnetic field is present. Figure 1 also shows model Ne profiles obtained by modifying the top boundary to non-zero flux condition to simulate the upward flow of plasma in the the upper ionosphere of Mars. Ness et al.36 using the Viking and Mariner Ne data showed that the magnitudes of the topside Ne scale heights are smaller for the regions of enormously large magnetic field either due to large solar wind dynamic pressure or remanace of crustal magnetization compared to those observed for weak crustal magnetic field. Later, Krymskii et al. 35 using the MGS/RS data derived Ne scale heights for the intervals; 145–165 km and 165–185 km and concluded that the average Ne scale height is about a factor of two smaller for the regions of large-scale localized magnetic field (southern hemisphere) compared to the regions of low or no magnetic field. One plausible reason for the small Ne scale height for the Martian topside ionosphere and its insensitivity to the variation of electron temperature is the presence of strong horizontal magnetic field. Such a magnetic field must inhibit the upward flow of plasma and reduce Ne scale height above the altitude of photochemical equilibrium (>140 km). Since O+ 2 is the major ion at and above the main ionospheric peak, + 7 2 −1 and 3 × 107 cm2 s−1 at we set fluxes of O+ 2 (Φ(O2 )) as 1 × 10 cm s the top boundary to investigate the impact of upward flow of ions due to the horizontal magnetic field on the topside Martian ionosphere. Clearly, the topside plasma with imposed fluxes at the top boundary would no longer be in diffusive equilibrium. It is also important to note that the ion densities in the transport region for altitudes >140 km are sensitive to these outflow fluxes while the ionosphere below this heights is still in photochemical equilibrium owing to the much faster ion recombination rate. The reduction of topside Ne density in response to upward fluxes of O+ 2 can be seen in is transported vertically Fig. 1. This reduction is expected because O+ 2 upward from the source region to the top boundary in the model. For the upward flux of 1 × 107 cm2 s−1 , the model Ne density is reduced by factor of 1.33, while this factor is increased to almost 2 for the upward flux of 3 × 107 cm2 s−1 . For the solar minimum conditions, these estimates are reasonably in agreement with those calculated by Fox43 who used the 7 2 −1 to explain the maximum allowed upward flux of O+ 2 of 4.7 × 10 cm s + Viking O2 density profile. The scale heights of the topside Ne profiles in

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch05

Plasma Transport Processes in the Topside Martian Ionosphere

65

our model are also reduced. For the upward flux of 3 × 107 cm2 s−1 , the Ne scale height of the lower part of the transport region (165–185 km) is estimated to be ∼24 km, consistent with an average scale height derived from the MGS Ne profiles.35 The Ne scale heights are also estimated for two regions: 200–250 km and 250–400 km for the two values of upward fluxes. We find that these scale heights are smaller by a factor of up to three compared to the Ne scale of the topside ionosphere in diffusive equilibrium.36 It is important to note that the maximum upward flux incorporated in the 7 −2 −1 43 s . model is comparable to the O+ 2 ionization rate of ∼3 × 10 ions cm 7 2 −1 An upward flux much larger than 3 × 10 cm s cannot be sustained in the model because we soon run into diffusion flow for these particular parameters. An alternative scenario for reducing the model scale height of the topside ionosphere is to use WD in the model as a free parameter. On the Earth the component of meridional wind parallel to the magnetic field can move the plasma up or down the field lines depending on which direction the wind is blowing. It can also be used to prevent the decay of the main ionospheric layer by maintaining the peak at higher altitudes where the chemical time constants are too long.47 Figure 2 shows the impact of downward drift on the Martian ionospheric structure. For the results shown in Fig. 2, the flux at the top boundary is assumed to be zero. Note that the effects of WD up to −50 m/s on the model Ne peaks are not quite large given the substantially short recombination time constant (∼250 s). The peak Ne is still in photochemical equilibrium. However, above the main Ne peak the plasma distribution is quite sensitive to the vertically propagating WD . As can be seen in Fig. 2 the effect of small WD of −10 m/s is substantial because of long transport time constants. As the magnitude of downward drift increases from −10 ms−1 to −50 ms−1 , plasma from the topside ionosphere is transported further and further down to the region of ionospheric peak where it recombines at much faster rate. This process reduces the topside Ne density while the peak Ne remains at around 9 × 104 cm−3 at about 135 km. Clearly, the topside ionospheric Ne for WD = −10 ms−1 , −20 ms−1 , and −50 ms−1 , in the transport region for altitudes >145 km are reduced by a factor of 2, 4, and 20, respectively, compared to the ionospheric Ne profile for diffusive equilibrium conditions. The corresponding Ne scale heights are also reduced. The estimated ratios of the Ne scale heights for the two altitude regions (145–165 km/165–185 km) are in the range 1.6–2.1 in reasonably good agreement with those derived from the MGS Ne data by Krymskii.35

July 13, 2012 9:4

66

ADGEO

9in x 6in

b1389-ch05

T. Majeed, S. W. Bougher & S. A. Haider

Fig. 2. A comparison of model Ne profiles with the MGS/RS Ne profile at SZA ∼ 83◦ . Solid curve is the model Ne profile with Φi = 0 and WD = 0. Dotted, dashed, and dashed-dotted curves are the model fits to the MGS/RS Ne profile with WD = −10 ms−1 , −20 ms−1 and −50 ms−1 respectively.

4. Summary We have developed a 1-D chemical diffusive model which includes plasma transport processes to investigate the variation of the topside Martian ionosphere observed with the radio occultation experiments on board MGS spacecraft. We show that topside ionosphere observed in the region with large-scale localized magnetic field appears to violate diffusive equilibrium condition that one can expect for the magnetic field free ionosphere. For such an ionosphere, the topside Ne scale height is smaller than that calculated with diffusive equilibrium condition (Φi = WD = 0). Thus, to explain the observed Ne profiles, one must invoke processes of vertical transport of plasma in the Martian upper atmosphere. We use a vertical drift velocity of −20 ms−1 in our 1-D model to obtain a reasonable fit to the MGS Ne

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch05

Plasma Transport Processes in the Topside Martian Ionosphere

67

profile. In addition, it appears that upward flow of plasma may play a role in interpreting the measured ionospheric Ne profiles. We use an upward 7 2 −1 to explain the measured MGS Ne profile. This flux of O+ 2 of 3 × 10 cm s flux is almost a factor of 1.5 smaller than that used by Fox43 to interpret the O+ 2 density profile measured by Viking spacecraft. We also find that the variation in the topside Ne can be modeled by using upward fluxes of 7 2 −1 . Because Ne profiles of the topside O+ 2 ions in the range 1−3 × 10 cm s ionosphere up to 400 km from anomalous regions of the Martian southern hemisphere are not available, these exploratory results are useful for future modeling and measurements of the upper ionosphere of Mars.

Acknowledgments We acknowledge support for this work through the Faculty Research Grant from the Office of Research at American University of Sharjah, UAE. TM also wishes to acknowledge support from NASA MDAP grant NNG04G192G. SWB wishes to acknowledge support from NASA MDAP grant NNX10A017G.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

A. J. Kliore et al., Science 149 (1965) 1243. G. Fjeldbo, W. C. Fjeldbo and V. R. Eshleman, Science 153 (1966) 1518. G. Fjeldbo, A. Kliore and B. Seidel, Radio Sci. 5 (1970) 481. M. A. Kolosov, V. M. Ivanov, D. S. Lukin and Y. G. Spiridonov, Space Res. 16 (1976) 1013. A. J. Kliore et al., Icarus 17 (1972) 484. M. B. Vasiliev et al., Kosm. Issled. 13 (1975) 48. G. Fjeldbo et al., J. Geophys. Res. 82 (1977) 4317. G. F. Lindal et al., J. Geophys. Res. 84 (1979). G. L. Tyler et al., J. Geophys. Res. 106 (2001) 23,327. D. D. Morgan et al., J. Geophys. Res. 113 (2008) A09303. M. P¨ atzold, et al., Science 310 (2005) 837. P. Withers, Radio Sci., 46 (2004). J. L. Fox and K. Y. Sung, J. Geophys. Res. 161 (2001) 21305. S. A Haider, S. P. Seth, V. R. Choksi, and K. I. Oyama, Icarus, 185 (2006) 102–112. J. L. Fox, J. Geophys. Res. 109 (2004) A11310. S. A. Haider, K. K. Mahaja, and E. Kellio, Rev. Geophys. 49 (2011) RG4001. S. W. Bougher, S. Engel, D. P. Hinson and J. M. Forbes, Geophys. Res. Lett. 28 (2001) 3091. J. M. Forbes et al., J. Geophys. Res. 107 (2002) 5113.

July 13, 2012 9:4

68

ADGEO

9in x 6in

b1389-ch05

T. Majeed, S. W. Bougher & S. A. Haider

19. R. J. Wilson, D. Banfield, B. J. Conrath and M. D. Smith, Geophys. Res. Lett. 29 (2002) 1684. 20. S. W. Bougher, S. Engel, D. P. Hinson and J. R. Murphy, J. Geophys. Res. 109 (2004) E03010. 21. J. M. Keating et al., Science 279 (1998) 1672. 22. J.-S. Wang and E. Nielsen, Planet. Space Sci. 52 (2004) 881. 23. S. W. Bougher, S. Engel, R. G. Roble and B. Foster, J. Geophys. Res. 105 (2000) 17,669. 24. Wang, J.-S and E. Nielsen, Planet. Space Sci. 51 (2003) 329. 25. M. B. McElroy, T. Y. Kong and Y. L. Yung, J. Geophys. Res. 82 (1977) 4379. 26. M. H. G. Zhang, J. G. Luhmann, A. J. Kliore, and J. Kim, J. Geophys. Res. 95 (1990) 14,829–14,839. 27. R. W. Schunk and A. F. Nagy, Ionospheres: Physics, Plasma Physics, and Chemistry (Cambridge Univ. Press, New York, 2000). 28. M. H. Hantsch and S. J. Bauer, Planet Space Sci. 38 (1990) 539. 29. S. J. Bauer and M. H. Hantsch, Geophys. Res. Lett. 16 (1989) 373. 30. C. Martinis, J. K. Wilson and M. J. Mendillo, J. Geophys. Res. 108 (2003) 1383. 31. H. Rishbeth and M. Mendillo, Planet Space Sci. 52 (2004) 849. 32. T. K. Breus et al., J. Geophys. Res. 109 (2004) A09310. 33. J. L. Fox and K. E. Yeager, J. Geophys. Res. 111 (2006) A10309. 34. Y. Ma, A. F. Nagy, I. V. Sokolov and K. C. Hansen, J. Geophys. Res. 109 (2004) A07211. 35. A. M. Krymskii et al., J. Geophys. Res. 109 (2004) A11306. 36. N. F. Ness, et al., J. Geophys. Res. 105 (2000) 15,991. 37. A. F. Nagy and T. E. Cravens, Solar system ionospheres, Atmospheres of the Solar System: Comparative Aeronomy, Geophys. Monogr. Ser., eds. M. Mendillo, A. Nagy and J. H. Waite (AGU, Washington, D. C., 2002) 30–54. 38. H. Shinagawa and T. E. Cravens, Geophys. Res. 94 (1989) 6506. 39. H. Shinagawa and T. E. Cravens, J. Geophys. Res. 93 (1988) 1027–1035. 40. Y.-J Ma and A. F. Nagy, Geophys. Res. Lett. 34 (2007) L08201. 41. R. Lundin et al., Nature, 341 (1989) 609. 42. H. Rosenbaur et al., Nature, 341 (1989) 612–614. 43. J. L. Fox, Geophys. Res. Lett. 24 (1997) 2901. 44. S. A. Haider et al., J. Geophy. Res. 115 (2010) A08310. 45. S. K. Jain and A. Bhardwaj, J. Geophys. Res. 116 (2011) E07005. 46. M. O. Fillingim et al., Icarus, 206 (2010) 112–119. 47. P. M. Banks and G. Kockarts, Aeronomy (Elsevier, N.Y., 1977).

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch06

Advances in Geosciences Vol. 30: Planetary Science and Solar & Terrestrial Science (2011) Ed. Anil Bhardwaj c World Scientific Publishing Company


THE 3D ANALYSIS OF THE HELIOSPHERE USING INTERPLANETARY SCINTILLATION AND THOMSON-SCATTERING OBSERVATIONS B. V. JACKSON Center for Astrophysics and Space Sciences, University of California, San Diego 9500 Gilman Dr. #0424, La Jolla, CA 92093-0424, U.S.A. [email protected] http://smei.ucsd.edu/ http://ips.ucsd.edu/

Both interplanetary scintillation (IPS) and Thomson-scattering observations from the U.S. Air Force/NASA Solar Mass Ejection Imager (SMEI) allow a determination of velocity and density in the inner heliosphere and its forecast from remote-sensing heliospheric observations. Recent solar missions, such as Hinode, STEREO, and SDO, and resultant modeling analysis using these data enhance our ability to measure detailed aspects of specific solar events, including their outflow and three-dimensional structure. Current success in this 3D heliospheric endeavor includes the analysis of heliospheric structures that are also measured in situ: interplanetary Coronal Mass Ejections (CMEs), shocks, solar co-rotating structures, and the energy transport provided by solar wind plasma throughout the heliosphere. This report highlights a portion of the work on this multi-faceted topic.

1. Introduction Beginning with observations from early coronagraphs (e.g., Jackson1 ), a variety of techniques have been explored to provide the three-dimensional (3D) structure of the corona and heliosphere. When heliospheric imaging first began2,3 using interplanetary scintillation (IPS) techniques,4 it was clear that views of heliospheric structure over time could provide information about their 3D extent (e.g., Gapper et al.,2 Hewish and Bravo,5 Behannon et al.6 ). The first of these techniques involved the use of different structure templates and the “by-eye” fitting of these according

69

July 13, 2012 9:4

70

ADGEO

9in x 6in

b1389-ch06

B.V. Jackson

Fig. 1. LOS weighting functions. (a) The weak-scattering IPS weighting function at 327 MHz assuming a source size of 0.1 arc seconds (from Jackson et al.19 ). (b) The Thomson-scattering weight function. Three samples are given at angles from the Sun of 16◦ , 31◦ , and 90◦ (from Jackson and Hick22 ).

to their line-of-sight (LOS) weighting response (Fig. 1). A second source of heliospheric remote sensing was introduced when it was proved that the same heliospheric structures registered a Thomson-scattering brightness response7 in the photometers8 of the Helios spacecraft. IPS measurements have been used to track heliospheric structures outward from the Sun since the beginning of their use in heliospheric imaging. A recent review of many of these early techniques is given in Jackson et al.9 and references therein. In the late 1970’s coronagraph techniques did not provide very clear images of Coronal Mass Ejections (CMEs). Ambiguities in IPS analysis were made using the 80 MHz Cambridge, England observations, and the comparisons of regions on the Sun, with in-situ measurements led Hewish and Bravo5 to describe most of the rapidly outward moving heliospheric structures observed as “coronal holes” to the consternation of many heliospheric physicists of that time. Later, measurements from more contemporary IPS arrays and data sets from the Large Angle Spectrographic COronagraphs (LASCO)10 flown on the Solar and Heliospheric Observatory (SOHO) spacecraft,11 clearly showed that CMEs, or perhaps the turbulent shocked plasma behind some CMEs, constituted the majority of rapidly-moving transient features observed in the heliosphere. The analysis of Solar Mass Ejection Imager (SMEI)12,13 data at the University of California, San Diego (UCSD) was developed to provide the same basic two-dimensional (2D) imaging input as IPS, but at a far higher temporal cadence, and precision in measuring density over

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch06

The 3D Analysis of the Heliosphere Using Interplanetary Scintillation

71

elongations (angular distances from the Sun) as great as 180◦ . Unlike IPS, the Thomson-scattering process is optically thin and directly related to heliospheric electron density by geometrical considerations alone.14 The originally-planned optical precision of SMEI observations is well-achieved in many sky locations most of the time,15 but the data are sometimes contaminated by aurorae near Earth above the polar-orbiting spacecraft.16 Thus SMEI performance is less than optimal at just those times of most interest for geoeffective studies and forecasts. Removing the auroral signatures from the SMEI images has been one of the most challenging aspects of these analyses to date. The LOS integration intrinsic to both IPS and Thomson-scattering observations precludes direct determination of the locations of outward-moving heliospheric structures, and thus a comparison with in-situ measurements requires additional more complicated analysis. A technique was developed at the UCSD over the years17 aimed at formalizing the determination of the 3D extent of heliospheric structures by using the LOS response in either IPS (Fig. 1a) or Thomson-scattered light (Fig. 1b) data, and iteratively fitting these from only a few viewing locations, while making as few assumptions about the structures as possible.18 Over the years this iterative fitting became a more formal computational procedure, given that such a system was necessary to yield the greatest information about 3D structure from heliospheric data sets such as IPS, Helios, and SMEI, and that the procedure could do this from a single point in space (Fig. 2 and next paragraphs). For a review, see Jackson et al.9 The UCSD tomographic analysis technique explicitly takes into account the 3D extent of heliospheric structures, including the fact that the greatest contribution comes from material closest to the Sun, but without any

Fig. 2. Depiction of the perspective views from a single observing location (Earth) as material moves outward from the Sun (from Jackson et al.23 ). This, and the LOS weighting change as the material flows outward, provide the necessary information about the shape of heliospheric structures.24

July 13, 2012 9:4

72

ADGEO

9in x 6in

b1389-ch06

B.V. Jackson

explicit assumption about the distribution of velocity and density along their lines of sight. Thus, it reconstructs 3D solar wind structures from remote-sensing data gathered at a single location, such as are available from the IPS and SMEI observations. Developing this technique19−22 was necessary in order to tap the full potential of IPS and visible-light heliospheric imagers and to enable subsequent analysis as a predictive tool for scientific research and space-weather purposes. Transients such as CMEs evolve on short time scales (hours to days). In the case of observations covering a wide range of solar elongations, heliospheric structures are seen from widely different perspectives as they move past Earth. This feature, which is absent from coronagraph and most other solar remote-sensing data, allows time-dependent 3D-reconstruction of transient structures. Presently, our time-dependent 3D-reconstruction incorporates a purely kinematic solar wind model. Given the velocity and density on an inner boundary (the “source surface”), a fully 3D solar wind model best fitting the observations can be derived by assuming radial outflow and enforcing conservation of mass and mass flux.19 Best fit is achieved iteratively. If the 3D solar wind model does not match the overall observations, the source surface values are suitably altered to minimize the deviations. This technique is employed to successfully analyze CME-associated structures in an exploratory sense using IPS observations. A website is operated at: http://ips.ucsd.edu that has utilized this technique since the year 2000 and this provides data analysis in near real time. Also, the technique has been extended for use with visible-light brightness data from SMEI, and most recently for use with IPS archival data from Ootacamund (Ooty), India,25,26 to analyze the 6–8 November 2004 CME sequence.27 See also Bisi et al.28 and Hara et al.29 for Solar-Terrestrial Environment Laboratory (STELab), Japan,30 and SMEI data analysis of these events. The abundant velocity measurements from the Ooty single-site radio telescope provide truly outstanding data that match in-situ velocities made during this complex series of events. The analyses from STELab, SMEI, and Ooty have been compared successfully with in-situ measurements and examples of these based on IPS analysis are presented on the UCSD website. The 3D analysis and comparisons with near-Earth in-situ monitors and also with results from the Mars Global Surveyor,31 and with measurements made aboard the Solar-Terrestrial Relations Observatory (STEREO)32 spacecraft confirm that the IPS analysis provides accurate heliospheric density and velocity measurements throughout the heliosphere.

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch06

The 3D Analysis of the Heliosphere Using Interplanetary Scintillation

73

Here we report on just a few of the latest analyses accomplished using this 3D technique. Section 2 briefly describes the time-dependent tomographic analysis routines developed by our group at UCSD for fitting STELab IPS velocity and g-level data, and SMEI brightness data. Section 3 provides a set of recent observations and analyses. These are discussed briefly in Sec. 4. We conclude in Sec. 5.

2. 3D-Reconstruction Analysis The mathematics of this technique are described in detail in Hick and Jackson,33 and Jackson et al.,24 and the reader can refer to these articles for more information than is given here. Also, computational aspects of the UCSD 3D-reconstruction program have been discussed in many other articles over the past decade.19,33,34 In early analyses it was assumed that the heliosphere co-rotates with the Sun. In more recent work20,21,35−39 this assumption has been relaxed. In the present case, LOS segments and their 3D weights are projected back in space and time to a solar wind inner boundary (a source surface) that is set at a given height (usually 15 RS ) that lies below the closest approach of all lines of sight to the Sun. Each LOS is mapped from Earth and each segment of it is projected to the source surface taking into account the relevant velocities and interactions from the model that provides the solar wind outward motion (see Fig. 3). In current analyses, the inversion process adjusts boundary conditions using a kinematic 3D solar wind model to best fit the observations and employing a least-squares fitting procedure. This minimizes the differences between modeled and observed SMEI brightness, or modeled and observed IPS g-value and velocity values, or a combination of these. As explained elsewhere,24,39 the mean solar wind Thomson-scattering signal from SMEI is difficult to distinguish from the very bright zodiacal light signal. Because of this, reconstructions based on SMEI data (unlike IPS g-level data) require that a mean ambient solar wind be included in the solar wind model as well as the observed Thomson-scattering brightness based on the average in-situ solar wind density at 1 AU. A least-squares fitting program developed specifically for this type of analysis inverts the weighted, projected model values on the 2D innerboundary source surface at different time steps, in order to provide solar wind model outflow parameters. These values are directly inverted on the source surface at the appropriate times to yield new solar wind parameters, and these latter are iteratively converged for each data set.

July 13, 2012 9:4

74

ADGEO

9in x 6in

b1389-ch06

B.V. Jackson

Fig. 3. Sample LOS segment mapping to the source surface over 360◦ (for one Carrington map at a given time). There is a slight displacement to the left over time from solar rotation; top to bottom. The mapping shows the coverage within the time interval from each segment. The Earth sub-point on these trace-back plots is at the approximate center. (left) Velocity LOS IPS projections from one day to another on 14 and 15 July 2000 (for time intervals of one day). (right) The Thomson-scattering LOS projections from SMEI from valid segments on ∼5◦ image centers for two half-day time intervals using this same plot format during Carrington rotation 2068 (21 March–17 April 2008). There is far more spatial coverage from the SMEI images shown in these plots, and thus a far higher spatial and temporal resolution possible in SMEI analysis (from Jackson et al.24 ).

In the fitting process, ratios of modeled-to-observed values and a modeled-to-observed χ2 are monitored to indicate the rate of convergence for the interval studied. Velocity and density corrections to the 3D model are made separately. First, the inversion changes are made to previous velocity conditions on the inner boundary surface. Second, the 3D solar wind model is updated and new projected locations of each LOS point on the inner-boundary surface are determined. Third, inversion changes are made to previous density boundary conditions on the inner boundary surface. Finally, the 3D model is again updated with all the newest boundary values. The inner-boundary Carrington maps of velocity and density are smoothed at every iterations using a 2D Gaussian spatial filter that incorporates equal-solar-surface areas, as well as a Gaussian temporal filter. Locations in the model that are not accessed by the above iterative procedure (and thus remain undetermined) are left blank in the final

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch06

The 3D Analysis of the Heliosphere Using Interplanetary Scintillation

75

result. For the analysis presented here, these blank places include sections of heliospheric volume on the opposite side of the Sun from Earth that cannot be accessed and thus not reconstructed with the resolutions of the 3D volume at the digital resolution used. For SMEI, this includes a large fraction of the region behind the Sun because the instrument does not view close to the solar surface. This blank volume is usually much smaller for the 327 MHz IPS data which can often view to within 11.5◦ of the Sun. The reconstruction program generally converges to an unchanging model within a few iterations, but operates for nine iterations to guarantee convergence.19 For a typical rotation and the digital resolutions of the current SMEI data sets, the density and velocity iterations generally take R Core i7 computer. The about 15 minutes to process using a 2.4 GHz Intel
IPS data sets normally take only a few minutes to process. Normally those IPS-velocity observations and SMEI-brightness lines of sight throughout the period that do not fit within a three-sigma limit of the mean ratio change ascribed at that location by the model (typically ∼1% of the SMEI brightness or the IPS velocity line of sight) are discarded. This provides a safeguard by removing outliers which do not fit the model values. The program then operates for nine more iterations (18 in total). The solutions are insensitive to the initial model values and, after a few iterations, any residue of the initial values has disappeared. Tests19 have shown that the 3D-reconstruction of a set of artificial observations using a known 3D input successfully reproduces the input.

3. Recent Observations and Analysis Examples of the 3D analysis procedure are presented below in four subsections: The first shows comparisons of velocity and density at Earth from the IPS data that have been verified by in-situ measurements; the second presents examples of our analysis that show both a CME and a co-rotating region using SMEI Thomson-scattering data compared with a current 3D-MHD model. The third gives measurements of a CME that was observed near the Sun in coronagraph observations and that was also measured in situ at the STEREO-B spacecraft situated 72◦ east of the Sun–Earth line; the fourth and final subsection describes a speculative recent analysis with tomography using IPS observations and the full SMEI data set. This latter example details a structure that was first observed in a coronagraph and then in the IPS and SMEI 3D-reconstructions, and that appears to be a manifestation of jets observed in Hinode data.

July 13, 2012 9:4

76

ADGEO

9in x 6in

b1389-ch06

B.V. Jackson

Fig. 4. Time-dependent reconstruction during Carrington rotation 1965 (10 July to 4 August 2000) using STELab IPS g-level data. (see Jackson et al.21 ). Upper Left: Density distribution as seen by an observer at 3 AU, 30◦ above the ecliptic, at the time indicated. The Sun, Earth, and Earth’s orbit are indicated in the image. The main structure near Earth is associated with a halo CME (the “Bastille-Day” CME) observed by LASCO on 14 July 2000. The reconstruction has a resolution of 20◦ × 20◦ in latitude and longitude, 0.25 AU in radial distance, and has a time cadence of one day. Lower Left: Ecliptic cut from the Sun outward to 1.5 AU of the same CME. In this and the previous image an r−2 density fall-off has been removed from the volume to better view structure from near the Sun to farther away. Center: Time series and correlation of reconstructed density at Earth and measured proton density by the Advanced Composition Explorer (ACE) spacecraft. ACE observations are combined into 18-hour averages commensurate with the resolution of the time-dependent model. Correlation has been limited to data times within five days of the event. Right: Same for the reconstructed velocity and ACE velocity observations.

3.1. The 14 July 2000 (Bastille-Day) CME The 3D time-dependent reconstruction technique was used to analyze CMEassociated structures using IPS g-level and velocity observations. Figure 4 shows an example of the “Bastille-Day” CME of 14 July 2000.21 Gaussian filters were applied to the data sets to restrict structure size to larger than the digital resolution.35,39 Here the dense structure reconstructed in the Bastille-Day CME event essentially traces the lower portion of a magnetic flux rope cylinder.40,41 The magnetic loop extent can only be inferred from its passage by Earth (using

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch06

The 3D Analysis of the Heliosphere Using Interplanetary Scintillation

77

Advanced Composition Explorer (ACE) and other near-Earth spacecraft in-situ observations), and by the Near Earth Asteroid Rendezvous (NEAR) spacecraft situated at ∼1.76 AU nearly on the Sun–Earth line. This cylinder was a huge structure with the same approximate orientation as the reconstructed density. Thus, the 3D-reconstructed density for this feature mimics the flux rope cylinder to the east and west of the Sun derived from the magnetic analysis of data measured at the Earth and at the NEAR spacecraft.

3.2. The 26 April 2008 CME and co-rotating region A simple kinematic solar wind model is presently the kernel of the UCSD 3D-reconstruction technique. At current resolutions, the kinematic modeling provides adequate 3D analysis near Earth. However, the physics behind this modeling becomes inadequate when either near the Sun or very distant from it, or when exploring shock processes. Numerical solar wind models based on the equations of magnetohydrodynamics (MHD) are currently the only self-consistent mathematical descriptions, capable of bridging many AU, from near the Sun to beyond Earth’s orbit. Although MHD provides only an approximation of actual plasma behavior, these models have successfully simulated many important space plasma processes. Some MHD algorithms are available from sources such as the NASA Goddard Community Coordinated Modeling Center (CCMC). One such model ENLIL (Odstrcil et al.42 ) is based on the ideal 3D MHD description, with two additional continuity equations for tracking the injected CME material and the magnetic field polarity (see Odstrcil and Pizzo43 ). Odstrcil has used boundary conditions available from our iteratively-fit kinematic 3D-reconstructions to drive the time-dependent heliospheric ENLIL model.44 The MHD modeling, adjusted to match in-situ parameters, agrees well with global 3D-reconstructions. With the addition of an input from measurements of observed near-solar CMEs, and a “cone model” approximation to these45 (for example, see Fig. 5a), the 3D-MHD densities and kinematic model 3D-reconstruction densities agree reasonably46,47 with SMEI 3Dreconstructions (Fig. 5b). The MHD modeling in this example shows the timing of the ICME density response at the STEREO-B spacecraft, and the interaction between the co-rotating interacting region and the ICME. SMEI also shows both structures. This analysis using our 3D-reconstruction is described further in other articles,47−50 which also describes various other

July 13, 2012 9:4

78

ADGEO

9in x 6in

b1389-ch06

B.V. Jackson

Fig. 5. Density comparison cuts in the ecliptic of the ENLIL 3D-MHD model that includes a cone model approximation, and the kinematic-model SMEI 3D-reconstruction of an ICME that arrives at STEREO-B on 29 April 2008 (from Jackson et al.47 ). An r−2 density fall-off has been removed from both ecliptic cuts. (a) The ICME that reaches STEREO-B at this time is encircled by a black line. Additional features are annotated on the plot. A dashed line “streamline” connects Sun to Earth. (b) Density plot from the SMEI 3D-reconstruction. The ICME ecliptic response is enclosed by an ellipse; co-rotating structure about to reach STEREO-B is shown by a dashed line.

ICME modeling techniques51,52 that have been used to derive the 26 April 2008 ICME event and other CME shapes, as well as the solar wind density and velocity both in and out of the ecliptic. Although we have highlighted the ENLIL 3D-MHD model in the above analysis, several other 3D-modeling efforts have compared well with past SMEI tomographic analysis. For the halo CMEs of 27–28 May 2003, comparisons between the Hakamada, Akasofu, and Fry, version 2 (HAF v2) kinematic model,53 and the UCSD sky map of the event at 18 UT 29 May show very similar results. Also, in future efforts we could compare our results with other 3D-MHD modeling efforts such as the BATS-R-US code,54 or with 3D-MHD codes by Wu et al.,55 Feng et al.,56 and Detman et al.57 These 3D-MHD modeling procedures all begin by using solar magnetic fields to derive background solar wind parameters then, as in the case of ENLIL with its cone model input, some use energy inputs near the solar surface to propagate transient solar wind parameters. While these are all defensible forward-modeling procedures applying near-solar-surface inputs in a 3DMHD model, they are used regularly only for larger CME events, and none match heliospheric remotely-sensed observations iteratively.

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch06

The 3D Analysis of the Heliosphere Using Interplanetary Scintillation

79

3.3. The January 2010 CME events SMEI 3D-reconstructions have been applied for many CME event studies, but nowhere has this analysis been more productive than for the CME events of January 2010 which erupted near the solar east limb. This disturbance arrived in situ at the STEREO-B spacecraft at ∼18 UT 20 January, and included a small density enhancement and a day-long magnetic field enhancement and rotation. The event passed the STEREO-B spacecraft at an average speed of ∼320 km s−1 . Using the SMEI 3Dreconstruction analysis technique and this speed we obtained an extremely good match with the density obtained at the time of the event in the data of STEREO-B, which was 72◦ east of the Sun–Earth line (Fig. 6). This insitu fit during the event provides good assurance that the density structure reconstructed in 3D has the actual shape of the CME at the STEREO-B location. Figure 7a presents a 3D-reconstructed difference image of the CME when it was centered on the STEREO-B spacecraft at ∼55◦ elongation to the east on the ecliptic. Figure 7b shows an ecliptic cut of the CME at the same time, and thereby provides a description of the shape of the CME density structure in the ecliptic.

Fig. 6. (a) Density time series from the time-dependent 3D-reconstruction using SMEI brightness data compared with STEREO-B density data during the 15-day interval that the CME that passed STEREO on 20–21 January 2010 72◦ east of the Sun–Earth line. The dashed line time series is the reconstruction result, and the continuous line presents the in-situ measurements from STEREO-B. Density is presented from STEREOB with a boxcar average of 0.5 day to approximately match the analysis from SMEI brightness. (b) Correlation between the two time-series over the interval shown.

July 13, 2012 9:4

80

ADGEO

9in x 6in

b1389-ch06

B.V. Jackson

Fig. 7. (a) SMEI brightness difference image obtained from the 3D-reconstruction analysis by subtracting one volumetric data set 12 hours prior to the one indicated. In this “fisheye” presentation the Sun is centered with the largest elongation shown ∼110◦ . An r−2 density fall-off has been removed from the volumes to better show structures distant from the Sun with the same brightness near it. (b) Density ecliptic cut at the same time showing the CME structure as it passes STEREO-B during the event. The Earth is indicated on its orbit to the right of center with the locations of STEREO-A (above and right) and STEREO-B (below and right) shown as small circles near Earth’s orbit. The ecliptic density manifestation of the CME is an arch that follows the ecliptic over more than 60◦ to the east and west of STEREO-B. A second CME is nearing Earth at this same time.

The SMEI 3D analysis when extended back to the solar surface at a constant speed of 320 km s−1 can be used to show what the CME looked like earlier in the LASCO coronagraphs and these reconstructions can be compared with LASCO C3 images obtained at the appropriate time periods. The density enhancement present in the STEREO-B in-situ analysis on 21 January can be traced back directly to the CME event near the solar surface that erupted late on 14 January 2010 (Fig. 8). A later density enhancement that passed STEREO-B on 23–25 January that is also shown nearing Earth in the Fig. 7b ecliptic cut, can be traced back toward the solar surface at the same speed and is shown to have left the Sun on 17 January. The coronagraph brightness is decreased in its inner portion near the Sun due to vignetting of the inner field of the image. The SMEI pseudo coronagraph observations approximately match this vignetted brightness fall-off by providing a measurement of the inner corona brightness from data that had an r−2 density fall-off imposed on the volumetric analysis.

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch06

The 3D Analysis of the Heliosphere Using Interplanetary Scintillation

(a)

81

(b)

Fig. 8. (a) Brightness obtained using the 3D-density reconstruction from SMEI traced back to near the solar surface at the speed of 320 km s−1 , and presented as Thomsonscattering brightness viewed in LASCO C3 observations. The density has had an r−2 density fall-off applied, and is shown calibrated in S10 relative to the brightness the density would have at 1 AU. (b) A SOHO LASCO C3 image obtained showing a CME over the solar east limb that erupted from the solar surface late in the day on 14 January 2010.

The density enhancement associated with the magnetic cloud at STEREO-B is of great interest in this case because this enhancement was observed at the center of classical magnetic structure58 associated with the CME, and can thus allow its orientation to be determined in 3D. We were indeed able to reconstruct this structure in 3D from the SMEI observations, determine its density orientation, and compare this with the 3D structure derived from in-situ magnetic field measurements as shown in these preliminary analyses. 3.4. The heliospheric response to jetting Hinode (Solar B),59 which was launched on 23 September 2006 to a polar Sun-synchronous orbit at about 600 km above the Earth, has a complement of three instruments that includes the X-Ray Telescope (XRT).60,61 With limited data downlink capability, the XRT has sometimes been run in campaign mode where images from the instrument from smaller area than the full solar disk are provided at a high temporal cadence. Figure 9 shows one image in a high-cadence (∼1 min per image) sequence from XRT. Operated in conjunction with the Solar Optical Telescope (SOT)62−64 on board Hinode, these images, especially in polar regions, show solar

July 13, 2012 9:4

82

ADGEO

9in x 6in

b1389-ch06

B.V. Jackson

07:14:48

07:14:48

07:18:48

07:22:48

Fig. 9. Hinode XRT images using an Al poly filter showing the region that produces jets over the north solar pole. The base of the most dominant of these at this time (at 56◦ E, 57◦ N ecliptic relative to Earth) is shown enlarged in the three bottom panels. The jet ejecta response moves away from the Sun and eastward relative to its solar surface location.

jets and the locations of vertically-oriented flux tube structures65 that are nearby. Figure 10 from Shimojo and Tsuneta66 depicts the general structure surrounding the jetting region. Tsuneta et al.65 describe the verticallyoriented flux tubes that have an average maximum field strength of 1.5 K Gauss, as “kG-patches”, and note that in any given polar region they all have the same sign, which is consistent with the polar magnetic field. If the flux tubes extend into interplanetary space, they have the possibility to serve as guide fields for X-ray jets, coronal plumes, and the fast solar wind.66 Tsuneta et al.65 remarked that the kG-patches probably fan out to provide all the open magnetic field in a coronal hole, and that these structures serve as the channels of the fast solar wind. However, Shimojo and Tsuneta66 also concluded that since X-ray jets occur near only a small portion of the kG-patches, it is unlikely polar X-ray jets provide sufficient

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch06

The 3D Analysis of the Heliosphere Using Interplanetary Scintillation

83

Fig. 10. The magnetic field configuration of the region around a jet as depicted by Shimojo and Tsuneta.57 The X-ray jet is shown centered in the top of the image and moves outward adjacent to the kG-patch.

energy for the acceleration of the fast solar wind, and that the energy of the solar wind is most likely provided by weak activity surrounding the kG-patches. Data sets from STELab (and SMEI) were used interactively at the CCMC in these 3D analyses67 to provide quick comparisons with other data sets. Time-dependent 3D-reconstruction analyses from the IPS data sets have fairly low spatial and temporal resolutions (cadences of about one day, and latitudinal and longitudinal resolutions of ∼20◦ ). Volumetric measurements feature a factor of four finer resolutions in order to smoothen up and provide somewhat higher spatial resolution and data cadence for the graphical displays at UCSD and the CCMC. In general these analyses do not show smoothly-flowing plasma, nor plasma that averages to the high polar speeds over the solar poles that are expected on the basis of Ulysses in-situ observations68 (made during polar passes at >1.5 AU). An average tomographic measurement made over times greater than one day has allowed the group at STELab to split up the IPS measurements into those obtained close to the Sun and those farther away and to thereby discover a general acceleration term in the IPS velocity data set.69 In a preliminary study of the solar jet response within the heliosphere, we chose a period for archival data analysis when Hinode observed the feature shown in Fig. 9. Within the period from 19 UT 5 September to

July 13, 2012 9:4

ADGEO

b1389-ch06

B.V. Jackson

84

1

9in x 6in

2

3

4

5

Fig. 11. (left) List of jet peak X-ray peak brightness energies on 14 September 2007 (column 4, event numbers 329 to 349) in DN s−1 from ∼850 events observed in the Hinode XRT observations during a three-week interval (from Sako et al.70 ). (right) The LASCO response for the brightest of these events #343 (shown as an enhancement to the north and a depletion to the south — arrow) in difference images.

08 UT 22 September, several particularly bright X-ray jets were observed on 14 September 2007 in the northern polar coronal hole. In a study of nearly 850 jets and their associated energy analysis from the X-ray brightening, over a three-week period, Sako et al.70 determine approximate energy inputs for each event measured, and these, including the example in Fig. 9, are highlighted in a list from the Hinode data for 14 September 2007 (Fig. 11a). The columns in the table from 1–5 respectively list the event number, latitude, longitude, peak brightness, and time. Though difficult to discern except in animations of the LASCO images, these events were each observed individually in LASCO C2 difference images to have surfaceprojected onsets commensurate within minutes to the times at which they were observed as the jet peak brightness responses in Hinode. All were situated at approximately the same location on the solar surface, and the most energetic of these events provided the largest C2 response while the smallest correspondingly provided the least. The LASCO C2 observation shown (Fig. 11b) highlights the brightest of the X-ray events that occurred during this interval. Figure 12 shows an IPS velocity ecliptic-cut coordinate plot from the CCMC interactive visualization at heliographic radius r = 0.3 AU (∼70 solar radii). Although the volumetric data are available from the 3Dreconstructions at the CCMC from r = 15 solar radii (3.75◦ elongation) out to 3.0 AU, we calculate that r = 0.3 AU is the approximate location of the greatest LOS weighting, and thus this is the heliographic height of greatest certainty in the 3D fit, especially for data measured over the

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch06

The 3D Analysis of the Heliosphere Using Interplanetary Scintillation

85

Fig. 12. An ecliptic-coordinate synoptic map presentation of IPS velocity volumetric data at a radial distance of 0.3 AU from the solar surface. The Sun–Earth line is centered at zero degrees longitude and latitude in the plot. A high-speed structure is marked by the arrow, and at this time is about 40◦ E, of the Sun–Earth line and 35◦ N latitude. Present a day earlier but as a smaller velocity enhancement, the feature increases to a maximum speed at this time and then diminishes. Contour intervals on the plot are placed from 500 to 800 km s−1 , at ∼10 km s−1 interval to show only features in the generally faster-speed polar holes.

solar poles. The highlighted higher velocity region associated with the jet response has decreased in latitude by about 20◦ and is shifted somewhat westward from the jet solar surface location. The duration of the response which exceeds one day indicates that it is a composite of several responses to jetting activity observed in coronagraph observations at slightly earlier times. In the SMEI analysis, the heliospheric response is clearer because of the higher data cadence and higher resolution available. By utilizing the full SMEI data set, some 4 × 106 LOS measurements over a Carrington rotation time interval, higher and more precise resolution analyses are provided for selected intervals than can be obtained normally (i.e., from our usual analysis at the CCMC or on the UCSD website at: http://smei.ucsd. edu). The jet response for this interval covered by Figs. 9, 11, and 12 is shown in the SMEI Thomson-scattering analysis where the observations from elongations of about 45◦ are back-projected to obtain the images shown in Fig. 13. These images show the SMEI observations near the solar surface presented as the C3 coronagraph would view the solar corona in white light. Although many radial structures are visible, enhancements at the correct position angles and times indicate that the SMEI instrument indeed viewed at least a remnant of the response observed in LASCO data at much larger distances from the Sun. That these structures correspond with the LASCO coronagraph observations implies that their 3D locations are

July 13, 2012 9:4

86

ADGEO

9in x 6in

b1389-ch06

B.V. Jackson

Fig. 13. Pseudo coronagraph images of the jet response shown in Figs. 9, 11, and 12 from the SMEI 3D-reconstruction analysis. The reconstructed back-projected image is as the C3 coronagraph would observe the sky in Thomson-scattered light. An r−2 density fall-off has been removed from the brightness response shown in order to mimic the coronagraph vignetting function.

properly known, and that they can be compared with their manifestations near the solar surface. 4. Discussion The foregoing is a brief description of the exploratory type of 3Dreconstruction analysis that has been developed at UCSD over the last two decades. The current reconstruction analysis technique utilizes observations from a single location in space and assumes no a-priori knowledge of the heliospheric structure other than its LOS weighting and that the radial outward solar wind flow behaves kinematically, conserving mass and mass flux. The preceding gives examples of these analyses using both IPS and SMEI Thomson-scattering data. Both of these data sets can be used in real time to forecast heliospheric 3D density and velocity in advance of its arrival at Earth and the inner planets. An article also submitted to this journal issue71 gives a more comprehensive account of the analysis technique used in IPS forecasting, and of the currently operating website, and details how this technique can be used in real time with the STELab system. Between the time this article was first conceived and its submission, the U.S. Air Force decided to stow the SMEI instrument, which is on board the Coriolis spacecraft, (on 28 September 2011), to save data access costs and civil-servant salaries during

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch06

The 3D Analysis of the Heliosphere Using Interplanetary Scintillation

87

a period of austerity in the U.S. Thus, for the foreseeable future there can only be access to SMEI archival data to support 3D analysis at UCSD, the CCMC, and at Nagoya, Japan.

5. Conclusion Analysis of data from IPS systems have provided heliospheric results for over five decades; presently this science has advanced to the point of yielding precise global measurements over much of the inner heliosphere. This in turn has spawned heliospheric imaging systems, of which SMEI was a prototype, for more advanced instruments on proposed NASA and ESA spacecraft. Current analyses that explore the 3D structure with most planned spaceborne instruments do not have the maturity of IPS, and generally go only so far as to look at time sequences of 2D images or at 3D stereographic approximations of known structure shapes. Such essentially 2D image analyses often provide a good-enough story for these current rudimentary comparisons; however, full 3D analysis can be obtained after careful instrument calibration and thorough background noise elimination using Thomson-scattering results. Eventually, the requirement for analyses that match most aspects of detailed in-situ measurements globally will compel further development of more precise systems. For more details and to look into the research currently presented using these global exploratory 3Dreconstructions, the reader is referred to the many journal articles cited here.

Acknowledgments I wish to thank the many colleagues who contributed to this work and for allowing it to be presented as a Distinguished Lecture for the SolarTerrestrial Section of the AOGS. At UCSD these colleagues are P.P. Hick, A. Buffington and J.M. Clover. Special thanks are due M. Kojima and M. Tokumaru for making available IPS data for these analyses under the auspices of a joint CASS/UCSD–STELab cooperative agreement. Additional thanks are due M. Shimojo and M. Sako who have furnished data from Hinode for use in the jet response analysis. SMEI was designed and constructed by a team of scientists and engineers from the University of California, San Diego, the U.S. Air Force Research Laboratory, Boston College, Boston University, and the University of Birmingham, U.K. The work of B.V. Jackson, and colleagues at the University of California,

July 13, 2012 9:4

88

ADGEO

9in x 6in

b1389-ch06

B.V. Jackson

San Diego was supported by NSF grants ATM-0852246 and AGS-1053766, and NASA grant NNX11AB50G and AFOSR grant 11NE043.

References 1. B. V. Jackson, A coronal hole equatorial extension and its relation to a high speed solar wind stream, Topical Conference on Solar and Interplanetary Physics, Tucson, Arizona, 12–15 January, 1977, pp. 7–9. 2. G. R. Gapper, A. Hewish, A. Purvis and P. J. Duffet-Smith, Nature 296 (1982) 633–636. 3. B. V. Jackson, IPS Observations of the 14 August, 1979 Mass Ejection Transient, Proc. of the Maynooth, Ireland Symposium on Solar/Interplanetary Intervals, 1982, pp. 169–173. 4. Z. Houminer, Nature Phys. Sci. 231 (1971) 165–167. 5. A. Hewish and S. Bravo, Solar Phys. 106 (1986) 185–200. 6. K. W. Behannon, L. F. Burlaga and A. Hewish, J. Geophys. Res. 96 (1991) 21,213–21,225. 7. I. Richter, C. Leinert and B. Planck, Astron. Astrophys. 110 (1982) 115–120. 8. C. Leinert, H. Link, E. Pitz, N. Salm and D. Knuppelberg, Raumfahart forschung 19 (1975) 264–267. 9. B. V. Jackson, P. P. Hick, A. Buffington, M. M. Bisi, J. M. Clover, M. Tokumaru, M. Kojima and K. Fujiki, J. Atmospheric and Solar-Terrestrial Phys. 73 (2011) 1–9. 10. G. E. Brueckner, R. A. Howard, M. J. Koomen, C. M. Korendyke, D. J. Michels, J. D. Moses, D. G. Socker, K. P. Dere, P. L. Lamy, A. Llebaria, M. V. Bout, R. Schwenn, G. M. Simnett, D. K. Bedford and C. J. Eyles, Solar Phys. 162 (1995) 357–402. 11. V. Domingo, B. Fleck and A. I. Poland, Space Sci. Rev. 72 (1995) 81–84. 12. C. J. Eyles, G. M. Simnett, M. P. Cooke, B. V. Jackson, A. Buffington, P. P. Hick, N. R. Waltham, J. M. King, P. A. Anderson and P. E. Holladay, Solar Phys. 217 (2003) 319–347. 13. B. V. Jackson, A. Buffington, P. P. Hick, R. C. Altrock, S. Figueroa, P. Holladay, J. C. Johnston, S. W. Kahler, J. Mozer, S. Price, R. R. Radick, R. Sagalyn, D. Sinclair, G. M. Simnett, C. J. Eyles, M. P. Cooke, S. J. Tappin, T. Kuchar, D. Mizuno, D. F. Webb, P. Anderson, S. L. Keil, R. Gold and N. R. Waltham, Solar Phys. 225 (2004) 177–207. 14. D. E. Billings, A Guide to the Solar Corona (Academic, New York, 1966) 150. 15. A. Buffington, D. L. Band, B. V. Jackson, P. P. Hick and A. C. Smith, Astrophys. J. 637 (2006) 880–888. 16. D. R. Mizuno, A. Buffington, M. P. Cooke, C. J. Eyles, P. P. Hick, P. E. Holladay, B. V. Jackson, J. C. Johnston, T. A. Kuchar, J. M. Mozer, S. D. Price, R. R. Radick, G. M. Simnett, D. Sinclair and D. F. Webb, J. Geophys. Res. 110 (2005) A07230. 17. B. V. Jackson, B. Rompolt and Z. Svestka, Solar Phys. 115 (1988) 327–343. 18. B. V. Jackson and H. R. Froehling, Astron. Astrophys. 299 (1995) 885–892.

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch06

The 3D Analysis of the Heliosphere Using Interplanetary Scintillation

89

19. B. V. Jackson, P. L. Hick, M. Kojima and A. Yokobe, J. Geophys. Res. 103 (1998) 12,049–12,067. 20. B. V. Jackson, A. Buffington and P. P. Hick, A heliospheric imager for solar orbiter, Proc. of Solar Encounter: The First Solar Orbiter Workshop, Puerto de la Cruz, Tenerife, Spain, 14–18 May, 2001, pp. 251–256. 21. B. V. Jackson, P. P Hick, A. Buffington, M. Kojima, M. Tokumaru, K. Fujiki, T. Ohmi and M. Yamashita, Time-dependent tomography of heliospheric features using interplanetary scintillation (IPS) remote sensing observations, CP679, Solar Wind Ten: Proceedings of the Tenth International Solar Wind Conference, eds. M. Velli, R. Bruno and F. Malara, 2002, pp. 75–78. 22. B. V. Jackson and P. P. Hick, Solar Phys. 211 (2002) 345–356. 23. B. V. Jackson, A. Buffington, P. P. Hick, M. M. Bisi and J. M. Clover, Solar Phys. 265 (2010) 257–275. 24. B. V. Jackson, P. P. Hick, A. Buffington, M. M. Bisi, J. M. Clover and M. Tokumaru, Adv. in Geosciences 21 (2008) 339–366. 25. P. K. Manoharan, Study of the solar wind using single-station interplanetary scintillation, PhD thesis, University of Bombay, 1991, pp. 53–95. 26. P. K. Manoharan, Solar Phys. 265 (2010) 137–157. 27. M. M. Bisi, B. V. Jackson, J. M. Clover, P. K. Manoharan, M. Tokumaru, P. P. Hick, and A. Buffington, Annales Geophysicae 27 (2009) 4479–4489. 28. M. M. Bisi, R. A. Fallows, P. K. Manoharan, G. D. Dorrian, B. V. Jackson, J. M. Clover, P. P. Hick, A. Buffington, and A. R. Breen, Adv. in Geosciences 21 (2008) 33–49. 29. L. K. Hara, N. U. Crooker, C. H. Mandrini, L. van Driel-Gesztelyi, S. Dasso, J. Wang, H. Elliott, G. Attrill, B. V. Jackson, and M. M. Bisi, Solar Phys. 244 (2008) 95–114. 30. M. Kojima and T. Kakinuma, J. Geophys. Res. 92 (1987) 7269–7279. 31. B. V. Jackson, J. A. Boyer, P. P. Hick, A. Buffington, M. M. Bisi, and D. H. Crider, Solar Phys. 241 (2007) 385–396. 32. M. M. Bisi, B. V. Jackson, A. Buffington, J. M. Clover, P. P. Hick and M. Tokumaru, Solar Phys. 256 (2009) 201–217. 33. P. P. Hick and B. V. Jackson, Heliospheric tomography: an algorithm for the reconstruction of the 3D solar wind from remote sensing observations, Proc. SPIE 5171, 2004, pp. 287–297. 34. M. Kojima, M. Tokumaru, H. Watanabe, A. Yokobe, K. Asai, B. V. Jackson and P. L. Hick, J. Geophys. Res. 103 (1998) 1981–1989. 35. B. V. Jackson and P. P. Hick, Three-dimensional tomography of interplanetary disturbances, Solar and Space Weather Radiophysics, Current Status and Future Developments (Astrophysics and Space Science Library, 315), eds. D. E. Gary and C. U. Keller, 315 (Kluwer Academic Publ., Dordrecht, The Netherlands, 2005) 355–386. 36. B. V. Jackson, A. Buffington, P. P. Hick, X. Wang and D. Webb, J. Geophys. Res. 111 (2006) A04S91. 37. B. V. Jackson, P. P. Hick, A. Buffington, M. M. Bisi, M. Kojima and M. Tokumaru, Astronomical and Astrophysical Transactions 26(6) (2007) 477–487.

July 13, 2012 9:4

90

ADGEO

9in x 6in

b1389-ch06

B.V. Jackson

38. M. M. Bisi, B. V. Jackson, P. P. Hick, A. Buffington and J. M. Clover, J. Geophys Res. — Space Physics Special Edition — Geomagnetic Storms of Solar Cycle 23 113 (2008) A00A11. 39. B. V. Jackson, M. M. Bisi, P. P. Hick, A. Buffington, J. M. Clover and W. Sun, J. Geophys. Res. 113 (2008) A00A15. 40. T. Mulligan, C. T. Russell, B. J. Anderson and M. H. Acuna, Geophys. Res. Letts. 28 (2001) 4417–4420. 41. M. Tokumaru, M. Kojima, K. Fujiki, M. Yamashita and A. Yokobe, J. Geophys. Res. 108(A5) (2003) 1220. 42. D. Odstrcil, J. A. Linker, R. Lionello, Z. Mikic, P. Riley, V. J. Pizzo, and J. G. Luhmann, J. Geophys. Res. 107(A12) (2002) SSH14-1-11. 43. D. Odstrcil and V. J. Pizzo, J. Geophys. Res. 104 (1999) 483–492. 44. D .Odstrcil, V. J. Pizzo, C. N. Arge, M. M. Bisi, P. P. Hick, B. V. Jackson, S. A. Ledvina, J. G. Luhmann, J. A. Linker, Z. Mikic, and P. Riley (2008), Numerical Simulations of Solar Wind Disturbances by Coupled Models, ASP Conference Series Proceedings, 385, 2008, pp. 167–173. 45. J. G. Luhmann, S. A. Ledvina, D. Odstrcil, M. J. Owens, X.-P. Zhao, Y. Liu and P. Riley, Adv. in Space Res. 46(1) (2010) 1–21. 46. D. Odstrcil, C. N. Arge, A. Rasca, A. Thernisien and H. Xie, Simulation of heliospheric disturbances initialized by various fitting techniques, NOAA Space Weather Workshop Boulder, CO, 27–30 April, 2010. 47. B. V. Jackson, A. Buffington, P. P. Hick, J. M. Clover, M. M. Bisi and D. F. Webb, Astrophys. J. 724 (2010) 829–834. 48. M. Bisi, B. Jackson, J. M. Clover, M. Tokumaru, A. Buffington, P. Hick and K. Fujiki, Coronal Mass Ejections in the Declining and Minimum Phase between Solar Cycles 23 and 24, Eos Trans. AGU, 14–18 December, 2009, Fall Meeting Suppl., Abstract SH43A-08, p. 374. 49. D. F. Webb, A. B. Galvin, N. Gopalswamy, T. A. Howard, A. A. Reinard, B. Jackson, and C. Davis, The 26 April 2008 CME; a Case Study Tracking a CME into the Heliosphere, Eos Trans. AGU, 14–18 December, 2009, Fall Meeting Suppl., Abstract SH41A-1640, p. 319. 50. B. V. Jackson, P. P. Hick, A. Buffington, J. M. Clover, D. B. Reisenfeld and T. R. Abell, 3-D Reconstruction of the Inner Heliosphere: A Global Solar Wind Boundary from Remote-Sensing Data, Invited presentation to the ninth annual CSPAR Conference, Maui, Hawaii, 14–19 March, 2010. 51. B. E. Wood and R. A. Howard, Astrophys. J. 702 (2009) 901–910. 52. N. Lugaz, J. N. Hernandez-Charpak, I. I. Roussev, C. J. Davis, A. Vourlidas and J. A. Davies, Astrophys. J. 715 (2010) 493–499. 53. W. Sun, C. S. Deehr, M. Dryer, C. D. Fry, Z. K. Smith and S.-I. Akasofu, Space Weather 6 (2008) S03006. 54. G. Toth, D. L. De Zeeuw, T. I. Gombosi and K. G. Powell, J. Comp. Phys., 217 (2006) 722–758. 55. C. C. Wu, C. D. Fry, S. T. Wu, M. Dryer and K. Liou, J. Geophys. Res. 112 (2007) A09104. 56. X. Feng, L. Yang, C. Xiang, S. T. Wu, Y. Zhou and D. K. Zhong, Astrophys. J. 723(1) (2010) 300–319.

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch06

The 3D Analysis of the Heliosphere Using Interplanetary Scintillation

91

57. T. R. Detman, D. S. Intriligator, M. Dryer, W. Sun, C. S. Deehr and J. Intriligator, J. Geophys. Res. 116 (2011) A03105. 58. L. F. Burlaga, J. Geophys. Res. 93 (1988) 7217–7224. 59. T. Kosugi, K. Matsuzaki, T. Sakao, T. Shimizu, T. Sone, S. Tachikawa, T. Hashimoto, K. Minesugi, A. Ohnishi and T. Yamada, et al., Solar Phys. 243 (2007) 3–17. 60. L. Golub, E. DeLuca, G. Austin, J. Bookbinder, D. Caldwell, P. Cheimets, J. Cirtain, M. Cosmo, P. Reid and A. Sette, et al., Solar Phys. 243 (2007) 63–86. 61. R. Kano, T. Sakao, H. Hara, S. Tsuneta, S. Matsuzaki, K. Kumagai, M. Shimojo, K. Minesugi, K. Shibasaki and E. E. DeLuca, et al., Solar Phys. 249 (2008) 263–279. 62. S. Tsuneta, K. Ichimoto, Y. Katsukawa, S. Nagata, M. Otsubo, T. Shimizu, Y. Suematsu, M. Nakagiri, M. Noguchi, and T. Tarbell, et al., Solar Phys. 249 (2008) 167–196. 63. Y, Suematsu, S. Tsuneta, K. Ichimoto, T. Shimizu, M. Otsubo, Y. Katsukawa, M. Nakagiri, M. Noguchi, T. Tamura and Y. Kato, et al., Solar Phys. 249 (2008) 197–220. 64. K. Ichimoto, B. Lites, D. Elmore, Y. Suematsu, S. Tsuneta, Y. Katsukawa, T. Shimizu, R. Shine, T. Tarbell and A. Title, et al., Solar Phys. 249 (2008) 233–261. 65. S. Tsuneta, K. Ichimoto, Y. Katsukawa, B. W. Lites, K. Matsuzaki, S. Nagata, et al., Astrophys. J. 688 (2008) 1374–1381. 66. M. Shimojo and S. Tsuneta, Astrophys. J. Letts. 706 (2009) L145–L149. 67. P. MacNiece, Solar and Heliospheric Modeling at the CCMC, SHINE 06 Workshop, Midway, UT, 31 July–4 Aug, 2006. 68. D. J. McComas, B. L. Barraclough, H. O. Funsten, J. T. Gosling, E. SantiagoMu˜ noz, R. M. Skoug, B. E. Goldstein, M. Neugebauer, P. Riley and A. Balogh, J. Geophys. Res. 105(A5) (2000) 10,419–10,433. 69. M. Kojima, A. R. Breen, K. Fujiki, K. Hayashi, T. Ohmi and M. Tokumaru, J. Geophys. Res. 109 (2004) A04103. 70. M. Sako, M. Shimojo and T. Kitabayashi, Statistical study of the polar X-ray jets, 38th COSPAR Scientific Assembly, 2010, E21-0018-10. 71. B. V. Jackson, P. P. Hick, A. Buffington, J. M. Clover and M. Tokumaru, Adv. in Geosciences, (in press 2012).

July 13, 2012 9:4

ADGEO

9in x 6in

This page intentionally left blank

b1389-ch06

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch07

Advances in Geosciences Vol. 30: Planetary Science and Solar & Terrestrial Science (2011) Ed. Anil Bhardwaj c World Scientific Publishing Company


FORECASTING TRANSIENT HELIOSPHERIC SOLAR WIND PARAMETERS AT THE LOCATIONS OF THE INNER PLANETS B. V. JACKSON, P. P. HICK, A. BUFFINGTON and J. M. CLOVER Center for Astrophysics and Space Sciences, University of California, San Diego 9500 Gilman Dr. #0424, La Jolla, CA 92093-0424, U.S.A. E-Mail: [email protected] http://smei.ucsd.edu/ http://ips.ucsd.edu/ M. TOKUMARU Solar-Terrestrial Environment Laboratory, Nagoya University Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan http://stesun5.stsw1.stelab.nagoya-u.ac.jp/index-e.html

Remotely-sensed interplanetary scintillation (IPS) from the solar-terrestrial environment laboratory (STELab) system, and Thomson-scattering observations from the U.S. Air Force/NASA Solar Mass Ejection Imager (SMEI) allow the determination of solar wind parameters at the locations of the inner planets. We show a 3D analysis technique developed to provide daily-cadence transient solar wind forecasts of velocity and density at Earth and the inner planets. These now include in-situ measurements near Earth available in real time. Where in-situ measurements are available these real-time analyses are compared with the predicted values. Using the global velocity measurements available from IPS analysis and daily updated magnetograms from the National Solar Observatory, we are also able to project outward solar-surface magnetic fields in order to provide reasonable global in-situ magnetic-field component trends from one day to the next. This paper summarizes the analysis available and current progress in using the STELab, Japan real-time data for validating these forecasts. A discussion is also provided as to how we can derive more meaningful future information from these remotely-sensed heliospheric measurements.

1. Introduction Two sets of heliospheric remote-sensing data are calibrated well-enough to provide accurate in-situ densities and velocities. One data set is made up of interplanetary scintillation (IPS) observations1 that have long been used to 93

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch07

B. V. Jackson et al.

94

remotely-sense small-scale (∼150 km) heliospheric density variations along the line of sight (LOS). IPS-array data,2−4 when analyzed, show structures that can be classified as either co-rotating or detached from the Sun, and these and other IPS observations have been studied over many decades.2−15 Large IPS radio arrays continuously operating around the world are currently at the Solar-Terrestrial Environment Laboratory (STELab), Japan4,16 (Fig. 1), at Ootacamund (Ooty), India17 ; and at Puschino, Russia.18 The STELab radio antennas measure both scintillation-level and velocity by combining observations from different radio sites. The velocity analysis from the STELab arrays are a primary source of information about this global heliospheric parameter, and thus also about the energy content of the solar wind in the inner heliosphere. The IPS scintillation-level data serve as a proxy for density to display large-scale heliospheric structures. A second source of calibrated heliospheric data is the Solar Mass Ejection Imager (SMEI).19,20 SMEI was launched (6 January 2003) on the Air Force Space Test Program satellite Coriolis. The instrument consists of three baffled cameras whose 3◦ × 60◦ fields of view are aligned in the long dimension to achieve a combined ∼160◦ wide field of view that scans most of the sky in the course of every 102-minute orbit. The cameras view the heliosphere in Thomson-scattered light with ∼0.5◦ angular resolution. Approximately 4,500 four-second exposure CCD-camera data frames per orbit are combined into a map of the sky. SMEI brightness has been calibrated to an absolute level of ∼4% using known stars viewed by the

(a)

(b)

Fig. 1. (a) STELab IPS radio array (one of three now operating) near Mt. Fuji. The arrays used singly measure scintillation intensity (or g-level). Scintillation signals cross-correlated between arrays give a robust sky plane IPS velocity determination. (b) A significantly larger system now measures g-levels at the Toyokawa radio site and integrates results with other arrays.16

July 13, 2012 9:4

ADGEO

9in x 6in

Forecasting Transient Heliospheric Solar Wind Parameters

b1389-ch07

95

instrument’s cameras.21 This brightness relates directly to the LOS density of heliospheric structures. The authors at the University of California, San Diego (UCSD) have developed a time-dependent tomographic analysis technique that allows solar wind temporal variations to be mapped in three spatial dimensions over time (3D). This analysis fits data in essentially the same manner whether IPS scintillation level, IPS velocity, SMEI brightness, or a combination of these data sets is used in the 3D reconstructions. The temporal and spatial cadences of solar wind outward-moving structures is predicated by the numbers and quality of the data, and for IPS observations from STELab these are set at about one day and 20◦ by 20◦ in latitude and longitude. Unlike other determinations such as Wang–Sheely– Arge modeling (see http://www.swpc.noaa.gov/ws/), this technique provides the 3D reconstruction of transient heliospheric structures such as coronal mass ejections (CMEs) as well as those that co-rotate, and updates these as they move outward from the Sun. The validity of the different techniques is certified by 3D-reconstruction results contained in the literature.9,11,12,22−28 Most of these publications compare remotely-sensed measurements with in-situ observations of density or velocity. The 3D reconstruction programs using STELab IPS data or SMEI data now reside at the Goddard Space Flight Center Community Coordinated Modeling Center (CCMC), for use by others besides our group in their analyses of heliospheric structures. IPS analyses alone have determined the structure location, mass, and solar wind speeds from ICMEs1,26 at Earth, and at the two Solar Terrestrial Relations Observatory (STEREO) spacecraft.12,15 Combining mass and speed, Jackson et al.27 demonstrated that IPS plasma ram pressure measurements from the Mars Global Surveyor (MGS) magnetometer observations compare satisfactorily with these results. The two solar wind parameters that are remotely-sensed in these analyses are proxies for heliospheric velocity and density. Observed as they progress outward from the Sun, and updated as frequently as possible, these are fundamental parameters we can measure and forecast. For some space weather effects, especially at Earth, both the onset of a shock and the magnetic field direction are important in determining how an interaction with planetary magnetic fields will proceed for transient heliospheric structures. The location of shocked plasma and the magnitude of the shock is often determined and forecast in our analyses by a calibrated density enhancement located behind the shock and associated with a higher velocity (e.g., Jackson et al.15 ). Results show that these enhancements are

July 13, 2012 9:4

96

ADGEO

9in x 6in

b1389-ch07

B. V. Jackson et al.

often “spotty” or non-continuous globally; we do not directly measure the shock strength. In addition, we do not directly measure all three magnetic field components remotely, but must use solar-surface magnetic fields to provide two field components (radial and tangential) propagated upward in a background solar wind modeling effort (non-rapid changes). We note that other 3D analyses model shock propagation effects and the associated production of high-energy particles from solar surface effects.29−32 We have used IPS velocity and scintillation-level data from 1999 onward to determine heliospheric structures in real time at periods when STELab data were available (generally from May through December each year). Since the middle of the last decade, these IPS observations forecast density, velocity, and magnetic field values at the locations of all the inner planets including Earth. From 2007 onward, we have made similar forecasts at the STEREO spacecraft. Forecasts using the IPS data set are somewhat different from those used to provide research results from archival data analysis because forecasts cannot use data retrospectively and effectively use only half the amount of data available for studies based on archival data. Here we report on the current state of UCSD IPS forecasting: This includes a new technique developed over several years13 that is presently used to forecast heliospheric structures at Earth in real time. Section 2 briefly describes the time-dependent tomographic analysis routines developed by our group at UCSD for fitting STELab IPS velocity and g-level (see Sec. 2.1) data, and SMEI brightness and its extension that includes realtime in-situ velocity and density in the forecasts. This section also describes the development that provides extrapolated magnetic-field variations into the heliosphere. Section 3 gives a brief description of the data sequencing and forecast-procedure programming at UCSD. Section 4 presents examples of these analyses for the real-time IPS data from STELab. We discuss these results and conclude in Sec. 5. 2. 3D-Reconstruction Analysis 2.1. The original 3D reconstruction technique The mathematics of this technique has been described in detail in Hick and Jackson,33 and Jackson et al.,10 and the reader can refer to these articles for more information than is given here. The computational aspects of the UCSD 3D-reconstruction program have been discussed in many other articles over the past decade.7,8,33 The early analyses assume that

July 13, 2012 9:4

ADGEO

9in x 6in

Forecasting Transient Heliospheric Solar Wind Parameters

b1389-ch07

97

the heliosphere co-rotates with the Sun. In more recent work1,11,22,25,28 this assumption has been relaxed. In this case, LOS segments and their 3D weights are projected back in space and time to a solar wind inner boundary (a source surface) that is set at a given height (usually 15 RS ) that lies below the closest approach of all lines of sight to the Sun. Each LOS is mapped from Earth and each segment of these is projected to the source surface assuming the velocity and interactions from the model providing the solar wind outward motion. In current analyses, the inversion process adjusts boundary conditions for the kinematic 3D solar wind model to best fit the observations using a least-squares fitting procedure. This minimizes the differences between modeled and observed SMEI brightness, or modeled and observed IPS g-level (source scintillation strength relative to the mean strength at given elongation) and velocity, or a combination of these. As explained elsewhere,10,28 the solar wind Thomson-scattering signal from SMEI is difficult to distinguish from the very bright zodiacal light signal. Because of this, reconstructions based on SMEI data (unlike IPS g-level data) require that a mean ambient solar wind is included in the model based on the average in-situ solar wind density at 1 AU. A least squares fitting program developed specifically for this type of analysis inverts the weighted, projected model values on the twodimensional (2D) inner boundary source surface, using different time steps, in order to provide solar wind model outflow parameters. These are directly inverted on the source surface at the appropriate times to yield new solar wind parameters, and these latter are iteratively converged for each data set. In the fitting process, ratios of modeled-to-observed values and a modeled-to-observed χ2 are monitored to indicate the rate of convergence for the interval. Velocity and density corrections to the 3D model are made separately. First, the inversion changes are made to previous velocity boundary conditions on the inner boundary surface. Second, the 3D solar wind model is updated and new projected locations of each LOS point on the inner-boundary surface are determined. Third, inversion changes are made to previous density boundary conditions at the inner boundary surface. Finally, the 3D model is again updated with all the newest boundary values. The inner boundary Carrington maps of velocity and density are smoothed for each iteration using a 2D Gaussian spatial filter that incorporates equal-solar-surface areas, and also a Gaussian temporal filter. Locations in the model that are unaffected by the above iterative procedure

July 13, 2012 9:4

98

ADGEO

9in x 6in

b1389-ch07

B. V. Jackson et al.

(and thus remain undetermined) are left blank in the final result. For the analysis presented here, these blank areas include sections of heliospheric volume on the opposite side of the Sun from Earth that are not observed and thus not reconstructed at the resolutions of the 3D volume in the digital resolutions adopted. For SMEI this area can cover a large fraction of the region behind the Sun because the instrument cannot view close to the solar surface. This area is usually much smaller in the case of the 327 MHz IPS when it accesses the sky to within 11.5◦ of the Sun. The reconstruction program generally converges to an unchanging model within a few iterations, but is operated for nine iterations to guarantee convergence.7 For a typical rotation and the digital resolutions of the current SMEI data sets, the density and velocity iterations generally take about 15 minutes to complete using a 2.4 GHz Intel
Core i5 computer. The IPS data sets normally take only a few minutes to process. Normally those IPS-velocity observations and SMEI-brightness lines of sight throughout the period that do not fit within a three-sigma limit of the mean ratio change ascribed at a particular location by the model (typically ∼1% of the SMEI brightness or IPS velocity lines-of-sight) are removed from the data set. This allows for the removal of lines of sight that are outliers which do not fit the model values. It is sometimes found that the overall χ2 fit, for g-level, brightness, or velocity is reduced when this criterion is imposed. The program is then operated for nine more iterations (18 in total). The solutions are insensitive to the initial model values and after a few iterations any residual signature of the initial values has disappeared. Other tests7 show that the 3D reconstruction of a set of artificial observations using a known 3D input successfully reproduces the input.

2.2. Integration of real-time in-situ observations in the reconstructions An innovation to the original 3D reconstruction technique for use in forecasting IPS velocity was introduced as a potential break-through13 in 2010. In this analysis in-situ data were used to constrain the reconstruction at the location closest to the Earth. Here, not only are the line of sight segments closest to Earth used in the inversion, but also the heavilyweighted in-situ measurements. The values at the source surface represent both the remote-sensing and in-situ measurements, and are projected outward to provide the 3D model over time at the resolution of the 3D reconstruction. In time-dependent velocity tomography, this produces the

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch07

Forecasting Transient Heliospheric Solar Wind Parameters

99

largest effect through modifying the radial along the Sun–Earth line, while leaving other regions of the 3D analysis relatively unaffected over time.13 The technique was refined during the early months of 2011 to include in-situ density measured at Earth in the volumetric analysis that is then used to modify the near-Earth segment IPS g-level values, or the SMEI brightness. In forecasting this has the effect of allowing a smooth transition between in-situ and remotely-sensed data, while making the change near Earth from the present to the future less abrupt. If the inclusion of in-situ measurements into the tomographic result were simply a way to smooth the transition from in-situ measurements to remotely-sensed observations, we would not expect much improvement in remotely-sensed forecast results beyond the immediate vicinity of the last in-situ measurements. However, in the case for the examples given in Jackson et al.13 there is also an improvement in the remote-sensing analysis. Thus, we speculate that the inclusion of the in-situ measurements near the observing point significantly improves each LOS measurement. It probably does this by refining information close to the observer, where small amounts of noise might seriously alter the result along the whole LOS. We now operate this tomographic technique regularly in our IPS velocity and density IPS forecasts, and are currently evaluating its overall effect on forecasting in general at UCSD and at the CCMC. We note that currently, this procedure only works at Earth to provide a smooth transition from in-situ measurements to those remotely-sensed and, while there should be an improvement at all the inner planets, this cannot be tested and certified unless there are plasma monitors at these locations that measure the values in situ. 2.3. Inclusion of the CSSS potential field model for outward extrapolation of the magnetic field In 2005, the Current Sheet Source Surface (CSSS) potential magnetic field model34 was introduced into our tomography analysis35 as an extension of the kinematic model (Fig. 2). In this regard, the CSSS model provides an accurate radial magnetic field at the inner source surface of the tomographic analysis that is then extrapolated upward (forward-modeled) to locations within the 3D matrix by using the volumetric velocity provided by (timedependent) UCSD tomography. The CSSS model is updated as frequently as there are appropriate data available, and this (as stated in the Introduction) generally provides daily updates to magnetic field determination made

July 13, 2012 9:4

100

ADGEO

9in x 6in

b1389-ch07

B. V. Jackson et al.

Fig. 2. The CSSS model34 as used in the UCSD tomography.35 In the inner region (1), the CSSS model calculates the magnetic field using photospheric measurements and a horizontal current model. In the middle region (2), the CSSS model opens the field lines by imposing a current at the source surface. In the outer region (3), the UCSD solar wind model convects the magnetic field outward radially as the source surface rotates below any given heliospheric location with the attached magnetic field.

near the sub-Earth solar meridian. Since the potential field model provides only radial fields and their change at the source surface, the extrapolation outward into the heliosphere provides only radial and tangential magnetic fields caused by the rotation of the source surface below any specific location within the 3D volume. Since the velocity changes over time and a kinematic model is imposed by the tomography, the magnetic fields that are “frozen in” to the outward-expanding solar wind retain this variation from the solar wind modeling. Dunn et al.35 show that even when a concerted effort is made to provide the most rapidly-updated solar-surface magnetic fields, it still does not reproduce rapidly-changing (daily) magnetic fields in situ. Absent from these analyses are the rapid changes that occur due to coronal currents, especially those during CMEs that may feature both radial and non-radial fields at the source surface.

3. Operation of the Real-Time UCSD Website It is one matter to have a conceptual process to show how forecasts are possible using a given data set and quite another to provide forecasts automatically. A great amount of work is required to provide consistent data in real time. These data are needed to provide input to scientific articles and advance the research that funds the analysis. However, the data are available at the end of each day, and it is often important to know which

July 13, 2012 9:4

ADGEO

9in x 6in

Forecasting Transient Heliospheric Solar Wind Parameters

b1389-ch07

101

data editing scheme works for each day’s data set in order to identify the best way to process it. The UCSD real-time analysis using the data from STELab is found at website http://ips.ucsd.edu. The procedure to provide real-time IPS visualizations proceeds automatically after a script that queries, from San Diego, the STELab, Japan computer, and indicates that new IPS data are available following the day’s data-taking and analysis (usually at about 4 a.m. local time in Japan). These data are downloaded, and a sequence of programming steps start. For an analysis that includes the real-time in-situ data, and magnetic field data from the National Solar Observatory (NSO), this requires that additional programs be run that query the National Oceanic and Atmospheric Administration (NOAA) computer every three hours for real-time Advanced Composition Explorer (ACE) spacecraft data and the NSO computer on a daily basis for the latest magnetic field maps. In the forecast analysis the tomography programs are run using the latest data. IPS data are taken over an approximately eight-hour period during each day and, at the time the tomography program is run, the data latency (time elapsed since the last data were taken) is about 10 hours for the last IPS sources recorded in Japan on each day. It takes approximately six minutes to complete the IPS tomography program forecast and to output the associated volumetric files. Once the tomography analysis is completed, a further procedure provides visualizations for the website. First, images are made from the current tomography analysis that are presented immediately on the website (within 18 minutes of the latest tomography run). Animations of the original tomographic analysis follow for use throughout the (daily) period before the next tomography run is available. The image data sets require updates that change time markers and updated in-situ measurements. These are refreshed hourly from the earlier-run tomography until there are no longer forecast volumetric files resident on the computer.

4. Examples of the 3D Real-Time Analysis Examples of the 3D real-time analysis are presented below in three subsections. The first shows comparisons of velocity and density at Earth to be verified by in-situ measurements. These are shown specifically for and around Earth where forecasts and results are regularly verified using spacecraft data and other measurements. The second subsection shows examples of real-time analysis available at other planets of the inner solar

July 13, 2012 9:4

102

ADGEO

9in x 6in

b1389-ch07

B. V. Jackson et al.

system. The third subsection gives current magnetic-field measurements and forecasts at Earth and at some of the inner planets.

4.1. Earth in-situ velocity and density real-time analysis Figure 3 shows our current IPS website — http://ips.ucsd.edu. Of primary importance and specific forecast interest is a time-series that includes the current velocity and density measured at a given in-situ location up to the present, and a forecast that shows how great a change is expected to occur at this location. Furthermore, this site should include some evaluation of the agreement with real-time data and recent forecast success. To provide this visualization, the tomographic program timeseries in velocity, density, and magnetic field is shown up to the present and compared with ACE real-time data. Figure 4 shows a sample of a

Fig. 3. The UCSD website front page. The CASS and STEL logos are links that allow access to other than the forecasting analysis. Date and time on the upper left of the webpage automatically update. Primary forecast products are accessed on the left for both the time-dependent tomography described here and an older version of the corotating analysis (see Sec. 2.1). Additional forecasting data products are obtained by links to additional pages on the upper portion of the left panel. These include the planetary and magnetic field forecasts by following the “Solar System Space Weather” link.

July 13, 2012 9:4

ADGEO

9in x 6in

Forecasting Transient Heliospheric Solar Wind Parameters

b1389-ch07

103

Fig. 4. Sample time-dependent tomography time series analysis obtained by pressing “Time Series” in the left panel under Time Dependent. The left time-series panel gives the IPS velocity result at Earth in blue. To the right, the IPS density is given in red. Dates are given in days during October 2011. ACE data are compared up to the present as the dashed line in each time series plot. ACE data are boxcar averaged over a one-day period. The presentation that is updated hourly shows the present time as a vertical line, and the future forecast to the right of this line.

time-dependent velocity and density time series as they are presented on the website. The tomographic analysis program is also operating at the CCMC, and is being evaluated there for use in real-time forecasting. Figure 5 compares a one-day forecast and a five-day “aft cast” of a density result. Specifically, the analysis compares a one-day-in-advance forecast value of the present time with ACE data (once its average values have become available). The correlation is also presented for data five days after the present to show how well the ACE data and the tomography agreed with each other in the past. A cross-correlation near 1.0 implies a good result. However, this is only one type of correlation presentation that can be made. To be meaningful, this type of correlation requires that the excursions in velocity and density are large, and do not simply retain the original value. This type of correlation does not support an assessment of how well the forecast can be made at times different from one day in advance, or how well actual change is forecast. In addition, no data criterion is presented that could assess the data quality at each point. More is known about the data sets (both those remotely-sensed and in situ) that input to this simple comparison, and some way to make a relatively more detailed assessment is desirable.

July 13, 2012 9:4

104

ADGEO

9in x 6in

b1389-ch07

B. V. Jackson et al.

Fig. 5. Sample time-dependent tomography time series correlation analysis obtained by pressing “Correlation” in the left panel under Time Dependent. The panels give the ACE averaged data (ordinate) correlated with the IPS tomography result one day in advance of the present (left panel) and five days after the present (right panel) in this example. The one large density in the right panel generates the extremely good correlation shown.

Other forecast data products on the UCSD IPS time-dependent website provide information as to how the remote-sensing process works. These are designed to show how well the IPS tomographic result appears globally relative to locations where in-situ measurements can be made as well as provide information about the global inner heliospheric response. The “Remoteview” link gives a graphical depiction of the overall 3D heliospheric structure (Fig. 6). In any given data sequence the threshold levels sometimes need to be adjusted to best show the CMEs and co-rotating features that are of most interest at a particular time. Also foreground structure can sometimes hide features behind them, so views from a given observing position can often obscure interesting features. A synoptic map in heliographic latitude and longitude at the location of Earth is often better for depicting the structure surrounding Earth, as shown in Fig. 7. Current maps (October 2011) generally show far more high velocity to the South and far more dense structures to the North as would be consistent with a southern coronal hole that is rapidly disappearing at the present stage of Solar Cycle 24. These maps can be used to view the temporal evolution of the structure surrounding Earth by playing the animations that present the analysis from the past to the future slightly a day from the time when the tomography program is run. Other forecast data products include a Sky Map and a Sky Sweep. The former is a view of the sky at a given instant in time (updated hourly) as seen from Earth. The latter shows the sky at the latest mid-day time in Japan

July 13, 2012 9:4

ADGEO

9in x 6in

Forecasting Transient Heliospheric Solar Wind Parameters

b1389-ch07

105

Fig. 6. Sample time-dependent tomography remote view obtained by pressing “Remoteview” in the left panel under Time Dependent. The panels show the heliosphere as an observer would see it in perspective from 3.0 AU, 30◦ above the ecliptic plane and ∼45◦ west of the Sun-Earth line. Velocity is depicted on the left over the range shown. The Sun, Earth, and Earth orbit are shown. Velocity higher than 400 kms−1 is not depicted and has no opacity in this remote-observer view, and thus all regions not observed in velocity and translucent have values >400 kms−1 . The right panel depicts density from the same location and time as velocity. The density has an r−2 density fall-off removed from it to display structure with approximately the same value whether near or far from the Sun.

Fig. 7. Sample time-dependent tomography synoptic map obtained by pressing “Synoptic Map” in the left panel under Time Dependent. The panels give a Carrington map of the heliosphere at 1 AU with the Earth centered in the map. Velocity is depicted on the left, density to the right.

when remotely-sensed IPS data sources were available (Fig. 8). This figure shows sample sky sweep maps for IPS velocity and g-level as presented on the website in either Hammer–Aitoff (full-sky) or fisheye projections. These sky maps give a view of the IPS solar wind model in Earth-centered ecliptic

July 13, 2012 9:4

106

ADGEO

9in x 6in

b1389-ch07

B. V. Jackson et al.

Fig. 8. Sky sweep maps to the left shown as Hammer–Aitoff and to the right as “fisheye” sky projections viewed from Earth (see text). The modeled solar wind is contoured in Earth-centered ecliptic coordinates with the Sun at the center of the map. Velocity perpendicular to the line of sight is contoured in the top maps. IPS g-level is contoured in the two lower maps. Superimposed on these are the locations of the radio sources that have provided the model values on this day. The source values are coded to indicate which sources have greater values (dark circumference, and a larger dot with a darker center than the nearby ambient) or lesser values (light circumference, and a larger dot with a lighter center than the ambient) than the model values at this location. A source value with the same value as the model level is shown as the smallest of these dots.

coordinates that has been fitted to the observed source measurements. The heliographic equator is shown as a curved line on the map. IPS source locations are also marked. A Hammer–Aitoff projection is an equal-area view of the whole sky and is similar to frequently-seen maps of the whole Earth surface. The fisheye view preserves the angular elongation of material as it moves outward from the Sun as if this material was observed in a photograph that, in this case, extends beyond half of the sky outward from the Sun. The Hammer–Aitoff map includes features beyond 90◦ ; the 90◦

July 13, 2012 9:4

ADGEO

9in x 6in

Forecasting Transient Heliospheric Solar Wind Parameters

b1389-ch07

107

boundary is shown as the curved lines mid-way between the East and West and the Sun is at the center of the map. Features in a fisheye map are most easily compared with images closer to the Sun such as are available from coronagraphs. The white circle surrounding the Sun in the map centers are regions within elongations of 11.5◦ of the Sun that cannot be accessed at 327 MHz (the strong-scattering regime36 ). For these analyses see the discussion of this limit in Jackson et al.10 Only those sources input to the tomography program within reasonable range limits are fitted in the analysis. Approximately 1% of the radio sources input are removed from the analysis midway through the iterative procedure when they deviate by more than three-sigma from the least squares fitting criteria established,7 and the radio sources shown in Fig. 8 are those that remain at the end of the analysis. 4.2. Inner planet in-situ velocity and density real-time analysis It is possible to present the analysis of each volumetric parameter at nearly any location within the 3D volume as well as provide a forecast from a few hours to several days in advance at this location. The time series at all the inner planets and at the STEREO A and B spacecraft are presented in this manner as accessed using the “Solar System Space Weather” link from the IPS front page or at: http://ips.ucsd.edu/index ss.html (Fig. 9). Only the time series plots of the density, velocity, and radial and tangential magnetic field are accessed from the left panel of the UCSD solar system space weather front page. In Fig. 10 are shown the time-series plots at Mercury and Mars from these presentations. 4.3. In-situ magnetic field real-time analysis As mentioned, the radial and tangential magnetic field analyses are presented at all the inner planets (including Earth), and at the STEREO A and B spacecraft as described in Subsec. 2.3 from the UCSD solar system space weather forecast front page left panel. Figure 11 shows a sample plot of these magnetic time series for the same two planets of Fig. 10, Mercury and Mars, in nano-Teslas. Confirming these values at the inner planets is difficult most of the time, but it is possible to confirm those present at Earth. Figure 12 shows the plots at Earth for radial and tangential fields where these values are presented along with values of the ACE radial and tangential magnetic fields measured up to the (local) present. The magnetic field comparisons clearly show similar excursions, and give the general

July 13, 2012 9:4

108

ADGEO

9in x 6in

b1389-ch07

B. V. Jackson et al.

Fig. 9. The Solar System Space Weather forecast front page accessed from a link from the IPS front page or at: http://ips.ucsd.edu/index ss.html. In this figure only the time-dependent tomography links are shown on the (scrolled) left hand panel of the above display. Time series links to all the inner planets from Mercury to Mars are given as well as to those same parameters determined at the STEREO A and B spacecraft. The same time series velocity and density plots at Earth are presented here. Additional forecasting data products are obtained by links to additional pages on the upper portion of the left panel. These include a link backward to the velocity and density front page, and a link to a front page containing more Earth-based magnetic field forecasts that are accessed by following the “Magnetic Field (NSO/NOAA)” link. The front page plots in the right panel are hourly-updated remote-observer views from above the north ecliptic pole showing the locations of all the inner planets and their orbits, and the locations of the two STEREO spacecraft.

amplitude of these magnetic field components, as well as the approximate current sheet crossing from positive to negative, but they almost completely miss the short-term transient features present in the ACE data. At this time our analysis provides another kind of presentation shown in Fig. 13 that accesses data products available from these magnetic field analyses linked from the “Magnetic Field (NSO/NOAA)” solar system space weather forecast page or: http://ips.ucsd.edu/index br nson bt nson.html. Although this website is unfinished at the time of writing, one of the data products available are Carrington synoptic maps of

July 13, 2012 9:4

ADGEO

9in x 6in

Forecasting Transient Heliospheric Solar Wind Parameters

b1389-ch07

109

Fig. 10. Sample planetary time series of velocity and density are shown; to the left, velocity, to the right, density. The upper two plots show the values from the past and forecast at Mercury, the lower two plots are values at Mars.

Fig. 11. Sample planetary time series of radial and tangential magnetic field are shown respectively to the left and right. The upper two plots show the values from the past and forecast at Mercury, the lower two plots are values at Mars.

July 13, 2012 9:4

110

ADGEO

9in x 6in

b1389-ch07

B. V. Jackson et al.

Fig. 12. Sample Earth time series of radial and tangential magnetic field are shown respectively to the left and right. ACE data are plotted in comparison up to the present.

Fig. 13. Sample magnetic field analysis front forecast page and Carrington synoptic maps at 1 AU of radial and tangential magnetic field shown respectively to the left and right in the two contour plots. These are evoked by pressing the “Synoptic Map” option under Time Dependent in the left panel. The Earth is shown centered in the contour plots. These plots are updated hourly with associated animations presented at six-hour intervals from six days before to one-day following the time of the tomography run.

the magnetic field (Fig. 13) similar to those of density and velocity presented relative to Earth as in Fig. 7. These maps give the magnetic field structure around Earth determined from the radial and tangential components of magnetic field at 1 AU.

5. Discussion and Conclusions In the above sections we described the UCSD forecast website at: http://ips.ucsd.edu/ that presents the UCSD 3D-reconstruction forecast time-dependent algorithm, and provided sample visualizations of some of the presentations found on this website. Such forecasts use only about half

July 13, 2012 9:4

ADGEO

9in x 6in

Forecasting Transient Heliospheric Solar Wind Parameters

b1389-ch07

111

of the available data that can be used in research projects where IPS data are used retrospectively. Many of these research projects in fact use these data averaged over days and weeks (co-rotating tomography), and are not able to show the short-term transient features that can often be observed in the best data. Thus, it is with considerable trepidation that these analyses are used by our group to forecast short-term (daily) transient events like CMEs as well as the more slowly-varying heliospheric structures that corotate. In forecast analyses every data source is significant, and one of the primary reasons to provide this analysis is to show to ourselves how well we perform, and to improve IPS data editing so that each value becomes more meaningful. Recent analyses have been validated in journal articles for short given time intervals,13,35 but because updates are continually being made to the analysis algorithms, to the visualizations, and to the data as new ways to present them are learned and better-editing techniques developed, we presently have no final best way to present this analysis. For the research purposes that fund the majority of this “extra work”, the analysis is extremely important because better than in any other way it validates the data and the techniques used to provide measurements that are only poorly-guessed at using other methods. Nowhere is there a more stringent criterion on data-editing than a forecast that is either accurate or poor following its presentation. The UCSD website is continually being updated as new techniques are developed, and thus these are but the present samples of presentations that may eventually become available using the data sets in space weather forecasting. One of the most-sought heliospheric parameters is that of magnetic field since this is the primary parameter that determines the interactions present between magnetized planetary bodies and the solar wind. This parameter fares poorly in our current transient presentations because it is generally very difficult to extrapolate it from photospheric magnetic fields. We expect that some inroad can be made in the current forecast analysis by matching derived radial and tangential magnetic fields to past measurements made in situ, as we do for velocity and density, but in general the CSSS model technique will never follow the rapid excursions present in the short-term, transient field in-situ components. Another remote sensing technique to forecast heliosphere magnetic fields can perhaps be realized using Faraday rotation observations,37,38 but these measurements (like those from vector magnetic fields and the best 3D-MHD analyses) are still distant future prospects. These analyses provide at best very low resolution forecasts. More abundant and precise IPS measurements would help greatly to achieve

July 13, 2012 9:4

ADGEO

112

9in x 6in

b1389-ch07

B. V. Jackson et al.

better results. More abundant IPS data from the Ooty radio array are currently available as archival data, and current plans to operate this array in real time are underway. When these data are available in real time these same programming techniques will be available to test the precision of the Ooty data. These combined with IPS data sets from other large IPS arrays have the potential to provide real break-throughs’ in this type of analysis. Even so, IPS data have a basic drawback in that they are only available above the location of a given observer and the fastest transient events could be missed entirely (or in part) by observations made at any given terrestrial longitude. While the construction of large radio arrays at different Earth longitudes is a possibility in order to rectify this situation, manning and calibrating each system around the world is a formidable undertaking. Less ground-based, labor-intensive and much higher resolution than the IPS analyses are the spacecraft imaging techniques developed using the SMEI instrument. Analysis of these data sets can provide the same type of remote-sensing analyses and, with careful design and enough throughput, such an instrument39 could provide not only a very high resolution Thomson-scattering density proxy, but also a velocity parameter proxy. This measurement could be made hourly, or even more frequently, from coronagraph heights out to elongations beyond 90◦ in order to provide higher-resolution and quantitative time series measurements and forecasts as already shown to be possible for density using SMEI archival data.40 However, such instruments also have a basic drawback in that they too are difficult to calibrate and maintain so that they provide quality data, and very often the uncalibrated image data give enough of a “good story” that prevents the full performance of such expensive systems (relative to those that are ground-based) from being realized. In short we see no quick solution to the best heliospheric space weather forecasting, and we do not know how these remote-sensing techniques will fare in the distant future. However, over the relatively short period of using data from STELab, there has been a significant improvement both in the forecast and research capability shown to be possible based on IPS analyses.

Acknowledgements We thank M. Kojima for his encouragement throughout the length of this UCSD IPS analysis project and for originally making available IPS data for these analyses under the auspices of a joint CASS/UCSD–STELab

July 13, 2012 9:4

ADGEO

9in x 6in

Forecasting Transient Heliospheric Solar Wind Parameters

b1389-ch07

113

cooperative agreement. The work of B.V. Jackson, and J.M. Clover was supported at the University of California at San Diego by NSF grants ATM0852246 and AGS-1053766, and NASA grant NNX11AB50G.

References 1. B. V. Jackson, P. P. Hick, A. Buffington, M. M. Bisi, M. Kojima and M. Tokumaru, Astronomical and Astrophysical Transactions 26(6) (2007) 477–487. 2. Z. Houminer, Nature Phys. Sci. 231 (1971) 165–167. 3. A. Hewish and S. Bravo, Solar Phys. 106 (1986) 185–200. 4. M. Kojima and T. Kakinuma, J. Geophys. Res. 92 (1987) 7269–7279. 5. K. W. Behannon, L. F. Burlaga and A. Hewish, J. Geophys. Res. 96 (1991) 21,213–21,225. 6. P. K. Manoharan, S. Ananthakrishnan, M. Dryer, T. R. Detman, H. Leinbach, M. Kojima, T. Watanabe and J. Kahn, Solar Phys. 156 (1995) 377–393. 7. B. V. Jackson, P. L. Hick, M. Kojima and A. Yokobe, J. Geophys. Res. 103 (1998) 12,049–12,067. 8. M. Kojima, M. Tokumaru, H. Watanabe, A. Yokobe, K. Asai, B. V. Jackson and P. L. Hick, J. Geophys. Res. 103 (1998) 1981–1989. 9. B. V. Jackson and P. P. Hick, Three-dimensional tomography of interplanetary disturbances, Solar and Space Weather Radiophysics, Current Status and Future Developments (Astrophysics and Space Science Library), eds. D. E. Gary and C. U. Keller, 314 (Kluwer Academic Publ., Dordrecht, The Netherlands, 2005) 355–386. 10. B. V. Jackson, P. P. Hick, A. Buffington, M. M. Bisi, J. M. Clover and M. Tokumaru, Adv. in Geosciences 21 (2008) 339–366. 11. M. M. Bisi, B. V. Jackson, P. P. Hick, A. Buffington and J. M. Clover, J. Geophys Res. — Space Physics Special Edition — Geomagnetic Storms of Solar Cycle 23, 113 (2008) A00A11. 12. M. M. Bisi, B. V. Jackson, A. Buffington, J. M. Clover, P. P. Hick and M. Tokumaru, Solar Phys. 256 (2009) 201–217. 13. B. V. Jackson, P. P. Hick, M. M. Bisi, J. M. Clover and A. Buffington, Solar Phys. 265 (2010) 245–256. 14. P. K. Manoharan, Solar Phys. 265 (2010) 137–157. 15. B. V. Jackson, M. S. Hamilton, P. P. Hick, A. Buffington, M. M Bisi, J. M. Clover, M. Tokumaru and K. Fujiki, J. Atmospheric and Solar-Terrestrial Phys. 73 (2011) 1317–1329. 16. M. Tokumaru, M. Kojima, K. Fujiki, Y. Maruyama, H. Ito and T. Iju, Radio Sci. 46 (2011) RS0F02. 17. P. K. Manoharan, Study of the solar wind using single-station interplanetary scintillation, PhD thesis, University of Bombay, 1991, pp. 53–95. 18. I. V. Chashei, V. I. Shishov, S. A. Tyul’bashev and V. V. Oreshko, IPS Observations Using the Big Scanning Array of the Lebedev Physical Institute:

July 13, 2012 9:4

114

19.

20.

21.

22.

23. 24.

25. 26. 27. 28. 29. 30.

31. 32.

33.

ADGEO

9in x 6in

b1389-ch07

B. V. Jackson et al.

Recent Results and Future Prospects, Asia Oceania Geosciences Society Meeting presentation, Taipei, 8–12 August, 2011, p. 270. C. J. Eyles, G. M. Simnett, M. P. Cooke, B. V. Jackson, A. Buffington, P. P. Hick, N. R. Waltham, J. M. King, P. A. Anderson and P. E. Holladay, Solar Phys. 217 (2003) 319–347. B. V. Jackson, A. Buffington, P. P. Hick, R. C. Altrock, S. Figueroa, P. E. Holladay, J. C. Johnston, S. W. Kahler, J. B. Mozer, S. Price, R. R. Radick, R. Sagalyn, D. Sinclair, G. M. Simnett, C. J. Eyles, M. P. Cooke, S. J. Tappin, T. Kuchar, D. Mizumo, D. F. Webb, P. A. Anderson, S. L. Keil, R. E. Gold and N. R. Waltham, Solar Phys. 225 (2004) 177–207. A. Buffington, J. S. Morrill, P. P. Hick, R. A. Howard, B. V. Jackson, and D. F. Webb, Analyses of the comparative responses of SMEI and LASCO, Proc. SPIE 6689, 66890B, 2007, pp. 1–6. B. V. Jackson, A. Buffington and P. P. Hick, A heliospheric imager for solar orbiter, Proc. of Solar Encounter: The First Solar Orbiter Workshop, Puerto de la Cruz, Tenerife, Spain, 14–18 May, 2001, pp. 251–256. B. V. Jackson and P. P. Hick, Solar Phys. 211 (2002) 345–356. B. V. Jackson, P. P. Hick, A. Buffington, M. Kojima, M. Tokumaru, K. Fujiki, T. Ohmi and M. Yamashita, Time-dependent tomography of heliospheric features using interplanetary scintillation (IPS) remote sensing observations, CP679, Solar Wind Ten: Proceedings of the Tenth International Solar Wind Conference, eds. M. Velli, R. Bruno and F. Malara, 2002, pp. 75–78. B. V. Jackson, A. Buffington, P. P. Hick, X. Wang and D. Webb, J. Geophys. Res. 111 (2006) A04S91. M. Tokumaru, M. Kojima, K. Fujiki, M. Yamashita and B. V. Jackson, J. Geophys. Res. 112 (2007) A05106. B. V. Jackson, J. A. Boyer, P. P. Hick, A. Buffington, M. M. Bisi and D. H. Crider, Solar Phys. 241 (2007) 385–396. B. V. Jackson, M. M. Bisi, P. P. Hick, A. Buffington, J. M. Clover and W. Sun, J. Geophys Res. 113 (2008) A00A15. Z. K. Smith, M. Dryer and S. M. Han, Astrophysics and Space Science, 119 (1986) 337–344. S. M. P. McKenna-Lawlor, M. Dryer, Z. Smith, K. Kecskemety, C. D. Fry, W. Sun, C. S. Deehr, D. Berdichevsky, K. Kudela and G. Zastenker, Real time predictions of three numerical models of the arrival at Earth of eleven flare/halo CME associated shocks compared with spacecraft measurements, 2002. Available at esa-spaceweather.net. M. Dryer, Z. Smith, C. D. Fry, W. Sun, C. S. Deehr and S.-I. Akasofu, Space Weather 2 (2004) S09001. C. D. Fry, T. R. Detman, M. Dryer, Z. Smith, W. Sun, C. S. Deehr, S.-I. Akasofu, C. C. Wu and S. McKenna-Lawlor, J. Atmospheric and SolarTerrestrial Phys. 69 (2007) 109–115. P. P. Hick and B. V. Jackson, Heliospheric tomography: an algorithm for the reconstruction of the 3D solar wind from remote sensing observations, Proc. SPIE 5171, 2004, pp. 287–297.

July 13, 2012 9:4

ADGEO

9in x 6in

Forecasting Transient Heliospheric Solar Wind Parameters

b1389-ch07

115

34. X. Zhao and J. T. Hoeksema, J. Geophys. Res. 100 (1995) 19–33. 35. T. Dunn, B. V. Jackson, P. P. Hick, A. Buffington and X. P. Zhao, Solar Phys. 227 (2005) 339–353. 36. A.T. Young, Astrophys J. 168 (1971) 543–562. 37. B. V. Jackson, SMEI (Solar Mass Ejection Imager) (a Facility?), invited presentation at the NSF UARS Facilities Workshop, 24 September, 2008, Haystack Observatory, MA. 38. E. A. Jensen, P. P. Hick, M. M. Bisi, B. V. Jackson, J. Clover and T .L. Mulligan, Solar Phys. 265 (2010) 31–48. 39. B. V. Jackson, A. Buffington, P. P. Hick, M. M. Bisi and J. M. Clover, Solar Phys. 265 (2010) 257–275. 40. J. M. Clover, B. V. Jackson, P. P. Hick, A. Buffington, J. C. Linford and M. M. Bisi, UCSD Time-dependent tomographic forecasting with interplanetary scintillation and white light observations, SHINE Conference, 3–8 July, 2011.

July 13, 2012 9:4

ADGEO

9in x 6in

This page intentionally left blank

b1389-ch07

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch08

Advances in Geosciences Vol. 30: Planetary Science and Solar & Terrestrial Science (2011) Ed. Anil Bhardwaj c World Scientific Publishing Company


RECENT PROGRESS OF SOLAR WEATHER FORECASTING AT NAOC HAN HE∗ , HUANING WANG, ZHANLE DU, LIYUN ZHANG, XIN HUANG, YAN YAN, YULIANG FAN, XIAOSHUAI ZHU, XIAOBO GUO and XINGHUA DAI Key Laboratory of Solar Activity, National Astronomical Observatories, Chinese Academy of Sciences, 20A Datun Road, Chaoyang District, Beijing 100012, China ∗[email protected]

The history of solar weather forecasting services at National Astronomical Observatories, Chinese Academy of Sciences (NAOC) can be traced back to 1960s. Nowadays, NAOC is the headquarters of the Regional Warning Center of China (RWC-China), which is one of the members of the International Space Environment Service (ISES). NAOC is responsible for exchanging data, information and space weather forecasts of RWC-China with other RWCs. The solar weather forecasting services at NAOC cover short-term prediction (within two or three days), medium-term prediction (within several weeks), and long-term prediction (in time scale of solar cycle) of solar activities. Most efforts of the short-term prediction research are concentrated on the solar eruptive phenomena, such as flares, coronal mass ejections (CMEs) and solar proton events, which are the key driving sources of strong space weather disturbances. Based on the high quality observation data of the latest space-based and ground-based solar telescopes and with the help of artificial intelligence techniques, new numerical models with quantitative analyses and physical consideration are being developed for the predictions of solar eruptive events. The 3-D computer simulation technology is being introduced for the operational solar weather service platform to visualize the monitoring of solar activities, the running of the prediction models, as well as the presenting of the forecasting results. A new generation operational solar weather monitoring and forecasting system is expected to be constructed in the near future at NAOC.

1. Introduction As a term to represent the conditions or activities in the Sun’s atmosphere that can influence or disturb the space environment, the Solar Weather

117

July 13, 2012 9:4

118

ADGEO

9in x 6in

b1389-ch08

H. He et al.

includes the following aspects on the Sun: • • • • • • • •

Solar active region (sunspot group), Solar flare, Coronal mass ejection (CME), Solar energetic particle (SEP), EUV, X-ray and radio radiation, Filament (prominence), Coronal hole, Solar cycle.

Solar activities are the origins of the space environment disturbances, and the solar weather forecasting is one of the important and key parts of the whole space weather services. The solar weather forecasting activities and services at National Astronomical Observatories, Chinese Academy of Sciences (NAOC) began in 1969, which were dedicated to the Chinese first satellite mission in 1970 for providing space environment support. The solar activity prediction services at NAOC continued ever since. In the early period of services, the solar weather forecasting were mainly for the short-wave communications and the series of Chinese space missions. In 1990, NAOC became a memeber of the International Space environment Services (ISES) (website: http://www.ises-spaceweather. org). The Regional Warning Center of China (RWC-China) was setup in 1991, which includes four sub-centers: Solar Activity Prediction Center (SAPC), Space Environment Prediction Center (SEPC), Ionospheric Disturbance Prediction Center (IDPC) and Geomagnetic Storm Prediction Center (GSPC), the SAPC is located in NAOC. NAOC is also the headquarters of the RWC-China, which is responsible for the data collection, user services and exchanges of data and information with other RWCs. NAOC has two solar observing bases, one is the Huairou Solar Observing Station in Beijing, the other is at Yunnan Astronomical Observatory in Kunming, Yunnan Province. (More solar observing bases of NAOC are under construction.) The most often used data types for daily solar activity forecasting include: The photospheric vector magnetogram, the Hα image of the chromosphere and the solar radio flux and spectrum. The data obtained by the international solar observing satellites and spacecrafts (e.g., GOES-series, SDO, STEREO, SOHO, etc.) are also employed for the solar weather monitoring and forecasting. The routine solar weather forecasting services at NAOC include short-term, medium-term, and long-term predictions,1 the items of each

July 13, 2012 9:4

ADGEO

9in x 6in

Recent Progress of Solar Weather Forecasting at NAOC

b1389-ch08

119

prediction are listed below: • Short-term prediction (within two or three days): (a) Maximum solar X-ray flare class in the following two days (options include none, C-, M-, X-class). (b) Solar proton event probabilities of the following three days. (c) Solar 10.7 cm radio flux (F10.7) daily values of the following three days. • Medium-term prediction (within several weeks): (a) Smoothed monthly mean sunspot number. (b) Solar X-ray flare activity level of the following week (options include very low, low, medium, high, very high). • Long-term prediction (in time scale of solar cycle): (a) Maximum value and phase of smoothed monthly mean sunspot number. The solar weather forecasting results are distributed by web pages as well as emails every weekday. Besides the routine services, special services are also provided according to the requirement of the Chinese national space missions, such as Shenzhou series of manned space flight and Chang’e series of moon exploration mission. During recent years, certain prediction models were established or improved for solar weather prediction at NAOC, new method and more physical considerations were introduced for prediction models development, and new computer system and tools as well as web platform are being constructed for solar weather monitoring, forecasting and data distribution. We will describe these contents in Secs. 2 and 3. In Sec. 4, we give a summary and future perspective of solar weather researches and services at NAOC.

2. Researches on Solar Weather Forecasting 2.1. Prediction models Currently available prediction models as well as models in development at NAOC are listed as following: • Currently available prediction models: (a) Solar flare short-term prediction model.2 (b) Solar proton event short-term prediction model.2

July 13, 2012 9:4

ADGEO

120

9in x 6in

b1389-ch08

H. He et al.

Solar 10.7 cm radio flux (F10.7) prediction model.1 Solar active longitude prediction model.3 Solar active level quantitative assessment model.4 3-D coronal magnetic field nonlinear force-free field (NLFFF) extrapolation model.5 (g) Solar cycle long-term prediction model.6

(c) (d) (e) (f)

• Prediction models in development: (a) CME prediction model. Most efforts of the prediction research are concentrated on the solar eruptive phenomena, such as flares, CMEs and solar proton events, which are the key driving sources of the strong space weather disturbances. 2.2. Physical measures of photospheric magnetic field It is believed that the sunspots (active regions) are the results of the concentrated magnetic flux emerged from the solar interior to the photosphere, the photospheric magnetic field and its evolution are the causes of the various solar weather phenomena. Compared to the traditional parameters that describe the shapes of the sunspots (e.g., area, magnetic type and Mclntosh type of the sunspot groups), physical analysis on the active region’s magnetic field can better reveal the nature of solar activities. Certain physical measures of the photospheric magnetic field are carefully selected and employed as the input factors for the short-term prediction models at NAOC. For longitudinal magnetograms (such as the magnetic field maps obtained by the MDI instrument aboard SOHO spacecraft), the three physical measures are the maximum horizontal gradient, the length of the neutral line and the number of singular points7 ; For vector magnetograms (such as the vector magnetic field data observed by Huairou Solar Observing Station of NAOC), the three additional physical measures are the length of strong-shear neutral line, the total unsigned current and the total unsigned current helicity.8 These measures reflect the complexities and energy storage level of the solar active regions.4 2.3. Artificial intelligence algorithm A new artificial intelligence algorithm, the support vector machine combined with K-nearest neighbors (SVM-KNN) classifying method,2 is employed to replace the traditional statistics method in short-term prediction modeling

July 13, 2012 9:4

ADGEO

9in x 6in

Recent Progress of Solar Weather Forecasting at NAOC

b1389-ch08

121

of solar eruptive events. Compared with traditional statistics mothod, the new SVM-KNN classifying method can utilize the input factors more effectively and can give more accurate results (69% average accuracy rate for 24-hour solar flare forecasts, with the predictors including the area of sunspot group, the magnetic classification of sunspot group, the McIntosh classification of sunspot group and the 10.7 cm solar radio flux).2 Very recently, a sequential supervised learning method is also employed for shortterm solar flare prediction by analyzing the evolutionary information of the active region’s magnetic field, which obtains the similar accuracy rate as the SVM-KNN model.9

2.4. Analyses of 3-D coronal magnetic fields It is commonly accepted that the solar eruptive phenomena, such as flares and CMEs, are the results of the coronal magnetic field reconnections that are driven by the magnetic field evolution in the photosphere. Understanding the 3-D coronal magnetic field structure will be much helpful to promote the accuracy of short-term solar weather prediction. A coronal magnetic field NLFFF extrapolation model has been established at NAOC.5 The extrapolation model uses the vector magnetograms of the photosphere as the bottom boundary, and calculates the 3-D coronal magnetic fields above solar active regions. By quantitative analyses of the 3-D magnetic field, physical and topological measures that describe the physical properties and topological structures of the coronal magnetic fields can be obtained (e.g., magnetic energy, magnetic helicity, current density, force-free factor α, null point, separatrix, quasi-separatrix layer, etc.). The coronal measures are believed to be more efficient and direct than the predictors of photosphere for short-term prediction models since the large eruptive events like X-ray flares and CMEs generally initiate in corona. Recent 3-D coronal magnetic field studies on the solar active region NOAA 10930 show some potentials of the 3-D coronal measures for the solar flare forecasting.10 As an example of the coronal measures studies, Fig. 1 exhibits the time series evolution of the current density distribution maps (left column of Fig. 1) calculated from the 3-D NLFFF modeling fields, together with the corresponding photospheric vector magnetograms (right column of Fig. 1) of NOAA 10930. As shown in Fig. 1, the magnitude of the current density increases over time along with the evolution of the photospheric magnetic fields, which reflects the activity enhancement of this active region.

July 13, 2012 9:4

122

ADGEO

9in x 6in

b1389-ch08

H. He et al.

Fig. 1. Time series evolution of the current density distribution maps (left column) calculated from the 3-D NLFFF modeling fields, together with the corresponding photospheric vector magnetograms (right column) of the solar active region NOAA 10930.10 The magnetic field data are obtained by the Hinode satellite.11

3. Operational Activities of Solar Weather Forecasting 3.1. Operational platforms for solar weather service Construction of computer platforms for operational solar weather service began in 2001 at NAOC. Basically, one platform will include three subsystems: the database system, the operational forecasting models and the web interface for system control and data distribution. Till now, there are three generations of operational platforms at NAOC, which are listed in Table 1. Currently, we mainly use the second generation platform for daily services (website: http://rwcc.bao.ac.cn). Figure 2 shows the home page and the background management interface of the second generation platform. The third generation platform is still in development, some functions of the third platform have been available for test or practical application (website: http://159.226.170.65).

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch08

Recent Progress of Solar Weather Forecasting at NAOC Table 1.

123

Three generations of operational platform for solar weather service.

Operational platform

Application period (year)

Distribution media

Supporting computer system

First generation

2001–2006

Web pages

Simple database system; Simple data table; Input observation data by hand; Run prediction models by hand

Second generation

2006–2011

Web pages;

Simple database system;

Simple English language pages

Complex data table; Grab observation data semi-automatically; Run prediction models through platform

Third generation (being developed)

2011

Dynamic and interactive web pages; Complete English language pages; 3-D computer simulation interface

Dedicated database server;

Mass storage devices; Grab and extract observation data automatically; Run prediction models automatically; 3-D computer simulation technique

3.2. 3-D computer simulation One of the most important improvements of the third generation operational platform for solar weather service is the introduction of 3-D computer simulation interface. This sub-system is named as Virtual-Sun. The VirtualSun is a Client–Server (C–S) system, the server part provide the solar activity data, the client part is installed on the user’s computer. The Virtual-Sun client obtains the solar data from the server through Internet, and displays the 3-D virtual image of the Sun as well as various solar activity components on the sun (e.g., sunspots, magnetograms, active region numbers, flares, warning or forecasting information, etc.). The 3-D virtual sun can be rotated and zoomed freely by using computer mouse or keyboard. Figure 3 shows the user interface of the Virtual-Sun client. The Virtual-Sun system is still in development, a test version of the client is available at http://159.226.170.65/virtual-sun/.

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch08

H. He et al.

124

Fig. 2. Web interface of the second generation platform for the operational solar weather services at NAOC. The left picture shows the home page of the platform, the right picture displays the background management interface (both in Chinese language).

Fig. 3.

The user interface of the Virtual-Sun client.

July 13, 2012 9:4

ADGEO

9in x 6in

Recent Progress of Solar Weather Forecasting at NAOC

b1389-ch08

125

4. Summary There are more than 40 years of solar weather services at NAOC, and 20 years of international data exchanges with other RWCs. Forecasting services at NAOC cover short-term, medium-term and long-term predictions of solar activities, and certain prediction models have been established for operational forecasting. In recent years, most efforts of the prediction research are concentrated on the solar eruptive phenomena, such as flare, CME and solar proton events, which are the key driving sources of the strong space weather disturbances. Based on the high quality near-real-time observation data of the latest space-based and ground-based solar telescopes and with the help of artificial intelligence techniques, new numerical models with quantitative analyses and physical consideration are being developed for the prediction of solar eruptive activities. The 3-D computer simulation technology is being introduced for the operational forecasting platform to visualize the monitoring of solar activities, the running of the prediction models, as well as the presenting of the forecasting results. A new generation operational solar weather monitoring and forecasting system is expected to be constructed in the near future at NAOC. Acknowledgments We are grateful to the reviewer for valuable comments and suggestions to improve the paper. Hinode is a Japanese mission developed and launched by ISAS/JAXA, with NAOJ as domestic partner and NASA and STFC (UK) as international partners. It is operated by these agencies in co-operation with ESA and NSC (Norway). This work is jointly supported by National Natural Science Foundation of China (NSFC) through grants 10803011, 40890161, 10921303 and 10733020, and National Basic Research Program of China (973 Program) through grants 2011CB811406. The authors also acknowledge the support of China Meteorological Administration grant (No. GYHY201106011).

References 1. H. He, H. Wang, Z. Du, R. Li, Y. Cui, L. Zhang and Y. He, Adv. Space Res. 42 (2008) 1450. 2. R. Li, Y. Cui, H. He and H. Wang, Adv. Space Res. 42 (2008) 1469. 3. L. Y. Zhang, H. N. Wang and Z. L. Du, Astron. Astrophys. 484 (2008) 523.

July 13, 2012 9:4

126

ADGEO

9in x 6in

b1389-ch08

H. He et al.

4. H. Wang, Y. Cui and H. He, Res. Astron. Astrophys. 9 (2009) 687. 5. H. He and H. Wang, J. Geophys. Res. (Space Physics) 113 (2008) A05S90. 6. Z. L. Du, H. N. Wang, H. He, L. Y. Zhang, R. Li and Y. M. Cui, Adv. Space Res. 42 (2008) 1457. 7. Y. Cui, R. Li, L. Zhang, Y. He and H. Wang, Solar Phys. 237 (2006) 45. 8. Y. Cui and H. Wang, Adv. Space Res. 42 (2008) 1475. 9. D. Yu, X. Huang, H. Wang and Y. Cui, Solar Phys. 255 (2009) 91. 10. H. He, H. Wang and Y. Yan, J. Geophys. Res. (Space Physics) 116 (2011) A01101. 11. T. Kosugi, K. Matsuzaki, T. Sakao, T. Shimizu, Y. Sone, S. Tachikawa, T. Hashimoto, K. Minesugi, A. Ohnishi, T. Yamada, S. Tsuneta, H. Hara, K. Ichimoto, Y. Suematsu, M. Shimojo, T. Watanabe, S. Shimada, J. M. Davis, L. D. Hill, J. K. Owens, A. M. Title, J. L. Culhane, L. K. Harra, G. A. Doschek and L. Golub, Solar Phys. 243 (2007) 3.

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch09

Advances in Geosciences Vol. 30: Planetary Science and Solar & Terrestrial Science (2011) Ed. Anil Bhardwaj c World Scientific Publishing Company


A NEW APPROACH FOR IDENTIFYING IONOSPHERIC GRADIENTS IN THE CONTEXT OF THE GAGAN SYSTEM∗ RAVI CHANDRA KUDALA Electronics Engineering Department, Defence College of Engineering, Defence University, Bishoftu, Ethiopia, [email protected]

The Indian Space Research Organization and the Airports Authority of India are jointly implementing the Global Positioning System (GPS) aided GEO Augmented Navigation (GAGAN) system in order to meet the following required navigation performance (RNP) parameters: integrity, continuity, accuracy, and availability (for aircraft operations). Such a system provides the user with orbit, clock, and ionospheric corrections in addition to ranging signals via the geostationary earth orbit satellite (GEOSAT). The equatorial ionization anomaly (EIA), due to rapid non-uniform electron-ion recombination that persists on the Indian subcontinent, causes ionospheric gradients. Ionospheric gradients represent the most severe threat to high-integrity differential GNSS systems such as GAGAN. In order to ensure integrity under conditions of an ionospheric storm, the following three objectives must be met: careful monitoring, error bounding, and sophisticated storm-front modeling. The first objective is met by continuously tracking data due to storms, and, on quiet days, determining precise estimates of the threat parameters from reference monitoring stations. The second objective is met by quantifying the above estimates of threat parameters due to storms through maximum and minimum typical thresholds. In the context GAGAN, this work proposes a new method for identifying ionospheric gradients, in addition to determining an appropriate upper bound, in order to sufficiently understand error during storm days. Initially, carrier phase data of the GAGAN network from Indian TEC stations for both storm and quiet days was used for estimating ionospheric spatial and temporal gradients (the vertical ionospheric gradient (σVIG ) and the rate of the TEC index (ROTI), respectively) in multiple viewing directions. Along similar lines, using the carrier to noise ratio (C/N0 ) for the same data, the carrier to noise ratio index (σCNRI ) was derived. Subsequently, the one-toone relationship between σVIG and σCNRI was examined. High values of σVIG were determined for strong noise signals and corresponded to minimal σCNRI , indicating poor phase estimations and, in turn, an erroneous location. On the other hand, low values of σVIG were produced for weak noise signals

127

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch09

K. Ravi Chandra

128

and corresponded to maximum σCNRI , indicating strong phase estimations and, in turn, accurate locations. In other words, if a gradient persists in the line of sight direction of GEOSAT for aviation users, the down link L- band signal itself becomes erroneous. As a result, the en-route aviation user fails to receive a SBAS correction message leading to deprivation for the main objective of GAGAN. On the other hand, since the proposed approach enhances the receivers of both the aviation user and the reference monitoring station in terms of their performance, based on σCNRI , the integrity of SBAS messages themselves can be analyzed and considered for forward corrections.

1. Introduction 1.1. An overview of the global navigation satellite system (GNSS) Ramakrishnan et al.1 stated that the aviation navigation community has been performing development testing and adopting new navigation systems. Such systems incorporate the GNSS for a wide variety of aircraft operations. The first generations of the GNSS-1 were represented by GPS and GLONASS for US and Russian military requirements, respectively. The second generation of the GNSS-2, the European global satellite navigation system (Galileo) and the Chinese COMPASS multistage satellite navigation BeiDou-2 systems, are still in the development stage. The operating frequencies of the entire GNSS constellation are located within the L-band. The selection of the L-band frequency is a compromise between higher band frequencies (which experience more path loss) and lower band frequencies (which experience more ionospheric time delay). According to Lee et al.2 and Rooney et al.,3 Galileo E1B/E1C shares a common central frequency with that of the GPS L1 frequency, such that users will be capable of determining position in the same receiver, assisted by any of these satellites in any combination. Since COMPASS signals overlap with other GNSS, there is a potential benefit for minimum service levels in the intermediate region for users. GNSS architecture comprises a space segment, a control segment, and a user segment (Fig. 1). The GNSS space segment is the union of the GPS, the GLONASS, the Galileo, and the Compass space segments, etc. When all of the GNSS constellations become operational, it will be possible to probe both the ionosphere and the troposphere in multiple directions (24 to 30) at any particular instant in time. GNSS’s control segment responsibilities include operation of the space segment, maintaining the space segment, predicting orbital and clock parameters (through the continuous tracking of satellites),

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch09

A New Approach for Identifying Ionospheric Gradients

129

GNSSSPACE SEGMENT

Ephemeris Clock corrections ionospheric Data

Pseudo-range data

CONTROL SEGMENT

Monitor station 2

Uplink Station

Monitor station 1

Monitor station N=6

Monitor station 3

Master Control Station

Monitor station 4

GPS (L1, L2, L2C, and L5), GLONASS (L1 band 1602.5625 MHz to 1615.5 MHz, L2 band :1246.4375 MHz to 1256.5 MHz) Galileo (E1B/E1C), COMPASS(1561 MHz (E2),1589 MHz (E1), 1268 MHz (E6) and 1207 MHz (E5b)), pseudo-range data current ephemeris clock corrections ionospheric data

USER SEGMENT GNSS Receiver Position, Velocity Time

Monitor station 5

Fig. 1.

Segments of the GNSS architecture.

monitoring ionospheric parameters (through precisely known monitoring stations), and uploading correction/data messages to the space segment. The purpose of uploading ionospheric data to the space segment is for onward transmission to the user segment and, in turn, ensuring improved user positional accuracy. The GNSS user segment comprises an antenna, a receiver, a control display unit, and a power-supply system. As Won et al.4 explains, a GNSS receiver that was developed based on the software designed radio (SDR)

July 13, 2012 9:4

130

ADGEO

9in x 6in

b1389-ch09

K. Ravi Chandra

approach can process a future satellite signal such as Galileo, GPS L2C, GPS L5, and GLONASS without additional hardware. A field programmable gate array or a digital signal processing that is based on a design approach is a promising solution for providing additional processing circuitry to a standalone SDR GNSS receiver in order to cope with multipath mitigation, interference cancellations, weak signal detection, etc. At any point in time, the user will be in the line of sight of roughly six to 12 satellites. Parkinson and Spilker5 stated that among multiple viewing satellite signals, four were considered the best for accurate positional information using the geometric dilution of the precision (GDOP) approach. Jean-Mari et al.,6 referred to the DOP as the effect of the geometry of satellites on measurement accuracy. Kaplan7 stated the effect as the ratio of the standard deviation of the position error to the standard deviation of the measurement error. The root mean square (RMS) position error of the user can be expressed as the product of the GDOP and the RMS ranging error.

1.2. The evolution of augmentation systems Several real-world effects intrude and degrade the accuracy of the GNSS. Hofmann-Wellenhof et al.,8 Parkinson and Spilker,5 Kaplan,7 Wilson et al.,9 and Seeber10 reported that prominent errors include ionospheric and tropospheric delays, multipath errors, ephemeris errors, instrumental bias errors, and clock errors. Due to all of these errors, standalone GPS service provides a horizontal accuracy of approximately 20 m, and a vertical accuracy of approximately 30 m. Using the differential GPS (DGPS) technique, the position of the user relative to a local reference station may be determined with an approximate 5 m accuracy. Blackwell11 proposed the basic idea associated with differential positioning by correcting the observed pseudorange of many GPS receivers (rover stations) with corrections provided from a single master receiver that was placed at a fixed point using precisely well-surveyed known coordinates (reference station). For post-processing operations using DGPS, pseudorange measurements for both rovers and a reference station are simultaneously observed and recorded. On the other hand, for real-time operations, according to Kai Borre and Gilbert,12 corrections from a reference station could be transmitted through VHF or UHF links to users for accurate position fixing. However, this system can only be used for

July 13, 2012 9:4

ADGEO

9in x 6in

A New Approach for Identifying Ionospheric Gradients

b1389-ch09

131

short baselines of up to approximately 100 km. Beyond this distance the ionosphere de-correlates and the errors due to spatial and temporal decorrelation grow as the user moves away from the reference station along the boundary of the coverage area. Furthermore, the system cannot provide continuity and availability to users with a high probability (of 99.999%). To overcome the coverage distance limitation of GNSS and DGPS around the world, various countries including the USA, the European Union, Japan, and India are augmenting the GNSS in order to support the navigation accuracy requirements of civil aviation (Horizontal: 16 m; Vertical: 6 m). As stated in ICAO guidelines, aviation operations include take-off, landing, and en-route flight. Operations are met using the following approach techniques: Precision Approach (PA) and Non-Precision Approach (NPA). In the case of NPA, the GNSS is used for horizontal positioning. For the PA case it is used for both horizontal and vertical positioning. PA is based on a constant rate of descent resulting in a smooth glide path. The glide path typically makes an angle of 30 with the horizontal and passes through a Decision Height (DH) (i.e., the height at which the pilot must decide whether or not to complete the landing operation). PA can be further classified into three categories based on DH and the Runway Visual Range (RVR), namely Cat-I, Cat-II, and Cat-III, as reported by Enge.13 The ICAO requirements for Cat-I were listed by Kibe.14 The GNSS is backed up broadly by three augmentation systems that are based on their mode of transmitting differential corrections (air, ground, and satellite) to aviation users. The first system is the aircraft based augmentation system (ABAS) for which error corrections are transmitted to a user through moving Aircraft. The second system is the ground based augmentation system (GBAS) for which differential corrections are transmitted to a user through ground-based controls such as LGF (LAAS Ground Facility) that service Category-II, Category-III, and PA civil aviation requirements. An example of the GBAS is the local area augmentation system (LAAS). The architecture of LAAS commonly comprises ground equipment and avionics. The ground equipment consists of four or more reference receivers that are separate from the LGF and broadcast transmitter. Information obtained from precisely located reference receivers is transmitted to the LGF. Ray et al.15 stated that the well-surveyed LGF, in turn, computes differential corrections and broadcasts, on VHF with integrity messages, to aircraft users within a radius of 20 to 30 miles from the airport. The aircraft user receives the LGF corrections and determines the integrity of any set of visible satellite

July 13, 2012 9:4

132

ADGEO

9in x 6in

b1389-ch09

K. Ravi Chandra

signals to the aircraft. The third system is the space based augmentation system (SBAS) for which differential corrections are transmitted to a user through a GEO-SAT payload. The operational SBAS’s are the US wide area augmentation system (WAAS), the European geostationary navigation overlay system (EGNOS), and the Japanese MTSAT satellite-based augmentation system (MSAS). In the development stage is the Russian system of differential correction and monitoring (SDCM), the Chinese satellite navigation augmentation system (SNAS), and the African satellite augmentation system (ASAS). GAGAN is one such SBAS and is planned for coverage over the Indian region to fill gaps between EGNOS and MSAS.

1.3. GAGAN deployment phases Details regarding the implementation of GAGAN were proposed by Kibe14 for the following three phases: the technology demonstration system (TDS), the initial experimental phase (IEP), and the final operational phase (FOP). GAGAN architecture comprises the following infrastructure (Fig. 2): • •

•

• •

Fifteen Indian Reference Stations (INRES) which receive and monitor GPS and GEOs signals. Two Indian mission control center (INMCC) which assesses the validity of signals from each satellite and computes ionospheric corrections; and develops ephemeris and clock information for GEOSAT. Three Indian navigation land up-link station (INLUS), which uplinks differential correction, and integrity information in GAGAN message format to GEOSAT. Associated Two communication links between INRES, INLUS, and INMCC. In the TDS and initial experiment phases, The total electron content (TEC) measurement network comprises 18 TEC stations for collecting TEC data over the Indian region (Fig. 3).

Ganeshan16 reported that GAGAN provides augmentation service beyond the Indian flight information region (FIR) through GEOSAT8, GEOSAT10 and GEOSAT18 footprint at 52, 82, and 83 degrees East longitude that spans a large proportion of the Asia-Pacific region. Since India is located within the low-latitude region, ionospheric variations are very predominant and ionospheric behavior is mainly characterized by spatial and temporal gradients.

July 13, 2012 9:4

ADGEO

9in x 6in

A New Approach for Identifying Ionospheric Gradients

Fig. 2.

b1389-ch09

133

The GAGAN architecture.

2. The Significance of Ionospheric Gradients 2.1. An introduction to ionospheric gradients The ionosphere is a dispersive medium located 50 to 1,000 km above the Earth in the upper atmosphere. Misra et al.17 reported that this layer is formed by ionization from solar radiation. Very high temperatures in the sun’s corona cause atoms of hydrogen and helium to escape from the sun’s gravity and to be transformed into a state of fully ionized plasma to neutral atoms within the Earth’s upper atmosphere. Ionospheric variations are fairly smooth within mid-latitude regions. Ratcliffe18 reported that electric fields within the ionosphere are more easily measured, based on the Faraday principle, from a geostationary satellite relative to a GNSS satellite. Fees et al.19 reported aspects of an ionospheric model using GPS and stated that, “the ionosphere status (condition) is a function of solar activity (Sun Spot Number), geo-magnetic latitude, local time of day, satellite elevation angle, Kp, and many other factors”. Kp is an index to determine the level

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch09

K. Ravi Chandra

134

40 35 Kull 30

Delhi

Dibrugar

Lucknow

Latitude in degree

Jodhpu 25

Raipur

Ahmedaba

Guwahatih

Bagdogr Kolkata

Aizwal

Bhopal Mumbai Hyderaba

20

15 Bangalor 10

Vishakhapatan

Chennai Port

Agatti Trivandrum

5 65

70

75

80

85

90

95

100

Longitude in degree Fig. 3.

The GAGAN region spanning the location of TEC receivers.

of geo-magnetic activity, and is a code related to the maximum fluctuations of the horizontal components observed on a magnetometer relative to a quiet day during a three-hour interval. The TEC is the key parameter for denoting ionospheric status and causes most of the effects within the GNSS signal. TEC is defined as the number of free electrons in a 1 m2 column along a signal path from the satellite to the receiver. Hoffman et al.8 stated that as the GNSS signal propagates through the ionosphere, the carrier experiences a phase advance, and the code experiences a group delay. The error contributed by this atmospheric layer is one of the largest sources of error that impacts GNSS positional accuracy, and, in turn, GAGAN performance. The ionospheric time delay is frequency dependent due to its disperse nature and can be eliminated using dual frequency GPS observations. Hence, there is a necessity for two carrier frequencies within the GPS design. Due to a certain ionospheric phenomenon (such as the equatorial anomaly, post-sunset TEC enhancements, structured depletions of TEC, etc.) medium and small-scale irregularities within the ionospheric

July 13, 2012 9:4

ADGEO

9in x 6in

A New Approach for Identifying Ionospheric Gradients

b1389-ch09

135

propagation path occur causing phase fluctuations and rapid changes in TEC. 2.2. Types of ionospheric gradients The GPS satellite moves with a speed of 3.9 km/sec in its orbit, whereas the Earth moves approximately 29 km/sec. Since a satellite and the Earth rotate with different speeds, the distance between these two will not be constant at any epoch. As a result, the difference in the distances with respect to time causes spatial and temporal variations within the ionosphere. Therefore, it is important to provide gradient information on both a spatial and temporal basis to aviation users. Therefore, ionospheric gradients can be classified as belonging to two types — temporal gradients and/or spatial gradients. 2.2.1. Temporal ionospheric gradients The irregular distribution of electron densities (i.e., the rate of TEC variation over a time occurrence in the line of sight (LOS) path between satellites to a user) can be referred to as the temporal gradient. Temporal gradients are expressed in TECU/minute or mm/min. Konno et al.20 used the divergence-free technique proposed by Hwang et al.21 for removing the temporal gradients of ionosphere delay in the context of LAAS. 2.2.2. Spatial ionospheric gradients The irregular distribution of electron densities (i.e., the rate of TEC variations over a distance in the LOS path either between a pair of satellites to a user or between a single SV and a pair of user stations) can be referred to as a spatial gradient. Spatial gradients are measured in TECu/km or mm/km. According to Konno et al.,20 the ionosphere-free technique proposed by Hwang et al.21 can be used to remove both spatial and temporal gradients.

2.3. The purpose of ionospheric gradient identification Since the Indian subcontinent spans the equatorial region, ionospheric behavior is erratic and severe. Scintillation often accompanies depletions in regions of steep ionospheric gradients. Therefore, sudden large changes in ionospheric time delays are expected to occur. Pulen et al.22 reported

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch09

K. Ravi Chandra

136

work in the context of the LAAS scenario in the US and affirmed that the largest observed gradient in slant ionospheric delay was 425 mm/km for higher viewing angles (> 65◦ ), while satellites experienced as high as 375 mm/km for lower viewing angles (< 15◦ ). On similar lines, during a solar storm on the 20th of November in 2003, Ramakrishnan et al.1 reported, while using the station pair method, that the largest observed gradient in slant ionospheric delay was 412 mm/km for higher viewing angles, whereas satellites experienced as high as 330 mm/km for lower viewing angles. As Luo et al.23 explains, severe ionospheric storm conditions can cause several meters of error in the range domain before being detected by the LGF in the context of LAAS. The effect of the gradient on the SV signal can be analyzed by classifying the ionospheric status in the following three ways: the ideal ionosphere (Ideal Iono), the nominal ionoshere (Nominal Iono), and the rare ionosphere (Rare Iono) (Fig. 4). In an ideal situation for the ionosphere (Fig. 4a), for a spatially uniform ionosphere bias, differential corrections from the INRES (reference station) can precisely cancel ionospheric errors for users (i.e., rover stations). In the Nominal Iono (Fig. 4b) scenario, the rate of change of ionospheric delay varies slightly, causing σVIG = 1−2 mm/km. In the Rare Iono (Fig. 4c) scenario, the rate of change of ionospheric delay varies very sharply, causing σVIG = 300−425 mm/km. Such a scenario arises on solar storm days. In the context of GNSS based civil aviation applications, Luo et al.24 stated that based on WAAS super truth data the ionospheric anomaly can be modeled as a linear semi-infinite “cloud” with a wave front pattern.

Fig. 4.

An illustration of ideal, nominal, and rare ionosphere.

July 13, 2012 9:4

ADGEO

9in x 6in

A New Approach for Identifying Ionospheric Gradients

b1389-ch09

137

The amplitude of such a wave front is the difference between maximum to minimum vertical delay. Accordingly, as an example of the scenario of the Wavefront model shown in Figs. 4b and 4c, the difference of maximum vertical Iono delay (24 m) and minimum vertical Iono delay (18 m — Nominal Iono and 23.75 m — Rare Iono) yielded approximately 25 cm and 6 m of user positional error for the Nominal Iono and Rare Iono, respectively. Gizawy and Skone25 reported that the presence of gradients in the TEC led to poor ambiguity resolution and degraded precise positioning accuracies. Mitigation of these errors is a vital task, especially for applications of SBAS and LAAS (i.e., for PA applications of category I, II, and III). Walter et al.26 reported a study on the effects of large ionospheric gradients on single frequency airborne smoothing filters for the WAAS and LAAS, as well as their mitigation techniques. Warnant and Pottiaux,27 also made an attempt to identify the possible correlation between ionospheric activity and unidentified problems within GPS data processing by duly detecting traveling ionospheric disturbances. Yoshihara28 reported that in the context of wave front models, spatial gradients were detected by a monitor station that could be used to mitigate gradient threats.

3. Measurement Methods for Ionsopheric Gradients 3.1. An estimation of temporal ionospheric gradients Since the broadcast message has a usable life on the order of a few minutes, the GAGAN must also place bounds on temporal gradients. Hence, with the help of the carrier phase differences between the L1 and L2 GPS fundamental frequencies, the rates of change of ionospheric structure over the specific line of sight satellite were determined with a sampling rate of 60 seconds. The rate of change of TEC (ROT) can be termed as an ionospheric gradient and is quantified statistically as ROTI. ROTI corresponds to the standard deviation of ROT. Basu et al.29 stated that ROTI measurements can be used to predict the presence of ionospheric scintillation. Du et al.30 reported that in the presence of weak scintillation, the scintillation index, S4 , is roughly proportional to the TEC fluctuation (i.e., S4 ≈ ROTI (2∼5 TECu/min)). The pre-processing and post-processing procedures involved in the estimation of the TEC by dual frequency GPS receivers, and the detailed algorithm developed in MATLAB code for measurements of temporal and spatial gradients were discussed in Chandra.31

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch09

K. Ravi Chandra

138

3.2. Estimations of spatial ionospheric gradients Due to the orbiting GPS satellite, the LOS between the satellite and the user moves through various regions of the ionosphere with respect to distance. In this work, the Time step method proposed by Lee et al.32 was used for estimations of ionospheric spatial gradients in the context of GAGAN. As reported in Misra and Enge,17 the ionosphere can be approximated as a thin shell of finite thickness with a condensed TEC at a height of 350 km. As such, the ionospheric thin shell concept is the basis for this method. In this work, due to the distinct mixed nature of spatial and temporal variation effects that may not be possible to decouple, gradients were derived using the time step method. 3.2.1. The time step method In this method, a receiver station that observes the single moving satellite over a period of time is configured. The configuration of this method is shown in Fig. 5. The ionospheric pierce point1 (IPP1 ) and the ionospheric pierce point2 (IPP2 ) are the points of intersection of the GPS satellite signal, with the ionosphere at an altitude of 350 km at time epochs t1 and t2 , respectively. In this scenario, initially, the ionospheric delay between the satellite and the receiver at the t1 epoch (say VTEC1) is compared with that of another t2 epoch (say VTEC2). Estimations of ionospheric delays at different epochs introduce the temporal de-correlation error. As a result, by determining the differences in these two vertical delays (as shown in Eq. (1)), differential

Fig. 5.

The time step method of configuration.

July 13, 2012 9:4

ADGEO

9in x 6in

b1389-ch09

A New Approach for Identifying Ionospheric Gradients

139

delay (τ ) is estimated (see Eq. (2)) as follows: VTEC = VTEC 1 − VTEC 2 ,
 1 1 τ = −(40.3VTEC ) 2 − 2 , fL1 fL2

(1) (2)

where fL1 = 1575.42 MHz

and

fL2 = 1227.60 MHz.

Subsequently, the distance between IPP1 and IPP2 is estimated (Eq. (3)) k and is used to obtain the vertical ionospheric gradient (σuVIG (ti )) in the line of sight direction between the user (u) and the satellite (k) at time epoch, ti (Eq. (4)), as follows: dIPP = (Re + h) cos−1 (sin ∅1 sin ∅2 + cos ∅1 cos ∅2 cos(λ1 − λ2 ))

(3)

where, (∅1 , λ1 ): (∅2 , λ2 ): Re : h:

Latitude and longitude of the IPP1 , in degrees Latitude and longitude of the IPP2 , in degrees Mean radius of the earth (6378 km) Altitude of the ionospheric shell (350 km) k σuVIG (ti ) =

VTEC ku (ti ) − VTEC ku (ti−1 ) mm/km, Dti ,ti−1

(4)

where ti : Local time of the ith epoch, and ti−1 : Local time of the (i − 1)th epoch Dti ,ti−1 : Distance between the IPPs corresponding to the time epoch’s t i and t i−1 VTEC ku (ti ): VTEC of the (ti )th epoch from the user (u) to the k th satellite k VTEC u (ti−1 ): VTEC of (ti−1 )th epoch from the user (u) to the k th satellite Datta-Barua et al.33 stated that the time step method was introduced in order to obtain sufficient sampling at distances of less than the physical separation of stations. Since the satellite moves with a speed of 3.9 km/sec approximately, it travels a distance of 234 km over a one minute period.

July 13, 2012 9:4

140

ADGEO

9in x 6in

b1389-ch09

K. Ravi Chandra

4. The Significance of the Carrier to Noise Ratio in the Context of GAGAN According to Ramalingam,34 the objective of GAGAN is to provide RNP parameters to GNSS users over the Indian service region. In order to improve the accuracy of GAGAN, an essential aspect is to ensure the careful processing of GNSS signals. Christie et al.35 stated that a lack of the thorough processing of received GNSS signals resulted in pseudorange measurement errors of a few meters to tens of meters at the zenith. Basically, a GNSS receiver provides two types of measurements, namely, the code phase and the carrier phase; and are called GNSS receiver observables that are used to estimate the range from the satellite to the receiver. According to Dyrud et al.,36 code derived TEC measurements provide a noisy but absolute measurement, and are plagued by instrumental biases that must be accounted for before the GNSS data can be reliably used for ionosphere characterization. Whereas, Jung et al.37 stated that a very accurate pseudorange and desired level of integrity could be acquired by using a carrier phase observable (measurement) of the GNSS signal, if integer ambiguity is resolved. Integer ambiguity refers to an unknown initial integer number (N) of cycles between the satellite and the receiver, soon after a receiver is turned on, and is the fractional part of the beat phase observed once an integer counter is initialized. During the tracking of a SV signal, the counter is incrementalized by one cycle whenever the partial phase changes from 2π to 0. As stated by El-Rabbany,38 as part of the modernization program a third civil signal known as L5 will be added to the first 12 Block IIF satellites along with L1 and L2 signals. Jung et al.39 reported a work using these three frequencies and proposed the instantaneous geometry-free Cascade Integer Resolution method for integer ambiguity resolution. Such an aspect, which contributes to positional accuracy and integrity is achieved among visible Satellite Vehicles (SVs) in multiple directions, examining the one-to-one relationship between signal strength and C/N0 . In other words, high values of C/N0 are produced for strongly received signals leading to better phase estimations and, in turn, location determination. At the same time, high values of C/N0 also signify gradient-free received signals. As a result, this C/N0 is considered as an important parameter that describes GPS receiver performance for identifying signals that suffer from severe ionospheric gradients. Du et al.,30 MacGougan et al.,40 Wieser and Brunner,41 and Fante42 have reported extensive work and have summarized that the value of C/N0 is a function of parameters such as noise, scintillations, multipaths,

July 13, 2012 9:4

ADGEO

9in x 6in

A New Approach for Identifying Ionospheric Gradients

b1389-ch09

141

and antenna characteristics — directivity, temperatures, viewing angles, etc. A normalized measure of the signal-to-noise ratio (SNR) is C/N0 . Sharawi et al.43 reported the general relationship relating the SNR to C/N0 as follows: C = 10 log10 (SNR × BW ), N0

(5)

where BW is the bandwidth of operation. The C/N0 is the ratio of the received carrier power level to the noise power level in a 1 Hz Band Width (BW), with units of dB-Hz. From derived C/N0 values, which correspond to every epoch, the arithmetic mean of the C/N0 values is evaluated (see Eq. (6)), as follows: C/N0  = Mean of the C/N0 =

N 1  (C/N0 )i , N i=1

(6)

where, N is the number of C/N0 values. From the obtained mean of C/N0 values, CNRI values can be estimated (see Eq. (7)), as follows: CNRI =

 (C/N0 )2  − C/N0 2 .

(7)

CNRI values sampled at intervals (N) of 5, 10, and 15 minutes were analyzed and compared. Local times corresponding to the sampling intervals of these CNRIs were estimated in comparison with temporal (ROTI) and k (ti )) ionospheric gradients in order to establish a one to one spatial (σuVIG epoch wise correspondence between C/N0 versus ROTI, and C/N0 versus k (ti ), respectively. σuVIG

5. Results In the GAGAN network, 18 TEC stations were installed in order to monitor ionospheric behavior. The processed GPS data obtained from the Space Application Center, Ahmedabad, was received through a scintillation monitor containing 23 parameters. Of these parameters, only eight, namely (as follows): (1) the user position, (2) the PRN number of satellites, (3) the GPS week number, (4) the GPS seconds of the week, (5) the elevation angle, (6) the azimuth angle, (7) the TEC in the slant direction

July 13, 2012 9:4

ADGEO

9in x 6in

K. Ravi Chandra

142

Vertical ionospheric gradient, mm/Km

b1389-ch09

PRN 4

PRN 5

PRN 9

PRN 10

PRN 11

PRN 16

PRN 19

PRN 20

PRN 25

PRN 26

PRN 29

PRN 30

(a) PRNs of Ahmdabad station (7