UPWELLING PROCESSES IN THE ARABIAN SEA WITH AN EMPHASIS ON THE WEST COAST OF
INDIA USING A NUMERICAL OCEAN MODEL
TANUJA NIGAM
CENTRE FOR ATMOSPHERIC SCIENCES INDIAN INSTITUTE OF TECHNOLOGY DELHI
APRIL 2020
© Indian Institute of Technology Delhi (IITD), New Delhi, 2020
UPWELLING PROCESSES IN THE ARABIAN SEA WITH AN EMPHASIS ON THE WEST COAST OF
INDIA USING A NUMERICAL OCEAN MODEL
by
TANUJA NIGAM
CENTRE FOR ATMOSPHERIC SCIENCES
Submitted
In fulfilment of the requirements of the degree of Doctor of Philosophy to the
INDIAN INSTITUTE OF TECHNOLOGY DELHI
APRIL 2020
Dedicated to My Family
Certificate
This is to certify that the thesis entitled “Upwelling processes in the Arabian Sea with an emphasis on the west coast of India using a numerical ocean model” being submitted by Ms. Tanuja Nigam to the Indian Institute of Technology Delhi for the award of the degree of DOCTOR OF PHILOSOPHY is a record of original bonafide research carried out by her. Ms. Tanuja has worked under my guidance and supervision and has fulfilled the requirements for the submission of this thesis. The results contained in this thesis have not been submitted in part or full to any other University or Institute for the award of any degree or diploma.
(Dr. Vimlesh Pant)
Associate Professor,
Center for Atmospheric Sciences
Indian Institute of Technology Delhi
New Delhi-110016, INDIA
Acknowledgements
With great pleasure, I express my deep sense of gratitude to my guide Prof. Vimlesh Pant, for giving me an opportunity to do my thesis work under his guidance and
supervision, and for valuable suggestions and encouragement during the course of the work. His patience, politeness and encouragement helped me to face and overcome all difficulties during my Ph.D.
I express my sincere thanks to Prof. A. D. Rao for providing me his useful advises to get clear understanding about the topic as well as for giving unending encouragement to proceed in future for achieving academic goals. I take this opportunity to thank Prof. Manju Mohan, Head of the Centre for Atmospheric Sciences (CAS) for providing necessary infrastructure facilities and other faculty members at CAS for their continuous support during the tenure of Ph.D.
I want to give special thanks to Kumar Ravi Prakash, CAS; Arulalan T, NCMRWF Noida; and Dr. Sachiko Mohanty, Scientist, IIRS Dehradun for their help to resolve many difficulties during my thesis work. I would like to thank all the members of the Ocean State Forecasting lab, who supported me throughout the duration of my
work.
Now, at the last but not the least, I want to thank from the bottom of my heart to My
Family who always stand behind me as my inner power and shower love and
affection during the Ph.D., their support and guidance make me enable to complete
this Ph.D. work successfully. Above all I am thankful to Almighty God for His eternal blessings and benevolence.
New Delhi (Tanuja Nigam)
i
Abstract
The upwelling is a process in ocean that brings the subsurface colder and often nutrient rich water to the surface. The upward vertical velocity in the water column initiated by the divergence of water mass on the surface due to the outward Ekman transport. This outward Ekman transport results from the combined effect of favorable wind stress and Coriolis force, an apparent force felt by any moving object/fluid on the rotating planet Earth. The colder upwelled water reduces sea surface temperature (SST) and, therefore, plays an important role in air-sea interaction by influencing the heat fluxes at the sea surface. Further, the colder near-surface water and supply of nutrients from subsurface depths to the upper-ocean photic zone lead to the phytoplankton bloom and enhances primary productivity of ocean. The oceanic upwelling can occur in coastal regions, equatorial oceans, and open ocean. The coastal upwelling process is primarily due to the alongshore southward (northward) wind that leads to the offshore Ekman transport at the eastern (western) boundary of the ocean basin. The equatorial upwelling takes place when the northeasterly and southeasterly trade winds generate Ekman transport northward in northern hemisphere and southward in southern hemisphere. The open ocean upwelling mainly caused by the cyclonic wind stress that supports radially outward Ekman transport and leads to upwelling.
The Indian Ocean is subjected to the wind reversal from northeasterly in winter (December to February) to southwesterly in summer (June to September) monsoon season. The intense coastal upwelling zones are found in the Arabian Sea (AS) i.e. along the Somalia, Oman and west coast of India. The poleward (equatorward) wind stress along the western (eastern) boundaries of the AS generates coastal upwelling at these coasts during the summer monsoon season. Several studies reported the coastal upwelling process at the west coast of India but, the underlying mechanisms responsible for its variability at different time scales is not well understood. The present thesis
ii
provides a detailed analysis of the climatological and interannual features of upwelling over the AS with a special emphasis on SEAS upwelling. A three dimensional ocean circulation model
‘Regional Ocean Modeling System’ (ROMS) is used to investigate the spatiotemporal variability of coastal upwelling in the AS. ROMS is a primitive equation model developed by Rutgers University, USA. The ROMS model is configured over a domain of 30°S-30°N, 30°E-90°E in the Indian Ocean with 40 sigma (bathymetry following) levels in the vertical and a horizontal resolution up to 0.125°. After a successful validation of the climatological features of the Indian Ocean, the model is used to examine the interannual variability of coastal upwelling for a period of 16 years from 1995 to 2010. The profiles of temperature and vertical velocity are analyzed as indices of upwelling intensity. The analysis revealed that the temporal variability of upwelling is dominating at annual and semiannual frequencies along these coasts. The semiannual frequency is caused due to the monsoonal wind reversal whereas, the annual frequency is associated with the variations in coastal upwelling process. Further, the fast Fourier transform analysis of SST shows interannual variability at 24-36 months (2-3 years) of comparatively weaker magnitude than annual and semiannual variability. The analysis of model results shows that 24-36 months variability is promoted by the Indian Ocean Dipole (IOD). The 2-3 years of variability, which is prominent in Somalia upwelling, is associated with the El-Niño Southern Oscillations (ENSO).
These interannual variabilities over the selected coastal upwelling regions are explained by the changes in the magnitude of alongshore wind under the influence of IOD and ENSO. The model sensitivity experiments are performed to understand the impact of IOD on the upwelling process over the west coast of India. The ROMS model is forced with composite surface atmospheric fluxes of pIOD years (1982, 1983, 1994, 1997, and 2006) and non-IOD years (from 1982 to 2010 excluding only pIOD years). The sensitivity experiments reveal that the upwelling intensity
iii
decreases during pIOD years comparative to the non-IOD years. The suppression of coastal upwelling process is identified by SST, depth of 20°C isotherm (D20), andzonal Ekman transport.
A detailed mixed-layer heat budget analysis over the upwelling region shows a drop in the vertical entrainment process during pIOD years as compared to the non IOD years. The weakening of coastal upwelling during pIOD years is found to be governed by the reduction in southward (or equatorward) meridional wind stress along the southwest coast of India. The cyclone-induced upwelling over the AS is studied in the cases of two tropical cyclones Phet and Nilofar occurred during the pre-monsoon and post-monsoon season, respectively. The impact of different ocean initial conditions (changing stratification) on the cyclone and its upper-ocean response is examined through the model sensitivity experiments.The interaction of Phet cyclone with coastal upwelling at the Oman coast is analyzed. Phet made its first landfall at Oman coast where it generated maximum cooling due its higher intensity at the landfall time. As Phet approached to the Oman coast, a combination of upward movement of subsurface colder water at the base of mixed layer (i.e. the cyclone-induced upwelling) and the upper-ocean mixing caused by cyclone resulted into a greater degree of cooling than the upwelling process alone. Analysis suggests that under the influence of Phet, the coastal upwelling process along Oman is replaced by a combination of cyclone-induced upwelling at the base of mixed layer and mixing in the upper-ocean column (70 to 80 m). The surface forcing for Phet and Nilofar cyclones are provided from atmospheric model
‘Weather Research and Forecasting’ (WRF). The analysis also highlights that the magnitude of cyclone-induced sea surface cooling is not only decided by the near-surface (upper 20 m) strong/weak stratification but also by the vertical profile of initial stratification in the deeper water column up to which the effect of cyclone-induced inertial oscillations is present.
iv
साराांश
समुद्र में अपवेल िंग एक प्रक्रिया है जो सतह पर उपसतह का ठिंडा और अक्सर पोषक तत्वों से
भरपूर पानी ाती है। क्रकनारों से दूर की तरफ जाने वा े एकमान ट्ािंसपोर्ट के कारण होने वा े ऊपरी सतह के द्रव्यमान के ववच न से पानी के स्तिंभ में ऊर्धवाटधर ददशा में एक धारा उत्पन्न होती है। यह बाहरी एकमैन ट्ािंसपोर्ट अनुकू हवा के दबाव ब और कोररओल स ब के सिंयुक्त प्रभाव से उत्पन्न होता है, यह कोररयोल स ब घूमती हुई पृथ्वी पर क्रकसी भी च ती वस्तु / तर पदार्ट द्वारा महसूस क्रकया गया एक आभासी ब है। ठिंडा अपवेल िंग का पानी समुद्र की
सतह के तापमान (एसएसर्ी) को कम करता है और इसल ए, समुद्र की सतह पर ऊष्मा के प्रवाह को प्रभाववत करके हवा-समुद्र की परस्पर क्रिया में महत्वपूणट भूलमका ननभाता है। इसके अ ावा, यह अपवेल िंग प्रक्रिया से ाया गया गहराई का ठिंडा पानी ऊपरी-महासागर के फोर्ोननक क्षेत्र में
फाइर्ोप ािंकर्न खि ने के ल ए आवश्यक पोषक तत्वों की आपूनतट करता है और महासागर की
प्रार्लमक उत्पादकता को बढाता है। महासागरीय अपवेल िंग प्रक्रिया तर्ीय क्षेत्रों, भूमर्धयरेिीय महासागरों और िु े समुद्रों में हो सकती है। तर्ीय अपवेल िंग प्रक्रिया मुख्य रूप से दक्षक्षण की
ओर की (उत्तर की ओर की) हवा के कारण होती है, जो महासागर बेलसन की पूवी (पश्श्चमी) सीमा
पर अपतर्ीय एकमान ट्ािंसपोर्ट को क्रकनारों से दूर की ओर े जाता है। भूमर्धयरेिीय अपवेल िंग तब होती है जब उत्तरपूवी और दक्षक्षणपूवी व्यापाररक हवाएिं (ट्ेड ववन्ड), उत्तरी गो ाधट में उत्तर की
ओर एकमान ट्ािंसपोर्ट और दक्षक्षणी गो ाधट में दक्षक्षण की ओर उत्पन्न होती हैं। मुख्य रूप से
चिवाती हवा के दबाव ब के कारण िु े समुद्र की अपवेल िंग प्रक्रिया जो रेडडय बाहर की ओर एकमैन ट्ािंसपोर्ट को सहारा देते है और ऊपर की ओर उपसतह का ठिंडा और पोषक तत्वों से भरपूर पानी ाती है। दहिंद महासागर को सददटयों में (ददसिंबर से फरवरी) उत्तरपश्श्चमी ददशा से आने वा ी और गलमटयों में (जून से लसतिंबर) दक्षक्षणपश्श्चमी ददशा से आने वा ी (मानसून के मौसम में) उ र्ी
हवाओिं का सामना करना पड़ता है। अरब सागर के सघन तर्ीय अपवेल िंग प्रक्रिया वा े प्रमुि क्षेत्र सोमाल या, ओमान और भारत के पश्श्चमी तर् पर पाए जाते हैं। गलमटयों में मॉनसून के मौसम के
दौरान इन तर्ों पर पश्श्चमी (पूवी) सीमाओिं के सार् ध्रुवों की ओर जाने वा ी (भूमर्धयरेिा की ओर जाने वा ी) हवाओिं का तनाव इन तर्ों पर तर्ीय अपवेल िंग प्रक्रिया पैदा करता है। कई अर्धययनों
ने भारत के पश्श्चमी तर् पर तर्ीय अपवेल िंग की प्रक्रिया की सूचना दी ेक्रकन, ववलभन्न समय की पररवतटनशी ता के ल ए श्जम्मेदार अिंतननटदहत तिंत्र को अच्छी तरह से समझा नहीिं गया है।
v
वतटमान र्ीलसस में अरब सागर की अपवेल िंग प्रक्रियाओिं और ववशेष जोर देने के सार् दक्षक्षण पूवी
अरब सागर की अपवेल िंग प्रक्रिया पर उनकी ज वायु और अिंतर वावषटक पैमाने पर एक ववस्तृत ववश् ेषण प्रदान करती है। यहााँ एक तीन आयामी महासागर सिंच न मॉड ‘रीजन ओशयन माडल िंग लसस्र्म’ (रोम्स) का उपयोग अरब सागर में तर्ीय अपवेल िंग प्रक्रियाओिं की स्र्ाननक- सामनयक पररवतटनशी ता जािंचने के ल ए क्रकया गया है। रोम्स एक वप्रलमदर्व समीकरण मॉड है
श्जसे रट्जसट ववश्वववद्या य, यु.एस.ए. द्वारा ववकलसत क्रकया गया है। रोम्स मॉड को दहिंद महासागर में 30°S-30°N, 30°E-90°E के क्षेत्र में कॉश्न्फगर क्रकया गया है, जो ऊर्धवाटधर और क्षैनतज ररजॉल्यूशन में 40 लसग्मा (बार्मीट्ी आधाररत) स्तरों के सार् 0.125° तक है। दहिंद महासागर की ज वायु सिंबिंधी ववशेषताओिं के एक सफ सत्यापन के बाद, मॉड का उपयोग 1995 से 2010 तक के 16 वषों की अवधध के ल ए तर्ीय अपवेल िंग प्रक्रियाओिं के अिंतर वावषटक पैमाने की पररवतटनशी ता की जािंच करने के ल ए क्रकया जाता है। तापमान और ऊर्धवाटधर वेग के
प्रोफाइ का ववश् ेषण अपवेल िंग प्रक्रियाओिं की तीव्रता के सिंकेत के रूप में क्रकया जाता है। ववश् ेषण से पता च ा क्रक इन तर्ों पर अपवेल िंग प्रक्रियाओिं की वावषटक और अधट-वावषटक आवृवत्तयों वा ी पररवतटनशी ता अधधक प्रभावी होती है। अधट-वावषटक आवृवत्त मानसूनी हवा के उ र् होने के कारण होती है, जबक्रक वावषटक आवृवत्त तर्ीय अपवेल िंग प्रक्रिया में लभन्नता से जुड़ी होती है। इसके अ ावा, एसएसर्ी का फास्र् फूररयर ट्ािंसफॉमट ववश् ेषण 24-36 महीनों (2-3 सा ) की पररवतटनशी ता
वावषटक और अधट-वावषटक पररवतटनशी ता की तु ना में तु नात्मक रूप से कमजोर पररमाण के में
पररवतटनशी ता ददिाता है। मॉड पररणामों के ववश् ेषण से पता च ता है क्रक 24-36 महीनों (2- 3 सा ) की पररवतटनशी ता इिंडडयन ओलशयन डाइपो (आइ ओ डी) के द्वारा उत्पन्न होती है।
2-3 सा की पररवतटनशी ता, जो सोमाल या में प्रमुि है, ए -नीनो साउदनट औलस ेशन (इ न स ओ) से जुड़ी है। यह अिंतर वावषटक पररवतटनशी ताएाँ इन चयननत तर्ीय अपवेल िंग प्रक्रियाओिं वा े क्षेत्रों पर आइओडी और इनसओ प्रक्रियाओिं के प्रभाव में होने वा े ए ोंगशोर पवन के पररमाण में
पररवतटन द्वारा पैदा होती है। भारत के पश्श्चमी तर् पर च रही अपवेल िंग प्रक्रिया पर आईओडी
के प्रभाव को समझने के ल ए मॉड सिंवेदनशी ता प्रयोगों का प्रदशटन क्रकया गया है। रोम्स मॉड को पाश्जदर्व इिंडडयन ओलशयन डाइपो सा ों (1982, 1983, 1994, 1997, और 2006) और गैर-आइओडी सा ों (1982 से 2010 तक केव पीआइओडी वषट को छोड़कर) के कम्पोश्जर् सतही
वायुमिंड ीय फ् क्स डार्ा के सार् फोसट क्रकया गया है। सिंवेदनशी ता प्रयोगों से पता च ता है क्रक गैर-आईओडी वषों की तु ना में पीआईओडी वषों के दौरान अपवेल िंग प्रक्रिया की तीव्रता कम हो
vi
जाती है। तर्ीय अपवेल िंग प्रक्रिया के तीव्रता की कमीिं को एसएसर्ी, 20 डडग्री सेश्ल्सयस आइसोर्मट (डी 20), और जोन एकमैन ट्ािंसपोर्ट के पररमाण की माप से पहचाना जाता है। अपवेल िंग प्रक्रिया
के क्षेत्र पर एक ववस्तृत लमधित- ेयर हीर् बजर् ववश् ेषण गैर-आईओडी वषों की तु ना में
पीआइओडी वषों के दौरान ऊर्धवाटधर इनट्ेनमेन्र् प्रवेश प्रक्रिया में एक धगरावर् ददिाता है।
पीआईओडी वषों के दौरान तर्ीय अपवेल िंग प्रक्रिया के कमजोर पड़ने को भारत के दक्षक्षण-पश्श्चमी
तर् पर प्रभावी दक्षक्षण की ओर की (भूमर्धयवती) मेररडाायोन वविंड स्ट्ेस में कमी से ननयिंत्रत्रत क्रकया जाता है। यहााँ अरब सागर में चिवात से प्रेररत अपवेल िंग प्रक्रिया का अर्धययन दो
उष्णकदर्बिंधीय चिवातों (फेर् और नी ोफर जो िमशः मानसून और प्री-मानसून के मौसम के
दौरान हुए र्े) के मार्धयम से क्रकया गया है। यहााँ चिवात और इसके ऊपरी-महासागरीय प्रनतक्रिया
पर ववलभन्न महासागरीय प्रारिंलभक श्स्र्नतयों (बद ते स्तरीकरण) के प्रभाव को मॉड सिंवेदनशी ता
प्रयोगों के मार्धयम से जािंच की गयी है। ओमान तर् पर तर्ीय अपवेल िंग प्रक्रिया के सार् फेर्
चिवात के परस्पर प्रभाव का ववश् ेषण क्रकया गया है। फेर् ने ओमान तर् पर अपना पह ा ैंडफॉ क्रकया, जहािं ैंडफॉ के समय इसकी तीव्रता अधधक होने के कारण इसने महासागर की
ऊपरी सतह पर अधधकतम शीत न उत्पन्न क्रकया। जैसे ही ओमान तर् के पास फेर् चिवात पहुाँचा, दो ववलभन्न प्रक्रियाओिं (यानी चिवात-प्रेररत अपवेल िंग प्रक्रिया और चिवात के कारण ऊपरी-सागर के लमिण की प्रक्रिया) के सिंयोजन से लमधित परत के अिंदरूनी लसरे पर उपसतह का
ठिंडा पानी ऊपर की ओर आता है और सतह को अत्यधधक ठिंडा करता है, श्जतना क्रक अपवेल िंग प्रक्रिया अके े कर सकती है। ववश् ेषण से पता च ता है क्रक फेर् के प्रभाव में, ओमान के तर्ीय अपवेल िंग प्रक्रिया को लमधित परत के आधार पर चिवात से प्रेररत अपवेल िंग के सिंयोजन से बद ददया जाता है और ऊपरी-सागर के लमधित पानी के स्तिंभ (70 से 80 मीर्र) में लम ाया जाता है।
फेर् और नी ोफर चिवातों के ल ए सतह पर वायुमिंड ीय फोलसिंग ‘वेदर ररसचट एिंड फोरकाश्स्र्िंग’(डब ू आर एफ) मॉड से प्रदान की गयी है। ववश् ेषण में यह भी बताया गया है क्रक चिवात से प्रेररत समुद्री सतह के ठिंडा होने का पररमाण न केव ननकर् सतह के (ऊपरी 20 मीर्र) मजबूत / कमजोर स्तरीकरण से तय होता है, बश्ल्क गहरे पानी के स्तिंभ में प्रारिंलभक स्तरीकरण के ऊर्धवाटधर प्रोफाइ से भी होता है। चिवात प्रेररत जड़त्वीय दो नों का प्रभाव भी
मौजूद होता है।
vii
Table of Contents
Page No.
Certificate
Acknowledgements Abstract
Table of Contents List of Figures
Chapter 1 Introduction 1-23
1.1 Background of the Study 2
1.2 Oceanic upwelling 4
1.2.1 Coastal upwelling 5
1.2.2 Equatorial upwelling 6
1.2.3 Open ocean (cyclone-induced) upwelling 6
1.3 North Indian Ocean circulation 7
1.4 Arabian Sea basin characteristics 11
1.4.1 Sea Surface Temperature 11
1.4.2 Sea Surface Salinity 13
1.4.3 Stratification 14
1.5 Upwelling features over the north Indian Ocean 15 1.6 Literature survey on the topic 18
1.7 Motivation of the study 21
1.8 Outline of the thesis 23
Chapter 2 Model, data and methodology 24-54
2.1 Introduction 25
2.2 ROMS model details 26
2.3 Model equations 28
2.4 Horizontal and vertical discretization 30 2.5 Initial and boundary conditions 34 2.5.1 Open lateral boundary conditions 34 2.5.2 Closed lateral boundary conditions 35 2.5.3 Surface and bottom boundary conditions 35
2.6 Mixing Schemes 36
2.6.1 Horizontal mixing schemes 36
2.6.2 Vertical mixing schemes 36
2.7 Data used 37
2.8 Model domain 38
2.9 Model configuration 40
2.10 Model validation 41
2.10.1 Sea Surface Temperature 43
viii
2.10.2 Sea Surface Salinity 44
2.10.3 Sea surface circulation 45
2.10.4 Temperature profile 50
2.10.5 Salinity profile 51
2.10.6 Vertical velocity 52
2.11 Summary 53
Chapter 3 Interannual Variability of Coastal Upwelling in Different Zones over Arabian Sea
55-83
3.1 Introduction 56
3.2 Model, data and methodology 58
3.2.1 Numerical experiments and data 58
3.3 Result and discussion 60
3.3.1 Sensitivity of the simulation of coastal upwelling to different heat fluxes during 1995 to 2010 over the Arabian Sea
60 3.3.2 Interannual variability of upwelling features 69
3.3.2.1 Model validation 69
3.3.2.2 Latitudinal variability of upwelling features 70 3.3.2.3 Spectral analysis of upwelling indices 76
3.4 Summary 82
Chapter 4 Impact of positive Indian Ocean Dipole (pIOD) on the coastal upwelling in the Southeastern Arabian Sea
84-109
4.1 Introduction 85
4.2 Model, data and methodology 86
4.2.1 Analysis of the different upwelling indices during positive and negative IOD events
86
4.2.2 Model details 91
4.2.3 Numerical experiments and data 93
4.2.4 Heat budget calculation 93
4.3 Results and discussion 94
4.3.1 Model validation 94
4.3.2 Surface forcing during normal and pIOD years 95 4.3.3 Impact of pIOD on upwelling at southwest coast of India
96
4.3.4 Mixed layer heat budget analysis 101
4.3.5 Impact on biological productivity 105
4.4 Summary 108
Chapter 5 Response of the cyclone-induced upwelling to different vertical ocean stratifications over the Arabian Sea
110-137
5.1 Introduction 111
5.2 Model, data and methodology 114
ix
5.2.1 Model and experiment details 114
5.2.2 Methodology 117
5.3 Results and discussion 118
5.3.1 Model validation 119
5.3.2 Cyclone-induced upwelling 119
5.3.3 Analysis of temperature and baroclinic kinetic energy profiles
129 5.4 Roles of upper-ocean stratification and baroclinic velocity shear
132
5.5 Summary 136
Chapter 6 Conclusions and Future scope of the work 138-144
6.1 Conclusion 140
6.2 Future scope of the work 144
References 145-167
Appendix A 168-174
Appendix B 175-178
List of Acronyms 179-180
Biographical Sketch 181-185
x
List of Figures
Figure No. Title Page
No.
Figure 1.1 The ‘Ekman spiral’ showing the changes in current direction and magnitude with the increasing depth in ocean. The directions of surface wind, surface current, and Ekman transport are marked in the figure.
3 Figure 1.2 Different types of upwelling processes (a) Coastal upwelling, (b)
Coastal downwelling, (c) Equatorial upwelling, (d) Cyclone-induced upwelling.
8 Figure 1.3 The wind stress (Nm-2) direction and magnitude over the Indian Ocean
during July (left) and January (right) (Adopted from Schott and McCreary, 2001).
10 Figure 1.4 The sea surface circulation features over the Indian Ocean during July
(left) and January (right) (Adopted from Schott and McCreary, 2001).
10 Figure 1.5 Sea surface temperature (°C) from Advanced Very High Resolution
Radiometer (AVHRR) in upper panel and Chlorophyll concentration (mg m-3) from Sea-Viewing Wide field-of- view Sensor (SeaWIFS) in lower panel over the Indian Ocean during winter (December-February) and summer (June-September).
17
Figure 2.1 The Arakawa C-grid chosen in the ROMS model for the horizontal discretization. u, v, ρ represent zonal velocity, meridional velocity, density, respectively.
31 Figure 2.2 The vertical staggered Arakawa C-grid arrangement in the ROMS
model.
32 Figure 2.3 The model domain and bathymetry (meter). The boxes are the
prominent upwelling zones.
39 Figure 2.4 The seasonal climatology of sea surface temperature (℃) from OISST
(lower panel) and ROMS model (upper panel). The spatial correlation between model and OISST sea surface temperature (below)
41-42 Figure 2.5 The seasonal climatology of sea surface salinity (psu) from NIO (lower
panel) and ROMS model (upper panel).
42 Figure 2.6 The seasonal climatology of sea surface current (ms-1) from SODA
(lower panel) and ROMS model (upper panel).
43 Figure 2.7 The seasonal evolution of temperature profile (℃) over coastal regions
off the Somalia (left), west coast of India (middle), Oman (right) simulated by ROMS model (upper panel) and SODA (lower panel).
46 Figure 2.8 The monthly evolution of salinity profile (psu) over the coastal regions
off the Somalia (left), west coast of India (middle), Oman (right) simulated by ROMS model (upper panel) and SODA (lower panel).
47 Figure 2.9 The monthly evolution of vertical velocity (ms-1×10-5) over the coastal
regions off the Somalia (left), west coast of India (middle), Oman (right) simulated by ROMS model (upper panel) and SODA (lower panel).
47
Figure 2.10 The spatially averaged temperature profile (℃) during summer monsoon season over the coastal regions off the Somalia (left), west
48
xi
coast of India (middle), Oman (right) simulated by ROMS model (upper panel) and SODA (middle panel) with the difference between observation and model temperature (lower panel).
Figure 2.11 The spatially averaged salinity profile (psu) during summer monsoon season over the coastal regions off the Somalia (left), west coast of India (middle), Oman (right) simulated by ROMS model (upper panel) and SODA (middle panel) with the difference between observation and model temperature (lower panel).
49
Figure 2.12 The spatially averaged vertical velocity (ms-1×10-5) during summer monsoon season over the coastal regions off the Somalia (left), west coast of India (middle), Oman (right) simulated by ROMS model (upper panel) and SODA (middle panel) with the difference between observation and model temperature (lower panel).
50
Figure 3.1 Model domain and bathymetry (m). White boxes mark prominent regions of coastal upwelling over the southwest coast of India [Box1:
74.5°E-76.5°E, 8°N-10°N], Somalia coast [Box2: 50°E-52°E, 7°N- 12°N], and Oman coast [Box3: 57°E-61°E, 17°N-21°N].
60
Figure 3.2 Sea surface temperature (°C) in 1st row for Exp_ERA (a, d, g), 2nd row for AVHRR (b, e, h) and 3rd row for Exp_Trop (c, f, i) for JJAS average of years 1999, 2004, and 2009. The bottom two panels show the difference between observed SST from AVHRR and modelled (Exp_ERA and Exp_Trop) SST.
64
Figure 3.3 The surface heat fluxes (in Wm-2) of ERA-Interim and TropFlux (a) shortwave radiation (b) Longwave radiation (c) latent heat flux (d) sensible heat flux (e) net heat flux averaged over the Box1 location off the west coast of India [74.5°E-76.5°E, 8°N-10°N].
65
Figure 3.4 The surface heat fluxes (in Wm-2) of ERA-Interim and TropFlux (a) shortwave radiation (b) Longwave radiation (c) latent heat flux (d) sensible heat flux (e) net heat flux averaged over Box2 region off the Somalia coast [50°E-52°E, 7°N-12°N].
66
Figure 3.5 The surface heat fluxes (in Wm-2) of ERA-Interim and TropFlux (a) shortwave radiation (b) Longwave radiation (c) latent heat flux (d) sensible heat flux (e) net heat flux averaged over Box3 region off the Oman coast [57°E-61°E, 17°N-21°N].
67
Figure 3.6 The surface net heat fluxes (in Wm-2) for (a) ERA-Interim and TropFlux averaged over whole model domain (b) difference between Exp_trop and Exp_ERA averaged over whole model domain (c) difference between Exp_Trop and Exp_ERA averaged over the Box1 region off the west coast of India [74.5°E-76.5°E, 8°N-10°N].
68
Figure 3.7 The time series of sea surface temperature (°C) from model simulations and observation (OISST) averaged over coastal upwelling regions of (a) west coast of India (b) Somalia coast (c) Oman Coast.
70 Figure 3.8 The time series of upwelling features averaged over coastal upwelling
region of west Coast of India (a) model simulated SST (b) meridional wind stress (c) zonal Ekman transport.
72
xii
Figure 3.9 The time series of upwelling features averaged over coastal upwelling region of Somalia (a) model simulated SST (b) meridional wind stress (c) zonal Ekman transport.
72 Figure 3.10 The time series of upwelling features averaged over coastal upwelling
region of Oman (a) model simulated SST (b) meridional wind stress (c) zonal Ekman transport.
73 Figure 3.11 The meridional transport (kg m-1s-1) at Somalia coast (upper panel),
west coast of India (middle panel), Oman coast (lower panel) with a constant value of meridional stress (N m-2)
74 Figure 3.12 Monthly variability of model simulated SST (°C) averaged over coastal
upwelling regions (a) West coast of India (b) Somalia (c) Oman for the period (1995-2010). The vertical bars show one standard deviation.
75 Figure 3.13 The power spectral analysis of observed and model simulated SST
(upper panel), model simulated vertical velocity (lower panel) averaged over the coastal upwelling regions (a) west coast of India (b) Somalia (c) Oman for the period of 1995-2010.
78
Figure 3.14 The power spectral analysis of observed components of surface wind stress (upper panel) and Ekman mass transport (lower panel) averaged over the coastal upwelling regions (a) west coast of India (b) Somalia (c) Oman for the period of 1995-2010.
79
Figure 3.15 Upper panel - Cross-correlation estimate (coherence estimation) with 95% (p= 0.05) significance level between DMI index (calculated by using HADISST SST anomalies) and meridional wind stress (Nm-2) over the model domain. Lower panel -The bootstrapped (N=2000) cross-correlation estimate with 95% (p= 0.05) significance level between DMI index (calculated by using HADISST SST anomalies) and meridional wind stress over the model domain.
80
Figure 3.16 Upper panel - Cross-correlation estimate (coherence estimation) with 95% (p= 0.05) significance level between ONI index calculated over Nino 3.4 region [170°W-120°W, 5°S-5°N] using HADISST SST anomalies and meridional wind stress (Nm-2) over the model domain.
Lower panel - The bootstrapped (N=2000) cross-correlation estimate with 95% (p= 0.05) significance level between ONI index calculated over Nino 3.4 region using HADISST SST anomalies and meridional wind stress (Nm-2) over the model domain.
81
Figure 4.1 Dipole Mode Index (DMI) calculated from the OISST data for years 1982–2010. The dashed lines are marked at one standard deviation of the SST anomaly.
87 Figure 4.2 The composite of meridional wind stress (Nm-2) for Normal years
(upper panel), pIOD years (middle panel), nIOD years (lower panel).
89 Figure 4.3 The composite of sea surface temperature (°C) for Normal years (upper
panel), pIOD years (middle panel), nIOD years (lower panel).
90 Figure 4.4 The composite of chlorophyll concentration (mg m-3) for Normal years
(upper panel), pIOD years (middle panel), nIOD years (lower panel).
90 Figure 4.5 Model domain along with bathymetry (m). The region of coastal
upwelling at the southwest coast of India is marked with a box.
92
xiii
Figure 4.6 Sea surface temperature (°C) in color shaded and surface currents (m s−1) in vectors simulated from ROMS model (a–d) in upper panel and observations from OISST and OSCAR surface currents (e–h) in lower panel.
95
Figure 4.7 Upper panel: (a–d) Climatology of meridional wind stress in shading (N m−2) overlaid with climatological wind stress vectors (m s−1). Lower panel: (e–h) Difference (pIOD-normal years) of meridional wind stress in shades (N m−2) derived from TropFlux data.
97
Figure 4.8 Comparison of model-simulated temperature profiles (°C) off the southwest coast of India along 9° N latitude for the normal years (a–d) in upper panel, pIOD years (e–h) in middle panel, and the difference (pIOD-normal) between them (e-a for June, f-b for July, g-c for August, h-d for September) in lower panel.
99
Figure 4.9 Zonal Ekman transport (in kg m−1 s−1) calculated for the (a) normal years and (b) pIOD years along the 9° N latitude off the southwest coast of India.
99 Figure 4.10 Model-simulated vertical velocity (× 10−5 m s−1) off the southwest coast
of India for the normal years (a–d) in upper panel and pIOD years (e–
h) in lower panel. The difference of the vertical velocity simulations of pIOD and Normal years (e-a to h-d).
103
Figure 4.11 Comparison of the depth of 23 °C isotherm (D23 in m) and sea surface temperature (°C) for the normal and pIOD years over the region marked with box in Fig.4.1 at the southwest coast of India.
104 Figure 4.12 Various terms of the mixed layer heat budget in units of °C day−1
calculated from model simulations carried out for the normal years (a) in upper panel and pIOD years (b) in middle panel. A closer view of the vertical entrainment term (°C day−1) for the normal and pIOD years shown in (c) in the bottom panel.
104
Figure 4.13 Composite of phytoplankton concentrations (× 101 μmol L−1) for August–October months of the normal years (left panel), pIOD years (middle panel), and climatology of phytoplankton concentrations (× 101 μmol L−1) in the right panel.
106
Figure 4.14 Composite of zooplankton concentrations (× 101 μmol L −1) for August–
October months of the normal years (left panel), pIOD years (middle panel), and climatology of zooplankton concentrations (× 101 μmol L−1) in the right panel.
107
Figure 4.15 Profile of nitrate concentration (μmol L−1) for normal years (left panel), pIOD years (middle panel), and climatology (right panel) at the southwest coast of India.
107 Figure 5.1 (a) Model domain and bathymetry (meter). (b) Tracks of cyclones Phet
and Nilofar cyclones. Locations of buoy AD06 (66.98°E, 19.00°N) and AD08 (68.67°E, 12.00°N) are denoted by blue circles. Argo profile locations are marked with purple and green circles for the Phet and Nilofar cyclones, respectively.
121
Figure 5.2 Sea Surface Temperature (°C) from TMI and AMSR-E merged optimum interpolated data (left panel) and ROMS model (right panel)
122
xiv
for Phet cyclone. The box marked in the lower panel represents the region of maximum cooling and brown dot on the track (thick black line) shows the location of cyclone at that day.
Figure 5.3 Sea Surface Temperature (°C) from TMI and AMSR-E merged optimum interpolated data (left panel) and ROMS model (right panel) for Nilofar cyclone. The box marked in the lower panel represents region of maximum cooling and brown dot on the track (thick black line) shows the location of cyclone at that day.
123
Figure 5.4 Time series of temperature profiles measured at the buoy (a, c) and model simulated (b, d) at the buoy locations and the difference between model and buoy (e, f). The dashed lines in a-d denote the depth of 26.5°C isotherm.
124
Figure 5.5 Temperature profiles (°C) from the ARGO measurements (green curves) and model simulations (blue curves) at the ARGO locations for Phet (left column) and Nilofar (right column).
124 Figure 5.6 Isopycnal displacement of thermocline (m) and vertical velocity (m s-1)
× 10-4 at the base of mixed layer for Phet (left) and Nilofar (right) at their respective selected cooling regions (marked by box in Fig. 5.2 and Fig. 5.3).
127
Figure 5.7 Daily surface divergence (s-1) × 10-5 for Phet (upper panel) and Nilofar (lower Panel)
128 Figure 5.8 Mixed layer depth (m) and mixed layer averaged temperature diffusion
(m2 s-1) for Phet (left) and Nilofar (right) over their respective regions of maximum cooling (marked by box in Fig. 5.2 and Fig. 5.3).
128 Figure 5.9 Daily average temperature profiles (°C) for (a) Phet(I), (b) Phet(R), (c)
Nilofar(R), and (d) Nilofar(I) at their respective regions of maximum cooling.
131 Figure 5.10 Daily average baroclinic kinetic energy profiles (m2 s-2) for (a) Phet(I),
(b) Phet(R), (c) Nilofar(R), and (d) Nilofar(I) over their respective regions of maximum cooling.
131 Figure 5.11 Time series of the Brunt Väisälä frequency (cycles hr-1) for (a) Phet(I),
(b) Phet(R), (c) Nilofar(R), and (d) Nilofar(I) over the region of maximum cooling.
134 Figure 5.12 Daily baroclinic velocity vertical shear (s-2) for (a) Phet(I), (b) Phet(R),
(c) Nilofar(R), and (d) Nilofar(I) over the region of maximum cooling.
134 Figure 5.13 Monthly climatological potential density (kg m-3) at surface (a, c) and
20 m depth (b and d) derived from Argo data for June and October. The surface currents (m s-1) from OSCAR data are overlaid as vectors in panels (a-d). Monthly climatological potential density (kg m-3, thick) and Brunt Väisälä Frequency (cycle hr-1, dashed) profiles derived from Argo data for June and October for Phet (black) and Nilofar (red) respectively over maximum cooling analysis boxes marked Fig 13e.
135
xv
List of Tables
Table No. Title Page
No.
Table 2.1
ROMS model configuration 40Table 3.1
Description of Experiments and their respective surface forcing datasets59
Table 3.2
Comparison of Model performance for both experiments and their respective error magnitudes61
