Master thesis

分类号 : _________________ 单位代码 : ____10300_______ 密级 : __________________ 学 号 : ____20175201014___ 硕士学位论文 The Relationship between the Trough Modes of the Bay of Bengal and Rainfall over Indochina Peninsula after South Asia Monsoon Onset 南亚季风爆发后孟加拉湾低槽模态与中南半岛降水关系研究 申 请 人 姓 名 : INMANY SOUTHIN 指 导 教 师 : PROFESSOR WANG JIANHONG 专 业 名 称 : CLIMATE 研 究 方 向 : METEOROLOGY 所 在 学 院 : COLLEGE OF ATMOSPHERIC SCIENCE 二〇一 九 年 六 月 十七 日

The Relationship between the Trough Modes of the Bay of Bengal and Rainfall over Indochina Peninsula after South Asia Monsoon Onset 南亚季风爆发后孟加拉湾低槽模态与中南半岛降水关系研究 Thesis Submitted to Faculty of Graduate Studies, Nanjing University of Information Science and Technology For the degree of Master of Meteorology by INMANY SOUTHIN (Meteorology) Supervisor: Professor Wang Jianhong 2019. 06. 17

独创性声明 本人声明所呈交的论文是我个人在导师指导下进行的研究工作及取得的研究成果。 本论文除了文中特别加以标注和致谢的内容外，不包含其他人或其他机构已经发表或撰写 过的研究成果，也不包含为获得南京信息工程大学或其他教育机构的学位或证书而使用过 的材料。其他同志对本研究所做的贡献均已在论文中作了声明并表示谢意。 学位论文作者签名：____________ 签字日期：__2019-06-17____ 关于论文使用授权的说明 南京信息工程大学、国家图书馆、中国学术期刊（光盘版）杂志社、中国科学技术 信息研究所的《中国学位论文全文数据库》有权保留本人所送交学位论文的复印件和电子 文档，可以采用影印、缩印或其他复制手段保存论文，并通过网络向社会提供信息服务。 本人电子文档的内容和纸质论文的内容相一致。除在保密期内的保密论文外，允许论文被 查阅和借阅，可以公布（包括刊登）论文的全部或部分内容。论文的公布（包括刊登）授 权南京信息工程大学研究生部办理。 □公开 □保密（_____年 _____月） (保密的学位论文在解密后应遵守此协议) 学位论文作者签名：______________ 签字日期：__2019-06-17____ 指导教师签名：______________ 签字日期：__2019-06-17____

DECLARATION I, INMANY SOUTHIN, declare that this thesis is my own unaided research work. It is being submitted in partial fulfillment of the requirement for the award of MASTER OF ATMOSPERIC SCIENCE in METEOROLOGY at Nanjing University of Information Science and Technology, Nanjing, Jiangsu Province, China. I further declare that this work never been submitted in part or whole for any purpose anywhere, and the thesis is presented with the consent of my supervisor. Works by the authors, which served as source of information, have been duty acknowledged by references to the authors. Signature of Candidate: __________________ Date: _____2019-06-17______ AGREEMENT ON AUTHORIZED USE OF THESIS Nanjing University of Information Science & Technology, China National Library, Chinese Academic Journal (CD) Magazine, China Degree Thesis Full Text Database of ISTIC have the right to keep the copies and electronic version of degree thesis, use methods of photocopy, reduction printing, etc. to keep the thesis, and provide information service to society by internet. The content of the electronic version thesis is in conformity with that of the paper thesis. The thesis is available for reference and borrowing and its full text or partial text can be published except those confidential theses during confidential period. Authorized publication of the thesis should be conducted by Graduate School, Nanjing University of Information Science and Technology. 口 Public 口 Confidential ( Year Month ) (Confidential thesis should obey this agreement when confidential period is ended) Signature of Thesis Author: __________________ Date: _____2019-06-17______ Signature of Supervisor: _____________________ Date: _____2019-06-17______

DEDICATION To my lovely family; my father BOUNSANG, my mother CHANSAVANH, my brother THANONGSIN, my sister POUNA.

TABLE OF CONTENTS TABLE OF CONTENTS.........................................................................................................a LIST OF ACRONYMS ...........................................................................................................e LIST OF FIGURES.................................................................................................................g LIST OF TABLES...................................................................................................................k ABSTRACT............................................................................................................................ l CHAPTER 1 INTRODUCTION AND LITERATURE REVIEWS..........................................1 1.1 Research Significance.......................................................................................................1 1.2 Background ......................................................................................................................1 1.2.1 Study Area ....................................................................................................................2 1.2.2 Problem Statement ........................................................................................................5 1.2.3 Objective of Study ........................................................................................................5 1.3 Literature Reviews ...........................................................................................................6 1.3.1 Precipitation Systems over ICP .....................................................................................6 1.3.2 South Asian Monsoon ...................................................................................................6 1.3.3 Monsoon Onset .............................................................................................................7 1.3.4 Bay of Bengal Trough or Monsoon Trough ...................................................................8 1.3.5 Influence of Tibetan Plateau on Monsoon Trough..........................................................8 1.3.6 Inter-tropical Convergence Zone ...................................................................................9 1.3.7 Sea-Surface Temperature...............................................................................................9 CHAPTER 2 DATA AND METHODOLOGY.......................................................................11 2.1 Data................................................................................................................................11 2.1.1 Tropical Rainfall Measurement Missions (TRMM) .....................................................11 a

2.1.2 NCEP-NCAR Reanalysis Dataset ...............................................................................11 2.1.3 Era-Interim Dataset .....................................................................................................12 2.1.4 Hadley Center Observations Dataset ...........................................................................12 2.2 Methodology ..................................................................................................................12 2.2.1 Monsoon Onset ...........................................................................................................12 2.2.2 Normalized precipitation water index ..........................................................................13 2.2.3 Climatological pentad means rainfall index .................................................................13 2.2.4 Webster and Yang circulation index.............................................................................14 2.2.5 Power Spectrum Analysis............................................................................................14 2.2.6 Empirical Orthogonal Function Analysis .....................................................................14 2.2.7 Composite analysis .....................................................................................................16 2.2.8 Correlation analysis.....................................................................................................17 2.2.9 Moisture Flux Convergence ........................................................................................17 CHAPTER 3 THE CLIMATE FEATURES ...........................................................................19 3.1 Climatic Features Over Study Domain............................................................................19 3.1.1 Rainfall Feature over ICP............................................................................................19 3.1.2 Wind Circulation.........................................................................................................21 3.1.3 Temperature ................................................................................................................23 3.1.4 Moisture Flux Convergence ........................................................................................25 3.2 Monsoon Onset and Its Index .........................................................................................26 3.2.1 Monsoon onset index ..................................................................................................26 3.2.2 Rainfall period after monsoon onset ............................................................................28 3.3 Climatic Features after Monsoon Onset ..........................................................................29 3.3.1 Rainfall distribution after monsoon onset ....................................................................29 3.3.2 Wind circulation after monsoon onset .........................................................................30 b

3.3.3 Air temperature at 1000 hPa after monsoon onset........................................................33 3.4 Chapter Summary...........................................................................................................35 CHAPTER 4 BAY OF BENGAL TROUGH AND ICP RAINFALL MODES........................37 4.1 EOF Modes of Vorticity and Rainfall after Monsoon Onset ............................................37 4.1.1 EOF of Vorticity at 500 hPa over the BoB ...................................................................37 4.1.2 EOF of Rainfall over ICP............................................................................................41 4.2 Bay of Bengal Trough in modes......................................................................................45 4.2.1 Based on Vorticity at 500 hPa over BoB EOF..............................................................45 4.2.2 Based on Rainfall over ICP EOF .................................................................................47 4.3 Structures of Bay of Bengal Trough................................................................................49 4.3.1 Based on Vorticity at 500 hPa over BoB EOF..............................................................49 4.3.2 Based on Rainfall over ICP EOF .................................................................................52 4.4 Chapter Summary...........................................................................................................54 CHAPTER 5 THE RELATIONSHIP OF THE BAY OF BENGAL TROUGH AND ICP RAINFALL ...........................................................................................................................56 5.1 Influencing factors under BoB vorticity modes ...............................................................56 5.1.1 SST and SST anomaly ......................................................................................................56 5.1.2 Temperature and temperature anomaly .............................................................................58 5.1.3 Geopotential height anomaly ............................................................................................61 5.1.4 Wind circulation anomaly and wind divergence ................................................................63 5.2 Influencing factors under ICP rainfall modes ..................................................................66 5.2.1 SST and SST anomaly ......................................................................................................66 5.2.2 Temperature and temperature anomaly .............................................................................67 5.2.3 Geopotential height anomaly ............................................................................................70 5.1.4 Wind circulation anomaly, omega, MFC and wind divergence ..........................................71 c

5.3 Chapter Summary...........................................................................................................76 CHAPTER 6 CONCLUSION ...............................................................................................78 6.1 Summary........................................................................................................................78 6.2 Conclusion .....................................................................................................................82 6.3 Theoretical and Practical Implications of Research .........................................................83 6.4 Original Contributions to Knowledge .............................................................................83 6.5 Suggestion for Future Research ......................................................................................84 ACKNOWLEDGEMENTS...................................................................................................85 REFERENCES .....................................................................................................................87 d

AIRI LIST OF ACRONYMS ARS AS All Indian Rainfall Index ASM Arabian Sea BoB Andaman Sea CPMI Asian Summer Monsoon EOF Bey of Bengal HadISST Climatological Pentad Mean Rainfall Index ICOADS Empirical Orthogonal Function ICP Hadley Center Sea Ice and Sea Surface Temperature Data Set IO International Comprehensive Ocean-Atmosphere Data Set ISM Indochina Peninsula ITCZ Indian Ocean JAXA Indian Summer Monsoon LLJ Inter-Tropical Convergence Zone MCAs Japan Aerospace Exploration Agency MFC Low-Level Jet MJJASO Maximum Co-variance Analysis MJO Moisture Flux Convergence MT May to October NASA Median-Julian Oscillation Monsoon Trough National Astronautics and Space Administration e

NCAR National Center for Atmospheric Research NCEP National Center for Environmental Prediction NEM Northeast Monsoon NPWI Normalized Precipitation Water Index PCA Principal Component Analysis Pentad Five Day Mean or Five Days Average of Data Set PO Pacific Ocean PSA Power Spectrum Analysis PW Precipitation Water RMS Root-Mean-Square SAM South Asian Monsoon Sd Standard Deviation SCS South China Sea SST Sea Surface Temperature SWM Southwest Monsoon TIO Tropical Indian Ocean TP Tibetan Plateau TPO Tropical Pacific Ocean TRMM Tropical Rainfall Measurement Miss WY Webster-Yang f

LIST OF FIGURES Figure 1. 1 Map of the study area showing the location of ICP and BoB .....................................3 Figure 1. 2 The study domain (a) and topography of study area (b). Domain A is the BoB trough mean position domain (10°N-22.5°N,80°E-92.5°E), B is an area of moisture flux that flux in to ICP (10°N-20°N,87.5°E-92.5°E), and C is the area of study (10°N-25°N,95°E-110°E). ..............4 Figure 3. 1 Monthly precipitation (mm) averaged over Laos (green solid line) and Indochina Peninsula (red solid line), during 1998-2017, where the black solid line is the demarcation for the wet season in the region. ...........................................................................................................20 Figure 3. 2 Spatial distribution of monthly mean precipitation (mm) over study domain during 1998-2017. ................................................................................................................................21 Figure 3. 3 Monthly mean wind circulation (m s-1) at 850 hPa over the study domain during 1998-2017. ................................................................................................................................22 Figure 3. 4 Monthly mean temperature (℃) over ICP (red solid line) and Laos (black solid line) during 1998-2017......................................................................................................................23 Figure 3. 5 Monthly mean temperature (˚C) at 1000 hPa over ICP during 1998-2017. ..............24 Figure 3. 6 Monthly mean moisture flux convergence by column of air (kg kg-1 m2 day-1) over the study domain. Wind vector showed the moisture transport and shaded area indicated convergence (positive) and divergence (negative)......................................................................25 Figure 3. 7 Power spectrum analyses of rainfall of 10 pentad data after monsoon onset over ICP during 1998-2018 (significant bounds at 90% confident interval (thick dash line) and at 95% confident interval (light dash line) )...........................................................................................28 Figure 3. 8 Spatial distribution of pentad mean precipitation (mm) over ICP in ten different pentad data after monsoon onset................................................................................................29 Figure 3. 9 Spatial distribution of pentad mean precipitation (mm) over ICP before and after monsoon onset. .........................................................................................................................30 Figure 3. 10 Mean wind circulation (m s-1) over ICP after monsoon onset based on 10 pentad reconstructed dataset. ................................................................................................................31 g

Figure 3. 11 Spatial distribution of pentad mean wind circulation (m s-1) over study domain before and after monsoon onset. ................................................................................................32 Figure 3. 12 Mean surface air temperature (˚C) over ICP at ten different pentads after monsoon onset. ........................................................................................................................................34 Figure 4. 1 Four EOF vector(a) first, (b) second, (c) third, (d) in the standardized anomaly field of the ten pentads reconstructed vorticity over BoB during 1998-2018. .....................................38 Figure 4. 2 Four PC: (a) PC1, (b) PC2, (c) PC3, (d) PC4 in the standardized time coefficient of the ten pentads reconstructed vorticity over BoB during 1998-2018. .........................................39 Figure 4. 3 Precipitation distribution over study domain for composite pentad from first EOF, (a); and second EOF (b). ...........................................................................................................40 Figure 4. 4 Fourth EOF vector (a) first, (b) second, (c) third, (d) in the standardized anomaly field of the ten pentads reconstructed precipitation over ICP during 1998-2018. ........................42 Figure 4. 5 Fourth PC: (a) PC1, (b) PC2, (c) PC3, (d) PC4 in the standardized time coefficient of the ten pentads reconstructed precipitation over ICP during 1998-2018. ................................43 Figure 4. 6 Precipitation distribution over study domain for composite pentad first EOF, (a); and second EOF (b) from ICP precipitation EOF analyses. ..............................................................44 Figure 4. 7 BoB trough features at different levels 850, 700 and 500 hPa based on EOF1 and EOF2 of vorticity EOF analysis.................................................................................................46 Figure 4. 8 BoB trough features at different levels 850, 700 and 500 hPa based on EOF1 and EOF2 of rainfall EOF analysis...................................................................................................48 Figure 4. 9 Vertical cross section of specific humidity (kg kg-1) along 90°E according to EOF1 (a) and EOF2 (b). ......................................................................................................................49 Figure 4. 10 Vertical cross section of specific humidity (kg kg-1) along 95°E according to EOF1 (a) and EOF2 (b). ......................................................................................................................50 Figure 4. 11 Vertical velocity (Pa s-1) along 90°E according to EOF1 (a) and EOF2 (b). ...........51 Figure 4. 12 Vertical cross-section of air temperature (°C) along 85 °E according to EOF1 (a) and EOF2 (b). ...........................................................................................................................51 Figure 4. 13 Vertical cross section of specific humidity (kg kg-1) along 90°E according to EOF1 (a) and EOF2 (b). ......................................................................................................................52 h

Figure 4. 14 Vertical cross section of specific humidity (kg kg-1) along 95°E according to EOF1 (a) and EOF2 (b). ......................................................................................................................53 Figure 4. 15 Vertical velocity (Pa s-1) along 90°E according to EOF1 (a) and EOF2 (b). ...........53 Figure 4. 16 Vertical cross-section of air temperature (°C) along 85 °E according to EOF1 (a) and EOF2 (b). ...........................................................................................................................54 Figure 5. 1 SST anomaly (˚C) for the composite pentad from (a) EOF1 and (b) EOF2 over northeast IO of the BoB vorticity at 500 hPa EOF analysis........................................................57 Figure 5. 2 spatial distribution of correlation coefficient between SST and ICP rainfall during EOF1 (a) and EOF2 (b). ............................................................................................................58 Figure 5. 3 Air temperature anomaly (˚C) for the composite pentad from (a) EOF1 and (b) EOF2 over BoB and ICP of the BoB vorticity at 500 hPa EOF analysis. ..............................................59 Figure 5. 4 spatial distribution of correlation coefficient between air temperature at 1000 hPa and ICP rainfall during EOF1 (a) and EOF2 (b).........................................................................60 Figure 5. 5 Sea surface temperature and air temperature at 1000 hPa (˚C) for the composite pentad from EOF1 and EOF2 over BoB, northeast IO and adjacent sea of the BoB vorticity at 500 hPa EOF analysis. ..............................................................................................................61 Figure 5. 6 Geopotential height (m2 s-2) for the composite pentad from EOF1 and EOF2 over BoB and ICP at three different levels of the BoB vorticity at 500 hPa EOF analysis. .................62 Figure 5. 7 Wind circulation anomaly (m s-1) for the composite pentad from EOF1 and EOF2 over BoB and ICP at three different levels of the BoB vorticity at 500 hPa EOF analysis. .........64 Figure 5. 8 Wind divergence (m s-1), shaded area indicated convergence (negative) and divergence (positive) for the composite pentad from EOF1 and EOF2 over BoB and ICP of the BoB vorticity at 500 hPa EOF analysis......................................................................................65 Figure 5. 9 SST anomaly (˚C) for the composite pentad from (a) EOF1 and (b) EOF2 over northeast IO of the ICP rainfall EOF analysis. ...........................................................................66 Figure 5. 10 spatial distribution of correlation coefficient between SST and ICP rainfall during EOF1 (a) and EOF2 (b). ............................................................................................................67 Figure 5. 11 Surface air temperature anomaly (˚C) for the composite pentad from (a) EOF1 and (b) EOF2 over study domain of the ICP rainfall EOF analysis. ..................................................68 i

Figure 5. 12 spatial distribution of correlation coefficient between air temperature at 1000 hPa and ICP rainfall during EOF1 (a) and EOF2 (b).........................................................................69 Figure 5. 13 Geopotential height (m2 s-2) for the composite pentad from EOF1 and EOF2 over BoB and ICP at three different levels of the ICP rainfall EOF analysis. .....................................70 Figure 5. 14 Wind circulation anomaly (m s-1) for the composite pentad from EOF1 and EOF2 over BoB and ICP at three different levels of the ICP rainfall EOF analysis...............................72 Figure 5. 15 Moisture flux convergence (kg kg-1 m2 day-1) shaded area indicated convergence (positive) and divergence (negative) for the composite pentad from EOF1 (a) and EOF2 (b). ....74 Figure 5. 16 Vertical velocity (Pa s-1) along 16°N for the composite pentad from EOF1 (a) and EOF2 (b)...................................................................................................................................75 Figure 5. 17 Wind divergence (m s-1), shaded area indicated convergence (negative) and divergence (positive) for the composite pentad from EOF1 and EOF2 over BoB and ICP of the ICP rainfall EOF analysis. .........................................................................................................75 j

LIST OF TABLES Table 3. 1 The monsoon onset which defined by normalized precipitation index (NPWI) column 2, Webster and Yang circulation index (WYI) column 3 and climatological pentad mean rainfall index (CPMI) column 4.............................................................................................................27 Table 4. 1 Strong and weak pentad obtained from standardized time coefficient of the four EOF modes of BoB vorticity at 500 hPa EOF analysis.......................................................................39 Table 4. 2 Strong and weak pentad obtained from standardized time coefficient of the four EOF modes of ICP rainfall EOF analysis. ..........................................................................................43 Table 5. 1 The critical values for correlation test base on 2 degree of freedom and two-tailed test. .................................................................................................................................................57 k

ABSTRACT The relationship between the trough modes of the Bay of Bengal and rainfall over Indochina Peninsula after South Asia Monsoon onset were studied. The climatological monsoon onset over ICP is on 28 or 29 April, with standard deviation of 10 days. Monsoon onset is characterized by abruptly increased in amount of rainfall coincides with pronounced northeastward progression of low-level southwesterly wind over IO and also BoB. The close relationship between the BoB trough and ICP rainfall was identified. The BoB trough hold a critical important relationship with precipitation over ICP. Physical structures and dynamical effects from the BoB trough have played significant role in ICP rainfall. On the other word, different modes of the BoB trough, shape-size and position, caused ICP to experience different rainfall patterns. The prolong south and depth extend low-levels BoB trough inducing heavy rainfall over west of ICP, on the other hand, weak, shallow and retreated north BoB trough, less rainfall overall the study area. East and southeast extend of BoB trough inducing heavy rainfall over east of ICP. The cyclonic circulation has also important roles in precipitation pattern over the area. It increased the surface convergence and maintained strong rising vertical motion over the area. Moreover, it intensified the southwesterly air flow and transfer more vapor and moisture toward the precipitation area. The SST and Air temperature from the adjacent sea provided evidence on lack of direct influence to the ICP rainfall. Nevertheless, there is a strongly relationship between ICP rainfall and TIO SST. Key words: Bay of Bengal (BoB) trough, Indochina Peninsula (ICP), Monsoon onset, Rainfall, EOF modes. l

CHAPTER 1 INTRODUCTION AND LITERATURE REVIEWS 1.1 Research Significance The climate variability over Indochina Peninsula (ICP) is critical important for people living in the region. ICP has always experienced of flood and drought, basically associated with the rainfall variability. According to Asian Development Bank report (2009), the socioeconomic conditions of South East Asia countries are highly dependent on agricultural product, which water is crucial important, for instance, Thailand, Vietnam, and Myanmar are among the top rice producing countries. Understanding of factors associated with regional rainfall is very useful for better prediction and more informed decisions in rainfall related disaster mitigation. The diagnosis of influence of monsoon trough over the Bay of Bengal (hereafter BoB trough) on the rainfall over ICP and the investigation of the physical structures and dynamical effects by which they influence in rainfall patterns, in terms of causative mechanisms, could be critical to the detection of key systems that associated with the rainfall over ICP. Because of insufficient statistical links between BoB trough and rainfall over ICP, this study would contribute to some knowledge that can be helpful in rainfall prediction. The findings of this study can also help local governments to enhance the monitoring and prediction skill of rainfall variability and thus can be applied in short to long-term adaptation strategies to mitigate losses due to rainfall related disasters. 1.2 Background The rainy season over ICP, in general, starts from middle May and withdraws by middle October. The geographical position makes ICP experience the long-lasting rainfall nearly 6 months each year. ICP is a unique region where the monsoon activity reflects a transitional feature of two distinct monsoon subsystems: South Asia Monsoon and East Asia Monsoon (Melorose, Perroy, & Careas, 2015)⁠. In other words, the rainfall in ICP is affected by two adjacent Pacific and Indian 1

Oceans, lying in the east and southwest respectively. The contrasting topography of the region also contributes to complex dynamic effects and determines the type of atmospheric circulation. The ICP rainfall has been in concern of scientists for decades. Rainy season over ICP starts as early as late April to early May, which was defined by pre-monsoon rain (Matsumoto, 1997)⁠ and the full establishment of the monsoon circulation is in mid-May, coexistence with active convective activity over the region. Rainfall increases gradually from the pre-monsoon rains to the summer after monsoon onset. According to the precious studies, on average, the monsoon onset first occurs in the eastern BoB and ICP in the second to third pentad of May, which is followed by the monsoon onset over the South China Sea in late May and Indian subcontinent monsoon onset in early June (Zhang, Li, Wang, & Wu, 2002).⁠ Recently, Nguyen-Le (2015) investigated the rainy season over eastern of ICP by determining the individual year data from 1958 to 2007 and found that, on average, the onset of summer rainy season was 6 May (Nguyen-Le, Matsumoto, & Ngo-Duc, 2015)⁠ with standard deviation of 13 days. On the other hand, variations of BoB trough during the winter induced increasing precipitation over southern periphery of Tibetan Plateau and southeastern China (T. Wang, Yang, Wen, Wu, & Zhao, 2011)⁠. However, a new dynamical index for the BoB trough (Liu, Zha, Yang, & Chen, 2018)⁠ is suggested that BoB trough had a relationship with upstream teleconnection. Although past studies have documented the variability of the rainfall related to the monsoon onset and other factors that has critical influence on the ICP rainfall, up to date limited studies have systematically and comprehensively examined the climatological aspect of the influence of different factors on the rainfall over ICP. As such, more research is desired (Barker, 2007)⁠. The global warming is strongly evidence and climate change, anomalies in rainfall and its influencing factors are intensifying. One of the greatest challenges is to determine the relationship between the factors that modulates the rainfall variability. 1.2.1 Study Area The ICP is a long narrow land located in between India and China, and was divided into two-part mainland and maritime continents. In this study we are focusing on mainland, lies between 95°E- 2

110°E and 10°N-25°N. It includes the country of Thailand, Vietnam, Laos, Cambodia, and eastern part of Myanmar (Figure 1.1). Figure 1. 1 Map of the study area showing the location of ICP and BoB. In Figure 1.2 (a) displayed the domains, which was defined based on the mean position of the BoB trough (A) lies between 10°N-22.5°N and 80°E-92.5°E, moisture transporting window (B) lies between 10°N-20°N and 87.5°E-92.5°E and ICP (C) lies between 10°N-25°N and 95°E-110°E. The topography of ICP is shown in Fig1.2(b). The topography of ICP projects southward from the Asia continent proper. It contains several mountain ranges extending from the Tibetan Plateau (TP) in the north, interspersed with lowlands largely drained by three major river systems running in the north-south direction: the Irrawaddy (Myanmar), the Chao Phraya (Thailand), and the Mekong (flowing through south China, Laos, Myanmar, Thailand, Cambodia, and Vietnam). It is bounded by the Indian Ocean (IO) to the west and southwest and by Pacific Ocean (PO) to the east. North center part of ICP has a higher elevation and slopes to the lower elevations toward south, east and west. TP to the north play an important role as a natural barrier against westerly wind, which flows around south and north periphery of TP. Because of the topography of the region, moisture advection toward ICP is contributed more by the IO during the monsoon season. 3

Figure 1. 2 The study domain (a) and topography of study area (b). Domain A is the BoB trough mean position domain (10°N-22.5°N,80°E-92.5°E), B is an area of moisture flux that flux in to ICP (10°N-20°N,87.5°E-92.5°E), and C is the area of study (10°N-25°N,95°E-110°E). The monsoon wind during the southwest monsoon month approach the west coast of India as southwesterly flow and are deflected by the Burmese mountains toward the Himalayan foothills (Goel & Srivastava, 1990) so that the flow over the plains of northern India is from east to west. This results in the formation of a quasi-stationary low-pressure area over the plains north and northeastern India, sometimes can be extended southeastward to the BoB or Myanmar, which is called BoB Trough. The BoB trough is a key semi-permanent feature of the southwest monsoon (Sikka & Narasimha, 1995), because the atmospheric boundary layer along the trough, in the light of the earlier studies, has played a significant effect on rainfall in adjacent area. The BoB trough has provided a rather special meteorological environment within the tropical region, wherein for the nearly 6 months period from May to October every year organized moist convection prevails over the subcontinental scale. The linkage between the monsoon trough and the equatorial Indian Ocean is found to be the main factor modulating the intensification of the rainfall along the trough during the summer monsoon (Annamalai, 2010). However, (Wang et al., 2011) suggested that the variations of the winter BoB trough played an important role in affecting the weather and climate over southern and eastern Asia. 4

1.2.2 Problem Statement The rainfall over the Indochina Peninsula (ICP) is of great concern not only for meteorologists that are interested in the phenomenon for perhaps purely scientific reasons, but there also lies a great deal of applications for national governments in protecting their citizens from natural disasters, naturally the citizens themselves and industries within the region. According to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change of 2007 (Harasawa et al., 2007),⁠ most of the region is expected to experience an overall increase of rainfall and more extreme downpours, resulting in more frequency floods, and an increase of strong winds caused by tropical cyclones. The most recent and striking example of extreme flooding occurred on 23 July 2018 in southern part of Laos, the continued rainfall for at least three consequent weeks cause an under- constructing dam collapsed, leading to widespread destruction and homelessness among the local population and more than 100 people was confirmed dead. Gaining a deeper understanding of past events will allow us to investigate the direct and indirect influences on rainfall variability throughout the region. Communicating this understanding with the scientific community, local governments and citizenry is crucial in coping with natural disasters that grow in severity with each passing decade. 1.2.3 Objective of Study The specific objectives of this study are: 1) To investigate the climatological feature of rainfall over ICP. 2) To analyze the key systems, especially the BoB trough structure, evolution and relationship with the precipitation over ICP after monsoon onset, mainly focus on the fifty days period. 3) To examine the pattern of BoB trough after monsoon onset and how it impacts on the precipitation over ICP. 4) To investigate the influence of other environment factors such as: sea-surface temperature, air temperature, and topographical effect. 5

1.3 Literature Reviews Several studies have been done over ICP in order to investigate the monsoon onset (Goel & Srivastava, 1990; Matsumoto, 1997; Potty, Mohanty, & Raman, 1997; Rajeevan, Gadgil, & Bhate, 2010),⁠ severity and variability of rainfall. The results suggested that different atmospheric and oceanic factors are linked to variation of rainfall over ICP. Such factors include: rainfall variation, south Asian monsoon, monsoon onset, monsoon trough, Tibetan Plateau, inter-tropical convergence zone, and sea-surface temperature. These factors are associated with anomalous rainfall resulting in flood and drought events over the region. 1.3.1 Precipitation Systems over ICP It is well-known that after the monsoon onset the abundant rainfall occurred. Nevertheless, (Harasawa et al., 2007),⁠ suggested that most of the region along the monsoon climate system is expected to experience an overall increase of rainfall and more extreme downpours, resulting in more frequent floods, and an increase of strong winds caused by tropical cyclones. During the previous decades numerous studies have been paid attention on rainfall variation. Using 210 rain gauge data observed in ICP (Yokoi, Satomura, & Matsumoto, 2007), suggested that the climatological characteristics of rainy season precipitation are 30-60 day and 10-20 day modes. However, these modes vary over the spatial area. The inter-annual variation of ICP rainfall is caused by the east-west inter-annual seesaw of the global divergent water vapor flux induced by the inter-annual variation in global divergent circulation (Chen & Yoon, 2000)⁠. 1.3.2 South Asian Monsoon The South Asian Monsoon system is the most developed among all other regional monsoon systems in the global tropic as it develops in response to the global-scale diabatic heating producing extremely large temperature gradients (between the equator to 30°N) in entire troposphere which are enhanced by heating of the elevated TP situated in the northern boundary. The plateau has the highest temperatures in the middle to the upper troposphere during the monsoon season. Wind in the low levels during summer monsoon season are characterized by the 6

strongest westerlies anywhere at 850 hPa over the Arabian Sea, known as the low-level westerly jet (LLJ) and a large-scale cyclonic vorticity extending from the north Bay of Bengal (BoB) to the western India known as monsoon trough (Rao, 1976). The easterly jet is around 5°N and the Tibetan anticyclone is around 30°N are important features of upper level winds over the monsoon region during northern summer. The monsoon, or the seasonal changes of winds and rainfall, in the region could be interpreted as a result of northward migration of the east-west oriented precipitation belt (ITCZ) from the southern hemisphere in the winter to the northern hemisphere in summer (Annamalai, 1999). Stronger (weaker) than average summer trade winds (Peter J. Webster & Yang, 1992)⁠ were associated with strong (weak) monsoon periods. 1.3.3 Monsoon Onset The onset of the monsoon is the key indicator characterizing the abrupt transition from the dry season to the rainy season and subsequent seasonal march. Numerous investigators have studied this problem from the regional perspective. It is to some extent difficult to obtain a unified and consistent picture of the climatological onset dates of the SAM due to differences in data, monsoon index, and the definition of the monsoon onset are used in this investigation. The onset of SAM is characterized by pronounced northeastward progression of the low-level south westerlies over the Indian Ocean and pole-ward extension of the intensified tropical convection. According to (Matsumoto, 1997)⁠, the initial onset stage of the SAM is characterized by the establishment of strong convection and abrupt inverse direction of the prevailing wind over the Bay of Bengal, the ICP and SCS in early and middle May. The temporal and spatial structure of the atmospheric circulation associated with the climatology and inter-annual variations of the summer monsoon were investigated in the period 1951-1996 (Zhang, Li, Wang, & Wu, 2002),⁠ and the climatological monsoon onset is on the second pentad of May, with standard deviation of 12 days. The study of (Nguyen-Le et al., 2015),⁠ has suggested that the most essential role in triggering the rainy season onset over ICP is tropical influence. Climatological summer monsoon rainfall outbreak is an abrupt northward extension of intense tropical convection and arrival of the southwesterly monsoon from the equatorial Indian Ocean. 7

The variation of the SAM is one of the strongest signals of the earth’s climate variability. In order to quantify the variability of the SAM, it is appealing to use representative variables as an objective measure, if such variables exist. The All Indian Summer Rainfall Index (AIRI) has long been used as a measure of the Indian summer monsoon (ISM), which is defined by the seasonally averaged precipitation over all the Indian subdivisions (Parthasarathy et al. 1992). AIRI has been widely used in the studies of the ISM. While AIRI is a good indicator of the strength of the monsoon rain over India, it is not clear how well it represents the summer monsoon in ICP region. 1.3.4 Bay of Bengal Trough or Monsoon Trough The monsoon trough (MT) is the most dominant feature of the monsoon season, situated along the Indo-Pakistan plains. Its eastern end is locked with the warm water of the northern BoB with dominantly moist processes operating on a day-to-day basis showing diurnal variability also. The western end is situated in the predominantly dry convective area of the western India and Pakistan and occasionally influenced by the moist processes associated with eastward moving lows/monsoon depressions/quasi-stationary subtropical cyclones along the Gujarat coast and southward extending western disturbances. The presence of this semi-permanent feature during the monsoon season (June to September) is considered as equatorial trough of the summer period. The study of (Potty, Mohanty, & Raman, 1997)⁠ shows that this trough can extend aloft up to about 4 km and slopes southward with height. The position of the trough changes day to day and has an important bearing on the SAM rainfall over the region. In the studying of the intensity of the trough over the Bay of Bengal and its association with the southern China precipitation were investigated on seasonal timescale (Hai-feng, Cholaw, Jie, & Lie-ting, 2012)⁠, and shown that southern China was affected more directly by the intensity of BoB trough to a greater degree than the Madden- Julian Oscillation (MJO). The latest study (Liu et al., 2018)⁠ based on the vertical velocity field of reanalysis datasets, has revealed the relationship between monsoon trough and upstream system. 1.3.5 Influence of Tibetan Plateau on Monsoon Trough The Asian summer monsoon (ASM) is the strongest element of the global monsoon system. In addition to land-sea contrast, it is affected by large-scale mountain ranges such as the Tibetan 8

Plateau (TP). (Yanai & Wu, 2006)⁠ gave a thorough review of the past studies about the effects of the TP. The review starts from the research in 1950s on the jet stream and the warm South Asian high, and the early progress of TP research. The review also cover the importance of thermal influences of the TP on seasonal transition and Asian summer monsoon onset based on different datasets (Wu & Zhang, 1998)⁠. The previous Atmospheric General Circulation Models studies suggested that the TP could produce downstream convergence in boreal spring, though the underlying mechanisms were not fully explored. (Wu et al., 2007)⁠ demonstrated that the TP altered the prevailing westerly wind in the spring, mechanically inducing convergence downstream of the plateau, and substantially increasing rainfall over northeastern ICP and east China. 1.3.6 Inter-tropical Convergence Zone The ITCZ is a low-pressure region that surrounds the Earth in the low latitudes, where the north east and southeast trade winds converge. In the northern summer, the surface wind field is dramatically different. The strong southeast trades from the Southern Hemisphere cross the equator and turn into a southwest current which is, therefore, called the Southwest Monsoon or Summer Monsoon. The southern boundary of the southwest monsoon is located near the equator. The SWM meets the trade wind system and forms the ITCZ. In this region, unstable air at the surface rises and the water vapor it holds condenses resulting in rainfall near the equatorial region. The position of the ITCZ varies throughout the year. (Goswami, Krishnamurthy, & Annamalai, 1999)⁠ said that, the most striking feature of northern hemisphere summer climate is the large shift of the ITCZ. While the mean precipitation zone associated with the ITCZ is zonal in winter, and fluctuated about 20°N in the summer over northern Indian Ocean. The position of the ITCZ there has a seasonal shift of about 20° latitude (Fu et al., 1983).⁠ The monthly total rainfall in these regions normally peak in July and August, which is partly related to the activity of the ITCZ (Nguyen, Katzfey, & McGregor, 2014). 1.3.7 Sea-Surface Temperature It is well-known that one of the starting mechanisms of the summer monsoon is the thermal contrast between land and ocean and that sea surface temperature (SST) and moisture are crucial 9

factors for its evolution and intensity. The Indian Ocean, therefore, may play a very important role in generation and evolution of the SAM. (Cherchi et al., 2007),⁠ has investigated the relationship between the Indian Summer Monsoon and tropical Indian Ocean (TIO) SST anomalies, as well as the ability of the coupled model to capture those connections. A new statistical technique is used to investigate the TIO SST variability and its relation with the tropical Pacific Ocean (TPO) and found that, the SST variability in the TIO contains a significant portion that is independence from the TPO. According to the studies of (Yang, Liu, Xie, Liu, & Wu, 2007)⁠ suggested that, IO warming induces robust climatic anomalies in the rainy season over ICP region. In response to the IO warming, precipitation increase over most of the basin, forcing a Matsuno-Gill pattern in the upper troposphere with a strengthened South Asian high. The relationship among the SST and TIO and the seasonal atmospheric circulation in the Asian Monsoon Region is examined by using the maximum covariance analyses (MCAs) (Yang, Liu, & Liu, 2010)⁠, the result shown that the Asian monsoon circulation is significantly correlated with two dominant SST anomaly modes, for instance, the Indian Ocean Basin mode and the Indian Ocean Dipole modes. 10

CHAPTER 2 DATA AND METHODOLOGY This chapter describes the data and methods that were employed in this study in order to achieve the objectives. 2.1 Data 2.1.1 Tropical Rainfall Measurement Missions (TRMM) The daily and monthly gridded precipitation total data (3B42 and 3B43) obtained at 0.25-degree horizontal grid resolution from the Tropical Rainfall Measurement Missions (TRMM) version 7 has a time span of 1998 to 2018, was used in this study. The TRMM is an international project of NASA and JAXA designed to provide improved estimates of precipitation in the Tropics, where the bulk of the Earth's rainfall occurs. TRMM began recording data in December 1997 and ended in April 2015. It flew in a (46-day) processing orbit at a 35° inclination with a period of about 91.5 min. This orbit allows TRMM to build up a complete view of the climatological diurnal cycle, as well as providing calibration for other precipitation-relevant sensors in Sun-synchronous orbits. The TRMM home page is located at http://trmm.gsfc.nasa.gov/. And ICP is located in tropical area so the TRMM is well described. 2.1.2 NCEP-NCAR Reanalysis Dataset The principal dataset used in this study is the daily mean global atmospheric data provided by National Centers for Environmental Prediction-National Centers for Atmospheric Research reanalysis. The zonal and meridional wind, geopotential height, air temperature, vertical velocity (omega), and specific humidity data have a horizontal resolution of 2.5-degree latitude and longitude at 17 standard pressure levels. All the climatological pentad means in this study are constructed from daily means for the period Jan-1998 – Jul-2018. 11

2.1.3 Era-Interim Dataset ERA-Interim is a global atmospheric reanalysis (Dee et al., 2011)⁠ which provides global atmospheric and surface parameters from 1979 to present, on 60 vertical levels (Dee et al. 2011). In this study, daily mean gridded data at 2.50° x 2.50° spatial resolution surface variables include total precipitation, divergence of moisture flux and SST were used and 2.50° x 2.50° spatial resolution pressure level variables starting from 1000 hPa to 200 hPa include zonal wind, meridional wind, geopotential height, air temperature, specific humidity, relative humidity, potential vorticity, relative vorticity and vertical velocity were used. 2.1.4 Hadley Center Observations Dataset The Hadley Centre Global Sea Ice and Sea Surface Temperature (HadISST) is a combination of monthly globally complete fields of SST for 1871-present. HadISST uses reduced space optimal interpolation applied to SSTs from the Marine Data Bank (mainly ship tracks) and ICOADS through 1981 and a blend of in-situ and adjusted satellite-derived SSTs for 1982-onwards. The \"bucket correction\" was applied to SSTs for 1871-1941. SSTs in boxes partially covered by sea ice were estimated from statistical relationships between SSTs and sea ice concentrations. SSTs were assigned a fixed value (-1.8°C) for areas with sea ice cover of greater than 90%. Version 1.1 is the current version. 2.2 Methodology Different analysis techniques were applied in the current study. This section gives specifics of methods that used. 2.2.1 Monsoon Onset The onset of the summer monsoon is generally concurrent with a reversal in the wind field and abrupt rise in precipitation and water vapor. According to the previous studies, different criteria have been used to define the monsoon onset (Matsumoto, 1997; Pai & Nair, 2009; Qian & Yang, 2000; B. Wang, Ding, & Joseph, 2009; Peter J. Webster & Yang, 1992; Zeng & Lu, 2004)⁠. In order 12

to define the starting time of the rainy season, many scientists had invented an index called monsoon onset index based on the climatological and atmospheric condition of the area of interest, such as all-Indian rainfall, normalized precipitation, climatological pentad mean rainfall and circulation index (Zeng & Lu, 2004)⁠. To quantify the relationship between the trough of the BoB and rainfall over ICP after monsoon onset using a single index is difficult. It requires an understanding of essential physics governing the BoB trough and rainfall variability over ICP. In this study we are going to apply the idea from normalized precipitation, climatological pentad means rainfall and Webster-Yang circulation index to compare and make onset index for this region. 2.2.2 Normalized precipitation water index The normalized precipitation water index is defined by: ������������������������ = ������������−������������������������������ (2.1) ������������������������������ −������������������������������ Where PW is the daily precipitable water at each grid, and PWmax and PWmin are the average of annual maximum and minimum daily PW at each grid, respectively. Zeng and Lu (2004) suggested the criterion to defined the global unified summer monsoon onset (retreat) date by the first day (d) or (d±1) when NWPI is greater (less) than the Golden Ratio (0.618) for the three consecutive days (Zeng & Lu, 2004).⁠ 2.2.3 Climatological pentad means rainfall index The climatological pentad means (5-day) rainfall index (CPMI) (B. Wang & Ho, 2001)⁠ is defined by constructed the data in a different year, which equal to 73 pentads each year. A decision has to be made as to how many harmonics should be used to retain both seasonal and sub-seasonal signals. In this study, we utilized the first 12 Fourier harmonics into consideration in order to avoid the noise and when the rainfall rate exceeds 6 mm per day the monsoon onset is defined. 13

2.2.4 Webster and Yang circulation index The Webster and Yang (1992) used a circulation index that is defined by a time-mean zonal wind (U) shear between 850 and 200 hPa, averaged over south Asia from the equator to 20°N and from 40°E to 110°E. And the positive value of U is westerly wind, which defined as monsoon onset (Peter J. Webster & Yang, 1992).⁠ ������������������ = ������850 − ������200 (2.3) 2.2.5 Power Spectrum Analysis Power Spectrum Analysis (PSA) is a type of frequency-domain analysis in which a structure is subjected to the probabilistic range of harmonic loading to obtain the probabilistic distributions for dynamic response measures. A root-mean-square (RMS) method converts the PSA curve for each response quantity into a single, most likely value. Because PSA curves represent the continuous probability density function of each response measure, most of the integrated area will occur near the resonant frequency of the structure. PSA was used in this study to obtain the period of the oscillation of dominant vorticity modes. The number of pentad period T is obtained from the time coefficient and f is a frequency using Equation 2.1; ������ = (1) (2.4) ������ 2.2.6 Empirical Orthogonal Function Analysis The Empirical Orthogonal Function (EOF) technique intend to find a new set of variables that capture most of the observed variance from a data through a linear combination of the orthogonal variable. The EOF is a widely used statistical method to minimize the multi-dimensional of complex climate data and identify the most important physical modes with a chance of minimum information to be lost (Habbachi et al., 2007). Therefore, it is a successful method to draw attention to a physical mechanism that can potentially contribute to climate variability (Mark et al., 2009). This technique, also called Principal Component Analysis (PCA), became popular for analysis of 14

atmospheric data following by Lorenz (1956) called the technique EOF analysis (Wilks S.D, 2006). The technique explains the variance-covariance of the data through a few modes of variability. The modes that account for the large percent of the original variability are considered significant. These modes can be represented by orthogonal spatial patterns (Eigenvectors) and corresponding time series (principal components). Two modes are spatially and temporally uncorrelated due to the orthogonal nature of the EOF. The orthogonal function of EOF is defined as follows: ������(������, ������, ������) = ∑������������=1 ������������ (������) × ������������������(������, ������) (2.5) Where ������(������, ������, ������)denotes the function of space (������, ������)and time (������), therefore, ������������������(������, ������)represents the spatial structure in relation to temporal variation of������. North et al., (1982) proposed that one should truncate EOFs only when the confident intervals for the corresponding eigenvalues do not overlap. North et al’s rule of thumb shows the 95% confidence intervals indicated as error bars. It would imply that could truncate between 1 and 2, or 2 and 3, but not anywhere else. It is more importance when discussing “physical important EOFs” rather than choosing truncation point. A collection of EOFs presents a convenient orthogonal basis set that is more tailored to the data than those that arise from classical differential equations. On the other hand, if want to interpret an individual EOF in physical terms, then is much more important to be certain that the EOF is not sensitively dependent on the sample. EOFs tend to give biased estimates of variance. For instance, the leading EOF is specially designed to explain maximal variance in the available sample. However, the detailed weighting required to achieve this variance is peculiar to sample and will not consistently hold in independent datasets. Consequently, the leading EOF generally will account for less variance in an independent dataset than in the sample from which it was derived. This reasoning suggests that if bias in variance is an important issue in particular analysis, then EOFs should be used with care. EOF analysis was used in this study to investigate the dominant modes of variability of vorticity. The data used was normalized to prevent area of maximum variance from dominating the eigenvectors (Walsh et al., 1980). The standardized anomaly������is computed as expressed in equation 2.6; ������ = ������−������ (2.6) ������������ 15

where ������ is observed, ������ is the long-term mean and ������������ is the standard deviation of vorticity, respectively. The value of ������provides immediate information about the significance of a particular deviation from the mean (Kabanda and Jury, 1999). 2.2.7 Composite analysis Composite analysis involves identifying and averaging one or more categories of fields of variable selected according to their association with key conditions. The key conditions in this study are the strong vorticity years as obtained from the EOF analysis time coefficient. Results of composites are then used to propose hypotheses for patterns which may be associated with individual cases (Folland 1983). This method was mainly be used to detect the anomaly associated with extreme events over the region. Composite for precipitation, was analyzed based on the observed vorticity extreme years obtained from the EOF analysis time coefficient, to assess the patterns of various systems linked to these anomalous years. In order to determine the significance of vorticity occurrence over the region, the composite analysis of various parameters was tested for statistical significance using t-test. This is based on the null hypothesis that for a given distribution of sample data, the sample mean is large enough for is sampling to be Gaussian distribution then the test statistic follows a distribution known as student’s t-test, given in Equation 2.7; ������ = ������−������0 (2.7) [������������������(������)]12 where, the denominator gives the variance of the sample mean and������0is a previously specified mean. The calculated values of������can then be compared with those of the theoretical t-distribution with N- 2 degree of freedom (N is sample size). If the calculated values of������is less than the theoretical value, then the significant area can be identified. The t-test in this study is at 90% confidence level. 16

2.2.8 Correlation analysis Correlation coefficients (r) are statistics that quantify the relationship between X an Y in unit- free terms. In order to calculate a correlation coefficient, we normally need three different sums of squares for variable X, Y and XY. The sums of squares for X variable is: ������������������������ = ∑(������������ − ������̅)2 (2.8) The sums of squares for Y variable is: ������������������������ = ∑(������������ − ���̅���)2 (2.9) The sums of squares for XY variable is: ������������������������ = ∑(������������ − ������̅)(������������ − ���̅���) (2.10) The correlation coefficient (r) is: ������ = ������������������������ (2.11) √(������������������������ )(������������������������ ) The correlation coefficient analysis is used to identify the relationship between two variables in the study domains. 2.2.9 Moisture Flux Convergence An analysis of Moisture flux convergence (MFC) is used to identify the physical mechanism responsible for producing extreme precipitation. It is a quantity that characterizes the strength of water vapor transport. Sufficient MFC supply is a central condition for heavy precipitation. Convergence over the precipitation area is also an important condition. The expression for MFC is derived from the conservation of water vapor in a pressure coordinate. However, the reanalysis of MFC involved vertical integrating the mathematical expression shown in Equation 2.8, where������is a gradient operator,������is the horizontal wind vector,������is specific humidity. ������������������ = −[������. (������������)] = −[������������. (������)] − [������. ������(������)] (2.12) The first term on the right-hand side of the equation (2.8) represents the convergent component of the MFC, which represent the product of the convergent component of wind with specific humidity. 17

The second term on the right-hand side of equation (2.8) represents the advective component, which describes the horizontal advection of specific humidity (Holman & Vavrus, 2012).⁠ 18

CHAPTER 3 THE CLIMATE FEATURES To determine the BoB trough and precipitation pattern over ICP, the mean state of the climate over the study area was first assessed. This analysis helped to understand how ICP’s precipitation varies after the monsoon onset. Therefore, this section is going to assess the patterns of the monthly averages for the various key parameters over ICP and BoB. 3.1 Climatic Features Over Study Domain 3.1.1 Rainfall Feature over ICP Distribution of precipitation over the ICP is investigated. Figure 3.1 shows an increase in precipitation from April to May, reaching its maximum in August, and then start to decrease until it reached its lowest point in February. The gradually increase in amount of precipitation from April is caused by several systems among which are the monsoon onset and northward migration of ITCZ starting from April and reached its highest point in August. Account for more than 80% of the total precipitation range (200 mm) as the threshold to determine wet season months, the 20- year average of monthly mean precipitation over ICP (Figure 3.1) presented May-October (MJJASO) as months with precipitation amount exceeding 200 mm. The wet season over ICP is therefore MJJASO. Chhin et al. (2017) also revealed that much of ICP receives rainfall over 75% of the total annual precipitation during MJJASO. 19

Figure 3. 1 Monthly precipitation (mm) averaged over Laos (green solid line) and Indochina Peninsula (red solid line), during 1998-2017, where the black solid line is the demarcation for the wet season in the region. Figure 3.2 showed the monthly spatial climatology of precipitation over ICP during 1998-2017, results which coexistence with its annual cycle. It also supports the results of past studies where the MJJASO is the wet season over ICP, with a south and southeast to north oriented precipitation zone dominating the region. The south and southwest noted to received more precipitation during the wet season, as compared to the center and north of the ICP. Magnificent point is southwest coastal area of Myanmar which receive high rainfall than the rest of the study domain. The precipitation observed over ICP is linked to the southwest monsoon which also peaks during August. During December to March, which is a dry seasonal month, the whole area receives less 300 mm precipitation with an exception to southeastern coastal area of Vietnam. 20

Figure 3. 2 Spatial distribution of monthly mean precipitation (mm) over study domain during 1998-2017. 3.1.2 Wind Circulation The mechanism by which water vapor flux is advected toward the continent can be understood by investigate of wind at low-level circulation 850 hPa (Figure 3.3). The wind circulation at this level also exhibits the semi-permanent cyclonic systems that are important features of the region. 21

Figure 3. 3 Monthly mean wind circulation (m s-1) at 850 hPa over the study domain during 1998-2017. The streamlines show the presence of BoB trough or the low-pressure system during the wet season months. Its position, which, is migrated in all dimension throughout the year, sometimes it moves northward into Indian subcontinent. 22

Another magnificent feature is the direction of the wind flow over ICP which has southwesterly characteristics during the wet season. From May to October, the wind flows from southwest to northeast over ICP while from November to April, the wind changes its direction to the northeasterly. When the wind flows from the southwest toward the peninsula, the topographical effect of the narrow mountains, cause the wind to move northwestward to the southern periphery of the Himalayan mountain and thus air rises and favors precipitation. 3.1.3 Temperature Figure 3.4 and 3.5 show the monthly average and temperature variation of air temperature during 1998-2017 over ICP. The temperature has characteristics of the southwest-northeast oriented temperature gradient, which much lower temperature in the northeast during November-March and then temperature start to increase from April to its highest in June and July. In contrast, southwestern part of the peninsula has nearly unified temperature year-round. ICP temperature is lowest during January, which has much temperature gradients northeast-southwest month due to the position of the sun. It makes averaged temperature of this month to be around 21.5°C. ICP experienced the hottest temperature in June and the temperature starting to decline in the followed months, which might be caused by the heavy downpours of precipitation over the area. However, an interesting feature to note is the high-temperature zone over the south of the peninsula. Figure 3. 4 Monthly mean temperature (℃) over ICP (red solid line) and Laos (black solid line) during 1998-2017. 23

Over the ICP, warm temperature covers entire region during the wet season. However, cool air temperature starts to creep into ICP as early as November, when the wind direction begins to shift from southwesterly to northeasterly. The northeastern part of ICP is noted to be cooler than the south. It is indicated that the northeast-southwest wind during this period carried the cool continental air mass from high latitude into the region. No big differences during the South Asia monsoon season Figure 3. 5 Monthly mean temperature (˚C) at 1000 hPa over ICP during 1998-2017. 24

3.1.4 Moisture Flux Convergence Figure 3. 6 Monthly mean moisture flux convergence by column of air (kg kg-1 m2 day-1) over the study domain. Wind vector showed the moisture transport and shaded area indicated convergence (positive) and divergence (negative). Moisture flux convergence is associated with precipitation as it has been used by researcher (Holman & Vavrus, 2012)⁠ to describe the extreme precipitation events and remote moisture sources are important for hydrologic budget. Figure 3.6 showed the monthly spatial climatology of moisture transport into ICP during 1998-2017. The maximum of MFC flux into the ICP is during MJJASO which also consistent with the precipitation analyses (Figure 3.2). The results indicated 25

that centered west and northwest have much more MFC than the rest, but according to precipitation analyses the area has more rainfall is south and southwest. From November the wind starts to flow northeasterly dry air from Asia continent and north PO flow into ICP and dry season is started. 3.2 Monsoon Onset and Its Index 3.2.1 Monsoon onset index The ICP climate and seasonal variation are influenced by the most dominant systems of the world climate called “monsoon”. The term monsoon stems from seasonal variations in winds, it is now more generally applied to tropical and subtropical seasonal reversal in both the atmospheric circulation and associated precipitation. These changes arise from reversals in heating and temperature gradients between continental regions and adjacent oceans with progression of the seasons, and the extremes are often best characterized as “wet” and “dry” seasons rather than summer and winter (Trenberth, Stepaniak, & Caron, 2000)⁠. Which ICP wet season is influenced by Asian Summer Monsoon (ASM) and dry season is influenced by Asian Winter Monsoon (AWM). During the wet season the wind at the low-levels are characterized by the strong westerlies at 850 hPa, and the large-scale cyclonic vorticity extending from the north BoB to the western India known as the “Monsoon Trough” (Rao,1976). The southwesterly wind current transport moisture from the tropical warm sea area toward south and southeast Asia, which coexistent with abruptly increase in amount of precipitation in this region. The dominant monsoon system is Asian- Australian monsoon, although has not been clearly identified with wind reversals (P. J. Webster et al., 1998)⁠. In order to define the starting time of the rainy season many scientists had invented an index called monsoon onset index based on climatological and atmospheric condition of the area of interest, such as, all-Indian rainfall, normalized precipitation, climatological pentad mean rainfall and circulation index (Zeng & Lu, 2004)⁠. To quantify the relationship between the trough of the BoB and rainfall over ICP after monsoon onset using a single index is difficult. It requires understanding of essential physics governing the BoB trough and rainfall variability. In this study we are going apply the idea from normalized precipitation, climatological pentad means rainfall and circulation index to compare and make onset index for this region. 26

According to Table 3.1, based on the reliability and deviation of monsoon onset of the three onset methods, we decided to apply the resulting of CPMI as the monsoon onset over the region, which has mean onset in pentad 23.81 (in between April 23 to April 28) and standard deviation of 1.91 (which approximated 10 day). Further data analyses will consider after these pentads. Table 3. 1 The monsoon onset which defined by normalized precipitation index (NPWI) column 2, Webster and Yang circulation index (WYI) column 3 and climatological pentad mean rainfall index (CPMI) column 4. Monsoon Onset Index index NPWI WYI CPMI Year Pentad 1998 Pentad Pentad 1999 26 2000 23 27 19 2001 21 2002 28 19 23 2003 23 2004 19 19 24 2005 24 2006 23 24 26 2007 24 2008 16 25 23 2009 23 2010 27 25 23 2011 26 2012 26 25 22 2013 24 2014 31 25 23 2015 26 2016 20 26 27 2017 26 2018 23 25 24 Mean 23 SD 18 24 23.81 1.91 28 21 26 26 21 24 29 23 31 26 20 26 18 22 24 27 33 27 16 22 23.81 24.19 5.11 2.40 27

3.2.2 Rainfall period after monsoon onset Figure 3.7 displays the power spectrum analysis of precipitation over ICP after monsoon onset 10 pentad data for 21 years during 1998-2018. The result (a) shows major peaks significance at 90% centered at 6, 7 and 8 pentad oscillation. It means that ICP rainfall oscillates by around 30-40 days period after the monsoon onset. The result (b) shows major peaks significance at 95% centered at 6.2 pentad oscillation. It means that ICP rainfall oscillates by around 30-35 days period after the monsoon onset. Figure 3. 7 Power spectrum analyses of rainfall of 10 pentad data after monsoon onset over ICP during 1998-2018 (significant bounds at 90% confident interval (thick dash line) and at 95% confident interval (light dash line) ). 28

3.3 Climatic Features after Monsoon Onset 3.3.1 Rainfall distribution after monsoon onset Figure 3. 8 Spatial distribution of pentad mean precipitation (mm) over ICP in ten different pentad data after monsoon onset. 29

To verify the significance of the oscillation of rainfall, climatological study of 10 pentad mean rainfall for 1998-2018 is displays in Figure 3.8. We could be observed high precipitation in second pentad then declined in following pentad and high precipitation again in around seventh to tenth pentad. This plot is well coherent with the power spectrum analyses as Figure 3.8. Figure 3.9 displayed a spatial distribution of rainfall and after monsoon onset. Which shows a significant difference in overall amount and spatially. In the two pentads before onset the amount of rainfall is nearly uniform less and dry, however, over the north part of Laos and Vietnam received relatively high rainfall. On the other hand, after monsoon onset, the amount of rainfall uniformly increased especially in the south and southwest. Figure 3. 9 Spatial distribution of pentad mean precipitation (mm) over ICP before and after monsoon onset. 3.3.2 Wind circulation after monsoon onset 30

Figure 3. 10 Mean wind circulation (m s-1) over ICP after monsoon onset based on 10 pentad reconstructed dataset. 31

Figure 3.10 shown the wind circulation over ICP at 850 hPa after monsoon onset. These indicated that, ICP dominated by southwesterly wind during this period. Which consistent with, some scientists, using wind as an indication to the monsoon onset. Moreover, wind at certain levels, had played a significant role in transported water vapor from the sea toward land, thus favor precipitation as in 3.8. Figure 3.11 shown the wind circulation at 850 hPa before and after the monsoon onset over ICP. Which cyclonic circulation are dominated over the study area during this time. However, before the onset, the cyclonic circulation is centered at southern tip of India and west BoB, therefore, the westerly wind over ICP which continental air mass occupied less moisture and so less rainfall over the region. On the other hand, after monsoon onset, south-westerly wind occupied more moisture from the IO toward the ICP, hence, abruptly increased in rainfall. Figure 3. 11 Spatial distribution of pentad mean wind circulation (m s-1) over study domain before and after monsoon onset. 32

3.3.3 Air temperature at 1000 hPa after monsoon onset Over ICP, surface air temperature is relatively high during the first three pentad after monsoon onset. However, the cooler surface air temperature starts to creep over ICP in the following pentad, and lowest surface air temperature observed on the last pentad. These situations indicated that, after monsoon onset southwesterly wind transport moisture from the sea toward continental land area, convergence of wind and topography effects of the height elevated mountains lift the air up, thus favor precipitation. Heavy downpour of precipitation, then cause dropping of surface air which relatively high before monsoon onset. On the other word, precipitation cooling effects as in Figure 3.12. 33