Integrating Remotely Sensed Satellite and Standardized Precipitation Index for Drought Assessment in the Climate Change affected Xe Bang Hieng River Basin = 기후변화의 영향을 받는 세방히엥 (Xe Bang Hieng) 강 유역의 가뭄 평가를 위한 원격탐사 위성자료와 표준강수지수(SPI)의 통합

다국어 초록 (Multilingual Abstract)

Climate change is heightening high-spatial resolution drought hazard exposure in Lao PDR's monsoon basins, but few long, continuous observations are available to assess at the basin-scale and plan. In this master’s thesis, meteorological drought in the Xe Bang Hieng River Basin has been assessed using the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) v3.0 daily rainfall (1981–2024) combined with the Standardized Precipitation Index (SPI). Daily CHIRPS data was quality controlled by QA/QC (transfer-date parsing, missing/duplicate files screening and unit checking), then summed to monthly sums and arranged into hydrologically relevant seasons for Laos: Wet (May–Oct) and Dry (Nov–Apr) in the start-year SeasonYear convention to treat cross-year dry seasons consistently. SPI was calculated at different timescales (SPI-3, SPI-6, SPI-9 and SPI-12), and seasonal drought metrics were calculated to reflect the frequency of drought occurrence (no. of months with precipitation index SPI ≤ −1 for each season) and intensity (the minimum value of the seasonal SPI). The non-parametric Kruskal–Wallis tests were used to test the differences between ‘within-year seasonal groups,’ and between ‘five-decade periods,’ with Dunn–Holm post-hoc when needed.
The results demonstrate strong monsoon control, as wet-season rainfall significantly exceeds dry-season totals (median 1746.5 vs 237.8 mm/season). The SPI series were nicely normalized (mean ≈ 0 and SD ≈ 1), and moderate drought conditions (SPI ≤ −1) emerged in ~16–17% of months over timescales. Kruskal–Wallis testing revealed no systematic differences in SPI among the seasonal phases in Laos, which aligns with the expectation that SPI reflects the deterministic seasonal cycle; however, it successfully detected significant decadal variation in wet-season rainfall input (p=0.0316) and persistence-sensitive drought behavior, particularly for wet-season SPI-12, where droughts occurred under normal precipitation levels (p=0.0191) and intensity (MinSPI; p=0.00979). Post-hoc inference survived one strong decade contrast (wet SPI-12 MinSPI: 1991–2000 vs. 2021–2024; pHolm=0.029947), but with caution because of the shorter recent period. Overall, the results emphasize the importance of wet-season persistence for timely drought early warning and water resources management in an adaptive fashion in the Xe Bang Hieng River Basin and Lao PDR.

국문 초록 (Abstract)

기후변화로 인해 라오스 (Lao PDR)의 몬순 유역 전반에서 가뭄 위험이 증가하고 있으나, 유역 규모의 평가와 계획 수립을 위한 장기간의 일관된 관측자료는 제한적이다. 본 학위논문은 CHIRPS v3.0 일강수량 자료(1981–2024)와 표준강수지수 (SPI)를 통합하여 세방히엥 (Xe Bang Hieng Basin)강 유역의 기상학적 가뭄을 평가하였다. 일 단위 CHIRPS 자료는 엄격한 품질보증/품질관리 (QA/QC) 절차(날짜 파싱, 결측 및 중복 기록 선별, 단위 검증)를 거친 후 월 누적 강수량으로 집계되었다. 또한 라오스의 수문학적으로 의미 있는 계절 구분인 우기(5–10월)와 건기(11–4월)로 자료를 조직하였으며, 연도를 넘나드는 건기를 일관되게 처리하기 위해 시작연도 기준의 SeasonYear 규약을 적용하였다. SPI는 여러 시간 규모 (SPI-3, SPI-6, SPI-9, SPI-12)에서 산정하였고, 계절별 가뭄 지표로 가뭄 빈도 (계절 내 SPI ≤ −1인 월수)와 가뭄 강도(계절 최소 SPI, MinSPI)를 도출하였다. 계절 내 구분 및 5개 10년(Decade) 기간 간 차이를 평가하기 위해 비모수 Kruskal–Wallis 검정을 적용하였고, 필요 시 Dunn–Holm 사후검정을 수행하였다.
분석 결과, 우기 강수량이 건기 강수량을 크게 상회하는 등 강한 몬순 지배가 확인되었다(중앙값: 우기 1746.5 mm/계절, 건기 237.8 mm/계절). SPI 시계열은 표준화 특성이 양호했으며(평균 ≈ 0, 표준편차 ≈ 1), 중등도 가뭄(SPI ≤ −1)은 시간 규모 전반에서 전체 월의 약 16–17%에서 발생하였다. Kruskal–Wallis 검정에서는 SPI가 결정론적 계절 주기를 제거하는 특성에 부합하게 라오스 계절 단계 간 SPI의 체계적인 차이는 나타나지 않았으나, 우기 강수 공급의 10년 규모 변동성(p=0.0316)과 지속성에 민감한 가뭄 거동에서 유의한 10년 규모 차이가 탐지되었다. 이는 특히 우기 SPI-12에서 가장 두드러졌으며, 가뭄 빈도(p=0.0191)와 강도(MinSPI; p=0.00979) 모두에서 유의했다. 사후 분석에서는 하나의 견고한 10년 대비가 유지되었는데(우기 SPI-12 MinSPI: 1991–2000년 vs 2021–2024년, pHolm=0.029947), 최근 기간이 더 짧다는 점을 고려해 해석에 주의가 필요합니다. 종합적으로, 이 연구는 세방히엥 강 유역과 라오스에서 가뭄 조기 경보와 적응적 수자원 관리를 위해 우기 강수의 지속성을 우선적으로 고려해야 함을 강조한다.

목차 (Table of Contents)

• CHAPTER 1. INTRODUCTION 1
• 1.1 BACKGROUND 1
• 1.2 PROBLEM STATEMENTS 2
• 1.3 RESEARCH OBJECTIVES 3
• 1.4 STUDY AREA 5
• 1.4.1 Overview of Xe Bang Hieng River Basin 5
• 1.4.2 Climate 7
• 1.4.3 Topography 9
• 1.4.4 Water Resources 10
• 1.4.5 Socioeconomic Status 11
• 1.5 THESIS STRUCTURE 11
• CHAPTER 2. LITERATURE REVIEW 13
• 2.1 FORECASTING DEFINATION 13
• 2.2 DROUGHT CONCEPT 16
• 2.2.1 Drought Definition 16
• 2.2.2 Drought Characteristics 17
• 2.2.3 Drought Types 18
• 2.2.4 Drought Frequency 20
• 2.2.5 Drought Impacts 21
• 2.2.6 Drought Indices 22
• 2.3 CONCEPT OF STANDARDIZED INDICES 24
• 2.3.1 Standardized Precipitation Index 25
• 2.3.2 Standardized Precipitation Evapotranspiration Index 26
• 2.4 REMOTELY SENSED SATELLITE 28
• 2.4.1 Optical Remote Sensing Indices 28
• 2.4.2 Remotely Sensed Satellite-based Meteorological Data 30
• 2.4.3 CHIRPS v3.0 33
• 2.5 CLIMATE CHNAGE 34
• 2.5.1 Global Climate Change & Spectral Impacts 34
• 2.5.2 Climate Characteristics & Hydrological Drivers 35
• 2.5.3 Global Warning & Climate Variability 36
• 2.5.4 Southeast Asia’s Climate Change 37
• 2.5.5 Mekong River Basin’s Climate 38
• 2.6 STATISTICAL CONCEPT 40
• 2.6.1 Pettitt Test 41
• 2.6.2 Kruskal Wallis Test 42
• 2.6.3 Chi-Square Test 43
• 2.7 RELATIVE STUDIES 44
• 2.8 RESEARCH QUESTIONS 54
• CHAPTER 3. METHODOLOGY 59
• 3.1 RESEARCH DESIGN 60
• 3.1.1 Data Acquisition and Spatial Boundary 61
• 3.1.2 Data Preparation 64
• 3.1.2.1 Data Screening 69
• 3.1.2.2 Temporal Aggregation 70
• 3.1.2.3 Season Construction 72
• 3.1.2.4 Periodization and Grouping 77
• 3.1.2.5 Chirps-based SPI Computation 81
• 3.1.2.6 SPI-based Seasonal Drought Metrics 85
• 3.2 DATA ANALYSIS TOOLS 88
• 3.3 STANDARDIZED PRECIPITATION INDEX 89
• 3.4 KRUSKAL-WALLIS TEST 98
• 3.5 POST-HOC ANALYSIS 101
• CHAPTER 4. RESULTS 103
• 4.1 METEOROLOGICAL DROUGHT 103
• 4.2 SEASONAL STRUCTURE OF RIANFALL AND DROUGHT 109
• 4.3 KRUSKAL-WALLIS RESULTS BY SEASONAL GROUPS 112
• 4.4 KRUSKAL-WALLIS RESULTS BY DECADES (5 GROUPS) 119
• 4.4.1 Data Acquisition and Spatial Boundary 119
• 4.4.2 SPI-derived Seasonal Drought Metrics across Decades 123
• 4.4.2.1 Key Decade Boxplots (Metrics) 134
• 4.4.2.2 Kruskal-Wallis H across SPI Scales 138
• 4.5 POST-HOC DUNN–HOLM DECADE CONTRASTS 143
• 4.6 INTEGRATED DISCUSSION AND IMPLICA-TIONS 146
• 4.7 UNCERTAINTY, ASSUMPTIONS, AND LIMITATIONS 148
• CHAPTER 5. CONCLUSION 151
• 5.1 DISCUSSION 151
• 5.1.1 Results’ Interpretation 151
• 5.1.2 Comparison with Extant Literature 153
• 5.1.3 Strategical Implications 154
• 5.2 SUMMARY 155
• 5.2.1 Strategical Implications 155
• 5.2.2 Limitations and Future Research 157
• BIBLIOGRAPHY 160