00051000980 Investigating compound drought and heatwave characteristics over Southeast Asia in recent decades
The necessity of research
Drought is a complex natural hazard that disrupts water supplies, agriculture, and socioeconomic conditions worldwide It is defined by extended periods of insufficient rainfall that cause water shortages and strain on resources In recent decades, droughts have become more intense and frequent, driven in part by climate change, with wide-ranging impacts on food security, livelihoods, and regional economies (WMO, 2006; Naumann et al., 2018).
Heatwaves are abnormal, prolonged periods of extreme heat, and global heatwaves are increasing in frequency and intensity due to climate change They threaten ecosystems, economies, and public health, with health risks intensifying even in temperate regions like Western Europe where populations are not yet adapted In tropical regions, especially Southeast Asia, health impacts are more pronounced owing to limited access to cooling equipment and medical care Therefore, understanding and mitigating heatwaves in Southeast Asia is essential to identify vulnerability hotspots and safeguard each country's economy, agriculture, ecosystems, and public health.
Compound drought-heatwave events, the simultaneous occurrence of prolonged dry spells and extreme heat, pose serious challenges to global economies, public health, agriculture, and ecosystems Climate change is driving these compound events to occur more frequently and with greater intensity, highlighting the urgent need for extensive research to understand their impacts and to develop effective adaptation and prevention strategies A holistic approach—integrating climate science, policy, and resilient agricultural practices—is essential to cushion economic losses, protect health, secure food supplies, and safeguard ecosystems against future drought-heatwaves Early warning systems and data-driven risk assessments can empower communities to prepare for and mitigate the socioeconomic costs of these events.
Global projections indicate that compound drought and heatwave (CDHW) events will become more frequent and severe as climate change continues These overlapping extremes are increasing in occurrence, exacerbating impacts in regions with already high climatic variability (IPCC, 2021) Understanding the evolving trends of CDHW events is essential for accurate climate risk assessment and for developing comprehensive adaptation strategies.
Drought and heatwaves interact to intensify other environmental stresses, such as wildfires, threatening ecosystem health and biodiversity When heatwaves and drought overlap, agricultural productivity declines and global food security is jeopardized, with the combined effects often exceeding the impact of either factor alone These synergistic extremes underscore the need for integrated climate resilience strategies that protect crops, ecosystems, and food systems worldwide.
Southeast Asia experiences a hot, humid tropical climate with high year-round rainfall, though rainfall patterns vary across the region due to the Asian monsoon system Droughts persist, particularly on the Southeast Asian mainland From 1979 to 2018, heatwaves became more frequent, longer, and more intense across Southeast Asia, driven by rising temperatures, high humidity, and El Niño events The co-occurrence of drought and heatwaves can trigger severe cascading impacts on ecosystems, economies, and human health.
To understand CDHW events in Southeast Asia and to design effective mitigation and adaptation plans, extensive, multidisciplinary research is crucial These compound hazards require comprehensive investigation and cross-sector collaboration because they threaten economies, public health, agriculture, and ecosystems Given the region’s extreme climate-change vulnerability, research and strategy development must be prioritized to protect the population, ecology, and economy An integrated, proactive approach is essential to build resilience against CDHW events and safeguard sustainable development in Southeast Asia.
Both drought and heatwave are natural hazards in their own right, but when they coincide as a drought–heatwave compound event, they pose a far greater threat This raises the question of how compound drought–heatwave events and their characteristics have evolved in Southeast Asia (SEA) in the context of climate change over recent decades Investigating these trends is crucial for understanding projected changes in natural disasters related to drought, heatwaves, and their combination, and for informing preparedness, resilience, and adaptation strategies in the region.
Literature review
Drought
Drought is defined as a prolonged natural deficit in precipitation over an extended period, typically a season or longer Along with reduced rainfall, higher temperatures, strong winds, and low relative humidity often accompany drought and can intensify its impacts Unlike storms and floods, drought generally causes non-structural damages and tends to be geographically widespread rather than localized Drought is a normal, recurring feature of climate that can occur in both high- and low-precipitation regions, and it should be distinguished from aridity, a permanent climatic characteristic confined to areas with inherently low precipitation.
Drought is categorized into several distinct types: meteorological drought, hydrological drought, agricultural drought, socioeconomic drought (Wilhite & Glantz,
Meteorological drought is defined by the degree of dryness, typically assessed by precipitation deficits relative to a historical average and the duration of the dry spell, and it is context-specific due to regional climate regimes Hydrological drought concerns deficits of water within the hydrological system, while agricultural drought relates to soil moisture deficits in conjunction with meteorological conditions and other climatic factors, and their impact on agricultural production and economic efficiency Socioeconomic drought arises from a water scarcity caused by an imbalance between water supply and demand within both natural and human socioeconomic systems, and these drought types are intrinsically linked to sustainable socioeconomic development In research, meteorological drought is often emphasized when studying a particular region; this thesis focuses on meteorological drought in Southeast Asia (SEA), with a particular emphasis on analyzing its characteristics when it co-occurs with heatwaves.
Beyond its categorization into distinct types, drought is characterized by severity, duration, and spatial extent, with additional defining features such as frequency, magnitude, and trends as described in established terminology Drought duration is region-specific, ranging from a few weeks to several years, and droughts can exhibit concurrent wet and dry periods when viewed across different temporal scales due to their complex and fluctuating nature Magnitude measures the cumulative water deficit below a defined threshold over the drought period, while severity denotes the depth of precipitation deficit or the extent of resulting impacts Spatial extent refers to the area affected by drought, which can include a single grid cell, river basin, or administrative region, and frequency—the average time between drought events meeting or exceeding a specified threshold—captures how often droughts recur.
Observing drought is challenging because its main indicator is a sustained decrease in precipitation over a period relative to long-term averages, which has driven the development of numerous drought indices for forecasting, monitoring, and planning The Percent of Normal (PON) is a simple climatic drought index that compares current precipitation to a 30-year average, but its lack of statistical transformation can make it less robust, as median–mean discrepancies can reduce accuracy and regional/seasonal distribution differences hinder cross-temporal or cross-regional comparisons The Deciles method partitions monthly precipitation into ten equal parts based on long-term data and typically emphasizes the lowest 10%, using two categories to describe precipitation deficits as severe and extreme drought The Standardized Precipitation Index (SPI) is a widely used drought index that converts precipitation data to a standard normal distribution with a mean of zero, where values above zero indicate wet periods and values below zero indicate dry periods, and drought is signaled when a value below −1 is reached for a sustained period Despite its strengths, the SPI is limited in detecting increasing drought duration and intensity driven by higher temperatures The Standardized Precipitation Evapotranspiration Index (SPEI) (Vicente-Serrano et al.) is another drought index in this framework that incorporates evapotranspiration to better reflect climate-driven moisture conditions.
Introduced around 2010, the SPEI extends the SPI by incorporating temperature data and explicitly accounting for the water balance and evapotranspiration, enabling a more robust assessment of drought under climate-change scenarios By addressing the SPI’s limitations in capturing drought characteristics linked to temperature, the SPEI provides a clearer measure of climate-change impacts on drought When temperature does not show significant trends, the SPEI converges with the SPI and yields essentially similar results The Palmer Drought Severity Index (PDSI) is another major drought metric, developed by Palmer.
Developed in 1965, the Palmer Drought Severity Index (PDSI) defines drought as an imbalance between water supply and demand rather than solely as precipitation anomalies It uses data on precipitation, temperature, and locally available soil water content to compute four components of the water balance: evapotranspiration, surface runoff, soil moisture recharge, and moisture loss The US Drought Monitor (USDM) is a composite drought indicator that integrates multiple indices, including the SPI and PDSI, as well as indicators such as vegetation health and hydrological conditions, to produce a weekly drought map.
Global warming since the 1980s has expanded drought-affected areas, increasing the global drought footprint by about 8% in the first decade of the 21st century, with the drying largely driven by enhanced evaporation in the mid- to high-latitudes of the Northern Hemisphere In Africa, Southeast Asia, eastern Australia, and southern Europe, reduced precipitation emerges as the primary driver of drying trends Dai (2012) and the PDSI analysis from 1870–2002 show that very dry regions (PDSI < -3.0) more than doubled since the 1970s, marked by a spike in the early 1980s due to ENSO-related precipitation deficits and continued expansion driven by surface warming, while very wet areas (PDSI > 3.0) declined somewhat in the 1980s These observations provide evidence that drought risk is increasing as anthropogenic warming continues, producing both higher temperatures and drier conditions Climate models project an intensification of drought in the 21st century across most of Africa, southern Europe and the Middle East, large parts of the Americas, Australia, and Southeast Asia, with regions such as the United States—previously buffered by natural variability—potentially facing persistent drought in the next 20–50 years.
In Southeast Asia (SEA), Phan-Van (2022) analyzed temperature and precipitation data to compute the Standardized Precipitation Evapotranspiration Index (SPEI) from 1960 to 2019, examining how drought relates to major climatic drivers The study identifies four distinct drought subregions in SEA, with the Maritime Continent and mainland Indochina exhibiting markedly different drought characteristics that have become more pronounced in recent decades It also reveals a substantial link between the El Niño–Southern Oscillation (ENSO) and SEA drought Separately, Nguyen-Ngoc-Bich (2022) used the Standardized Precipitation Index (SPI) to evaluate drought conditions and employed two Representative Concentration Pathways (RCPs) to explore drought features under different greenhouse gas emission scenarios Under the RCP8.5 scenario, projections point to shorter yet more severe droughts in the late 21st century, especially in mainland Indochina.
Tropical Southeast Asia is divided into two climate zones: the Monsoon Climate Region (MCR) and the Equatorial Climate Region (ECR) The MCR, which includes Myanmar, Thailand, Laos, Cambodia, and Vietnam, shows pronounced seasonal differences in temperature and precipitation, while the ECR, encompassing Malaysia and Indonesia, exhibits substantial regional variation with Malaysia typically experiencing higher rainfall and warmer temperatures In Tropical Southeast Asia, key variables such as temperature, precipitation, crop season, and monsoon patterns shape drought indices, with the monsoon season playing a central role by driving rainfall variability that governs regional drought conditions (Zaki & Noda, 2022).
Heatwaves
Heatwaves are frequently defined by temperature thresholds and duration, using either absolute limits (for example a daily maximum temperature exceeding a fixed value on several consecutive days) or percentile-based thresholds that reflect the local climate To detect and analyze heatwaves, researchers employ several methodologies, including the Constant Threshold, the Average Temperature plus 5°C method, the Upper Tail percentile approach, Summer-derived thresholds, the Excess Heat Factor (EHF), and the Summer Heat Index (SHI) An alternative method identifies heatwaves through the local 90th percentile of observed daily maximum temperatures (Tx) within a specified period These definitions and methods provide a framework for consistent heatwave detection and assessment of impacts across regions and time.
Recent research by Qiu and Yan (2020) shows that heatwave events were historically relatively infrequent across large parts of Asia, North America, Europe, and North Africa, but a notable increasing trend has emerged, especially in East Asia The analysis reports annual frequency increases of 0.1 events in China, 0.07 in Japan, and 0.09 in South Korea It also highlights the need for preparedness in the South-Central United States and Southern Mexico to mitigate potential impacts from extreme heat and suggests a possible rise in heatwave frequency in North-Central Europe, particularly in Germany and Sweden.
Building on prior findings, Han et al (2022) examined global heatwave characteristics across a range of emission scenarios Their projections indicate that by the 2050s, heatwaves could last roughly four times longer than in a reference period, with the longest-lasting events expected to concentrate in Central Africa, northern South America, and Southeast Asia Under high-emission conditions, individual heatwaves may extend substantially, highlighting the potential for markedly increased exposure to extreme heat in vulnerable regions.
Over 44 days, projections indicate rapid upward trends in key climate-impact metrics under a high emission scenario, with intensity doubling, total annual duration increasing eightfold, and temperature magnitude rising ninefold relative to the reference period When socio-economic vulnerability is considered, Han et al (2022) identify hotspots in Western Europe, eastern North America, and northern China facing heightened future risk, underscoring the urgent need to strengthen adaptive capacity in these regions.
Yin et al (2022) analyze contemporary global patterns and impacts of heatwaves, identifying North Africa, Northern Australia, South Asia, and the Arabian Peninsula as regions currently experiencing severe heatwave impacts, defined by more than three events per year and a total of over 15 heatwave days annually The study also highlights that North Africa, Australia, and Brazil are regions with particularly rapid increases in heatwave frequency.
Global warming is driving a rise in both the frequency and intensity of heatwaves worldwide In Southeast Asia, Mukherjee and Mishra (2021) analyze heatwaves by examining their frequency, duration, and intensity, finding that heatwaves are becoming more severe as the climate warms While regional impacts vary, the overall pattern shows more frequent and intense heatwaves Perkins-Kirkpatrick and Lewis (2020) highlight that extreme heat events are now far more likely due to human-caused climate change, with future projections indicating this trend will continue if greenhouse gas emissions remain high However, more research is needed to fully understand the specific effects of heatwaves on vulnerable populations and key sectors in Southeast Asia.
Investigating heatwave trends across Southeast Asia (SEA) for the period 1979-
Heatwaves are becoming more frequent, longer, and more intense across most of the region, as shown by Li (2020) The study also identifies a significant link between heatwave characteristics and the El Niño index, indicating that large-scale climate patterns modulate regional heatwave activity In Southeast Asia, Li et al (2022) report that metrics of extreme annual events exhibit particularly strong trends, with the magnitude of the hottest heatwave each year increasing faster than the corresponding average intensity Additionally, the duration of the longest annual heatwave shows a statistically significant upward trend across a broader area in Southeast Asia than the metric for average heatwave duration, suggesting that the most extreme heatwaves are intensifying more rapidly than mean conditions would imply.
Compound drought and heatwave events (CDHW)
Compound events are defined as more than one extreme event occurring simultaneously or successively, or as the combination of extreme events with underlying conditions that amplify their impacts A complementary definition describes compound events as the result of nonextreme events interacting to produce an extreme event or impact, as stated by IPCC (2021).
Heatwaves and droughts reinforce each other through a self-amplifying cycle: higher temperatures boost evaporative demand, pulling water from soils and vegetation and causing further drying, while arid conditions lack enough moisture for evaporative cooling, letting heat accumulate As soils and vegetation dry, evapotranspiration declines, producing drier air that reduces rainfall chances and sustains aridity More solar radiation is converted into sensible heat instead of latent heat, warming the environment and intensifying heatwaves This creates a positive feedback loop—dry soil raises temperatures, which increases atmospheric water demand and accelerates soil drying In a warming climate, the feedback between soil moisture and temperature is a major driver of hot extremes.
Compound drought and heatwave (CDHW) events occur when a heatwave overlaps with drought, requiring separate identification of drought and heatwave indicators before detecting CDHW For CDHW detection, indices such as the Standardized Precipitation Index (SPI) for meteorological droughts and the Standardized Heatwave Index (SHI) for heatwaves are used (Shan et al., 2024) Additional drought indicators include the Palmer Drought Severity Index (PDSI) and the Standardized Precipitation Evaporation Index (SPEI), while heatwave indicators can comprise daily maximum temperatures or the Excess Heat Factor (EHF) (Mukherjee et al., 2020; Pan et al., 2024).
Shan et al (2024) propose a method to identify CDHW (concurrently dry and hot) events on a daily scale across the four seasons in Belgium The study reveals seasonality in CDHW events, showing a negative dependence between droughts and heatwaves in winter and a positive dependence in the other seasons.
CDHW is characterized by several attributes, including frequency, duration, severity, and spatial extent Among these, frequency is the most commonly assessed feature and can be interpreted as the joint period of drought and heatwave events The dry and hot indicators exhibit functional correlations that help quantify the severity of CDHW, as discussed by Hao et al (2022) For example, based on the severity levels of drought and heat extremes, the Dry-Hot Magnitude Index (DHMI) was introduced by Wu et al (2019) to provide a unified measure of CDHW severity.
Abella and Ahn's 2024 study analyzes the pan-Asian region and identifies Southeast Asia and other areas as particularly vulnerable to compound dry hazards (CDHW) The findings indicate changes over time in the frequency, severity, or duration of CDHW events, signaling a rising climate-related risk The research examines how heatwaves, drought, and wildfires interact and intensify one another, creating compounded impacts Although the study provides a broad overview of CDHW risks across the pan-Asian region, it also notes that more in-depth research is needed to fully understand the distinctive characteristics and effects of these hazards in specific countries or subregions.
Southeast Asia has emerged as a major hotspot for coexisting climate extremes, attracting attention from researchers worldwide Regional studies focusing on Malaysia (Muhammad et al., 2018) and the larger Indochina and Maritime Continent sub-regions (Phan-Van et al., 2022) have been prominent in the literature The research timeframe generally extends from the late 20th century to the present, aligning with growing concerns about the impacts of climate change in the region.
Compound events caused by drought and heatwaves (CDHW) are significantly influenced by climate change, as shown by several studies CDHW events can lead to intensified damage to ecosystems, economies, and societies, especially in a warming climate Zscheischler et al (2018) highlight the heightened occurrence and intensity of these events in relation to global warming.
Drought-related heatwaves have been more frequent in recent times, as has impacted the global land area (Mukherjee & Mishra, 2021)
Gridded datasets and station observations are widely used in the statistical analysis of historical climate data to identify trends and patterns in heatwaves and droughts (CDHW) Climate modeling, including regional and global models, is essential for projecting future CDHW under different climate-change scenarios In addition, many studies leverage remote sensing to monitor surface temperatures and drought conditions, delivering valuable geographical data The range of methodologies reflects the multifaceted nature of CDHW events and the need for a comprehensive understanding through diverse analytical tools.
To accurately project changes in compound dry and hot days (CDHW), it is essential to use both global and regional climate models Global climate models (GCMs), including CMIP5 and CMIP6, can detect long-term trends in CDHW frequency on a large scale (Wu et al., 2021) In contrast, high-resolution regional climate models (RCMs), such as those developed under CORDEX, better capture the observed frequency and variability of CDHW events (Hao et al., 2022) Regional simulations across China indicate that RCMs generally align with the observed regional pattern of CDHW frequency (Lu et al., 2018).
Despite progress in understanding climate extremes at global and regional scales, there is a critical need for precise regional-scale assessments within Southeast Asia to capture the unique characteristics and impacts of these events Few studies quantify the spatio-temporal dynamics and trends of compound drought-heat waves (CDHWs) across Southeast Asia Most SEA research analyzes heatwaves and drought separately, overlooking the compounded impacts when these events coincide Consequently, only a limited number of studies map the large-scale occurrence of CDHWs in SEA This research aims to fill that gap by examining how CDHW characteristics have evolved over the past decade and how ENSO phases modulate them By elucidating the spatial and temporal patterns and projected trends of CDHWs in SEA, the work can improve early warning systems and support country-specific adaptation strategies.
Linkage between El Niủo- Southern Oscillation (ENSO) and CDHW events
Hao et al (2022) indicated that various large-scale climate drivers, for example,
El Niño–Southern Oscillation (ENSO), the Pacific Decadal Oscillation (PDO), the Atlantic Multidecadal Oscillation (AMO), and the North Atlantic Oscillation (NAO) can drive the formation of compound droughts and hot temperature extremes More specifically, ENSO has been demonstrated to influence the seasonality of these compound droughts and heat events across diverse regions.
In tropical and subtropical regions, particularly Southeast Asia, ENSO strongly influences precipitation and temperature, thereby altering the frequency of compound dry and hot events during summer Evidence indicates that during El Niño, Southeast Asia tends to experience drier and warmer conditions, whereas during La Niña the region tends toward wetter and cooler conditions (Hao et al., 2018).
ENSO exerts its most direct influence on Southeast Asia through its modulation of precipitation El Niño episodes typically weaken the region’s trade winds and suppress rainfall, increasing agricultural and meteorological drought risk, while La Niña events can bring above-average rainfall Building on Phan-Van et al (2022), this study analyzes drought variability and its links to major climate drivers in Southeast Asia over a 60-year record.
It is demonstrated that drought is strongly influenced by ENSO than other large-scale climate drivers over SEA
El Niủo is one of the drivers that cause drought conditions across SEA Strong
El Niño episodes can trigger severe water shortages by reducing rainfall, threatening ecosystems, water supplies, and agricultural productivity In Southeast Asia, major drought events occurred during the 1982–83 and 1997–98 El Niño years, underscoring regional vulnerability to climate variability Research by Juneng and Tangang (2005) shows a strong association between El Niño phases and meteorological droughts across Malaysia and the broader Maritime Continent These findings highlight the critical link between El Niño and drought risk, with important implications for water management and agricultural planning in the region.
SEA precipitation responds to ENSO with substantial regional and intra-country variability For example, some studies indicate that western Indonesia may experience increased rainfall during El Niño, contrary to the general drying trend associated with ENSO The strength of the ENSO influence also varies seasonally, with La Niña during June–July–August showing a strong positive impact across several regions of Asia However, observational data and ERA5 reanalysis reveal noticeable discrepancies in Southeast Asia across all four seasons (Sourav Mukherjee et al 2020).
ENSO phases have a significant effect on surface air temperatures in Southeast Asia, amplifying the region’s long-term warming trend El Niño is generally associated with warmer-than-average temperatures across Southeast Asia, as reduced cloud cover and rainfall allow more solar radiation to reach the surface Globally, El Niño events often trigger temporary spikes in the global mean surface temperature and can lead to record warm years, with future projections (Nguyen–Le, 2024) suggesting amplified positive temperature anomalies during El Niño, especially over land and notably during winter in northern Indochina and Myanmar In contrast, La Niña tends to cool temperatures, bringing near-average or below-average conditions in parts of Southeast Asia due to increased cloud cover and rainfall that limit warming; observations during the Jan–Mar 2021 La Niña phase showed below-average temperatures over parts of Mainland Southeast Asia Globally, La Niña events commonly cause temporary dips in the global average temperature.
Southeast Asia’s geography splits into a Maritime Continent and Mainland region, producing distinct ENSO-driven rainfall responses; Nguyen-Thanh et al (2023) document notable differences in how ENSO phases affect precipitation variability across SEA The Maritime Continent, including Indonesia, Malaysia, and the Philippines, shows a strong, direct rainfall response to ENSO driven by the western Pacific arm of the Walker Circulation, with severe rainfall deficits during El Niño and surpluses during La Niña, a vulnerability supported by historical drought data (Herho, 2018) and reliable model projections (Nguyen-Le, 2024) Mainland Southeast Asia’s rainfall response is more complex and less direct, heavily modulated by the Asian monsoon, with ENSO impacts varying by season and ENSO type; Indochina may experience alternating wet/dry patterns during El Niño, and the drying signal from El Niño Modoki can be stronger there than in the Maritime Continent Although La Niña generally brings wetter conditions, some parts of mainland Indochina experience drier boreal summers (JJA) during La Niña, as noted by Nguyen-Thanh et al (2023).
ENSO-related temperature anomalies affect both land and sea, yet future projections indicate a land-sea contrast in which ENSO’s warming influence intensifies more on land areas (including both mainland and islands) than in the surrounding oceans (Nguyen-Le, 2024) Notably, the amplification of El Niño warming is projected to be especially strong for winter temperatures in northern Indochina and Myanmar.
The Maritime Continent remains a core region where ENSO's influence on precipitation is most direct and pronounced due to its pivotal location relative to the Pacific Walker Circulation's convective centers Mainland Southeast Asia's climate response, while clearly modulated by ENSO, is shaped by a more complex interplay with the Asian monsoon system and other climate drivers like the Indian Ocean Dipole (IOD) (Juneng & Tangang, 2005) Understanding these sub-regional distinctions is vital for tailoring climate adaptation and mitigation strategies across Southeast Asia's diverse landscapes Further studies are needed to investigate the relationship between ENSO and CDHW in SEA.
Introduction of the research site
Southeast Asia is a strategically important region located between the Pacific and Indian Oceans, spanning roughly 30°N to 18°S latitude and 90°E to 150°E longitude The region encompasses eleven countries: Brunei, Myanmar, Cambodia, Timor-Leste, Indonesia, Laos, Malaysia, the Philippines, Singapore, Thailand, and Vietnam.
Figure 1.1 The eleven countries within the boundaries of SEA
The region comprises two distinct zones: the mainland—home to Thailand, Vietnam, Cambodia, Laos, and Myanmar—and the maritime realm, including Indonesia, the Philippines, Malaysia, Singapore, Brunei, and East Timor It is defined by extensive coastlines, major river systems such as the Mekong, Irrawaddy, and Chao Phraya, and a diverse terrain featuring significant mountainous areas alongside vast archipelagos.
Figure 1.2 The elevation map of SEA
Southeast Asia's climate is dominated by tropical monsoon patterns that drive distinct wet and dry seasons under the Southwest and Northeast monsoons In recent decades, the region has experienced an increase in the frequency and intensity of extreme weather events—such as droughts and heatwaves—often amplified by the El Niño–Southern Oscillation (ENSO).
Research questions, hypotheses, objectives, tasks
This study has 03 hypotheses corresponding to 03 research questions shown in the following table
Table 1.1 Research questions and hypotheses
Q1: What methods can be used to identify a compound drought-heatwave (CDHW) event? Additionally, what are the characteristics of CDHW events, and how can they be determined?
H1: It is possible to determine CDHW based on the indices used to identify drought and heatwave events
Q2: What is the spatial variability of the characteristics of CDHW events across
H2: The weather and climate of the SEA region are influenced by various natural climate drivers, resulting in a complex distribution of CDHW events
Q3: What are the influences of the El
Niủo–Southern Oscillation (ENSO) on
In the context of ongoing climate change, weather conditions are continually fluctuating and influenced by diverse climatic phenomena; among them, ENSO (El Niño-Southern Oscillation) significantly shapes the temperature and rainfall patterns of Southeast Asia (SEA) and exhibits a notable relationship with CDHW.
Based on the research questions and hypotheses, the research objectives and the sub-division tasks are presented in the following table
Table 1.2 Research objectives and tasks
CDHW events based on existing methodologies and established drought and heatwave indices
- Task 1.1: Conduct a comprehensive review of existing methodologies to identify CDHW events
- Task 1.2: Selection of drought and heatwave indices
- Task 1.3: Define clear criteria for what constitutes a CDHW event
O2: To analyze the spatial and temporal characteristics of CDHW events in SEA in recent decades
- Task 2.1: Choose data sources which include long-term and high-resolution gridded climate data (temperature, precipitation) across SEA in recent
- Task 2.2: Apply the chosen methodologies to identify CDHW events to the climate datasets
- Task 2.3: Determine key characteristics of CDHW events and analyze the spatial distribution of CDHW event frequency, duration, and severity across SEA in recent decades
- Task 2.4: Determine temporal characteristics of CDHW events across SEA in recent decades
O3: To investigate the influence of
ENSO on the characteristics of
CDHW events across SEA in recent decades
- Task 3.1: Analyze historical ENSO index data and categorize ENSO phases
- Task 3.2: Compare the characteristics of CDHW events between different ENSO phases.
Scope of the research
This study analyzes the spatial and temporal characteristics of compound drought and heatwaves (CDHW) in Southeast Asian countries, using high-resolution reanalysis data CHIRTS and CHIRPS to assess droughts and heatwaves from 1983 to 2016 Additionally, the Oceanic Niño Index (ONI) is employed to identify ENSO phases, enabling evaluation of ENSO's influence on CDHW.
Framework of research
Asia has reportedly experienced an increase in the frequency and intensity of extreme weather events, including droughts and heatwaves, often exacerbated by phenomena such as
H1: Based on the indices used to identify drought and heatwave events, it is possible to determine CDHW
H2: The weather and climate of the SEA region are influenced by various climate drivers, resulting in a complex distribution of CDHW events
Amid ongoing climate change, weather conditions are increasingly variable and shaped by a range of climatic phenomena Among these, ENSO—the El Niño-Southern Oscillation—significantly influences temperature and rainfall patterns in Southeast Asia, and it exhibits a notable relationship with CDHW.
O1: To identify and characterize CDHW events based on existing methodologies and established drought and heatwave indices
This study analyzes the spatial and temporal characteristics of CDHW events in Southeast Asia (SEA) over the past several decades It also investigates the influence of ENSO on the characteristics of these CDHW events across SEA during the same period.
Using high resolution reanalysis data of daily minimum, maximum temperature from CHIRTS and precipitation from CHIRP to identify and calculate CDHW’s characteristics events over SEA during 1983 -2016
Data
This study uses two gridded datasets to assess CDHW events: daily maximum and minimum surface temperatures from CHIRTS (Climate Hazards Center InfraRed Temperature with Stations) and daily precipitation from CHIRPS (Climate Hazards Group InfraRed Precipitation with Stations) Both datasets offer a 0.05° × 0.05° spatial resolution and serve as historical observations for the analysis from 1983 to 2016 The data cover 11 Southeast Asian countries, and this thesis specifically evaluates compound CDHW events for these 11 nations.
CHIRTS-daily, developed by the Climate Hazards Center (CHC), is a high-resolution daily maximum and minimum temperature dataset at 0.05° × 0.05° covering 60°S to 70°N for 1983–2016, available at www.chc.ucsb.edu/data/chirtsdaily It leverages the relative monthly variability of daily temperatures from ERA5 to build accurate daily estimates, with CHIRTS-daily’s native resolution finer than ERA5’s 0.25° × 0.25° To merge the two sources, ERA5 data are downscaled via bilinear interpolation in the Interactive Data Language (IDL7) using the CONGRID command Studies show CHIRTS-daily aligns more closely with observational data for extreme temperatures than ERA5’s global reanalysis, as noted by Parsons et al (2022).
Amou et al (2021) indicated that the phenomenon of heatwaves and their impacts have been inadequately considered due to several factors, including unreliable datasets and inconsistencies between meteorological data and heatwave detection metrics Through their evaluation, this study utilized the CHIRTS dataset to investigate heatwaves in Kenya from 1987 to 2016 Iris et al (2021) assessed the quality of reanalysis and satellite-based datasets (CHIRPS/CHIRTS, ERA5, MERRA-2, PERSIANN-CDR), demonstrating that CHIRPS aligns well with Global Precipitation Climatology Center’s (GPCC) precipitation data over Central America Pyarali et al
CHIRTS data show close agreement with observed data, particularly in regions with sparse gauge information, and effectively capture heatwave trends Amou et al (2021) used CHIRTS daily maximum temperatures to calculate the Heatwave Magnitude Index daily (HWMId), defined as the maximum intensity of heatwave events within a year, where a heatwave is a period of at least three consecutive days with daily maximum temperatures exceeding the 90th percentile of a reference period Li et al (2022) utilized CHIRTS to identify heatwaves based on relative thresholds, specifically TX90pct and TN90pct, representing the 90th percentile for each day of the year, calculated using a 15-day moving window centered on the day of interest for daily maximum and minimum temperatures, respectively Heatwaves are identified when temperatures exceed the corresponding thresholds for a minimum of three consecutive days, yielding thresholds that are relative to both the time of year and geographical location.
In 2024, researchers estimated potential evapotranspiration (PET) from daily minimum and maximum temperatures and then used PET to compute the Standardized Precipitation-Evapotranspiration Index (SPEI), providing a robust basis for drought assessment and monitoring in West Africa.
The UC Climate Hazards Group (CHG) released CHIRPS version 2.0, a high‑resolution, quasi‑global rainfall dataset that blends station data with satellite imagery to produce gridded precipitation time series Spanning 1981 to the present and covering 50°S to 50°N, CHIRPS provides 0.05° x 0.05° resolution rainfall data similar to CHIRTS Data processing draws on multiple private archives plus five public sources—SASSCAL, GTS, GHCN monthly, GHCN daily, and GSOD—with GTS updated daily and the other sources updated monthly Supplementary monitoring data are also contributed by national meteorological agencies, especially in Mexico, Central America, South America, and sub‑Saharan Africa.
Since 2015, the Climate Hazards Group (CHG) has accumulated nearly 1.2 billion daily observations from more than 200,000 locations, forming a robust data foundation for the CHIRPS rainfall product Numerous studies have demonstrated the suitability of CHIRPS for drought assessment (Didi et al 2020, Du et al 2024).
Within the Vietnamese region, Le et al (2020) found that CHIRPS shows only a small deviation from observed data and is advantageous for long-term water-resource planning and drought preparedness In examining high-resolution global drought indices, Gebrechorkos et al (2023) highlighted CHIRPS’s advantage due to its homogeneity The CHIRPS dataset is built from a continuous series of thermal infrared observations from geostationary satellites, enabling consistent precipitation monitoring Guo et al (2017) note that CHIRPS can adequately capture drought characteristics at multiple scales across the Mekong region, including Vietnam Additionally, CHIRPS precipitation data are frequently used together with evaporation data from the Global Land Evaporation Amsterdam Model (GLEAM) to compute the Standardized Precipitation-Evapotranspiration Index (SPEI) (Peng et al.).
2020, Gebrechorkos et al., 2023, Pyarali et al., 2022) or to compute the Standardized Precipitation Index (SPI) (Guo et al., 2017).
Method
Drought
Regarding drought identification, the Standardized Precipitation Index (SPI) quantifies observed precipitation as a standardized departure from a chosen probability distribution, offering a simpler calculation than the Palmer Drought Severity Index (PDSI) and the Standardized Precipitation Evapotranspiration Index (SPEI) (McKee et al., 1993; Palmer, 1965; Vicente-Serrano et al., 2010) However, SPI relies solely on precipitation and neglects other drought drivers such as temperature and soil moisture By contrast, PDSI is more complex to compute over large areas because it depends on precipitation, temperature, soil moisture, and potential evapotranspiration (PET) SPEI, which uses precipitation and PET, provides a more comprehensive drought assessment In this study, SPEI (Vicente-Serrano et al., 2010) is calculated with PET estimated by Hargreaves & Samani (1985) to identify drought occurrences at a 90-day timescale, defining a drought period when SPEI is below -1.
The calculation of the Standardized Precipitation-Evapotranspiration Index (SPEI) involves the following steps:
To quantify the daily gap between precipitation and potential evapotranspiration, we derive Pdaily from CHIRPS precipitation data and compute daily mean temperature from CHIRTS data The daily mean temperature, together with the grid-point latitude, is then used to estimate PETdaily using the Hargreaves & Samani (1985) method This approach produces a gridded, day-by-day PET and precipitation dataset that supports moisture-balance analyses and regional comparisons of water availability.
The daily difference (Ddaily) is then calculated as Ddaily=Pdaily−PETdaily
Then, the three-parameter log-logistic distribution is fitted to the accumulated precipitation-evapotranspiration differences, and these are subsequently transformed into a standard normal distribution
Probability-weighted moments (PWMs) are calculated using the "alpha PWMs” following Hamed, K., & Rao, A.R (2000) resulting in PWMs = [w0, w1, w2]
From these, the L-moments [γ,α,β] are computed Subsequently, the cumulative distribution function (CDF) of the three-parameter log-logistic distribution is calculated, where 'x' in the formula represents Ddaily
Where Г(β) is the gamma function of β
The SPEI can be easily derived by standardizing the values of F(x):
Let W be computed as W = -2 ln(P) for P ≤ 0.5, where P denotes the probability of exceeding a given drought value D, with P = 1 − F(x) If P > 0.5, this probability is replaced by 1 − P, and the sign of the resulting SPEI (Standardized Precipitation-Evapotranspiration Index) is inverted The constants used are defined within the model.
Heatwaves
In this study, heatwave events are identified using two distinct approaches The first approach aligns with research conducted in Southeast Australia, where experts advocate using the Excess Heat Factor (EHF) as the metric to identify heatwaves, as described by Nairn et al.
A 2009 study defines a heatwave as at least three consecutive days with the Extreme Heat Factor (EHF) H1 component greater than zero To calculate EHF, the approach follows Nairn’s framework by incorporating two additional indices: Excess Heat and Heat Stress The Excess Heat Index (EHI) is defined as the average of the daily maximum and minimum temperatures over a three-day period, relative to the 95th percentile of daily temperatures for the climate reference period.
The second index, heat stress, compares the average temperature of three consecutive days with the average temperature of the previous 30 days This index represents a short-term (acclimatization) temperature anomaly
The combination of Excess Heat and Heat Stress as an index of Excess Heat Factor provides a measure of heatwave event on severity, duration, and spatial distribution
The second way to identify a heatwave, suggested by Perkins & Alexander
The 2013 study employs relative thresholds to define extreme temperature events It uses TX90pct, the calendar-day 90th percentile derived from a 15-day moving window centered on the day of interest for daily maximum temperatures This thresholding approach allows the threshold to vary with season and location A heatwave is defined as the exceedance of these thresholds for at least three consecutive days.
Compound drought and heatwaves (CDHW)
This study identifies events in Southeast Asia where heatwaves and droughts occur simultaneously A heatwave is defined as a period of at least three consecutive days with either EHF > 0 or Tx > the 90th percentile of daily Tx Droughts are identified based on the SPEI value When both heatwave and drought thresholds are exceeded for three or more consecutive days, the event is classified as a CDHW event.
For example, Figure 2.1 illustrates the occurrences of CDHW events identified by SPEI < -1 and Tx > Tx90pct during the period from day 101 to day 150 of the year
1983 The blue line represents SPEI, the orange line represents (Tx - Tx90pct), and the red line indicates SPEI = -1
Figure 2.1 Compound drought and heatwave events identified based on SPEI and Tx at the location (105°E, 15°N) from day 101 to day 150 of the year 1983
During this period, the location registered three CDHW events with durations of 7, 3, and 3 days, respectively While the CDHW criteria were met around day 120, the event lasted fewer than three days and was not classified as a CDHW event.
Figure 2.2 illustrates the occurrences of CDHW events identified by SPEI < -1 and EHF > 0 during days 101 to 150 of 1983 The blue line represents SPEI, the orange line represents EHF, and the red line marks SPEI = -1 The results indicate that during this period the location experienced a single CDHW event lasting 7 days.
Figure 2.2 Compound drought and heatwave events identified based on SPEI and EHF at the location (105°E, 15°N) from day 101 to day 150 of the year 1983
Following the calculation of CDHW identification, the importance of each method will be examined in further detail after calculating CDHW’s characteristics
The characteristics of CDHW will be calculated are: (1) CDHWF; (2) CDHWN;
1) CDHWF: The mean frequency of CDHW events
2) CDHWN: The mean number of CDHW’s days
3) CDHWD: The mean duration of the CDHW events
4) CDHWL: The mean duration of the longest CDHW event of the years
5) CDHWS: The mean severity of CDHW events
If a heatwave is identified through EHF, the severity will be calculated following Laz et al 2023 The severity of CDHW is identified by:
If a heatwave is identified through Tx, the severity will be calculated as follows: CDHWS2 = - Drought index × (Tx – The 90th percentiles of daily Tx)
6) CDHWM: the mean severity of the most severe CDHW event of the years
Additionally, the entire time series from 1983 to 2016 will be divided into four periods to evaluate the temporal increase or decrease of these two characteristics Period
1 spans 1983–1991, Period 2 spans 1992–2000, Period 3 spans 2001–2008, and Period
From 2009 to 2016, the study first divides the data into four sub-periods After this four-segment division, the time series is further split into two broader periods to enable a direct comparison between 2000–2016 and 1983–1999 In preparing the 90-day SPEI drought index, the analysis excludes the first 90 days of 1983 This two-tier segmentation supports a consistent, scale-aware assessment of drought dynamics across the selected historical windows.
The different between the period from 2000–2016 (X) compared to the period from 1983–1999 (Y) is tested by Mann–Whitney U (Mann & Whitney, 1947) Where:
H0: X and Y have the same distribution
H1: X tends to be larger than Y
Statistically significant are marked by hatching when p_value < 0.05
Details of the application of Mann–Whitney U are as follows:
The study supposed that X has n1 samples, Y has n2 samples Combine the two groups and rank them in ascending order, then calculate the rank of each group:
Calculate the standard deviation of U:
P_value calculated from the standard normal distribution N (0,1).
ENSO years
This thesis compares the characteristics of CDHWF and CDHWN across El Niño, La Niña, and neutral years to assess how ENSO influences the increase or decrease of CDHW To evaluate ENSO’s effects on CDHW characteristics, the study employs the Oceanic Niño Index (ONI) to identify El Niño years, La Niña years, and neutral years, using ONI as the primary metric to categorize these ENSO phases.
El Niño and La Niña events are tracked using the Niño 3.4 region (5°N–5°S, 120°W–170°W) and the Ocean Niño Index (ONI) An event is defined as at least five consecutive ONI values with anomalies at or above +0.5°C for warm El Niño conditions, or at or below −0.5°C for cool La Niña conditions.
Extensive studies have shown that ENSO, associated with sea surface temperature anomalies, exerts a lagged influence on temperature and precipitation in Southeast Asia In this thesis, a warm (cold) ENSO year is identified when the five-month average of the Oceanic Niño Index (ONI)—covering October–November–December (OND) of the previous year and January–February (JF) of the year under study—exceeds +0.5°C (or falls below −0.5°C).
Using the Oceanic Niño Index (ONI), this study investigates the impact of the El Niño–Southern Oscillation (ENSO) on the occurrence of the CDHW event The ONI values for each year are arranged in ascending order This ascending ONI-based year ranking is then used to reorder the 34-year period from 1983 to 2016 Finally, linear regression coefficients are calculated for each grid point using the reordered 34-year series.
This study examines how the characteristics of El Niño years (X) differ from those of La Niña years and Neutral years (Y), and it also compares Neutral years (X) with La Niña years (Y) using the Mann–Whitney U test to determine whether the observed differences are statistically significant.
H0: X and Y have the same distribution
H1: X tends to be larger than Y
Statistically significant are marked by hatching when p_value < 0.05
CDHW’s characteristics in SEA
Figure 3.1 shows CDHWF from 1983 to 2016 across the computational domain covering 11 Southeast Asian countries included in this analysis, evaluated with two criteria: (a) SPEI < -1 and EHF > 0, and (b) SPEI < -1 and an additional condition This dual-criterion approach highlights the spatial and temporal development of drought vulnerability and hydrological drought risk throughout Southeast Asia during the study period.
Figure 3.1 The mean frequency of CDHW events from 1983 to 2016 (CDHWF) a) SPEI < -1 and EHF > 0, b) SPEI < -1 and Tx > Tp90pct
Two subfigures use separate color scales to reflect differences in frequency ranges Overall, the spatial distribution patterns of CDHWF are comparable between the EHF-based and Tx-based definitions, yet CDHWF values are higher when Tx is used, indicating that CDHW defined by SPEI and Tx occurs more frequently Regions between 10°S and 5°N—excluding several central Indonesian islands—show relatively low CDHWF, generally below 0.6 events per year with EHF and below 1.2 events per year with Tx By contrast, areas north of 5°N exhibit higher CDHWF, ranging from 0.4 to 1.0 events per year under EHF and 0.8 to 2.0 events per year under Tx Notably, eastern Thailand near the borders with Laos and Cambodia emerges as a hotspot where compound drought and heatwave events occur most frequently.
Figure 3.2, continuing the trend from the previous figure, illustrates CDHWN and reveals a consistent spatial distribution between the EHF- and Tx-based definitions Nevertheless, the CDHWN values derived from the Tx definition are generally higher than those obtained with the EHF definition, indicating a systematic difference in magnitude while preserving overall spatial patterns.
Figure 3.2 The mean number of CDHW’s days from 1983 to 2016 (CDHWN), a) SPEI < -1 and EHF > 0, b) SPEI < -1 and Tx > Tp90pct
Across most of Southeast Asia, CDHWN values range from about 1 to 4 days per year when calculated with the EHF, and from 2 to 8 days per year when using the Tx The CDHWN pattern follows the frequency shown in figure 3.1, with regions that have higher CDHWF also experiencing higher CDHWN In Indonesia, the eastern, western, and central western regions show the lowest values, with fewer than 1 day per year on the EHF and under 2 days per year on the Tx By contrast, Thailand, Laos, and Cambodia exhibit higher CDHWN levels, where CDHWs occur around 3–5 days per year under the EHF and 6–10 days per year under the Tx.
Figure 3.3, similar to Figure 3.2, illustrates CDHWD Under the EHF-based definition, CDHW events typically last between 3 to 6 days
Figure 3.3 shows the mean duration of CDHW events (CDHWD) from 1983 to 2016, with panel a for SPEI < -1 and EHF > 0 and panel b for SPEI < -1 and Tx > Tp90pct The highest CDHWD values occur mainly over central Indonesia and Timor-Leste, exceeding 6 days In contrast, northern Vietnam and the Sarawak region show shorter durations.
Across Malaysia, western and central Kalimantan, and the Papua region of Indonesia, event durations are relatively uniform, typically 3-4 days per event CDHWD defined using Tx lasts on average 3-5 days, with very few instances exceeding 6 days, and these durations are shorter than those identified using EHF This suggests that although the Tx-based method yields higher CDHWF and CDHWN in many areas, the duration of individual events is generally shorter compared to those defined by EHF Regions north of latitude 5°N show longer average durations, ranging from 4-6 days per event, than those in the south, where durations typically fall between 3-4 days per event.
Figure 3.4, like Figure 3.2, illustrates CDHWL When using EHF, CDHWL typically ranges from 4 to 8 days per event across most parts of Southeast Asia CDHWL exceeding 8 days is primarily observed in central Indonesia, with additional occurrences in southern Vietnam, where CDHWL also surpasses 8 days per event.
Figure 3.4 The mean duration of the longest CDHW event of the years from 1983 to
2016 (CDHWL), a) SPEI < -1 and EHF > 0, b) SPEI < -1 and Tx > Tp90pct
Under the Tx-based definition, CDHWL similarly ranges from 4 to 8 days However, compared to Tx, EHF-based CDHW more frequently exhibits durations in the
6 to 8 days range, indicating that EHF tends to capture longer-lasting compound events more often than Tx
Figure 3.5, similar to Figure 3.2, illustrates CDHWS When using EHF, the spatial distribution of CDHWS is more pronounced
Figure 3.5 shows the mean severity of CDHW events (CDHWS) from 1983 to 2016 under two conditions: SPEI < -1 with EHF > 0 (panel a) and SPEI < -1 with Tx > Tp90pct (panel b) Under the EHF-based criterion, CDHWS values exceed 8°C² mainly in Laos, northern Myanmar, Thailand, Cambodia, and Vietnam, while most other countries remain lower, with many areas below 4°C²; notably, northern Vietnam records severity above 18°C² despite relatively low CDHWF and CDHWN By contrast, with the Tx-based threshold, most CDHWS values stay under 10°C, indicating generally lower severity than the EHF-based method, although northern Vietnam and Myanmar still show higher levels in the 10–14°C range Overall, the EHF approach highlights a clearer contrast between high and low severity regions than the Tx-based method.
Figure 3.6, like Figure 3.4, illustrates CDHWM and shows that CDHWM remains generally below 25°C² under the EHF-based definition and below 15°C under the Tx-based definition Furthermore, regions with severity values below 5°C² are more prevalent in low-latitude countries when using the EHF-based approach.
Figure 3.6 The mean severity of the most severe CDHW event of the years from 1983 to 2016 (CDHWM), a) SPEI < -1 and EHF > 0, b) SPEI < -1 and Tx > Tp90pct
Results show that the number of CDHW events, defined by CDHWF and CDHWN, is higher when heatwaves are defined using Tx than with the EHF index This finding is in part consistent with Perkins et al (2012), who analyzed global heatwave characteristics from 1950 to 2011 The underlying reason is that the EHF index is referenced against a climatological threshold, which tends to capture anomalous conditions primarily during warmer periods, thereby reducing the likelihood of detecting unusual warm events outside the summer Furthermore, Perkins et al (2012) and the accompanying figure illustrate these relationships.
6 demonstrate that EHF consistently shows the longest duration of CDHW across many regions
Figures 3.5 and 3.6 show that CDHW events identified by the EHF framework have higher severity values than those identified by Tx, reflecting the role of human resilience in heat exposure The EHF formula incorporates the background temperature that people must adapt to over roughly 30 days, so ongoing warming leads to greater severity In contrast, Tx merely flags the occurrence of extreme heat without accounting for adaptability, effectively measuring only the discomfort caused by heat waves.
Figures 3.5 and 3.6 show that low-latitude regions exhibit lower climate severity than higher-latitude areas The underlying thesis attributes this to the relatively constant solar radiation these regions receive throughout the year, which minimizes temperature variability Consequently, temperature differences are less pronounced, leading to milder environmental fluctuations This steadier thermal profile enhances human adaptability and contributes to the overall lower severity observed in low-latitude regions.
CDHW exhibits a degree of similarity with both the EHF and Tx methods, suggesting shared dynamics among these approaches To evaluate the rising and falling trends of CDHW during 1983–2016 and to assess the influence of ENSO on this phenomenon, the study conducts a more in-depth analysis of the CDHWF and CDHWN characteristics using the Tx method.
The evolution of CDHW from 1983 to 2016
Figure 3.7 depicts the period-to-period changes in CDHWF values Positive differences, highlighted with warm colors, indicate that CDHWF increases in the later period compared with the preceding one, while negative differences, represented by cool colors, indicate a decrease.
Across Southeast Asia, CDHWF shows an overall upward trend from 1983 to 2016, with P21, P43, and P41 registering positive values in most areas This increase, however, is not continuous: Figure 3.7.b documents a slight decline in CDHWF across most regions during 2001–2008 compared with 1992–2000 Despite the dip, a pronounced rise in CDHWF—about 2 events per year—occurs in central Indonesia and across large parts of Vietnam, Laos, Cambodia, Thailand, and Myanmar While additional regions also exhibit increasing trends, their magnitudes are generally smaller, typically less than 0.8 events per year.
Figure 3.7 The variation in CDHWF values across the periods from 1983 to 2016 a) 1992-2000 relative to 1983-1991 (P21), b) 2001–2008 relative to 1992-2000 (P32), c) 2009–2016 relative to 2001–2008 (P43), d) 2009–2016 relative to 1983-1991 (P41)
Figure 3.8 closely resembles Figure 3.7 while highlighting differences in CDHWN values across different periods The spatial pattern of CDHWN mirrors the distribution of CDHWF depicted in Figure 3.7 Regions where the number of events increases (or decreases) correspond to areas where the number of days likewise increases (or decreases).
Figure 3.8 The variation in CDHWN values across the periods from 1983 to 2016, a) 1992-2000 relative to 1983-1991 (P21), b) 2001–2008 relative to 1992-2000 (P32), c) 2009–2016 relative to 2001–2008 (P43), d) 2009–2016 relative to 1983-1991 (P41)
Overall, an increasing trend in CDHW is dominant during the period from 1983 to 2016, except for a slight decrease observed during 2001–2008 compared to 1992–
In 2000, analysis shows that days per year increased by more than 8 in areas near central Indonesia and across parts of Vietnam, Laos, Cambodia, Thailand, and Myanmar, while in other regions the annual increase ranges from about 2 to 6 days, indicating regional variations in the trend.
Figure 3.9 The variation in CDHWS values across the periods from 1983 to 2016, a) 1992-2000 relative to 1983-1991 (P21), b) 2001–2008 relative to 1992-2000 (P32), c) 2009–2016 relative to 2001–2008 (P43), d) 2009–2016 relative to 1983-1991 (P41)
An assessment of the severity of compound droughts and heatwaves (CDHWs) across Southeast Asia for 1983–2016 reveals a clear, though nuanced, trend toward stronger events, with substantial variation by period and sub-region In aggregate, CDHW severity intensified across Southeast Asia Following a phase of neutral conditions or only minor gains in the southern islands (P32: 2001–2008 vs 1992–2000), the pattern then shifted to a generally moderate increase (P43: 2009–2016 vs 2001–2008).
Among mainland Southeast Asia, Vietnam, Laos, Thailand, and Myanmar show the most pronounced changes, with a substantial net increase in CDHW (compound drought and heatwave) severity despite ongoing fluctuations In the most recent period examined (P43: 2009–2016 vs 2001–2008), this intensification frequently exceeds 3 to 5 units Even with inherent variability, the overall trend indicates that Southeast Asia faces a much higher risk of more severe CDHW events in the coming decades, with the mainland recording a particularly severe and recent increase.
According to Sourav Mukherjee and Ashok Kumar Mishra (2020), in a warming world CDHWF increases by about 1 to 3 events per year and CDHWN by roughly 2 to 10 days per year, findings that align closely with the results of this thesis However, the period 2001–2008 witnessed declines in both indicators across most areas of Southeast Asia relative to 1992–2000 While multiple factors may contribute to this pattern, the role of ENSO—a dominant climate driver that strongly influences temperature and precipitation in Southeast Asia—warrants particular attention Notably, 2001–2008 was characterized by a prolonged La Niña beginning in 2000 and a very strong La Niña episode in 2007–2008, with no strong El Niño events during this period, whereas two very strong El Niño events occurred during 1992–2000.
To minimize the influence of ENSO variability, the study splits the analysis into two phases As shown in Figure 3.10, the interval 2000–2016, relative to 1983–1999, is marked by a pronounced intensification of compound drought and heatwave events across most of Southeast Asia.
Figure 3.10 The variation of CDHW’s characteristics from 2000-2016 relative to
Most of Southeast Asia experienced an increase in the occurrence of these climate events Moderate increases of about 0.4 to 0.8 events per year are evident in mainland Southeast Asia, including Myanmar, Laos, and Cambodia, while Vietnam shows a mixed pattern with a tendency toward a significant rise in the northern region The Philippines, by contrast, witnessed a decrease in frequency In Myanmar and Laos, event days increased by roughly 2 to more than 5 days per year, indicating that these events are becoming not only more frequent but also longer lasting when comparing 2000–2016 with 1983–1999 Southeast Asia also shows a widespread and often stronger rise in severity, with increases of 1 to 3 units in much of mainland Southeast Asia, including Vietnam, Laos, Thailand, and Myanmar Taken together, these trends point to a growing climate challenge for the region and a higher risk of concurrent drought and extreme heat in recent decades.
The connection between CDHW characteristics and ENSO
Figure 3.11 displays how CDHWF values differ across ENSO phases Positive values, shown in warm colors, indicate higher CDHWF during El Niño years than during La Niña years, while negative values, shown in cool colors, indicate the opposite, with the same interpretation applying to comparisons between other ENSO phase pairs.
During La Niña years (Figure 3.11a), almost the entire Southeast Asia region exhibits positive CDHWF values, with small areas in the western part of central Indonesia, western Malaysia, and northern Myanmar showing negative anomalies Regions with significantly higher CDHWF during El Niño years occur at roughly two events per year and are observed over central Indonesia, most of Malaysia, the Philippines, Vietnam, Laos, Cambodia, and Thailand A similar pattern is evident when comparing El Niño and Neutral years; however, in this case, regions with strongly positive CDHWF anomalies exceed a defined threshold.
CDHW events peak during El Niño years, with about two events per year becoming localized in certain regions, while areas showing slight negative anomalies (0 to −0.6 event/year) are more common in western Indonesia Compared to Neutral and La Niña years, the differences are smaller, with CDHW changes typically between 0 and 1 event/year across Southeast Asia Slight negative anomalies of around −0.2 event/year are observed over northern Vietnam, Laos, and Myanmar Overall, CDHW events occur most frequently in El Niño years, less frequently in Neutral years, and least frequently in La Niña years.
Figure 3.11 displays the differences in CDHWF values across ENSO years from 1983 to 2016 Panel a compares El Niño years with La Niña years, illustrating the differential CDHWF outcomes under warm-phase versus cool-phase ENSO conditions Panel b contrasts El Niño years with Neutral years to show how CDHWF shifts when transitioning from neutral to El Niño conditions Panel c compares Neutral years with La Niña years, detailing the CDHWF differential between neutral and La Niña regimes Collectively, the figure underscores the ENSO-driven variability of CDHWF over the 1983–2016 period.
Figure 3.12 mirrors Figure 3.11 by illustrating how the CDHWN values vary across ENSO phases, with the CDHWN pattern aligning with the spatial distribution of CDHWF shown in Figure 3.11 Regions with a higher number of events correspond to more CDHWN days, while areas with fewer events show fewer CDHWN days During El Niño years, CDHWN is generally more than 4 days per year higher than during La Niña years, with several regions exceeding 10 days Compared to Neutral years, regions with positive CDHWN differences still dominate, with CDHWN values during El Niño years remaining higher in many zones.
El Niño years typically increase CDHWN by 2 to 10 days per year, but in some regions CDHWN during El Niño years can decrease by 0 to 4 days per year, with these reductions occurring more frequently over western Indonesia The difference in CDHWN between Neutral and La Niña years ranges from -2 to 6 days per year, with negative values primarily observed in northern regions.
Figure 3.12 shows the differences in CDHWN values across ENSO years from 1983 to 2016, with three panels: a) El Niño years relative to La Niña years, b) El Niño years relative to Neutral years, and c) Neutral years relative to La Niña years The figure highlights how CDHWN responds differently during El Niño events compared with La Niña events, indicating both the magnitude and direction of the contrasts It also demonstrates how El Niño years compare to Neutral years, capturing the shift in CDHWN when transitioning from neutral to El Niño conditions Finally, the comparison of Neutral years relative to La Niña years reveals baseline differences in CDHWN under La Niña influence Together, these panels summarize the ENSO-related variability of CDHWN over the 1983–2016 period, providing a clear basis for interpreting ENSO-phase effects in the data.
El Niño years bring a marked rise in CDHWF and CDHWN across most of Southeast Asia when compared with La Niña and Neutral years The region’s precipitation and temperature variability is shaped by major teleconnections including ENSO, the Pacific Decadal Oscillation (PDO), the Indian Ocean Dipole Mode (IOD), and the North Atlantic Oscillation (NAO), with ENSO exerting the strongest influence Hao et al (2018) show global-scale evidence of strong negative correlations between precipitation and NINO34 and significant positive correlations between temperature and NINO34, particularly in Southeast Asia, where compound dry and hot events are more likely during periods of high NINO34 values (El Niño).
Figure 3.13 The difference in CDHWS values in ENSO years from 1983 to 2016, a) El Niủo years relative to La Niủa years, b) El Niủo years relative to Neutral years, c) Neutral years relative to La Niủa years
One notable finding is the significantly heightened severity of compound drought and heatwave events during El Niủo years across extensive parts of mainland Southeast Asia, including the Indochina Peninsula, as well as large areas of the Philippines These regions are characterized by strong positive values, often exceeding +3 to +5 °C/year, indicating a substantial exacerbation of extreme conditions under El Niủo In contrast, the maritime continent, particularly Indonesia and surrounding islands, displays a more heterogeneous response, with some areas showing a positive (El Niủo-induced increase in severity), neutral, or even slightly negative differential Limited areas, such as parts of southern China, exhibit negative values, suggesting a reduced impact of El Niủo or a comparatively greater severity during La Niủa periods for these specific locales
Most notably, the Indochina Peninsula (encompassing significant portions of
El Niño episodes heighten the risk of compound drought and heatwaves across Southeast Asia and the Philippines, with deep red and purple shading indicating strong positive anomalies and substantially greater severity during El Niño than during ENSO-Neutral periods The cross-hatched areas emphasize these zones of intensified impact Northern Borneo shows a clear positive anomaly, pointing to increased drought and heatwave severity in El Niño years In contrast, southern China and northern Vietnam (north of about 20°N) exhibit negative anomalies, suggesting that these northern regions experience less severe compound drought and heatwaves during El Niño than in Neutral conditions Parts of New Guinea and several eastern Indonesian islands display negative or near-neutral values The core maritime continent—Sumatra, Java, and Peninsular Malaysia—generally shows a slight positive to near-neutral difference, meaning El Niño conditions are only marginally more severe or comparable to Neutral conditions for these compound events.
Recent analysis shows a strong signal that Neutral years produce greater severity of compound drought and heatwaves than La Niña years across much of mainland Southeast Asia, notably the Indochina Peninsula (Thailand, Laos, Cambodia, and southern Vietnam), and the Philippines These regions, often cross-hatched in the data, indicate Neutral conditions have a substantially larger impact than La Niña on drought-heatwave events A similar, pronounced positive difference is observed in Papua New Guinea and nearby eastern Indonesian islands By contrast, northern Vietnam shows negative values, indicating La Niña years bring greater drought-heatwave severity there Across most of the maritime continent—Sumatra, Java, Borneo, and Peninsular Malaysia—the differences are more moderate, with Neutral conditions being slightly more severe than, or roughly comparable to, La Niña conditions.
Figure 3.14 displays the linear regression coefficients for CDHWF frequency at each grid point, with years ordered by ascending ONI values Warm-colored regions indicate that increasing ONI values, signaling stronger El Niño events, are associated with a higher frequency of CDHWF Cool-colored regions show that higher ONI values correspond to a lower occurrence of CDHWF.
Figure 3.14 Linear regression coefficient of CDHWF on each grid point in increasing order of ONI
Across Southeast Asia, most areas show warm colors corresponding to linear regression coefficients above 0.05, signaling a broad rise in CDHWF as El Niño intensifies A few isolated regions—primarily in eastern and western Indonesia—display cool colors with slightly negative coefficients Coefficients exceeding 0.1 appear clearly over Thailand, Laos, Cambodia, and the Philippines, indicating the strongest increase in CDHWF during stronger El Niño events.
Figure 3.15 is similar to Figure 3.14 but illustrates the CDHW number (CDHWN)
Figure 3.15 Regression coefficient of CDHWN on each grid point in ascending order of ONI
Linear regression coefficients are generally above 0.2, with some areas in eastern and western Indonesia displaying coefficients from -0.1 to 0.2, consistent with Figure 3.14 As ONI values rise, signaling a transition from the El Niño–Southern Oscillation toward the El Niño phase, CDHWN shows an increasing trend across most of Southeast Asia, most pronounced in Thailand, Laos, Cambodia, and the Philippines.