7.1 Introduction

Droughts are generally defined as prolonged periods with well below normal rainfall in a region, leading to an extreme shortage of water (Gonzalez-Hidalgo et al., 2009), and floods are defined as intensive rainstorms which can quickly increase streamflow in rivers where excessive water overflows out of the river and leads to surface water flooding. The flood process has several characteristics, including the flood peak, flood volume and flood duration. The frequency of those characteristics during the flood process are often different but mutually correlated (Luo et al., 2021).

Droughts and floods have received widespread attention because they could have substantial social and economic consequences. With the unprecedented impact of climate change and human activities in recent decades, the terrestrial water cycle is becoming non-stationary, leading to more extreme events. This will affect the hydrological characteristics of rivers, with impacts on industrial, economic and social development. As one of the most important transboundary rivers in the world, the Lancang Mekong River Basin (LMRB), especially the Mekong River Basin (MRB), suffers frequent hydrological disasters. According to the Emergency Events Database (EM-DAT), the LMRB recorded 173 floods and 23 droughts between 1990 and 2016, affecting 148.5 million people and causing a total economic loss of 61.4 billion US dollars.

At the same time, the LMRB will face two major challenges in the twenty-first century: changes to hydrological characteristics brought by climate change, and the impact of rapid hydropower expansion on extreme events (Liu et al., 2022). A significant increase in basin-wide temperatures and changes in monsoon patterns has been suggested by future climate projections, which is expected to cause an increase in extreme rainfall intensity and frequency and ultimately drive flood changes in the basin. Meanwhile, rapid hydropower expansion at an unprecedented rate in this century also has a huge impact on the LMRB. Considering that this change will affect the lives of nearly 237 million people, it is necessary to comprehensively review the combined impacts of climate change and reservoir operation on floods and droughts in the LMRB.

This chapter aims to provide a comprehensive review of the impact of climate change and reservoir expansion on floods and droughts in the LMRB. Section 7.2 explains the characteristics of LMRB droughts and floods during the historical period, and Sect. 7.3 explores the influence of future climate change on droughts and floods. Section 7.4 reviews the impact of reservoir operation on droughts and floods, and discusses the potential for reservoirs to adapt to and mitigate extreme events.

7.2 Characteristics of Droughts and Floods in the LMRB

7.2.1 Drought Characteristics

Drought is a natural disaster, defined as a significant deviation from normal hydrologic conditions such as rainfall, soil moisture or runoff (Mishra & Singh, 2010). Generally, droughts can be divided into four categories according to their nature and effects as meteorological, agricultural, hydrological and socio-economic droughts. Drought is one of the most serious disasters in the Mekong River Basin (Zhang et al., 2020a, 2020b) and the middle LMRB is trending to intensified drying (Fig. 7.1). Since the beginning of the twenty-first century, the Mekong River has been affected by several major drought events (Guo et al., 2017; Son et al., 2012). Understanding the spatial and temporal characteristics of drought can greatly facilitate drought management and risk reduction.

Fig. 7.1
A color-coded map of the Lower Mekong River Basin depicts the Standardized Precipitation Index change trend from 1979 to 2019. The colors indicate different levels of precipitation changes from drier to wetter areas.

Trend of Standardized Precipitation Index (SPI) change from 1979 to 2019 in the LMRB. Data was obtained from the European Centre for Medium-Range Weather Forecasts Reanalysis v5 (ERA5)

Full size image

7.2.1.1 Spatial Characteristics of Drought in LMRB

Drought in the Mekong River has apparent spatial heterogeneity and high correlation with latitude (Li & Chen, 2015; Li et al., 2013). Based on 0.25° × 0.25° resolution daily precipitation data from the Global Land Surface Data Assimilation System (GLDAS), Zhang et al., (2020a, 2020b) found that high-incidence areas of extreme agricultural drought were in Yunnan Province, China and northwestern Thailand and high-vulnerability areas were distributed in the middle and southern LMRB. Combining high incidence with high vulnerability, the middle of the LMRB and the Sesan, Srepok and Sekong river basins (3S) are high-risk areas of agricultural drought. Liu et al. (2020) found that dry extreme events have increased significantly over the northeastern Thailand, most of Cambodia and Myanmar, particularly for southern Cambodia and the Mekong Delta where the frequency of extreme drought is around 10%. In general, the occurrences of drought events are mostly in the lower LMRB, followed by the upper and middle LMRB (Tang & Cao, 2020).

Through principal component analysis and K-means clustering of daily precipitation observation data from 35 weather stations from 1960 to 2005, it was found that there is a strong linkage between climate zones and spatial characteristics of drought, while the high-risk areas of drought are mainly located in the middle and southern LMRB (Li et al., 2013). Because rainfall stations in the LMRB are sparse and unevenly distributed, satellite precipitation data are usually used as alternative sources. The satellite precipitation products from the Tropical Rainfall Measuring Mission (TRMM) and Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) have been proven to be reliable data sources for studying drought in this area (Luo et al., 2019; Zeng et al., 2012).

Son et al. (2012) used the Normalized Difference Vegetation Index (NDVI) from the Moderate Resolution Imaging Spectroradiometer (MODIS) and monthly surface temperature data from the National Aeronautics and Space Administration (NASA) to monitor agricultural drought from November 2001 to April 2010, and found that moderate and severe droughts occurred over the whole lower LMRB. Based on two long-term satellite-based precipitation products, namely the Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Climate Data Record (PERSIANN-CDR) and CHIRPS, Guo et al. (2017) found that hydrological drought occurred more frequently in the north and south of the LMRB, especially in the Mekong Delta which had experienced long-term and extreme drought events. Further, the spatial distribution of drought was mainly affected by precipitation and temperature, while precipitation was the dominant factor in the distribution of dry extremes (Li et al., 2013). Kang and Sridhar (2021) found that 68.4–76.1% of incidences of increased drought were caused by decreased precipitation or increased temperature. In terms of water vapor fluxes, the Tibetan Plateau Monsoon (TPM) and South Asian Monsoon (SAM) are the major factors that affect the occurrences of drought in the upstream and downstream regions of the LMRB respectively (Tang & Cao, 2020).

7.2.1.2 Temporal Characteristics of Drought in LMRB

Besides the obvious spatial characteristics, drought in the LMRB also has distinct temporal characteristics, and there are two major modes of drought development. One is the evolution from severe drought at the beginning of the dry season to moderate drought at the end of the dry season occurring in the Mekong Delta. The other is the gradual evolution of drought that intensifies and expands and occurs in the upper LMRB (Zhang et al., 2020a, 2020b).

Through wavelet transformation, it was found that drought in most of the LMRB shows an evolution with a major period of 3–7 years, which is likely related to El Niño-Southern Oscillation (ENSO) (Li et al., 2013). Meanwhile, the results of Empirical Mode Decomposition (EMD) showed that more than 60% of drought changes are caused by inter-decadal changes in precipitation (Li et al., 2013). Through monitoring of dry extremes, the main drought events identified were in 1983, 1991–1994, 1998–1999, 2005 and 2015–2016 (Guo et al., 2017). Focusing on droughts in different seasons, agricultural drought occurs mainly in the dry season, corresponding to the ripening stage of rain-fed rice and the heading stage of winter-spring rice (Son et al., 2012). However, generally, drought occurs most frequently in the boreal spring, especially in March and April (Li et al., 2013; Sridhar et al., 2019).

When considering the entire basin, drought in the LMRB decreased during 1977–2010. However, the Mekong Delta, with the most drought events in the LMRB, is trending towards drier conditions in some areas. Zhou et al. (2011) analyzed the precipitation data of 38 weather stations in the LMRB from January 1977 to August 2010 from the National Oceanic and Atmospheric Administration (NOAA) Data Sharing Network and the Mekong River Commission (MRC), and found that extreme droughts in the LMRB have decreased, especially in the dry season. After reconstructing the runoff of the Mekong river from 1557 to 2005 using tree rings, it was found that there has been a significant increase in runoff in the last 30 years (Yang et al., 2019a, 2019b). Kang and Sridhar (2021) applied the Soil and Water Assessment Tool (SWAT) model to simulate soil moisture, runoff and evapotranspiration in the LMRB using meteorological forcing data from the Coupled Forecast System Model version 2. Based on the results of the Modified Palmer Drought Severity Index (MPDSI), Standardized Soil Moisture Index (SSI) and Multivariate Standardized Drought Index (MSDI), drought was shown to increase in most of the lower LMRB while decrease in west of the mid LMRB during 1953–2016. Furthermore, a study based on the latest version of TRMM Multi-Satellite Precipitation Analysis (TMPA) real-time product (3B42RTv7) to achieve real-time Variable Infiltration Capacity (VIC) Macroscale Hydrologic modeling showed that 30% of areas experienced severe hydrological drought from January 2015 to December 2018 and severe drought would become normal in the LMRB (Zhang et al., 2020a, 2020b). Similarly, Jing et al. (2020), using Gravity Recovery and Climate Experiment (GRACE) data, found that drought in the upper part of LMRB had a slight increasing trend, while drought in the lower had an insignificant increasing trend from 2003 to 2016.

The change in drought is also related to human activities especially in the Mekong Delta. Based on the Temperature Vegetation Dryness Index (TVDI), most of the Mekong Delta has experienced moderate and even severe drought in the dry season (Phan et al., 2020). In 2016–2018, drought intensity became more severe and the increase in severity mainly occurred in rice fields (Tran et al., 2019). The increase of drought intensity in the Mekong Delta was attributed to human activities. Transforming land from perennial trees and forests to residential and public transportation land, led to the increase of drought in the central part of the Mekong Delta, while an increase of aquaculture land and mangrove forests led to a decrease in drought in coastal areas (Phan et al., 2020).

7.2.2 Flood Characteristics

Due to the sporadic nature of floods, the characteristics of floods in the LMRB are summarized as follows in order to provide a basis for the countries along the LMRB to formulate flood control measures.

7.2.2.1 Causes of Floods

The causes of floods in the main stream of the LMRB are mainly determined by factors such as topography and landforms, precipitation and runoff. The topography and geographical location of the LMRB determine different characteristics of floods generated in the basin. Due to geographical location and the influence of southwest monsoon, precipitation in the upper and lower reaches of the LMRB is significantly different (Kingston et al., 2011). Spatially, the annual precipitation amount increases gradually from north to south, with local patterns due to topographic effects. Furthermore, surface runoff can be divided into rainwater runoff and snowmelt runoff (Wang et al., 2021). The former is caused by rainfall and the latter by melting snow. Runoff is a basic element of the hydrological cycle, which causes the change of river water regimes. Rainfall contributes to 80–90% of runoff in the LMRB, which is the major factor influencing flood occurrence (Delgado et al., 2012; Lauri et al., 2012; Wang et al., 2022).

The LMRB has a monsoon climate, and torrential rains are the direct cause of floods (Darby et al., 2016; Yang et al., 2019a, 2019b). The Lancang River basin is a transitional climate zone from plateau temperate climate to subtropical climate. The middle and lower reaches of the Lancang River basin are typical alpine and canyon landforms with alternating mountains and gorges, and the terrain is undulating. The intensity of the heavy rain in this area is relatively large and it is the main flood-hit area in the Lancang River basin. The Mekong River Basin is a subtropical or tropical climate zone, and the weather systems that cause torrential rain are mainly tropical convergence zones, tropical cyclones and tropical depressions. The torrential rains in the Mekong River basin mainly occur from July to October. The upper reach of the Mekong River basin is a mountainous and hilly area, with a small number of mountain plains and basins intermittently distributed. The width of the river valleys alternates repeatedly. The valleys in the basins and alluvial plains in the dam area are open and gentle. The terrain is low and flat, and the downstream canyons are easily blocked by water. Therefore, it is easy to be flooded here, during the flood season. The middle and lower reaches of the LMRB and the Mekong Delta are mainly plains and lowlands, which are also vulnerable to flooding (Chen et al., 2020b; Wang et al., 2022).

The LMRB in Phnom Penh receives a large amount of water from its main tributary, Tonle Sap Lake. Tonle Sap Lake is an important flood buffer and natural reservoir (Chang et al., 2019; Try et al., 2022). The depth and area of the Tonle Sap Lake varies greatly between the rainy and dry seasons. During May to September, when the water level of the Tonle Sap Lake is lower than that of the LMRB, 10% to 18% of flood water in the LMRB will flow backward into the Tonle Sap Lake through the Tonle Sap River, greatly reducing the peak flow of the LMRB (Chang et al., 2019; Try et al., 2019). Thus, the regulation of Tonle Sap Lake, the storage of many river branches and the backflow of tides play a considerable role in influencing the main river channel. With Phnom Penh as its apex, the Mekong Delta has dense river networks. The river channels have changed from alluvial to siltation, and the water system is unstable, complex and volatile. The flow rate is slow, and the sediment is gradually silting up, forming a series of wide alluvial plains and wetlands. Due to the low and flat terrain, the discharge capacity of the river channel is seriously insufficient relative to the huge flood volume of the river section above Stung Treng. When the water level at the three stations of Kratie, Kampong Cham, and Takhmau is high, the flood of the LMRB passes through a large number of distributary channels or siltation channels to the floodplain behind the embankment, and the flood lasts for on average 19–48 days per year (Xu et al., 2020).

7.2.2.2 Temporal Characteristics of Floods

Floods in the LMRB are mostly formed by continuous heavy rains or rainstorms. Rainstorm volume increases from north to south in the basin. Downstream areas are likely to flood even in December.

In the Lancang River basin, the annual maximum flood is likely to occur from July to October, as precipitation during this period exceeds 80% of the annual precipitation (Kingston et al., 2011; Yang et al., 2019a, 2019b). At the same time, there are occasional floods in early- and mid-October. This is due to the relatively stable atmospheric circulation in the central and western regions of Yunnan Province in early- and mid-September, and the probability of heavy rain is relatively small. In October, cold air from the north becomes active, and the southwest airflow retreats to the west of Yunnan. The two air currents converge, resulting in heavy rain, which then causes floods. Tropical cyclones from the South China sea are one of the reasons for torrential rains in the Mekong River basin (Chen et al., 2019, 2020a). The flood period in the upper reaches of the Mekong River basin is from June to November, among which August is the most likely month for the annual maximum flood to occur. The flood period in the middle and lower reaches of the Mekong River basin is from June to December, and the probability of occurrence of the annual maximum flood is almost the same in August and September.

7.2.2.3 Spatial Characteristics of Floods

Floods caused by rainstorms in the LMRB show obvious spatial differences. The peak discharge per unit area is close to the limit value for global rain flood rivers (O'Connor & Costa, 2004). To consider spatial variation, we selected long-term daily streamflow observation data for seven hydrological stations from north to south in the LMRB, including Yunjinghong, Chiang Sean, Luang Prabang, Vientiane, Mukdahan, Pakse and Stung Treng. Table 7.1 shows the statistics of flood peak and flood volumes at these selected stations. The mean annual flood (MAF) of these seven stations ranges from 6710 to 54,000 m3/s. The measured maximum peak discharge (MAX) over the measured years ranges from 13,900 to 78,100 m3/s, while the mean annual 30-day flood volume (MAX30d) ranges from 13.8 to 112.9 billion m3. From the above statistical data, it could be found that the peak and volume of floods were not only large, but also increasing from north to south in the LMRB.

Table 7.1 Statistics of flood peak and flood volume at seven selected stations in LMRB
Full size table

The LMRB has a large latitudinal span, and the direction of the river and the shape of the basin are essentially parallel to the monsoon activity route. Therefore, the LMRB is dominated by regional floods, and the probability of basin-wide floods is small (Wang et al., 2022). Floods at upstream and downstream of Vientiane are basically discontinuous (MRC, 2007). Since hydrological measurement data became available, the flood from August to September 1966 was the most extensive flood in the LMRB, and its impact was mainly limited to the Lancang River basin and the upper and middle reaches of the Mekong River basin. Vientiane City and Nongkhai City were the regions most severely affected by the flood, and they experienced a catastrophic flood. Figure 7.2a shows distribution of the flood range and severity in the LMRB from 1985 to 2019. During the flood season, flood-prone zones are mainly distributed in the Mekong River basin, especially the downstream (Hoang et al., 2019). The most frequent flood-prone zone is mainly located in the “3S” river basin (i.e., Sekong, Se San, Sre Pok). Moreover, severe floods with high flood peaks or large flood volumes are more often to occur in the “3S” river basin (Wang et al., 2022). The flood-prone zones are mainly distributed in Tonle Sap Lake and the Mekong Delta, followed by the Mun-Chi River Basin in eastern Thailand and the Songkhram River Basin, upstream of Nakhon Phanom station.

Fig. 7.2
Two maps of the Lower Mekong River Basin. A depicts flood extent and severity from 1985 to 2019. B highlights operating, planned, and under-construction reservoirs and the main river.

a The distributions of flood extent and severity in LMRB during 1985–2019. Data were obtained from http://floodobservatory.colorado.edu/, for definitions of flood severity see Chen et al. (2020b). b The distributions of reservoirs in LMRB

Full size image

7.3 Climate Change Impacts on Droughts and Floods

7.3.1 Climate Change Impact in LMRB

Climate change is now considered to be one of the main threats facing the planet in the twenty-first century. The IPCC AR6 report pointed out that global surface temperature increased from 1850–1900 to 2010–2019 with a best estimate of 1.07 °C, and the rate of global warming has continued to accelerate (IPCC, 2021). In general, this warming intensifies the global hydrological cycle, which means an increase of global average precipitation and evaporation (Mishra & Singh, 2010). As a basin with a typical monsoon climate, the LMRB has also been greatly affected by climate change.

Projections based on climate models suggest a significant increase in basin-wide temperatures and changes in monsoon patterns (Pokhrel et al., 2018). The annual average temperature and precipitation respectively was predicted to increase by 0.6–1.4 ℃ and by 1.2–8.6%, respectively, for the years 2032–2042 compared to the baseline of 1982–1992 (Lauri et al., 2012). Compared to the baseline of 1971–2000, the daily average temperature during 2036–2065 was predicted to increase by 2.4 ℃ and 1.9 ℃ from Representative Concentration Pathway (RCP) 8.5 and RCP4.5 ensembles under CMIP5 climate projections, respectively; while annual precipitation from the two ensembles increased by −3 to 5% (RCP8.5) and 3–4% (RCP4.5) with an average of 3% across all scenarios (Hoang et al., 2016). The change patterns for the 95th and 99th percentile precipitation showed more prominent increases under Inter-Sectoral Impact Model Intercomparison Project (ISIMIP) projections, though the change patterns for the 90th percentile precipitation were comparable to that of the annual precipitation (Wang et al., 2017). Under high-resolution atmospheric GCMs (AGCMs), the annual precipitation of the future climate (2075–2099) would increase by 6.6–14.2% (Try et al., 2020a).

The temperature increase tends to be greater in the southern and northern parts of the basin, whereas the patterns for annual precipitation varied with GCMs and emission scenarios but increases were more likely in the Lancang River basin (Hoang et al., 2016; Lauri et al., 2012). Patterns for precipitation changes were different even under the same emission scenario: e.g., the largest increase was in the middle basin for three GCMs, while in the northernmost and southern parts for the remaining GCMs (Lauri et al., 2012). Under CMIP5, precipitation was projected to increase in some areas while decrease in others, though the overall pattern was increasing (Hoang et al., 2016).

7.3.2 Climate Change Impacts on Drought

GCMs are an advanced tool for assessing the impacts of climate change and have been widely used for drought projections (Abbasian et al., 2019; Tabari et al., 2021). Several climate change scenarios are used to describe the likely effects of future climate change, such as the Special Report on Emission Scenarios (SRES), the Representative Concentration Pathways (RCPs) and the Shared Socioeconomic Pathways (SSPs) (O’Neill et al., 2014). The effect of climate change under different scenarios is assessed below.

The SRES were originally used to describe climate change and its future effects (IPCC, 2007). Although the SRESs are currently rarely applied, the research method of combining the impact of population, energy, and economy on climate change is still valuable. SRES A1B represents very rapid economic growth with increasing globalization and a balance of fossil intensive and non-fossil fuels. The impact of climate change in this scenario on the drought of the Mekong is controversial. Based on precipitation and temperature from Japan Meteorological Agency’s GCM in the periods of 1979–1998 and 2080–2099, Kiem et al. (2008) simulated future hydrology by using the grid-based University of Yamanashi distributed hydrological model (YHyM) and found that there were fewer days of drought in the LMRB, meaning that drought will be alleviated in the future. However, when Falloon and Betts (2006) used the Hadley Centre Global Environment Model version 1—Total Runoff Pathways (HadGEM1-TRIP) to simulate global river flow changes under this A1B scenario, they found that although the projected annual runoff of the Mekong River increases by 40.3%, this increase reflects in a large increase in the monthly maximum flow and a large decrease in the monthly minimum flow. Hirabayashi et al. (2008) claimed the increase of drought from 2001 to 2030 mainly occurs in the middle LMRB, but the drought increases in the entire basin from 2017 to 2030, especially in the last 20 years of the twenty-first century. Scenario SRES A2 represents less rapid economic growth than SRES A1 but more rapid population growth. The consensus on the future drought of the Mekong in this scenario is that although climate change has increased annual runoff, there may be water deficit in the dry season because the increase in runoff occurs mainly in the wet season. Van Huijgevoort et al. (2014) found a decrease in low discharge for most models and by combining with three GCMs and five large-scale hydrological models, indicated drought would intensify in the LMRB. SRES B2 means development following environmentally, economically, and socially sustainable pathways. Under this scenario, Deb et al. (2018) estimated water resources based on five commonly used GCMs and the Hydrologic Engineering Center—Hydrologic Modeling System (HEC-HMS) models. Their conclusions were basically consistent with the conclusions under SRES A2, which is that the available water resources will decrease in the dry season but increase in the wet season. However, drought may become serious even in the wet season. Yamauchi (2014) found that drought duration and severity will generally increase in the wet season in the lower Mekong River under a scenario similar to SRES B2 proposed by the MRC.

In recent research, the SRES used frequently in earlier studies, have been replaced by RCP scenarios, describing radiative forcing and concentrations of greenhouse gases until the year 2100. Hoang et al. (2016) found that climate change reduced the frequency and extent of extreme drought of the Mekong River, when simulating extreme drought using the distributed hydrological Visual MODFLOW (VMod) model (Lauri et al., 2012), forced by five GCMs and under two RCPs. Their results suggest that the projected drought will decrease in 2036–2065 compared to 1971–2000. However, Sridhar et al. (2019) used SWAT and VIC models to simulate flows under the same RCP scenarios, and found that all results showed a significant reduction in the dry season flow and more severe drought, although the overall flow of the Mekong River would increase. Meanwhile, Thilakarathne and Sridhar (2017) found that most GCMs indicated increased probabilities of severe drought scenarios in the entire Mekong Basin, and the lower Mekong Basin was forecast to experience a higher risk of drought based on precipitation of 15 GCMs from NASA Earth Exchange Global Daily Downscale Predicted (NEX-GDDP). Based on four regional climate models (RCMs) (HadGEM3-RA, SNU-MM5, RegCM4 and YSU-RSM) and SWAT, the severity, duration, and frequency of drought in the Srepok basin located in the 3S area would increase. Meteorological, agricultural, and hydrological drought events were predicted to increase from 13% to 43%, 14% to 44%, and 22% to 40% respectively, under RCP 8.5 compared to 1980–2005 (Sam et al., 2019). Reduced drought but increased spatial heterogeneity were suggested by the result of a geomorphology-based hydrological model (GBHM), and southwestern China and the Mekong River estuary may suffer more severe drought (Li et al., 2021).

In the latest research into the effects of climate change on drought, SSPs, which consider the changes in global society, demographics and economics, are now being used as important inputs for the latest climate models. However, only a few studies have analyzed Mekong drought based on SSP scenarios. Zampieri et al. (2019) used the Annual Green Water Resources indicator (defined as the squared mean divided by the squared standard deviation of annual precipitation) time series to study droughts on a global scale under the SSP5-RCP8.5 scenarios, and found that the LMRB will become drier despite the increase in precipitation. Similarly, anthropogenic forcing was suggested to increase the risk of extreme and severe drought under SSP3-RCP7.0 and SSP5-RCP8.5 scenarios (Zhang et al., 2021). Whitehead et al. (2019) integrated RCP and SSP scenarios to simulate the flow and water quality of the Mekong River, but different SSP scenarios had no effect on lower flow because the model mainly focused on water quality. Future research on the impact of climate change on the drought in the LMRB should consider different SSP scenarios. By 2100, the range of changes in various socio-economic scenarios may be greater than the range of changes in various forcing levels (Arnell et al., 2019).

7.3.3 Climate Change Impacts on Floods

Climate change induced flood risks have been one of the challenges affecting global safety and sustainable development. The LMRB is one of the many flood-prone areas in Asia and has the highest flood-induced fatality in the world. Climate change has been one of two major challenges to its water resources in the twenty-first century. Understanding the impacts of climate change on floods in this region will help to plan and manage its water resources and provide guidance to disaster prevention and mitigation.

7.3.3.1 Relationship Between Climate Change and Floods

Climate change affects floods, both directly and indirectly, i.e., rainfall, sub-surface flow, and groundwater (Fang et al., 2014). For the direct effects (i.e., the main way), the anomaly in atmospheric circulation such as ENSO and monsoon changes can cause regional rainfall change (Räsänen & Kummu, 2013; Yang et al., 2019a, 2019b), while moisture increases in the atmosphere under global warming, causes an increase in the magnitude and frequency of rainfall (Kunkel et al., 2013). For the indirect effects, landcover (e.g., desertification) and soil property changes (e.g., soil erosion) under climate change scenarios can affect rainfall-runoff processes and thus can produce quicker and larger flood peaks (Bronstert, 2003).

In the LMRB, the monsoon climate dominates its hydro-climate conditions (Yang et al., 2019b). Two monsoon systems, i.e., the Indian summer monsoon (ISM) and Western North Pacific Monsoon (WNPM) (Delgado et al., 2012; Yihui & Chan, 2005), regulate the monsoon rainfall in the rainy season, leading to over 80% of the annual precipitation (Costa-Cabral et al., 2008) during May to October in the basin, and 80–90% of the discharge in the lower Mekong (Delgado et al., 2012). More importantly, the interannual variability of the rainy season precipitation in the LMRB is significantly modulated by co-variability of the ISM and WNPM (i.e., monsoon combined effect, donated as ISWN), other than individual effects (Holmes et al., 2009; Yang et al., 2019a, 2019b). When the ISM and WNPM are stronger (weaker) than normal, the combined effect is stronger (weaker) than normal. Also, the monsoons in this basin further interact differently with ENSO and are coupled to ENSO cycles where the coupling strength can be changed over time and with the Australian monsoon (Delgado et al., 2012; Lau & Wang, 2006). For example, during the warm phase of ENSO, the ISM and WNPM are found to weaken; the WNPM is most affected during the decay of the warm phase of ENSO, while the ISM is mostly affected during the ENSO development. Further, the linkage of the ISM and ENSO have dramatically reduced, but the linkage of the WNPM and ENSO have strengthened since the 1970s (Räsänen & Kummu, 2013; Wang et al., 2001; Wu & Wang, 2002).

Usually, the annual flood period, flood volume and annual flood peak decreased during El Niño and increased during La Niña (Räsänen & Kummu, 2013). During El Niño (La Niña) years, the flood start date was also delayed (advanced) from the average, while the flood end dates advanced (delayed). For the monsoons, the WNPM positively connects annual maximum discharge and flood season average discharge in Kratie and other stations in the southern parts of the LMRB, while the ISM has less influence on the interannual flood regime in these stations (Delgado et al., 2012). On average, the flood start date is advanced (delayed) by 8–12 days, Q10 increases (decreases) by 7.4–14.4%, and flood volume increases (decreases) by 9.0–17.5% during the strong (weak) monsoon years in over half of the monsoon impacted regions (Wang et al., 2022).

7.3.3.2 Flood Change in the Future

Under CMIP5 projections, the high flow (Q5) during the period of 2036–2065 was projected to increase at all considered stations in the LMRB with the range of 5–8%, when compared to the baseline period of 1971–2000, but this was also projected to slightly reduce in some scenarios with −6 to −1% (Hoang et al., 2016). Further, the extremely high flow represented by yearly peak discharges also exhibited substantial increases, meaning both the magnitude and frequency of annual peak flows were projected to increase in the future. For the ISIMIP projections, the flood frequency increased at a higher rate (10–140%) than that of the flood magnitude (5−55%) (Wang et al., 2017).

By using AGCMs, the flood magnitude in the LMRB would be more severe than in the present climate by the end of this century. The increase of precipitation could lead to an increase of the high flow (Q5) by 13–30%, of the peak inundation area by 19–43% and the increase of peak inundation volume by 24–55%, while no significant change was predicted to occur on peak flood timing (Try et al., 2020b). Under the scenario of the global average temperature increasing by 4 °C, different patterns of sea surface temperature significantly affected the variation of flood inundation in the future (2051–2110) in the Tonle Sap Lake and Vietnamese Mekong Delta. Extreme flood events (50-year, 100-year, and 1000-year return periods) showed the discharge, inundation area and inundation volume increased by 25–40%, 19–36%, and 23–37%, respectively (Try et al., 2020a).

In the Cambodian lowlands and Vietnamese Mekong Delta, where climate change and sea-level rise strongly alter the delta flood dynamics, RCMs based simulations projected that average and maximum water levels and flood duration would increase in 2010–2049, when compared to the baseline period of 1997–2000 (Västilä et al., 2010). When compared to the historical baseline period of 1971–2000, climate change based on CMIP5 suggested that the annual maximum water level in the future period (2036–2065) increased by 10–15% at Chau Doc, 2–8% at Long Xuyen, and less than 5% at Can Tho, with higher changes in wet years (Triet et al., 2020). The flood extent also separately increased by 1, 3 and 7% in the dry, normal and wet years for the future period of 2036–2065, while the inundation depth increased by 10–40 cm during the same period.

Climate change will remarkably alter flooding in the LMRB, which can be revealed from the model projections. A generally increasing pattern in flood peak inundation will potentially cause greater economic losses and death. Flood control and disaster reduction strategies including flood forecasting and flood control improvement are therefore urgently needed.

Fig. 7.3
Nine multi-line graphs plot S S P- 126, 370, and 585 versus years for drought probability and flood probability. The graphs have fluctuating increasing and decreasing trends.

Time series of drought probability and flood probability at the three representative stations during 1981–2100 in LMRB under the impact of climate change (CMIP6). The probability is calculated from the 30-year moving value (Fig. 7.3 was redrawn according to the data of Yun et al., 2021a)

Full size image

7.4 Impact of Reservoir Regulation on LMRB

In order to tackle the increase in extreme events (e.g., drought and floods) under climate change, as well as to meet the increased demands of energy and agricultural irrigation from these developing countries in the basin due to rapid urbanization and population explosions, reservoirs in LMRB have expanded with an unprecedented rate in the past decades (MRC, 2017). Before 2008, the LMRB was one of the basins least affected by human activities in the world, with the active reservoir storage capacity accounting for only 2% of the annual streamflow (Kummu et al., 2010). In the following ten years, a large number of reservoirs were successively constructed and put into operation. The total storage capacity of the 103 large reservoirs under operation by the end of 2021 in the LMRB reached a staggering 100.3 km3, accounting for 23% of the annual flow (according to the Greater Mekong Dam Database (GMDD), https://wle-mekong.cgiar.org/maps/). These reservoirs have profoundly altered the hydrological systems of the LMRB.

7.4.1 Observed Changes in Streamflow

The most discernible impact of reservoir operation in the LMRB is the changes to the seasonal flood pulse. Researchers have been concerned about the hydrological impact of the Lancang River basin dams because of their transboundary effects. Although many news and media outlets (Campbell, 2009; Stone, 2010) claimed that drought in the downstream basin might be caused by the construction of the upstream dams, most studies (Li et al., 2017; Yun et al., 2020) based on actual observations indicated there were limited impacts from upstream reservoirs (active storage capacity of 0.72 km3) before 2010.

Cochrane et al. (2014) showed that the dry season (February to May) flow of Chiang Sean station during 1991–2010 increased significantly, while there was almost no change in the wet season flow, compared to the period 1961–1990, and a similar trend was observed at Vientiane station. At the same time, Pakse and Stung Treng stations located in the downstream LMRB experienced a limited increase in dry season flow and an abnormal decrease in wet season flow, which may be caused by ignoring the difference in precipitation and land use changes during different periods. In addition, other studies (Li et al., 2017; Lu et al., 2014) have also reached similar conclusions, and attributed the increased dry season streamflow at Chiang Sean station to the upstream dams.

Subsequently, more reservoirs had a greater impact on the streamflow of the LMRB (Han et al., 2019; Räsänen et al., 2017; Yun et al., 2020). Li et al. (2017) reported the critical impact of the upstream dams on downstream dry-season flows, which have increased dramatically as far as Kratie station in Cambodia. Räsänen et al. (2017) found that the streamflow at the upstream LMRB decreased by 32–46% during the wet season, and increased by 121–187% during the dry season in 2014 compared with 1960–1990. Moreover, Han et al. (2019) concluded a 95% contribution to streamflow changes from human activities at Yunjinghong station in China since 2008. Yun et al. (2020) pointed out that the impact of newly constructed reservoirs in the LMRB during 2009–2016 was tremendous in the upstream, and reservoir construction in the downstream LMRB had a greater impact on streamflow in the lower LMRB.

The total storage capacity of the LMRB’s reservoirs under operation in 2021 was 57.7 km3, which is six tenths of the total reservoir capacity in the LMRB. Most of these reservoirs are located in Laos, Thailand, Cambodia, and Vietnam. Some of these reservoirs are used to divert water to hydropower plants, downstream river sections, and adjacent tributary basins, to increase water head for more hydropower generation. For example, some reservoirs in Thailand are used to divert water, which increases the complexity of downstream research. In general, the numerous downstream tributary reservoirs have reduced seasonal streamflow changes. Piman et al. (2013a) found that when considering 23 reservoirs aimed at maximizing hydropower generation in the 3S area (Laos, Cambodia, and Vietnam), the dry season flow will increase by 63%, and the wet season flow will decrease by 22%. Yun et al. (2020) pointed out that the annual streamflow changes downstream of Mukdahan station would be affected more by the reservoirs downstream of Vientiane station (−3 to 8%) instead the reservoirs in China (−2 to 4%) during 2009–2016.

In addition, diversions have been associated with reduced flows downstream of dams, as well as flow augmentation in tributaries (e.g., Baird et al., 2015; Chanudet et al., 2016). For example, the Nam Theun 2 dam (completed in 2010, Laos) enables diversions from the Nam Theun river into the Xe Bang Fai river, resulting in an 83% (from 220 to 486 m3/s) increase in natural mean annual flow at the Xe Bang Fai river. Today, downstream releases from Nam Theun 2 are just 2 m3/s, less than 1% of its mean annual inflow of 238 m3/s, which affects the livelihoods of over 110,000 people (Hecht et al., 2019). At the same time, similar diversion has also been observed in other tributaries, including Houay Ho (Laos) and Yali Falls (Vietnam). Diversions to raise water levels in the Nam Ngun 1 reservoir have drastically reduced dry-season flows in the Nam Song River (Hecht et al., 2019). Ruiz-Barradas and Nigam (2018) indicated that extensive irrigation diversions in the Mun-Chi basin of northeastern Thailand resulted in an abnormal phenomenon of increased precipitation and decreased flow at same time in this area.

When evaluating the comprehensive impact of the LMRB’s reservoirs, basin-wide hydrological models agreed that the reservoir will mainly affect seasonal flow fluctuations, including reducing the wet season streamflow and increasing dry season streamflow. Hoanh et al. (2010) investigated the impacts of 6 upstream dams in China and 81 downstream dams in Thailand, Laos, Cambodia, and Vietnam (including 11 mainstream dams), and concluded that the wet season flow will be reduced by 8–17% and the dry season flow will increase by 30–60% by the 2030s. Simulation results from Piman et al. (2013b) showed that 88 future reservoirs will increase Kratie’s dry season flow by 28% and decrease wet season flow by 9%. Lauri et al. (2012) indicated that 126 future reservoirs with hydropower generation strategy will lead to an increased dry season (+160%) and decreased wet season (−24%) flow at Kratie station. Based on comparison with actual observations and simulation results considering 86 reservoirs, Yun et al. (2020) pointed out that reservoirs in the LMRB reduced the streamflow variability by increasing dry season streamflow (+15 to +37%) and decreasing wet season streamflow (−2 to −24%) between 2009–2016. Shin et al. (2020) used a basin-wide river-floodplain-reservoir modeling system to directly evaluate the impacts of 86 existing dams across the LMRB and found that while the effects of dams on downstream flood patterns was minimal until 2010, the impacts has substantially increased since then because of new dam construction.

7.4.2 Reservoir Impact on Extreme Events

Streamflow change caused by the rapid hydropower expansion have affected many aspects of the LMRB’s river ecosystem, including changes in hydrological extreme events (droughts, floods).

Recent observations (Li et al., 2017; Räsänen et al., 2017) indicated that the flood season in the LMRB has been delayed due to the seasonal storage buffer created by the reservoirs. At the same time, Yun et al. (2020) based on observation, found that reservoir operation during 2009–2016 reduced flood risk in the LMRB, while climate change increased the flood magnitude and frequency by up to 14% and 45%, respectively; reservoir operation reduced flood magnitude and frequency by 16% and 36%, respectively. Reservoirs in the LMRB have a greater influence during the dry season (Dang et al., 2020; Piman et al., 2013a). It is worth mentioning that recent research (Ji et al., 2018; Wang et al., 2020) based on remote sensing and hydrological modeling pointed out that China’s reservoirs have minimal impact on the reduction of the Tonle Sap Lake. Instead, the large-scale construction of dams on downstream tributaries is the significant contributor to hydrological alterations in the Tonle Sap Floodplain (Arias et al., 2014).

In the twenty-first century, climate change and hydropower expansion have brought new challenges to the LMRB. Kiem et al. (2008) and Hoang et al. (2016) pointed out that future increased precipitation will change streamflow patterns and increase flood risks in the LMRB. At the same time, due to the demand for energy, rapid expansion of hydropower will more directly alter river flow. Yang et al., (2019a, 2019b) and Yun et al. (2020) reported that reservoir regulation will reduce streamflow in the wet season and increase streamflow in the dry season in the LMRB. Under the combined impact of future climate change and hydropower development, streamflow and water hazards in LMRB will change drastically.

A large number of studies have evaluated the hydrological impact due to future precipitation and temperature changes under a changing climate. Based on CMIP5 forcing data, Hoang et al. (2016) found that future water vapor content will increase during the wet season (8 out of 10 scenarios) and dry season (all 10 scenarios) in the LMRB. At the same time, the extreme high streamflow of the LMRB in the twenty-first century (Hoang et al., 2019), as well as the magnitude and frequency of floods, will also exhibit a continuously increasing trend (Wang et al., 2017). However, projections of future drought based on GCMs has greater uncertainty. Kiem et al. (2008) projected that drought events would decrease due to increased precipitation, while Thilakarathne and Sridhar (2017) projected that greater interannual variability will exacerbate drought. The evaluation results indicate that future climate change will dominate the annual flow changes, including the increase in average annual flow and inter-annual fluctuations, which can lead to more severe droughts and floods.

Different to climate change, reservoir regulation mainly dominates the change of seasonal runoff, including decreasing intra-annual fluctuations and reducing flood events. Lauri et al. (2012) reported that climate change may exacerbate or offset the wet season flow (−21 to +4%) at Kratie station during 2032–2042. Wang et al. (2017) reported that the reservoirs in the upper LMRB can alleviate the increasing flood risk upstream of Luang Prabang station. At the same time, many GCMs project that streamflow will increase during the dry season (e.g., Hoang et al., 2016). Reservoir operation in the LMRB can also reduce dry/wet hydrological extremes increased by climate change. On the whole, after considering the combined effect of climate change and dams, reservoirs show a strong effect in regulating seasonal streamflow change, including an increase in dry season flow (+150%) and a decrease in wet season flow (−25%), as well as a reduction of flood magnitude (−2.3 to −29.7%) and flood frequency (−8.2 to −74.1%) (Yun et al., 2021a).

7.4.3 Adaptation and Mitigation of Extreme Events

Climate change will bring new challenges to water resource management. In the future period, the LMRB will likely encounter more hydrological extreme events (droughts and floods), and reservoirs are regarded as one of the most important measures to deal with future uncertainties. Under the impact of climate change, the LMRB’s future extreme events will continue to increase, including more frequent meteorological/hydrological droughts (Sam et al., 2019) and higher flood risks (Hoang et al., 2016; Wang et al., 2017). Based on the latest CMIP6 projections, Yun et al. (2021a) showed a dramatic increase in drought events during the mid-twenty-first century and flood events at the end of the twenty-first century in the LMRB. Recent studies (Guo et al., 2020; Wu et al., 2018) suggest that properly functioning reservoirs can delay the propagation from meteorological extreme events to hydrological extreme events by releasing/storing water in different situations. It is estimated that the LMRB’s reservoirs can mitigate the impact of climate change on droughts and floods, reducing most of the basin-wide dry hydrological extremes as well as the 32% of wet hydrological extremes. However, the effect of reservoirs in mitigating long-term extreme events (return periods of more than 6 years) is relatively limited.

Fig. 7.4
Nine multi-line graphs plot S S P- 126, 370, and 585 versus years for drought probability for natural and with dam, and flood probability for natural and with dam. The graphs have fluctuating increasing and decreasing trends.

Time series of drought probability and flood probability at the three representative stations during 1981–2100 in LMRB under the impact of climate change and reservoir operation. The probability is calculated from the 30-year moving value (Fig. 7.4 was redrawn according to the data of Yun et al., 2021a)

Full size image

The LMRB is one of the basins with the least amount of irrigation water during the dry season (Haddeland et al., 2006). Different from some other basins in the world where reservoir functions are in direct competition, the hydropower generation of the reservoirs in LMRB is complementary to agricultural irrigation (Lacombe et al., 2014). Reservoir regulation has increased the dry season flow and reduced the wet season flow in the LMRB. The increased dry season flow will supplement the dry season's irrigation, and the reduction in flood risk caused by decreased wet season flow, will benefit agricultural production (Yun et al., 2020). Under the dry and hot climate conditions in the future, reservoir regulation will reduce the ecological, agricultural and fishery economic losses in the midstream and downstream LMRB (Yang et al., 2019a, 2019b).

With the continuous construction of reservoirs in the LMRB, the riparian countries recognize that a well-established Lancang-Mekong cooperation mechanism will facilitate the deployment and cooperation of transboundary water resources to cope with future drought and flood events (Kittikhoun & Staubli, 2018; Li et al., 2019). For example, emergency releases from upstream reservoirs mitigated severe drought in the downstream countries in March 2016 (Hecht et al., 2019), and this case also confirms that increased dry season flow can alleviate the constraints of salt and acid groundwater on delta agriculture (Piman et al., 2013a; Smajgl et al., 2015). The latest assessment results (Yun et al., 2021b) carried out in the LMRB showed that while climate change would increase flood risk, adaptive reservoir operation can reduce flood magnitude by 5.6–6.4% and frequency by 17.1–18.9% at the cost of 9.8–14.4% of hydropower generation. In particular, upstream reservoirs will suffer more hydropower loss (5.4 times that of downstream reservoirs) to benefit downstream flood control in the LMRB (Yun et al., 2021b).

Reservoir operations established through in-depth international water cooperation can mitigate hydrological extremes in transboundary rivers (Wheeler et al., 2018; Yu et al., 2019). However, reservoir operations that lack international cooperation will prioritize their own water use in upstream countries, for example, reservoir over-discharge during floods or unrestrained water storage during drought, which will exacerbate flood and drought disasters in downstream countries. Existing organizations (such as the Mekong River Commission) and emerging basin-wide organizations (such as the Lancang-Mekong Cooperation) will conduct integrated regulation and cooperation in the LMRB to achieve transboundary management and coordination of water resources. China has shared hydrological data at Jinghong Station (near the China-Myanmar border) since November 2020, which will provide an important basis for supporting LMRB cooperation.

7.5 Conclusion and Recommendation

This chapter comprehensively reviews the characteristics of drought and flood changes during historical and future periods in the LMRB under the impact of climate change and human intervention.

In the past decades, there were obvious spatial distribution patterns of drought in the LMRB. Drought has increased significantly over the middle and the lower LMRB during 1979–2019. Similarly, floods have obvious temporal and spatial distribution patterns affected by topography and precipitation. The further downstream, the later the rainstorms end, and the later the flood season ends. Floods mainly occur in July–October, with the highest probability of flooding in August, and the flood peak values downstream are much higher than those upstream. Floods are now one of the most threatening hazards in the LMRB.

With future climate change-induced increases in temperature and precipitation, results based on GCMs and hydrological models showed that the annual flow will increase but the flow in dry season will decrease, it is estimated that the LMRB will face a severe drought threat during the mid-twenty-first century. Meanwhile, climate model projections indicate that the flood frequency will increase by 10–140%, the flood magnitude will increase by 5–55%, and the peak inundation will enlarge by 19–43% in the twenty-first century. This will potentially cause economic losses and human fatalities. The basin-wide adaptive strategies including flood forecasting and enacting new flood protection standards are therefore urgently need to be planned and carried out.

To tackle the increasing future droughts and floods, reservoirs in the LMRB have expanded at unprecedented rate over the past twenty years. By 2021, the total storage capacity of the 103 huge reservoirs in the LMRB reached a staggering 100.3 km3, accounting for 23% of the annual flow. These reservoirs are regarded as one of the most important measures to mitigate future extreme events. Reservoirs show an effective regulating effect for streamflow and extreme events, including an increase in dry season flow and a decrease in wet season flow, as well as the reduction of future droughts and floods.