Abstract
Climate change alters hydrological cycles and streamflow regimes at the local, regional and global levels. In this study, we aimed to assess the change in water balance change and hydrological extremes in the Prek Thnot River Basin of the Lower Mekong in Cambodia through a hydrological model (SWAT) under the two climate change scenarios (RCP2.6 and RCP8.5) following three different GCMs. An ensemble of 3 GCMs included GFDL-CM3, GISS-E2-R-CC and IPSL-CM5A-MR models and was applied to a well-calibrated SWAT model through climate change factors. Annual precipitation under RCP2.6 likely decreases by 0.1–0.5% for the near future (2021–2040) and mid-future (2051–2070) and decreases by 0.2–1.3% under RCP8.5. The decrease in precipitation will lead to reductions in water yield by 1–4% (RCP2.6) and 2–5% (RCP8.5). However, peak flow is expected to increase, while the low flow was projected to decrease (1–2% for RCP2.6 and 8–9% for RCP8.5). The study further found that high flow events will increase in both magnitude and frequency. The finding highlights water resources management issues in the Prek Thnot River Basin, including the frequency of future flood events.
HIGHLIGHTS
We estimated the climate change impact on water balance and hydrological extremes.
Annual water balance in the Prek Thnot River Basin is projected to decrease in the future.
Future streamflow is expected to increase in the wet season while decreasing in the dry season.
Peak flows have an increasing trend while low flows have a decreasing trend in the future.
Graphical Abstract
INTRODUCTION
Climate change is one of the major global threats that strongly affect the environment, ecosystem, and human society. In recent decades, the study on climate change impact on pervasive global warming driven by anthropogenic emissions of greenhouse gases (GHGs) has been given ample attention worldwide (IPCC 2014). In the past few decades, intensification of precipitation and increases in temperature due to global warming have been observed in Asia (Tan et al. 2021a, 2021b; Xu et al. 2021). These changes significantly impact different hydrological systems and consequently increase regional water hazards such as floods and drought (Chattopadhyay & Jha 2016; François et al. 2019). Moreover, climate change alters hydrological cycles and streamflow regimes at the regional and global levels (IPCC 2014; Martínez-Retureta et al. 2021), especially at the basin scale (Zhang et al. 2016). The predicted effects are precipitation patterns, increased intensity of extreme weather events, and glacier retreats (IPCC 2014). They consequently alter the river discharge regimes and directly affect ecosystems, water security. Hence, quantifying the climate change impact on water balance components and hydrological extremes is essential to watershed management as well as to the development of adaptation strategies, more effective mitigation and greater resilience against water hazards (Bhatta et al. 2019; Blanco-Gómez et al. 2019; Oeurng et al. 2019; Touch et al. 2020; Tan et al. 2021b; Yun et al. 2021).
A set of scenarios known as the Representative Concentration Pathways (RCPs), published in the Fifth Assessment Report (AR5) of the Intergovernmental Panel on Climate Change (IPCC) has recognized the significant indications of climate change: global temperature rises due to greenhouse emissions and water cycle changes (IPCC 2014). IPCC (2014) reported that global ocean and land temperatures have increased by 0.85 °C throughout 1880–2012. The global average temperature increase is projected to be between 0.3 °C and 4.8 °C, and the global mean sea level may rise by 0.26–0.82 m by the end of the 21st century (Stocker 2014). Across Southeast Asia, the mean annual temperature by the end of the 21st century is expected to rise from 0.8 °C in the lowest emissions scenario to 3.2 °C in the highest emissions scenario. At the same time, a moderate increase in precipitation is also projected for this region from 1% to 8% by 2100 (IPCC 2013). However, the change in global future precipitation is not uniform (both an increasing and/or decreasing trend) in terms of regions and seasons. The RCPs integrate several scenarios of policy-level interventions, adaptations and vulnerability mitigation practices (Van Vuuren et al. 2011). Specifically, IPCC uses General Circulation Models (GCMs) for future climate change projections based on different Representative Concentration Pathways (RCPs) scenarios. GCM projections indicate a substantial variability of future precipitation and temperatures in terms of trends and extremes (Reshmidevi et al. 2018). Many studies have been applied and tested together with a range of viable GCM models and emission scenarios allowing a comprehensive understanding of the impacts of climate change on water resources management (Islam et al. 2005a; Touch et al. 2020), water availability (Islam et al. 2005b, 2007b; Oeurng et al. 2019), and water scarcity including virtual water trading (Islam et al. 2007a; Gosling & Arnell 2016).
The effective development of water resource use and protection under changing environmental conditions requires hydrological models that can simulate water balance components and hydrological extremes under different change scenarios. Hydrological models are the principal tools used to explore the potential effects of climate change on water resources. The present study used the semi-distributed model, Soil and Water Assessment Tool (SWAT) model (Arnold et al. 1998), to assess water balance components. SWAT has been successfully and widely used around the world to simulate basin hydrology under different climate change scenarios (Senent-Aparicio et al. 2017; Ji & Duan 2019; Zhang et al. 2019) for understanding the effects of future development and management activities. Many studies have applied and tested SWAT performance in different Mekong River basins for climate change impact assessment on water availability and even extend to water quality (Oeurng et al. 2016; Shrestha et al. 2016; Ang & Oeurng 2018; Oeurng et al. 2019; Sok et al. 2020, 2022; Touch et al. 2020; Chim et al. 2021). Moreover, the SWAT model was used to simulate the water balance in different physiographic regions of the West Seti River Basin of Nepal and to analyze the impact of climate change on hydrological extremes (Budhathoki et al. 2021). In Thailand, studies of climate change impacts on water balance and hydrological extremes were in Bang Pakong-Prachin Buri river basin (Okwala et al. 2020) and flow regimes in the Chao Phraya River Basin (Ligaray et al. 2015).
Several previous studies have been carried out in the Prek Thnot River Basin by concerned authorities and stakeholders. Most of these studies were concentrated on flood management, agricultural development, irrigation and drainage system, land-use change, and integrated water resources management (Lee et al. 2014; Fujimoto & Tomotsugu 2019; Sothea et al. 2019; UNDP 2019; Khorn et al. 2020a, 2020b; Koem & Tantanee 2020; Sagara 2021). However, none of them has focused on the overall change in water balance components and extreme hydrological events in future under the climate change scenarios.
Thus, this study aims to bridge the knowledge gap regarding climate change and its impacts on water balance components and extreme hydrological events utilizing the SWAT model's most recent data and research approach in the Prek Thnot River Basin. Furthermore, this study also analyzed the changes in extreme hydrological events due to the changing climate, which affects the water resources development projects in the basin. Three different GCMs with two different emission scenarios, RCP 4.5 and RCP 8.5, were used during the study for a baseline of 2000–2011 and two future periods: near-term future from 2021 to 2040 or 2030s and medium-term future from 2051 to 2070 or 2060s. These GCMs were selected based on previous applications for similar purposes within the region (Oeurng et al. 2019; Heng et al. 2020; Touch et al. 2020; Chim et al. 2021; Sok et al. 2022).
MATERIALS AND METHODS
Study area
The Prek Thnot River Basin experienced maximum flood with discharge up to 1,371 m3/s in 1991 and 1,304 m3/s in 2000 (JICA 2012). In August 1991, floods affected more than 150,000 people in Kompong Speu and Kandal provinces, inundated over 15,000 hectares of paddy fields and damaged thousands of households and main roads connecting the provinces to the Phnom Penh capital (UNDP 2019). Severe floods occurred again in 2000, 2001, 2002, 2008, 2011 and 2013 (UNDP 2019). Within the Prek Thnot River Basin in 1995, drought damaged over 4,000 hectares of crops in Kampong Speu province (MoE 2001). Chhinh & Millington (2015) conducted a study on drought monitoring for rice production in the Prek Thnot River Basin in Kampong Speu province, where droughts between 1994 and 2006 damaged more than 1,000 hectares of paddy rice.
Hydro-meteorological data used in the study
Recorded streamflow and rainfall at daily scale data used in this study were obtained from the data recorded by the Department of Hydrology and River Works (DHRW) of the Ministry of Water Resources and Meteorology (MOWRAM), Cambodia. This study has eight rainfall and one streamflow gauge station along the main Prek Thnot River (Table 1).
Recorded rainfall, streamflow, and its descriptive statistic were used in this study
Name of Station . | Record used . | Timestep of Measurement . | Monthly (mm) & (m3/s) . | |||
---|---|---|---|---|---|---|
Period . | Min . | Mean . | Max . | STDV . | ||
Rainfall Record Used | ||||||
Kampong Speu | 1997–2011 | Daily | 6 | 106 | 232 | 76 |
Kong Pisey | 1997–2011 | 10 | 93 | 243 | 72 | |
Oral | 1997–2011 | 16 | 98 | 251 | 72 | |
O Taroat | 1997–2011 | 2 | 91 | 195 | 66 | |
Peam Khley | 1997–2011 | 5 | 92 | 208 | 64 | |
Phnom Srouch | 1997–2011 | 11 | 98 | 233 | 67 | |
Prey Pdao | 1997–2011 | 5 | 101 | 228 | 68 | |
Trapeang Chor | 1997–2011 | 13 | 94 | 242 | 71 | |
Streamflow Record Used | ||||||
Peam Khley | 2000–2011 | Daily | 5 | 47 | 192 | 54 |
Name of Station . | Record used . | Timestep of Measurement . | Monthly (mm) & (m3/s) . | |||
---|---|---|---|---|---|---|
Period . | Min . | Mean . | Max . | STDV . | ||
Rainfall Record Used | ||||||
Kampong Speu | 1997–2011 | Daily | 6 | 106 | 232 | 76 |
Kong Pisey | 1997–2011 | 10 | 93 | 243 | 72 | |
Oral | 1997–2011 | 16 | 98 | 251 | 72 | |
O Taroat | 1997–2011 | 2 | 91 | 195 | 66 | |
Peam Khley | 1997–2011 | 5 | 92 | 208 | 64 | |
Phnom Srouch | 1997–2011 | 11 | 98 | 233 | 67 | |
Prey Pdao | 1997–2011 | 5 | 101 | 228 | 68 | |
Trapeang Chor | 1997–2011 | 13 | 94 | 242 | 71 | |
Streamflow Record Used | ||||||
Peam Khley | 2000–2011 | Daily | 5 | 47 | 192 | 54 |
SWAT conceptual model

SWAT model input
The SWAT model was set up to cover about 5,600 km2 from the most upstream of the Prek Thnot River Basin (Figure 1). After the preparation of forcing data, four significant steps must be done to set up the model: (i) watershed discretization and sub-watershed characteristics derivation, (ii) define the HRU, (iii) run the model and analyze the parameter sensitivity; and (iv) calibrate and validate the SWAT model, including uncertainty analysis. Each data input is obtained from different sources, which is summarized in Table 2.
Summary of data and their corresponding sources in the SWAT model
Data Types . | Description . | Spatial Resolution . | Temporal Resolution . | Data Sources . |
---|---|---|---|---|
Topography map | DEM | 30 m | Shuttle Radar Topography Mission (SRTM) (http://srtm.csi.cgiar.org.) | |
Landuse/land-cover (LULC) map | Landuse classification | 250 m×250 m | 2002 | MRC |
Soil map | Soil types | 250 m×250 m | 2002 | MRC |
Meteorology | Observed rainfall | 8 stations | Daily, 1997–2011 | DHRW of MOWRAM |
Hydrology | Observed Streamflow | 8 stations | Daily, 2000–2010 | DHRW of MOWRAM |
Climate Weather | Gridded climate data | 0.25° | Daily, 1997–2011 | Global Weather Data (http://globalweather.tamu.edu) |
General Circulation Models (GCMs) data | Climate Change Scenarios RCP2.6&8.5 | Change factor in the subbasin | Monthly, 2030 and 2060s | MRC |
Data Types . | Description . | Spatial Resolution . | Temporal Resolution . | Data Sources . |
---|---|---|---|---|
Topography map | DEM | 30 m | Shuttle Radar Topography Mission (SRTM) (http://srtm.csi.cgiar.org.) | |
Landuse/land-cover (LULC) map | Landuse classification | 250 m×250 m | 2002 | MRC |
Soil map | Soil types | 250 m×250 m | 2002 | MRC |
Meteorology | Observed rainfall | 8 stations | Daily, 1997–2011 | DHRW of MOWRAM |
Hydrology | Observed Streamflow | 8 stations | Daily, 2000–2010 | DHRW of MOWRAM |
Climate Weather | Gridded climate data | 0.25° | Daily, 1997–2011 | Global Weather Data (http://globalweather.tamu.edu) |
General Circulation Models (GCMs) data | Climate Change Scenarios RCP2.6&8.5 | Change factor in the subbasin | Monthly, 2030 and 2060s | MRC |
Spatial maps of the Prek Thnot River Basin: DEM (Top), land use distribution (bottom left), and soil type distribution (bottom right).
Spatial maps of the Prek Thnot River Basin: DEM (Top), land use distribution (bottom left), and soil type distribution (bottom right).
Model calibration and validation
SWAT is a physically-based, semi-distributed, agro-hydrological simulation model that operates on a sub-daily to annual scale time step on a watershed scale, and it simulates the overall water balance for each HRU. SWAT was developed to predict the impact of management on water, sediment, and agricultural chemical yields in ungauged catchments (Arnold et al. 2012). The calibration was performed by using the manual and a semi auto-calibration procedure (SWAT-CUP), which was applied using a sequential uncertainty fitting algorithm (SUFI-2) detailed by Abbaspour (2014) with the comparison of observed data and literature review information for overall hydrology components. In order to calculate potential evapotranspiration, the Penman-Monteith method was selected. The parameters controlling groundwater behavior in the model and dependent on spatial data have been calibrated with literature daily (Table 3). The parameters were calibrated/validated in the gauge station, namely Peam Khley. The calibration results showed the importance of parameters, such as Soil_AWC, Soil_K, and ALPHA_BF (groundwater parameter) in the studied flow of the analyzed in Prek Thnot River Basin (Table 3). The parameter CN2 is significantly related to the runoff quantity and depends on soil utilization. Soil_K and Soil_AWC are related to the quantity of soil-water relationships in various soil types of the region. The model performance was measured by comparing the simulated with the observed components using the Nash–Sutcliffe model efficiency factor (NSE) as the objective function. The Nash-Sutcliffe efficiency (NSE), percent bias (PBIAS), the ratio of mean squared error to the standard deviation of the measured data (RSR), and coefficient of determination (R2) are frequently used measures in hydrological modeling studies (Moriasi et al. 2015).
Calibrated values of SWAT parameters for the Prek Thnot River Basin
Parametera . | Description of parameters . | Literature range . | Calibrated value . |
---|---|---|---|
v_ALPHA_BF.gw | Baseflow alpha factor (days) | 0–1 | 0.975 |
a_GW_DELAY.gw | Groundwater delay (days) | 0–500 | 1.75 |
v_GW_REVAP.gw | Groundwater ‘revap’ coefficient | 0.02–0.20 | 0.0596 |
v_GWQMN.gw | Threshold depth of water in the shallow aquifer required for return flow to occur (mm) | 0–5,000 | 570 |
a_RCHRG_DP.gw | Deep aquifer percolation fraction | 0–1 | 0.2 |
a_REVAPMN.gw | Threshold depth of water in the shallow aquifer for ‘revap’ to occur (mm) | 0–5,000 | 103.75 |
v_ESCO.hru | Soil evaporation compensation factor | 0–1 | 0.595 |
r_CN2.mgt | SCS runoff curve number | 35–98 | [54.12–88.85]2 |
r_CH_K2.rte | Effective hydraulic conductivity in main channel alluvium (mm/h) | −0.01 to 500 | 386 |
r_SOL_AWC.sol | Available water capacity of the soil layer (mm H2O/mm soil) | 0–1 | [0.33–0.49]3 |
r_SOL_K.sol | Depth soil surface to bottom of layer (mm/hr) | 0–2,000 | [76.41; 769.43] |
r_SOL_Z.sol | Saturated hydraulic conductivity (mm/hr) | 0–3,500 | [390.50; 2603.60] |
Parametera . | Description of parameters . | Literature range . | Calibrated value . |
---|---|---|---|
v_ALPHA_BF.gw | Baseflow alpha factor (days) | 0–1 | 0.975 |
a_GW_DELAY.gw | Groundwater delay (days) | 0–500 | 1.75 |
v_GW_REVAP.gw | Groundwater ‘revap’ coefficient | 0.02–0.20 | 0.0596 |
v_GWQMN.gw | Threshold depth of water in the shallow aquifer required for return flow to occur (mm) | 0–5,000 | 570 |
a_RCHRG_DP.gw | Deep aquifer percolation fraction | 0–1 | 0.2 |
a_REVAPMN.gw | Threshold depth of water in the shallow aquifer for ‘revap’ to occur (mm) | 0–5,000 | 103.75 |
v_ESCO.hru | Soil evaporation compensation factor | 0–1 | 0.595 |
r_CN2.mgt | SCS runoff curve number | 35–98 | [54.12–88.85]2 |
r_CH_K2.rte | Effective hydraulic conductivity in main channel alluvium (mm/h) | −0.01 to 500 | 386 |
r_SOL_AWC.sol | Available water capacity of the soil layer (mm H2O/mm soil) | 0–1 | [0.33–0.49]3 |
r_SOL_K.sol | Depth soil surface to bottom of layer (mm/hr) | 0–2,000 | [76.41; 769.43] |
r_SOL_Z.sol | Saturated hydraulic conductivity (mm/hr) | 0–3,500 | [390.50; 2603.60] |
aThe qualifier (r_) states to relative change, i.e., the existing parameter must be multiplied by (1+ the value found in calibration), (v_) refers to the value of the current parameter must be replaced by the value found in calibration, and (a_) means that absolute modification (i.e., the calibrated value must be added to the original value of the parameter). 2 Varies by HRU. 3 Varies by soil layer.
Analysis of water balance components
Analyzing hydrological changes
Two techniques were employed to analyze different aspects of hydrological changes. First, annual, seasonal, and monthly streamflow statistics were calculated to understand the river's flow regime changes. The second technique, the Q5 and Q95, was calculated to analyze the high-flow and low-flow conditions. The high flow condition exceeding 5% of the time (Q5) and low flow condition exceeding 95% (Q95) were calculated to evaluate the changes in flood and drought characteristics, respectively. The flow duration curve was used to identify the Q5 and Q95. These results were analyzed at the basin scale for 12 years (2000–2011). Hydrological changes were calculated under individual scenarios and ensembles, i.e., average changes from three different GCMs under both RCPs in two-time horizons.
Analyzing extreme streamflow
The range of x depends on the value of b; it is bounded by c+(a/b) from above for b>0, i.e., bounded from below for b<0, i.e., the shape parameter b determines which extreme value distribution is represented. Depending on the value of b, Equation (2) corresponds to the Fisher-Tippett distribution types I, II, and III.



Selected climate change scenarios and GCMs
Climate change scenarios were developed using different Global Circulation Models (GCMs), emission scenarios, time horizons, and locations (MRC 2015, 2017). The most important source of uncertainty in flow, irrigation demand and sediment modeling attributed to the different GCMs have been previously identified (Shrestha et al. 2016; Oeurng et al. 2019; Heng et al. 2020; Touch et al. 2020; Sok et al. 2022). Hence, the GCMs selection for climate change scenarios is an essential model development procedure. This study selected the future climate change scenarios based on the MRC (2015) that recommended three different GCMs (GISS-E2-R-CC, IPSL-CM5-MR and GFDL-CM3). For each GCM, we focused on two emission scenarios of Representative Concentration Pathways (RCPs), namely RCP2.6 and RCP8.5 were suitable for the widespread impacts of climate change. RCP 2.6 is a very low scenario in which radiative forcing peaks by the mid-21st century and gradually declines to 2.6 W/m2 by 2100 (Taylor et al. 2012). RCP8.5 is a scenario of high greenhouse gas emissions increasing throughout the twenty-first century before reaching an 8.5 W/m2 radiative forcing level by 2100 (Riahi et al. 2011). Moreover, RCP 8.5 is a plausible and accurate representation of the concentrations of atmospheric carbon that would be reached on the business-as-usual path. The difference between RCP 2.6 and 4.5 remains relatively small until the end of the century. Therefore, these RCP scenarios were selected based on their projects, which are designed to synthesize impact projections in the river flow, water temperature, precipitation, agriculture, biome, health, and infrastructure sectors at different levels of global warming. Furthermore, RCP2.6 and RCP8.5 are commonly selected in Mekong River Basin and Cambodia (Heng et al. 2020; Kayiranga et al. 2021; Wang et al. 2021; Sok et al. 2022). Based on an MRC (2017) study, two time horizons, 2030s (near-term future 2021–2040) and 2060s (medium-term future 2051–2070), were considered in this study, as the MRC is using these time horizons to maximize the amount of uncertainty.
The Department of Climate Change and Adaptation Initiative (CCAI) of the Mekong River Commission was obtained and downscaled climate change data sets (IPCC 5th Assessment Report). This dataset contains a monthly SWAT model (change factors) for precipitation, temperature, and climate weather (solar radiation and relative humidity). MRC CCAI uses SIMCLIM software to downscale the climate dataset. SimCLIM downscales GCM outputs using pattern scaling and a bilinear interpolation method. MRC CCAI conducts the change factor technique to quantify projected climate changes because it is the most helpful and straightforward technique to generate scenarios depending on multiple GCMs, emission scenarios, sensitivities, time horizons, and locations (MRC 2015, 2017). The detailed information in future climate scenarios used for this study was summarized in Table 4.
Future climate scenarios used for this study
GCMs (Model ID) . | Emission Scenario . | Pattern of Change . | Spatial Resolution (longitude x latitude) . |
---|---|---|---|
Geophysical Fluid Dynamics Laboratory Climate model version 3 (GFDL-CM3) | RCP2.6 RCP8.5 | Wetter overall | 2.5°×2° |
Institute Pierre-Simon Laplace Coupled Model, version 5A, coupled with NEMO, mid resolution (IPSL-CM5A-MR) | RCP2.6 RCP8.5 | Increased seasonal variability | 2.5°×1.27° |
Goddard Institute for Space Studies Model E2, coupled with the Russell Ocean model, with the carbon cycle (GISS-E2-R-CC) | RCP2.6 RCP8.5 | Drier overall | 2.5°×2° |
GCMs (Model ID) . | Emission Scenario . | Pattern of Change . | Spatial Resolution (longitude x latitude) . |
---|---|---|---|
Geophysical Fluid Dynamics Laboratory Climate model version 3 (GFDL-CM3) | RCP2.6 RCP8.5 | Wetter overall | 2.5°×2° |
Institute Pierre-Simon Laplace Coupled Model, version 5A, coupled with NEMO, mid resolution (IPSL-CM5A-MR) | RCP2.6 RCP8.5 | Increased seasonal variability | 2.5°×1.27° |
Goddard Institute for Space Studies Model E2, coupled with the Russell Ocean model, with the carbon cycle (GISS-E2-R-CC) | RCP2.6 RCP8.5 | Drier overall | 2.5°×2° |
RESULTS AND DISCUSSION
Streamflow calibration and validation of the SWAT model
Model performance evaluation of daily streamflow during calibration and validation period in the Peam Kley station
Statistical indicators . | Calibration (2000–2006) . | Model Performance . | Validation (2007–2010) . | Model Performance . |
---|---|---|---|---|
NSE | 0.65 | Satisfactory | 0.48 | Unsatisfactory |
R² | 0.66 | Satisfactory | 0.60 | Satisfactory |
PBIAS | −4.79% | Very good | −11.85% | Satisfactory |
RSR | 0.59 | Satisfactory | 0.72 | Good |
Statistical indicators . | Calibration (2000–2006) . | Model Performance . | Validation (2007–2010) . | Model Performance . |
---|---|---|---|---|
NSE | 0.65 | Satisfactory | 0.48 | Unsatisfactory |
R² | 0.66 | Satisfactory | 0.60 | Satisfactory |
PBIAS | −4.79% | Very good | −11.85% | Satisfactory |
RSR | 0.59 | Satisfactory | 0.72 | Good |
Observed and simulated daily streamflow for the Prek Thnot River Basin: (a) calibration (2000–2005) and (b) validation (2007–2010).
Observed and simulated daily streamflow for the Prek Thnot River Basin: (a) calibration (2000–2005) and (b) validation (2007–2010).
Correlation between observed and simulated daily river discharge of Prek Thnot River Basin: (a) calibration period (2000–2005) and (b) validation period (2007–2010).
Correlation between observed and simulated daily river discharge of Prek Thnot River Basin: (a) calibration period (2000–2005) and (b) validation period (2007–2010).
Water balance components in the Prek Thnot River Basin
Water balance components on the Prek Thnot River Basin: Precipitation, Potential Evapotranspiration, Actual Evapotranspiration, Lateral flow, Percolation, Groundwater, Surface runoff, and Water yield of baseline period from 2000 to 2011.
Water balance components on the Prek Thnot River Basin: Precipitation, Potential Evapotranspiration, Actual Evapotranspiration, Lateral flow, Percolation, Groundwater, Surface runoff, and Water yield of baseline period from 2000 to 2011.
The basin is dominated by the tropical monsoon season, distinguished into two seasons, the rainy season from May to October and the dry season from November to April. The monthly contribution of water balance components is high in October during the wet season and contributes low value in February (Table 6). The annual average rainfall during the study period in the Prek Thnot River Basin was 1,169 mm; 72% (848 mm) of the average rainfall was withdrawn by actual evapotranspiration, 23% (272 mm) returned flow to the streamflow, while another 8% recharged to deep aquifer groundwater (Table 7). Water yield of 272 mm has contributed to surface runoff (proportion of 16%), lateral flow (proportion of 48%), and groundwater (proportion of 35%). These results revealed that actual evapotranspiration contributed to the basin's most considerable amount of water loss.
Monthly average water balance components (precipitation, potential evapotranspiration, actual evapotranspiration, surface runoff, lateral flow, and water yield) of the Prek Thnot River Basin for the baseline period from 2000 to 2011
. | PRECIP . | PET . | AET . | SURQ . | LAT_Q . | WYLD . |
---|---|---|---|---|---|---|
Monthly . | (mm) . | |||||
Jan | 16.2 | 148.1 | 48.1 | 1.1 | 1.3 | 6.1 |
Feb | 12.8 | 140.3 | 54.4 | 0.6 | 0.8 | 3.4 |
Mar | 59 | 150.3 | 70.0 | 3.2 | 3.7 | 7.8 |
Apr | 106.4 | 139.0 | 68.4 | 4.6 | 6.7 | 10.4 |
May | 100.7 | 144.9 | 73.6 | 2.3 | 8.6 | 11.6 |
Jun | 102.6 | 128.7 | 70.3 | 2.5 | 7.1 | 9.7 |
Jul | 128.3 | 122.7 | 75.3 | 8.7 | 12.3 | 22.2 |
Aug | 143.4 | 115.7 | 76.7 | 9.3 | 12.2 | 22.6 |
Sep | 167.1 | 104.3 | 79.4 | 14.5 | 19.2 | 39.6 |
Oct | 242.7 | 110.7 | 86.8 | 41.6 | 45.0 | 104.1 |
Nov | 70.5 | 119.1 | 84.1 | 6.4 | 11.8 | 25.8 |
Dec | 19.6 | 134.3 | 61.2 | 1.2 | 2.5 | 8.5 |
. | PRECIP . | PET . | AET . | SURQ . | LAT_Q . | WYLD . |
---|---|---|---|---|---|---|
Monthly . | (mm) . | |||||
Jan | 16.2 | 148.1 | 48.1 | 1.1 | 1.3 | 6.1 |
Feb | 12.8 | 140.3 | 54.4 | 0.6 | 0.8 | 3.4 |
Mar | 59 | 150.3 | 70.0 | 3.2 | 3.7 | 7.8 |
Apr | 106.4 | 139.0 | 68.4 | 4.6 | 6.7 | 10.4 |
May | 100.7 | 144.9 | 73.6 | 2.3 | 8.6 | 11.6 |
Jun | 102.6 | 128.7 | 70.3 | 2.5 | 7.1 | 9.7 |
Jul | 128.3 | 122.7 | 75.3 | 8.7 | 12.3 | 22.2 |
Aug | 143.4 | 115.7 | 76.7 | 9.3 | 12.2 | 22.6 |
Sep | 167.1 | 104.3 | 79.4 | 14.5 | 19.2 | 39.6 |
Oct | 242.7 | 110.7 | 86.8 | 41.6 | 45.0 | 104.1 |
Nov | 70.5 | 119.1 | 84.1 | 6.4 | 11.8 | 25.8 |
Dec | 19.6 | 134.3 | 61.2 | 1.2 | 2.5 | 8.5 |
PRECIP, precipitation; AET, actual evapotranspiration; PET, potential evapotranspiration; SURQ, surface runoff contribution to streamflow; LAT_Q, lateral flow; and WYLD, the net amount of water that contributes to streamflow (SURQ+LAT_Q+groundwater contribution to streamflow – transmission losses).
Annual average water balance components of the Prek Thnot River Basin for the baseline period (2000–2011)
Water Balance Component . | Amount (mm) . |
---|---|
Precipitation | 1,169 |
Evapotranspiration | 848 |
Surface runoff | 96 |
Lateral flow | 131 |
Groundwater (Shallow Aquifer) flow | 85 |
Groundwater (Deep Aquifer) flow | 34 |
Groundwater contribution to streamflow (Baseflow) | 119 |
Water yield | 272 |
Water Balance Component . | Amount (mm) . |
---|---|
Precipitation | 1,169 |
Evapotranspiration | 848 |
Surface runoff | 96 |
Lateral flow | 131 |
Groundwater (Shallow Aquifer) flow | 85 |
Groundwater (Deep Aquifer) flow | 34 |
Groundwater contribution to streamflow (Baseflow) | 119 |
Water yield | 272 |
Impact of climate change on water balance components
Projected annual change of the water balance components at a basin scale
Projected annual average water balance components and its percentage change compared to the baseline condition (from 2000 to 2011) based on GCMs under emission scenarios RCP2.6 and RCP8.5 in the 2030s and 2060s for Prek Thnot River Basin
Scenario . | Time Period . | GCMs Model . | PRECIP . | AET . | SURQ . | LAT_Q . | GW_Q . | WYLD . |
---|---|---|---|---|---|---|---|---|
(mm/year) . | ||||||||
Projected Annual Water Balance Components for Prek Thnot River Basin | ||||||||
Baseline | 2000–2011 | 1,169 | 848 | 96 | 131 | 119 | 272 | |
RCP 2.6 | 2030s | GFDL | 1,192 | 859 | 102 | 136 | 124 | 282 |
(2021–2040) | GISS | 1,144 | 844 | 87 | 123 | 109 | 253 | |
IPSL | 1,166 | 847 | 95 | 131 | 118 | 270 | ||
2060s | GFDL | 1,200 | 863 | 104 | 135 | 125 | 285 | |
(2051–2070) | GISS | 1,136 | 842 | 85 | 125 | 106 | 248 | |
IPSL | 1,165 | 846 | 94 | 131 | 118 | 270 | ||
RCP 8.5 | 2030s | GFDL | 1,259 | 892 | 119 | 144 | 137 | 311 |
(2021–2040) | GISS | 1,072 | 826 | 65 | 108 | 83 | 205 | |
IPSL | 1,157 | 840 | 93 | 130 | 115 | 267 | ||
2060s | GFDL | 1,362 | 936 | 140 | 158 | 159 | 360 | |
(2051–2070) | GISS | 960 | 777 | 46 | 86 | 60 | 147 | |
IPSL | 1,140 | 837 | 89 | 129 | 112 | 265 | ||
Annual Change of Water Balance Components for Prek Thnot River Basin | ||||||||
RCP 2.6 | 2030s | GFDL | 2.0 | 1.3 | 6.3 | 3.4 | 4.2 | 3.7 |
(2021–2040) | GISS | −2.1 | −0.5 | −9.4 | −6.1 | −8.4 | −7.0 | |
IPSL | −0.3 | −0.1 | −1.0 | −0.3 | −0.8 | −0.7 | ||
2060s | GFDL | 2.7 | 1.8 | 8.3 | 2.6 | 5.0 | 4.8 | |
(2051–2070) | GISS | −2.8 | −0.7 | −11.5 | −4.7 | −10.9 | −8.8 | |
IPSL | −0.4 | −0.3 | −1.8 | −0.3 | −1.3 | −0.9 | ||
RCP 8.5 | 2030s | GFDL | 7.7 | 5.2 | 24.0 | 9.8 | 15.1 | 14.3 |
(2021–2040) | GISS | −8.3 | −2.6 | −32.3 | −17.5 | −30.3 | −24.6 | |
IPSL | −1.0 | −0.9 | −3.1 | −0.9 | −3.4 | −1.8 | ||
2060s | GFDL | 16.5 | 10.4 | 45.8 | 20.6 | 33.6 | 32.4 | |
(2051–2070) | GISS | −17.9 | −8.4 | −52.1 | −34.3 | −49.6 | −46.0 | |
IPSL | −2.5 | −1.3 | −7.3 | −1.2 | −5.9 | −2.6 |
Scenario . | Time Period . | GCMs Model . | PRECIP . | AET . | SURQ . | LAT_Q . | GW_Q . | WYLD . |
---|---|---|---|---|---|---|---|---|
(mm/year) . | ||||||||
Projected Annual Water Balance Components for Prek Thnot River Basin | ||||||||
Baseline | 2000–2011 | 1,169 | 848 | 96 | 131 | 119 | 272 | |
RCP 2.6 | 2030s | GFDL | 1,192 | 859 | 102 | 136 | 124 | 282 |
(2021–2040) | GISS | 1,144 | 844 | 87 | 123 | 109 | 253 | |
IPSL | 1,166 | 847 | 95 | 131 | 118 | 270 | ||
2060s | GFDL | 1,200 | 863 | 104 | 135 | 125 | 285 | |
(2051–2070) | GISS | 1,136 | 842 | 85 | 125 | 106 | 248 | |
IPSL | 1,165 | 846 | 94 | 131 | 118 | 270 | ||
RCP 8.5 | 2030s | GFDL | 1,259 | 892 | 119 | 144 | 137 | 311 |
(2021–2040) | GISS | 1,072 | 826 | 65 | 108 | 83 | 205 | |
IPSL | 1,157 | 840 | 93 | 130 | 115 | 267 | ||
2060s | GFDL | 1,362 | 936 | 140 | 158 | 159 | 360 | |
(2051–2070) | GISS | 960 | 777 | 46 | 86 | 60 | 147 | |
IPSL | 1,140 | 837 | 89 | 129 | 112 | 265 | ||
Annual Change of Water Balance Components for Prek Thnot River Basin | ||||||||
RCP 2.6 | 2030s | GFDL | 2.0 | 1.3 | 6.3 | 3.4 | 4.2 | 3.7 |
(2021–2040) | GISS | −2.1 | −0.5 | −9.4 | −6.1 | −8.4 | −7.0 | |
IPSL | −0.3 | −0.1 | −1.0 | −0.3 | −0.8 | −0.7 | ||
2060s | GFDL | 2.7 | 1.8 | 8.3 | 2.6 | 5.0 | 4.8 | |
(2051–2070) | GISS | −2.8 | −0.7 | −11.5 | −4.7 | −10.9 | −8.8 | |
IPSL | −0.4 | −0.3 | −1.8 | −0.3 | −1.3 | −0.9 | ||
RCP 8.5 | 2030s | GFDL | 7.7 | 5.2 | 24.0 | 9.8 | 15.1 | 14.3 |
(2021–2040) | GISS | −8.3 | −2.6 | −32.3 | −17.5 | −30.3 | −24.6 | |
IPSL | −1.0 | −0.9 | −3.1 | −0.9 | −3.4 | −1.8 | ||
2060s | GFDL | 16.5 | 10.4 | 45.8 | 20.6 | 33.6 | 32.4 | |
(2051–2070) | GISS | −17.9 | −8.4 | −52.1 | −34.3 | −49.6 | −46.0 | |
IPSL | −2.5 | −1.3 | −7.3 | −1.2 | −5.9 | −2.6 |
PRECIP, precipitation; AET, actual evapotranspiration; SURQ, surface runoff contribution to streamflow; LAT_Q, lateral flow; GW_Q, groundwater contribution to streamflow; and WYLD, the net amount of water that contributes to streamflow (SURQ+LAT_Q+GW_Q – transmission losses).
Projected changes of water balance components in different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline in the 2030s (left) and 2060s (right).
Projected changes of water balance components in different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline in the 2030s (left) and 2060s (right).
Under the projected climate scenarios, the GISS-E2-R-CC was projected to decrease between 2 and 18% (precipitation), 0.5–8% (actual evapotranspiration), 9–52% (surface runoff), 5–34% (lateral flow), 8–50% (groundwater flow), and 7–46% (water yield) (Figure 6). Likewise, a considerable reduction in recharges to deep aquifers is also predicted. On the other hand, the IPSL-CM5A-MR was also expected to decrease by 0.3–2.5% (precipitation), 0.1–1% (actual evapotranspiration), 1–7% (surface runoff), 0.8–4% (lateral flow), 0.8–6% (groundwater flow), and 0.7–3% (water yield). However, the GFDL-CM3 is projected to increase in range from 2–17%, 1–10%, 6–46%, 3–21%, 4–34%, and 4–32% of precipitation, actual evapotranspiration, surface runoff, lateral flow, groundwater flow and water yield, respectively. Generally, the GFDL-CM3 model exhibited an increase in annual streamflow, while the GISS-E2-R-CC and IPSL-CM5A-MR models revealed a significant decrease for all time slices. In this study, annual precipitation in the Preck Thnot River Basin was projected to increase by 8–17% for the GCMs under the two RCP scenarios, for the near and medium term, respectively, compared to the previous study; it was 5–11% according Sok et al. (2022) in Sen River Basin, the largest sub-basin of Tonle Sap Lake in the Lower Mekong Basin.
Future precipitation at the sub-basin scale
Changes in annual average precipitation under different future projection in the 2030s.
Changes in annual average precipitation under different future projection in the 2030s.
Changes in annual average precipitation under different future projection in the 2060s.
Changes in annual average precipitation under different future projection in the 2060s.
Future water yield at the sub-basin scale
Changes in annual average water yield under different future projections in the 2030s.
Changes in annual average water yield under different future projections in the 2030s.
Changes in annual average water yield under different future projections in the 2060s.
Changes in annual average water yield under different future projections in the 2060s.
Impact of climate change on the flow regime
The comparison of baseline and future monthly streamflow in different GCMs (top) and the range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios in the 2030s (left) and 2060s (right).
The comparison of baseline and future monthly streamflow in different GCMs (top) and the range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios in the 2030s (left) and 2060s (right).
Projected change of monthly streamflow in different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline in the 2030s (left) and 2060s (right).
Projected change of monthly streamflow in different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline in the 2030s (left) and 2060s (right).
Changes in hydrological extremes
Climate changes can dramatically change river flow regimes and extreme events such as floods and droughts. The high (Q5, flow equals or exceeds 5% of the time) and low (Q95, flow equals or exceeds 95% of the time) flows were extracted from flow duration curves derived from simulated daily flows. The Q5 and Q95 provide some pieces of evidence to prove that streamflow is expected to increase in magnitude through GFDL-CM3 and IPSL-CM5-MR (except GISS-E2-R-CC) under both RCP scenarios in the 2030s and 2060s (Table 9). The previous studies were conducted using the same GCMs model under RCP2.6 and RCP8.5 scenarios in the 2030s and 2060s in the Sen River Basin (Heng et al. 2020; Sok et al. 2022). Also, Oeurng et al. (2019) and Touch et al. (2020) utilized the same GCMs model but different emission rates (RCP6.0) in all rivers of the Tonle Sap Lake Basin and Pursat River Basin, respectively. The results showed that there are more incredible percentage changes in streamflow in the 2060 than in the 2030s, and the projected changes under RCP8.5 are more extensive than those by RCP2.6.
Absolute and percentage changes in future flows, relative to the baseline period for the three GCMs under RCP2.6 and RCP8.5 in the 2030s and 2060s
Percentage changes under RCP2.6 . | ||||||||
---|---|---|---|---|---|---|---|---|
Time Horizons . | Exceedance Probability . | Baseline . | GFDL-CM3 . | GISS-E2-R-CC . | IPSL-CM5A-MR . | |||
Years . | (%) . | (m3/s) . | (m3/s) . | %Change . | (m3/s) . | %Change . | (m3/s) . | %Change . |
2030s | 5 | 944 | 1,000 | 6 | 884 | −6 | 952 | 1 |
95 | 182 | 183 | 1 | 165 | −9 | 187 | 3 | |
2060s | 5 | 944 | 986 | 4 | 894 | −5 | 949 | 1 |
95 | 182 | 185 | 2 | 169 | −7 | 186 | 2 | |
Percentage changes under RCP8.5 . | ||||||||
. | . | . | RCP8.5 . | |||||
Time Horizons . | Exceedance Probability . | Baseline . | GFDL-CM3 . | GISS-E2-R-CC . | IPSL-CM5A-MR . | |||
Years . | (%) . | (m3/s) . | (m3/s) . | %Change . | (m3/s) . | %Change . | (m3/s) . | %Change . |
2030s | 5 | 944 | 1,080 | 14 | 776 | −18 | 975 | 3 |
95 | 182 | 186 | 2 | 127 | −30 | 189 | 4 | |
2060s | 5 | 944 | 1,400 | 48 | 609 | −35 | 1,020 | 8 |
95 | 182 | 224 | 23 | 71 | −61 | 204 | 12 |
Percentage changes under RCP2.6 . | ||||||||
---|---|---|---|---|---|---|---|---|
Time Horizons . | Exceedance Probability . | Baseline . | GFDL-CM3 . | GISS-E2-R-CC . | IPSL-CM5A-MR . | |||
Years . | (%) . | (m3/s) . | (m3/s) . | %Change . | (m3/s) . | %Change . | (m3/s) . | %Change . |
2030s | 5 | 944 | 1,000 | 6 | 884 | −6 | 952 | 1 |
95 | 182 | 183 | 1 | 165 | −9 | 187 | 3 | |
2060s | 5 | 944 | 986 | 4 | 894 | −5 | 949 | 1 |
95 | 182 | 185 | 2 | 169 | −7 | 186 | 2 | |
Percentage changes under RCP8.5 . | ||||||||
. | . | . | RCP8.5 . | |||||
Time Horizons . | Exceedance Probability . | Baseline . | GFDL-CM3 . | GISS-E2-R-CC . | IPSL-CM5A-MR . | |||
Years . | (%) . | (m3/s) . | (m3/s) . | %Change . | (m3/s) . | %Change . | (m3/s) . | %Change . |
2030s | 5 | 944 | 1,080 | 14 | 776 | −18 | 975 | 3 |
95 | 182 | 186 | 2 | 127 | −30 | 189 | 4 | |
2060s | 5 | 944 | 1,400 | 48 | 609 | −35 | 1,020 | 8 |
95 | 182 | 224 | 23 | 71 | −61 | 204 | 12 |
Flow duration curves in different GCMs (top) and ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 emission scenarios compared with the baseline in the 2030s (left) and 2060s (right).
Flow duration curves in different GCMs (top) and ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 emission scenarios compared with the baseline in the 2030s (left) and 2060s (right).
Percentage changes of Q5 and Q95 in different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline during the future periods in the 2030s (left) and 2060s (right).
Percentage changes of Q5 and Q95 in different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline during the future periods in the 2030s (left) and 2060s (right).
For the Q95, in the 2030s, streamflow is predicted to decrease about 9 and 30% by the GISS-E2-R-CC model and, at the same time, to increase from 1–2% and 3–4% by the other two models for RCPs, respectively. In the 2060s, Q95 is expected to increase and decrease at a higher magnitude than in the 2030s. It would increase from 2 to 23% (GFDL-CM3) and 2–12% (IPSL-CM5A-MR), but it would decrease by about 7 and 61% for the GISS-E2-R-CC model for RCP2.6 and RCP8.5, respectively. The significant decrease in Q95 projected by the GISS-E2-R-CC model indicates that the drought event would significantly increase in the 2030s and 2060s. These projected decreases in Q95 would have implications for future drought for policymakers to set up measures for sustaining ecosystems and biodiversity during the dry season. In this study, the GISS-E2-R-CC model projected a decreasing trend of Q5 and Q95 under most RCP scenarios in most future periods, indicating that both the low and high flows will be lower than their baseline values similar to the previous studies (Oeurng et al. 2019; Heng et al. 2020; Touch et al. 2020; Sok et al. 2022).
Changes in frequency and magnitude of the annual peak flow
The frequency and magnitude of the annual peak flow for Prek Thnot River Basin in the 2030s and 2060s were conducted using the Consolidated Frequency Analysis (CFA) under three different GCMs and two emission scenarios (RCPs). This study simulated daily streamflow under climate change scenarios for 2-, 5-, 10-, 20-, 50-, and 100-year return periods, which correspond to the Sok et al. (2022). Based on Pradhanang et al. (2013), 2-year return periods are responsible for small floods, 10-year return periods are for moderate floods, and 100-year return periods are for large floods. The GCMs ensemble means were used to quantify the average changes in different time slices.
Projection of peak flow return period curves with different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline for the future periods in the 2030s (left) and 2060s (right).
Projection of peak flow return period curves with different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline for the future periods in the 2030s (left) and 2060s (right).
Percentage change of annual peak flow for different return periods with different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline for the future periods in the 2030s (left) and 2060s (right).
Percentage change of annual peak flow for different return periods with different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline for the future periods in the 2030s (left) and 2060s (right).
Changes in frequency and magnitude of the annual low flow
Projection of base flow return period curve with different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline for the future periods in the 2030s (left) and 2060s (right).
Projection of base flow return period curve with different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline for the future periods in the 2030s (left) and 2060s (right).
Percentage change of baseflow or low flow for different return periods with different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline for the future periods in the 2030s (left) and 2060s (right).
Percentage change of baseflow or low flow for different return periods with different GCMs (top) and range of projected change represented as an ensemble of selected GCMs outputs (bottom) under RCP2.6 and RCP8.5 scenarios compared with the baseline for the future periods in the 2030s (left) and 2060s (right).
It is important to note that changes in flow regime, increasing peak flow, and decreasing low flow could threaten river ecosystems and socio-economic development, particularly in agriculture and aquaculture. However, climate change is not the only driving factor for the water balance components and hydrological extremes. Land-use change and other human activities (e.g., dam, reservoir, withdrawal, diversion, etc.) in the river system also play a crucial role in such hydrological extremes and could be the research interest for further investigations.
CONCLUSIONS
This study is the first and most detailed climate change assessment of the Prek Thnot River Basin's hydrology. The SWAT model was applied to assess the impacts of climate change on water balance components and extreme hydrological events in the Prek Thnot River Basin. The model was successfully built with good performance to examine the watershed properly. The hydrological response of the basin was simulated using the daily streamflow for three GCMs under two RCP scenarios for the near future (2021–2040) and mid-future (2051–2070).
Under climate change scenarios, the annual water balance components mostly decrease spatially across the basin. Individual GCMs and scenarios showed considerable differences in patterns and magnitude changes in spatiotemporal precipitation and total water yield. The ensemble means scenario showed decreases from the baseline condition in annual precipitation and total water yield under both RCP scenarios for the near- and mid-future periods.
The results further found the changes in the annual distribution of streamflow of the Prek Thnot River Basin. Climate change profoundly affecting the flow regime would be positively altered (increase compared to the baseline) with the GFDL-CM3 model under both RCPs in all future periods; at the same time, GISS-E2-R-CC and IPSL-CM5A-MR models would be negatively altered (decrease compared to the baseline). Generally, future monthly streamflow was projected to increase during the wet season while decreasing during the dry season.
Through the effect of climate change on hydrological extremes, peak flow would be more substantial in magnitude and frequency with GFDL-CM3 and IPSL-CM5A-MR models under both RCP scenarios for all future periods. However, the low flow with GISS-E2-R-CC and IPSL-CM5A-MR models is projected to decrease significantly in both magnitude and frequency. Under these projected changes, simulated peak flow increased slightly from the near to mid-future under both RCPs, while the low flow decreased. Additionally, extreme peak flow events increased in both magnitude and frequency, and extremely low flows are projected to occur less often under climate change.
This study discussed that the prediction results depend on the climate model used. The increasing peak flow and decreasing low flow could threaten river ecosystems and socio-economic development, particularly in agriculture. However, climate change is not the only driving factor for the water balance components and hydrological extremes. Land-use change and other affected by human impact (e.g., dam, reservoir, withdrawal, diversion, etc.) in the river system also play a crucial role in such hydrological extremes, which can be a part of the further investigation caused by these factors that were not included in the study. This study suggests the careful and efficient selection of structural and non-structural measures and water management plans to mitigate the impacts of climate change in the water resources system.
ACKNOWLEDGEMENTS
The authors would like to acknowledge the Cambodia Higher Education Improvement Project (Credit No.6221-KH) of the Ministry of Education, Youth and Sports (MOEYS) for financial support of the study. The authors would also like to thank the Ministry of Water Resources and Meteorology (MOWRAM) and the Mekong River Commission (MRC) of Cambodia for providing the data.
DATA AVAILABILITY STATEMENT
Data cannot be made publicly available; readers should contact the corresponding author for details.
CONFLICT OF INTEREST
The authors declare there is no conflict.