We use the MIKE HYDRO BASIN model developed by the Danish Hydraulic Institute (DHI) to assess the balance and allocation of water resources in the Srepok River Basin. This model was inherited from the Srepok River Basin Integrated Planning project [49] . The model’s inputs include the water demand of sectors (including environmental flows), flow data in sub-basins, reservoir parameters, and hydropower works. These are operated according to the basin’s inter-reservoir operation process, regulated under Decision No. 1201/QD-TTg, dated July 24, 2014, of the Prime Minister. The basin model includes 58 sub-basins and water used for domestic, industrial, irrigation, husbandry farming, aquacultural, and tourism purposes in Viet Nam ( Figure 1 ). Water demand for domestic, industrial, livestock, aquacultural, and tourism purposes is calculated according to Viet Nam’s standards and estimated for 2017, assuming this demand will stay the same over the years. Agricultural irrigation demand is calculated according to crop growth stages for rice and the FAO 56 Dual Crop Coefficient method for other crops (corn, sugarcane, potato, cassava, coffee, and fruit trees). Crop irrigation demands are established in the MIKE HYDRO BASIN model, depending on annual rainfall. The total water demand was about 2301 million m³ in 2017 [49] . Environmental flows are approximately 10% of the water demand of all sectors, corresponding to approximately 230 million m³ for 2017 [49] . Environmental flows are assumed to be constant across calculation years.

The water scarcity index is calculated using the following formula [14] [51] .

$WSI=\frac{WU}{MAR-EWR}$ (1)

in which: WU: Water demand;

MAR: Total available quantity;

EWR: Environmental flow.

Accordingly, WSI ≥ 1 is considered scarce in raw water resources, which represents an over-exploited basin; 0.6 ≤ WSI < 1 is an environmentally stressed basin; 0.3 ≤ WSI < 0.6 is a moderately exploited basin; WSI < 0.3 is an environmentally safe basin.

This study used the 2016 climate change scenarios of Viet Nam [52] to project the WSI of the basin for 2065. The period of 1986-2005 was used as the baseline to evaluate the level of water scarcity by month. The RCP4.5 scenario was employed to assess changes in streamflow using the MIKE NAM model derived from the water balance model set for the Srepok River Basin [49] . Table 1 presents projected changes in seasonal rainfall in the basin under the RCP 4.5 scenario for the two periods, 2016-2035 and 2046-2065.

Water demand for the domestic, industrial, livestock, aquacultural, and tourism sectors was determined based on baseline year data and development scenarios for the respective periods [49] . The population is projected to increase by an average of 1.1% - 1.5%/year, the industrial production value is expected to grow at 14%/year, and the livestock growth rate is expected to be about 1%/year. Cultivation areas are projected to continue expanding until 2020 and then remain stable. Water demand for agricultural irrigation will fluctuate annually based on rainfall, assuming the cultivation area remains constant at the 2020 level.

Using cultivation area data for 2017 and the water demand of the remaining sectors to calibrate and verify the MIKE HYDRO BASIN model for the basin, we set the calibration period to 2005-2010 and the verification period to 2011-2015. The monthly average flow series was calibrated and verified in the MIKE HYDRO BASIN model. Observed and simulated flow data at the Cau 14, Giang Son, and

Table 1. Seasonal rainfall changes in the Srepok River Basin under the RCP 4.5 scenario [52] .

Ban Don stations were used for the evaluation according to the criteria. The calibration and verification criteria were the Nash-Sutcliffe coefficient (NSE), BIAS, and the correlation coefficient (R).

$\text{NSE}=1-\frac{{{\displaystyle \sum }}_{k=1}^n{\left(X_k^{Sim}-X_k^{Obs}\right)}^2}{{{\displaystyle \sum }}_{k=1}^n{\left(X_k^{Obs}-{\bar{X}}^{Obs}\right)}^2}$ (2)

$\text{Bias}=\frac{1}{n}{\displaystyle {\sum }_{t=1}^n\left(X_{Sim,t}-X_{Obs,t}\right)}$ (3)

$R=\frac{{{\displaystyle \sum }}_{t=1}^n\left(X_t^{Obs}-{\bar{X}}^{Obs}\right)\left(X_t^{Sim}-{\bar{X}}^{Sim}\right)}{\sqrt{{{\displaystyle \sum }}_{t=1}^n{\left(X_t^{Obs}-{\bar{X}}^{Obs}\right)}^2{{\displaystyle \sum }}_{t=1}^n{\left(X_t^{Sim}-{\bar{X}}^{Sim}\right)}^2}}$ (4)

A comparison of the simulated and observed results at the Cau 14, Giang Son, and Ban Don stations ( Table 2 ) shows a strong correlation between the values.

Figure 3 presents the flow simulation results at Ban Don station, highlighting the similarity between the observed and simulated data. After calibration and

Table 2. Model calibration and validation statistics for Cau 14, Giang Son and Ban Don stations.

verification, the model demonstrates good performance, making it suitable for simulating flow in the baseline period and two projected periods, 2016-2035 and 2046-2065, based on climate change scenarios.

The baseline period for our calculations was from 1986 to 2005, with assessments conducted monthly. The level of water scarcity was estimated each month for each year within this timeframe. Average monthly values were used to determine the monthly water scarcity levels across the Srepok River Basin sub-basins during the baseline period (1986-2005) ( Figure 4 ). The areas most affected by water stress (0.6 ≤ WSI < 1) and water scarcity (WSI ≥ 1) are primarily located in the sub-basins of the Ea H’leo, Ea Krong Ana Rivers and part of the Srepok tributary, mainly within Dak Lak province. Water scarcity can be predominantly observed in February, March, and April, with the lowest average flow of the year.

During the 1986-2005 period, at least 32/58 sub-basins (0.83 million people) were affected by water scarcity on average for at least 1 month per year, and 25/58 sub-basins (0.65 million people) were affected by water scarcity for at least 3 months ( Figure 5 , Table 3 ), with about 0.83 million people (46% of the population) affected in total. These sub-basins belong to three main river branches: the Ea Hleo River, the Ea Krong Ana River, and the Srepok River. Water scarcity mainly occurs during the dry season, from late November to the end of April. In years with low flows in the dry season, the water scarcity level increases. For example, in 1995, 35/58 sub-basins had water scarcity for at least 1 month, and 31/58 sub-basins had water scarcity for at least 3 months; in 1998, 34/58 sub-basins had water scarcity for at least 1 month, and 29/58 sub-basins had water scarcity for at least 3 months; and in 2005, 35/58 sub-basins had water scarcity for at least 1 month, and 31/58 sub-basins had water scarcity for at least 3 months.

The proportion of water demand in the Srepok River Basin indicates that the agricultural sector had the most prominent water demand from 1986 to 2005, accounting for 95% of the total demand in the basin. The total water shortage in this period was about 839.1 million m³, accounting for a 48.6% deficit, to which water shortage during the dry season contributed 51.6%. Agriculture experienced a significant water shortage of approximately 836 million m³ ( Table 4 ), presenting a shortfall of about 51.1%. Among 58 sub-basins, sub-basins in Ea Hleo River, Ea Krong Ana River, and Srepok River, which are characterized by intensive agricultural development, experienced the highest water shortage rates at 65.19%, 54.98%, and 26.19%, respectively ( Figure 6 ).

These findings align with research conducted by Tran Van Ty and colleagues in 2012 [48] , which indicated that most sub-basins in the Srepok River Basin had faced specific water shortage rates, especially in dry years from 1997 to 2007. The agricultural water deficit was high. However, the water shortage levels in that study were considerably lower than those in our findings, which revealed shortfalls of roughly 19.4% and 28.6% in the dry season. The difference between the two studies can be explained by two main factors: 1) Including environmental flow in our water scarcity assessment and 2) variations in the average dry season flow at Ban Don station. The average flow from 1997 to 2007 was 121.5 m³/s, higher than the 106.2 m³/s from 1986 to 2005.

On the one hand, this difference highlights the importance of considering environmental flow in water resource management to accurately reflect an area’s total water requirements and emphasizes the need to balance economic development

Table 3. Number of sub-basins experiencing water scarcity for one and three months in the tributaries of the Srepok River.

Table 4. Projected water demand and shortage in the 2016-2035 and 2046-2065 periods.

with environmental conservation. On the other hand, the flow during the dry season significantly affects the water shortage rate. Therefore, integrating environmental flow requirements in water scarcity assessments with accurate dry season flow forecasts is crucial for developing effective strategies to address water scarcity. This holistic approach promotes sustainable water management and improves preparedness for future challenges related to water scarcity.

We examined both the RCP4.5 Climate Change scenario for the two periods—2016 - 2035 and 2046 - 2065—and their associated socio-economic development scenarios to assess the level of water scarcity in the Srepok River Basin. Both periods project increased rainfall in the basin. From 2016 to 2035, rainfall increased by an average of 6%/year, and from 2046 to 2065, it increased by an average of 8.24%/year. Compared with the baseline period, the average annual flow increases by 15.3% from 2016 to 2035 (the dry season increases by 12.2%) and 29.4% from 2046 to 2065 (the dry season increases by 23.5%). Increased flow is more significant than increased rainfall in these two periods owing to construction and reservoirs that will increase this factor in the dry season [53] [54] .

The two periods have different water use demands, altering the basin’s flow and leading to changes in water scarcity. In both periods, water scarcity increases ( Figure 5 , Table 4 ). Water scarcity occurs in January and July in both periods, whereas there is no water scarcity during the baseline period. Areas with similar water scarcity levels in the baseline period are mainly concentrated in sub-basins of the Ea Hleo and Ea Krong Ana Rivers and part of the Srepok River branch. In both periods, about 1.4 million people are affected by water scarcity for at least one month, and about 1 million are affected for at least three months.

The total water demand also increases compared with the baseline period. In the 2016-2035 period, the demand is about 3376.76 million m³/year (1.95 times), and in the 2046-2065 period, it is about 3801.39 million m³/year (2.19 times) ( Table 4 ). This change is due to variations in different socio-economic conditions for these periods. Water demand in the 2046-2065 period decreases compared with the 2016-2035 period because, in the former, the crop area does not change much, and increased rainfall in the basin reduces the demand for water for agriculture. The water shortage levels significantly rise in these two periods, of which the agricultural sector still accounts for the highest proportion ( Table 4 ).

The total water shortage levels for the agricultural sector increase compared with the baseline in both periods: the 2016-2035 period has a shortage of 1787.24 million m³ (67.2%), and the 2016-2035 period has a shortage of 1,406.75 million m³ (60.58%) ( Table 3 ). In the main river branches, the agricultural water shortage significantly increases compared with the baseline and is most severe in the 2016-2035 period ( Figure 6 ).

The Srepok River Basin has experienced high-density agricultural intensification, resulting in significant water demand and a substantial water shortage. According to these climate change and socio-economic development scenarios, agricultural water shortages are estimated to exceed 60%. This water shortage significantly affects crop productivity and the socio-economic development of the region, so it is necessary to have measures and policies in place to minimize these impacts, such as water demand management strategies, water reuse, water storage measures, and construction (reservoirs and dams). We should also apply efficient water-saving irrigation technologies such as drip irrigation.

One limitation of our study is that our climate change scenarios do not consider the future development of reservoirs, hydropower projects, and new operating procedures, which may change water scarcity levels in the basin. In future studies, we will evaluate measures to reduce water scarcity and its damage to the socio-economic well-being of the Srepok River Basin. In addition, the impacts of irrigation, hydropower, and reservoir projects on water scarcity were not considered. We will supplement these assessments in subsequent studies.