Climate Change-Driven Hydrological Shifts in the Kon-Ha Thanh River Basin

Abstract

Climate change is projected to bring substantial changes to hydroclimatic extremes, which will affect natural river regimes and have wide-ranging impacts on human health and ecosystems, particularly in Central Highland Vietnam. This study focuses on understanding and quantifying the projected impacts of climate change on streamflow in the Kon-Ha Thanh River basin, using the Soil and Water Assessment Tool (SWAT) between 2016 and 2099. The study examined projected changes in streamflow across three time periods (2016–2035, 2046–2065, and 2080–2029) under two scenarios, Representative Conversion Pathways (RCPs) 4.5 and 8.5. The model was developed and validated on a daily scale with the model performance, yielding good performance scores, including Coefficient of Determination (R²), Nash-Sutcliffe Efficiency (NSE), and Root Mean Squared Error (RMSE) values of 0.79, 0.77, and 50.96 m³/s, respectively. Our findings are (1) streamflow during the wet season is projected to increase by up to 150%, particularly in December, under RCP 8.5; (2) dry season flows are expected to decrease by over 10%, beginning in May, heightening the risk of water shortages during critical agricultural periods; and (3) shifts in the timing of flood and dry seasons are found toward 2099 that will require adaptive measures for water resource management. These findings provide a scientific foundation for incorporating climate change impacts into regional water management strategies and enhancing the resilience of local communities to future hydroclimatic challenges.

1. Introduction

In recent years, Vietnam has faced significant challenges due to increasing climate extremes, reflecting a global trend of climate-related disruptions [1,2,3]. Vietnam is among the countries most vulnerable to the changing climate [4,5] that could be further exacerbated by human intervention (e.g., dam development, land use/land cover (LULC) change, urbanization). Therefore, the critical need for effective responses to climate change’s impact is considered one of the prioritized missions in recent decades. Over the past century, rising temperatures and shifting precipitation patterns have been observed [4]. Over the past five decades, average annual temperatures have risen by approximately 0.5 to 0.7 °C. While annual precipitation has declined in northern regions, it has increased in southern areas, and sea levels have experienced a 20 cm rise [4]. However, according to the Intergovernmental Panel on Climate Change (IPCC), global temperatures are predicted to increase further by the end of the century under various Greenhouse Gas (GHG) emission scenarios [6]. This warming is likely to result in increased hydrological extremes, altered rainfall patterns [7], and higher evaporation rates, leading to heightened risks of drought [8,9] and flooding [10,11]. Since 2016, climate change has profoundly impacted Vietnam, with more than 7 tropical depressions and 10 typhoons in the East Sea, exceeding previous averages. These changes are having serious effects on the economy [12,13], human lives, and the environment [14]. Vietnam’s National Climate Change Strategy (NCCS) is a climate action plan that outlines the nation’s endeavors to mitigate climate change and reduce greenhouse gas emissions, underscoring the need for data-driven strategies to manage changing hydrologic conditions [15,16]. Due to the complexity of climate change and the limited understanding of its effects, Vietnam is striving to develop adaptive measures to manage extreme weather events and enhance resilience [5]. Specifically, one of the most significant consequences of climate change is on regional hydrologic conditions, leading to increased severity of climate extremes, changes in water regimes, and more frequent flood events. These impacts could severely pressure human well-being, agricultural productivity, recovery costs after disasters, and efforts to maintain freshwater supplies [17]. These trends are also evident at the regional level, where river basins such as the Kon-Ha Thanh are facing intensified hydrological extremes. Therefore, understanding how climate change affects regional water resources is crucial for planning strategies to mitigate its impact.

The Kon-Ha Thanh rivers in Vietnam suffer from severe flooding during the rainy season and are incredibly low flow during the dry season [18]. Climate change has intensified these issues, increasing the basin’s vulnerability to flash floods from intense storms and annual typhoons. Vietnam’s National Steering Committee for Natural Disaster Prevention and Control reported, in 1999, that the flooding of the Kon River inundated over 30,000 households under 1 to 3 m of water, disrupting local transportation. Additionally, climate change is expected to worsen these effects, leading to more frequent and severe flooding and drought events [5]. During the dry season, low rainfall, prolonged heat, and poor water storage capacity, steep slopes in riverbeds lead to water shortages and increased salinization in agricultural production. The most vulnerable areas, particularly in the Kon and Ha Thanh estuaries, were found to be experiencing these impacts [18]. Moreover, prolonged drought periods negatively impact livestock and poultry, heightening the likelihood of disease outbreaks during warmer seasons [19]. These effects are anticipated to worsen as climate change continues. Additionally, Quy Nhon—a downstream city of the Kon-Ha Thanh River—is expected to face severe flooding, which disproportionately impacts low-income communities, including farmers and fishermen. This situation poses major socio-economic challenges, such as reduced agricultural productivity and damage to fishing infrastructure [19,20]. Currently, local authorities and communities face numerous challenges, including limited access to hydrological data [5,21], financial constraints, and the need for sustainable agricultural practices to adapt to climate change impacts [19,21]. Therefore, assessing the impacts of climate change and proposing appropriate response measures are important for developing effective strategies for this region [22]. However, there is a lack of comprehensive studies on climate impacts in this basin, hindering the ability of authorities and stakeholders to develop effective solutions against natural disasters.

Nowadays, numerical models have frequently been used to investigate the impacts of climate change. Specifically, outputs from General Circulation Models (GCMs) are widely used in hydrologic simulations to assess changes in streamflow and other hydrological outcomes [12,13]. Additionally, hydrological models, including distributed and lumped types, have been developed to simulate watershed processes. However, selecting an appropriate model is crucial for accurately capturing watershed responses to climate change. The Soil and Water Assessment Tool (SWAT) is widely used in climate-related studies [13]. It was first applied in complex watersheds to forecast how land management strategies will affect water resources and agricultural chemical yields [23]. Generally, SWAT is a sophisticated hydrological model that requires various inputs, such as topographic information, LULC data, meteorological data (e.g., precipitation and temperature), and soil types [24]. While other hydrologic models, such as MIKE SHE [25], TOPMODEL [26], and PCSWIM [27], could also be used to perform similar hydrologic simulations, SWAT was chosen in this study due to its ability to handle large watersheds, its integration of agricultural land use impacts, and its proven reliability in climate-related studies globally [28,29,30,31].

In general, this study quantifies projected changes in future extreme events (2016–2099) using two RCPs (4.5 and 8.5) from the Vietnam Assessment Report on Climate Change [4]. Specifically, the study aims to (i) analyze projected changes in extreme events and (ii) quantify changes in the monthly frequency of high and low flow. While investigating the projected impacts of climate extremes toward 2099 is the main objective of this work, we would also propose general solutions to enhance regional preparedness and resilience against climate change impacts. This study fills a critical gap by providing the first comprehensive assessment of future hydrological extremes in the Kon-Ha Thanh basin under different climate change scenarios, offering valuable insights for local adaptation strategies. These findings will support local stakeholders, including agricultural agencies and disaster management authorities, and serve as a valuable scientific basis for mitigating future natural disasters.

2. Materials and Methods

2.1. Study Area

The Kon-Ha Thanh River basin is among the largest in South Central Vietnam, covering most of Binh Dinh Province and a small area of Gia Lai Province (Figure 1). The two main rivers of this basin, the Kon and Ha Thanh, originate from high mountains, accumulating numerous tributaries before flowing into the East Sea. The Ha Thanh River covers a basin area of 549 km² with a length of 38 km, while the Kon River has a basin area of 2582 km² and extends approximately 178 km [32]. The considerable size of the Kon-Ha Thanh River basin, covering 3131 km², highlights its significance in regional hydrological processes and the potential scale of climate change impacts. The topography is complex (narrow in the hilly area and flat in the coastal region), with altitudes ranging from 0 to 1400 m above mean sea level.

The monthly mean temperatures range from 18 °C to 20 °C, in which the average annual rainfall varies between 1800 mm and 3300 mm, with at least 65% and a maximum of 80% of this rainfall occurring from September to December. In general, the dry season is defined between January and August and the wet season is between September and December. In addition, the regional rainfall distribution is uneven. Specifically, the highland and northern mountainous regions receive the most rainfall, with annual averages ranging from 2220 to 3030 mm [32]. The next highest precipitation occurs in the Vinh Kim mountains in the midstream of the Kon River, Van Canh district, upstream of the Ha Thanh River, and the northern coastal districts, with rainfall ranging from 2000 to 2180 mm. In contrast, the coastal area, Tay Son district, and the downstream region of the Kon River receive an average annual rainfall of 1610 to 1880 mm. The region also experiences two to four typhoons annually, leading to significant inundation issues [32].

2.2. Hydrological Soil and Water Assessment Tool (SWAT) Model

This model was originally developed by USDA-ARS in the early 1990s to predict the environmental impact of land management practices in watersheds [23]. SWAT is commonly used for watershed-scale studies and its ability to integrate different hydrological components to evaluate changes in streamflow, sediment, and water quality. It is widely applied to topics such as agricultural management practices [33], climate change [34,35,36], hydrological processes in rivers [37,38,39], nutrient cycling and transport [14], pesticide monitoring [40], LULC changes [41], and the performance of satellite products [42].

2.3. Model Data, Setup, and Workflow

In this study, ArcGIS (version 2.1.5) was used to prepare the SWAT inputs. The Shuttle Radar Topography Mission (SRTM) 1 Arc-Second (30-m resolution) data were used, while the LULC and soil type maps in 2017 were obtained from the Binh Dinh Province Department of Planning (http://skhdt.binhdinh.gov.vn/, accessed on 1 January 2017). The LULC map indicated that mixed forest was the major land cover type that covers 42.47% of the entire study area, followed by agricultural land at 24.38%, land for annual crops at 16.87%, developed areas at 15.82%, and water bodies at 0.46%. In terms of soil types, the region consisted of Clay (6.08%), Silt Loam (10.41%), Light Clay (18.77%), Sand (23.12%), and Loam Sand (41.62%) (Figure 1).

The daily precipitation and temperature data were collected over a 23-year period (1986–2008). Moreover, the precipitation data were extracted from 11 stations, with 9 stations located in mountainous areas and 2 stations in lowland areas (Figure 1), at an average density of 285 km² per station.

Figure 2 presents the proposed framework used in this study, which consists of four main stages: (1) input data preparation, (2) model setup, run, calibration, and validation, (3) scenario testing (i.e., historical and future), and (4) hydrological assessments and recommendations. First, we prepare the main datasets (i.e., DEM, LULC, and soil types) for watershed delineation using the Terrain Analysis Using Digital Elevation Models (TauDEM) V5.0 [43], slope classification, and HRU generation. We used observed streamflow records from 1986 to 2008 to calibrate and validate the model, though data length was limited by availability constraints from the Binh Dinh Province Department of Planning.

In this study, the results and analysis section were shown for the historical (1986–2005) and future (2016–2099) scenarios under RCPs 4.5 and 8.5, utilizing projected daily precipitation and maximum and minimum temperatures (Figure 2). Finally, we conducted hydrological assessments focused on future projected changes in seasonal streamflow (i.e., dry and wet seasons), flood peaks, and low flows, with changes evaluated at the sub-catchment level and the respective discussions would be given to reveal the difference between projected climate and the historical regional conditions.

2.4. Model Calibration and Validation

In this work, the calibration and validation periods were chosen according to the availability of continuous streamflow data. Within the basin, the streamflow was available continuously at Binh Tuong station during the period from 1990 to 2008 (Figure 1a). Therefore, we performed a 10-year calibration from 1990 to 1999, followed by a 9-year validation from 2000 to 2008, with a 5-year warm-up period (1986–1989). Additionally, commonly used performance metrics were chosen in this study to evaluate the model’s performance, including Root Mean Squared Error (RMSE), Coefficient of Determination (R²), and Nash-Sutcliffe Efficiency (NSE) [44], as shown in Equations (1), (2) and (3), respectively.

$$\mathit{RMSE}=\sqrt{{\displaystyle \frac{{{\displaystyle \sum }}_{i=1}^n{\left(X_{\mathit{obs},i}-X_{\mathit{model},i}\right)}^2}{n}}},$$

(1)

$$R^2={\left({\displaystyle \frac{{{\displaystyle n\sum }}_{i=1}^n\left(X_{\mathit{obs},i}·X_{\mathit{model},i}\right)-\left({{\displaystyle \sum }}_{i=1}^n{X}_{\mathit{obs},i}\right)\left({{\displaystyle \sum }}_{i=1}^n{X}_{\mathit{model},i}\right)}{\sqrt{\left({{\displaystyle n\sum }}_{i=1}^n{\left(X_{\mathit{obs},i}\right)}^2-{\left({{\displaystyle \sum }}_{i=1}^n{X}_{\mathit{obs},i}\right)}^2\right)·\left({{\displaystyle n\sum }}_{i=1}^n{\left(X_{\mathit{model},i}\right)}^2-{\left({{\displaystyle \sum }}_{i=1}^n{X}_{\mathit{model},i}\right)}^2\right)}}}\right)}^2,$$

(2)

$$NSE=1-{\displaystyle \frac{{{\displaystyle \sum }}_{i=1}^n{\left(X_{\mathit{obs},i}-X_{\mathit{model},i}\right)}^2}{{{\displaystyle \sum }}_{i=1}^n{\left(X_{\mathit{obs},i}-{\overline{X}}_{\mathit{obs}}\right)}^2}},$$

(3)

where the X_obs is observed value and X_model is a modeled value at time/place i.

The SWAT Calibration and Uncertainty Procedures (SWAT-CUP V2019) software was used for model calibration and validation. SWAT-CUP provides a robust framework for automating parameter optimization and assessing model uncertainty [45]. Specifically, the Sequential Uncertainty Fitting (SUFI-2) algorithm was employed for model calibration. Specifically, SUFI-2 was chosen due to its efficiency in minimizing uncertainties while optimizing the match between simulated and observed streamflow that was proved as efficient in previous works [46,47]. In addition, SWAT-CUP also supports sensitivity analysis, which identifies the most influential parameters affecting model performance, thereby ensuring a more accurate calibration.

2.5. Climate Change Scenarios

Two climate change scenarios from the Vietnam Government [4] were used. These products are built from CMIP5 (Couple Model Intercomparison Project Phase). Specifically, these climate projections presented here are derived from the World Bank’s Climate Change Knowledge Portal (CCKP) datasets [48] that support the IPCC Assessment Report 5 (AR5) [49]. The CMIP iteration of models uses these datasets, which are processed outputs from simulations conducted by 16 General Circulation Models (GCMs) created by climate research institutions across the world and assessed by the IPCC for quality assurance (e.g., CanESM2, EC-Earth, FGOALS-g2, MIROC-ESM, etc.) [50]. Vietnam uses five global and regional climate models (AGCM/MRI, PRECIS, CCAM, RegCM, and clWRF) to create climate change scenarios. Specifically, these projections are based on an ensemble of these GCMs for indicators such as average temperature anomalies and annual precipitation anomalies for Vietnam under RCPs 4.5 and 8.5 [50]. Before extracting those data for specific regions, they were downscaled and bias was corrected at a resolution of 10 km. The daily rainfall is adjusted from the model using observation data and the Quantile Mapping method [51]. Temperature bias correction has been implemented as recommended by Amengual et al. [52]. The government recommends utilizing both RCP 4.5 and RCP 8.5 greenhouse gas emissions to assess water resources for river basins within Vietnam.

Additionally, regional downscaling methods using Statistical Downscaling (SD) and bias correction are employed for these projections [50]. This approach involves deriving empirical relationships that link large-scale atmospheric variables (predictors) with local or regional climate variables (predictands). The projected changes in temperature and precipitation, as summarized in

Table 1. Temperature and precipitation change at the Kon-Ha Thanh River basin according to the scenario of the Vietnam Government released in 2016.

		Temperature (°C)				Precipitation (%)
		Winter	Spring	Summer	Autum	Winter	Spring	Summer	Autum
		December– February	March– May	June– August	September– November	December– February	March– May	June– August	September– November
RCP4.5	2016–2035	0.7	0.7	0.7	0.7	5.3	10.4	1.5	19.0
	2046–2065	1.3	1.3	1.6	1.4	12.6	−2.9	−4.3	27.9
	2080–2099	1.5	1.8	2.1	1.8	54.5	22.5	4.3	22.0
RCP8.5	2016–2035	0.8	0.8	0.8	0.8	1.2	2.9	26.1	18.2
	2046–2065	1.7	1.8	2.0	1.8	11.8	−8.9	5.2	24.5
	2080–2099	2.8	3.1	3.5	3.2	23.9	17.7	3.2	16.9

, are used for future simulations in the SWAT model (see Section 3.2, Section 3.3, Section 3.4, Section 3.5 and Section 3.6) across three periods: 2016–2035, 2046–2065, and 2080–2099.

2.6. Climate Change Assessments

The Kon-Ha Thanh River basin covers diverse regions with different characteristics in topography, which partly affects the rainfall distribution over the region, leading to varying levels of climate impacts on the natural regime (Figure 1). To fully observe this variation, streamflow at seven locations across the catchment is considered (

Table 2. Characteristics of sub-catchments over the Kon-Ha Thanh River basin.

No	Location	River	Sub-Catchment Area (km2)
1	Vinh Kim	Kon	1012.6
2	Binh Tuong	Kon	1615
3	Tay Son	Kon	302.6
4	Thuan Ninh	Kon	115.6
5	Nui Mot	Kon	188.7
6	Kon outflow	Kon	2582
7	Ha Thanh outflow	Ha Thanh	549.3

, Figure 1a). The climate change impact is analyzed through the variation in peak flow, base flow, and potential hydrological shifts at each sub-catchment.

3. Results

3.1. SWAT Calibration and Validation

The calibration period focused on three parameter groups that physically represent (i) overland flow, (ii) groundwater flow, and (iii) streamflow. The calibration strategy first focused on groundwater flow by adjusting two parameters, including baseflow alpha factor (ALPHA_BF) and groundwater delay (GW_DELAY). Then, overland flow and streamflow parameters were adjusted, such as the SCS runoff curve number for moisture condition II (CN2), the surface runoff lag coefficient (SURLAG), the effective hydraulic conductivity in the main channel alluvium (CH_K2), the effective hydraulic conductivity in the tributary channel alluvium (CH_K1), and overland flow Manning’s N value (OV_N). These parameters were chosen based on findings from Neitsch et al. [53] and Arnold et al. [45] (

Table 3. Selected SWAT parameters used in model calibration and validation.

No	Parameter Description	Parameter	Default Range	Fitted Value
1	SCS runoff curve number for moisture condition II	CN2	−0.2–0.2	0.15
2	Moist soil albedo	Sol_Alb	0–0.25	0.13
3	Average slope length (m)	SLSUBBSN	10–150	100
4	Saturated hydraulic conductivity (mm/hour)	Sol_K	−0.3–0.3	0.12
5	Depth from soil surface to bottom of layer (mm)	Sol_Z	−0.5–0.5	0.20
6	Effective K (hydraulic conductivity) in the tributary channel alluvium (mm/hour)	CH_K1	0–150	100
7	Effective hydraulic conductivity in the main channel alluvium (mm/hour)	CH_K2	0–500	185
8	Available water capacity of the soil layer (mm/mm soil)	SOL_AWC	−0.25–0.25	0.15
9	Maximum canopy storage (mm)	CANMX	0–100	8
10	Manning’s n value for main channel	CH_N1	0.01–30	15
11	Manning’s n value for main channel	CH_N2	0–0.3	0.12
12	Manning’s N value for overland flow	OV_N	0.01–30	0.5
13	Baseflow alpha factor (l/day)	ALPHA_BF	0–1	0.9
14	Groundwater delay (day)	GW_DELAY	0–500	30
15	Groundwater ‘revap’ coefficient	GW_REVAP	0.02–0.2	0.05
16	Threshold depth of water in the shallow aquifer for revap to occur (mm)	REVAPMN	0–500	150
17	Threshold depth of water in the shallow aquifer required for return flow to occur (mm)	GWQMN	0–5000	100
18	Surface runoff lag coefficient (day)	SURLAG	0.05–24	20

).

The comparison of the observed and simulated streamflow during the validation and calibration period is shown in Figure 3. In general, our simulation was categorized as having good agreement compared with the observed data. Moreover, the discharge peaks can be captured by the model (although some peaks are smaller than observed, dots below the 1:1 line, Figure 3b,c), which is useful to simulate future flood peaks.

The SWAT model achieved R² and NSE scores of 0.77, 0.76, 0.79, and 0.77, respectively, for calibration and validation (Figure 3b,c), which were categorized as good based on the model’s performance using efficiency criteria (see Section 2.4). Additionally, the RMSE values for both periods are relatively low, at 93.38 m³/s and 50.96 m³/s, respectively, indicating a good performance that would then be used use in performing future scenarios in this study.

3.2. Projected Changes in Monthly Streamflow

Figure 4 presents a comparison of the average monthly streamflow between historical data and future projections under RCPs 4.5 and 8.5 for the Kon-Ha Thanh River. Specifically, the streamflow increase during the rainy season is noticeable, with this increase found starting from September until December. This change might be explained by the result of a precipitation increase higher than evapotranspiration during the rainy season. During the 2046–2065 period, the highest streamflow occurs in the rainy season (November), reaching 280.6 m³/s under RCP 4.5 and 272.6 m³/s under RCP 8.5. In contrast to other periods, the dry season (May) has the least amount of streamflow. The streamflow for RCPs 4.5 and 8.5 scenarios are 27.6 m³/s, and 24.2 m³/s, respectively. Overall, the monthly streamflow across all three periods under RCP 4.5 tends to be higher compared to it under RCP 8.5.

On the other hand, we also found a clear trend of increasing streamflow across all sub-catchments as climate change progresses toward 2099. The projected increases are more pronounced under RCP 8.5, particularly between 2080 and 2099, indicating a stronger hydrological response to higher emissions that correlate with the higher precipitation projected (see Table 1). Moreover, while the seasonal pattern of streamflow remained similar over future periods (i.e., 2016–2035, 2046–2065, and 2080–2099), with peak streamflow occurring during the end of the year, the magnitude of these peaks is projected to significantly rise under the RCP 8.5 scenario. For example, the Kon sub-catchment (at the outlet of the Kon-Ha Thanh River basin) exhibits the highest projected streamflow peaks, surpassing 500 m³/s, followed by the Binh Tuong sub-catchment with around 250 m³/s, while sub-catchments like Tay Son and Thuan Ninh show more moderate increases. Moreover, under RCP 8.5, the late-century projections show increases (39.5–84.6%) in streamflow over sub-catchments (Figure 4), highlighting the potential for more severe water management challenges and increased flood risks in the region if higher emissions continue.

Additionally, our findings suggest that streamflow variation differs across the Kon-Ha Thanh River catchment due to variations in the rate of precipitation change (Table 1). Specifically, the greatest variation occurs in sub-catchments originating from mountainous areas in the southern part of the basin, such as the Tay Son and Ha Thanh sub-catchments (Figure 1). The change tendency in this area is relatively high, particularly in the Ha Thanh area. The monthly variation rate in the rainy season at this midland sub-catchment reaches double compared to those located in downstream regions (Figure 1 and Figure 4). With the streamflow increase from 36.8% to 47.8%, the flood disaster in the catchment is expected to become more severe toward 2099, and this tendency was found to be similar to those indicated for Vietnam’s coastal central region [18,54].

3.3. Projected Seasonal Streamflow Changes

Figure 5 shows the projected monthly and seasonal changes in streamflow for the Kon River and Ha Thanh River under different RCP scenarios (i.e., 4.5 and 8.5) across two future periods, 2046–2065 and 2080–2099. In general, the results indicate significant variability in monthly and seasonal streamflow.

First, in Kon River, the wet season demonstrates pronounced increases in streamflow under both RCP 4.5 and 8.5 scenarios. Between 2046 and 2065, RCP 8.5 projects the highest increases during the wet season (September to December), with peak streamflow reaching 405.8 m³/s (RCP 8.5) and 372.7 m³/s (RCP 4.5), reflecting notable rises of 54.3% and 41.7%, respectively, compared to the baseline (1986–2005) (Figure 5a,b,e,f). During the dry season months (January to August), streamflow changes are less pronounced, though a significant increase is still projected under the higher emissions scenario. However, RCP 8.5 shows an expected increase of 39.6%, compared to 10.8% under RCP 4.5 scenario (Figure 5e). In the later period (2080–2099), the trends intensify, with monthly streamflow continuing to rise, particularly in the wet season, with the highest increases found at 54.7% (RCP 8.5) and 47.5% (RCP 4.5). Interestingly, RCP 4.5 shows a slightly higher increase ratio (+5.8%) compared to RCP 8.5 (+0.4%). Moreover, peak flows during the wet season are found in November, with streamflow values reaching up to 554.6 m³/s (RCP 4.5) and 538.9 m³/s (RCP 8.5). The lowest flows are found in March and April for both RCPs (Figure 5a,b).

For the Ha Thanh River, similar trends are observed, but at a smaller magnitude. Monthly average streamflow peaks are found to rise substantially during the wet season, especially in October, November, and December (Figure 5c,d), under the influence of different GHG scenarios. Specifically, under RCP 8.5, streamflow is projected to rise to 113.3 m³/s (+74.7%) by 2046–2065 and 114.4 m³/s (+76.4%) by 2080–2099, compared to the historical period (1986–2005). The highest flow peaks occur in November, with streamflow reaching up to 154.2 m³/s and 149.8 m³/s (2046–2065) under the RCPs 4.5 and 8.5, respectively, following similar trends observed in the Kon River (Figure 5a,b). Moreover, the dry season shows more moderate increases, which is similar to the Kon River, with rises of up to 36.3% and 77.2% under the RCPs 4.5 and 8.5, respectively, during the 2046–2065 period, and 82.9% and 109.4% under the RCPs 4.5 and 8.5, respectively, between 2080 and 2099 (Figure 5g,h).

In general, these results are consistent with projected changes in rainfall patterns under the different RCP scenarios (see Section 2.6; Table 1). The large increases in wet season streamflow, particularly in the later months of the year, align with projected increases in wet season precipitation (Table 1) with higher GHG emission (i.e., RCP 8.5) showing a higher increase in streamflow compared to RCP 4.5. In addition, the more modest increases in dry season flow suggest that while rainfall increases are likely, they are not profound during the dry months. These findings are similar to those indicated by Nguyen et al. [55] in their study of the Srepok River Basin, Central Highland Vietnam. Through this analysis, we also found a shift in hydrological conditions across sub-catchments. Specifically, while the peaks in average streamflow were found in November during the 2046–2065 period, these peaks shifted a month forward to December in the 2080–2090 period (Figure 4 and Figure 5). Overall, these projected changes highlight the vulnerability of the Kon-Ha Thanh River basin to climate change, with both rivers showing high sensitivity to higher RCP scenarios in terms of seasonal climate variations.

3.4. Projected Changes in the Frequency of High-Flow

Figure 6 shows the changes in flood peak frequency under RCPs 4.5 and 8.5 scenarios across various sub-catchments that show a clear upward trend in streamflow and peak flows across all projected periods. Specifically, to accurately assess the frequency of flood flows, we conducted the analysis based on a 20-year historical streamflow (Figure 6).

Overall, under the RCP 4.5 scenario, streamflow increases gradually and consistently across the three periods (2016–2035, 2046–2065, and 2080–2099), suggesting moderate growth in flood risk over time. However, in each sub-catchment, peak streamflow values increase at a slower rate under RCP 4.5 compared to the RCP 8.5 scenario. The RCP 8.5 scenario shows a sharper increase in streamflow, particularly during the 2080–2099 period compared to the historical period (Figure 6). Moreover, this increase is found to become higher in higher GHG emission scenarios and become more profound when progressing toward the 2080–2099 period (Figure 6). It means that under this higher GHG scenario, the sub-catchments are found to experience significantly more intense flood peaks. This trend is especially noticeable in Vinh Kim and Binh Tuong sub-catchments, where peak flows in the 2080–2099 period surpass those of earlier periods and are significantly higher than those projected under the RCP 4.5 scenario (Figure 6a,b). On the other hand, the Kon River outlet (Figure 6k,l) shows a peak flow of over 10,000 m³/s under RCP 8.5 and over 7500 m³/s (RCP 4.5) during the 2080–2099 period while they are around 9000 m³/s under RCP 8.5 and 7400 m³/s under RCP 4.5 during 2046–2065, highlighting an increasing in flood magnitude and frequency that indicates a higher chance of experiencing more flood disasters in the future.

In general, these results are consistent with previous studies that highlight the increased vulnerability of regions to flood disasters under higher-emission pathways, as a result of more intense and frequent precipitation events. In the context of the Kon and Ha Thanh River basins, the pronounced rise in streamflow under RCP 8.5 during the 2080–2099 period underscores the urgency of implementing effective climate adaptation measures to mitigate the heightened flood risk.

3.5. Projected Changes in the Frequency of Low-Flow

Drought conditions in this region are complex with higher pressure caused by the rapid socio-economic development, urbanization, and population growth [5,55]. Thus, while the evaluation of peak-flow changes has been presented in Section 3.4, it is necessary to examine the changes in low-flow in this section. In general, we found a notable decline in low flow within the Kon-Ha Thanh River basin over examined future periods (i.e., 2016–2035, 2046–2065, and 2080–2099) (Figure 7).

First, under both RCP 4.5 and 8.5 scenarios, the low-flow conditions in different sub-catchments exhibit a declining trend (Figure 7). For example, in the RCP 4.5 scenario, sub-catchments like the Vinh Kim (Figure 7a) and Kon River outlet (Figure 7k) show significant decreases in low flow as it moves from the current period toward future scenarios, particularly during the 2080–2099 period. Moreover, this trend is more pronounced under the RCP 8.5 scenario (Figure 7b,l), where the impacts of increased temperatures possibly lead to even lower streamflow values during the dry season (January to August).

On the other hand, similar to those discussed above, the Ha Thanh River outlet (Figure 7m,n) also shows a continuous decline in low flow. These findings suggest that the basin’s water availability during drought periods will likely reduce in the future, posing further challenges for water resource management, especially under the RCP 8.5 scenario.

In general, the reduction in low flow reflects the expected changes in climate variables, particularly temperature increases (see Table 1), which exacerbate evapotranspiration losses. These findings align with recent works [19,22], indicating that regions with similar climatic and socio-economic pressures are likely to experience more severe drought conditions in the future [28].

3.6. Hydrological Shift Under Projected Climate Impacts

While the evaluation of low- and high-flow conditions was presented in previous sections, this section aims to highlight the hydrological shifts under the impacts of the future climate across the Kon-Ha Thanh River basin (Figure 8). In general, the flood season is projected to shift earlier, while the dry season may experience notable reductions in water availability.

First, we found that this basin is likely to experience significant increases in streamflow, particularly in December, with flows expected to rise by up to 150% under the RCP 8.5 scenario in various sub-catchments (2080–2099), such as Vinh Kim, Binh Tuong, Tay Son, Thuan Ninh, and Nui Mot (Figure 8b,d,f,h,j). This trend is also observed in the Kon and Ha Thanh River outlets, where increases of more than 200% are projected for future periods (Figure 8k,m), with Ha Thanh River likely to experience a greater increase compared to the Kon River. In contrast, the dry season may see reductions in water availability across most sub-catchments, except for the Kon and Ha Thanh Rivers, where streamflow is projected to decrease by more than 10% starting in May under both RCP 4.5 and RCP 8.5 (Figure 8). These reductions could lead to water scarcity during critical agricultural periods, affecting crop yields and water resource management.

In general, the seasonal shifts in water availability will have far-reaching consequences, not only for agriculture but also for other sectors that rely on a consistent water supply, such as domestic consumption and industrial activities. The results suggest that policymakers and local stakeholders will need to adapt their water management strategies to mitigate the risks associated with both increased flood intensity and reduced dry seasonal flows. Investments in adaptive agricultural practices, flood control infrastructure, and sustainable water use will be critical for managing these future changes effectively.

4. Discussion

The analysis of hydrological shifts in the basin highlights significant impacts caused by the future climate, particularly under high GHG emission scenarios. Our results show that under RCP 8.5, the wet season is likely to experience substantial increases in streamflow, especially in December, where flows could rise by as high as 150%, as observed in sub-catchments such as Vinh Kim and Binh Tuong (Figure 8). Similarly, the Ha Thanh River outlets could experience increases of over 200%, which is consistent with other studies in Vietnam, such as Khoi and Suetsugi [56], who indicated pronounced streamflow increases in the Be River basin due to rising rainfall intensities during the wet season under future climate impacts.

The dry season presents the opposite trend, with decreases in streamflow starting as early as May under both RCPs 4.5 and 8.5. The projected decline, particularly in May, where water availability could drop by over 10%, poses a threat to water security for agricultural and domestic purposes. This finding aligns with the recent work by Huong et al. [57], in which they found similar reductions of up to 16.9% in streamflow in the Nam Rom River basin, and Khoi et al. [58], which reported a decrease of up to 19.6% in Ho Chi Minh City. In general, these reductions pose challenges for crop irrigation and industrial water supply, especially in downstream areas that heavily rely on water availability during the dry months.

Furthermore, we found a hydrological shift in both dry and flood seasons, with the earlier arrival of the regional flood season by the end of the 21st century. This shift could have far-reaching consequences, disrupting agricultural cycles, water resource planning, and flood mitigation strategies. Similar shifts in seasonality were highlighted by Vu et al. [59], who demonstrated comparable changes in drought patterns across the Central Highlands. These results underscore the growing urgency for adaptive strategies that address these projected shifts, including investments in early warning systems, improved flood protection infrastructures, and changes in agricultural practices to accommodate the changing hydrological regime.

In general, the basin is expected to face significant hydrological disruptions due to climate change, with severe implications for water management, agriculture, and flood control. The findings of this study offer a scientific foundation to support local policymakers in developing preparedness strategies to ensure water sustainability and enhance the region’s resilience against future climatic shifts.

5. Conclusions and Future Work

This study utilized the SWAT model to project future changes in streamflow—specifically high and low flows—in the Kon-Ha Thanh River basin under the impacts of different climate change scenarios, i.e., RCPs 4.5 and 8.5. The calibration and validation process was conducted using 19 years of observed data from the Binh Tuong hydrological station. Yet, there are uncertainties in the model (e.g., input data, climate change scenarios, calibration/validation period, limited data availability for model evaluation). However, the model’s results have supported a comprehensive investigation of the impact of climate change on the hydrological regime of the Kon-Ha Thanh River basin. Our findings are summarized as follows:

(1)

The Kon-Ha Thanh River basin is projected to experience substantial increases in wet-season streamflow, particularly in December, with potential rises of up to 150% under RCP 8.5. Meanwhile, the dry season could see reductions in water availability starting as early as May, with streamflow decreasing by over 10%. Moreover, while the average streamflow peaks are found in November during the 2046–2065 period, these peaks shift a month forward to December in the 2080–2090 period. In addition, the higher GHG emission scenario (i.e., RCP 8.5) shows higher impacts on seasonal and monthly streamflow compared to RCP 4.5;

(2)

Flood and dry seasons are projected to arrive earlier and last longer by the end of the 21st century; this shift could alter flood dynamics, potentially intensifying flooding events and driving broader hydrological changes across the region, significantly affecting agriculture, infrastructure, and local communities. Additionally, the projected decline in water availability during the dry season may elevate the risk of water shortages, potentially affecting crop irrigation and domestic water use;

(3)

We found that the earlier onset of the rainy season and extended dry periods may disrupt traditional agricultural cycles, potentially forcing changes in crop planning, harvest schedules, and water usage patterns. These hydrological shifts could also strain local infrastructure and water management systems, highlighting the need for improvements in flood control, water storage, and distribution systems to meet the growing demands of the population and economy.

In general, this study demonstrates an attempt to incorporate flood protection measures with the effects of climate change, which will be helpful for sustainable planning and strategy. Specifically, the findings of this study will support authorities and stakeholders in planning sustainable solutions for the region to mitigate climate impacts. It provides a valuable scientific basis for combatting disasters, thereby assisting stakeholders, regional authorities, and officials in promoting human resilience. To be more relevant in the present context, the authors will evaluate the new climate change data in future work using CMIP6 GCMs under SSPs.