Climate change and watershed hydrology: assessing variability in water balance components and groundwater flow patterns

In order to steer long-term planning and management of water resources, it is crucial to develop and assess various forecasts of future water availability (Najafi et al. 2012; Taylor et al. 2013; Nam et al. 2015; Woznicki et al. 2015). One of the key challenges surrounding the impacts of a changing climate is the effect that it will have on the sustainability of water supplies. Climate change and variability impact the hydrology of a river system, ultimately affecting ecological and human health by altering the structure and function of the aquatic environment (Sadio & Faye 2023). There is a strong consensus that climate change has resulted in noticeable changes in water resource availability, including groundwater and surface water (Malekinezhad & Banadkooki 2018).

Nowadays, several studies have been conducted on the impact of climate change on hydrological processes at various spatial and temporal scales (Scibek & Allen 2006; Woldeamlak et al. 2007; Kollet & Maxwell 2008; Candela et al. 2009; Goderniaux et al. 2011; Green et al. 2011; Taylor et al. 2013; Wanders & Van Lanen 2015; Meixner et al. 2016; Ayada et al. 2024). Climate change variables, particularly variations in rainfall and temperature, have a considerable impact on various hydrological components as well as the long-term sustainability of water resources (Arab Amiri & Gocić 2021; Soundala & Saraphirom 2022). Changing climate variables influence the hydrological processes by changing the amount of evapotranspiration, surface runoff, replenishment of groundwater, soil moisture content, surface water, and groundwater levels, the quality of water, and snow cover (Jalota et al. 2018; Kovács & Jakab 2021).

The world's fastest-growing population areas have significant issues related to water resources and sustainability, mostly in arid and semi-arid regions where the resource is essential for household, industrial, and agricultural purposes. Climate change, along with rapid population growth, may have an impact on sustainable development and exacerbate the water security issue (Conway & Schipper 2011; Otieno & Anyah 2012; Hirpa et al. 2019). The effects of climate change are widespread in Africa due to the continent's predominately arid and semi-arid regions and its limited ability to modify its agriculture-based economic structure (Nicholson et al. 2018; Hirpa et al. 2019). Rainfed agriculture predominates in the region, which is vulnerable to climatic fluctuation (Shiferaw et al. 2014; Thomas et al. 2019).

A number of studies have been carried out recently to assess how the hydrological process is affected by climate change (e.g., Eckhardt & Ulbrich 2003; Nyenje & Batelaan 2009; Touhami et al. 2015; Hughes et al. 2021). Hydrological modeling with physically based tools is often used in assessing past and future water resources as a consequence of climate change (Chu et al. 2010; Moradkhani et al. 2010; Warburton et al. 2010; Gosling et al. 2011; Tong et al. 2012; Pratoomchai et al. 2014). Nevertheless, a majority of these studies were performed on single hydrological components (e.g., stream flow, soil moisture, recharge, and groundwater flux). However, relatively little study has been conducted on the potential impact of climate change on the overall hydrological processes of a watershed. Therefore, this study attempts to evaluate the impact of climate change and variability on various hydrological processes in the area through integrated modeling approaches. In our work, WetSpass (Batelaan & De Smedt 2007) and a modular three-dimensional finite-difference ground-water flow model (MODFLOW) (Harbaugh 2005) were used to study the impact of climate change on hydrological processes. Climate change analysis was performed using the Mann–Kendall (MK) statistical test (Mann 1945; Farlie & Kendall 1971), the most commonly used method for assessing the significance of trends in hydrological time series, to identify statistically significant decreasing or increasing trends. We used Sen's slope estimator to calculate the trend's magnitude.

Walga–Darge watersheds are the main tributaries of the Omo-Gibe River basin. Coupled with burgeoning demand due to urban expansion and rapid population growth, climate change could pose a threat to the sustainable use of water resources. Therefore, this paper aims to present the potential impact of climate change on surface water balance and groundwater flow systems over the Walga–Darge watersheds, upstream of the Omo-Gibe basin. The specific objectives of the present study are: (1) to assess the trend and magnitude of change in precipitation and temperature in the study area watersheds; and (2) to evaluate the impact of precipitation and temperature variation on surface runoff, evapotranspiration, recharge, and groundwater flow systems of the study watershed. Therefore, the study's findings can contribute to sustainable water resource development and support future research projects of a similar nature in the watershed and surrounding area.

Preeminently, the geology of the study area is related to the results of volcanic activities of trap serious volcanics of the Ethiopian plateau, tectonic movement, and the development of the Ethiopian rift system. The dominant lithologic units in the study area are Paleogene fissural flood basalt with minor trachyte, rhyolite, and pyroclastic flow; Neogene pyroclastic, rhyolite, and trachytic flow; and Pleistocene-Holocene basic to acidic volcanic and phreatomagmatic deposits, which include the recent lacustrine and alluvial sediments (GSE 2010, 2014). There are deposits of alluvial sand and silt along the lower course or braided river channels, as well as some lacustrine loamy soil at higher elevations in the intermountain plains. The simplified geological map of the study area is given in Figure 2.

The groundwater occurrences in the study area are predominantly characterized by volcanic aquifers. The groundwater flows dominantly through weathered surfaces, fractures, and fissures. The aquifer constitutes fractured and weathered basalt, ignimbrite, rhyolite and tuff, and pyroclastic. Regional groundwater flow direction in the study area enlarges following surface water flow directions. Accordingly, most of the rivers in the study area are perennial, which shows the presence of significant base flow.

The WetSpass model requires various climatic and other physical parameter data to generate monthly, seasonal, and annual estimates of water movement across earth systems. The input data sets required for the models are listed in Table 1.

Trend analysis aims to determine if the climate data show an increasing, decreasing, or no trend at all. The climatic data set used in the present study includes precipitation and annual average temperature from the period of 1985–2020. The nonparametric MK statistical test (Mann 1945; Farlie & Kendall 1971) the most commonly used method for assessing the significance of trends in hydrological time series, was applied to identify statistically significant decreasing or increasing trends. Sen's slope estimator was used to calculate the trend's magnitude. The MK trend test is based on two hypotheses: the null hypothesis (H₀), and the alternative hypothesis (H₁). H₀ expresses the absence of a trend, whereas H₁ indicates a significant rising or declining trend over time in hydro-climatic data (Agbo et al. 2021; Gadedjisso-Tossou et al. 2021).

The surface water system in the study area – interception, surface runoff, actual evapotranspiration, and recharge – was simulated with the spatially distributed hydrologic model WetSpass (Batelaan & De Smedt 2007). The model takes into account the spatial variability of basin parameters. The spatial input data required to run the model are soil, land use, slope (%), elevation (m), depth of groundwater (m), rainfall (mm/year), potential evapotranspiration (mm/year), temperature of the air (°C), and wind speed (m/s). Rainfall, air temperature, potential evapotranspiration, wind speed, and groundwater depth are interpolated to gridded data sets using an inverse distance weighted interpolation technique. The data inputs for the model are ASCII files. The raster maps were converted to ASCII files using ArcGIS's conversion tool to make them compatible with the WetSpass model. The model is connected to parameters of soil types and land use using attribute databases (Batelaan et al. 2003).

Calibration is the process of matching the calculated head value with the observed head value. In this study, the trial-and-error method was used to calibrate the model using the head values of 88 wells.

To evaluate the potential impact of climate change on hydrological processes, this study employed both historical and hypothetical climate data. Using 30 years of actual historical climate data as a baseline, three scenarios were generated for the next 60 years, assuming it increases or decreases linearly over time. For instance, a rainfall decrease of −3.334 mm per year is predicted to result in a decrease of 333.4 mm in 100 years. Based on the assumption, these perturbed data are then fed into a hydrological model, and the resulting changes in hydrological processes are assessed. This assumption has been used in previous studies including those of Jiang et al. (2007); Zhang et al. (2016); and Chuko & Abdissa (2023).

First, the baseline WetSpass model was constructed from current climate data (2010–2020). Second, the MODFLOW model was constructed and calibrated simultaneously with WetSpass. Third, historical and future water balances were simulated using the WetSpass model by varying the two main climatic inputs – precipitation and temperature – following the projected future climate, despite keeping the same values for the other inputs. Fourthly, WetSpass-generated historical and future recharge scenarios were fed into the fully calibrated MODFLOW. The groundwater system's sensitivity to climate change has been quantified as the difference in groundwater flow and level values between the baseline period and future points of time.

The nonparametric MK test and Sen's slope estimator techniques were adopted to ascertain whether there is an increasing or declining trend in long-term temporal data, with their statistical significance at a 95% level of confidence. The results of the statistical analysis of annual rainfall and temperature are presented in Table 2.

For the majority of the stations, the annual rainfall decreased, as the results show (Table 2). A statistically non-significant (p > 0.05) increasing trend in the amount of rainfall has only been observed at one meteorological station (Dilela). The annual rainfall at five stations (Wolkite, Dariyan, Woliso, Chitu, and Ameya) has experienced decreasing trends. However, the decreasing trend was only statistically significant at the Wolkite and Dariyan stations (p-value < 0.05). On the other hand, a non-significant decreasing trend was observed at the Woliso, Chitu, and Ameya stations, being that the computed p-value is higher than the significance level α = 0.05. These findings indicated that annual rainfall decreased by −2.514, −14.637, −13.722, −3.334, and −3.778 mm/year at Woliso, Wolkite, Dariyan, Chitu, and Ameya, respectively (Table 2). However, it increased by 5.2 mm/year at Dilela station. Concerning the annual temperature data (maximum temperature, minimum temperature, and mean temperature), the analyses revealed significantly increasing trends for all stations (Wolkite and Woliso). The highest increasing trend in temperature and the decreasing trend in precipitation have been detected at Wolkite station. Overall, these results indicate that the rainfall amount shows a decreasing trend and the temperature shows an increasing trend over the catchments.

The output of the WetSpass hydrological model for different climate scenarios is presented in Table 3. From the result in Table 3, it is apparent that the study revealed that the volume of surface water fluxes has substantially decreased historically and in the future due to climate change. The first set of three models (1980s, 2000s, and 2020s) is based on actual historical climate data, while the second three (2040s, 2060s, and 2080s) are based on hypothetical climate data. All water balance components – precipitation, interception, surface runoff, actual evapotranspiration, and recharge – show decreasing trends except actual evapotranspiration in the 2000s (Table 3).

The spatiotemporal variation of actual evapotranspiration rates is caused by changes in vegetation cover, air temperature, and an uneven distribution of rainfall. Actual evapotranspiration is expected to increase accompanying the temperature rise. The simulated actual evapotranspiration rates of the 1980s, 2000s, 2020s, 2040s, 2060s, and 2080s are 66.11, 66.65, 72.26, 73.39, 76.42, and 79.36%, respectively. The rate shows increasing trends. However, concerning the overall water flux, evapotranspiration shows a decreasing trend throughout time due to a decrease in precipitation, except in the 2000s. Despite the substantial increase in the annual mean temperature in the future, the annual evapotranspiration revealed slight reduction tendencies for all simulations due to a decline in rainfall.

As shown in Figure 5(a), the mean annual groundwater recharge values indicate that the highest groundwater recharge is more concentrated in soil textural classes of sandy loam, combined with forest and agricultural land use classes. This is essentially due to the high permeability of the soils, the high amount of precipitation, and the lower temperature. However, due to limited precipitation, increased temperature, and clay soil textural classes in the southern parts of the catchment, it received a relatively small amount of recharge. The variation in recharge between the simulated time intervals is shown in Table 4. Looking at Figure 5(b), it is apparent that the responses of annual groundwater recharge to temperature and rainfall variations appear to be non-uniform throughout the watershed.

The combined effect of decreasing rainfall and higher temperatures shows a noticeable negative influence on groundwater recharge in the study catchment. The first three simulations were based on actual historical data, and the others depended on hypothetically constructed future climate data. The result shows groundwater recharge declined historically and, in the future, which is consistent with the historical and predicted decrease in rainfall as seen in Table 4. Based on the study result, the simulated mean annual recharge over the study area for the 1980s, 2000s, 2020s, 2040s, 2060s, and 2080s is 263.83, 256.79, 161.4, 129.52, 101.38, and 64.67 mm/year, respectively. Consequently, a reduction in rainfall and increased temperature result in a relative change in recharge that can range from a reduction of 2.65, 38.82, 50.91, 61.57, and 75.49% in the 2000s, 2010s, 2040s, 2060s, and 2080s, respectively, relative to the 1980s. Furthermore, the result indicated groundwater recharge is much more vulnerable to climate change than other components of the water balance summarized in (Table 3).

The potential impact of climate change on groundwater flow systems was evaluated under six different groundwater recharge scenarios. Because a decrease in recharge could lead to a decrease in the amount of renewable groundwater, it remains crucial for controlling the volume of groundwater. The steady-state simulation results indicated that groundwater discharge to the stream gradually decreased due to variations in aquifer recharge over time. Table 4 illustrates the overall impact of climate change on the groundwater flow budget, which was calculated by the MODFLOW model.

The reduction in groundwater recharge magnitude shows a significant decline in the water table and groundwater discharge to the stream. According to the simulation results, the water level is declining by an average of +0.00067, −14.79, −18.76, −18.69, and −18.69 by 2000, 2020, 2040, 2060, and 2080, respectively (Figure 7). Consequently, annual river seepage to the aquifer for the simulation period was increased by 2.422, 28.50, 33.333, 35.75, and 36.71%, respectively. However, annual groundwater discharge to the river shows a decreasing trend of +3.09, −26.46, −30.76, −32.82, and −34.02%, respectively.

Assessing the trend of climate change in the Walga–Darge watershed was one of the study's objectives. Based on the analysis of climate data, annual rainfall has been greatly varied over the last 30 years. The results from this study indicated that there has been a statistically significant decreasing trend of annual rainfall at two stations, a statistically non-significant decreasing trend at three stations, and a statistically non-significant increasing trend at one station. Generally, the annual rainfall trend analysis indicates a decline in the amount of rainfall over the watersheds. Similarly, the study conducted by Cheung et al. (2008) revealed a statistically significant decreasing trend of summer rainfall (which is the rainy season of Ethiopia) in the Omo-Gibe watersheds at the 0.05 level. Conway & Schipper (2011) and Viste et al. (2013) reported that there are no clear trends that could be detected in central and northern Ethiopia. In contrast to earlier findings, however, the findings of the current study are able to demonstrate the decreasing trend in rainfall amount.

In previous time series studies, examining rainfall and temperature trends in Ethiopia has been conducted at various spatiotemporal scales. Several studies have revealed that precipitation varies in terms of its direction, and temperature almost similarly experiences significant increasing trends (Seleshi & Camberlin 2006; Berihun et al. 2023). The minimum, maximum, and mean temperature trends are consistent with previous studies. For instance, the study conducted by Conway & Schipper (2011) indicated that minimum and maximum temperatures show increasing trends from 1951 to 2002 of 0.4 and 0.2 °C/decade, respectively, during the periods 1898–2002. Wagena et al. (2016) concluded that future annual rainfall will increase as the maximum and minimum temperatures increase in the Beles basin. Similarly, the analysis of rainfall trends in the Lake Tana region indicated that the annual rainfall is increasing; however, the rate of change is not statistically significant (Weldegerima et al. 2018). Nevertheless, other studies reported that the mean, maximum, and minimum temperature had a generally increasing trend, whereas the amount of rainfall showed a generally decreasing trend in the Lake Tana Sub-basin (Addisu et al. 2015). Overall, there is no consensus on the direction of the rainfall trend, and there is a consistent increase in temperature.

This study set out to examine how climate change affects hydrological processes at the catchment scale. In the present study, there has been a considerable shift in the catchments' hydrological regime as a result of climate change. Based on the simulated annual water balance, mean annual evapotranspiration, runoff, interception, and recharge show decreasing trends, which are mostly caused by a decrease in relatively excess precipitation and an increase in temperature. Furthermore, decreasing the precipitation rate is likely to result in a decrease in all four hydrologic parameters – surface runoff, evapotranspiration, interception, and recharge – because there is less water available in the system. On the other hand, increasing temperatures affect hydrologic processes in both negative and positive ways. Temperature, followed by rainfall, plays a significant role in the loss of water over a catchment. The rise in temperature causes increases in evaporative demand (Huang et al. 2015; Berg et al. 2016), although a reduction in seasonal rainfall affects the soil moisture content that is available for evaporation (Hovenden et al. 2014; Caretta et al. 2022).

Climate change affects groundwater resources directly through changes in the natural replenishment of groundwater by recharge and indirectly through changes in groundwater withdrawals – demand for groundwater is likely to increase (Taylor et al. 2013; Lall et al. 2020). In the current study, climate change and variability were found to have a considerable effect on the replenishment of groundwater. Based on the results presented in Table 4, a reduction in groundwater recharge causes a decrease in groundwater discharge to the river, while annual river seepage to the aquifer increases significantly. There are two ways in which climate change affects streamflow: (1) by reducing rainfall amount, as less rainfall is expected to cause lower average streamflow; (2) by reducing baseflow contribution to streamflow as a result of variations in rainfall and temperature that negatively impact groundwater recharge, which in turn affects groundwater levels and stream leakage to the aquifer. Therefore, changes in groundwater systems may also impact surface water systems, as groundwater is a major contributor to streamflow in areas with relatively shallow water tables (Earman & Dettinger 2011). However, increasing rainfall amounts affect hydrologic processes positively. For example, in contrast to our findings, Nyenje & Batelaan (2009) reported that groundwater recharge and baseflow are predicted to increase in the future in the upper Ssezibwa catchment, Uganda, due to rising rainfall. Therefore, climate change and variations have a significant role in controlling the hydrological processes of a catchment.

Understanding how climate change affects water resources is essential to planning, managing, and developing effective adaptation strategies. The application of WetSpass and MODFLOW model output to a hydrologic model allows for comparisons between simulated past, recent, and potential futures and provides an important understanding of the hydrological dynamics in response to climate change. To understand the effects of climate variation on hydrological processes, six different scenarios were simulated using the WetSpass and MODFLOW models. The direct impact of climate change indicates that the increasing trends in temperature and decreasing rainfall amounts pose a significant threat to the sustainable use of water resources. Therefore, adaptation to climate change impacts on water resources should be seriously considered in the water management policy.

In general, besides uncertainties in climate and hydrologic models, the authors hope that this study provides a sound base for future studies on understanding and reassessing the impact of climate change on water resources and management in the future. However, it is essential to note that this study does not account for potential changes in land use in the future and effects due to future water demand, which could have an even greater impact than climatic trends. Therefore, future studies should examine how land use change and future water demand affect the sustainability of water resources in the catchments.

The authors are grateful to Wollega University for financing this study. We also express our gratitude to the following institutions for providing the data used to conduct this study: the Ethiopian Geological Survey, Ethiopian Construction Works Corporation, Oromia Construction Corporation, and the National Meteorological Agency.

A. G. conceptualized the whole article, developed the methodology, arranged the software, prepared the original draft, reviewed the manuscript, and edited the article. F. W. developed the methodology, arranged the software, reviewed the manuscript, and edited the article.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.