Impact of climate change on groundwater potential and recharge in the drought prone Runde catchment of Zimbabwe

Groundwater, the world's largest freshwater resource, is an important source of water for: meeting growing domestic water demands, irrigation, and coping in times of drought, whose occurrence seems to be increasing due to climate variability and change (Calow et al. 2010; MacDonald et al. 2012; Shivakoti et al. 2019). Globally, more than two billion people depend on groundwater for daily water supply (Wu et al. 2020). In Africa, groundwater is the largest freshwater resource, with the first continent-wide groundwater estimates showing that the volume of groundwater is about 0.66 million km³, which is 100 times more than the annual renewable freshwater resources (MacDonald et al. 2012). However, these groundwater resources are unevenly distributed and underutilized (SADC-GMI 2019). About 70% of people in the Southern African Development Community (SADC) are estimated to depend on groundwater resources. In Zimbabwe, groundwater is a source of water for above 70% of the population (UNDP 2021). In semi-arid and drought prone parts of Zimbabwe, as in other parts of Southern Africa, the dependence on groundwater is more acute. In these semi-arid and drought prone areas groundwater plays an important role in meeting household and productivity-related water needs and more critically sustains people during the long-dry spells and droughts (Jannis et al. 2021). Essentially, groundwater is an integral component of their coping or adaptation mechanisms. Climate change presents an exogenous challenge to the availability, use and dependence on groundwater in drought-prone areas to the detriment of their livelihoods, productivity, and coping strength. Rural communities are the most vulnerable to climate change-induced drought and variability due to their high reliance on groundwater and groundwater-fed systems for potable water use (Amraoui et al. 2019). Furthermore, as climate change-induced droughts persist in the SADC region, developing resilience is strongly linked to groundwater: access, development, and resource management (Amraoui et al. 2019; Gailey et al. 2018).

Runde catchment in Zimbabwe is a drought-prone catchment that encapsulates the dependance on groundwater in drought-prone areas. Groundwater is the main source of water for drinking and production in most rural communities. In Zimbabwe, aquifers known to have significant groundwater resources include the Lomagundi dolomites and forest sandstone aquifers, none of which are in the Runde catchment. Alluvial aquifers are also known to have good water resources in Zimbabwe (Owen 1989), but their occurrence in spatial extent is limited as they follow rivers (Lambert & Faulkner 1991) and the significant ones shown in Owen (1989) are not located in areas where most of the rural populace is. Runde catchment is underlain by the basement complex rock, which is known for low to medium groundwater yields. Generally, groundwater occurrence depends primarily on geology, geomorphology, weathering, and climate. In drought prone areas, such as Runde, groundwater is not only used to meet rural water demands but is also used to cope with extreme hydrometeorological events such as droughts (Arabameri et al. 2019). In severe droughts, massive groundwater programmes are implemented to save humans and livestock, e.g. the drought of 1992. Climate change presents a new challenge in terms of groundwater: water use, dependability, and management in the Runde Catchment as anywhere else in Africa because climate changes results in long-term changes in rainfall and temperature (Barua et al. 2020; Arabameri et al. 2019). Rainfall and temperature (evaporation) are hydrologic cycle variables which are key in sustaining groundwater resources (MacDonald et al. 2013). In Runde catchment, recharge is mainly through rainfall and hence long-term changes in rainfall will impact the groundwater availability, potential and sustainable utilization (Barua et al. 2020). Furthermore, in Africa, climate change is increasing the intensity and frequency of droughts (WMO 2021) and thus there is a need to develop better adaptation strategies. Unfortunately, there is limited catchment specific data and information to understand the impacts of climate change on groundwater resources to enable better planning and programming related to groundwater use, potential, adaptation, and disaster response. It is imperative that water resources managers, development actors and communities in the Runde catchment understand the potential impacts of climate change on groundwater potential and recharge (Barua et al. 2020).

Groundwater recharge refers to the process through which water is added to the groundwater storage below the land surface, which is marked by a change in the water table level. Quantification of groundwater recharge is important for sustainable groundwater utilization. Ideally, to avoid groundwater mining people should only utilize groundwater that can be replenished by annual recharge (Adhikari et al. 2018). However, the estimation of recharge is very challenging and a number of methods have been developed that include: chloride mass balance, isotopic tracers, integrated and groundwater water balance methods, groundwater level fluctuation and streamflow/baseflow methods (Xu & Beekman, 2003; MacDonald et al. 2021) and a number of these methods have been used in Southern Africa. In Zimbabwe, studies in small headwater catchment by MacDonald & Edmunds (2014) estimated recharge by the chloride mass balance method to be about 3.7% of long-term mean annual rainfall (22 mm/year) on melanocratic bedrock and about 1.1% (6.7 mm/year) on leucocratic felsic bedrock. Sibanda et al. (2009) evaluated the recharge of four methods, namely chloride mass balance, Darcian flownets, MODFLOW modelling and water table fluctuation over the Nyamandhlovhu aquifer in western Zimbabwe. They pointed out that reasonable estimates of recharge are in the range of 15–20 mm/year, which is about 2.7–3.6% of mean annual rainfall. Earlier, Larsen et al. (2002), after sampling and analysing data from 22 boreholes in Western Zimbabwe, estimated recharge to be about 20–25 mm/year (3.6–4.5%). Houston (1988) applied baseflow analysis, simulation and hydro-chemical analysis to estimate groundwater recharge in Masvingo province, an area in Runde catchment. Houston (1988) concluded that recharge estimates were about 2–5% of annual rainfall. It is against this background that 2% of mean annual rainfall is used as a recharge estimate for national groundwater resources planning in Zimbabwe. Whilst using 1D or 3D groundwater modelling is ideal for recharge assessment, the availability of data to develop the conceptual and numerical models is a challenge. Thus, the use of a percentage of a more commonly measured variable (i.e. rainfall) provides a method that can be applied over wider areas and can be integrated with satellite datasets. Hence, in some studies such as Misi et al. (2018) recharge is determined as percentage of mean annual rainfall. This approach can serve as a first-order estimate of groundwater recharge and is useful in estimating the impact of climate change on recharge.

Groundwater potential, from an exploration perspective, refers to the possibility of groundwater occurrence or availability (Rahmati et al. 2015; Misi et al. 2018). Such an assessment can be done using geological and geophysical methods, but for regional scales these methods tend to be costly and time consuming. On the other hand, Geographic Information Systems (GIS) and remote sensing can be effectively used for groundwater potential mapping at catchment or regional scales (Ahmadi et al. 2021). GIS-based groundwater potential mapping combines spatial thematic layers such as elevation using some multi-criteria decision-making framework (MCDA). One of the more commonly applied MCDA methods is the Analytical Hierarchy Process (AHP) (Saaty 1980). In AHP, the relative importance of thematic layers is included based on expert knowledge. Misi et al. (2018) used a GIS based approach to map groundwater potential in Upper Manyame Sub-catchment in Zimbabwe using the AHP method (Barua et al. 2020). This showed that the method worked considerably well. The Upper Manyame Sub-catchment was deemed to be of moderate groundwater potential. Rahmati et al. (2015) also demonstrated the use of GIS and Saaty's AHP for groundwater potential mapping. They defined a Groundwater Potential Index (GWPI) which they used to map groundwater potential using spatial layers such as drainage density, rainfall, and lineament density amongst others. Recently, Abrar et al. (2021) used the AHP method to assess the groundwater potential of the western escarpment of the Ethopian rift valley. Some of the factors that they considered include slope, rainfall, geomorphology and lithology (Adhikari et al. 2018). Overall, studies show that an area's groundwater potential can be mapped using GIS and earth observation data (Rahmati et al. 2015).

Water resources in Southern Africa are highly vulnerable to climate change and adaptive capacity is low (Kusangaya et al. 2014), which inherently makes the people vulnerable as well. To understand climate change impacts on groundwater potential and recharge, climate change projections must be used. Climate change projections are computed using General Circulation Models (GCMs) under future scenarios. Some current projections are from the following GCMs: CCCma-CanEM2, CSIRO, MOHC-HadGEM3, MPI-ESM-LR and IPSL-CMSA-MR under RCP 4.5 and RCP 8.5 for the period 2020–2080. Unfortunately, the climate change projections are provided at a coarse spatial scale for hydro-geological analysis (Amraoui et al. 2019; Masimba et al. 2019; Adhikari et al. 2018). Therefore, the projections should be downscaled. There are two broad climate variable downscaling approaches: i.e. Statistical Downscaling and Dynamic Downscaling used to account for point weather station variable. Dynamic downscaling involves the use of a regional climate model or numerical weather model to make use of the model's physics to produce high resolution data. As an example, at a continental scale, the Coordinated Regional Downscaling Experiment (CORDEX) program of the World Climate Research Program (WRCP) provides dynamically downscaled data for the Africa region (Nikulin et al. 2012). On the other hand, statistical downscaling involves the establishment of empirical relationships between historical/future large-scale atmospheric variables and their local scale equivalents. The statistical relationships are then used to produce high resolution data. In Zimbabwe's Upper Manyame Sub-catchment, Masimba et al. (2019) demonstrated the use of a statistical method to downscale GCM climate projections. However, this work did not extend to an impact assessment as shall be done in this study. Whilst a number of studies have worked on climate change impact studies in Zimbabwe (Maviza & Ahmed 2021), it is noted that only 12% of the impact assessment work (n=52) over the past 29 years has focused on hydrologic studies in general, but groundwater has not been studied. This is probably due to the dearth of data to fully develop conceptual and numerical models as highlighted above. However, we propose that GIS and earth observation methods can be extended to climate change impact assessment at regional scales for getting first-order insights that can guide programming, policy and decision making.

Given the above context and challenges in the Runde catchment, this research sought to assess the impact of climate change on groundwater potential and recharge in a data-scarce region, using GIS and earth observation data (including remote sensing). For the first time in Runde catchment we generate catchment scale insights on climate change impacts on groundwater recharge and potential using earth observation methods. The methodologies presented in this study are applicable to region-scale assessments. Specifically, the objectives were: (i) to determine the spatial variability of groundwater potential, (ii) to analyse the spatial and temporal variation of groundwater recharge, (iii) to analyse downscaled climate variables (rainfall, temp, and ET) from CORDEX-Africa for the Runde catchment, and (iv) to assess the impact of climate change on groundwater potential and recharge.

The research was carried out in the Runde catchment which lies in the south-eastern arid regions of Zimbabwe (Figure 1). Geographical location is at 20° S and 30° E.

The area is approximately 42,364.75 km². Runde catchment is divided into five sub-catchments according to its main river systems namely: Runde, Lower Runde, Tokwe, Mutirikwi and Chiredzi. Runde covers part of the Midlands to the Lowveld of Masvingo provinces of Zimbabwe.

Runde catchment is also home to several overcrowded communities (Shurugwi, Chivi, Zaka, Zimuto and Chiredzi) that makes it vulnerable to access potable water in the form of groundwater. Runde groundwater is practically affected by seasonal major river flows and the catchment is underlined by rocks of the basement complex which forms localized aquifers (Misi et al. 2018). Hence the base flows are very limited. Thus, most rivers flow between mid-November to March, the rainy season, while there are practically no flows on major rivers for the rest of the year (Gumindoga et al. 2018). The catchment is significant to Zimbabwe since it houses the country's largest surface water bodies, such as Tokwe-Mukosi which is the largest inland and Lake Mutirikwi, a centre for tourist attraction together with Great Zimbabwe ruins. The water bodies are mostly fishing grounds, and sources of drinking and irrigation water to the sugarcane plantations in the southern low-veld of the country. The water bodies also offer great potential in mitigating the negative impacts of climate change (Barua et al. 2020).

Illegal gold panning along the river stream banks play a major role in socioeconomic activities and threatens water flows and quality due to mercury, violating the Mines and Mineral Act of 1996 (SADC-GMI 2019).

The obtained values are converted from raw digital number (DN) values to solar electromagnetic radiation reflectance in either band. NDVI values range from −1 to +1, and values close to +1 has more vegetation cover.

Soil data was retrieved from https://soilgrids.org/#/?layer=ORCDCM_M_s12_250&vector=1.

The groundwater recharge was obtained from using rainfall data. Groundwater recharge can be determined as a fraction of rainfall considered (MacDonald et al. 2012). To get spatial data, satellite rainfall estimates were used. The CHIRPS satellite rainfall data was used.

Sen's slope estimator

Q=Q_(N+1)/2, if N=odd of ½ Q_N/2=Q_(N+2)/2 if N=even.

This study used various ways of evaluating the homogeneity of hydro-meteorological data. To verify the conformity of projected hydrometeorological data to observed data a homogeneity classification was done. The significant changes in the mean value of the time series were evaluated using the homogeneity classification of data (Warnatzsch & Reay 2019; Pohlert 2020).

The soil groups in the Runde catchment were ferralic (1168 km²), chromic (8571 km²), harplic (10,499 km²), Lithic (9661 km²), sodic (3002 km²) and others (1098 km²) (Figures 3–5). The results demonstrated that areas with clay (ferralic) were assigned the least value of GWP because of the presence of clay horizons in the area considerably restricting percolation whereas the highest value was assigned for sandy loam (chromic). This was due to their low water holding capacity and high permeability allowing fast percolation for groundwater recharge (Interconsult, 1985; Crosbie et al. 2019). Spatial distribution of geological information ranged from sandstone to intrusive granite. The sandstone, sediments, limestone, and alluvium soils are sedimentary rock (shamvaian and karoo, 5537.071 km², 12.6% and 3116.128 km², 7.1%) which are porous materials that have high infiltration capacity. The basaltic, ultramafic lavas, intrusive granite (great dyke, 28,556.528 km², 65%) and granophyre are produced from volcanic and highly impervious (sebakwian, 882.379 km², 2.0%), inland water covering 46.844 km², 0.11%). The Andesitic, gneisses, Older Gneiss complex, kimberlite, paragneiss's (Beitbridge, 5584.521 km², 12.7%) and serpentine and pyroxenites (bulawayan, 124.675 km², 0.28%) are also impervious. Areas around Mogenster are surrounded by young intrusive granite and Zimuto is surrounded by limestone, shale, and quartzite resulting in low to moderate GWP (SADC-GMI 2019). The areas around Renco, Ngundu and Neshuro are surrounded by gneisses, metamorphic rocks which are impervious and porosity is only achieved through rock fracture by overlay burden (Bonsor et al. 2018). Rock water-bearing is increased by the interconnection of the fractures within the rock structure. Dolerites and gabbros are more exhibited around Chivi, Mandamahwe and Zvishavane 14,198.323 km², 32.3% south-west of the catchment area GWP achieved by weathering and fractures. Hydrogeological volcanic areas are common in most drought-prone areas of Southern Africa (Ketema et al. 2016; Bonsor et al. 2018). The results exhibit that the hard rock aquifers of heterogenous and discontinuous weathered fractures pose a high risk of drilling failure (Rahmati et al. 2015; Rajaveni et al. 2017; Bera et al. 2021). The results show grasslands are spatially distributed throughout the Runde catchment from Chilonga, Lowveld to Guinea Fowl, highveld. The grasses varied from annual, Sporoborous to Hyperhenia at Lowveld and highveld respectively.

Slope relates to the general direction of the groundwater flow and its influence on groundwater recharge. Gentle slope of <2% indicates the presence of high GWP zones whereas steep slope >15% shows the presence of poor GWP zones as water runs rapidly off the surface and does not have sufficient time to infiltrate the surface, keeping other parameters constant, ceteris paribus (Ahmed & Al-Manmi 2019). The low GWP areas show poor soil water infiltration and hard basement rock (Crosbie et al. 2015). The topographical wetness index illustrated areas of high moisture content in the soil. The wetness is spatially distributed evenly around the whole Runde catchment. The lowest pockets are dotted across the whole study area distinctively at areas around Chivi, Zimuto, and Ngundu of low to moderate GWP (Ahmed & Al-Manmi 2019).

Rainfall was bias corrected with CHIRPS 29 data up to 2006. Spatially distributed rainfall had a minimum of less than 20 mm/year for Chivi, Mberengwa and Chilonga, and a maximum of more than 140 mm/year for Guinea and Mogenster in Runde catchment. Minimum rainfall was received in 2001/2002, while maximum was in 2002/2003.

The precipitation trend, exhibiting variation of precipitation, show the impacts of climate change being felt (Amraoui et al. 2019; Warnatzsch & Reay 2019; Jannis et al. 2021).

Groundwater recharge is directly proportional to received precipitation, hence the groundwater potential (Figure 6). The recharge map showed a range from the lowest at 16 mm/year to the highest at 24 mm/year which is 3.5% of the total precipitation (Stanzel et al. 2018; Crosbie et al. 2019; Wu et al. 2020). Areas with the lowest groundwater recharge are around Chilonga, Boli, Chivi, Triangle, Mandamahwe and Mberengwa. The recharge was mapped from the precipitation trend pattern (Lekula & Lubczynski, 2018). High drainage density and low infiltration results in low groundwater recharge (Misi et al. 2018). High sloping areas have increased runoff with less water retention on the ground surface, and hence reduced groundwater recharge (Rajaveni et al. 2017).

The spatial variation of groundwater potential map showed to be low mostly in areas with low elevation where precipitation is low as well, this concurred with previous studies (Figure 7). On the other hand, some portions are observed with low groundwater potential at rock formation inhibiting infiltration and permeability (Misi et al. 2018; Interconsult, 1985). However, some portions of high groundwater potential are dotted through the catchment probably following the pattern of the topographic wetness index (Misi et al. 2018). This is an indication of high accumulation of soil moisture due to rising water table (Gumindoga et al. 2020). In the Pearson correlation presented, this research showed the highest r for borehole yield capacity, 0.773, lithology, 0.11, rainfall 0.006 which are positive, however drainage density, −0.236, elevation, −0.040 and wetness, −0.001 had an inverse relationship. Thus, an increase in drainage density, elevation and wetness index reduces the expected GWPI.

The low to moderate groundwater potential is from a granitic formation called Shamvaniah (Misi et al. 2018; Chikodzi & Mutowo, 2014) due to poor weathering and fracture formation. The high to very high yield groundwater potential is at Bulawayan aquifers (Misi et al. 2018; Rajaveni et al. 2017; Macdonald et al. 2008). The findings are supported by studies that deduced that the higher the drainage density, the higher the runoff and the lower the infiltration water into the subsurface (Rajaveni et al. 2017).

The validation results have shown that out of 62 boreholes 50.1% of the borehole capacity yield fell in regions of low GWP, 1.6% very high GWP, 43.5% moderate GWP and 4.8% in the region of high groundwater potential (Figure 8). The GWP map for the study show four different zones, low, moderate, high and very high, classified as in previous studies (Misi et al. 2018). The validation results showed that Gwenhoro, Tongogara exhibits very high (>7 L/s) GWP, Mutema, Matenda, Batsire and Eastlea have high (4–7 L/s) GWP, Taru, Rukasi and New town have medium (1–4 L/s) GWP whereas Chivi, Chirozva and Mushandike are low (<1 L/s) GWP. The low to moderate groundwater potential is from a granitic formation called Shamvaniah (Misi et al. 2018; Chikodzi & Mutowo, 2014). The high to very high yield groundwater potential are in Bulawayan aquifers (Misi et al. 2018; Rajaveni et al. 2017; Macdonald et al. 2008).

A significant correlation exists between borehole yield capacity and groundwater potential index, r=0.63, n=62 and p-value >0.05. The high GWP is supported at quartzite, schist and psammite that behaves like dolomite and limestone due to weathering and fractures.

There is good confidence in temperature projections, although precipitation scenarios remain uncertain though it has a great impact on groundwater recharge (Figure 9 and Table 1). The changes in precipitation amount and intensity are more important than changes in temperature on groundwater recharge (Stanzel et al. 2018). The likely future changes for climate change impacts assessments are more related to precipitation (Warnatzsch & Reay 2019).

The future projections (2020–2080) and historical observed for Runde catchment rain gauge stations are plotted to validate the trends. The projected climate variables show no trend among themselves, but there is a similar visual trend against observed data.

GCM projected rainfall was analysed for trends and change-point detection over the meteorological stations used in this study (Tables 2 and 3).

The Mann Kendall trend test and Sen's slope were used to detect the presence of monotonic trends (Moses & Ramotonto 2018). The absence of Sen's slope for precipitation indicates variation of intensity and amount of projected rainfall for Runde catchment. A decline in precipitation with no statistical significance was noted for Hwange in Zimbabwe. Jannis et al. (2021); Touma et al. (2015) and Bera et al. (2021) highlighted precipitation and temperature as major climate variables which directly affect GWP. The results showed no trend for the simulated variables. The p-value is greater than the significance level alpha (0.05). There is variation to the extreme events, droughts and floods.

The change-point analysis (homogeneity) showed that p-value is generally less than the level of significance. Thus the increase in temperature has an effect on the groundwater potential.

The SNHT showed sensitive signal change results with breaks near the start and end of the time series also observed in studies. The Von Neumann ratio depicted more sensitivity to loss of homogeneity of nature compared to strict stepwise shifts that could not detect the location of the shift (Pohlert 2020).

Trends analyses results showed an average decrease of 3.23 to 0.23 °C for the catchment. There is a negative Sen's slope, supporting the decreasing trend in the test. There is no temperature trend.

These results differ from temperature analysis results for meteorological stations, Beitbridge, Bulawayo and Harare highlighted a statistically significant positive trend for daily maximum temperature in the 20th century (Molina & Bernhofer 2019).

The trend analysis conformed with the study carried out for Rozwa dam in Bikita district of Zimbabwe showing an increased variation in precipitation data trend, 1953–2008 (Molina & Bernhofer 2019; Warnatzsch & Reay 2019; Moses & Ramotonto 2018). The result opposed precipitation decline highlighting ultimately reduced surface water ponding across Zimbabwe (Masimba et al. 2019). The results for temperature analysis showed p-value great than alpha 0.05, with Sen's slope ranging from −8.27 × 10⁻⁸ to −1.36 × 10⁻⁸ predicting some decreasing trends. The results concurred that Beitbridge, Bulawayo and Harare had a statistically significant negative trend for daily maximum temperature in the 20th century (Molina & Bernhofer 2019). The same report predicted future scenarios to increase mean global temperature from 1.3 to 4.6 °C by 2100, resulting in global warming of 0.1–0.4 °C per decade. Climate change and variability in south-eastern Zimbabwe studied for 1960–2010 data for Zaka weather stations showed a mean temperature increase (Jannis et al. 2021). The study highlighted RCP 4.5 had the least temperature increase compared to RCP 8.5 with the highest warming rates and their mean produced a range of 1.5–5.4 °C (Molina & Bernhofer 2019; Warnatzsch & Reay 2019). The RCP 4.5 and RCP 8.5 scenarios have shown the average increase of maximum temperature in the range of 1.5–5.4% (Jannis et al. 2015; Masimba et al. 2019). GWP has shown to be directly proportional to precipitation change within which the aquifer is replenished (Shivakoti et al. 2019). The anticipated increase in precipitation from downscaled data resulted in altering the GWP, which concurs with previous studies (USAID 2020).

The groundwater recharge occurrence is a combination of several favourable influencing factors like slope, land use, lithology, rainfall, elevation, soil and drainage density that reduced runoff generation and encouraged infiltration (Fenta et al. 2015; AIH, 2019) (Figure 10 and Table 4). Groundwater recharge is controlled by long term climate conditions hence any changes in climate can affect groundwater storage and availability in the long term (Fenta et al. 2015; Taylor et al. 2019). The projected general circulation model, precipitation from downscaled data resulted in altering the groundwater recharge, concurring with previous studies (USAID 2020).

Variations in precipitation timing, form, and quantity result in increased frequency of flash floods which drained away from high elevation and mountainous places for the 2080s. The projected future scenarios of an increased mean temperature supported by 1.3–4.6 °C by 2100, resulting in global warming of 0.1–0.4 °C per decade, also influence potential evapotranspiration, reducing groundwater recharge (Touma et al. 2015; Masimba et al. 2019; Warnatzsch & Reay 2019).

The projected precipitation from downscaled data directly altered the groundwater potential (Figure 11). Table 5 shows the groundwater potential projected to the 2020s, 2040s, 2060 and 2080s using the RCP 8.5 and RCP 4.5 scenario.

Figure 10 illustrates the projected GWR for 2020–2080 for RCP 4.5. The areas are downstream of the largest water bodies in the country: Bangala, Tokwe Mukosi and Mutirikwi dam. In these low-lying areas, drainage density is low with high infiltration expected, resulting in high groundwater potential (Rajaveni et al. 2017; Mahato et al. 2021; MacDonald & Edmunds 2019; Misi et al. 2018). It was shown that igneous and metamorphic rocks upland produced negligible amounts due to the unweathered and non-fractured basement. The groundwater cycle of drought-prone areas like the Runde catchment largely depends on rainfall, which directly influences recharge (Jassas & Merkel, 2014). The projected groundwater recharge for 2020 s, 2030 s, 2040 s, 2060 s and 2080 s.

The GWP was influenced by projected precipitation through weighted combination analysis with other constant predictors. The low-lying areas show an increase in groundwater potential. The low-lying areas are at the south-eastern parts of the catchment. The highland areas are in the north-western parts of the catchment including Gwenoro stretching to Mogenster at the eastern part of the catchment.

The projected results (Figure 12) showed variation in precipitation patterns differently for RCP 8.5, hence increasing groundwater potential (Touma et al. (2015); NUST, 2019; SADC-GMI 2019).

The areas showing an increase in groundwater potential are around Chilonga. The groundwater potential is due to predicted amount of rainfall and percolation from climate change impacts. In these low-lying areas drainage density is low, resulting in high infiltration and high groundwater potential (Rajaveni et al. 2017; Mahato et al. 2021; MacDonald & Edmunds 2019; Misi et al. 2018). The anticipated variation of climate variables (precipitation, temperature and evapotranspiration) from simulations predicted an increase in precipitation, directly influencing groundwater recharge for GWP (Taylor et al. 2019).

In this study the spatial distribution of groundwater potential areas was assessed using thematic layers derived from satellite images and existing data in a GIS environment. The thematic layers included soil types, lithology, drainage density, slope, rainfall, land cover and topographic wetness.

• (i)

  The most promising groundwater potential zones in the Runde catchment are found in alluvial deposits which are highly weathered and have high permeability. Hence the development of productive boreholes can safely target these areas for agricultural purposes. Areas with poor groundwater potential are around structural hills and escarpments. This was validated to low (<1), moderate (1–4), high (4–7) and very high (>7) GWP.

• (ii)

  The spatial variation of the groundwater recharge map shows large areas with low groundwater recharge in drought-prone areas. The situation is being exacerbated by climate change. The current groundwater recharge was 3.5% of rainfall, which ranged from 16 to 24 mm per day.

• (iii)

  Climate variables are downscaled from the 2020s, 2040s, 2060 to 2080s to predict groundwater potential map from the rainfall amount and are the source of the results. Rainfall shows an increase in amount in the lowlying area and a slight decrease in highland areas.

• (iv)

  The research concluded that there is an effect of spatial and temporal variability of average GWP area of 11.05% in drought-prone areas of Runde catchment. Low and moderate GWP decreased 30.8 and 5.8% respectively, high to very high GWP increased by 34.8 and 1.8% respectively.

• (i)

  GIS and Remote Sensing must be prioritized and embraced by planning institutions for groundwater resources management and planning since it is a cheap, fast, and convenient way of analysis

• (ii)

  Due to the sensitivity of groundwater potential and recharge to climate change and climate variability, it is imperative for the government of Zimbabwe and other water resources planning authorities to reform their land and water conservation policies.

• (iii)

  The authorities must improve the rural water information management system (RWIMS) database which has some borehole Global Positioning System (GPS) location without borehole yield capacity.

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