Modeling the Hydrological Response of the Dano Catchment, in the Volta Basin to Landuse Landcover Change and Climate Change

Abstract

This study investigated the scenario-based impacts of land-use landcover change and climate change on water resources of the Dano Catchment, in the Volta basin. Combining remote sensing images, field investigation, and historical climate variables, a scenario analysis was complemented by a spatially-distributed and physically process-based modeling approach to simulate the impact of LULCC and climate change on the catchment governing hydrological processes. First, CORDEX-Africa framework RCMs outputs were thoroughly investigated to gain knowledge on patterns and spatial variations in precipitation and temperature of the Volta basin and determine the model that best replicates the climate of the Volta Basin. Subsequently, climatic scenarios were constructed from these set of Regional Climate Models (RCMs) outputs. Representative concentration pathways (RCP) 4.5 and 8.5 ensemble of four datasets of the CORDEX-Africa project were then used to study the future change in temperature and precipitation. The predicted climate change signal was investigated in detail by comparing the future period 2020 – 2049 with the historical period 1971 – 2005. With regard to the future climate, the used climate models’ ensemble projected a clear trend of reduced rainfall over the catchment and consistent increase of temperature between 1.5°C and 1.7°C. The past LULC change was studied based on classified land use maps of the years 1990, 2000, 2010 and 2016 produced using supervised Maximum Likelihood Algorithm. Based on these maps two future LULC scenarios, viz, Business-as-usual and Afforestation were developed for the year 2050. The LULCC investigation from 1990 to 2016 suggests a decrease of savanna at annual rates of 2.53% while cropland and settlement areas have increased. However, the study noted about 5.3% forest recovery and grassland increased between 2000 and 2010. Prior to impact assessment, data retrieved from CHIRPS and NASA-POWER for Precipitation and Temperature (after validation with station climate data), were used to set up and drive the hydrological model SWAT. The model correctly reproduces the observed runoff regime at the basin outlet. Statistical performance measures (R², NSE, KGE) range between 0.71 and 0.94 for the calibration and validation of surface runoff indicating satisfactory. The modeled future LULC maps and climate scenarios were used to feed SWAT. The isolated and combined influence of LULC and climate change is investigated. The simulations that assumed constant LULC and climate as changing factor indicated decreases in surface runoff of about 77% and 3.6% to 3.9% increase in potential evapotranspiration. The simulation that assumed a constant climate and a changing LULC showed increasing water yield between 0.18% – 0.28% and mainly increasing runoff (0.02 - 3.30%). Actual evapotranspiration was estimated to slightly increase by 0.21% from the baseline condition. The combined application of LULC change and climate change signals gave a clear potential evapotranspiration increase (3.2% to 3.7%) causing surface runoff reduction between 72% - 77% and other components. The application of the historical and future climate data to SWAT showed that future changes in surface runoff follow the predicted precipitation signal. SWAT tends to produce larger hydrological changes under climate change scenarios and runoff showed more sensitivity to changes in precipitation. In summary, there is a strong relationship between runoff, climate, and human factors in the catchment that could result to increase in drought and flood risks, signifying the vulnerability of the catchment to future water availability. Therefore, efforts should be geared towards strengthening the operation of existing soil-water conservation techniques to enhance water infiltration and aquifer recharge.