Publications

Oussou, FE; Ndehedehe, CE; Yalo, N; Asiwaju-Bello, YA; Oloukoi, J; Odeloui, DTJ; Boukari, M; Diaw, AT; Kunstmann, H (2025). Improving the depth-to-bedrock maps for groundwater-to-atmosphere modeling in Africa. JOURNAL OF HYDROLOGY, 656, 132964.

Abstract
In recent years, large-scale depth-to-bedrock (DTB) maps have been produced at the global level, but none have integrated field borehole measurements from Africa. This study addresses this gap by using over 4,000 field depth-to-bedrock measurements, suggesting new computation approaches, and providing DTB maps. After calculating the likelihood of thick layer occurrence from the digital elevation model, sixteen different past and proposed approaches were used to compute the depth-to-bedrock. Across different distance classes from three terrain-related features (geological lineaments, depression zones, and river networks), the average DTB values for means, medians, standard deviations, and maximums are respectively 16.5 f 1.2 m, 15.5 f 1.8 m, 10.8 f 1.3 m, and 54.6 f 16.6 m. To evaluate the reliability of the estimated DTB, the widely reported Parflow-CLM (PF-CLM) model is setup for a 100 km2 domain around an Eddy Covariance station in the Nazinga Park located at the border between Ghana and Burkina Faso. The static variables used to run the model are the HydroSHED elevation products, MODIS land use, FAO Soil data, and geological dataset derived from GHLYMPS while the meteorological forcing are made of the air temperature, surface pressure, specific humidity, shortwave radiation downwards, longwave radiation downwards, 10 m u and v components of wind, and total precipitation. The main purpose of the test case is to check whether the PF-CLM model setup with the estimated DTB can offer any reliability for further hydrological problems implementation in the data scarce area. Therefore, the model is run at a hourly time step from January to February 2014. It is remarkable that despite the relatively small size of the domain, PF-CLM model captures very well the water and energy fluxes occurring at the land surface with NSE and R2 values reaching respectively 0.69 and 0.6 for the latent heat, 0.77 and 0.77 for the sensible heat, 0.67 and 0.86 for the outgoing longwave radiation, and 0.34 and 0.71 for the ground heat. Both the spatial distribution and temporal dynamics of the water and energy fluxes displays values that are within the expected for the domain. These field DTB statistics and the simulation accuracy contribute to reducing uncertainties regarding the depth of basin bottom boundaries, especially in Hard Rock areas. The DTB maps produced in this study are relevant for large-scale groundwater modeling and will aid in improving the integration of the groundwater components into Land Surface Models (LSMs) especially in the data scarce areas of Africa.

DOI:
10.1016/j.jhydrol.2025.132964

ISSN:
1879-2707