Evaluation and selection of a set of CMIP6 GCMs for water resource modeling in the poorly gauged complex terrain of the Tana River basin in Kenya

The Tana River basin is among the least monitored in terms of meteorological data in Kenya. The Kenya Meteorological Department (KMD) provided data on a ten-day timescale, which is not adequate for water resource evaluation. To bridge this data gap, there is a growing need to leverage General Circulation Models (GCMs) and global datasets to assess current and future water resources in this basin. This study focused on evaluating the performance of 19 CMIP6 GCMs concerning precipitation (pr), maximum temperature (tasmax), and minimum temperature (tasmin) for the complex terrain of the Tana River basin. This involved a rigorous process of disaggregating the data provided by the KMD into a daily timescale for downscaling. The GCMs’ historical output was prepared using the Climate Data Operator (CDO) in Cygwin. The Kling Gupta Efficiency (KGE) was computed for each variable at three stations: Nyeri (upstream), Kitui (midstream), and Bura (downstream). The KGE results were validated using Taylor statistics. Five GCMs, CMCC-ESM2, MPI-ESM1-2-HR, ACCESS-CM2, NorESM2-MM, and GFDL-ESM4, performed best with a multivariable Multi-station KGE statistic of 0.455–0.511. The outputs from these selected GCMs were subsequently downscaled for later use in assessing the water resources and crop water demand in the basin.