Assessment of global reanalysis precipitation for hydrological modelling in data-scarce regions: A case study of Kenya

19 flood prone catchments in Kenya, Eastern Africa Flooding is a major natural hazard especially developing countries, and the need for timely, reliable, actionable hydrological forecasts paramount. Hydrological modelling essential to produce but challenging task, poorly gauged catchments, because of inadequate temporal spatial coverage hydro-meteorological observations. Open access global meteorological reanalysis datasets can fill this gap, however they have significant errors. This study assesses performance four (ERA5, ERA-Interim, CFSR JRA55) over Kenya period 1981–2016 on daily, monthly, seasonal, annual timescales. We firstly evaluate by comparing them against observations from Climate Hazards group Infrared Precipitation with Station. Secondly, we ability these simulate streamflow using GR4J model considering both parameters sensitivity identifiability. While ERA5 best performing dataset overall, varies season, catchment therefore there are marked differences suitability products forcing models. Overall, wetland western regions highlands obtained relatively better scores compared those semi-arid regions, inform future applications setting up models that be used forecasting, early warning, action data scarce such as Kenya.