942 Optimizing WRF as a Regional Climate Downscaling Tool for Hydro-Climatological Applications in the Eastern Nile Basin

Wednesday, 10 January 2018
Exhibit Hall 3 (ACC) (Austin, Texas)
Mahmoud Osman, JHU, Baltimore, MD

Most of the Nile water inflow at Aswan, Egypt, originates from the Ethiopian plateau, which provides the main source of water for the Eastern Nile (EN) basin. Due to climate change and variability, the hydrological attributes of the Nile may fluctuate, mainly due to changes in precipitation and temperature patterns. This study contributes to the development of an integrated hydro-climate model for the EN for the impact assessment of the Nile inflow at Aswan. The climate component consists of a regional atmospheric model for the EN basin, planned to simulate climate scenarios with adequate precision. The Weather Research and Forecasting (WRF) model is configured for a domain centered over the EN with a parent domain defined for the Middle East – North Africa (MENA) by the (CORDEX). WRF physics parameterization sensitivity experiments are carried out to select an optimum combination of physics schemes to reproduce observed climate conditions. The model skill is also examined by downscaling the ERA-Interim reanalysis dataset from 1980 to 2009 over the EN prior to the model use in simulating climate projection scenarios.
The WRF performance is assessed using gridded observational datasets for precipitation, temperature and evapotranspiration. The model is tested for four different configurations in two-year simulations to determine the optimal combination of physics parameterizations prior to the 30-year downscaling experiment. The results indicate accurately modelled temperature and evapotranspiration fields, however, with significant positive precipitation biases, especially over the highlands. The bias-corrected precipitation data is coupled to the semi-distributed hydrological rainfall-runoff model (SWAT), previously configured for the Baro-Akobo-Sobat sub-basin. The simulated flow hydrograph based on bias-corrected WRF simulations yields high correlations and Nash-Sutcliffe Efficiency coefficients for the observed flow hydrographs. Results indicate that the simulated precipitation fields from WRF should be subject to bias correction prior to use in hydrological models especially for impact studies.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner