Jump to Main Content
Sensitivity of different physics schemes in the WRF model during a West African monsoon regime
- Gbode, Imoleayo E., Dudhia, Jimy, Ogunjobi, Kehinde O., Ajayi, Vincent O.
- Theoretical and applied climatology 2019 v.136 no.1-2 pp. 733-751
- air temperature, data collection, model validation, monsoon season, rain, satellites, surface temperature, troposphere, weather research and forecasting model
- A 2-month (August–September) regime of the year 2007 West African monsoon (WAM) was simulated with 27 physics combinations using the Weather Research and Forecasting model at 20-km horizontal grid. The objective is to examine WAM sensitivity to parameterization of microphysical, convective, and boundary layer processes for long-term simulation. The model precipitation was evaluated against the TRMM, CMORPH, and GPCP satellite rainfall products. The surface temperature was compared against the ERA-Interim, NCEP, MERRA, and global surface air temperature, an ensemble of the three reanalysis datasets. Model skill score (MSS) computed from a synthesis of the normalized correlation coefficient, mean bias, and mean absolute error was used to rank the model performance. Results show the model adequately simulates the diurnal cycles of surface temperature than precipitation, as well as the westward propagation of intense precipitation associated with the African easterly waves. The new Grell-Freitas (nGF) cumulus parameterization scheme (CPS) outperforms its predecessor especially when combined with the Mellor-Yamada-Nakanishi-Niino 2.5 (MYNN) planetary boundary layer scheme. The new simplified Arakawa-Schubert (nSAS) and Tiedtke CPSs produced better simulation of precipitation and surface temperature, respectively. The simulation of observed peak of diurnal precipitation in nSAS and nGF highlights success made towards a more realistic representation of convective processes by the schemes. Goddard microphysics and MYNN performed better for both variables. Based on the MSS, some relatively good and poorly performing combinations for precipitation and surface temperature were identified. The optimal combinations are however not separated in a statistically significant way and, thus, could be used for long-term simulation of WAM.