Adapting the Weather Research and Forecasting Model for the Simulation of Regional Climate in East Africa

Abstract

The regional climate model (RCM) is a tool to fill the gap between the outcome of global climate model and the demand of regional government and society for regional climate information. The current Weather Research and Forecasting (WRF) RCM inherits several advantages of the original WRF model. For example, (1) it can be used for multiple scale applications from large eddy to planetary scale; (2) it can be coupled to other climate system component models; (3) it has multiple physical options for different purposes of research; and (4) it has parallel infrastructure to distinguish the scientific problems from engineering problems.

In order to adapt WRF for long term integration for climate application, the model and its pre- and post-processors are modified in several aspects, including SST and other surface parameters treatment, and exponential relaxation lateral boundary. In this study, we also transferred a pre-released version of CAM3 radiation scheme that was first implemented into WRF by NCAR Regional Climate Modeling team to the WRF standard release versions (2.1.1 released as of November 8, 2005 and 2.1.2 released as of January 30, 2006). The preliminary investigation demonstrates that the model has shown the encouraging and promising results on simulating regional climate.

The objective of this study is to investigate the hydrological cycle in the East Africa region. Through data analysis, the water channel in the East Africa region is determined. The moisture transport from the Indian Ocean is the main source of the water vapor for precipitation in "short rain" season from October to December (OND). The variability of the moisture transport directly results in the variability of precipitation in OND. The local soil moisture, which directly connects to evaporation, does not substantially contribute to precipitation in the region.

To investigate the physical processes of the variability of the moisture transport, five seasonal simulations from September to December for 1994 to 1998 are conducted by WRF RCM. The model successfully simulates the interannual variability of two positive Indian Ocean Dipole Mode (IODM) episodes (1994 and 1997), two negative IODM episodes (1996 and 1998), and the meridional mode in 1995. The analysis further demonstrates that the model can reproduce many surface features of the interannual variability related to IODM changes, such as the mean sea level pressure, wind regime, and surface air temperature. The seasonal change of the monsoon system is also well represented. The results show the interannual variability in East Africa and the adjacent region is mainly related to the variability of the Indian Ocean SST. Through sensitive experiments of different resolution SST forcing, we suggest that the SST forcing itself and SST gradient plays a role on exerting effect on atmosphere. The current conclusions are very preliminary and the results should be further carefully examined in future publications.