Sub-grid variability in rainfall distribution has been widely recognized as an important factor in include in the representation of land surface hydrology within climate models. In this paper, using West Africa as a case study, we investigate how the sub-grid variability in rainfall distribution affects the modeling of rainfall intersection and other processes within a coupled dynamic biosphere0atmosphere model, According to our results, even the evapotranspiration is tuned to be consistent with observations while neglecting spatial variability of rainfall, significant errors may result in the presentation of surface hydrological processes and surface energy balance. However, the extent of the resulting errors ins not limited to the surfaces processes. They extend to the atmosphere via the low-level clouds feedback to impact solar radiation, boundary layer energy, atmospheric circulation and the distribution of precipitation. The same errors also propagate into the biosphere through vegetation dynamics and can eventually lead to a significantly different biosphere-atmosphere equilibrium state. This study provides a good example for the need to have physical realism in modeling most of the details of the complex biosphere-atmosphere-ocean system.
