The contribution of fine scale atmospheric numerical models in improving the quality of rainfall forecasts during heavy precipitation episodes

Abstract

The atmospheric numerical models have known great advances with the ongoing development of numerical weather prediction and computing resources. The spatial and temporal resolutions of the global atmospheric models have improved and therefore, the accuracy and reliability of their results have substantially increased. Emphasis was made on the improvement of models dynamics and physical aspects, but also on data assimilation and input data diversification using new numerical schemes and new physical parameterizations that better assess the small-scale weather phenomena. However, these models were not able to overcome their physical limitations and therefore, some small-scale processes are far from being thoroughly apprehended. To overcome these limitations, new numerical models applied to limited areas and finer scale numerical weather models have been developed including additionally, the microphysics of clouds, atmospheric chemistry and soil characteristics (vegetation index, albedo, roughness property, etc.). The output of atmospheric models, particularly in the field of precipitation, is the main input of hydraulic numerical models and flood warning systems. Better control of the rainfall forecasts, especially during extreme weather events, will have a positive impact in improving the quality of weather early warning systems and thus the quality of numerical hydraulic models predictions. The present work illustrates, through two recent cases studies, a real demonstration of the contribution of fine-scale atmospheric numerical models in improving the quality of rainfall forecasts taking into accounts that time series of predicted rainfall amounts issued from the fine scale atmospheric models provide more precise information at the scale of water basin and also that these models have demonstrated an ability to better predict stormy situations that are causing floods in many parts of the country.