8.3 High-resolution seasonal precipitation in a semi-arid complex terrain area using WRF-FDDA

The present work demonstrates the capability of the Weather Research and Forecasting (WRF) model with four dimensional data-assimilation (WRF-FDDA) to create a high-resolution climatography of seasonal precipitation over the south-eastern Mediterranean. The system was used to dynamically downscale global Climate Forecast System (CFS) reanalysis with continuous assimilation of conventional and unconventional meteorological observations. Precipitation seasons (December-January-February) in 7 years, including an extreme dry and an extreme wet season observed in the past decades, were generated at a 2-km spatial resolution. Verification against rain-gauge observations at 5 hydrological basins shows that the WRF-FDDA system successfully reproduces the spatial and inter-annual variability of the accumulated seasonal precipitation, as well as the timing, intensity, and length of wet and dry spells. The best agreement between model and observations was obtained at areas dominated by complex terrain, illustrating the benefit of the high-resolution lower-boundary forcing in the dynamical downscaling process. On the other hand, some model positive biases were found over coastal flat terrain, raising the possible need for more accurate sea-surface temperature data. The model was able to reproduce some of the extreme events, but exhibited limitations in the case of rare events. This specific disagreement between model and observations suggests that different fine tuning and model configurations may be needed to correctly simulate rare events.