Item

Abstract

Ten planetary boundary-layer (PBL) parameterization schemes in the Weather Research and Forecasting (WRF) model have been evaluated with respect to temperature and wind forecasts for typical summer conditions in the eastern Mediterranean island of Cyprus. An ensemble of twenty-two simulations was performed, for a combination of the ten PBL and compatible surface layer parameterization schemes and each of the setups has been evaluated in view of forecasting skill and other statistical indices. Comparison of the model results with measurements from eight sites in Cyprus revealed significant differences in forecast skill, during both daytime (unstable conditions) and nighttime (stable conditions). The time series produced by the simulations were assessed against observations, in an effort to identify biases and skills in the forecasting ability of the models, employing commonly used statistical performance measures, such as the Mean Error, Mean Absolute Error, Mean Squared Error and Root Mean Squared Error. In summary, most simulations exhibit a negative bias for all stations concerning temperature and nearly all have positive biases for wind speed and direction. Statistical analysis of the results identified the best performing parameterization configurations for further investigation, to determine the most suitable setup for summertime forecasting in Cyprus, focusing on coastal stations, where land – sea breeze circulations govern the diurnal weather characteristics.