LAUSR

Abstract In our previous work, a practicable objective three-parameter multivariate calibration method build on a quadratic meta-model was introduced leading to the optimization of COSMO model. The process had been applied initially for a regional climate model (RCM) providing a considerable alternative to the, more standard, expert tuning techniques. Based on the successful implementation of this method, the calibration of the model on the operational level is undertaken in the present work. A larger number of parameters are selected in order to optimize the model performance for a high horizontal resolution (~2 km) over the domain of Switzerland and Northern Italy as well an extensive period of one year. The influence of the variables is associated to daily forecasts such as daily minimum and maximum 2 m temperature as well as 24 h accumulated precipitation. The increased computational demands and the substantial complexities regarding the application of a meta-model for a large parameter number are successfully tackled. The forecast skill of the 2-m temperature was improved by lowering the temperature bias by 0.2 °C. Contradictory, a small decrease in forecast skill for precipitation and dew point was observed. There is practically no change in forecasted wind speed when using the optimum set of parameters as none of the parameters tested is directly affecting model's ability to forecast wind. Currently, the results indicate a relatively low benefit with respect to the computational cost of the method as it remains expensive for a regular usage of the calibration procedure.