Verification of HARMONIE-AROME, ECMWF-IFS and WRF: Visibility and Cloud Base Height

Abstract: Visibility and cloud base height are two important parameters within aviation meteorology and verification of operational numerical weather prediction (NWP) models with regard to these two parameters is therefore important. Visibility is often parameterized through an extinction coefficient, which is dependent on for example liquid or ice water content, however it also depends on the humidity, precipitation intensity and particle concentration. The cloud base height is defined as the lowest level where the cloud fraction exceeds 4 octas and is governed by the parameterizations of for example microphysics, turbulence and radiation. The horizontal and vertical resolutions are also of importance when forecasting the visibility and cloud base height and more advanced parameterizations, higher resolutions and better data assimilation methods may be necessary to improve the forecasts. In this study the three operational NWP models HARMONIE-AROME (HIRLAM-ALADIN Research on Mesoscale Operational NWP in Euromed – Applications of Research to Operations at Mesoscale), ECMWF-IFS (European Centre for Medium-Range Weather Forecasts – Integrated Forecast System) and WRF (Weather Research and Forecasting) are verified with regard to visibility and cloud base height. For occasions with fog HARMONIE-AROME is the most accurate, possibly due to finer resolution, while ECMWF-IFS has a better aerosol representation and is the most accurate at higher visibility ranges. For the cloud base height WRF and ECMWF-IFS provide the best forecasts, despite their low resolutions, and superior parameterizations may therefor be the reason for their better forecasts. A verification of HARMONIE-AROME is also conducted to investigate how model changes may have affected the forecasts. The visibility forecasts does not change noticeably due to any model change, instead a strong seasonal dependence is evident. At a couple of model changes the cloud base height forecasts deteriorated. The wind directions at which the model generally struggles more to provide correct forecasts are those not in the prevailing wind direction.