Browsing by department "Ilmakehätieteiden keskus"

The thesis examines the use of a cloud initialization method in which Meteosat satellite observation-derived cloud products are used to adjust the initial post-data assimilation model cloud fields in numerical weather prediction (NWP) models. The method was developed by van der Veen (2013) and uses cloud masks from Meteosat geostationary weather satellites together with surface-based cloud base height observations to improve the initial clouds, and through them the cloud forecast. Initially developed with HIRLAM, a short-range NWP model, it has later been used with nowcasting-range NWP models based on the HIRLAM and HARMONIE-AROME NWP models. Distinguishing features of nowcasting NWP models and the use of satellite data in NWP models are discussed. The cloud initialization method and later changes to it are described, and the effect of the method on the quality of cloud forecasts is examined. New verification scores for the current pre-operational MetCoOp-Nowcasting (MNWC) model, which includes the latest version of the cloud initialization method, are examined for the months of June and September 2020. The primary weather parameter verified was cloud cover, but other key parameters are also discussed. The results of the pre-operational MNWC were promising, due to monthly-average cloud cover biases being reduced. MNWC scored higher than the reference model in forecasting cloud-free and mostly cloud-free situations in particular, but some improvements were also seen for situations with denser cloud cover. Results for other model fields were mixed. Further verification of the MNWC model is needed. In the future, there is high potential for the further development of the method, with ongoing development of HARMONIE-AROME, the cloud initialization method, the cloud products’ algorithms, and as the next generation of Meteosat satellites enters operational use in the next few years.