Model validation and verification

Precipitation forecasts

The prediction of rainfall still poses a major problem to todays weather forecast models, especially in areas with a prominent topography such as the Alps in Switzerland. A quantification of these errors is a coercible prerequisite for the improvement of these forecasts.

High - resolution NWP models

A higher resolution of weather forecast models allows to increase the number of implicitly simulated atmospheric processes and thereby improve the model.

Forecast error quantification based on potential vorticity

Potential vorticity can be used to identify model errors.

• Validation and intercomparison of global assimilation systems as the ECMWF with focus on the jet-stream representation and stratospheric temperatures. (Sandro Buss)

• Johannes Jenkner investigates Quantitative Precipitation Forecasting on the basis of the limited area model of MeteoSwiss. Here, forecasted precipitation fields are compared to a novel observational dataset based on both rain gauge measurements and RADAR images. The application of conventional and novel verification methods renders possible a distinctive conditional verification of the model output and facilitates the investigation of dynamical impacts on the forecast quality. Error climatologies are produced for different synoptical settings and case studies are investigated by sensitivity experiments. (Johannes Jenker)
  

• Concluded research