MeteoSwiss - COSMO-LEPS forecasts

The imprecise knowledge of the state of the atmosphere is one of the reasons for inaccuracies in weather forecasts. Ensemble prediction systems try to account for these uncertainties by calculating several forecasts with slightly varying initial conditions. The resulting ensemble of forecasts is used to estimate the probability of occurrence for a certain event. Besides that, the ensemble forecast gives an indication of the predictability of an event, hence, a measure for the reliability of a forecast.

Figure 1: COSMO-LEPS forecast for precipitation, wind gusts and temperature. The ensemble is composed of 16 individual forecasts. The plot shows the mean as well as the uncertainty range of the forecast.

New forecast products

COMSO-LEPS (limited area ensemble prediction system) forecasts are used to create meteograms, compilations of forecast parameters for the next days at a given location (Figure 1). The meteograms show the mean of all forecast members (red line) and an uncertainty band (gray area) that envelops 50% of all forecast members. The width of the uncertainty band is a direct measure for the uncertainty of the prognosis.

A different way of using ensemble forecast are probability plots, e.g. for precipitation (Figure 2). These plots show the probability that a certain event will occur. A forecasted probability of e.g. 70% to exceed 20mm precipitation means that this event will happen in 7 out of 10 cases. Compared to deterministic forecasts (one single forecast, giving only probabilities of 0% or 100%) this is additional information which is especially important for decision making processes. The forecast user is able to decide if he wants to take preventive action or not, depending on his personal risk profile and therefore profit from using the forecast.

Figure 2: COSMO-LEPS probability forecast for exceeding 50,100,150 and 200 mm precipitation in 72h