Validation of time-lagged ensemble forecasts relevant for predicting aircraft wake vortices

Abstract

Time-lagged ensemble (TLE) forecasts of the high-resolution model COSMO-FRA with an hourly update cycle and a lead time of 6 hours are validated against measurements of a Wind-Temperature Radar (WTR) for three cases representing different weather regimes, a frontal passage, a stormy situation and a high pressure system. Forecasts and measurements are available every 10 minutes. The present study focuses on parameters relevant for forecasting aircraft wake vortices, namely wind, virtual potential temperature and Turbulent Kinetic Energy (TKE) for altitudes below 1600 m. When compared with a reference run started once a day at 00 UTC with a lead time of 24 h a reduction of forecast error in the first two forecast hours is observed for all wind components. For wind speed at low levels around 400 m this reduction can be of order 1.7 ms-1 for the 1-hour forecast member and 0.9 ms-1 for the equally weighted ensemble mean depending on weather situation. TLE forecasts of virtual potential temperature show an improvement of close to 1 K except for the high pressure system while forecast errors of TKE are reduced below 900 m in all cases. Since no data assimilation is used, the improvements of the forecasts are the result of the hourly update cycle which benefits from more up-to-date initial- and boundary conditions provided by the driving COSMO-EU model. The results of this study may contribute to improve the quality of wake vortex predictions and encourage the assimilation of temporal high resolution wind data to further improve the very short range forecasts.