NWP Model Specific Humidities Compared with CHAMP/GPS and TERRA/MODIS Data

Abstract

Limb sounding observations of the Earth’s atmosphere by the GPS receiver onboard CHAMP allow to determine vertically resolved profiles of the specific humidity. The specific humidity shows a high variability and is a critical parameter for short term numerical weather prediction (NWP). Specific humidity profiles calculated with the hydrostatic NWP model HRM (High resolution Regional Model) of the Deutscher Wetterdienst over the BALTEX (Baltic Sea Experiment) modelling area are compared with profiles from CHAMP/GPS radio occultation data (14th of May to 10th of June 2001) and from the MODIS detector onboard the TERRA spacecraft (dataset from 1st of December 2001, 1100 UTC). Radiosonde data from the station Lindenberg/Germany (1st of May to 10th of June 2001) and ECMWF data (14th of May to 10th of June 2001) are used to evaluate the results. The HRM model used as boundary conditions 6-hourly analyses of the European Centre for Medium-Range Weather Forecast (ECMWF). Consecutive 30 h forecast starting each day at 0 UTC were performed. The mean difference between the HRM and the radiosonde data or the ECMWF data is below 0.4 g/kg while the biases between HRM and CHAMP/GPS are up to 1.7 g/kg and the biases between HRM and TERRA/MODIS are around 0.5 g/kg. The standard deviations of the specific humidity differences between HRM and CHAMP/GPS, TERRA/MODIS, ECMWF or the radiosonde data from the station Lindenberg/Germany are about 1 to 2 g/kg in the lower troposphere (below about 4 km) and decrease with increasing height.