Validation of Integrated Water Vapor from Numerical Models Using Ground-Based GPS, SSM/I, and Water Vapor Radiometer Measurements

Abstract

Time series of vertically integrated water vapor (IWV) derived for 25 ground-based global positioning system (GPS) sites in Sweden and Finland are used for validating the IWV from weather forecast models for a 3-month period from August through October of 1995. Results from analyses and forecasts of the Baltic Sea Experiment (BALTEX) Model and Europa Model [operational until December of 1999 at the German Weather Service (DWD)] at horizontal resolutions of 1/6° and 1/2° are evaluated. IWV retrievals from the Special Sensor Microwave Imager (SSM/I) for the Baltic Sea area and, for a second 6-month period in 1997, from a ground-based water vapor radiometer (WVR) at Potsdam are used additionally. The models reproduce the observed temporal and spatial variability of the IWV fields accurately, but a systematic humid bias of 2.5?3 kg m?2 is diagnosed versus GPS. However, radiosondes in the same area indicate a smaller bias of about 1.5?2 kg m?2 and hint at a slight dry bias in the GPS data. The standard deviation of six-hourly model time series versus GPS data is 2.8 kg m?2; the correlation is 0.92 on average. No model drift can be found within the 48-h forecast range, but, as expected, the fit to the observations decreases, for instance, because of a less accurate timing of frontal passages. Differences in results at the different model resolutions are generally negligible. The model bias slightly depends on the amount of humidity, being smaller in relative and absolute terms for higher IWV values. A comparison of the results from the models versus GPS and SSM/I retrievals shows an encouraging correspondence between the two independent data sources. The WVR also confirms the findings from the GPS time series and is in accordance with results obtained from GPS. A preliminary evaluation of the new operational DWD Lokal Model indicates a reduced bias of about 1 kg m?2 when compared with observations from German GPS sites.