Impact of GPS-based water vapor fields on mesoscale model forecasts

Ground-based GPS networks, supplemented by in situ and remote sensing instruments, can observe the four dimensional distribution of atmospheric water vapor with unprecedented spatial and temporal resolution. Such data have the potential to improve significantly operational numerical weather prediction. Determination of the optimal configuration of such observing systems for short-range predictions of convective initiation and rainfall requires assimilation and testing by comprehensive analysis and forecasting tools.

Radiometrics, Inc., ENSCO Inc., UCAR, and the Forecast Systems Laboratory (FSL) are conducting a data assimilation experiment using slant delays sensed by a Global Positioning System (GPS) network along with Doppler Radar Wind Profiler (DRWP) and microwave radiometer data. The GPS slant water sensor operates by measuring the delay in the GPS signal caused by atmospheric integrated water vapor. Through the determination of the delays for 8 to 12 satellites in view as each track across the sky, the GPS sensor obtains a large number of slant path measurements of integrated water vapor. When such sensors are deployed in arrays, three-dimensional water vapor distributions can be obtained by variational methods.

This paper will focus on the assimilation of GPS-based measurements of atmospheric water vapor and the influence of water vapor profiles on short-range forecasts of moisture and convective initiation over the Southern Plains. The data assimilation is conducted using the Advanced Regional Prediction System (ARPS) Data Analysis System (ADAS) from the University of Oklahoma. Short-range forecasts are generated using the ARPS mesoscale model and predictions are validated against available observational data.