Troposphere Modeling in a Regional GPS Network

Abstract

By using a regional network of Global Positioning System (GPS) reference stations, it is possible to recover estimates of the slant wet delay (SWD) to all GPS satellites in view. SWD observations can then be used to model the vertical and horizontal structure of water vapor over a local area, using a tomographic approach. The University of Calgary currently operates a regional GPS real-time network of 14 sites in southern Alberta. This network provides an excellent opportunity to study severe weather conditions (e.g. thunderstorms, hail, and tornados) which develop in the foothills of the Rockies near Calgary. In this paper, a 4-D tomographic water vapor model is tested using the regional GPS network. A field campaign was conducted during July 2003 to derive an extensive set of truth data from radiosonde soundings. Accuracies of tomographic water vapor retrieval techniques are evaluated for 1) using only ground-based GPS input, and 2) using a ground-based GPS solution augmented with vertical wet refractivity profiles derived from radiosondes released within the GPS network. Zenith wet delays (ZWD) are computed for both cases, by integrating through the 4-D tomography predictions, and these values are compared with truth ZWD derived from independent radiosonde measurements. Results indicate that ZWD may be modeled with accuracies at the sub-centimeter level using a ground-based GPS network augmented with vertical profile information. This represents an improvement over the GPS-only approach.