4D GPS water vapor tomography: new parameterized approaches

Abstract

Water vapor is a key variable in numerical weather prediction, as it plays an important role in atmospheric processes. Nonetheless, the distribution of water vapor in the atmosphere is observed with a coarse resolution in time and space compared to the resolution of numerical weather models. GPS water vapor tomography is one of the promising methods to improve the resolution of water vapor measurements. This paper presents new parameterized approaches for the determination of water vapor distribution in the troposphere by GPS. We present the methods and give first results validating the approaches. The parameterization of voxels (volumetric pixels) by trilinear and spline functions in ellipsoidal coordinates are introduced in this study. The evolution in time of the refractivity field is modeled by a Kalman filter with a temporal resolution of 30 s, which corresponds to the available GPS-data rate. The algorithms are tested with simulated and with real data from more than 40 permanent GPS receiver stations in Switzerland and adjoining regions covering alpine areas. The investigations show the potential of the new parameterized approaches to yield superior results compared to the non parametric classical one. The accuracy of the tomographic result is quantified by the inter-quartile range (IQR), which is decreased by 10–20% with the new approaches. Further, parameterized voxel solutions have a substantially smaller maximal error than the non parameterized ones. Simulations show a limited ability to resolve vertical structures above the top station of the network with GPS tomography.