GNSS observations as a numerical weather prediction data source, a way forward to enhanced forecast quality; aims, challenges and plans for 2014-2017

Abstract:

The GNSS signal propagating from the satellite to the receiver is subjected to the phase delay due to the presence of the atmosphere. The signal’s troposphere phase delay is linked with the density of all gaseous constituencies, including one of the most important - water vapour. Current GNSS processing methodology does not provide a framework for effective estimation of line-of-sight troposphere delay between satellite and receiver because of that a new functional and stochastic modelling should be introduced. Coherently, assimilation of the GNSS observations is relatively new, but very promising approach, to improve the short range forecasts (especially in terms of medium and heavy rainfall systems). With these data it is possible to provide significant amount of information about the 3D structure of the atmosphere. However, there are still many unresolved problems related to the data assimilation; such as, modelling of signal propagation (forward model) as well as correlation in time and space between GNSS observations.

This paper introduce the challenges that are going to be addressed within the course of this project: 1) The unique methodology for GNSS Slant Total Delay (STD) estimation will be developed, 2) The method to effectively assimilate the STDs into the NWP model will be investigated, 3) The impact of the GNSS data assimilation on NWP models performance will be derived for the area of Poland.

This project requires extensive GNSS signal propagation simulations to establish effective functional and stochastic models of Slant Delay. The impact of additional artefacts (ionosphere, clocks, ambiguities and multipath) on the troposphere estimates will be assessed using synthetic observations derived from numerical weather prediction model fields. This part of research is also linked with establishment of the forward operator that transforms NWP variables into the GNSS observations space. The extensive covariance and auto-covariance analysis of NWP model fields along with the GNSS troposphere estimates will help to build stochastic models in the GNSS processing software, moreover provides an error estimates for assimilation package in Weather Research and Forecasting (WRF) model. In a long run we expect this research will have large impact on the operational weather forecasting.