Abstract
GPS tomography has been investigated since 2000 as an attractive tool for retrieving the 3D field of water vapour and wet refractivity. However, this observational technique still remains a challenging task that requires improvement of its methodology. This was the purpose of this study, and for this, GPS data from the Australian Continuously Operating Research Station (CORS) network during a severe weather event were used. Sensitivity tests and statistical cross-comparisons of tomography retrievals with independent observations from radiosonde and radio-occultation profiles showed improved results using the presented methodology. The initial conditions, which were associated with different time-convergence of tomography inversion, play a critical role in GPS tomography. The best strategy can reduce the normalised root mean square (RMS) of the tomography solution by more than 3 with respect to radiosonde estimates. Data stacking and pseudo-slant observations can also significantly improve tomography retrievals with respect to non-stacked solutions. A normalised RMS improvement up to 17% in the 0–8 km layer was found by using 30 min data stacking, and RMS values were divided by 5 for all the layers by using pseudo-observations. This result was due to a better geometrical distribution of mid- and low-tropospheric parts (a 30% coverage improvement). Our study of the impact of the uncertainty of GPS observations shows that there is an interest in evaluating tomography retrievals in comparison to independent external measurements and in estimating simultaneously the quality of weather forecasts. Finally, a comparison of multi-model tomography with numerical weather prediction shows the relevant use of tomography retrievals to improving the understanding of such severe weather conditions.
Highlights
Global Positioning System (GPS) tomography has been investigated since 2000 as an attractive tool for retrieving the tomography
Based on the radiosonde data, the following calculations were made to retrieve wet refractivity and water vapour content: (1) the formula used by Sargent [47] was applied to obtain the relative humidity from the dew point temperature [48]; (2) the formula used by Sonntag [49] was used to calculate the partial pressure of water vapour pw from RH and T; and (3) the atmospheric state parameters pw, p, and T were used to calculate wet refractivity Nw and water density profiles ρwv using equations proposed by Davis et al, [50], i.e., N = k1 Rd ρ + k0 2 pw pw
A sensitivity test of the perturbation of the quality of slant observations, inputs of GNSS tomography, has been presented and was combined with the two others improvements to investigate the interest in this approach in meteorological applications
Summary
Global Positioning System (GPS) tomography considers the use of slant-integrated estimates, wet delays, or corresponding water vapour content estimated from the data records of ground-based GPS stations to respectively retrieve the 3D field of wet refractivity or water vapour density, as introduced by [1,2]. Comparisons of tomography retrievals with other techniques (e.g., those using a water vapour radiometer, radiosonde, raman lidar, or atmospheric emitted radiance interferometer) and with numerical weather models have shown relevant results and an encouraging understanding of meteorological situations ([3,4,5,6,7,8,9,10,11,12,13,14,15,16,17]) The resolution and configuration/geometry of the network of GPS stations are critical parameters with which to obtain the best scenario for applying GPS tomography to retrieve water vapour density or wet refractivity. The last section summarises and highlights conclusions and perspectives for future works
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
