Development of a hybrid tomography model based on principal component analysis of the atmospheric dynamics and GPS tracking data

Sadeghi, Elaheh; Hossainali, Masoud Mashhadi; Safari, Abdolreza

doi:10.1007/s10291-022-01264-w

Cited by 4 publications

(1 citation statement)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hence, it has become one of the most promising techniques to reconstruct the four-dimensional distribution of atmospheric WV [6]. Plenty of Global Navigation Satellite System (GNSS) tropospheric WV tomography (GNSS tomography) studies and experiments have been carried out in different regions [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25]. A GNSS tomographic domain covering the ground-based GNSS receiver network is usually cuboid-shaped or box-shaped within the troposphere, and it is first needed to discretize into many tomographic voxels in the process of tomographic solving.…”

Section: Introductionmentioning

confidence: 99%

Development of a New Vertical Water Vapor Model for GNSS Water Vapor Tomography

Wan

Zhang

et al. 2022

Remote Sensing

View full text Add to dashboard Cite

One of the main challenges of Global Navigation Satellite System (GNSS) tomography is in solving ill-conditioned system equations. Vertical constraint models are typically used in the solution procedure and play an important role in the quality of the GNSS tomography, in addition to helping resolve ill-posed problems in system equations. In this study, based on a water vapor (WV) parameter, namely IRPWV, a new vertical constraint model with six sets of coefficients for six different WV states was developed and tested throughout 2019 in the Hong Kong region with four tomographic schemes, which were carried out with the model and the traditional vertical constraint model using three different types of water vapor scale height parameters. Experimental results were numerically compared against their corresponding radiosonde-derived WV values. Compared with the tests that used the traditional model, our results showed that, first, for the daily relative error of WV density (WVD) less than 30%, the new model can lead to at least 10% and 49% improvement on average at the lower layers (below 3 km, except for the ground surface) and the upper layers (about 5–10 km), respectively. Second, the skill score of the monthly root-mean-square error (RMSE) of layered WVD above 10 accounted for about 83%, 87%, and 64%. Third, for the annual biases of layered WVD, the new model significantly decreased by 1.1–1.5 g/m3 at layers 2–3 (about 1 km), where all schemes showed the maximal bias value. Finally, for the annual RMSE of layered WVD, the new model at the lower (about 0.6–3 km) and upper layers improved by 13–42% and 5–47%, respectively. Overall, the new model performed better on GNSS tomography and significantly improved the accuracy of GNSS tomographic results, compared to the traditional model.

show abstract