Regional spatio-temporal zenith tropospheric delay modelling using independent component analysis

Accurate modelling of the zenith tropospheric delay (ZTD) is of significant value in many applications such as Global Navigation Satellite System (GNSS) positioning. Most ZTD models require meteorological parameters, which are difficult to obtain in many practical applications. Though some empirical models can predict the ZTD values without meteorological parameters, their performances are usually unsatisfactory. To overcome these issues, a spatio-temporal ZTD modelling method based on independent component analysis (ICA) without using meteorological parameters is proposed in this study. In this method, multi-dimensional ZTD data can be reconstructed by several temporal independent components and corresponding spatial response values. Regional ICA ZTD models for Hong Kong, UK and China mainland were established using the daily GNSS ZTD data. Results show that ICA regional ZTD models perform better than Saastamoinen, UNB3m and GPT2w models in general. The proposed ICA ZTD modelling method is proven to have advantages of simpler modelling process, fewer model parameters and higher applicability. This study suggests that ICA is very promising in modelling spatio-temporal variation of ZTD, especially for a region with dense GNSS network.