A segmented function-based model for ZTD estimation over China

Tropospheric delay is a key error source in high-precision GNSS positioning and atmospheric sensing. However, conventional models using a single functional expression fail to capture the non-uniform vertical variation of delay, especially in the lower troposphere. This study proposes a segmented vertical correction model for ZTD over China, named CZTD-P. The atmosphere is divided into four altitude intervals, and piecewise functions are fitted using ERA5 reanalysis data. Model validation against radiosonde data shows that CZTD-P significantly outperforms the widely used GPT3 model, with a 20% reduction in RMS error. The CZTD-P model demonstrates improved spatial, vertical, and seasonal accuracy, offering enhanced support for regional tropospheric delay modeling and GNSS positioning applications.