.png){kind=icon}
.png){kind=icon}
Return
A Novel Enhancement of the MIMIC Remote Sensing Water Vapor Product and Its Application in GNSS Positioning
.png){kind=icon}
OA
AI
R
N
Z
H
DOI：10.1109/JSTARS.2026.3662775.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Tropospheric delay plays a crucial role in global navigation satellite systems (GNSS) positioning, climate monitoring, and weather forecasting. In particular, in high-precision GNSS applications such as precise point positioning (PPP), prior tropospheric information can significantly accelerate convergence and improve positioning accuracy. This study derives zenith tropospheric delay (ZTD) from the Morphed Integrated Microwave Imagery at CIMSS–Total Precipitable Water Version 2 (MIMIC-TPW2) product and applies them to PPP. To improve the accuracy of MIMIC precipitable water vapor (PWV), generalized regression neural network (GRNN) and random forest (RF) models were developed with both spatial and temporal validation. Results show that RF achieves the best spatial performance, reducing the PWV root mean square error from 2.89 to 2.16 mm, while GRNN performs better temporally, improving winter PWV RMSE by up to 79.3%. The ZTDs derived from the enhanced PWV were then applied to PPP and compared with both the traditional PPP using estimated ZTD and the PPP employing GPT3-derived ZTD. The results demonstrate that the proposed method significantly improves convergence and positioning accuracy, with vertical convergence time reduced by 36.8% and positioning accuracy enhanced by 34.7% compared to traditional approaches in winter.
Keywords:
Generalized regression neural network (GRNN)
MIMIC PWV
precise point positioning (PPP)
random forest (RF)
zenith tropospheric delay (ZTD)
Journal
IF:
5.3
Papers:
1.2K
Citations:
3.0W