A regression algorithm for the retrieval of wet troposphere path delay from dual-channel water vapor radiometer measurements

Abstract The dual-channel water vapor radiometer provides a real-time tropospheric wet delay for electromagnetic signal delay calibration by measuring atmospheric radiation signals. This paper introduces the linear regression algorithm (LRA) inversion method for improving the accuracy of the zenith wet delay (ZWD) values from water vapor radiometers. Compared to the traditional fixed coefficient algorithm (FCA), which has previously been applied for ZWD retrieval, LRA can improve the accuracy of ZWD retrieval performance from water vapor radiometer measurements in different seasons. LRA has a greater capacity to describe the relationship between water vapor radiometer measurement and atmospheric refractive structure information. Compared with the ZWD obtained from radiosonde data, the root mean square error of the ZWD using LRA was 1.77 mm for the tested year. This research contains a detailed comparison of LRA and the traditional FCA using the same test datasets. The comparison demonstrates that the LRA significantly improves the ZWD accuracy, particularly in seasons with high water vapor content.