Improved one/multi-parameter models that consider seasonal and geographic variations for estimating weighted mean temperature in ground-based GPS meteorology

Abstract

In ground-based GPS meteorology, weighted mean temperature is the key parameter to calculate the conversion factor which will be used to map zenith wet delay to precipitable water vapor. In practical applications, we can hardly obtain the vertical profiles of meteorological parameters over the site, thus cannot use the integration method to calculate weighted mean temperature. In order to exactly calculate weighted mean temperature from a few meteorological parameters, this paper studied the relation between weighted mean temperature and surface temperature, surface water vapor pressure and surface pressure, and determined the relationship between, on the one hand, the weighted mean temperature, and, on the other hand, the surface temperature and surface water vapor pressure. Considering the seasonal and geographic variations in the relationship, we employed the trigonometry functions with an annual cycle and a semi-annual cycle to fit the residuals (seasonal and geographic variations are reflected in the residuals). Through the above work, we finally established the GTm-I model and the PTm-I model with a $$2^{\circ }\times 2.5^{\circ }(\mathrm{lat}\times \mathrm{lon})$$ resolution. Test results show that the two models both show a consistent high accuracy around the globe, which is about 1.0 K superior to the widely used Bevis weighted mean temperature–surface temperature relationship in terms of root mean square error.