Remote sensing of atmospheric water vapor using GPS data in the Hong Kong Region

Abstract

GPS Meteorology has been the most active research topic recently. Remote sensing of water vapor content in the atmosphere is an important objective for the ground based GPS Meteorology. This thesis focuses on the methodology for accurately remote sensing of the Precipitable Water Vapor using Hong Kong GPS data. The precipituble water vapor is converted from the wet zenith delay in GPS signal. The estimated wet zenith delay is often affected by the azimuthal asymmetry of water vapor distribution. A horizontal gradient model is therefore developed to simulate this effect. This model is proven to be beneficial to the GPS height determination and the estimation of wet zenith delay. Determination of dry zenith delay is very important for separating wet zenith delay from total tropospheric zenith delay. Three common hydrostatic zenith delay models, namely Saastamoinen. Hopfield and Black, are calibrated for the accurate determination of dry zenith delay using Hong Kong Radiosonde data. The test results indicated that the revised models could remove a 15 millimeter systematic error in local conditions. The mapping scale factor is a bridge between wet zenith delay and precipitable water vapor. The mapping scale factor varies in season and geography, and is dominated by a weighted mean tropospheric temperature. A real-time method for the calculation of the weighted mean tropospheric temperature, which is suitable for the Hong Kong region, is developed by using the Sequential Regression Analysis method. Radiosonde data are treated as a standard to calculate the precipitable water vapor, dry zenith delay and weighted mean tropospheric temperature. Their accuracy caused by observed errors is analysed. The results show that the observed error cause 1.2 millimeter uncertainty for precipitable water vapor, 2 millimeter dry zenith delay error and 1 K uncertainty for weighted mean tropospheric temperature. One-month GPS data have been used to derive the precipitable water vapor in Hong Kong. The GPS-derived result is in good agreement with that from Radiosonde data. The actual difference of the precipitable water vapor is smaller than 2 millimeter in the two techniques. This result also indicates that the ground-based GPS remote sensing technique used in this thesis is applicable to GPS Meteorology in Hong Kong.