GPS Meteorology with Single Frequency Receivers
- Left: BfS ODL monitoring network with more than 1800 stations; Right: probe of a monitoring station
- © BfS
Zhiguo Deng (Successful
finalization June 2012)
Atmospheric water vapor plays a significant role in atmospheric convection and in the development of clouds and precipitation. As one of the key parameters for modern weather prediction, the atmospheric water vapor has high temporal and spatial variability. The lack of observations of the atmospheric water vapors in space and time limits the accuracy of short-term weather forecasts. Therefore, the spatial and temporal resolution of the atmospheric water vapor observations needs to be improved. Using the Global Positioning System (GPS) in meteorology provides a unique opportunity for this need. Radio signals emitted by the GPS satellites are bent and delayed depending on the temperature, pressure and water vapor. Based on the tropospheric delay, the water vapor distribution within the troposphere can be determined. However, to detect the water vapor distribution with a resolution of kilometer scale in horizontal and sub-kilometer scale in vertical direction, the existing GPS networks must be densified. Due to economic reasons, this densification is recommended with single frequency (SF) receivers. For normal dual-frequency (DF) GPS receivers the observations of the second frequency L2 can be used to eliminate the ionospheric delay by forming a linear combination with the observations of the first frequency L1. In the SF data processing a different ionospheric delay handling is required.
In this thesis it is shown that the epoch-differenced ionospheric delay correction is sufficient for estimating the tropospheric delay, e.g., the Zenith Total Delay (ZTD), from SF GPS data. Based on this result, the Satellite-specific Epoch-differenced Ionospheric Delay model (SEID) was developed. In the SEID model the ionospheric corrections for SF data are generated from the observations of surrounding reference stations equipped with DF receivers. With the derived ionospheric corrections and the SF data, pseudo L2 data are generated, which can be processed using existing GPS processing software packages without any changes.
In order to evaluate the performance of the SEID model, 24 simulated densification scenarios with different reference station densities and varying numbers of reference stations were defined and investigated. The validations showed very promising results: for densification scenarios with mean distances of the SF station to reference stations below 80 km, the ZTD accuracy of the SF receivers is comparable with those of the DF receivers. The study shows that the ZTD reliability of the SF data is improved with decreased reference station distance and increased number of reference stations.
The approach is validated with data from a very dense GPS network with mixed SF and DF receivers in Germany. The ZTDs derived from the SF and DF data were compared. Their differences in Root Mean Square (RMS) are about 3 mm which is negligible compared to the differences due to processing with various state-of-the-art software packages of about 7 mm.
To assess the possibility of densifying an existing GPS network with low-cost SF GPS receivers, an evaluation study was carried out. Observations from 258 German DF GPS
stations are treated as observations from SF GPS stations, i.e., only L1 GPS observations are used. ZTD, Slant Total Delay (STD) and Slant Water Vapor (SWV) products, derived from the SF data using the SEID model, are validated using tropospheric products derived from DF data, a Water Vapor Radiometer (WVR) and European Centre for Medium-Range Weather Forecasts (ECMWF) analyses. The three validation studies show that the ZTD, STD and SWV products obtained from SF data are almost of the same high-quality as those from the DF data. Compared to the tropospheric products from the DF data the ZTD from the SF data have an accuracy of 3 mm in RMS, and the relative accuracy of SF STDs is almost constant for all elevation angles and equals ~ 0.18%, which is not degrading with decreasing elevation angles. The SWV between GPS and WVR agree equally well; the standard deviation increases almost linearly from 1.3 kg∙m-2 near the zenith to about 2 kg∙m -2 at 20° elevation. The quality of the tropospheric products derived from SF data is fully adequate for atmosphere sounding. The easy implementation and the accuracy of the SEID model can speed up the densification of existing networks with SF receivers.