Methodologies in GPS Data Processing for Environmental Sciences: Contributions to the Study of Monsoon in West Africa

Abstract

Global Positioning System (GPS) is now very useful for meteorology and environmental sciences. This thesis focuses on the study of the water cycle (atmospheric and continental) in West Africa, as part of AMMA (African Monsoon Multidisciplinary Analyses) and GHYRAF (Gravity and Hydrology in Africa) projects.In the first part, we analyze the precision of GPS solutions in Africa. The integrated water vapor (IWV) is especially important in understanding the key atmospheric processes of the monsoon from subdiurne to multiyear periods. We pay much attention to sources of errors and the strategy of GPS data processing. A recall of the main theoretical elements allows us to identify the major sources of error that may affect the GPS estimates. We quantify the sensitivity of IWV for each of these sources of error and obtain a long-term precision of 1 kg m−2 on IWV. Meanwhile, station heights are really sensitive to errors and choices of parameterization. They are mainly subject to bias between 2 and 6 mm (choice of antenna models and the cut-off angle) and to seasonal signals whose amplitude is lesser than 2 mm (choice of tropospheric modeling). In the second part, we focus on estimates of station positions in order to quantify crustal deformation caused by continental hydrology. We compare the GPS solutions to estimates calculated from geophysical data and from GRACE space gravimetric data. Through this study, we also evaluate the quality of representation of the seasonal variation of soil moisture in West Africa with hydrological models and GRACE products. The three data sets are rather matching with the annual range of vertical deformations. An additional GPS signal is however detected, on the heights of AMMA GPS stations which appears to be an oscillation occurring between September and March with maximum amplitude in Ouagadougou evaluated between 12 and 16 mm. The hypothesis of an artifact GPS is rejected given previous sensitivity tests. The additional