Correlating optical, microwave and thermal remote sensing signals with groundwater head measurement time series

Due to global climate change, population growth and an ever increasing demand for fresh water, monitoring groundwater changes over large areas becomes very important. Current groundwater monitoring still heavily depends on groundwater head data that are often not available for large parts of the non-western world. Consequently, groundwater assessment, especially for large areas comprising several aquifers and basins, is very complex and not very accurate. The main research question of this study is to check whether optical, microwave and thermal remote sensing signals carry information on groundwater levels. The advantage of remote sensing data is that it provides synoptic overview of semi-quantitative maps of surface properties, such as vegetation, surface temperature, and soil moisture signals, where we seek the relation with shallow groundwater tables. In this study, we analyze and compare groundwater head measurement data and various remote sensing signals. Our datasets comprise over 5000 groundwater head measurement points in the Rhine-Meuse basin. These groundwater head data are compared to the signals of soil moisture (AMSR-E and ERS), thermal (MODIS land surface temperature) and vegetation (MODIS NDVI, LAI and FPAR). We investigated the correlation among these variables by performing straightforward statistical analyses, such as scatter-plots, multiple regression techniques and analyses of temporal variability between the variables. Results show that particularly the soil moisture demonstrates significant relations with groundwater depth, specifically for shallow water table depths and areas with sparse vegetation. We argue that remote sensing signals should be considered as important information for groundwater assessment in data-poor environments.