Monitoring surface deformation associated with groundwater abstraction in the Western Cape using SAR Interferometry

Abstract

Aquifers are able to contain significant amounts of water, which is typically stored in the pore spaces between sand particles or in the fissures and matrix of rocks. This groundwater is an increasingly important source of water in many parts of South Africa, particularly along the arid south-west coast where the relatively deep (<150m) Elandsfontein and Langebaan road aquifers are found. Groundwater is extracted here for agriculture, industry and domestic use. Recent borehole data shows that increased abstraction combined with the low rainfall experienced between 2015 and 2018 is resulting in reduced groundwater levels in the area. Groundwater abstraction may lead to ground subsidence due to pressure reduction and can lead to permanent storage potential loss due to compaction. Ground surface monitoring is therefore valuable for sustainable aquifer management, preventing damage to infrastructure and providing input to geo-hydrological models. Such monitoring can be done with differential-GPS or levelling surveys but these are labour-intensive, expensive and do not provide comprehensive spatial coverage of an aquifer. Synthetic Aperture Radar (SAR) is an active remote sensing system capable of imaging the earth in the microwave spectrum and capturing both the intensity and phase of the returning signal. Comparing the phase of the same surface over time with a technique called differential interferometry (DInSAR) allows spatial monitoring of earth surface movements to fractions of a wavelength (mm to cm). This work investigates DInSAR monitoring of the south-west coast using Sentinel-1 C-band data in an attempt to detect small-scale ground subsidence over large areas. Early results show that Sentinel-1 is an appropriate data source. However, limitations inherent to conventional DInSAR techniques were encountered namely: 1) the dynamic land-cover of the study area, 2) the slow deformation rate and 3) atmospheric noise similar in scale to the expected deformation. Advanced time-series approaches are recommended for effective monitoring of aquifer related deformation in the area.