GNSS observations of the land uplift in South Africa: Implication for water loss estimation

Global Navigation Satellite System (GNSS) networks in South Africa indicate a spatially coherent uplift. The cause of this uplift is not clear, but one hypothesis is a crustal deformation due to mantle flow and dynamic topography (Hammond et al., 2021, JGR Solid Earth). We provide an alternative evidence based on elastic loading modelling and independent observations, suggesting that land water loss due to multiple drought periods is a dominant driver of land uplift in South Africa.

The use of continuously measuring GNSS stations has proven to be a successful method for quantifying terrestrial water mass changes, by inverting the observed vertical displacements of the Earth’s crust. Depending on the density of the GNSS network, this method has the potential to derive not only temporal but also spatial higher-resolution total water storage change (TWSC) than the Gravity and Climate Experiment (GRACE) and GRACE Follow-On (GRACE-FO) missions. Since vertical displacements in GNSS data are not only affected by water mass changes, extensive time series analyses are required to reduce or eliminate non-hydrology-related deformations, such as non-tidal oceanic and atmospheric loading. In this way, GNSS also offers an alternative method to monitor the frequently occurring droughts in South Africa, like the severe “Day Zero” drought in Cape Town from 2015-2017.

In this study, daily GNSS time series of vertical displacements (2000-present) are analysed. A long-term trend as well as annual and semi-annual signals are separated from the noisy observations using Singular Spectral Analysis (SSA). The final time series of all stations are inverted into water mass loading over a uniform grid, with the deformation properties of the Earth’s crust being defined by the Preliminary Reference Earth Model (PREM). An experimental approach shows that a 2° x 2° grid resolution of the GNSS-derived TWSC provides appropriate solutions over most of South Africa. The GNSS solution agrees with a GRACE-assimilated solution and a hydrological model at monthly scale over different provinces, with correlations up to 93% and 94%, respectively. The long-term trend averaged over the entire country is correlated with 80% and 54%, respectively. Negative long-term TWSC trends are evident in all data sets and provide compelling evidence that long-term land uplift in South Africa has a hydrological origin.