Quantifying riparian total evaporation along the Groot Letaba River: A comparison between infilled and spatially downscaled satellite derived total evaporation estimates

- The SEBS model was applied to derive ET using Landsat and MODIS data.
- Two approaches were then applied to estimate daily ET at a moderate spatial scale.
- These estimates were evaluated against eddy covariance ET measurements.
- Satellite-derived ET and measured ET was shown to be in poor agreement.
- This was due to a limitation when applying SEBS for water stressed conditions.

The use of satellite earth observation data for the estimation of evapotranspiration has been well documented and represents a viable approach for the quantification of riparian water use at landscape to regional scales. However, the trade-off between the spatial and temporal resolution associated with imagery can limit the reliability of satellite-based evapotranspiration modelling. This study investigated two approaches to quantify evapotranspiration at a moderate spatial resolution (30 m) on a daily time step, for a perennial river flowing through a semi-arid, savanna landscape. The Surface Energy Balance System (SEBS) Model was used to derive daily evapotranspiration from satellite imagery. The actual crop coefficient (Kcact) and output downscaling with linear regression (ODLR) approaches were then evaluated by comparing their respective estimates against Eddy covariance (ECET) measurements at two locations. Comparisons of ET estimates acquired using the Kcact and ODLR approaches against ECET, yielded RMSE values of; 1.88 and 2.57 mm d−1 and 1.10 and 2.39 mm d−1 (for two replicate transects), respectively. The poor performance of these techniques was largely attributed to the SEBS ET estimates used as inputs to these techniques, as SEBS may overestimate evapotranspiration during conditions of water stress.