Using Landsat satellites to assess the impact of check dams built across erosive gullies on vegetation rehabilitation

Gully erosion, a process of soil removal due to water accumulation and runoff, is a worldwide problem affecting agricultural lands. Building check dams perpendicular to the flow direction is one of the suggested control practices to stabilize this process. Though there are many studies on the effect of erosive controls on land stabilization, few examine its effect on the rehabilitation of vegetation. Here we use information from the satellites Landsat-7 (1999–2018) and Landsat-8 (2013–2018) to assess the effect of soil check dams built during 2012 across three gullies with distinct structures in a dryland area on vegetative cover and water status. We use a time series analysis technique to decompose Landsat-derived soil adjusted vegetation index (SAVI) into woody (SAVIW) and herbaceous (iSAVIH) contributions. The integral over the seasonal signal of the normalized difference water index (iNDWI) was used to assess changes in water status in the gully. We used herbaceous biomass collected in the field in 2014–2017 to validate iSAVIH as a proxy of herbaceous biomass. Our results show that following the construction of the check dams, the change in woody vegetation cover is best described by a sigmoid model with an increase of ~57% (95% CI: 39%–76%; p < 0.0001), while the herbaceous vegetation increases linearly at a rate of ~71% per year (95% CI: 48%–93% y⁻¹; p < 0.0001). The correlation between iSAVIH and herbaceous biomass (R² = 0.56; n = 16; p < 0.001) corroborates this increase. We found higher herbaceous productivity in the deeper gully compared to the shallower gullies but not statistically different increase rates. An increase in iNDWI of ~68% (95% CI: 43%–95%; p < 0.0001) likely implies an improved water infiltration rate that favored the vegetation expansion. Our satellite-based approach can be used to assess the impact of erosive control practices on vegetation rehabilitation in heterogeneous gullies.