A combined field and remote sensing approach to assess spatio-temporal changes in vegetation cover for erosion modelling

Land degradation in African tropical mountain environments, driven by water erosion and unsustainable land use, is further exacerbated by climate change. This leads to reduced crop productivity, and significant delivery of sediment to downstream rivers and lakes. A thorough understanding of the erosion processes and its drivers, including its spatial and temporal patterns, as well as the quantification of erosion rates is key to mitigate soil erosion, reduce sediment delivery, and ensuring agricultural and environmental sustainability. Such quantitative temporal and spatial data are required to run spatially distributed erosion models that are capable of simulating the impact of various management scenarios. However, these data are often missing for tropical mountain environments such as the southern Ethiopian highlands. Indeed, typical soil erosion models such as the RUSLE or WaTEM/SEDEM require an assessment of the cover management factor, and most applications of these model use standard tabulated values that are not region-specific and are thus not representative for the spatio-temporal vegetation developments. New developments using remote sensing, however, provide an opportunity to better parameterize the crop management factor.

Here we present spatio-temporal data on vegetation and crop types using optical remote sensing for two small catchments (Charcharo, 145ha and Zaga, 87ha) in the Gamo highlands of southern Ethiopia, with the aim to better assess the changing erosion risk and to quantify local cover management factors. We have digitalized all the crop parcels in these catchments and obtained 1538 field observations of vegetation cover for different fields and crop types throughout the year. In addition, we obtained ground-based NDVI values for 264 field parcel-crop combinations at different time intervals using a crop sensor. These field-based observations were compared with the satellite-based time series per crop to produce a better assessment of temporal changes in crop cover management. We also digitized all the soil conservation measures and started monitoring stream flow and sediment sampling using a high-resolution sampler for validating erosion model predictions.

The Charcharo catchment situated at an elevation of 2890 - 3039 meters is dominated by grazing land, with barley and potato as primary crops. Vegetation cover varies between 0% to 70-85% over the course of the growing season. Terraces are the main soil conservation method used in the area with a density of 100 m/ha. Zaga catchment is situated at an elevation of 1898 to 2166 meters and is primarily agricultural land with maize, sorghum and teff as dominant crops. Vegetation cover varies between 0% and 60-80% over a four to six month growing season. Here, stone bunds are the predominant soil conservation practice with a density of 34 m/ha. Upon completion, the study will provide quantitative information on soil erosion and the crop cover management factor in particular. This information will be used to run erosion models and to simulate the impact of climate and land use change. This research will also highlight the benefits of different conservation measures, aiding the local community and governmental stakeholders.