That’s how we looked at bare ground dynamics in time and space. However, the flip side of the medal is to look at how vegetation dynamics behaved in the same area. To do so, we took the MODIS calibrated time series to create the second set of maps. We estimated vegetation indices such as NDVI, EVI, and NPP for each growing season between 2000 and 2020, and checked whether they showed a statistically significant trend change.
We spotted some areas in the landscape with a decreasing trend as well as areas with an increasing trend. This second analysis picks up particularly well changes in land cover such as farmland conversion, expanding settlements, and shifting vegetation due to meandering rivers (see Map below).
The mapping was followed by field validation. In July 2021, we explored most parts of SORALO’s landscape with the printed maps in our hands and discussions with key informants and other community members. We noticed that field observations matched our mapping outputs.
A few interesting trends were detected:
A common feature in the area between lake Magadi and the bottom of the rift valley is the presence of parallel narrow valleys separated by rocky ridges. These valleys are used seasonally by the community as livestock corridors (see figure below) to move between one area and the next one. They show a higher bare ground frequency increase than other places (see Figure 4).
Rangeland health is an elusive concept that various actors can perceive differently. Although solely focused on vegetation cover and bare ground occurrence, our tool showed a good match with both field observations and community inputs. The tool worked well to identify potentially critical areas, facilitate discussions on what was happening where and why, and ultimately identify some priority areas for restoration. The same tool can be adapted to cover other regions where extensive livestock rearing is the primary livelihood strategy.