Remote Sensing Solutions to Assess Land Degradation and to Enhance Smallholder Farming in West Africa

Abstract

Africa is facing different challenges in the 21. century. One of them is the changing climate and its impact on managed and not managed land. Land degradation and desertification increases the pressure on agriculture land and droughts threaten traditional rain-fed agriculture. Extreme weather events like heavy rainfall and storms further jeopardize agricultural stability, heightening the risk of crop failures within an already vulnerable system. Remote sensing has a high potential to monitor impacts like land degradation. Therefore, we are utilizing Sentinel-1A/B C-Band VV/VH-SAR time series to detect surface changes, based on the recognition of irregularities in the radar backscatter or coherence dynamics. The results from Sentinel-1 and -2 timeseries regarding woody cover mapping and breakpoint analyses of the complex savanna systems shows the high potential of quantifying land degradation using Sentinel-1 and Sentinel- 2 data. Remote sensing can also support smallholder farmers in finding suitable adaptation practices against the negative impacts of climate change. These measures vary in complexity, feasibility, and cost, ranging from changes in land management and crop selection to technological innovations such as weather forecasting and irrigation systems. The evaluation of those implemented adaptation practices is crucial for a transfer to other regions. Remote sensing-based analyses, using Sentinel-1, Sentinel-2, and Planet data are employed to monitor and evaluate adaptation measures. Therefore, we applied different methodologies like random forest regression, convolutional neural networks (CNN), fuzzy logic approaches, and time series analysis. The results can serve as a basis for local decision makers and stakeholders supporting the implementation of effective adaption strategies to increase resilience and promote sustainable agricultural practices in the face of a changing climate.