Using Vegetation Sensitivity to Climate Change to Predict Agriculture Climate Fragility in the Atlas Mountains Oases of Morocco

Swiss partners

• Universität Bern: Henri Rueff (main applicant)

Partners in the MENA region

• Université Hassan II - Casablanca, Morocco: Hassan Rhinane (main applicant)

Presentation of the project

The Atlas Mountains (High Atlas and Anti Atlas) of Morocco form a unique socio-ecological system where people and nature work in harmony for ecosystem services, and supporting livelihoods from farming. Climate fragility is however a reality for farmers due to rising temperatures, longer, intensified, and more frequent climate hazards such as drought or floods. Water, a scarce commodity, is likely to further decrease by up to 15% by 2050, altering agriculture production. This water stress not only puts the Atlas oases’ farmers at risk, but also affects lowland farming highly relying on mountain waters for irrigation. Policy makers need evidence-informed decision for future climate hazards response and early warning systems for improved integrated water management supporting Morocco’s “Generation Green” agricultural development strategy. Recent resolution improvements and features development in remote sensing imagery have potential to make dry spell predictive tools more robust. This research will test and validate dry spell predictive ability of a model using 12 spectral feature bands, 8 index bands, and 10 texture bands extracted from Sentinel-2 super resolution satellite imagery applied to the Atlas Mountains of Morocco. Machine learning will allow processing the data and run the predicting operation by using various regression types. The output of the model will be a yearly drought map of standard precipitation index (SPI) values, up to 5 years and at the farm scale (10m resolution). While misclassifications of land cover change in drylands can occur, ground validation and visualization can ensure statistically significant classification and help adjusting machine learning automation. This research aims at first, improving climate hazards prediction accuracy, and second at modelling the future occurrences of climate hazards in the region in order to inform farmers (rainfed and irrigated farming systems) and policy makers. Addressing water management thematically, this research will also anchor the foundations for a long-term partnership between Hassan II University and the University of Bern through the joint supervision of a Moroccan doctoral student and research grant acquisitions. Scientific visits in the respective laboratories, joint field trips, and hosting the Moroccan doctoral student at University of Bern for a month will reinforce the partnership.