Evaluation of Agroforestry Suitability in Tazekka National Park: A geospatial approach based on Google Earth Engine

Abstract. Climate change is having a major impact on the world's forests, compromising their health and resilience. Agroforestry, which consists of integrating trees into agricultural systems, appears to be a solution for strengthening this resilience. This study assesses the suitability of agroforestry in Tazekka National Park, Morocco, using its ecological features to explore the contribution of agroforestry to reducing the effects of climate change. Using Google Earth Engine (GEE), the Land Use and Land Cover (LULC) classification was generated by comparing two major machine learning algorithms: Support Vector Machine (SVM) and Random Forest (RF). The most accurate LULC classification, as determined by the algorithms, was integrated with various environmental variables such as rainfall, temperature, soil pH, soil texture, slope, vegetation indices (NDVI, NDWI), population density, erosion risk, and tree cover. These factors were incorporated into the analysis using Multi-Criteria Analysis (MCA) to calculate and generate a suitability index for agroforestry. The Analytical Hierarchy Process (AHP) was employed to assign weights to the variables based on their relative importance. The results of the LULC classification revealed that SVM outperformed RF, achieving an accuracy of 95.05% compared to 85.15% for RF. The final results produced a more accurate suitability map for agroforestry, effectively identifying areas with low, medium, and high suitability for agroforestry interventions. This research highlights the potential of agroforestry at the local level and proposes a strategic framework for sustainable land management in the face of climate change at regional and global levels.