Mapping smallholder cashew plantations to inform sustainable tree crop expansion in Benin

Cashews are grown by over 3 million smallholders in more than 40 countries worldwide as a principal source of income. As the third largest cashew producer Africa, Benin has nearly 200,000 smallholder growers contributing 15% country's national export earnings. However, lack information on where and how trees grow across country hinders decision-making that could support increased production poverty alleviation. By leveraging 2.4-m Planet Basemaps 0.5-m aerial imagery, newly developed deep learning algorithms, large-scale ground truth datasets, we successfully produced first map characterized expansion plantations between 2015 2021. In particular, SpatioTemporal Classification with Attention (STCA) model to distribution plantations, which can fully capture texture from discriminative time steps during growing season. We further Clustering Augmented Self-supervised Temporal (CASTC) distinguish high-density versus low-density automatic feature extraction optimized clustering. Results show STCA an overall accuracy 85% CASTC achieved 76%. found area almost doubled 2021 60% new plantation development coming cropland or fallow land, while encroachment into protected areas 55%. Only half were 2021, suggesting high potential for intensification. Our study illustrates power combining high-resolution remote sensing imagery state-of-the-art algorithms better understand tree crops heterogeneous landscape.