Burkina Faso's W National Park (WNPB) is one of the country's biodiversity hotspots. Despite this Park's protection, illegal human activities threaten its vegetation, and little information about the on-site causes of degradation is available. Here we explored the extent and drivers of changes in land use/land cover (LULC) in the WNPB. For this purpose, we used Landsat (1988, 2001 and 2016), and ancillary data to determine the spatial structure and changes in LULC with a random forest classifier. We identified socio-economic factors influencing LULC by conducting household surveys, focus groups, and field observations. We used the ranking and binary logistic regression methods to analyze the socio-economic data. A logistic regression model assessed the relationship between land cover change and socio-economic factors. The results showed that multi-temporal classification performed significantly better than mono-temporal one in the study area. However, the combination of mono-temporal reflectances and vegetation indices for LULC classification significantly enhanced the accuracy to the level of multi-temporal classification. This result indicates that this combination is an efficient alternative to multi-temporal classification in the study area where cloud-free imagery is rare. In the WNPB, LULC dynamics from 1988 to 2001 was mainly characterized by reduction of tree savannas (9.4%) and woodland (8.8%) and expansion of bare soil (0.1%) and mixed vegetation (18.1%). From 2001 to 2016, croplands increased by 102.1%. Bush fire (77.19%), wood harvesting (48.25%), lack of rain (48.25%) and elephants’ impact (20.18%) were the leading direct causes of LULC change. In general, the perceptions of local populations about vegetation trends were in line with the observations from remote sensing. Altogether, the tree vegetation in the study area decreased significantly between 2001 and 2016, and the cause of this change appears to be unsustainable land-use rather than rainfall conditions. According to these findings, land-use management practices in this park should be more sustainable and provide education (and changing habits) to the surrounding population. It is therefore suggested that the government regulate illegal activities through forest law enforcement to protect the WNPB from over-exploitation.

Biodiversity, Change detection, GIS, Protected areas, Random forest, Remote sensing, SDGs, Vegetation, West Africa