Assessing the driving forces of Guinea savanna transition using geospatial technology and machine learning in Old Oyo National Park, Nigeria

Abstract

Savannization and de-savannization are two land cover transition processes that are yet to be understood at a local scale in sub-Saharan Africa. We studied the patterns across different ecological, anthropogenic and climatic factors to infer the trends of savannization and de-savannization in Old Oyo National Park, Nigeria. Geospatial technology and two machine learning methods were used to determine the pattern, dynamics, and importance of savannization and de-savannization with a set of thirteen eco-climatic variables. The savannized areas with total land cover masses of 934.27 km² in 1986–2003 decreased to 901.01 km² in 2003–2019, and vice versa for de-savannized areas. The thirteen eco-climatic variables contributed to savannization and de-savannization within the two transition intervals in varying degrees. Relevant stakeholders should employ mild burn severity to manage savannization at locations close to host communities, rivers, and roads within elevations above 450 m and low rainfall in the dry season below 6 mm.

Keywords:

• Climate change
• savannization
• de-savannization
• anthropogenic
• computational intelligence

Acknowledgements

We appreciate the Nigerian National Park Service for issuing the entry permits into the study area. We also like to thank Messers Peter Ajiibi and Umar Yusuf Tanko for their tremendous contributions throughout the data collection stage.

Disclosure statement

The authors declare that they have no competing interests.