Forest Loss around Sacred Forests in Central Benin between 1999 and 2024: a Landsat-based Assessment of Land-cover Change and Its Proximate Drivers
KPOGNIWE Zouvèwou Joël
Applied Biosciences and Ethnobotany Unit (UBEA), Laboratory of Medical and Pharmaceutical Biotechnology, National School of Applied Biosciences and Biotechnologies (ENSBBA), National University of Science, Technology, Engineering, and Mathematics (UNSTIM), Abomey, Benin.
ASSONGBA Yédjanlognon Faustin *
Applied Biosciences and Ethnobotany Unit (UBEA), Laboratory of Medical and Pharmaceutical Biotechnology, National School of Applied Biosciences and Biotechnologies (ENSBBA), National University of Science, Technology, Engineering, and Mathematics (UNSTIM), Abomey, Benin and Applied Ecology Laboratory (LEA), University of Abomey-Calavi, Abomey, Benin.
TOBADA Pléchakov
Applied Biosciences and Ethnobotany Unit (UBEA), Laboratory of Medical and Pharmaceutical Biotechnology, National School of Applied Biosciences and Biotechnologies (ENSBBA), National University of Science, Technology, Engineering, and Mathematics (UNSTIM), Abomey, Benin.
DJEGO G. M. Julien
Applied Ecology Laboratory (LEA), University of Abomey-Calavi, Abomey, Benin.
*Author to whom correspondence should be addressed.
Abstract
Sacred forests are customary conservation sites that persist within the agricultural landscapes of West Africa, but their long-term fate under demographic, religious and economic change remains poorly quantified. This study assessed land-cover change within 1-km-radius windows centred on 15 sacred forests in three communes of central Benin (Savalou, Dassa-Zoumè and Glazoué) between 1999 and 2024. Landsat images acquired in December 1999 and November 2024 were classified into forest, cropland, built-up and water classes with a support vector machine, and class transitions were derived by cross-tabulation. Forest area across the windows declined from 21.79 to 6.97 km² (−68%), equivalent to a compound rate of −4.5% per year, and decreased significantly at the site level (Wilcoxon signed-rank test, n = 15, p < 0.001). The net gain in cropland balanced 83% of the net forest loss and the net gain in built-up area 17%, the latter concentrated in the Glazoué sites. Between 51% and 87% of the forest present in 1999 was converted to cropland (gross transition) in the 14 sites with transition data. Only one site, Kpossozoun, showed a net forest gain. Field observations documented grazing, tree felling and charcoal production as additional pressures. These estimates are derived from pixel counts of two classified maps, without area-adjusted confidence intervals, and should be considered indicative; the drivers discussed are inferred from observations and the literature and were not tested statistically. These results indicate that the landscapes surrounding sacred forests are being converted rapidly and are consistent with customary status alone being insufficient to prevent forest loss. Integrating sacred forests into local land-use planning, together with community-based monitoring of land-cover change, is needed to slow further decline.
Keywords: Sacred groves, deforestation, support vector machine, transition matrix, customary governance, sudanian zone