Approche cartographique par le SIG GRASS pour l'analyse de la structure du paysage au Libéria, Afrique de l'Ouest

Algorithms of GIS are increasingly used for automatic extraction of landscape characteristics due to development of advanced methods of geospatial data analysis. This paper explores the potential of GRASS GIS for analysis of geometry of landscape patches using calculation of raster data. The data were obtained from the classified satellite images covering Liberia, West Africa, for the period of 2014 and 2023. Landscape dynamics was analysed in changes in six computed patch indices indicating deforestation in coastal Liberia: patch density index, shape index, patch number index, standard deviation, coefficient of variation and patch range. The numerical analysis was technically performed using scripting methods of GRASS GIS by the following key modules: r.li.patchdensity, r.li.shape, r.li.patchnum, r.li.padsd, r.li.padcv and r.li.padrange. Based on the analysis of patch sizes in forest areas for the estimated periods, the shape index increased from 2.86 in 2014 to 4.09 in 2023, indicating an increase in the sum of landscape edge lengths. This means that the divergence of the shape of a landscape patch becomes higher during this period which indicates increased landscape fragmentation. The study also revealed the increased curvature of the landscape area and separability of individual elements indicate forest fragmentation processes. In contrast to this phenomenon, lower values ​​of the patch density index decreased from the average of 1.68 to 1.15, indicating a decrease in the dense forest class over this period. In addition, the forest area was reduced by 12%, suggesting an average annual deforestation rate of 0.9% in Liberia.