Remote Sensing for Analyzing Forested Landscape Structure and Land-Use Histories in Guyana’s Bauxite Mining Landscapes

Abstract

Monitoring secondary forest regrowth is a priority in forest restoration strategies. A site history helps in understanding the present status of natural regeneration in the three landscapes impacted by bauxite mining. Nonetheless, high rainfall in bauxite residue storage areas can facilitate natural regeneration of forest. This research analyzed the natural regeneration of forest after thirty years of different land use histories at three bauxite mining areas of the Upper Demerara—Berbice region of Guyana. There are no man made forest plantations in the three landscapes being reviewed. The methodology included: 1) the selection of three sampling landscapes with different land use histories 2) the generation Land Use/Land Cover maps using KMeans unsupervised classification of satellite images in each landscape and 3) the assessment of landscape structure of the land cover classes for year 2020 at class and landscape level using landscape metrics. The assessment of landscape structure of the land cover classes was carried-out with landscape metrics for the comparisons at class and landscape level. Principal component analysis enables the identification of main patterns among landscape-level metrics and land cover classes. Discriminant classification of the landscape classes was analyzed with the different metrics. The results suggest that Normalized Difference Vegetation Index and KMeans unsupervised classification can be used to evaluate the difference in natural forest regeneration among landscapes with differing land use histories. The landscape metrics revealed secondary stages of forest succession. The Landscape Shape Index and Edge Density were the most significant for landscape differentiation. The result of the various land uses reveals a mosaic of early, intermediate, and late successional sequences.