Code and reproducibility materials for "Mapping and projecting land-use and land-cover change in semi-arid Tanzania: Implications for potential Loss and Damage

Abstract

This repository contains reproducibility materials associated with the study “Mapping and projecting land-use and land-cover change in semi-arid Tanzania: Implications for potential Loss and Damage.” The materials include Google Earth Engine scripts for Landsat preprocessing, spectral-index calculation and Random Forest land-use and land-cover classification; R scripts for land-cover area summaries, temporal change analysis, transition matrices, accuracy assessment, population co-location and household-pressure summaries; and documentation of the Artificial Neural Network–Cellular Automata scenario-modelling workflow. Historical land-use and land-cover classifications cover 1994, 2009 and 2024, with a conditional scenario projection for 2039 in Simanjiro District, Tanzania.