Using the Multi-Temporal Landsat Data for Detecting Land Cover Change in Nampula City, Mozambique

Abstract

     The change in land cover eventually occurs as the population increases, resulting in high human activities. Monitoring this change can be beneficial for territorial planning and ecosystem monitoring. Therefore, this study aimed to evaluate the spatial-temporal patterns and rates of land cover change in Nampula City, Mozambique, over the past three decades (1989-2020). For this purpose, data from the Landsat-5 TM and Landsat-8 OLI / TIRS satellites were applied as input to two classification systems: (1) thresholding-NDVI and MNDWI and (2) supervised classification. The results showed that the supervised classification method performed better than the thresholding system, with an overall accuracy of 92.4% and a kappa coefficient of 0.89. Estimates pointed to a reduction of 0.04% in the water area and 20.3% in cultivated land. In contrast, barren rock and urban areas experienced an increase of 18.2%, while shrubs and grasslands showed a growth of 2.1% of their area. The results showed a considerable change over the study period and that the spatial dynamics of crop and barren rock and urban areas resulting from human interventions require special consideration. This study provides an opportunity for further studies on the spatial dynamics of land cover change in Nampula City, facilitating effective land management and sustainable development strategies in the region.