Spatio-Temporal Analysis of the Pattern of Agricultural Land Use Change in Tanzania Using U-net Deep Learning Model

Keywords: Agricultural change, Land use, Deep learning, Tanzania
Abstract Type: Paper Abstract

Abstract

In Sub-Saharan Africa, cropland expansion is expected to increase in the next thirty years to meet the continent’s food needs, which calls for a large-scale agricultural land transformation. To gain insight into these changes, this study aims to analyze the spatial pattern of agricultural land use change in Tanzania over a 6-year (2017-2022) period, using advanced machine learning to map annual cropland cover in high resolution (~5 m) PlanetScope imagery combined with Sentinel-1 radar imagery over the entire country. The study will identify the areas and determine the cropland expansion rate during the analyzed time period while identifying temporal patterns in cultivation that indicate variations in cropping/fallow cycles. Ultimately, this study will provide further insights into the nature and extent of agricultural transformation, which can help raise awareness of food security and biodiversity and enable better spatial planning and agricultural policies.