Modelling of land use land cover changes using machine learning and GIS techniques: a case study in El-Fayoum Governorate, Egypt

Abstract

Land use/land cover (LULC) changes can occur naturally or due to human activities. In this study, the maximum likelihood algorithm (MLH) and machine learning (random forest algorithm (RF) and support vector machine (SVM)) were investigated for image classification to oversight spatio-temporal land use changes in El-Fayoum governorate, Egypt. The Google Earth Engine has been utilized to pre-process the Landsat imagery, and then upload it for classification. Each classification method was evaluated using field observations and high-resolution Google Earth imagery. LULC changes were assessed, utilizing Geographic Information System (GIS) techniques, over the last 20 years in three different periods: 2000–2012, 2012–2016, and 2016–2020. The results showed that socioeconomic changes occurred during these transitions. The SVM procedure provided the most accurate maps in terms of the kappa coefficient (0.916) compared to MLH (0.878) and RF (0.909) procedures. Therefore, the SVM technique was adopted to classify all available satellite imagery. The results of change detection showed that urban sprawl has occurred and most of the encroachments were on agricultural land. The results showed that agricultural land area decreased from 26.84% in 2000 to 26.61% in 2020 and urban area increased from 3.43% in 2000 to 5.99% in 2020. In addition, urban land expanded rapidly on account of agricultural lands by a total of 4.78% from 2012 to 2016, while it expanded slowly by a total of 3.23% from 2016 to 2020. Overall, this study offers useful insight into LULC changes that might aid shareholders and decision makers in making informed decisions.