Temporal detection and prediction of agricultural land consumption by urbanization using remote sensing

Abstract

The most fertile land in Egypt is under continuous loss by urbanization. In some countries, this could be considered normal, but in Egypt, it is a severe violation of the country’s limited agricultural land. The main objectives of this research are to quantify the agricultural land consumption by urbanization and predict the future pattern of this process in some parts of Alexandria and El-Behiera governorates using the Landsat satellite images of 1987, 2001, 2015, and 2019. Image classification was performed using the Maximum likelihood supervised classification technique. Five land use/cover (LULC) types have been identified, which are agriculture, urban, gardens, the international road, and water. Classification results were further enhanced by employing post-classification and filtering processes which increased classification accuracy and minimized urban land use overestimation. Four LULC maps have been developed for 1987, 2001, 2015 and 2019. A post-classification comparison method was applied to detect the LULC changes among the independently classified images and quantify the loss of agricultural land by urbanization. Prediction of urban encroachment to agricultural land was performed by Land Change Modeler basing on both the past change trend of urbanization and the influencing distance-to-urban factor. The future LULC prediction maps for 2019, 2050, and 2100 have been generated. Prediction validity was applied using actual vs predicted LULC 2019 maps. The results indicated that the agricultural and urban categories are the prevailing LULC types in the investigated area. Over the last three decades significant loss in the agricultural land was observed together with remarkable growth in the urban area. The agricultural land was decreased by 11.03%, while the urban area was increased by 11.84%. The rate of agricultural land consumption by urbanization after 2001 was more remarkable than the rate before 2001. The results also indicated that the agricultural and urban land use categories have the highest possible probability of change. It is predicted that agricultural land will decrease by 11.5% and 24.5% in 2050 and 2100, respectively. Conversely, urban areas will increase by 12% and 25% in 2050 and 2100, respectively. This process will become the main feature of LULC change in the future, which means that the agricultural land could completely vanish within the next two hundred years at the current degree of agricultural land consumption by urbanization. The outcomes of this research could help policy makers to adopt appropriate strategies towards better management of LULC and increase the sustainability of the highly productive agricultural land through preventing violation on such valuable land.