Utilizing Spatio-temporal Satellite Images and Machine Learning to Examine the Impacts of Changing Land Use and Land Cover on Sohar City

The growing urbanization of cities and towns in the Gulf Cooperation Council (GCC) states as a result of socioeconomic change has primarily resulted in a great deal of strain on the few natural resources and the loss of productive areas. In fact, there hasn't been much focus on the spatial patterns of urbanization and how they affect the environment and agricultural resources, especially in Oman. It may be possible to better understand the connection between spatial growth patterns and its effects on agricultural production by predicting urban growth in Sohar City. This study aims to analyze spatiotemporal dynamics of land use/land cover (LULC) for the last six years (2017–2022) and its impacts especially on the reduction of agricultural land. The spatiotemporal data was obtained from Sentinel-2 with a cell size of 10 meters which was then cropped using the vector boundaries of Sohar city. Part of the data was labelled manually which was used to train supervised machine learning technique for classifying the data into different land covers such as water bodies, trees, crops, etc. Accuracy assessment was carried out by com-paring the automatic classification results with ground truth and an accuracy of 93.4% was obtained. It has been found that built up area is increasing rapidly while hugely affecting the arable land. It's also interesting to see that during the COVID19 pandemic, this urban growth was stopped as evident from the satellite data. The results of this study could address the dangers and declines of urban sustainability for Sohar city, as well as provide geographical guidelines for observing future changes in LULC dynamics. The national strategy for future urban development in Oman also benefits from identifying regions of bare soil and vegetation that are amenable to urbanization.