Spatio-temporal analysis of Land Use and Land Cover Changes in the Mitidja plain (1983–2023): impact of urbanization on agricultural land

Changes in land use and land cover (LULCC), particularly the conversion of agricultural land to urban areas in the Mitidja plain over the past century, have significantly impacted agriculture and the economy of Algeria, compromising the crucial role of this fertile region in ensuring national self-sufficiency in food production. These concerns have motivated the present study, aimed at analyzing a series of satellite images from Landsat and Sentinel-2 satellites, covering the period from 1983 to 2023, on the Google Earth Engine platform. The objective is to differentiate between various land classes through supervised classification and assess the temporal variation of their surface area using the modified Mann–Kendall test. Two algorithms, namely Random Forest (RF) and Smile CART, were employed to generate LULC classes and map the transition from agricultural to urban landscapes. The results demonstrated the superiority of the RF classifier in detecting land cover changes associated with the expansion of urban areas at the expense of agricultural land compared to the Smile CART algorithm. Over the period 1983–2023, the periodic evaluation of the surface areas of each land class on a 5-year basis provided valuable insights into the spatial evolutions of urbanization at the expense of cultivated lands. During this period, cultivated lands experienced a significant decrease of − 7.1 km2/year, representing a decrease of 28.04%. Conversely, built-up areas saw a significant increase of 6.84 km2/year, corresponding to a 25.67% increase. The improved accuracy achieved with RF classifier can help urban planners better monitor agricultural land conversion, enabling timely interventions to preserve critical farmland.