Analysis of Spatial and Temporal Changes in Land Cover and Land Use: The Case Study of Kocaeli-Kandıra

Abstract

The ability to accurately determine land cover change and the direction of change provides the scientific basis for policy implementation and predictable urban planning, and is one of the most reliable sources of data. The utilisation of cloud-based platforms, such as Google Earth Engine (GEE), facilitates the acquisition of all satellite data monitoring our planet, thereby offering substantial advantages. This facilitates specific or global analyses of each region, enabling such studies to be conducted with minimal effort. This study examines the district of Kandıra in Kocaeli province, which has undergone significant land use transformation in recent years due to increasing pressure from transport, urbanisation, industrialisation for food production and tourism. The amount and spatial character of the transformation in question were analyzed between 2017 and 2024. In this regard, the Random Forest classifier executed on Sentinel-2 satellite images on the Google Earth Engine platform was cross-validated with the Landsat-based LandTrendr time series algorithm. The analysis revealed that 2,503 hectares of agricultural land in the study area had been converted to artificial land. When comparing the results of two different algorithms, the spatial overlap between the methods was calculated as 0.1041 using the Jaccard Similarity Index; the parcel-based details of the classification method and the complementary role of LandTrendr in capturing vegetation damage were revealed. When examining the spatial characteristics of the change, it was determined that new developments were concentrated at an average distance of 727 meters from the Rural Housing Areas designated in the 1/25,000 scale Master Development Plan. This finding proves that the transformation in Kandıra is not a random scattering (leap-frog/outlying), but rather follows a spread model defined in the literature as Edge Expansion, which extends beyond the boundaries of planned rural settlements, and that the pressure of unplanned construction is concentrated on agricultural lands.