Impact of Airport Construction on the Spatial Pattern of Land Use in Agricultural Area: A Case Study of Majalengka Regency, Indonesia

Land is a fundamental and finite resource that supports the physical development of various sectors, including agriculture, housing, industry, mining, and transportation. As population growth accelerates, the demand for land increases accordingly, leading to a dynamic competition among different land use types. This condition contributes significantly to the transformation of land use and land cover (LULC), especially in rapidly developing areas. These transformations often involve the conversion of productive agricultural land into non-agricultural uses, raising concerns over long-term food security, ecological sustainability, and spatial equity. Majalengka Regency, located in West Java Province, Indonesia, has become one of the most dynamic regions in terms of land transformation due to its inclusion in the national development strategy through the Rebana Triangle Special Economic Zone (SEZ). This strategic zone, connecting Cirebon, Patimban, and Kertajati, is envisioned to accelerate industrial growth, infrastructure development, and regional connectivity. Within this framework, Majalengka plays a critical role, particularly with the construction of West Java International Airport (BIJB) and the surrounding industrial corridors. While these developments offer promising economic potential, they also exert significant pressure on existing land resources, particularly agricultural areas such as rice fields and drylands. The objective of this study is to analyze spatial and temporal changes in LULC across Majalengka Regency over a ten-year period from 2011 to 2021. The research utilizes multispectral remote sensing data from the Sentinel-2A satellite, which provides high temporal resolution and medium spatial resolution suitable for regional-scale analysis. The image data were processed using the Google Earth Engine (GEE), a cloud-based geospatial analysis platform that enables efficient access, processing, and analysis of large satellite datasets without requiring local storage or high-end computing infrastructure. GEE's capability to conduct multi-temporal analysis over extended periods makes it a valuable tool for environmental monitoring and spatial planning. Land cover classification was conducted using the Smile-Random Forest (RF) algorithm, a supervised machine learning approach known for its accuracy in handling multidimensional remote sensing data. To improve the accuracy and thematic detail of the classification, two additional spectral indices were incorporated: the Normalized Difference Built-Up Index (NDBI) and the Normalized Difference Water Index (NDWI). These indices enhanced the model’s ability to distinguish between built-up areas, water bodies, and vegetated or agricultural land. Furthermore, base maps of rice field distribution were integrated into the classification process to refine the delineation of agricultural zones, particularly paddy fields that are crucial to local food systems. The classification model achieved high levels of accuracy, with an Overall Accuracy (OA) of 98.81% and a Kappa coefficient of 95.91%, indicating a strong level of agreement between predicted and ground truth data. The spatial analysis revealed a considerable decline in agricultural land over the ten-year period, with a net reduction of approximately 4,457.36 hectares in rice fields and drylands. Conversely, there was a marked increase in built-up land, including residential settlements, industrial areas, and transportation infrastructure associated with the expansion of BIJB and its surrounding economic zones. These findings underscore the complex relationship between regional development initiatives and environmental sustainability. The ongoing land conversion, if left unregulated, poses risks to agricultural productivity, local food security, and ecological resilience. The study emphasizes the need for integrated land use planning and policy interventions that balance economic growth with the conservation of essential land resources. Spatial planning efforts must be aligned with long-term sustainability goals, particularly in regions that are designated for strategic economic development. Moreover, this research highlights the utility of combining remote sensing technology, spectral indices, machine learning algorithms, and cloud-based platforms as an efficient and scalable methodology for monitoring land use change. The use of GEE significantly reduces processing time and technical barriers, allowing for real-time analysis and easy replication in other regions facing similar developmental pressures. In conclusion, this study provides a comprehensive assessment of land use changes in Majalengka Regency within the context of rapid economic transformation. The methodological framework presented here offers practical implications for policymakers, planners, and environmental managers in designing land governance strategies that are data-driven, forward-looking, and sensitive to both developmental and environmental dimensions.