Strategic siting of desalination plants and design of an optimal water distribution network

Water scarcity is an increasing challenge in arid and semi-arid regions, making desalination an essential solution for meeting rising freshwater demand. Effective placement and design of desalination plants within water distribution systems are crucial for ensuring long-term sustainability and cost efficiency. Focusing on Bhavnagar District, Gujarat, where rapid population growth, industrial development, and climate variability intensify water stress, this study introduces an integrated framework combining Geographic Information System (GIS), Remote Sensing (RS), and Random Forest (RF) machine learning to identify optimal desalination plant sites and design an efficient distribution network. Using multi-criteria spatial datasets—land use/land cover, slope, elevation, groundwater salinity, coastal proximity, and accessibility—the RF model classified site suitability with high accuracy (AUC > 0.9). The district (6514.33 km²) was categorized into five suitability zones: very low (21.87 %), low (29.05 %), moderate (20.36 %), high (16.69 %), and very high (12.02 %). Highly suitable zones were concentrated along coastal belts, supporting seawater intake needs and infrastructure efficiency. An optimized distribution network considering settlement density, demand centers, and pipeline accessibility was developed to minimize losses and costs. This integrated GIS–RS–RF framework provides a scalable model for desalination planning and offers actionable insights for enhancing water security in coastal, water-stressed districts.