Optimizing a Network of Weather Stations in Arid Regions using a Fuzzy AHP-Based Approach

Abstract Accurate hydrological analyses largely depend on precipitation data from weather stations (WSs). The stations’ density and spatial distribution are essential for ensuring data accuracy. However, in regions with diverse climatic conditions, WS network often fail to meet the World Meteorological Organization (WMO) standards. This study evaluates the WS network in Isfahan Province, an arid to semi-arid region, to develop an optimized system for reliable meteorological data collection. A novel framework is introduced to achieve this goal by integrating statistical analysis with the Fuzzy Analytical Hierarchy Process (Fuzzy AHP). Initially, a correlation equation was fitted to annual precipitation records, and the absolute relative error was distributed using the Kernel Density Function. Next, the Fuzzy AHP algorithm was employed to generate a weighted overlay layer based on seven critical physical and environmental factors: elevation, slope, proximity to existing stations, land use, proximity to roads, distance from streams, and population centers. These two outputs were combined to produce a refined suitability map, identifying 17.9% of the land area as highly or fairly suitable for WS establishment. According to WMO guidelines, an additional 72 WSs are required throughout the province. Results demonstrated that incorporating error-informed weighting from spatial rainfall uncertainty into a GIS-based multi-criteria framework significantly improved the spatial accuracy in data-scarce regions. The scalable framework offers a practical tool for meteorological agencies, planners, and researchers, supporting more resilient networks and accurate hydrological analyses.