:: Optimal‎ ‎Selection of Rain Gauge Numbers and Locations for Estimating Mean Rainfall in Fars Province Using a Hybrid Geostatistical and Artificial Bee Colony Approach (Dr. Mohammad Javad Abedini)

The optimal design of rain gauge station networks is one of the classic challenges in‎ ‎applied hydrology‎. ‎In research and studies over the past three decades‎, ‎methods with simplifying‎ ‎assumptions have been used to overcome the curse of dimensionality‎. ‎In the present study‎, ‎a new‎ ‎and efficient method—without any simplifying assumptions—has been employed for the optimal design of‎ ‎rain gauge station networks‎. ‎For the first time in this study‎, ‎geostatistical methods have been used‎ ‎to formulate the objective function‎, ‎combined with the artificial bee colony (ABC) algorithm‎, ‎to‎ ‎design and prioritize stations in the rainfall monitoring network‎. ‎To evaluate the effectiveness of‎ ‎the proposed method‎, ‎the design of the rain gauge station network in Fars Province—comprising 163‎ ‎existing non-recording rain gauges and 46 proposed potential stations—was examined using this‎ ‎approach‎. ‎The artificial bee colony algorithm‎, ‎due to its limited number of control parameters‎, exhibits relatively high speed in determining the optimal solution‎. ‎After running the optimization‎ ‎algorithm for three networks—53 stations (rain gauges installed in fences)‎, ‎163 stations ‎(non-recording rain gauges)‎, ‎and 209 stations (non-recording rain gauges plus proposed locations)—the‎ ‎results demonstrated the method's efficiency and confirmed its performance compared to existing‎ ‎paradigm in prior studies‎. ‎While traditional classical methods‎, ‎relying on simplifying assumptions‎, fail to address the curse of dimensionality‎, ‎the proposed method—leveraging the nature of‎ ‎intelligent‎, ‎randomized‎, ‎and domain-specific search—can be applied to a wide range of small‎, ‎medium‎, or large values of n and is capable of providing an optimal rain gauge network configuration‎. ‎As‎ ‎the number of rain gauges increases‎, ‎the optimization algorithm encounters more favorable options for‎ ‎reducing error variance‎. ‎Given its reasonable speed in obtaining optimal solutions‎, ‎the proposed‎ ‎method can serve as a reference for validating hypotheses used in other methods designed for medium‎ n  values and across a wide range of C(N,n) values in the scientific community‎. ‎The results of‎

‎applying this method across three distinct networks in Fars Province will provide decision-makers‎ ‎with more effective options for daily rainfall monitoring‎, ‎enhancing the precision of water balance‎ ‎components in annual water balance equations.