New optimization methods for designing rain stations network using new neural network, election, and whale...

Area of Science:

• Hydrology
• Environmental Science
• Data Science

Background:

• Accurate spatial precipitation estimation is crucial for water resource management.
• Existing rain gauge networks often require optimization for efficiency and cost-effectiveness.

Purpose of the Study:

• To optimize the rain gauge network in Sistan and Baluchestan Province.
• To identify redundant stations without compromising rainfall estimation accuracy.
• To introduce a novel neural network algorithm for network optimization.

Main Methods:

• Development of a new meta-heuristic optimization algorithm based on artificial neural networks (ANNs), termed the neural network algorithm (NNA).
• Evaluation of NNA against election and whale optimization algorithms.
• Analysis of Kriging variance and topography in relation to station network optimization.

Main Results:

• The NNA demonstrated superior performance with a mean error of 0.06 mm compared to other algorithms.
• Identified 22 out of 49 existing rain gauge stations as having no significant impact on rainfall estimation.
• Recommended reducing the network to 27 stations for optimized spatial precipitation estimation.

Conclusions:

• The proposed NNA is effective in optimizing rain gauge networks.
• Network optimization can lead to cost savings without sacrificing data quality.
• The study provides a data-driven approach for improving hydrological monitoring networks.