Deployment optimization method for hydrological sensor network to maximize spatial coverage

Abstract

A hydrological sensor network is a comprehensive monitoring system which consists of varieties of hydrological stations. The spatial optimization of the deployment of these gauging stations is critical for obtaining hydrological information effectively, given that sensing resources are often limited. In this paper, we consider the hydrological station siting problem as a continuous location problem and propose a general method for siting hydrological stations in space to maximize the total area for monitoring. Considering the coverage characteristics of different stations, the target watershed area is discretized into a series of area-based or line-based demand objects with specific weights. The effective coverage range of a station is determined based on the minimum density required and terrain conditions. Two types of finite dominant sets (FDS) are designed to identify the critical potential locations for line-coverage and area-coverage stations. By formulating a modified maximal covering location problem (MCLP) model with distance constraints which follow the general regulations for hydrological sensor network design, the siting problem can be solved using the MCLP-based model and FDS. The precipitation stations and streamflow stations in the lower reach of the Jinsha River Basin are selected as examples to test the performance of the proposed optimal siting method. Results show that the model solution achieves better coverage than that of the real-world deployment. The applicability of the proposed method, criteria for candidate site selection, and impacts of different weighting schemes are also discussed.