Constrained optimization of sensor locations for existing light-pollution monitoring networks

The high levels of nocturnal artificial light emissions from urban areas induce adverse effects on the environment and human health. Having adequate monitoring provides information for planning public policies and measuring their degree of achievement. Consequently, there is a need for optimal designing or redesigning of Light-Pollution Monitoring Networks focused on sensing the most impacting regions such as protected natural areas, and delivering representative measurements of the whole region of interest. This article addresses the optimum redesign and improvement of existing monitoring networks constrained to relocating a limited number of sensors. Furthermore, we introduce metrics to evaluate the covering region of a network and its accuracy for recovering artificial light emission levels of the whole area. Then, we find the sensor locations that maximize such metrics and, consequently, the monitoring performance. Our results are novel designs that use punctual measurements to estimate light levels of the whole region of interest for different numbers of sensors. Finally, the method, metrics, and a study case provide clear guidelines for the re-design of monitoring networks regarding sensitive zones for health, urban, or ecological planning, as well as the intensity of light emissions and existing monitoring networks.