Prototyping low-cost automatic weather stations for natural disaster monitoring

Weather events put human lives at risk mostly when people might occupy areas susceptible to natural disasters. Deploying Professional Weather Stations (PWS) in vulnerable areas is key for monitoring weather with reliable measurements. However, such professional instrumentation is notably expensive while remote sensing from a number of stations is paramount. This imposes challenges on the large-scale weather station deployment for broad monitoring from large observation networks such as in Cemaden-The Brazilian National Center for Monitoring and Early Warning of Natural Disasters. In this context, in this paper, we propose a Low-Cost Automatic Weather Station (LCAWS) system developed from Commercial Off-The-Shelf (COTS) and open-source Internet of Things (IoT) technologies, which provides measurements as reliable as a reference PWS for natural disaster monitoring. When being automatic, LCAWS is a stand-alone photovoltaic system connected wirelessly to the Internet in order to provide real-time reliable end-to-end weather measurements. To achieve data reliability, we propose an intelligent sensor calibration method to correct measures. From a 30-day uninterrupted observation with sam-pling in minute resolution, we show that the calibrated LCAWS sensors have no statistically significant differences from the PWS measurements. As such, LCAWS has opened opportunities for reducing maintenance costs in Cemaden's observational network.

Automated summary

This paper proposes a low-cost automatic weather station system (LCAWS) that provides reliable weather measurements through an intelligent sensor calibration method. The calibrated LCAWS sensors show no significant differences from the reference PWS measurements, reducing maintenance costs in Cemaden's observational network.