Development and Assessment of Methods for Predicting Rainfall in Unmeasured Watersheds Using the Relation Between Observed and Sensor-Measured Rainfall

Abstract

Recently, with the development of information and communication technology and the Internet of Things (IoT), observation technology using sensors is being applied in a variety of ways, such as using a sensor to observe rainfall in an unmeasured area. In this study, the relationship between the rainfall sensor signal (S) and the amount of rainfall (R) was developed through an experiment in an artificial rainfall generator, and the applicability was evaluated through outdoor observation. The coefficient of determination of the relational expression developed through the indoor experiment was 0.95, the mean absolute error was 2.66 mm/hr, the root mean square error was 3.87 mm/hr, the efficiency coefficient was 0.89, and the concordance index was 0.97, showing very high reliability. In the outdoor test results, the error rate was 7.96% when comparing the data from the rainfall sensors in vehicles and the precipitation station, which were not observed at the same location. Despite such errors, it is judged that accurate rainfall observation using a rainfall sensor is possible in an area where a precipitation station is not installed.