EDM: An ensemble decision method for introducing meteorological forecast uncertainty into the disaster response decision process

Abstract

In this study, an ensemble decision method (EDM) that ensures that disaster risk information includes sufficient meteorological forecast uncertainty and that decision supporting information retains sufficient disaster risk is proposed, thereby resolving the current issue of incomprehensive information. Visual presentations of real-time, multidimensional information are created to help decision-makers discern comprehensive information directly from ensemble warning charts, thereby enhancing the efficiency of operations and the quality of the decision process. The EDM comprises three steps: developing the disaster data model, developing ensemble warning charts, and developing the decision process to utilize the charts. In the first step, simulations are run using the forecast data of several ensemble members, rainfall inundation thresholds, and flooding risk hotspots in weather forecasts. In the second step, ensemble warning information is visualized by adopting three types of charts to present the disaster simulation results: a counties and cities inundation warning threat graph, a towns inundation warning time series chart, and a multidimensional inundation warning chart. In the third step, a decision process is designed to support decision-makers in assessing inundation threats, evaluating activation of disaster response operations, and determining response levels. To validate the proposed method, it is tested using data from three torrential rain events in 2019. The test results of the three torrential rain events show that the inundation warnings produced by the EDM better reflect actual inundation than those produced by the deterministic forecasts.