River Forecast Application for Water Management: Oil and Water?

Abstract

Managing water resources generally and managing reservoir operations specifically have been touted as opportunities for applying forecasts to improve decision making. Previous studies have shown that the application of forecasts into water management is not pervasive. This study uses a scenario-based approach to explore whether and how people implement forecast information into reservoir operations decisions in a workshop setting. Although it was found that participants do utilize both forecast and observed information, they generally do not utilize probabilistic forecast information in a manner to appropriately minimize risks associated with the tail end of the forecast distribution. This study found strong tendencies for participants to wait for observed information, as opposed to forecast information, before making decisions. In addition, study participants tended to make decisions based on median forecast values instead of considering forecast probability. These findings support the development of quantitative decision support systems to optimally utilize probabilistic forecasts as well as for forecast agencies such as NOAA/NWS to continue investments in work to better understand contexts and environments where forecasts are used or have the potential for use in supporting water management decisions.