P6.2
The presentation of risk and uncertainty in hydrologic forecasts by the U.S. National Weather Service
Thomas Adams, NOAA/NWS, Wilmington, OH
All forecasts are uncertain and many factors contribute to their uncertainty. With hydrologic forecasting, these uncertainties or errors arise principally from the meteorological inputs, including quantitative precipitation and temperature forecasts (QPF and QTF, respectively), as well as quantitative precipitation estimates (QPE) and temperature estimation. Since models and initial model states are imperfect, they add to forecast uncertainty, as do forecaster inputs. The relative magnitude of these contributing errors varies from forecast to forecast. To varying degrees some of the errors associated with the hydrologic modeling process are being conveyed in current forecast products within the U.S. National Weather Service (NWS) Advanced Hydrologic Prediction Services (AHPS) web-based graphics and text products. The nature of these text and graphical products are continuing to evolve. The central question that arises, however, is how best to convey this information to the general public, emergency managers, and decision makers to be most efficacious in their assessment of risk. If the information contained in the raw forecasts is not easily understood, little or no information will be conveyed to the desired audiences, the level of risk will not be interpreted correctly, and inappropriate decisions will be made.
This paper presents the range of NWS hydrologic products available at present, which conveys forecast uncertainty. However, since we are in the infancy of expressing hydrologic forecast uncertainty to the general public, emergency managers, and decision makers, there is ample room for improvement. One of the desired outcomes of this paper is to bring the discussion of the presentation of forecast uncertainty to the fore and to elevate the need for end-user understanding of forecast uncertainty and risk. If this is achieved, end-users of hydrologic forecasts will understand that deterministic forecasts often provide misinformation, especially during rapidly changing, extreme conditions when forecasts are most critical, and are correct only by random chance.
Poster Session 6, New Approaches to Hydrologic Prediction and Forecast Communication
Thursday, 15 January 2009, 9:45 AM-11:00 AM, Hall 5
Previous paper Next paper