Open AccessA Laboratory Study of the Benefits of Including Uncertainty Information in Weather Forecasts
Author(s) -
Mark S. Roulston,
Gary E. Bolton,
Andrew N. Kleit,
Addison L. Sears-Collins
Publication year - 2006
Publication title -
weather and forecasting
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.393
H-Index - 106
eISSN - 1520-0434
pISSN - 0882-8156
DOI - 10.1175/waf887.1
Subject(s) - predictability , computer science , numerical weather prediction , forecast skill , model output statistics , public information , consensus forecast , data assimilation , econometrics , environmental science , meteorology , operations research , statistics , economics , mathematics , geography , internet privacy
Modern operational methods of numerical weather prediction, such as “ensemble forecasting,” allow assessments of state-dependent predictability to be made. This means that forecast-specific estimates of the forecast standard errors are possible. Quantitative estimates of forecast uncertainty are often not communicated to the public as it is unclear what the value of this information will be to people who must make weather-dependent decisions. Using laboratory-based methods developed by experimental economists to study individual choice it is found that nonspecialists are able to make better decisions that increase their expected reward while reducing their exposure to risk, when provided with information about the day-to-day uncertainty associated with temperature forecasts. The experimental framework used herein may provide a useful tool for evaluating the effectiveness with which weather forecasts can be communicated to end users.