Dealing with forecast uncertainty in short-range weather forecasts: user needs, scientific requirement and economic values

An example of public criticism about the accuracy of TV weather forecasting will be presented and analyzed to illustrate the user needs about forecast uncertainty. What can meteorologists learn from it to better serve public and end-users on a solid scientific basis? Given the chaotic nature of atmospheric system and imperfect observations and numerical models, it's scientifically impossible to predict the weather in 100% accuracy. Examples of such uncertainty in operational models will be demonstrated. Therefore, reform is needed in both the ways of producing and broadcasting weather forecast information. A forecast without explicitly describing quantitative uncertainty information is incomplete. Inclusion of forecast uncertainty can, instead, maximize the economical value of a forecast and satisfy the needs for a wider range of users. Two examples will be demonstrated to support the above claim. The principle of how to use probabilistic information in decision-making and how to measure a forecast's economic value will also be described. It's time to quantitatively add forecast uncertainty to weather, climate, water and any environmental prediction. Training and education to forecasters, end-users and public is the key to the success of this revolutionary transition from a deterministic to a stochastic point of view about weather forecasting and service.