A confidence index for numerical weather prediction forecasts

A method for assessing the confidence in a current weather forecast is presented. This method, The Confidence Index (CI), uses not only recent forecast performance but also detects the presence of complicating synoptic scale features (e.g., vorticity maxima, jet streaks, closed lows and highs) that may create sudden weather changes and potential error in the forecast. The Index creates a spectrum of forecast confidence ranging from a low-risk forecast through progressively higher risks in which additional decision-making may be required on the part of the user. CI uses forecast and analysis fields available in most global meteorological models and can be applied to most nations' forecast products. The index can be specialized to a given region or belt of latitude, or globalized with varied levels of precision against historical forecast error. It can also be specialized to focus on specific meteorological fields to accommodate specific forecasting requirements such as day-to-day forecasting, fire meteorology, and wind forecasting. The development of the CI methodology, its tuning, and examples of its application and use over the continental United States will constitute the core of this presentation.