Reading Room

Motivations for probabilistic forecasting

The scientific basis for expressing weather forecasts in terms of event probabilities is given by Zhu et al. (2002), along with examples showing how decision-making can be optimized to maximum economic benefit. The NWS commissioned the National Research Council to write Completing the Forecast, which outlines societal needs, institutional responsibilities, etc.

The NOAA Forecast Uncertainty Steering (NFUSE) Team conducted a survey of NWS forecasters and managers to assess the readiness of the NWS to add uncertainty information to the current suite of products (Novak et al. 2008). This survey was drawn on heavily to assess requirements that appear in OSIP documentation.

Operational Forecaster Uncertainty Needs and Future Roles
David R. Novak, David R. Bright, Michael J. Brennan
Weather and Forecasting
2008 early online release, posted January 2008
DOI: 10.1175/2008WAF2222142.1
[Abstract] [PDF(1.17M)]

Click here to view a recent State of the Science FACT SHEET for Weather Forecast Uncertainty. This is a draft document and will be revised another couple of times, so please check back again for additional updates.

Workshop Reports

This section contains reports generated from the 1st and 2nd workshops held at GSD. These reports contain an overview of the workshops, plus the findings, outcomes, and recommendations for future planning.

AWIPS II

The software being developed at GSD is ultimately intended for implementation in AWIPS II. Planning information and other documentation about AWIPS II can be found here.

References

Scientific/Economic basis for probabilisitic forecasting:

EKDMOS:

Reforecast-based methods:

A method based on Bayes' Theorem:

Forecaster survey, and a very nice bibliography on ensemble methods and probabilistic forecasting: