Research Summaries

Back Testing and Refinement of Bayesian Data Analytics for Tactical Prediction Capabilities in Data Denied Areas

Fiscal Year 2020
Division Research & Sponsored Programs
Department NPS Naval Research Program
Investigator(s) Nuss, Wendell A.
Sponsor NPS Naval Research Program (Navy)
Summary Battlespace decisions rely on: a) accurate forecasts of environmental conditions and their impact on operations; and b) information about the uncertainty or probability that the conditions and impacts will occur as forecast. Numerical weather prediction (NWP) models generally provide forecasts of environmental conditions but lack probability information unless run as ensemble systems. Error in these systems and inaccurate probability characterizations can severely affect operational decisions, especially in data denied regions where little observational ground truth data is available. Bayesian post-processing of NWP fields can be done to reduce errors and provide a complete characterization of the probability of predicted conditions. Post-processing of short-range NWP predictions winds, temperatures, clouds and other important environmental information lends itself to machine learning (ML) approaches but require observations from which to learn. The Bayesian multivariate approach has been applied to a limited number of wind and cloud forecasts successfully using minimal learning data. This approach will be extended to additional variables and refined to more fully exploit the probability information in those predictions.
Keywords
Publications Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal
Data Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal