132 ensembleBMA: An R package for probabilistic weather forecasting

Tuesday, 25 January 2011
Chris Fraley, University of Washington, Seattle, WA; and A. E. Raftery, T. Gneiting, and M. Sloughter
Manuscript (1.3 MB)

The ensembleBMA software package, written in the R language and available through CRAN, provides functions for modeling using ensemble postprocessing via Bayesian model averaging (BMA), and for forecasting and verification from these models. The modeling functions estimate model parameters from training data via the EM algorithm for mixture models, and can accommodate data that may have exchangeable and/or missing ensemble members. Current modeling options include mixtures of normal distributions (appropriate for temperature or pressure), mixture of gamma distributions (appropriate for maximum wind speed), and mixtures of gamma distributions with a point mass at 0 (appropriate for quantitative precipitation).

Supplementary URL: http://cran.r-project.org/web/packages/ensembleBMA/index.html

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner