. Extended-range probabilistic fire-weather forecasting based on Ensemble Model Output Statistics and Ensemble Copula Coupling

Abstract
Probabilistic fire-weather forecasts provide pertinent information to assess fire behavior and danger of current or potential fires. Skill of these forecasts can affect fire-management practices and resources, and firefighter safety. Currently, operational fire weather guidance is only provided for lead times less than seven days, with most products only providing day 1-3 outlooks. None of the fire forecasts are statistically post-processed to correct for model biases. Extended fire weather forecasts out to two weeks can aid in decisions regarding placement of in- and out-of-state resources, ideal conditions for prescribed burns, and to assess preventative wildland fire strategies. Ensemble Model Output Statistics (EMOS) and ensemble copula coupling (ECC) postprocessing methods are used to provide locally calibrated and spatially coherent probabilistic forecasts of the Hot-Dry-Windy index, a purely meteorological fire-weather index out to two weeks. The univariate post-processing approach uses transformations of the truncated normal distribution to generate probabilistic forecasts of future observations conditional on the current ensemble forecasts. Twenty years of reforecasts and ERA-Interim reanalysis data over CONUS from the European Centre for Medium-Range Weather Forecasts are used to perform the postprocessing methods. Verification of the resulting probabilistic forecasts is quantified with the continuous ranked probability score (CRPS), and performance is evaluated for wildfire case studies.