88th Annual Meeting (20-24 January 2008)

Monday, 21 January 2008: 11:15 AM
Automatic verification of the BlueSky smoke forecasting system
220 (Ernest N. Morial Convention Center)
Robert Solomon, USDA Forest Service, Seattle, WA; and N. Larkin, S. O'Neill, T. Strand, and B. Potter
BlueSky is a modeling framework that provides real-time predictions of surface smoke concentrations from prescribed fire, wildfire and agricultural burn activities to aid land managers. The framework is flexible enough to allow for varying combinations of the relevant physical models required to predict smoke and/or various other emissions from fire (e.g., a variety of fuelload maps, consumption, emissions and dispersion models). BlueSky has been implemented through the United States by the Forest Service as well as other agencies.

An essential component of any model/modeling system is a means for users to evaluate performance. For smoke modeling, a verification system is necessary because land managers need to evaluate their burn decisions against potential National Ambient Air Quality Standard (NAAQS) exceedances, which requires a clear understanding of the accuracy of the predicted particulate matter (PM) concentration fields.

An automated system for evaluating BlueSky smoke concentrations has been developed and made web accessible. Each day, observed air quality data throughout the United States is obtained via the EPA's AirNow site and compared to the prior day's BlueSky forecast; currently, only PM2.5 is shown on the web site. The web site gives a qualitative measure of how well predictions and observations compared and using some of the more advanced features, quantitative, statistical measures can be given for each measurement location.

Supplementary URL: