AAAR 34th Annual Conference
October 12 - October 16, 2015
Hyatt Regency
Minneapolis, Minnesota, USA
Abstract View
Sensitivity of Ambient PM2.5 Concentration to Prescribed Burning and Fire Weather Forecast Data Using Principal Components Regression Analysis
KARSTEN BAUMANN, Sivaraman Balachandran, Jorge Pachon, James Mulholland, Armistead G. Russell, Atmospheric Research & Analysis
Abstract Number: 550 Working Group: Source Apportionment
Abstract Fire weather forecasts published daily by NOAA’s National Weather Service, are used by land and wildlife managers to determine when meteorological and fuel conditions are suitable to conduct prescribed burning (PB). In this work, we investigate the sensitivity of ambient PM$_(2.5) to various fire and meteorological parameters in a spatial setting that is typical for the Southeast. We use the method of principle components regression (PCR) to estimate the sensitivity of PM$_(2.5), measured at a regulatory monitoring site in Jacksonville, North Carolina, to fire data and, to observed and forecast meteorological parameters. Fire data was gathered from routine PB operations at Marine Corps Base Camp Lejeune, extending 10-50 km south from Jacksonville. Principal components analysis (PCA) was run on ten data sets that included PB activity data along with meteorological forecast data alone or in combination with observations. For each data set, PCA scores from the first seven principal components (explaining > 80% of total variance) were regressed against observed PM$_(2.5). PM$_(2.5) showed significant sensitivity to PB, with a unit-based sensitivity of 3.6 ±1.1 µg m$^(-3) per 1000 acres burned at the investigated distance scale. Applying this sensitivity to the available PB activity data revealed a PB source contribution to measured PM$_(2.5) of up to 25%. As expected, PM$_(2.5) had a negative sensitivity to dispersive parameters, and was sensitive to wind direction. The PCR method showed positive sensitivity to forecast precipitation, likely reflecting land managers’ decision to conduct PB on days where rain can naturally extinguish fires. Perhaps most importantly for land managers, our PCR analysis suggests that instead of relying on the forecasts from a day before, their PB decisions should be based on the forecasts released the morning before the burn, since these data were more stable to PCR treatment and yielded more statistically robust results.