Wind climate analysis for prescribed fire planning: case study of Tallahassee, Florida

Abstract

Background: The wind rose is the main tool available for summarizing winds of a location for prescribed burners, but its ability to express variability in wind speed and direction is limited because winds are confined to speed and direction bins. The Offset Elliptical Normal mixture model uses joint bivariate normal distributions to capture variability in wind speed and direction. This mode of wind classification enhances the ability to link winds of similar speed and direction to driving dynamics. We apply the Offset Elliptical Normal mixture model as a case study to winds in Tallahassee, Florida, a national center for prescribed burning. The output provides probability and variability of wind direction and speed by month and hour. Results: Winter in Tallahassee is dominated by northern frontal winds, spring by weak prevailing western winds, summer by weak multi-directional winds, and autumn by easterly anticyclonic winds. The Offset Elliptical Normal mixture model provides relative frequency of each wind, so prescribed burners can anticipate how often favorable or unfavorable winds may occur. The statistical underpinnings of the mixture model allow random selection of empirically correlated speed and direction for any ellipse, which could prove useful for prescribed fire modeling. Conclusion: The substantive improvement of the Offset Elliptical Normal mixture model compared to the wind rose argues for its wider use by the wildland fire community.