Abstract
Wild and prescribed fire injury to trees can produce mortality that is not immediately apparent, and environmental stress subsequent to a fire may also contribute to tree mortality in the years after a fire (Hood and Bentz 2007). In order to predict post-fire tree mortality from fire injury variables before tree mortality is clearly apparent, dozens of statistical logistic regression models have been developed (see Woolley and others 2012 for a review), and some are incorporated within larger fire behavior and effects models and computer models used to support land management decisions (Hood and others 2007, Reinhardt and others 1997). In the Wallowa and Blue Mountains of northeastern Oregon, a polychotomous field key has been developed to predict tree mortality (Scott and others 2002). For this project, post-fire tree mortality and fire injury variables were collected for 26 wild and prescribed fires across Oregon and Washington, providing an opportunity to test specific published models for 16 species of conifers. Our objectives with these data are (1) to assess the ability of previously published logistic regression models and other guidelines or methods to predict 3-year post-fire tree mortality in Oregon and Washington State and (2) to identify suites of fire injury variables that best discriminate between live and dead trees in that region.
Parent Publication
Citation
Progar, Robert A.; Ganio, Lisa; Grayson, Lindsay; Hood, Sharon M. 2018. Chapter 9 - Monitoring survival of fire-injured trees in Oregon and Washington(Project WC-F-08-03). In: Potter, Kevin M.; Conkling, Barbara L., eds. 2018. Forest health monitoring: national status, trends, and analysis 2017. General Technical Report SRS-233. Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research Station. Pages 143-150.
