Fire history database of the western United States: preliminary report covering Oregon and Washington

Human activities are changing the chemical composition of the atmosphere at an unprecedentedrate, that may lead to significant changes in climate (Bolin and others 1986; Houghton and others1990, 1992). These climatic changes could directly alter fire frequency, extent and severity bychanging the amount, distribution and seasonality of precipitation and other factors that influencefire (Clark 1990; Flannigan and Van Wagner 1991). Climate may also change the rate of forestproduction, mortality and decomposition that will change the amount and distribution of fuel,hence indirectly alter fire regimes (Clark 1990). In addition, changes in climate could alter theglobal distribution of forest life-zones (Emanuel and others 1985; Leverenz and Lev 1987; Smithand others 1992). The rate at which forest communities adjust to climate change will becontrolled in part by disturbance processes, primarily fire and land use practices (Overpeck andothers 1990). Fire is the dominant natural disturbance in many parts of the western United States(Pyne 1982; Agee 1993) where steep topographic and climatic gradients result in a great varietyof fire regimes. A continental-scale fire frequency model is being developed by the U.S.Environmental Protection Agency (EPA) as an essential component of a broad-scale vegetationmodel used to predict the response of vegetation of global climate change. This workplan detailsa procedure to assemble existing tree-ring information on past forest fire regimes in thecontinental United States (west of 100°W, exclusive of Alaska) that can be used to calibrate andverify the EPA model.
Fire frequency in forested areas can be reconstructed by dating the annual ring in which a fire scarforms and/or estimating the year of origin of a tree that regenerates after fire (e.g. Barrett andArno 1988, Sheppard and others 1988). The annual rings are either dated dendrochronologicallyusing prepared samples (by crossdating patterns of wide and narrow rings (Stokes and Smiley1968)) or by ring-counting, in the field or laboratory, using minimally prepared samples. Fire sizeis estimated from the number and spatial distribution of trees or sites recording fire in a given year(e.g. Agee and others 1990; Swetnam and Dieterich 1985; Baisan. and Swetnam 1990).Numerous fire histories, reconstructed from tree-rings for small portions of the western UnitedStates, are available from both published and unpublished sources. These reconstructions cover awide variety of vegetation types and topographic settings.