ABSTRACT: "Historical range of variability" and "reference conditions" are two concepts that attempt to characterize ecosystem conditions as they may exist in the absence of RESEARCH ARTICLE pervasive human impacts. However, to define reference conditions from reference landscapes, such as U.S. Forest Service Research Natural Areas, requires a long-term perspective by which to assess whether existing ecosystem conditions are driven by predominately natural rather than human factors. We used fire-scarred trees to reconstruct • centuries-long chronologies of surface fires in four research natural areas (three established and one proposed) that contain ponderosa pine (Pinus ponderosa Laws.) forests in South Historical Surface Dakota, Wyoming, and Colorado. Fire frequency was variable among research natural areas, but recent fire-free periods in three of the four areas were up to approximately 2.5 Fire Frequency in times longer than any presettlement intervals. Loss of surface fires most likely is related indirectly to recent land and resource use in areas outside of the research natural areas and related directly to fire suppression and livestock grazing in the research natural areas Ponderosa Pine themselves. Studies that attempt to define reference conditions for ponderosa pine ecosystems from existing conditions in these Research Natural Areas will need to consider Stands in Research changes that may have occurred in these areas as the result of loss of historical fire patterns. Determination of historical fire frequency also should provide useful information Natural Areas, for the management or restoration of ecosystem processes and conditions in these or similar natural areas. Central Rocky Index terms: fire chronologies, cross-dating, fire scars, historical range of variability, Pinus ponderosa Mountains and Black Hills, USA INTRODUCTION agement agencies are among the factors Peter M. Brown that have disrupted surface fire regimes in While change is normal in all ecosystems, ponderosa pine forests (Savage 1991; Cov- Rocky Mountain the rate or direction of ecological changes ington and Moore 1992, 1994; Touchan et Tree-Ring Research, Inc. brought about by land and natural resource al. 1995; Brown and Sieg 1996; Swetnam 2901 Moore Lane use during recent decades may be outside and Baisan 1996; Rile et al. 1997). Shifts Fort Collins, CO 80526 USA the range of what many ecosystems are from historical patterns also have resulted [email protected] able to sustain before they shift into new in altered fire regimes, with high-severity modes of behavior (Covington et al. 1994, crown-destroying fires replacing low-se- Michael G. Ryan Holling and Meffe 1996). A well-docu- verity surface fires (Swetnam 1990, Cov- Thomas G. Andrews mented example is changes that have oc- ington and Moore 1994, Arno et al. 1995), curred in ponderosa pine (Pinus pondero- and have led some researchers to question USDA Forest Service sa Laws.) forests of the western United whether ponderosa pine ecosystems are Rocky Mountain Research Station States over the past century (Cooper 1960; sustainable under these new disturbance 240 W. Prospect Road White 1985; Covington and Moore 1992, regimes (Covington et al. 1994). Ft. Collins, CO 80526 USA 1994; Mutch et al. 1993; Arno et al. 1995; • Rile et al. 1997). Historically open forests For management of ecosystems, it is crit- with scattered trees and high understory ical to know when human manipulation of biomass and diversity have shifted to for- ecosystem components may be compro- ests dominated by closed-canopy stands mising ecosystem function and resilience with reduced understory production and (Holling and Meffe 1996). Two species composition. Historical conditions concepts—"historical range of variability" were maintained by frequent, episodic (HRV; Morgan et al. 1994) and "reference surface fires that killed ponderosa pine conditions" (e.g., Kaufmann et al. seedlings and saplings before they were 1994)—are used to characterize able to reach the forest overstory (Cooper ecosystem processes and patterns in the 1960, White 1985). Livestock grazing, absence of pervasive human impacts. logging, geographical fragmentation HRV refers to long-term patterns of Natural Areas Journal 20:133—139 caused by road and fence construction, ecosystem conditions that prevailed and active fire suppression by land man- before, generally, the middle to late 1800s, when widespread non-Native Amer- Volume 20 (2), 2000 Natural Areas Journal 133 ican—primarily Euro-American—settle- We reconstructed historical ranges of vari- considered to be representative of condi- ment in the western United States began in ability in surface fire frequency in ponde- tions driven by predominately natural rather earnest. Although Native Americans are rosa pine stands at four RNAs (three es- than human agents. known to have altered ecosystems to a tablished and one proposed) in Colorado, limited extent prior to settlement, their Wyoming, and South Dakota. Fire frequen- At three of the four study areas, we col- effects do not match the predominance cy in the RNAs was reconstructed using lected 10 to 13 fire-scarred ponderosa pine and ubiquity of impacts that have occurred fire-scar records in dendrochronologically trees each from two stand-level sites to during the recent century as a result of cross-dated tree-ring series. These recon- provide replication of both spatial and tem- population increases and industrialization. structions were made to assess if fire fre- poral patterns in fire frequency (Table 1; Reference conditions are ecosystem con- quency was different in these reference Brown and Sieg 1996, 1999). At the fourth ditions defined either from analysis of landscapes during the postsettlement peri- study area (Hot Creek), we collected 18 presettlement patterns or from present-day od, or if recent fire frequency and likely fire-scarred ponderosa pine trees from one ecosystems that have been altered mini- associated ecosystem processes in the site (Table 1). Sites ranged in size from 10 to mally by human impacts. Present-day eco- RNAs are within the HRV. 20 ha and were selected based upon the systems useful for defining reference con- presence of adequate old-age fire-scarred ditions may be found in wilderness areas, METHODS trees or remnant (dead) material. The goal of many National Park Service landscapes, collection at each site was to obtain a and Forest Service Research Natural Areas Study Areas comprehensive, long-term inventory of past (RNAs; e.g., Ryan et al. 1994). RNAs are fires using fire scar records from individual permanently protected landscapes that have Four study areas were selected in South trees (Brown and Sieg 1996, Swetnam and been set aside principally for the purposes Dakota, Wyoming, and Colorado. These Baisan 1996). Long-term sequences of fire of maintaining certain aspects of biological included three established RNAs—Upper scars often are recorded on individual trees. diversity and conducting Pine Creek in the Black Hills National However, individual trees may be missing non-manipulative ecological research Forest of South Dakota (Ryan et al. 1994), fire records because not all fires that (Ryan et al. 1994). Lone Pine in the Arapahoe-Roosevelt Na- burned in the vicinity of a tree may be tional Forest in northern Colorado, and recorded as scars, and scars may be burned While both of these concepts, HRV and Hot Creek in the Rio Grande National off by subsequent fires or other-wise eroded reference conditions, are crucial for scien- Forest in southern Colorado—and one from the scar record. tifically based ecosystem management pro-posed RNA, Ashenfelder Basin in the (Kaufmann et al. 1994, Holling and Meffe Medicine Bow National Forest in central Upper Pine Creek was established as an 1996, Leslie et al. 1996), reference land- Wyoming (Table 1). Each of the RNAs RNA in 1931, the first in this region of the scapes can be used to characterize refer- contained unroaded, old-growth pondero- U.S. Forest Service (Ryan et al. 1994). The ence conditions only if one uses a sa pine stands that have had a history of 482-ha RNA is in the rocky, isolated center long-term perspective to a sse ss whether little or no logging or other significant range of the Black Hills and contained present-day conditions in these landscapes human disturbances during the twentieth within the Black Elk Wilderness Area. As represent predominately natural rather than century. Native American impacts before in many areas of the Black Hills, ponderosa human factors. Probably no ecosystem on the twentieth century probably occurred pine regeneration in Upper Pine Creek has Earth has escaped at least some minimal to some extent in each of the areas. been heavy during the twentieth century, level of impact from industrial society How-ever, the RNAs are for the most part and there are many areas of dense, often (Swanson et al. 1993, Kaufmann et at. located in areas with difficult access and suppressed trees (McAdams 1995). 1994, Vitousek et al. 1997). Although log- each is ging, road construction, or other landscape fragmentation may not have occurred in a reference landscape with a history of sur- Table 1. Fire history study sites in Research Natural Areas in South Dakota, Wyoming, and face fires, fire cessation in adjacent land- Colorado. scapes may have excluded landscape-scale fires from the reference area. In addition, Site Elevation Latitude (N)/ active fire suppression by land manage- Code Range (m) Longitude (W) ment agencies has occurred in virtually all Site Name ecosystems of the western United States. Exclusion of surface fires from a reference Upper Pine Creek, S.D. UPC 1660–1690 43° 53.0' / 103° 30.0' area may result in changes in ecosystem Upper Pine Creek Middle, S.D. UPM 1670–1720 43° 52.5' / 103° 30.5' structure and function that are as pro- Ashenfelder Basin Lower, Wyo. ASL 1920–1950 42° 20.0' / 105° 23.0 nounced as those in more directly impact- Ashenfelder Basin Upper, Wyo.