Western Pine Forests with Continuing Frequent Fire Regimes: Possible Reference …nature.berkeley.edu/stephenslab/wp-content/uploads/2015/... · 2015. 5. 22. · Rollins et al. 2002

Western Pine Forests withContinuing Frequent Fire Regimes:Possible Reference Sites forManagement

Scott L. Stephens and Peter Z. Fule

In contrast to a few isolated forests in northern Mexico, most forests in the western Untied States havebeen significantly modified by fire suppression, harvesting, and livestock grazing. This has increasedtheir fire hazards and many are in need of restoration. Understanding reference conditions ischallenging because we have few intact forests functioning under the continuing influence of climatevariation, insects, diseases, and frequent fires. This article summarizes information from reference sitesand argues for incorporating natural heterogeneity in restoration targets across similar forests in theUnited States.

Keywords: forest restoration, desired conditions, wildfire, Jeffrey pine, ponderosa pine, Mexico

T o describe reference conditions insurface-fire–adapted forests, man-agers increasingly rely on photo-

graphic evidence, written historical ac-counts, early forest inventory data, andstudies that reconstruct pre–Euro-Americanconditions from living and dead woody ma-terial. In doing so, often, it is assumed thatthese historical conditions would prevail to-day in the absence of fire exclusion. How-ever, the interaction of climatic variation,forests, and fire is complex, making it uncer-tain if similar prehistorical conditions wouldhave occurred in the present environment(Millar and Wolfenden 1999, Stephens et al.2003). As a result, there are few forests inwestern North America that could serve asmodels or “controls” for forests functioningunder the continuing influence of climatevariation, insects, diseases, and frequentfires.

Conifer forests in isolated ranges of

northern Mexico have not experienced sys-tematic fire suppression and some have notbeen harvested (Table 1). Although mosthave been grazed to varying levels of inten-sity, the absence of systematic fire suppres-sion suggests these forests may provide in-formation useful for developing descriptionsof reference conditions for similar forests inthe southwestern United States and else-where. A handful of US forests, mostly re-mote sites in National Parks or WildernessAreas, also exhibit relatively undisrupteddisturbance regions (Table 1).

In this article we summarize informa-tion on pine-dominated forests in south-western North America (northwestern Mex-ico and southwestern United States) thathave relatively intact disturbance regimes.This information could be used by managersinterested in restoring similar forests in theregion, particularly in western forests thathave experienced a significant increase in

area burned over the last 60 years (Stephens2005). To enhance the clarity of the com-parisons, we present information frompaired fire/no-fire sites or others that arecross-border comparisons. This approachattempts to isolate the effect of fire manage-ment on these ecosystems. We focus espe-cially on our research in Baja California andthe Grand Canyon, Arizona.

Reference AreasReference or “relict” sites are rare be-

cause most landscapes have been impactedheavily by industrial human society in thepast 100–150 years. The term “relict” refersto isolated places that were unimpacted orless impacted, although it does not meanthat these ecosystems are “museum pieces”maintained in a constant state. Grazing andtimber harvesting have been nearly ubiqui-tous in northwestern Mexico, especiallysince the creation of communal landhold-ings called ejidos in the early 20th century(Heyerdahl and Alvarado 2003). However,organized fire suppression has not been a na-tional priority in Mexico. As a result, fre-quent fire regimes persisted into the late20th centuries in a few protected areas suchas the Sierra San Pedro Martir (SSPM), BajaCalifornia, and La Michilıa Biosphere Re-serve, Durango (Table 1; Figure 1).

Jeffrey pine (Pinus jeffreyi)–mixed coni-fer forests have been studied in the SSPM, anisolated mountain range located approxi-

Journal of Forestry • October/November 2005 357

AB

ST

RA

CT

mately 85 mi southeast of Ensenada, BajaCalifornia, Mexico. Vegetation of the SSPMis composed of conifer forests and shrub-lands of the Californian floristic provincethat occur in no other area of Mexico (Min-nich and Franco 1998). The SSPM is the

southern terminus of the Peninsular Moun-tain Range that begins at the boundary be-tween the San Jacinto and San BernardinoMountains in California; approximately215 mi separates the SSPM from the SanBernardino Mountains and the floras andsoils in these areas are very similar.

The mixed conifer forests of the SSPMhave experienced neither tree harvesting nora policy of large-scale fire suppression. Lim-ited fire suppression began in the 1970s buthas only consisted of one or two four-personhand crews in the summer and fall periods(Stephens et al. 2003). The climate in theSSPM is Mediterranean with annual precip-itation of approximately 22 in. (Minnich etal. 2000). Median fire return intervals in for-ests in the SSPM are shorter than 15 years(Table 1); similar fire return intervals onceoccurred in the Jeffrey pine and mixed coni-fer forests in the eastern Sierra Nevada andSan Bernardino Mountains before fire sup-

pression was initiated early in the 20th cen-tury (Stephens et al. 2003). SSPM fire sea-son is approximately June–early Aug.,unlike most California forests that were oncedominated by late season or dormant seasonfires (Stephens et al. 2003).

In the United States, true relicts includephysically inaccessible areas, such as steepmesas at Zion National Park (Madany andWest 1983) or forest patches isolated bysharp lava flows at El Malpais, New Mexico(Grissino-Mayer and Swetnam 1997). Al-though protected from human impacts,these unusual sites may be inherently differ-ent from larger forestlands. For example,one Zion site had fire-free intervals muchlonger than those at any other ponderosapine (Pinus ponderosa) reference site, per-haps because the small mesa had few ignitionopportunities (Table 1). The larger pro-tected landscapes in the southwesternUnited States, such as Grand Canyon and

Figure 1. Jeffrey pine–mixed conifer forestsin the SSPM, Baja California, Mexico. Highheterogeneity in forest structure is commonin this ecosystem. (Photo by Scott Stephens.)

Table 1. Fire-adapted reference sites in forests in northwestern Mexico (Jeffrey pine-mixed conifer and Durango pine) and thesouthwestern United States (ponderosa pine).

Site Management status Fire regime Forest attributes studied Reference

SSPM, Baja California, Mexico National Park MFI � 3.9–22 yr (range ofvalues at two study sites inthe 18th–20th centuries);last fire �1970

Forest structure, treemortality, snag density,fuel loads, regeneration,fire-climate relationships,soil chemistry andstructure, grazing history

Stephens et al. 2003,Stephens 2004, Stephensand Gill 2005, Stephensand Fry 2005, Minnich etal. 1995, Minnich andFranco 1998

Topia (Arroyo Laureles),Durango, Mexico

Communally owned (ejido)forest

MFI � 4.7 yr; last fire, 1991(site was harvested afterstudy).

Forest structure and spatialpattern, fuels

Fule and Covington 1997

La Michilia (Playa Grande),Durango, Mexico

Biosphere reserve MFI � 7.9 yr; last fire, 1998 Grazing history Fule and Covington 1999

Sierra Madre Occidental,Durango, Mexico

Communally owned (ejido)forest

MFI � 3–6 yr (range ofvalues at four sites); lastfires, �1970–1990

Land tenure changes Heyerdahl and Alvarado2003

Sierra Los Ajos, Sonora, Mexico Communally owned (ejido)forest

MFI � 4.0–5.9 yr; last fires,1916–1972

Forest structure, soilchemistry

Swetnam et al. 2001,Villanueva-Diaz andMcPherson 2002

Tutuaca, Chihuahua, Mexico Conservation reserve on ejidolands

MFI � 3.9–5.2 yr (range ofvalues at two study sites);last fire, 1995

Fire-climate relationship Fulé et al. 2005

Gila Wilderness, NM Wilderness area (GilaNational Forest)

47,500 ac burned at leastonce in 20th century;24,400 ac burned at leasttwice

Spatial patterns of fire,landscape ecology

Rollins et al. 2002

Grand Canyon (Powell Plateau,Fire Point, Rainbow Plateau),AZ

National Park MFI � 3.2 yr (PowellPlateau), 3.7 yr (FirePoint), 4.0 yr (RainbowPlateau) through 1879;several large surface fires ateach site after 1879

Forest structure, composition,regeneration, plantcommunity response towildfire, comparison offire-scar data withhistorical records, fire-climate relationship

Fule et al. 2002, 2003, Gildaret al. 2004

Zion National Park (ChurchMesa, Greatheart Mesa), UT

National park MFI � 69 yr on ChurchMesa

Tree regeneration, grass Madany and West 1983

El Malpais (Mesita Blanca,Hidden Kipuka), NM

National monument WMPIa � 8.6 yr (MesitaBlanca), 13.3 yr (HiddenKipuka)

Grazing history Grissino-Mayer and Swetnam1997

a WMPI, Weibull median fire interval, an alternative to MFI for describing average fire return intervals. WMPI and MFI values are usually nearly equal.Fire regimes were documented from fire-scarred trees and historical records at all sites (MFI). The MFI values shown include all fires; the intervals between larger fires (those that scarred 25% or moreof the samples) usually average two to five times longer. Other forest attributes also were studied at most sites.MFI, mean fire interval; WMPI, Weibull median fire interval.

358 Journal of Forestry • October/November 2005

the Gila Wilderness, were grazed by live-stock and most fires were suppressed untilnatural fire policies were implemented in the1970s. However, even with some disruptionto the fire regime, these quasi-referencelandscapes conserve features such as oldtrees, native species, and resilience to severefires.

At Grand Canyon National Park innorthern Arizona, fires ceased in ponderosapine forests as early as 1880 because of live-stock grazing, but occasional fires continuedafter 1880 in isolated areas near the canyonrim. Although the post-1880 fire regimechanged from about two dozen fires per cen-tury to only a few, the forests were neverharvested and livestock grazing was elimi-nated in the 1930s. Fule et al. (2002) arguedthat evidence from forest structure, fuelloads, and native plant communities (Gildaret al. 2004) indicated that the relict GrandCanyon forests remained relatively similarto their pre–Euro-American conditions (seeFigure 3). The climate in this region is mon-soonal.

Forest Structure, Variability,and Species Composition

Forests in the SSPM are composed ofJeffrey pine, white fir (Abies concolor), sugarpine (Pinus lambertiana), lodgepole pine(Pinus contorta var. murrayana), and limitedamounts of incense-cedar (Calocedrus decur-rens) and quaking aspen (Populas tremu-loides; Figure 1). The most common foresttypes are Jeffrey pine, Jeffrey pine–mixed co-nifer, and mixed white fir forests, respec-tively (Minnich and Franco 1998). Soils inour sampled areas are Entisols (derived fromdiorite parent materials) that are shallow,well to excessively drained, and relativelyacidic (Stephens and Gill 2005). Our re-search is located on the northern plateau ofthe SSPM at an approximately 8,500-ft ele-vation.

Forest structure has been inventoriedusing a systematic grid of plots (Stephens2004, Stephens and Gill 2005). High vari-ability characterized all live tree, snag, andfuel attributes measured in the forests of the

SSPM. This high amount of variability isprobably the result of the relatively intactfrequent surface fire regime and because noharvesting has ever occurred in this forest.High variability in surface fuel loads wouldproduce equally diverse fire behavior and ef-fects, and this would maintain high spatialheterogeneity if the forest continues to burnunder a frequent fire regime (Stephens2004).

Examples of the high variation in foreststructure are summarized in the followingparagraphs from an area of approximately500 ac (Table 2). In this area, average dbh inJeffrey pine–mixed conifer forests was 13 in.(range, 1–44 in.; Stephens and Gill [2005]).Average tree density was 59 trees � ac�1

(range, 12–130 trees ac). Average basal areawas 86 ft2 � ac�1 (range, 25–220 ft2 � ac�1).Approximately 20% of the sampled plotshad a basal area below 60 ft2 � ac�1, 60% ofplots had basal areas from 60 to 95 ft2 � ac�1,and 20% of plots had basal areas of 95–220ft2 � ac�1. Average canopy cover was 25.3%(range, 14–49.5%; Stephens and Gill

Table 2. Comparison of forest structural data from paired reference/nonreference sites.

Site Type of site Living tree structure Dead tree structure Forest floor fuels

SSPM, Baja California, Mexicoa Reference Average tree densities ranged from12 to 130 trees � ac�1 for alltrees �1-in. dbh; average basalarea ranged from 25 to 220 ft2 �ac�1

Density of snags � 1-in. dbhranged from 0 to 10.1 snags �ac�1, 26% of sampled plotshad no snags; average snagdensity 2.06 ac�1

Large woody fuels (1000-hr fuels,�3-in. diameter) ranged from 0to 67 tons � ac�1, total woodyfuels varied from 0.01 to 69tons � ac�1

San Bernadino Mountains,Californiab

Nonreference Average tree density 68 trees � ac�1

in 1992 for all trees �4-in. dbh,which represented a 79%increase in density whencompared with a 1932inventory

Density of snags in some areasaveraged 50 trees � ac�1 frommultiyear drought that endedin 2004

NA

Grand Canyon (Powell Plateau,Fire Point, Rainbow Plateau),AZc

Reference Average pine densities rangedfrom 85 to 101 trees � ac�1 anddeciduous species varied from24 trees � ac�1 (N.M. locust atPowell Plateau) to as many as264 trees � ac�1 (Gambel oak atRainbow Plateau)

Density of snagse �19.7-in. dbhranged from 1.6 to 2.8 trees �ac�1

Rotten woody fuelse (1000-hrfuels, �3-in. diameter) rangedfrom 0.7 to 2.8 tons � ac�1,total woody fuels from 8.2 to20.8 tons � ac�1

Grand Canyon (Grandview)c Nonreference Average pine density 261 trees �ac�1, average oak density 118trees � ac�1

Density of snagse �19.7-in. dbhaveraged 1.0 trees � ac�1

Rotten woody fuelse (1000-hrfuels, �3-in. diameter) averaged1.6 tons � ac�1, total woodyfuels averaged 7.6 tons � ac�1

Topia (Arroyo Laureles), Durango,Mexicod

Reference Average pine density 111 trees �ac�1, average density of otherspecies (mostly oaks) 151 trees �ac�1

Density of all snags � breastheight, 61 trees � ac�1


Topia (Arroyo Verde), Durango,Mexicod

Nonreference Average pine density 607 trees �ac�1, average density of otherspecies (mostly oaks) 500 trees �ac�1

Density of all snags � breastheight, 71 trees � ac�1


a Stephens (2004); Stephens and Gill (2005).b Minnich et al. (1995); Sims (2004).c Fule et al. (2002).d Fule and Covington (1997).e P.Z. Fule, unpublished data.NA, not available.


[2005]) and the amount of area in regener-ation patches was 3.8% (Stephens and Fry2005).

Average surface fuel loads were 6.6 tons� ac�1 (range, 0.01–69 tons � ac�1; Table 2;Stephens [2004]). Total surface fuel loadwas less than the average load in 73% of thesampled area. Surface fuel load was greaterthan 8 tons � ac�1 on 24% of plots andgreater than 16 tons � ac�1 on 8% of plots.Average 1,000-hour fuel load was 5.9 tons �ac�1 (Stephens 2004). Thirty-seven percentof plots had no 1,000-hour fuels and 67%had less than the average 1,000-hour load.Most structural attributes measured in theSSPM varied by approximately one order ofmagnitude even though the area sampledhas similar aspects, slopes, soils, and domi-nant tree species.

Average snag density in 2003 was 2.06snags � ac�1 (range, 0–10.1 snags � ac�1; Ta-ble 2) and 26% of sampled plots had nosnags (Stephens 2004). Mortality during thesevere drought from 1999 to 2003 increasedsnag density by approximately 0.5 snags �ac�1. Less than the average snag density(2.06 snags � ac�1) was recorded on 71% ofthe sampled area. The foregoing data under-score the high degree of variability for arange of stand parameters in relict forests innorthern Mexico. Mixed conifer forests inthe San Bernardino Mountains of southernCalifornia also have experienced increasedmortality during a similar drought (1999–2003), but the increase in snag density wasmuch greater, estimated at approximately 50trees � ac�1 in many areas (Sims [2004]; Ta-ble 2; Figure 2). Fire suppression has in-creased tree densities in this area, and whenthis was coupled to a multiyear drought, na-tive bark beetles killed millions of trees(Eatough-Jones et al. 2004). In contrast,similar bark beetle–induced tree mortalitydid not occur in the SSPM during the1999–2003 drought (an average of 0.5 newsnags � ac�1 was produced in the SSPM dur-ing the 1999–2003 drought; Stephens[2004]).

Analysis of relict and nonreference sitesin the United States affords informationabout potential effects of management (es-pecially fire suppression) on stand character-istics. In 1932 mixed Jeffrey pine forests inthe San Bernardino Mountains of southernCalifornia had an average density of 38 trees� ac�1 (including trees with more than 5-in.dbh; Minnich et al. [1995]). Tree densityfrom the SSPM for all trees with more than4-in. dbh was 44 trees � ac�1, which is simi-

lar to that reported in the San BernardinoMountains before large-scale fire suppres-sion or harvesting. In 1992, mixed Jeffreypine forests in the San Bernardino Moun-tains had tree densities 79% larger thanthose in the early 1930s mainly because offire suppression (Table 2; Minnich et al.[1995]).

Ponderosa pine forests at relict GrandCanyon sites retain an open stand structure,resistant to stand-replacing fire, while exhib-iting high variability especially in the decid-uous forest component (Table 2; Figure 3).These forests are located on high plateaus atan approximately 7,500-ft elevation withwell-drained soils derived from Kaibab lime-stone. Gambel oak (Quercus gambellii) andNew Mexican locust (Robinia neomexicana)regenerate effectively by sprouting after fire,creating a shrubby midstory in some re-gions. Average pine densities ranged from 85to 101 trees � ac�1 and deciduous speciesvaried from 24 trees � ac�1 (N.M. locust atPowell Plateau) to as many as 264 trees �ac�1 (Gambel oak at Rainbow Plateau; Fuleet al. [2002]). On individual plots, oak den-sities reached over 1,000 trees � ac�1 (Fule etal. 2002).

The contribution of deciduous trees tothe forest basal area at the relic Grand Can-yon sites is much less—oaks and locust av-

eraged 4–24 ft2 � ac�1, compared with 95–133 ft2 � ac�1 for ponderosa pine and thepines comprise the great majority of canopyfuels—but the structural heterogeneity ofthe broad-leaved trees is likely very impor-tant for wildlife habitat and plant diversity(Table 2; Gildar et al. [2004] and Laughlinet al. [2004]). Canopy cover ranged from 48to 52% (Fule et al. 2002). Surface fuel loads(all woody fuels) ranged from 8.2 to 20.8tons/ac; rotten coarse woody debris (1,000-hour fuels) made up only 8–36% of the to-tal, likely because of repeated fires that fore-stalled the development of large quantities ofrotten downed wood (P.Z. Fule, unpub-lished data).

For comparison to nonreference condi-tions, at a paired site in the park where fireexclusion had occurred since 1887 (Grand-view), pine density averaged 261 trees � ac�1

and oaks averaged 118 trees � ac�1 (Table 2;Fule et al. [2002]). Pine density at Grand-view was similar to the overall average pon-derosa density for the state of Arizona, 250trees � ac�1 (O’Brien 2002). Basal area atGrandview was less than at any of the refer-ence sites, however (average, 73.3 ft2 � ac�1

for pine), indicating that most pines were ofsmall size.

Modern Fire Behavior inReference Sites

Are severe fires in recent years primarilya function of increased fuel or of worseningfire weather? Modern fires burn differentlyin reference sites than in neighboring fire-excluded forests, offering anecdotal insightinto this question, although the number ofexamples is very limited. In June 2003, theAspen wildfire burned with stand-replacingintensity near Tucson, Arizona (not a refer-

Figure 2. Tree mortality in mixed coniferforests in the San Bernardino Mountains ofsouthern California influenced by past for-est management, native bark beetles, and a5-year drought. Tree density shown is sim-ilar to that of many fire-excluded forests.(Photo by Scott Stephens.)

Figure 3. Open forest dominated by maturetrees at Powell Plateau, Grand Canyon Na-tional Park, is resistant to intense fire be-havior because of light surface fuels andhigh canopy base height. (Photo by PeterFule.)


ence site), whereas a wildland-use fire onPowell Plateau, Grand Canyon, Arizona (areference site) burned with low intensity(Figure 4). The Aspen fire burned over84,000 ac, destroying 333 structures. Over1,200 fire personnel and 14 aircraft wereused in the firefighting effort, for a total sup-pression cost exceeding $16,000,000 (datafrom Coronado National Forest). In con-trast, the 3,500-ac Powell fire damaged noresources and was managed by a staff ofmore than 40 with one helicopter (data fromGrand Canyon National Park). Similarly, afire at La Michilıa Biosphere Reserve, Du-rango, during the severe El Nino fire year of1998 caused minimal tree mortality, but afirefighter was tragically killed in a needlesssuppression effort (Rodrıguez-Trejo andFule 2003).

Observations of reduced fire severity inreference sites are rare but similar findingscome from more extensive dendroecologicaland fire modeling studies. Compared withnearby fire-excluded forests, reference for-ests at Grand Canyon were less dense andhad the smallest 20th century increases incanopy biomass, canopy bulk density, andcrownfire hazard as measured by crowningindex (windspeed needed to maintain can-opy burning; Fule et al. [2004]). Fire mod-

eling showed that in 1880, only 6% of theentire landscape of North Rim of GrandCanyon was vulnerable to stand-replacingfire at a windspeed of 28 mi � hr�1. By 2000,increased forest density in ponderosa andhigher-elevation forests outside the relict ar-eas caused 33% of the landscape to be sus-ceptible to stand-replacing fire at the samewindspeed threshold (Fule et al. 2004).

Management ImplicationsIn many areas of the western United

States with ponderosa pine, Jeffrey pine, andmixed conifer forests, past harvesting has re-moved the largest trees and fire suppressionand livestock grazing have increased under-story tree density and fuel loads. Informationfrom forests with relatively intact disturbanceregimes can assist in the development of de-sired conditions in some US forests. Referenceforests are especially useful because they havebeen functioning under the continuing in-fluence of climate variation, insects, diseases,and frequent fires. These reference sites areincredibly valuable and should be protectedin the future. At the same time, it is impor-tant to recognize the limitation of relict areasfor making broader inferences. Often, thesesites occur in areas where biophysical condi-tions are somewhat different from the dom-inant areas (steep slopes, isolated areas, andfire seasonality), so fire history and effectsalso may be different from historical patternsin target management areas. Other factorsinclude the relatively small size of most re-licts, affecting landscape-scale processes, andthe lack of appropriate relicts for many foresttypes.

The patchy distribution of forest struc-ture observed from the forests in the SSPM,Grand Canyon, and elsewhere arguesagainst the application of uniform targets forsnag retention, fuels, and live trees acrosssimilar forest landscapes. An improvementin management guidelines would be to man-age for forest structures over moderate spa-tial scales (hundreds to thousands of acres)instead of on a stand (10–20 ac) basis (Ste-phens 2004). An example of fine-scaled for-est restoration was undertaken by Coving-ton et al. (1997), where the pre–fire-exclusion structure of a ponderosa pineforest was used as a template to guide thin-ning. Flexible and heterogeneous ap-proaches would enable managers to meetmanagement standards by maintaining highsnag densities and large fuel loads over only aportion of the forested landscape. Thiswould require flexibility in the application

of restoration treatments, something that israre in US federal land management (Ste-phens and Ruth 2005). It is important torestrict our interpretation of these findingsto areas with similar species, disturbance re-gimes, and abiotic environments (Dibbleand Rees 2005), because forests have a greatdeal of variation.

Because of forest variation and our in-sufficient knowledge of the ecological, eco-nomic, and social effects of different restora-tion treatments that can be used to reducefire hazards in western US forests (Stephensand Moghaddas 2005), we need to developvigorous adaptive management programs.Such programs would use a rigorous systemof monitoring, experimentation, and re-search to improve fire and fuels managementpolicies, strategies, and projects (Stephensand Ruth 2005). Information from the ref-erence forests in northern Mexico and thesouthwestern United States can assist in thedevelopment of desired conditions and willallow scientists to understand how intactforests will respond to future conditions.

Literature CitedCOVINGTON, W.W., P.Z. FULE, M.M. MOORE,

S.C. HART, T.E. KOLB, J.N. MAST, S.S. SACK-ETT, AND M.R. WAGNER. 1997. Restoration ofecosystem health in southwestern ponderosapine forests. J. For. 95(4):23–29.

DIBBLE, A.C., AND C.A. REES. 2005. Does thelack of reference ecosystems limit our science?A case study in nonnative invasive plants asforest fuels. J. For. 103(7):329–338.

EATOUGH-JONES, M., T.D. PAINE, M.E. FENN,AND M.A. POTH. 2004. Influence of ozone andnitrogen deposition on bark beetle activity un-der drought conditions. For. Ecol. Manage.200:67–76.

FULE, P.Z., AND W.W. COVINGTON. 1997. Fireregimes and forest structure in the Sierra Ma-dre Occidental, Durango, Mexico. Acta Bot.Mex. 41:43–79.

FULE, P.Z., AND W.W. COVINGTON. 1999. Fireregime changes in La Michilıa Biosphere Re-serve, Durango, Mexico. Conserv. Biol.13:640–652.

FULE, P.Z., W.W. COVINGTON, M.M. MOORE,T.A. HEINLEIN, AND A.E.M. WALTZ. 2002.Natural variability in forests of Grand Canyon,USA. J. Biogeogr. 29:31–47.

FULE, P.Z., T.A. HEINLEIN, W.W. COVINGTON,AND M.M. MOORE. 2003. Assessing fire re-gimes on Grand Canyon landscapes with firescar and fire record data. Int. J. Wildl. Fire 12:129–145.

FULE, P.Z., J.E. CROUSE, A.E. COCKE, M.M.MOORE, AND W.W. COVINGTON. 2004.Changes in canopy fuels and potential fire be-havior 1880–2040: Grand Canyon, Arizona.Ecol. Model. 175:231–248.

Figure 4. Two fires burning in Arizona pon-derosa pine forests in June 2003. (Top) thePowell fire in Grand Canyon National Park(photo by National Park Service). (Bottom)the Aspen fire, Coronado National Forest(photo Arizona Daily Star, Azstarnet.com).


FULE, P.Z., J. VILLANUEVA-DıAZ, AND M. RAMOS-GOMEZ. 2005. Fire regime in a conservationreserve, Chihuahua, Mexico. Can. J. For. Res.35:320–330.

GILDAR, C.N., P.Z. FULE, AND W.W. COVING-TON. 2004. Plant community variability inponderosa pine forest has implications for ref-erence conditions. Nat. Areas J. 24:101–111.

GRISSINO-MAYER, H.D., AND T.W. SWETNAM.1997. Multi-century history of wildfire in theponderosa pine forests of El Malpais NationalMonument. New Mex. Bureau Mines MineralResourc. Bull. 156:163–171.

HEYERDAHL, E.K., AND E. ALVARADO. 2003. In-fluence of climate and land-use on historicalsurface fires in pine-oak forests, Sierra MadreOccidental, Mexico. P. 196–217 in Fire andclimatic change in temperate ecosystems of thewestern Americas, Veblen, T.T., W.L. Baker,G. Montenegro, and T.W. Swetnam (eds.).Springer-Verlag, New York.

LAUGHLIN, D.C., J.D. BAKKER, M.T. STODDARD,M.L. DANIELS, J.D. SPRINGER, C.N. GILDAR,A.M. GREEN, AND W.W. COVINGTON. 2004.Toward reference conditions: Wildfire effectson flora in an old-growth ponderosa pine for-est. For. Ecol. Manage. 199:137–152.

MADANY, M.H., AND N.E. WEST. 1983. Live-stock grazing—fire regime interactions withinmontane forests of Zion National Park, Utah.Ecology 64:661–667.

MILLAR, C.I., AND W. WOLFENDEN. 1999. Therole of climate change in interpreting historicalvariability. Ecol. Applic. 9:1207–1216.

MINNICH, R.A., M.G. BARBOUR, J.H. BURK, AND

R.F. FERNAU. 1995. Sixty years of change inCalifornian conifer forests of the San Bernar-dino Mountains. Conserv. Biol. 9:902–914.

MINNICH, R.A., M.G. BARBOUR, J.H. BURK, AND

J. SOSA-RAMIREZ. 2000. Californian mixed-co-nifer forests under unmanaged fire regions in

the Sierra San Pedro Martir Baja, California,Mexico. J. Biogeograph. 27:105–129.

MINNICH, R.A., AND E. FRANCO. 1998. Land ofchamise and pines: Historical accounts andcurrent status of northern Baja California’svegetation. Univ. Calif. Publ. Bot. 80. 166 p.

O’BRIEN, R.A. 2002. Arizona’s forest resources,1999. USDA For. Serv. Resource Bull.RMRS-RB-2, Ogden.

RODRıGUEZ-TREJO, D.A., AND P.Z. FULE. 2003.Fire ecology of Mexican pines and a fire man-agement proposal. Int. J. Wildl. Fire 12:23–37.

ROLLINS, M.G., P. MORGAN, AND T. SWETNAM.2002. Landscape-scale controls over 20th cen-tury fire occurrence in two large Rocky Moun-tain (USA) wilderness areas. Landsc. Ecol. 17:539–557.

SIMS, T. 2004. Dead tree count totals in 2004. SanBernardino and San Jacinto Mountains Vege-tation Mortality Project, San Bernardino Na-tional Forest, CALMAST, and ESRI, Red-lands, CA.

STEPHENS, S.L. 2004. Fuel loads, snag density,and snag recruitment in an unmanaged Jeffreypine-mixed conifer forest in northwesternMexico. For. Ecol. Manage. 199:103–113.

STEPHENS, S.L. 2005. Forest fire causes and ex-tent on United States Forest Service lands. Int.J. Wildl. Fire 14:213–222.

STEPHENS, SL, C.N. SKINNER, AND S.J. GILL.2003. Dendrochronology-based fire history ofJeffrey pine-mixed conifer forests in the SierraSan Pedro Martir, Mexico. Can. J. For. Res.33:1090–1101.

STEPHENS, S.L., AND D.L. FRY. 2005. Spatial dis-tribution of regeneration patches in an old-growth Pinus jeffreyi-mixed conifer forest innorthwestern Mexico. Journal of VegetationScience (in press).

STEPHENS, S.L., AND S.J. GILL. 2005. Foreststructure and mortality in an old-growth Jef-

frey pine-mixed conifer forest in northwesternMexico. For. Ecol. Manage. 205:15–28.

STEPHENS, S.L., AND J.J. MOGHADDAS. 2005. Ex-perimental fuel treatment impacts on foreststructure, potential fuel behavior, and pre-dicted tree mortality in a California mixed co-nifer forest. For. Ecol. Manage. 215:21–36.

STEPHENS, S.L., AND L.W. RUTH. 2005. Federalforest fire policy in the United States. Ecol. Ap-plic. 15(2):532–542.

SWETNAM, T.W., AND J.L. BETANCOURT. 1998.Mesoscale disturbance and ecological responseto decadal climatic variability in the AmericanSouthwest. J. Clim. 11:3128–3147.

SWETNAM, T.W., C.H. BAISAN, AND J.M. KAIB.2001. Forest fire histories of the sky islands ofLa Frontera. P. 95–123 in Changing plant lifeof La Frontera, Webster, G.L., and C.J. Bahre(eds.). University of New Mexico Press, Albu-querque, NM.

VILLANUEVA DIAZ, J., AND G.R. MCPHERSON.2002. Estudios dendroclimaticos en montanasdel suroeste de los Estados Unidos de Americay del norte de Mexico [in Spanish]. Cien. For.Mex. 86(24).

Scott Stephens ([email protected])is assistant professor of fire science, Division ofEcosystem Science, Department of Environ-mental Science, Policy, and Management,University of California, Berkeley, CA 94720-3110. Peter Fule ([email protected]) is asso-ciate professor and associate director, School ofForestry, Ecological Restoration Institute,Northern Arizona University, Flagstaff, AZ86011. This research was funded by the Uni-versity of California Agricultural ExperimentStation, and the USDA–USDI Joint Fire Sci-ences Program.


Documents

Western Pine Forests with Continuing Frequent Fire Regimes: Possible Reference …nature.berkeley.edu/stephenslab/wp-content/uploads/2015/... · 2015. 5. 22. · Rollins et al. 2002