Attachment G Bunchgrass Communities-A.pdf

United States Department of Agriculture

Forest Service

Pacific Northwest Research Station

General Technical Report

PNW-GTR-641

August 2005

Bunchgrass Plant Communities of the Blue and Ochoco Mountains:

A Guide for Managers Charles Grier Johnson, Jr., and David K. Swanson

The Forest Service of the U.S. Department of Agriculture is dedicated to the principle of multiple use man-agement of the Nation’s forest resources for sustained yields of wood, water, forage, wildlife, and recreation.

Through forestry research, cooperation with the States and private forest owners, and management of the national forests and national grasslands, it strives—as directed by Congress—to provide increasingly greater service to a growing Nation.

The U.S. Department of Agriculture (USDA) prohibits discrimination in all its programs and activities on the basis of race, color, national origin, age, disability, and where applicable, sex, marital status, familial status, parental status, religion, sexual orientation, genetic information, political beliefs, reprisal, or because all or part of an individual’s income is derived from any public assistance program. (Not all prohibited bases apply to all programs.) Persons with disabilities who require alternative means for communication of program information (Braille, large print, audiotape, etc.) should contact USDA’s TARGET Center at (202) 720-2600 (voice and TDD). To file a complaint of discrimination, write USDA, Director, Office of Civil Rights, 1400 Independence Avenue, SW, Washington, DC 20250-9410 or call (800) 795-3272 (voice) or (202) 720-6382 (TDD). USDA is an equal opportunity provider and employer.

Authors Charles Grier Johnson, Jr. was the area ecologist (retired), and David K. Swanson is the area ecologist for the Malheur, Umatilla, and Wallowa-Whitman National Forests, 1550 Dewey Avenue, Baker City, OR 97814.

Cover photograph: Bluebunch wheatgrass near Bean Creek, Hells Canyon National Recreation Area.

Unless otherwise noted, all photographs were taken by Charles G. Johnson.

iii

Abstract Johnson, Charles Grier, Jr.; Swanson, David K. 2005. Bunchgrass plant communities of the Blue and Ochoco Mountains: a guide for managers. Gen. Tech. Rep. PNW-GTR-641.

Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. 119 p.

A classification of bunchgrass vegetation is presented for the Malheur, Ochoco, Umatilla, and part of the Wallowa-Whitman National Forests. It includes grassland vegetation as well as shrubland and forest land where the herbaceous layer is dominated by bunchgrasses. It is based on potential vegetation, with the plant association as the basic unit. Diagnostic keys and descriptions are presented for each type. Descriptions include information about plant species occurrence, environment and soils, states and transitions, forage productivity, management considerations, and relationships to other classifications.

Keywords: Blue Mountains, Ochoco Mountains, plant association, plant ecology, range management, steppe.

iv

Contents 1 Introduction 1 What Is a Bunchgrass?

1 History of Bunchgrasses and Livestock Grazing in the Pacific Northwest 2 Study Area 2 Physiography and Geology 2 Climate 4 Soils 4 Vegetation Overview 4 Plant Association Concept 5 State-and-Transition Concept 5 Phases and States 6 Fire Regimes and Fire Effects 6 Ponderosa Pine Forest 6 Juniper Woodlands 6 Shrublands 7 Grasslands 8 The Spread or Decline of Woody Plants 8 Newcomers to Bunchgrass Communities 9 Indicator Species 9 Sampling and Field Methods 10 Classification and Data Analysis Methods 10 Use of the Management Guide 10 Cover Percentages 10 Estimating Cover 11 Format of the Vegetation Type Descriptions 12 Keys for Determining the Bunchgrass Plant Associations and Plant Community Types 12 Limitations in the Use of the Keys 12 Rules for Use of the Keys 12 Key to Series 13 Key to Ponderosa Pine Series Vegetation 13 Key to Western Juniper Series Vegetation 14 Key to Shrubland Vegetation 16 Key to Grassland Vegetation 18 Ponderosa Pine Series 18 PIPO/CELE3/FEID-AGSP Plant Association (CPS234) 20 PIPO/CELE3/PONEW Plant Association (CPS233) 21 PIPO/PUTR2/FEID-AGSP Plant Association (CPS226) 22 PIPO/PUTR2/AGSP-POSA12 Plant Association (CPS229) 24 PIPO/ARTRV/FEID-AGSP Plant Association (CPS131) 26 PIPO/FEID Plant Association (CPG112) 28 PIPO/AGSP Plant Association (CPG111) 29 PIPO/PERA4 Plant Community Type (CPS8) 29 PIPO-JUOC/CELE3-SYOR2 Plant Community Type (CPC212) 30 Western Juniper Series 30 JUOC/CELE3/FEID-AGSP Plant Association (CJS41) 32 JUOC/PUTR2/FEID-AGSP Plant Association (CJS321) 34 JUOC/ARAR8/FEID Plant Association (CJS112) 35 JUOC/FEID-AGSP Plant Association (CJG111) 36 JUOC/ARRI2 Plant Community Type (CJS811) 36 JUOC/AGSP Plant Community Type (CJG113) v

37 Shrubland Series 37 CELE3/FEID-AGSP Plant Association (SD4111) 38 CELE3-PUTR2/AGSP Plant Community Type (SD4115) 38 CELE3/AGSP Plant Community Type (SD4112) 39 CELE3/PONEW Plant Community Type (SD4114) 39 PUTR2-ARTRV/FEID-AGSP Plant Association (SD3124) 41 PUTR2/FEID-AGSP Plant Association (SD3111) 42 PUTR2-ARTRV/FEID Plant Community Type (SD3125) 42 PUTR2/ERDO Plant Community Type (SD3126) 43 ARTRV/FEID-KOCR Plant Association (SD2929) 45 ARTRV/FEID-AGSP Plant Association (SD2911) 47 ARTRV/AGSP-POSA12 Plant Association (SD2918) 48 ARTRV-SYOR2/BRCA5 Plant Community Type (SD2917) 48 ARTRV-SYOR2 Plant Community Type (SD2919) 49 ARTRV/BRCA5 Plant Community Type (SS4914) 49 ARTRV-PERA4 Plant Community Type (SD3010) 49 ARTRV/ELCI2 Plant Community Type (SD3011) 50 ARAR8/FEID-AGSP Plant Association (SD1911) 52 ARAR8/AGSP Plant Association (SD1924) 54 ARAR8/POSA12 Plant Association (SD9221) 55 ARRI2/POSA12 Plant Association (SD9111) 57 ARRI2/PEGA Plant Community Type (SD9141) 57 ARTRT4/POSA12-DAUN Plant Community Type (SD2401) 57 SYOR2 Plant Community Type (SM32) 57 BERE/AGSP-APAN2 Plant Community Type (GB4915) 58 Grassland Series 58 Green Fescue Communities (FEVI-LULA3, FEVI-JUPA, FEVI-PENST) 58 FEID-GETR Plant Community Type (GB5923) 59 FEID-AGSP-CYTEF Plant Association (GB5925) 60 FEID-AGSP-FRALC2 Plant Association (GB5926) 61 MEBU-STOC2 Plant Community Type (GB5011) 61 STOC2-SIHY (Alpine) Plant Community Type (GS50) 62 FEID-KOCR (Mound) Plant Association (GB5912) 64 FEID-KOCR (Ridge) Plant Association (GB5911) 66 FEID-AGSP-LUPIN Plant Association (GB5916) 67 FEID-AGSP-BASA3 Plant Association (GB5917) 69 FEID-AGSP-PHLOX Plant Association (GB5931) 71 FEID-AGSP Plant Association (GB59) 72 FEID-DAUN Plant Community Type (GB5932) 73 AGSP-POSA12-LUPIN Plant Association (GB4119) 75 AGSP-POSA12-BASA3 Plant Association (GB4123) 77 AGSP-POSA12-ERHE2 Plant Association (GB4124) 78 AGSP-POSA12-ASRE5 Plant Association (GB4125) 80 AGSP-POSA12-TRMA3 Plant Association (GB4126) 82 AGSP-POSA12-SCAN3 Plant Association (GB4112) 83 AGSP-POSA12-APAN2 Plant Association (GB4127) 84 AGSP-POSA12-DAUN Plant Association (GB4911) 86 AGSP-POSA12 Plant Association (GB4121) 87 AGSP-BRCA5 Plant Community Type (GB4131) 88 AGSP-ERUM Plant Community Type (GB4132) 88 AGSP-CYTEF Plant Community Type (GB4133) 88 ERDO-POSA12 Plant Community Type (FM9111) 88 POBU-MAGL2 Plant Community Type (GB4411) vi

89 DAUN-LOLE2 Plant Association (GB9114) 91 POSA12-DAUN Plant Association (GB9111) 93 Glossary 94 Acknowledgments 94 Metric and English Equivalents 94 References 98 Appendix A: Plants Listed by Life Form and Scientific Name 108 Appendix B: Plants Listed by Common Name 118 Appendix C: Vegetation Types by Ecoclass Code

Bunchgrass Plant Communities of the Blue and Ochoco Mountains: A Guide for Managers

Introduction This management guide is provided for use in determin-ing the bunchgrass plant associations and plant communi-ty types encountered on lands administered by the Malheur, Ochoco, Umatilla, and Wallowa-Whitman National Forests (NF) within the Pacific Northwest Region (Region 6). Clas-sification of new plots along with bunchgrass plots includ-ed in Plant Associations of the Blue and Ochoco Mountains (Johnson and Clausnitzer 1992) provide the basis of this study. The purpose of this publication is to provide the field investigator with floristic indicator species that are diagnos-tic of a particular environment and potential vegetation.

Successional relationships and disturbance effects are portrayed by using the state-and-transition approach (Westoby et al. 1989), following the current trend in scien-tific thinking. This way of describing vegetation change on rangelands has largely replaced the successional model, in which ecological status was described as late seral, mid ser-al, or early seral based on an assumed successional path-way toward potential natural vegetation. The successional model in turn provided an alternative to the earlier condi-tion classes of “good,” “fair,” and “poor” based on suitabil-ity for domestic livestock grazing. To assist in conversion to the state-and-transition model, the user might consider the phases in a state-and-transition model as analogous to ser-al stages or condition classes, and the transitions as multiple pathways between them resulting from succession or distur-bance by grazing, fire, and so forth.

What Is a Bunchgrass?

Grasses belong to the grass family Poaceae (Gramineae).

Bunchgrasses are those grasses that have the characteristic growth habit of a clump, tuft, or bunch, in contrast to sod-forming grasses that spread via stolons or rhizomes. Bunch-grasses have fibrous root systems that occupy a belowground area approximating the aboveground foliar canopy. Studies by John Ernst Weaver (1915) gave us the first overall picture of the bunchgrass plant—above and below ground—by ex-huming them carefully. The rooting depths of our principal bunchgrasses are linked to the environments they occupy.

Following are key bunchgrasses (see apps. A and B for a full list of scientific and common names of species) and a description of their adaptations to some of the harshest sites in eastern Oregon:

Sandberg’s bluegrass (Poa sandbergii, syn. P. secunda)— This plant relies on lateral roots (3 to 5 in long) to utilize surface soil moisture, with deep roots extending down only 8 in. The root structure explains the plant’s ability to occu-py droughty scabland sites. Surface soil moisture is depleted more rapidly than moisture in deeper layers, and thus Sand-berg’s bluegrass usually matures and dries up earlier than more deep-rooted bunchgrasses (Sauer and Uresk 1976).

Short stature and a brief green period make this grass highly resistant to grazing.

Idaho fescue (Festuca idahoensis)—This deep-rooted bunchgrass extends to an average depth of 18 in. Because this is not a great depth for bunchgrass, Idaho fescue re-quires relatively moist sites.

Prairie junegrass (Koeleria cristata syn. K. macrantha)— Similar to Idaho fescue in site requirements, the roots of this grass extend downward about 15 in.

Bluebunch wheatgrass (Agropyron spicatum, syn.

Pseudoroegenaria spicata)—This tenacious bunchgrass has roots that extend to an average depth of 4 ft, allowing it to occupy some of the driest and hottest sites in eastern Oregon. Bluebunch wheatgrass is a very nutritious grass that is relished by livestock and wildlife. It is, however, easily damaged if grazed heavily during its period of active growth (Blaisdell and Pechanec 1949, Miller et al. 1986).

Most bunchgrasses in our region are cool-season grasses;

they accomplish most of their growth during cooler, moist-er weather before the summer drought. Most cool-season grasses are able to grow in the fall if rains occur, go dor-mant during the winter, then resume growth in early spring.

They complete their yearly growth and go to seed before the extreme drought of late summer begins. Warm-season grasses, which grow mainly during midsummer, occur at low elevations in the Hells Canyon region (see Johnson and Simon 1987) but were not recorded on the plots used in this study.

History of Bunchgrasses and Livestock Grazing in the Pacific Northwest Early European explorers gave vivid accounts of the bunch-grass landscape they encountered prior to the arrival of Eu-ropean settlers (the following quotes are from McArdle and Costello 1936). In 1841 Commander Wilkes wrote from north central Oregon: “These hills, as well as the country nearer at hand, were covered with a natural hay or bunch-grass, which affords very nutritious food for cattle.” Cap-tain Fremont wrote when passing through eastern Oregon in 1843: “The mountains were covered with good bunchgrass.”

Granville Stuart wrote

We crossed the Rocky Mountain Divide on the 10th day of October, 1857. As soon as we had crossed the divide a wonderful change appeared in the country. Instead of the gray sagebrush covered plains of Snake River, we saw smooth rounded hills and sloping benchland covered with yellow bunch-grass that waved in the wind like a field of grain.

GENERAL TECHNICAL REPORT PNW-GTR-641

Grazing by domesticated livestock in the Blue Mountains began with Indian horses in the 1700s, but widespread im-pacts began with expansion of cattle and sheep ranching af-ter 1870 (Galbraith and Anderson 1971, Humphrey 1943).

Numbers of cattle and sheep peaked around 1900. This was a time of true open range, with no government regulation of numbers of animals, season of use, or which livestock own-ers were entitled to use the land. Today we still observe ar-eas with grazing-resistant and invasive weedy species on sites that we believe should support bunchgrass vegetation;

these are former livestock driveways or bedding areas from the early 1900s. Livestock numbers declined after 1906 as the USDA Forest Service began to regulate grazing through its allotment system. Livestock grazing, primarily by cattle, continues to this day on most bunchgrass-dominated plant communities in the Blue Mountains.

Study Area The Blue and Ochoco Mountains of eastern Oregon are part of the Blue Mountains physiographic province (Orr and Orr 1999). This broad area includes the Ochoco Mountains to the west; the Strawberry, Aldrich, and Green-horn Mountains in the southern part of the Blue Mountains;

the Elkhorn Mountains and the tristate uplands of the north-ern Blue Mountains; and the Wallowa Mountains on the east end of the province (fig. 1). The Blue Mountains segment is a northeast- to southwest-trending axis of geologically old mountains that extend south from Pomeroy, Washing-ton, 200 mi to the south near Burns, Oregon. A 50-mile-long spur begins west of Ukiah, Oregon, with an east-west axis toward Fossil, Oregon. The Ochoco Mountains are an-other east-west trending axis of ridges that begin west of the Aldrich Mountains near Dayville, Oregon, and culmi-nate north of Prineville, Oregon. The study was conducted on USDA Forest Service-administered land, including the mountains and intervening canyons in the Malheur, Ochoco, Umatilla, and Wallowa-Whitman NF. Additional plots were installed on the Crooked River National Grasslands between Prineville and Madras, Oregon.

The study area was divided into four segments to assist in portraying the distribution of vegetation as described in this management guide. These segments are as follows with key physiographic features listed:

1. Northern Blue Mountains—North of Interstate 84

(running between La Grande and Pendleton)

• High points—Oregon Butte ( 6,401 ft) and

Mount Emily (6,064 ft)

• Low points—Tucannon River, Umatilla River, Wenaha River

2. Central Blue Mountains—South of Interstate 84 and north of U.S. Highway 26 (running between Unity and John Day).

• High points—Vinegar Hill (8,131 ft) and Rock Creek Butte (9,106 ft)

• Low points—Powder River, Grande Ronde River, North Fork John Day River

3. Southern Blue Mountains—South of U.S. Highway 26

• High points—Strawberry Mountain (9,038 ft) and

Fields Peak (7,363 ft)

• Low points—South Fork John Day River, Malheur

River

4. Ochoco Mountains

• High points—Round Mountain (6,753 ft) and Spanish Peak (6,885 ft)

• Low points—Crooked River

Physiography and Geology The study area is dominantly an extensive lava plateau at elevations of 3,500 to 6,500 ft, punctuated by a few deep canyons, with elevations as low as 2,000 ft, and mountains rising to about 9,000 ft. Most bunchgrass communities here occur over volcanic rocks that erupted during the Cenozoic Era—primarily basalt, but also andesite and rhyolite.

Interbedded in these volcanic rocks are some relatively thin layers of tuffaceous sedimentary rocks. Locally in the southern half of the study area, the older sedimentary and metasedimentary rocks that occur beneath the volcanics are exposed. These consist mostly of graywacke, shale, and argillite deposited in a marine environment. Included with these sedimentary rocks in a few places are ultramafic rocks, including serpentine, that originated as deep ocean crust. These ultramafic rocks occur in environments with bunchgrass vegetation in the Strawberry Mountains and in the vicinity of Mine Ridge southwest of Unity, Oregon. One large area of granitic rock also occurs in the study area—the Bald Mountain batholith—which is centered in the vicinity of Anthony Lakes, northwest of Baker City, Oregon. This high-elevation area has only incidental occurrences of bunchgrass vegetation. (This section is based on Orr and Orr 1999 and Walker and McLeod 1991.)

Climate The climate of bunchgrass plant communities in the Blue Mountain region is temperate and semiarid (USDA NRCS 2004a, Western Regional Climate Center 2004). Precipita-tion is highly seasonal, with most arriving between Novem-ber and June; drought conditions are common in late summer. Temperatures generally decrease with increasing elevation, although low temperature extremes in all seasons tend to occur at moderate elevations on valley bottoms surrounded by higher terrain. Summer highs are typically in the high 70s and 80s °F, and lows in the 40s. Winter highs average in the 30s °F and lows in the teens. Summer temperatures in the 90s °F are not unusual, and winter temperatures occasionally fall below 0 °F. The lowest temperature ever recorded in Oregon or Washington (-54 °F) was recorded at both Seneca and Ukiah, not far from our bunchgrass study plots. Precipitation is generally lower in the south and at lower elevations in our study area. At lower elevations (2,000 to 3,000 ft in the north and 4,500 to 5,000 ft in the south), where bunchgrass communities occur on a variety of sites, annual precipitation averages between 10 and 20 in. Stations at higher elevations in the vicinity of bunchgrass communities have annual precipitation of 20 to 40 in, but effective precipitation on sites where bunchgrass communities actually occur is typically reduced by wind, which removes snow in winter, and intense solar radiation on steep southerly aspects, which greatly increases evapora-tion.

Daubenmire (1956) investigated in detail the relationship of climate to several of the vegetation series that are clas-sified here. The zone dominated by sagebrush with blue-bunch wheatgrass is generally drier and has colder winters than bluebunch wheatgrass areas without sagebrush, al-though no simple statistic defines that boundary. The tran-sition from the sagebrush to ponderosa pine zone occurs at mean annual precipitation values of about 15 in. The blue-bunch wheatgrass and Idaho fescue zones show consider-able overlap in precipitation, but the latter are consistently cooler (which reduces drought severity) and in some cases have higher precipitation also. The July mean temperature in all of Daubenmire’s bluebunch wheatgrass zone stations was greater than 70 °F, as opposed to less than 70 °F in the fescue zone. Note that local site characteristics often allow a vegetation type to occur outside of its zone as determined by the macroclimate.

Pomeroy

Walla Walla

Pendleton

La Grande Enterprise

Ukiah

Baker City

Unity John Day

Dayville

Seneca

Burns

Prineville

Madras

Fossil

Heppner

Sample plots

National forest

O R E G O N

0 25 50 Miles

Figure 1—Location of the study area.

Soils Soils of bunchgrass plant communities are a mixture of loess, volcanic ash, and weathered bedrock. The parent materials are noncalcareous, and basalt is the most common lithology. Most soils have little or no stone-free loamy material at the surface, and coarse fragments increase with depth, reaching bedrock or extremely gravelly/cobbly material within about 2 ft of the surface. A common soil class is Lithic Argixerolls, which are soils with a brown surface layer rich in organic matter from herbaceous understory plants (a mollic epipedon), a seasonally dry (xeric) moisture regime, a clay-enriched subsoil horizon (an argillic horizon), and bedrock near the surface (a lithic contact). Where bedrock is very close to the surface, one or both of the horizons mentioned above may be missing, resulting in Lithic Haploxerolls (soils where the clay subsoil is thin or absent) or Lithic Xerorthents (soils where both the clay subsoil and brown surface layer are thin or absent). For more information on these soil classification terms see

USDA NRCS (1998).

On level to gently sloping sites with shallow soils, both the clay-rich subsoil and underlying bedrock retard infiltra-tion of spring meltwater, resulting in temporary soil satura-tion. However, as a result of the dry summer climate and the low available water capacity of these soils, soil moisture is depleted during the summer. These strongly contrasting soil moisture conditions present a special challenge for plants, and the result is sparse vegetation of stiff or low sagebrush (Artemisia rigida or A. arbuscula), Sandberg’s bluegrass (Poa sandbergii), onespike oatgrass (Danthonia unispicata), and occasionally rushes (Juncus spp.).

On soils with better drainage, thanks to a sloping surface, well-fractured bedrock, or gravelly colluvial parent materi-al, spring saturation is not a problem, but summer depletion of soil moisture is ubiquitous. Plant communities are strong-ly dependent on the severity of drought as determined by the interaction between macroclimate (i.e., precipitation and temperature), soil climate/microclimate (as determined by slope aspect and steepness, and wind deflation of snow), and soil moisture-holding capacity. Thus a plant species may be restricted to deeper soils near the climatically dry limit of its range.

Vegetation Overview Climate, elevation, aspect, soils, and disturbance all influ-ence vegetation. In the Blue and Ochoco Mountains, tem-perature and moisture are fundamental determinants of where plant species grow. The bunchgrass vegetation series can be arrayed according to their moisture and temperature adaptations from “cool, moist” to “hot, dry” as follows:

FOREST

Ponderosa pine/bunchgrass Western juniper/bunchgrass

SHRUBLANDS

Mountain mahogany/bunchgrass Bitterbrush/bunchgrass Mountain big sagebrush/bunchgrass Low sagebrush/bunchgrass Stiff sagebrush/bunchgrass

GRASSLANDS

Green fescue (subalpine) Idaho fescue (subalpine) Idaho fescue (montane) Bluebunch wheatgrass Onespike oatgrass Sandberg’s bluegrass

The temperature and moisture environment for plant growth is determined by the combination of climate (strong-ly influenced by elevation) and local soil/site conditions. For example, a site at low elevation with very shallow soils will be dry for most of the growing season and might support stiff sagebrush or Sandberg’s bluegrass series vegetation.

A site at higher elevation, on a deeper soil, or on north as-pects is cooler and moister and may sustain ponderosa pine or Idaho fescue series vegetation.

Plant Association Concept In this study we sampled existing vegetation and the environmental attributes of each site to help identify the potential natural vegetation, which in turn was used to define plant associations. Existing vegetation is the plant cover, floristic composition, and vegetation structure occurring at a given location at the current time (Winthers et al. 2004). Potential natural vegetation (PNV) is that plant community that would become established if all successional sequences were completed without human interference under the present environmental and floristic conditions, including those created by man (Winthers et al.

2004). Historical vegetation is that vegetation believed to have existed in the study area prior to the impact made by Euro-Americans beginning about 1850 (USDA Forest Service 2003). Prior to the arrival of Euro-Americans, the primary disturbance agents on bunchgrass sites were fire (including fire caused by Native Americans), wild ungu-lates, small animals, insects, diseases, and slope-driven soil movements. As a result of these disturbance factors, historical vegetation did not consist entirely of late-succes-sional vegetation; in other words, historical vegetation is not synonymous with PNV. Instead the historical vegetation landscape generally contained a mix of vegetation commu-nities resulting from disturbance events. The historical range of variability represents this mix of different seral conditions that are believed to have occurred in the histori-cal vegetation.

This publication uses the conventions for classifying po-tential vegetation that have been in use in Region 6 of the Forest Service for several decades. The plant association, the fundamental unit of classification, is identified by deter-mining the potential vegetation of a site. Plant associations are named by using both dominant late-seral overstory spe-cies and other species that indicate environmental conditions such as moisture supply and temperature. A plant associa-tion includes the relatively rare pristine late-seral vegetation for which it is named and all the disturbed and early seral vegetation that would succeed to it.

This definition differs somewhat from the usage of “plant association” in the National Vegetation Classification Sys-tem (Jennings et al. 2003) and the USDA Forest Service Terrestrial Ecological Unit Inventory Technical Guide (Winthers et al. 2004), where plant associations are units in the classification of existing vegetation. Our concept of a plant association is similar to the “habitat type” concept used in Forest Service Northern Region (Region 1) (Pfister and Arno 1980). The habitat type is a unit of land classified as capable of supporting a specific plant association.

The plant community type is a less formal vegetation type, simply an aggregation of plant communities with flo-ristic and structural similarities (Society for Range Man-agement 1989). We established plant community types to describe (1) highly disturbed communities where the PNV could not be determined, and (2) possible plant associations where our sample size was too small to adequately under-stand the range of variation.

Series are aggregations of plant associations and plant community types based on the dominant overstory plants.

For example, bunchgrass vegetation that occurs with an overstory of big sagebrush is placed in the big sagebrush series.

The concept of potential natural vegetation can become confusing where communities have been greatly altered from the historical condition. Two problematic situations are particularly common: (1) juniper invasion and (2) re-placement of bunchgrasses by introduced species such as cheatgrass. In the first case, many sites that historically were treeless will succeed to juniper woodland if a seed source is available and they are protected from fire. We classify shrubland or herbland vegetation into shrubland or herb-land plant associations if they currently have few or no trees, even if they could theoretically support junipers. Once ju-nipers have become established, we classify the vegetation into a juniper plant association, even if historically junipers were absent (and we note the latter fact in the plant asso-ciation description). Secondly, in some cases, severe dis-turbance combined with invasion by nonnative plants has produced communities that no longer would succeed natu-rally back to the PNV. If the PNV can be determined, these communities are classified into a plant association. In a few cases where the PNV cannot be identified, we have estab-lished plant community types to accommodate them.

State-and-Transition Concept This study portrays the condition and successional status of vegetation by the state-and-transition model (Westoby et al.

1989). According to this model, the vegetation present on a given site is determined by complex interactions of the history of disturbance, climate, management, and avail-ability of propagules (e.g., seed). Vegetation is modeled as a set of states or phases, and transitions describe how vegetation can change from one state to another. Examples of transition agents include fire, succession, grazing, and invasion by exotic plant species. This model has largely superseded the range succession model (Dyksterhuis 1949) in range management.

In current usage, “phases” are separated from one another by transitions, most of which are reversible. Phases are sep-arated into “states” when a transition crosses an ecological threshold that is nearly irreversible (Laycock 1991, String-ham and Kreuger 2001). For example, overgrazing may re-duce cover by native grasses somewhat, but after a change in grazing regime the bunchgrasses might recover their for-mer dominance; the pristine and somewhat degraded com-munities would then be considered two phases of a single state. If degradation were to proceed to the point where the native grasses were eliminated and were replaced by highly competitive exotic grasses that would persist even if grazing ceased, then a threshold would be crossed and transition to a new state would have occurred.

Phases and States Plant associations were subdivided into phases and states that represent significantly different cases from the perspec-tive of ecology and management. For a given plant associa-tion we typically recognized four phases of vegetation as follows:

A Relatively pristine vegetation close to the hypothetical

PNV. For some shrubland plant associations with ade-quate data we were able to subdivide the A phase into a transient postfire phase (A1) and a late-successional phase with larger and older shrubs (A2).

B Vegetation that has been moderately altered by grazing, such that certain grazing-sensitive species have been af-fected but are still present in substantial amounts.

C Vegetation that has been greatly altered by grazing but still retains native species in sufficient amounts that they could presumably recover to approximate the PNV phase given sufficient time. Some plant associations had more than one phase of this kind (C1, C2, etc.): e.g., a phase where bunchgrasses have been partly replaced by exotic annuals vs. a phase where they have been partly replaced by grazing-resistant perennial forbs. The au-thors believe that C phases represent communities in peril of crossing a threshold into a new state via loss of native species, invasion of other species, or soil erosion.

D Vegetation that has been altered by loss of native species and invasion of exotics to the point that succession back to the original PNV is now unlikely and a new state has been established. For some plant associations there are more than one of these phases (D1, D2, etc.), depending on what invasive species has gained dominance.

Thus A, B, and C represent different phases within a sin-gle state (fig. 2). Transitions between them are caused by heavy grazing, succession, and fire, and are believed to be reversible. The D phases are separated by an ecological threshold from the others and thus belong to a different state.

This threshold is formed by loss of native species, invasion of exotic plants, soil erosion, or some combination of these factors. There may also be thresholds separating some of the D phases from one another, but we have not studied these transitions in detail.

Fire Regimes and Fire Effects Fire has historically been an important natural disturbance process in bunchgrass communities and remains so today.

Historical fire frequency is difficult to determine in bunch-grass vegetation because of the lack of trees that could pro-vide a record of fires. Fire effects and postfire succession in this region are fairly well known thanks to local monitoring (see Johnson 1998b, Swanson 2005, and USDA Forest Ser-vice 2005 for more information)

Ponderosa Pine Forest Considerable fire history work has been done in ponderosa pine forests, including locally in the Blue Mountains (Heyerdahl et al. 2001). These studies generally agree that ponderosa pine forests historically had frequent light underburns, with a return interval of about 10 to 20 years (Heyerdahl et al. 2001) or even less (Brown and Smith 2000). However, nearly all of these studies have been in more mesic pine communities, such as those with an understory of pinegrass (Calamagrostis rubescens), elk sedge (Carex geyeri), or snowberry (Symphoricarpos spp.), and not in pine forests with a bunchgrass understory. The exception is Miller and Rose’s (1999) study in south-central Oregon of fire scars on scattered pines in bunchgrass ecosystems that were being invaded by juniper. This study showed fire-return intervals of 12 to 15 years, i.e., not unlike the moister pine sites. Note that such short fire-return intervals would make survival of young pines a very rare event. A few investigators believe that methodological problems have led us to underestimate fire-return intervals in pine, and that they could actually have been much longer (e.g., Baker and Ehle 2001).

The frequent light underburns of pines probably caused little tree mortality except for seedlings and saplings, but consumed much of the understory biomass. Fire effects on the understory vegetation are similar to those of analogous treeless vegetation types discussed below.

Juniper Woodlands Available evidence suggests that low-severity underburns are not common in these woodlands and that most fires kill the trees (Baker and Shinneman 2004). Juniper woodlands in our study area include (1) woodlands that contain old trees and probably existed before the change in fire regimes that accompanied European settlers and (2) woodlands that are the result of postsettlement juniper invasion (Burkhardt and Tisdale 1969, Miller and Rose 1995, Young and Evans 1981).

The juniper woodlands where juniper has probably been present for many centuries contain old trees and are pri-marily in environments with little fuel because of shallow, droughty soils. In our classification, these are the JUOC/

CELE3/FEID-AGSP, JUOC/ARAR8/FEID, JUOC/ARRI2,

and JUOC/AGSP types. Historically, fires in these types were probably not only rare but quite variable in frequency and severity. We presume that fire-return intervals were at least 50 years, because shorter return intervals would have eliminated the junipers (Burkhardt and Tisdale 1976).

The juniper types that have resulted from recent invasion are typically on deeper soils with more productive herba-ceous layer and hence more fuels: the JUOC/PUTR2/ FEID-AGSP and JUOC/FEID-AGSP types. These sites historically had fires frequent enough to exclude junipers;

mean fire return intervals may have been as low as 12 to 15 years (Miller and Rose 1999). Reduction of fine fuels by juniper competition may have made the vegetation less able to carry fires today than before juniper invasion (Petersen 2004).

Shrublands We can subdivide our shrublands into two groups based on their fuel characteristics and historical fire regimes: (1) those with relatively dense herbaceous understory and his-torically frequent fires and (2) those with light fuels and rare fires. The former group includes most of our bitterbrush and big sagebrush plant associations. Historical fire return inter-vals as low as 12 to 15 years have been estimated for moun-tain big sagebrush steppe (Miller and Rose 1999). Given the rate of sagebrush regeneration, this fire frequency would have resulted in widespread grassland communities with little sagebrush. A longer fire return interval of 20 to 70 years has also been suggested for sagebrush steppe (Brown and Smith 2000).

The second group includes all plant associations in the mountain mahogany (CELE3), stiff sagebrush (ARRI2), and low sagebrush (ARAR8) series, plus PUTR2/ERDO, ARTRV/AGSP-POSA12, and ARTR4/POSA12-DAUN.

Fires were historically (and remain today) infrequent and patchy in these communities because the poor soils produce little fuel.

Mountain mahogany, sagebrush, and bitterbrush are readily killed by fire, although in some environments bit-terbrush shows limited ability to resprout (Clark et al. 1982, Driscoll 1963) and mature mountain mahogany can survive some light fires (Arno and Wilson 1986). All three spe-cies recolonize burned areas by seed, generally seed car-ried from surviving plants by wind (mountain mahogany and sagebrush) or by animals (bitterbrush). Thus the rate of shrub recolonization depends on the size and severity of the burn, and weather for seed production and seedling surviv-al in postfire years. Our monitoring of big sagebrush com-munities after fire shows that under good conditions, this shrub can reestablish 10 percent canopy cover as early as 10 years after complete kill by fire. Our limited observations of bitterbrush and mountain mahogany suggest much slower recovery of these species.

Response of the herbaceous layer of shrublands to fire is like that of grasslands described below, with one important exception. When the shrub canopy is very dense, e.g., in big sagebrush with canopy cover of 50 percent or more, fires can burn very hot and cause more death of herbaceous plants than would otherwise occur. We have observed Idaho fescue eliminated from a site by severe fire in dense sagebrush, cre-ating openings for weedier grass and forb species to expand.

Grasslands Historical fire frequencies in grasslands of the study area re-main largely unknown. There is consensus that fires were frequent (return intervals less than 35 years) and consumed most aboveground plant parts (Brown and Smith 2000). Our observations of modern fires show that patchy and light fires are not unusual in grasslands, especially on low-biomass types.

Plant associations of the Idaho fescue series contain enough biomass to burn nearly every year, but actual fires were not nearly this frequent because of low flammability of the vegetation early in the season and lack of ignitions (Agee 1996). Our monitoring shows that Idaho fescue is often suppressed for a few years after the fire, after which

Figure 2—Generalized state-and-transition diagram used in this study. Boxes symbolize phases and arrows are transitions between them. The A, B, and C phases result from increasing grazing-related degradation that is probably reversible by natural succession. The D phases are in a second “state” separated from the first state by an ecologic threshold. Transitions to D phases are caused by loss of native species, invasion of exotic species, and soil erosion; they are probably not reversible by natural succession.

A1 B

A2

C1 C2

C3, etc.

D1 D2

D3, etc.

5 4 1, 2 5

State 1 State 2

Transitions:

(1) Heavy grazing and trampling

(2) Exotic species invasion

(3) Soil erosion

(4) Fire

(5) Succession

Threshold

1, 2, 3 it regains its former cover; other species in Idaho fescue communities return to prefire cover in the first postfire season. There is a flush of annual plants on some sites in the first year after fire, but annuals subsequently decline to prefire levels.

Response of bluebunch wheatgrass series vegetation to fire is even weaker, with plants typically regaining their prefire covers in postfire year 1. Occasionally bluebunch wheatgrass is suppressed a little and Sandberg’s bluegrass or annual forbs increase after fire, but effects are subtle and transient.

Some grassland types on very poor soils (AGSP-ERUM, AGSP-CYTEF, ERDO-POSA12, DAUN-LOLE2, and POSA12-DAUN) produce little biomass and probably had less continuous and less frequent fires than the other grass-land types. Our limited monitoring suggests that, as in the case of the other grassland types, post- and prefire vegeta-tion are very similar.

The Spread or Decline of Woody Plants As mentioned above, juniper has become more widespread than it was in the past. Juniper expansion in our area be-gan in the late 1800s and continues today (Miller and Rose 1999). The main causes of juniper expansion are exclusion of fire, introduction of domestic livestock (which reduces fu-els for fires and reduces vegetation that competes with juni-per), and possibly an unusually moist climate during some recent decades. Burkhardt and Tisdale (1969) believe that juniper could potentially colonize any site that currently supports big sagebrush and Idaho fescue. Competition by ju-nipers usually causes thinning of understory vegetation, es-pecially shrubs (Miller et al. 2000). This effect has been linked to increased soil erosion in ecosystems invaded by junipers (Petersen 2004). As mentioned above, our JUOC/ PUTR2/FEID-AGSP and JUOC/FEID-AGSP types have re-sulted from juniper invasion. The presence of a few seedling junipers in several of our big sagebrush types (ARTRV/ AGSP-POSA12, ARTRV/FEID-AGSP, and ARTRV/FEID- KOCR) suggests that local juniper invasion of these types is likely in the future.

Big sagebrush is widely believed to be denser today than it was prior to settlement by Europeans. Likely reasons for this increase include fire suppression and removal of grass com-petition by livestock grazing. Controversy over this issue has been heated (see, e.g., Welch and Criddle 2003). There is little information about presettlement sagebrush densities;

however, one study of old photographs in Wyoming and Ida-ho showed changes since the 1870s in sagebrush abundance to be variable and localized (Johnson 1986). Owing to the lower fire frequency today than in the past, there is probably proportionally less area of low-density, postfire regenerating sagebrush communities in our study area today than there was historically.

Some woody species of bunchgrass ecosystems have ex-perienced declines owing to browsing by wildlife and live-stock, notably bitterbrush (Purshia tridentata) and mountain mahogany (Cercocarpus ledifolius). Seedlings of these spe-cies are quite rare in our study area, and existing plants are often old and severely hedged, or (in the case of moun-tain mahogany) defoliated up to the height that animals can reach. These species are long lived and quite tolerant of de-foliation once they are established (Garrison 1953), but re-cruitment of new plants is needed because productivity of individual plants drops off (after about 80 years in the case of bitterbrush) and older plants gradually die off (McCon-nell and Smith 1977). At least a temporary reduction in den-sities of all the large ungulates that feed on them—livestock, deer, and elk—will probably be needed at some point to en-sure the survival of the next generation of these plants.

Newcomers to Bunchgrass Communities Bunchgrass plant communities in the Western United States have undergone major changes as a result of the introduction of plants from Eurasia. These newcomers are mostly annu-als that colonize disturbed soil. Among the most important are the annual bromegrasses, including cheatgrass (Bromus tectorum) and several other species (B. japonicus, B. mollis, B. brizaeformis, and B. commutatus). These grasses were in-troduced in the late 1800s, and by about 1930 had colonized most of the Pacific Northwest and intermountain region (Mack 1981). They are cool-season grasses with a winter-annual life cycle: seeds germinate in the fall and grow until forced into dormancy by cold weather, then resume growth in the spring, flower, and produce seed before dying in early to mid summer (Klemmedson and Smith 1964). Cheatgrass and its relatives have displaced native annual grasses such as the annual fescues, as well as bluebunch wheatgrass, which was more successful at colonizing disturbed soil without such a strong competitor as cheatgrass (Mack 1981).

Cheatgrass is highly palatable to livestock when green and can be kept in check to some degree by spring grazing (Daubenmire 1940). Unfortunately, heavy spring and early summer grazing will also weaken native perennial bunch-grasses, and if grazing pressure is subsequently decreased, cheatgrass will outcompete the native grasses to fill any new gaps.

Cheatgrass is known to invade sagebrush communities and increase fuels and fire frequency, resulting in annual grasslands that are perpetuated by fire (Young and Evans 1978). We have not observed this cheatgrass-fire cycle in our study area. Flushes of annual plants (including annual bromes like cheatgrass) sometimes occur in the first year after fires in our bunchgrass communities, but our observa-tions to date show cover by annuals declining back to prefire levels within a year or two after the initial postfire flush without greatly altering the fire regime. Our densest stands of annual bromes occur in communities of the bluebunch wheatgrass and Idaho fescue series that have been highly disturbed by grazing and trampling, or in a few cases, by tillage on abandoned farm fields.

Another aggressive introduced annual grass has appeared in bunchgrass communities of our area, Ventenata dubia.

We have even observed it to displace cheatgrass on some of our plots. Ventenata forms a dense thatch that probably in-hibits reestablishment of native bunchgrasses. It is short, wiry, and unpalatable to livestock. Ventenata appears to be spreading in our area, and the limits to conditions it can tol-erate are still unknown.

Medusahead (Elymus caput-medusae) is another inva-sive annual grass that has infested large areas in the Pacific Northwest but occurs only in a few places in our study area.

Medusahead expansion here remains a possibility, at least locally on the disturbed, clay-rich soils that it prefers (Dahl and Tisdale 1975).

Many introduced perennial grasses are also present in our study area, some that were seeded onto burned areas or de-graded rangeland, and some that have spread on their own.

The only one that occurs widely in bunchgrass communi-ties in our study area is Kentucky bluegrass (Poa pratensis).

This grass is occasionally present in the more mesic bunch-grass plant associations, i.e., those containing Idaho fes-cue. This grass is highly tolerant of close grazing and can invade overgrazed sites if sufficient soil moisture is present (Daubenmire 1972, Mack and Thompson 1982). Interme-diate or pubescent wheatgrass (Agropyron intermedium) is also present at scattered locations in mesic bunchgrass com-munities, but it does not appear to have spread far beyond where it was originally seeded.

Indicator Species The habitat needs of plant species are evident by the environments in which they persist. Some species require stable conditions, others thrive on recurring instability.

Some species have adapted to a particular locale owing to long-term climatic conditions of the area; others have colonized and thrived owing to changes in the micro-climate of a particular site.

Plants that designate thresholds of environmental changes along gradients are called indicator plants. The indicator plants selected to define the plant community type or the plant association are those deemed to be the most diagnostic of a particular environment, i.e., those that have a high fidelity and constancy to the type. Although they do not necessarily indicate the sum of all environmental condi-tions, they are considered the best candidates of the associ-ated flora within a classified type to indicate the occurrence and distribution of that vegetation unit.

The indicator species selected are those sought by field in-vestigators to help determine proper assignment to a given plant association or plant community type. These indicator plants appear in the keys and in the names of the plant asso-ciations and community types. Information on identification of these plants may be found in Johnson (1998a).

Sampling and Field Methods Sampling of bunchgrass plots used in this study began in 1956 by Fred Hall (Hall 1973) and continued with the senior author’s fieldwork from the 1980s until 2004. Beginning in 1993, the rangelands of the Blue Mountains were systemati-cally covered to more adequately sample vegetation of vary-ing types and conditions. In 1999 and 2000, this study was extended to include vegetation in the Ochoco Mountains. To assist in interpreting changes over time (trend), many Park-er Three-Step Condition and Trend (Parker 1951) sites were used for plot locations. A total of 487 plots were sampled.

At each sample site a 375-m2 (72 ft in diameter) circu-lar reconnaissance plot was established. Within this area, a total plant species list was derived. Next the canopy cover-age was estimated to the nearest 5 percent for all species on the list. Additionally ground surface cover was estimated for bare ground, bedrock, rock, gravel, erosion pavement, moss-es, lichens, and litter. Site environmental attributes were ex-amined and noted as well. These included elevation (feet), aspect (degrees), slope (percent), lithology, soil depth, posi-tion, and slope shape.

Photopoints were established at each plot by locating a steel angle iron stake at the plot center. From this location, varying views were taken by using 35mm single lens re-flex cameras employing color slide film and black and white print film.

The plots were intended for repeated monitoring. There-fore, reference markers and information necessary for relo-cation are provided for each sample location. Plot locations were described on plot cards and marked on U.S. Geological Survey quadrangle maps. Reference signs placed at the site and on adjacent points of access lead the investigator to the plot center stake. In the office, the plot locations were digi-tized from the topographic maps and the resulting coordi-nates stored in the plot database and geographic information system.

Plot sizea

Canopy 375 m2 1/10 acre cover (%) xb rc x r

1 1.94 m (6.36 ft) 1.09 m (3.58 ft) 6.