Designing Native Seed Mixes for
Restoring Crested Wheatgrass Fields in Eastern Montana
Prepared by:
Peter Lesica and Stephen V. Cooper
Conservation Biology Research
929 Locust
Missoula, Montana 59802
Prepared for:
Bureau of Land Management
Lewistown District
920 NE Main Street
Lewistown, Montana 59457-4079
December 2011
Executive Summary
Old fields dominated by crested wheatgrass are common throughout much of eastern Montana. Many of these fields are virtual monocultures, providing relatively little benefit to livestock or wildlife. Understanding secondary succession in these old fields in relation to soils can aid in planning reclamation activities. Our study addresses whether soil properties are related to the degree to which crested wheatgrass fields have been invaded by natives and the degree to which crested wheatgrass has invaded native vegetation. We also determine which native species are most successful at coexisting with crested wheatgrass and on which soils.
We conducted our study at 24 sites in central Montana where there were 75-year-old crested wheatgrass fields adjacent to unplowed native vegetation. Soil textures ranged from clay to loamy sand. We located three 0.8 ha macroplots at each site, one in the native vegetation and two in the crested wheatgrass field, one adjacent to native vegetation and one in the center of the field. We measured canopy cover of all species and density of native species in randomly located plots within each macroplot. We also sampled and analyzed soil in native and crested wheatgrass macroplots.
There was no difference in measured soil variables between native vegetation and crested wheatgrass fields. Greater canopy cover of crested wheatgrass was associated with heavier, especially siltier soils, but the associations were not strong (r20.30). Mean crested wheatgrass cover in native vegetation was low (2%) with the exception of one site with 56%. Total canopy cover of all native species at the edge of crested wheatgrass fields declined significantly as the canopy cover of crested wheatgrass increased, due primarily to the decline in total canopy cover of native grasses. Only 36 native perennials were common at the edge of crested wheatgrass fields. The majority (75%) of the native perennials had lower abundance in the center of crested wheatgrass fields compared to the edge, indicating that most species had limited dispersal.
Our results suggest that there is little reason to choose among crested wheatgrass fields for restoration based on soils. The general paucity of native species in crested wheatgrass fields more than seven decades after they were planted and the negative relationship between cover of native species and crested wheatgrass indicates that cover of crested wheatgrass will have to be diminished, at least temporarily, in order to establish significant number of native plants. We recommend using herbicide to weaken crested wheatgrass stands. This protocol will minimize disturbance and crested wheatgrass emergence and provide a low-competition window for native plant establishment.
Our study suggests that the majority of native species have a limited ability to disperse so native species will have to be seeded into crested wheatgrass fields after they have been treated with herbicide. Results of our study provide guidance on which species are able to successfully coexist with crested wheatgrass. Although there was broad overlap among the native species present in crested wheatgrass fields with either heavy or light soil, there were some species, especially forbs, that occurred preferentially on one of the two soil types. Our recommendations omit some species that were common in crested wheatgrass fields. These include prickly pear cactus and narrow-leaved sedge. The former has large flowers that are attractive to pollinators but is unpalatable to livestock and most wildlife. The latter produces only very small amounts of palatable forage. Forbs in recommended seed mixtures show a range of phenologies. Some species flower and produce fruit early in the growing season, while others are active throughout most of the summer. Having a mixture of forbs that are active at different times helps provide adequate habitat for birds such as sage grouse and generalist pollinators such as bumblebees.
Introduction
Much public land in eastern Montana is retired agricultural fields planted to crested wheatgrass 70-80 years ago. These old fields provide little benefit for livestock or wildlife. Many crested wheatgrass fields have resisted invasion by native species and are still virtual monocultures, even after more than seven decades (Christian and Wilson 1999). Monoculture crested wheatgrass fields lack the shrubs and forbs to support strong populations of antelope, sage grouse and Baird`s sparrow compared to native prairie or sagebrush steppe (Reynolds and Trost 1980, Urness 1986). Lloyd and Martin (2005) found that nestlings of chestnut-collared longspurs grew more slowly and had a 17% lower chance of surviving in crested wheatgrass fields compared to native prairie of northeast Montana. However, studies in Washington suggest that crested wheatgrass can provide suitable nesting habitat for sage grouse when it has been reinvaded by sagebrush (Schroeder and Vander Haegen 2006).
Some crested wheatgrass fields are returning to native composition, but most are not (McHenry and Newell 1947, Smoliak et al. 1967, Looman and Heinrichs 1973, Box 1986, Marlette and Anderson 1986, Wilson 1989). Crested wheatgrass has been successfully planted on a wide variety of soils (Knowles and Buglass 1980), and edaphic differences are one possible explanatory variable for differential rates of old-field secondary succession. Furthermore, the identity of native species able to invade crested wheatgrass fields may depend on soil texture. Understanding secondary succession in these old fields in relation to soils can aid in planning reclamation activities.
In some cases crested wheatgrass has invaded native vegetation from adjacent planted fields (Christian and Wilson 1999, Heidinga and Wilson 2002, Bakker and Wilson 2004, Henderson and Naeth 2005, Romo 2005, Vaness and Wilson 2007). Knowing which native habitats are more prone to invasion by crested wheatgrass will help managers prioritize old-field
restoration efforts.
Our study describes secondary succession of crested wheatgrass fields to native prairie in central Montana with an experimental design inadvertently begun ca. 75 years ago when crested wheatgrass fields were planted in a mosaic of native steppe. Our study addresses whether soil properties are related to the degree to which crested wheatgrass fields have been invaded by natives and crested wheatgrass has invaded native vegetation from nearby fields. Our study will also determine which native species are most successful at coexisting with crested wheatgrass and on which soils. These results have important implications for restoration of these fields to more diverse native composition.
Methods
Study area and the history of crested wheatgrass introductions
Crested wheatgrass was first introduced into the U.S. in 1898, and the first successful experimental plantings were at U.S. Department of Agriculture stations in 1907 to 1910 (Rogler and Lorenz 1983, Lorenz 1986). The first planting in Montana was in 1915. The crested wheatgrass cultivar, Fairway, was developed in Canada at the same time as a similar selection in
Montana. It is a diploid and can be referred to Agropyron cristatum var. pectinatum (M. Bieb.) Roshev. ex B. Fedtsch. (Dewey 1986). Fairway was the only cultivar in mass production at the time, and was used almost exclusively in reclaiming Montana LU lands in the late 1930`s.
Crested wheatgrass has been planted on 6 to 10 million ha in the northern Great Plains (Lesica & DeLuca 1996); much of this occurred in the first half of the last century as a response to abandoned marginal cropland (Lorenz 1986). The region was settled by homesteaders in the early 1900`s and millions of hectares of native prairie were plowed and planted to wheat. Wheat farming was profitable until the great drought of 1928-1937. During the drought there was a mass exodus from the region, and most of the farm fields were abandoned. The Agricultural Adjustment Act of 1933 and the Bankhead/Jones Act of 1937 authorized the U.S. Department of Agriculture to purchase these abandoned homesteads (Land Utilization or LU lands), and the Soil Conservation Service was given the responsibility of stabilizing the soil and returning the land to productivity (Lorenz 1986).
LU lands administered by the Bureau of Land Management in Montana are found mainly from the north-central portion of the state just north of the Missouri River through Fallon County in the southeast. Our study sites were all located on LU lands planted to crested wheatgrass ca. 75 years ago and located primarily in Fergus and Petroleum counties with three sites in Musselshell and one each in Chouteau and Yellowstone counties in the east-central part of the state.
Field Methods
The Lewistown District Office of BLM provided a list of 34 potential study sites. We randomly selected 16 sites for sampling in 2010 and four sites in 2011, stratified by soil type based on existing soil surveys. Since most of the study sites in Fergus and Petroleum counties were on heavy soils, we also sampled four sites in Musselshell, Choteau and Yellowstone cos.
that had light soil. Locations are provided in Table 1.
Table 1. Location of study sites and soil variables from the center of the crested wheatgrass fields.
We located three circular, 0.8 ha macroplots (50 m radius) at each site (Table 2). The
Native macroplot was in native vegetation at the edge of the crested wheatgrass field. The Edge
Site name County Location pH EC (S) Soil texture class Soil texture
Erie Fergus S 21 & 20, T20N., R19E 6.64 77 heavy Loam
Maruska Fergus S 28, T19N., R23E 7.56 330 heavy Clay
Chippewa Fergus S 11, T15N., R22E 6.62 132 heavy Clay-loam
Milwaukee Fergus S 8, T20N., R19E 6.95 340 heavy Clay-loam
GrassRange Fergus S 27, T15N., R23E 6.41 132 heavy Loam
Boxelder Petroleum S 23, 24 & 25, T17N., R25E. 7.06 85 heavy Loam
Hansom Petroleum S 25, T18N., R26E 7.72 184 heavy Clay-loam
East Christina Fergus S 25, T18N., R26E 6.81 46 heavy Clay-loam
GrassRangeE Fergus S 8, T15N., R24E 7.01 70 heavy Clay
Wolf Fergus Ss 9, 10, 11, 12, T18N., R24E 6.62 152 heavy Clay
Kosir Fergus S 8, T18N., R24E 8.05 120 heavy Clay
Moulton Fergus S 23, T15N., R23E. 6.91 158 heavy Clay
Duck Creek Fergus S 34 & 35, T17N., R24E 7.46 130 heavy Silty-clay
Windmill Petroleum S 5 & 6, T15N., R28E 7.55 290 light Loam
Maidenhead Fergus S 26 & 35, T16N., R21E 7.12 170 heavy Clay-loam
Manuel Petroleum S 3, T15N., R28E 7.94 160 heavy Loam
Carl Spring Fergus S 7, 8 & 17, T14N., R23E 7.56 227 light Loamy-sand
RifleRange Yellowstone S 8, T3N, R26E 8.13 185 light Loamy-sand
Loma Chouteau S 13, T25N, R9E 7.85 224 light Sandy-loam
West Bohemia Fergus S 7, T18N., R23E 7.32 445 heavy Sandy-clay-loam
Gallatin Fergus Ss 25, 36, 31 T 23 N, R 17,18 E 8.12 185 light Sandy-loam
Milton 1 Musselshell S 3, T9N, R26E 7.83 296 light Loamy-sand
Milton 2 Musselshell S 14, T9N, R26E 7.45 152 light Sandy-loam
Milton 3 Musselshell S 11, T9N R26E 8.2 199 light Sandy-loam
We located three circular, 0.8 ha macroplots (50 m radius) at each site (Table 2). The Native macroplot was in native vegetation at the edge of the crested wheatgrass field. The Edge macroplot was in the crested wheatgrass field near the edge of the native vegetation. In all but three cases the centers of these two macroplots were within 150 m of each other (mean=114m). The Center macroplot was either in the center of the crested wheatgrass field or as far from native vegetation as possible. In all but one case the center of the Center macroplot was greater than 150 m from the center of the Native macroplot (mean=314 m). The Edge and Center macroplots were almost always at least 100 m apart. In majority of cases the three macroplots were in the same grazing pasture.
Table 2. Distance (m) between Native and Edge and between Native and Center macroplots at the 24 study sites.
Site Native-Edge Native-Center Site Native-Edge Native-Center
Erie 116 227 Duck Creek 129 298
Maruska 100 435 Windmill 495 897
Chippewa 60 154 Maidenhead 119 339
Milwaukee 110 530 Manuel 117 286
Grass Range 120 271 Carl Spring 121 247
BoxElder 158 342 Rifle Range 147 373
HansomDam 77 106 Loma 94 223
EChristina 75 201 West Bohemian 645 554
GrassRangeE 95 276 Gallatin 110 327
Wolf Ind 147 315 Milton 1 100 168
Kosir 72 203 Milton 2 115 198
Moulton 232 395 Milton 3 103 162
We located four 0.02 ha sample plots (8 m radius) within each macroplot, one on each principal compass bearing at randomly-chosen distances from the macroplot center. We estimated canopy cover of all vascular plants observed in each sample plot using the following cover classes: T- = <0.05%, T=0.5%, T+=0.6-0.9%, 1%, 2%, 5%, 10%, 15%, etc. We located
three 1-m2 microplots in each sample plot 4 m from plot center, one along each of three evenly spaced radii. The number of individuals of all native perennial species was counted in each microplot. Individual stems or clumps of stems were counted for rhizomatous species.
We used a hand-held GPS device to record the latitude and longitude of the center of each macroplot at each site. We collected three soil samples from haphazardly-chosen locations near plot center in the Native and Center macroplots at each site. Macroplot soil samples were combined and analyzed for particle-size distribution by a modified Bouyoucos hydrometer method (Day 1965). pH and conductivity (a measure of salinity) were measured using hand-held meters. Nomenclature follows Dorn (1984). Wyoming big sagebrush (Artemisia tridentata ssp. wyomingensis) is hereafter referred to as sagebrush. We did not attempt to quantify Selaginella densa in density microplots because individual ramets and/orgenets cannot be distinguished, and the plant is inappropriate for restoration. We attempted to distinguish western wheatgrass (Elymus smithii) from thick-spike wheatgrass (Elymus lanceolatus), but the two species can be very similar (Lavin and Seibert 2011), and we may have erred occasionally.
Data analysis
We used the total canopy cover of perennial native species in Edge macroplots to represent the degree of native plant invasion into crested wheatgrass fields. We used the Edge Density Ratio (EDR = Mean DensityEdge/Mean DensityNative) as the measure of a species` success coexisting with crested wheatgrass. We assumed that dispersal limitation was minimal because
Edge macroplots were immediately adjacent to native vegetation. Density in the Edge macroplots alone is a poor measure of success in the crested wheatgrass field because it does not take into account how large the propagule source is nor the size or growth form of the species compared to other species. For example, a rhizomatous species such as Elymus smithii, could have a mean density of 50 stems/m2 in the Native macroplot and only 5 stems/m2 in the crested field, a 90% reduction. However, a bunchgrass like Stipa viridula might have 4 plants/m2 and 2 plants/m2 in Native and Edge macroplots respectively, only a 50% reduction, and yet have a much lower Edge density than the sod grass. Dividing Edge macroplot density by Native macroplot density takes the source population into account. Occasionally a species was recorded in an Edge microplot but was not recorded in Native microplots. For these species, if the mean Edge macroplot density was <1.0, the species was considered an accidental and dropped from analysis for that site. If mean Edge macroplot density was >1.0, mean Native macroplot density
was assigned the lowest possible value (0.083, 1 plant in 12 microplots). Only species occurring in density microplots in at least two Edge macroplots with a mean density 1.0 in one of them were used in analyses. This helps insure that species are appropriate for a wide variety of sites.
We quantified the ability of native species to disperse using Center Density Change, the mean density in the Edge macroplot subtracted from the mean density in the Center macroplot (CDC = Mean DensityCenter l Mean DensityEdge).
We used principal components analysis (PCA) to explore the relationship among soil variables and paired-sample t-tests to compare soil variables between native vegetation and adjacent crested wheatgrass fields across all sites and plant cover between Edge and Center macroplots. Statistical significance was assigned at P0.05; P-values were not adjusted for
multiple comparisons.
We use linear regression analysis to assess the relationship between canopy cover of crested wheatgrass in Edge and Native macroplots as well as soil variables and the cover of sagebrush and other native species. We used a two-sample t-test to assess the difference in crested wheatgrass canopy cover and cover of sagebrush between heavy and light soils. One-sample
t-test were used to determine which plants had greater density in edge compared to center macroplots. We employed analysis of variance (ANOVA) to test the difference in Edge macroplot sagebrush cover between heavy and light soil with Native macroplot cover as a covariate.
We used the nonparametric Mann-Whitney test to determine the effect of soil texture on the ability of native species to coexist with crested wheatgrass and sagebrush density between native and edge macroplots because the raw data did not meet the assumptions of parametric tests. If the test statistic had a P-value <0.20, the species was considered to have a preference for soil texture. Fisher`s exact test was used to assess the distribution of native species on the two soil texture classes. If the P-value was <0.20, the species was considered to preferentially occur on one soil texture.
Results
Soils
There was no difference between pairs of Native and Center crested wheatgrass macroplots across all sites (n=24) for pH (P=0.64), electrical conductivity (P=0.21), % sand (P=0.78), % silt (P=0.12), or % clay (P=0.34). There were strong correlations among all soil variables except salinity. The first principal component in the Center macroplot soil PCA explained 65% of the variation in the five soil variables and had strong positive loadings by % sand (0.97) and pH (0.75) and strong negative loadings by % silt (-0.93) and % clay (-0.85). The second principal component explained 20% of the variation in the soil variables and was dominated by positive loading of salinity (electrical conductivity, 0.89).
Percent sand in soils of crested wheatgrass fields was bimodally distributed with 50% the low point between the two modes of the distribution. Thus we considered soils with <50% sand to be "heavy," these include clays, clay loams and loams. Soils with >50% sand were considered "light" and included sandy loams and loamy sands.
Crested wheatgrass and soil
Canopy cover of crested wheatgrass was positively associated with % silt in Edge (r2=0.30, P=0.006) and Center macroplots (r2=0.13, P=0.08) but negatively associated with % sand in Edge (r2=0.18, P=0.039) and Center (r2=0.13, P=0.09) macroplots. Canopy cover of crested wheatgrass was also negatively associated with pH in Edge macroplots (r2=0.28, P=0.008) but not in Center macroplots (P=0.61). Salinity (electrical conductivity) and % clay were not strongly associated with crested wheatgrass cover (r<0.29). Mean canopy cover of crested wheatgrass was 63% in heavy soils and 49% in light soils in Edge macroplots across all sites, but this difference was not statistically significant (t=1.57, P=0.14). The difference in crested wheatgrass cover between texture classes was even smaller in Center macroplots (P=0.70).
Canopy cover of crested wheatgrass in Native macroplots varied from 0 to 14% (mean=2%) with one outlier (Milwaukee) having 56%. Canopy cover of crested wheatgrass in Native macroplots was positively associated with its cover in Edge macroplots (r2=0.17, P=0.046). Crested wheatgrass cover was greater than 5% in Native macroplots at only three sites(Milwaukee, Wolf, Grass Range), and these all had heavy soils. Mean crested wheatgrass cover was 5.5% in Native macroplots with heavy soils but only 0.5% in those with light soil, but this difference was not statistically significant (t=1.4, P=0.18). Canopy cover of crested wheatgrass was weakly positively associated with % silt (r2=0.15, P=0.058) in Native macroplots.
Sagebrush
Sagebrush occurred in microplots in either or both Edge and Center macroplots at ten sites. Mean microplot density of sagebrush at these ten sites was 0.61 and 0.31 in Edge and Center macroplots respectively and this difference was marginally significant (Mann-Whitney U=29, P=0.10). There was a marginally significant trend for sagebrush cover at the edge of
crested wheatgrass fields to increase as the cover increased in adjacent native vegetation (r2=0.21, P=0.059).
Canopy cover of sagebrush in Native macroplots varied from 0 to 29% across the study sites (mean = 13%) and was greater than 1% at 18 of 24 sites. Half of the six no-sagebrush sites were on heavy soil and half on light soil, and there was no difference in mean sagebrush canopy cover between the two types of soil in Native macroplots where sagebrush occurred (P=0.71). Correlations between sagebrush cover and soil variables in native vegetation were all weak (r0.15). Sagebrush canopy cover at the edge of the crested wheatgrass fields did not differ between heavy and light soils after correcting for adjacent native cover (P=0.25). There were no strong correlations between sagebrush canopy cover at the edge of crested wheatgrass fields and any crested wheatgrass field soil variables (r0.22).
Native species and crested wheatgrass
Total canopy cover of all native species in Edge macroplots declined significantly as the canopy cover of crested wheatgrass increased (r2= 0.35, P=0.002). This decline was due primarily to the decline in total canopy cover of native grasses with increasing canopy cover of crested wheatgrass (r2= 0.37, P=0.002). There was also a negative association between the cover of crested wheatgrass and cover of native forbs and native shrubs, but these relationships were not statistically significant (P>0.17).
We recorded 191 species of vascular plants (not including Selaginella densa) in the macroplots across all sites. Of these, 152 were native perennials. Forty-one species of native perennials occurred in microplots of Native macroplots at one or more sites, but only 36 of these occurred at more than one site with a mean density >1 in the Edge macroplot. We relativized the
Edge Density Ratio values within each site (EDR/sum of EDRs) to give each site equal weight in comparing the performance of native species in occupying the edge of crested wheatgrass fields across all sites (sum of relativized CI values equals 1.0 for each site). The sum of these relativized EDR values for an individual species is a good measure of how well that species was
able to coexist with crested wheatgrass because it integrates mean species abundance in Edge macroplots with frequency across these macroplots. The sum of EDR values and number of sites in which each species occurred in Edge macroplots is presented in the two tables below.
Table 3. The co-occurrence of native species with crested wheatgrass in Edge macroplots with heavy soil based on the sum of relativized EDR values across all sites and the number of sites (N) at which each species occurred.
Shrubs N Sum EDR Graminoids N Sum EDR Forbs N Sum EDR
Opuntia polyacantha 7 0.8592168 Koeleria macrantha 12 1.4411399 Vicia americana 14 3.1073337
Artemisia frigida 9 0.6551899 Poa secunda 13 1.0424688 Sphaeralcea coccinea 7 1.3962516
Artemisia tridentata 6 0.2709372 Carex stenophylla 4 0.9727545 Aster falcatus 3 0.9404662
Gutierrezia sarothrae 1 0.118008 Stipa comata 6 0.4873855 Gaura coccinea 1 0.903789
Artemisia cana 1 0.0012113 Stipa viridula 6 0.2323152 Iva axillaris 2 0.6961672
Rosa woodsii 1 0.0002884 Elymus smithii 8 0.1701284 Achillea millefolium 3 0.6681698
Bouteloua gracilis 4 0.1000402 Bahia opposittifolia 1 0.5454068
Aristida longiseta 1 0.0860646 Antennaria parvifolia 2 0.4070356
Sporobolus cryptandrus 1 0.0472512 Lomatium foeniculaceum 4 0.2837278
Elymus lanceolatus 3 0.006359 Allium textile 3 0.1386688
Potentilla pensylvanica 1 0.1299714
Psoralea argophylla 2 0.1132627
Phlox hoodii 1 0.0714286
Lygodesmia juncea 1 0.0531576
Comandra umbellata 2 0.0522589
Cerastium arvense 1 0.0020188
Table 4. The co-occurrence of native species with crested wheatgrass in Edge macroplots with light soil based on the sum of relativized EDR values across all sites and the number of sites (N) at which each species occurred.
Shrubs N Sum EDR Graminoids N Sum EDR Forbs N Sum EDR
Opuntia polyacantha 5 1.035327 Poa secunda 4 1.064049 Psoralea argophylla 1 0.726392
Gutierrezia sarothrae 2 0.415943 Sporobolus cryptandrus 2 0.742689 Sphaeralcea coccinea 6 0.492516
Artemisia tridentata 3 0.226351 Koeleria macrantha 4 0.422414 Achillea millefolium 1 0.400376
Artemisia frigida 4 0.216029 Bouteloua gracilis 5 0.293246 Astragalus missouriensis 1 0.283447
Stipa comata 5 0.212543 Bahia opposittifolia 1 0.23382
Carex filifolia 1 0.121065 Vicia americana 3 0.199408
Elymus smithii 6 0.111379 Astragalus gracilis 2 0.152555
Aristida longiseta 2 0.106539 Aster falcatus 1 0.149243
Carex stenophylla 3 0.086255 Allium textile 3 0.074317
Stipa viridula 2 0.065148 Cerastium arvense 1 0.073244
Lygodesmia juncea 1 0.065045
Cymopterus acaulis 1 0.025088
Gaura coccinea 1 0.005575
Several species that occurred in ten or more sites had higher Edge macroplot density ratios (EDR) in heavy compared to light soils. These were Gaura coccinea (U=23.5, P=0.063), Koeleria macrantha (U=79, P=0.06), Opuntia polyacantha (U=13, P=0.067), and Psoralea argophylla (U=24, P=0.073). Only Allium textile had a significantly higher EDR in light
compared to heavy soils (Mann-Whitney U = 5.0, P=0.025).
Dispersal ability of native plants
Mean total native plant cover was 11.4% in Edge macroplots and only 7.8% in Center macroplots, but this difference was not statistically significant (paired t=1.55, P=0.14). Canopy cover of all three growth forms (shrubs, graminoids, forbs) of native species were greater in Edge compared to Center macroplots, but none of the differences were statistically significant
(P>0.11). The canopy cover of crested wheatgrass did not differ between Edge and Center macroplots across all sites (paired t=1.24, P=0.23).
Of those species that occurred in Center macroplots more than twice, most declined in density across all sites (2=4.27, P=0.04), and the median change in plant density between Edge and Center macroplots was negative for 75% of the 32 species. All shrubs were less common in Center macroplots compared to Edge macroplots. Carex stenophylla and Poa secunda were two grasses that were more common in Center compared to Edge macroplots more often than not. Several forbs were more abundant in Center compared to Edge macroplots across all sites, including Allium textile, Bahia oppositifolia, Lygodesmia juncea, Psoralea argophylla and Sphaeralcea coccinea (see Table 5).
Table 5. Mean density difference and median percent difference in density between Edge and Center macroplots (DensityCenter l DensityEdge / DensityEdge x 100) for native species across N study sites. Species with a negative median value were less common in Center macroplots more often than they were more common. An asterisk (*) indicates P-value 0.10 by one-sample t-test.
Shrubs N Mean density change Median % change
Artemisia frigida 16 -0.656* -80
Artemisia tridentata 10 -0.300 -87
Gutierrezia sarothrae 6 -0.077 -100
Opuntia polyacantha 13 -0.160* -88
Graminoids
Elymus lanceolatus 3 -0.639 -100
Elymus smithii 18 -0.266 -47
Aristida longiseta 4 -0.177 -91
Bouteloua gracilis 16 -0.617 -79
Carex stenophylla 14 +0.607 75
Koeleria macrantha 19 -0.355 -41
Poa secunda 22 +1.821 17
Sporobolus cryptandrus 3 -0.042 -40
Stipa comata 16 -1.737* -77
Stipa viridula 13 -0.006 -80
Forbs N Mean density change Median % change
Achillea millefolium 8 -0.844* -100
Allium textile 10 -0.008 100
Antennaria parvifolia 5 -0.050 -50
Aster falcatus 9 -1.676 -85
Astragalus gracilis 4 -0.094 -75
Astragalus missouriensis 5 -0.267 -100
Bahia oppositifolia 4 +1.104 125
Comandra umbellata 3 -0.667 -100
Gaura coccinea 10 +0.185 -38
Iva axillaris 3 -0.583 -91
Liatris punctata 3 -0.361 -100
Lomatium foeniculaceum 6 -1.514 -90
Lygodesmia juncea 4 +0.667 121
Phlox hoodii 3 -0.194 -100
Psoralea argophylla 8 +0.028 150
Sphaeralcea coccinea 18 +1.704 87
Vicia americana 19 +0.014 0
14
Discussion
Crested wheatgrass and soil
There was no evidence that soils in the center of crested wheatgrass field differed in any
systematic way from those still supporting native vegetation (unplowed) at the edge of the fields.
This result suggests that the overall dearth of native species in crested wheatgrass fields was due
to either the presence of the crested wheatgrass or poor dispersal of the natives. There may have
been differences in soil organic matter between soils supporting native and crested wheatgrass,
but we did not measure this variable.
Our results suggest that crested wheatgrass can attain and maintain dense stands across
the range of soil textures encompassed by our study, clay to loamy sand. Within this range
crested wheatgrass canopy cover increased with the proportion of silt and showed a tendency to
decrease with increasing sand. These results are consistent with reports that crested wheatgrass
can be established on a wide range of soils (Holechek 1981, Rogler and Lorenz 1983, Johnson
1986, Krzic et al. 2000).
Crested wheatgrass invasion from planted fields into adjacent native vegetation averaged
2% canopy cover in spite of the long time period since introduction in the 1930`s. This is
surprising because crested wheatgrass is reported to be a pernicious invader in the prairie
provinces of Canada (Heidinga and Wilson 2002, Henderson and Naeth 2005, Hansen and
Wilson 2006). Our results suggest that heavier, especially siltier soils are more likely to support
greater invasion of crested wheatgrass into native prairie, but the relationship between soil
texture and invasiveness was not strong.
Native species and crested wheatgrass
The negative relationship between the canopy cover of crested wheatgrass and cover of
native vegetation at the margins of crested wheatgrass fields suggests that crested wheatgrass is
interfering with native species, especially grasses, establishing and/or persisting in the fields.
This assumes that dispersal limitation is unimportant because of the proximity of Edge
macroplots to native vegetation and the amount of time over which plants had the opportunity to
disperse (but see below). The paucity of native species in crested wheatgrass fields is a common
observation (Looman and Heinrichs 1973, Anderson and Marlette 1986, Wilson 1989, Heidinga
15
and Wilson 2002), and the ability of crested wheatgrass to deter the colonization of native
species has been demonstrated experimentally (Wilson and Pärtel 2003).
Canopy cover of sagebrush in native vegetation was variable and not strongly related to
soil texture, pH or salinity across our study sites. These results are at odds with previous reports
stating that Wyoming big sagebrush occured more often on fine-textured compared to sandy
soils in eastern Montana (Morris et al. 1976) and was found on sandy loams in southeast
Montana only if soil depths were shallow (Vanderhorst et al. 1998). In the northern Great Basin
Davies et al. (2007) found the cover of Wyoming big sagebrush was negatively associated with
the sand content of the upper 15 cm of the soil profile. We speculate that the presence of
sagebrush in native vegetation depends more on fire history (Cooper et al. 2007) and perhaps
grazing history than on edaphic differences. The positive relationship between the canopy cover
of sagebrush in native vegetation and Edge macroplots and the fact that sagebrush density
decreased between the edges and centers of crested wheatgrass fields suggests that sagebrush
invasion into crested wheatgrass fields is limited by poor dispersal (Beetle 1960) and/or an
inability to establish well from seed except under special conditions.
Thirty-six native species that occurred in microplots in crested wheatgrass fields were
ranked by the sum of Edge Density Ratios (EDR) across all sites (Table 3, 4). Sum EDR
integrates the abundance of a species in crested wheatgrass fields and the number of sites it
occurred in. Several of these species occurred in only one soil textural class (heavy or light) or
performed differently on the two types of soil. Consequently we ranked species by sum EDR
separately for the two soil types. Opuntia polyacantha, Artemisia frigida and A. tridentata were
the most successful shrubs and Koeleria macrantha, Poa secunda, Carex stenophylla, Stipa
comata and S. viridula were the most successful graminoids to invade crested wheatgrass fields
with heavy soil. Vicia americana, Sphaeralcea coccinea, Aster falcatus, Gaura coccinea, Iva
axillaris, Achillea millefolium, Bahia opposittifolia, Antennaria parvifolia, and Lomatium
foeniculaceum were common forbs invading crested wheatgrass fields on heavy soil. Many of
these mgood invadersn were the same for crested wheatgrass fields with light soil. However the
grasses Sporobolus cryptandrus and Bouteloua gracilis ranked higher in light soils as did the
forbs Psoralea argophylla, Astragalus missouriensis, and A. gracilis. Many of these same
species, especially Artemisia frigida, Koeleria macrantha, Stipa comata, Sphaeralcea coccinea,
Elymus smithii, Bouteloua gracilis and Poa secunda, are reported to occur in old crested
16
wheatgrass fields in southern Canada (Looman and Heinrichs 1973). Poa secunda is reported to
increase as crested wheatgrass increased (Heidinga and Wilson 2002).
Our results differ from previous research in one significant way. Studies conducted in
southern Saskatchewan (Looman and Heinrichs 1973, Heidinga and Wilson 2002, Bakker and
Wilson 2004) indicate that Elymus smithii and Bouteloua gracilis were common in old crested
wheatgrass fields. This is what would be expected based on competition/niche theory (Keddy
1989, Chasde and Leibold 2003) because crested wheatgrass is a caespitose, cool-season grass
while E. smithii is rhizomatous, and B. gracilis is a warm-season grass. Our results were
strikingly different; the cool-season bunchgrasses, Koeleria macrantha and Stipa comata were
much more abundant in the crested wheatgrass fields than either E. smithii or B. gracilis. Soils
in the Canadian studies were reported to be loam or clay loam (Wilson et al. 2004) which is
similar to soils in the majority of our sites. We have no explanation for this disparity in results
unless residual native vegetation of the study sites in Saskatchewan were strongly dominated by
E. smithii and B. gracilis and these two species provided the majority of native grass propagules.
Native species dispersal ability
Poor dispersal may also partly explain the paucity of native species in crested wheatgrass
fields. Macroplots in the center of crested wheatgrass fields had similar canopy cover of crested
wheatgrass but were, on average, three times farther from native vegetation than Edge
macroplots. The great majority of native species were less abundant in Center compared to Edge
macroplots, suggesting that dispersal is limiting reinvasion of many native species. Our results
suggest that this mlimited dispersal effectn is not strong because it did not result in a statistically
significant difference in overall native canopy cover between vegetation in the edge and centers
of the fields.
Our results also suggest that some native species are able to disperse better than others
(higher values in Table 5), and this effect has persisted even after many decades. Native species
showing little difference in mean microplot density between Edge and Center macroplots or with
greater density in Center macroplots across sites (mean density change >-0.10) are assumed to
have good dispersal abilities. Poa secunda, Carex stenophylla, Sporobolus cryptandrus and
Stipa viridula were grasses that showed good dispersal as did the shrub Gutierrezia sarothrae
(Table 5). The forbs Allium textile, Antennaria parvifolia, Astragalus gracilis, Bahia
17
oppositifolia, Gaura coccinea, Lygodesmia juncea, Psoralea argophylla, Sphaeralcea coccinea
and Vicia Americana appeared to be as common in Center macroplots as Edge macroplots.
There is no reason to believe that dispersal of these aforementioned species into crested
wheatgrass fields is limited.
Large differences between Center and Edge macroplot densities indicate poor dispersal
ability (mean density change -0.30). Such species include the shrubs, Artemisia frigida, A.
tridentata, and Opuntia polyacantha, the grasses, Elymus lanceolatus, Bouteloua gracilis,
Koeleria macrantha and Stipa comata, and the forbs, Achillea millefolium, Aster falcatus,
Comandra umbellate, Iva axillaris, Liatris punctata and Lomatium foeniculaceum. It is possible
that the abundance of native species in Edge macroplots may be limited by poor dispersal to
some extent even though Edge macroplots were adjacent to Native macroplots. Poor dispersal of
native species into crested wheatgrass fields has been reported for the Intermountain region of
Idaho (Marlette and Anderson 1986). There is accumulating evidence that range expansions of
many species is limited by dispersal (Primack and Miao 1992, Eriksson 1998, Marisco and
Hellmann 2009), and restorations will require active seeding (Bush et al. 2007).
Restoration recommendations
Crested wheatgrass was somewhat more common and more invasive on heavier and
especially siltier soils, but these differences were not great. Furthermore, crested wheatgrass
invasion of adjacent native vegetation was minimal, and there was little difference in invasion
between heavy and light soils. These results suggest that there is little reason to choose crested
wheatgrass fields for restoration based on soils.
The general paucity of native species in crested wheatgrass fields more than seven
decades after they were planted and the negative relationship between cover of native species
and crested wheatgrass indicates that cover of crested wheatgrass will have to be diminished, at
least temporarily, in order to establish significant number of native plants. It is unlikely that any
treatment outside of draconian measures will eliminate crested wheatgrass from these fields
(Ambrose and Wilson 2003, Hulet et al. 2009, Fansler and Mangold 2011). However, temporary
reduction of crested wheatgrass dominance through herbicide application may allow the early
establishment of natives (Bakker and Wilson 2004) that will allow conversion of fields from
virtual monocultures to more diverse grasslands that support increased native vertebrate and
18
invertebrate diversity. Crested wheatgrass cover can be diminished through plowing or herbicide
application. Plowing or disking is unlikely to be very successful because these methods distuirb
the soil, and older crested wheatgrass stands have a large seed bank and re-establish quickly from
seed (Anderson and Marlette 1986, Henderson and Naeth 2005). Herbicide application will not
eliminate crested wheatgrass (Ambrose and Wilson 2003). However, we recommend using
herbicide to weaken crested wheatgrass stands. This protocol will minimize disturbance and
crested wheatgrass emergence and provide a low-competition window for native plant
establishment.
Our study suggests that the majority of native species have a limited ability to disperse
even distances as little as 300 meters over a period of several decades. Native species will have
to be seeded into crested wheatgrass fields after they have been treated with herbicide. This will
be especially true for poor dispersers such as sagebrush, Artemisia frigida, Stipa comata, Aster
falcatus and Lomatium foeniculaceum (Table 5). Repeated seeding may be needed if weather
conditions are suboptimal to any great extent (Wilson et al. 2004).
Results of our study provide guidance on which species are able to successfully coexist
with crested wheatgrass. We recommend that species with high sum relativized EDR values
(Tables 3, 4) should be seeded into herbicide-treated fields in order to provide enhanced
biological diversity. Native grass species will be used by ungulates as well as livestock.
Increased abundance and diversity of forbs provide habitat for birds, including sage grouse
(Peterson 1970, Drut et al. 1994), as well as insects, including native pollinators (Dramstad and
Fry 1995). Although there was broad overlap among the native species present in crested
wheatgrass fields on different soils, there were some species, especially forbs, that occurred
preferentially on one of the two soil types. Our annotated species lists for light and heavy soils
are presented in Table 6.
Our recommendations omit some species that had high sum EDR values. Prickly
pear cactus (Opuntia polyacantha) was one of the most common woody plants in crested
wheatgrass fields, and it has large flowers that attract insects. However, planting cactus may not
be desirable in pastures that are being grazed by livestock. Narrow-leaved sedge (Carex
stenophylla) was one of the most common native species in crested wheatgrass fields. Although
it is palatable, it is small and provides little plant cover. Sandberg bluegrass (Poa secunda) is an
early, small bunchgrass that provides minimal forage and could be left out of seed mixtures.
19
However, it is competitive with introduced annual bromes, and several cultivars are
commercially available.
Forbs in both seed mixtures show a range of phenologies. Some species such as Vicia
americana and Astragalus missouriensis flower and produce fruit early in the growing season,
while others, such as Bahia oppositifolia and Aster falcatus are active throughout most of the
summer. Having a mixture of forbs that are active at different times helps provide adequate
habitat for birds such as sage grouse and generalist pollinators such as bumblebees.
Table 6. Species recommended for crested wheatgrass restoration seed mixes for heavy and light
soils (see Methods). Forbs are categorized according to when they flower: early (E), mid (M), or
late (L) phenology classes.
Heavy soil
Shrubs
Artemisia frigida
Artemisia tridentata
Gutierrezia sarothrae
Grasses
Koeleria macrantha
Poa secunda
Stipa comata
Stipa viridula
Elymus smithii
Forbs
Vicia americana (E-M)
Sphaeralcea coccinea (M)
Aster falcatus (L)
Gaura coccinea (M)
Iva axillaris (M)
Achillea millefolium (M)
Bahia oppositifolia (M-L)
Antennaria parvifolia (E)
Lomatium foeniculaceum (E)
Allium textile (M)
Potentilla pensylvanica (M)
Psoralea argophylla (M)
Light Soil
Shrubs
Gutierrezia sarothrae
Artemsisia tridentata
Artemisia frigida
Grasses
Koeleria macrantha
Poa secunda
Bouteloua gracilis
Stipa comata
Sporobolus cryptandrus
Elymus smithii
Forbs
Psoralea argophylla (M)
Sphaeralcea coccinea (M)
Achillea millefolium (M)
Astragalus missouriensis (E)
Bahia oppositifolia (M-L)
Vicia americana (E-M)
Astragalus gracilis (M)
Aster falcatus (L)
Allium textile (M)
Cerastium arvense (E)
Lygodesmia juncea (M)
20
Acknowledgements Adam Carr, Dan Brunkhorst, Mike Barrick and Dustin Crowe of BLM
identified potential study sites for us. Bill and Dana Milton allowed us to collect data on their
ranch. We are grateful to the many ranch families that allowed us to conduct our study on their
grazing allotments.
Literature Cited
Ambrose, L. G. and Wilson, S. D. 2003. Emergence of the introduced grass Agropyron cristatum and the native
grass Bouteloua gracilis in a mixed-grass prairie restoration. Restoration Ecology 11: 110l115.
Bakker, J. D. and Wilson, S. D. 2004. Using ecological restoration to constrain biological invasion. Journal of
Applied Ecology 41: 1058l1064.
Beetle, A. A. 1960. A study of sagebrush: the section Tridentatae of Artemisia. Wyoming Agricultural Experiment
Station Bulletin 368.
Box, T.W. 1986. Crested wheatgrass: Its values, problems and myths; where now? In:
K.L. Johnson (ed.), Crested wheatgrass: Its values, problems and myths. Symposium
proceedings. Utah State University, Logan. pp. 343-345.
Bush, R. T., T. R. Seastedt and D. Buckner. 2007. Plant community response to the decline of diffuse knapweed in
a Colorado grassland. Ecological Restoration 25: 169-174.
Chase, J. M. and M. A. Leibold. 2003. Ecological niches. University of Chicago Press, Chicago.
Christian, J. M., and S. D. Wilson. 1999. Long-term ecosystem impacts of an introduced grass in the northern Great
Plains. Ecology 80:2397l2407.
Cooper, S. V., P. Lesica and G. M. Kudray. 2007. Post-fire Recovery of Wyoming Big Sagebrush Shrub-steppe in
Central and Southeast Montana. Report to USDI Montana BLM, Montana Natural Heritage Program, Helena.
Davies, K. W., J. D. Bates and R. F. Miller. 2007. Environmental and vegetation relationships of the Artemisia
tridentata ssp. wyomingensis alliance. Journal of Arid Environments 70: 478-494.
Day, P. A. 1965. Particle fractionation and particle-size analysis. In, C. A. Black, editor-in-chief mMethods of soil
analysis, Part 1n American Society of Agronomy, Inc., Madison, WI.
Dewey, D. R. 1986. Taxonomy of the crested wheatgrasses. In: K.L. Johnson (ed.), Crested wheatgrass: Its values,
problems and myths. Symposium proceedings. Utah State University, Logan. pp. 31-44.
Dorn, R. D. 1984. Vascular Plants of Montana. Mountain West Publishing, Cheyenne, Wyoming.
Dramstad, W. and G. Fry. 1995. Foraging activity of bumblebees (Bombus) in relation to flower resources on
arable land. Agriculture, Ecosystems and Environment 53: 123-135.
Drut, M. S., W. H. Pyle and J. A. Crawford. 1994. Diets and food selection of sage grouse chicks in Oregon.
Journal of Range Management 47: 90-93.
Eriksson, A. 1998. Regional distribution of Thymus serpyllum: management history and dispersal limitation.
Ecography 21: 35-43.
Hansen, M. J. and Wilson, S. D. 2006. Is management of an invasive grass Agropyron cristatum contingent on
environmental variation? Journal of Applied Ecology 43: 269l280.
21
Heidinga, L. and S. D. Wilson. 2002. The impact of an invading alien grass (Agropyron cristatum) on species
turnover in native prairie. Diversity and Distributions 8: 249-258.
Henderson, D. C. and Naeth, M. A. 2005. Multi-scale impacts of crested wheatgrass invasion in mixed-grass prairie.
Biological Invasions 7: 639l650.
Holchek, J.L. 1981. Crested wheatgrass. Rangelands 3:237-250.
Johnson, K. L. 1986. The social values of crested wheatgrass: pros, cons and tradeoffs. In: K.L. Johnson (ed.),
Crested wheatgrass: Its values, problems and myths. Symposium proceedings. Utah State University, Logan. pp.
331-335.
Keddy, P. A. 1989. Competition. Chapman & Hall, London.
Knowles, R.P., and E. Buglass. 1980. Crested wheatgrass. Agriculture Canada
Publication 1295, Ottawa.
Maja Krzic, M., K. Broersma, D. J. Thompson and A. A. Bomke. 2000. Soil Properties and Species Diversity of
Grazed Crested Wheatgrass and Native Rangelands. Journal of Range Management 53: 353-358.
Lavin, M. and C. Seibert. 2011. Grasses of Montana. http://www.montana.edu/mlavin/herb/mtgrass.pdf.
Lesica, P. and T. H. DeLuca. 1996. Long-term harmful effects of crested wheatgrass on Great Plains grassland
ecosystems. Journal of Soil and Water Conservation 51: 408l411.
Lloyd, J. D. and T. E. Martin. 2005. Reproductive success of Chestnut-collared Longspurs in native and exotic
grasslands. Condor 107: 363-374.
Looman, J., and D.H. Hendrichs. 1973. Stability of crested wheatgrass pastures under
long-term pasture use. Canadian Journal of Plant Science 53: 501-506.
Lorenz, R.J. 1986. Introduction and early use of crested wheatgrass in the Northern Great
Plains. In: K.L. Johnson (ed.), Crested wheatgrass: Its values, problems and myths.
Symposium proceedings. Utah State University, Logan. pp. 9-19.
Majerus, M., Holzworth, L., Tilley, D., Ogle, D., Stannard, M. 2009. Plant Guide for Sandberg bluegrass (Poa
secunda J. Presl). USDA-Natural Resources Conservation Service, Idaho Plant Materials Center, Aberdeen, ID.
Marlette, G. M. and J. E. Anderson. 1986. Seed banks and propagule dispersal in crested wheatgrass stands.
Journal of Applied Ecology 23: 161-175.
McHenry, J.R., and L.C. Newell. 1947. Influence of some perennial grasses on the
organic matter content and structure of an eastern Nebraska fine-textured soil. Journal of the American Society of
Agronomy 39: 981-994.
Morris, M. S., R. G. Kelsey and D. Griggs. 1976. The geographic and ecological distribution of big sagebrush and
other woody Artemisias in Montana. Proceedings of the Montana Academy of Sciences 36: 56-79.
Peterson, J. G. 1970. Food habits and summer distribution of juvenile sage grouse in central Montana. Journal of
Wildlife Management 34: 147-154.
Primack, R. B. and Miao, S. L. 1992. Dispersal can limit local plant distribution. Conservation Biology 6: 513-519.
Reynolds, T. D. and C. H. Trost. 1980. The response of native vertebrate populations to crested wheatgrass
planting and grazing by sheep. Journal of Range Management 33: 122-125.
22
Rogler, G.A., and R.J. Lorenz. 1983. Crested wheat-grass-early history in the United States. Journal of Range
Management 36: 91-93.
Romo, J. T. 2005. Emergence and establishment of Agropyron desertorum Fisch. (crested wheatgrass) seedlings in
a sandhills prairie of central Saskatchewan. Natural Areas Journal 25: 26-35.
Schroeder, M. A. and W. M. Vander Haegen. 2006. Use of Conservation Reserve Program fields by greater sage
grouse and other shrubsteppe-associated wildlife in Washington state. Technical report prepared for USDA Farm
Service Agency. Washington Department of Fish and Wildlife, Olympia, WA.
Smoliak, S., A. Johnston, and L.E. Lutwick. 1967. Productivity and durability of crested
wheatgrass in southeastern Alberta. Canadian Journal of Plant Science 47: 539-548.
Urness, P. J. 1986. Value of crested wheatgrass for big game. Page 147-153 in K. L. Johnson (ed.), Crested
wheatgrass: its values, problems and myths. Utah State University, Logan, UT.
Vanderhorst, J., B. L. Heidel, and S. V. Cooper. 1998. Botanical and vegetation survey of Carter County. Montana.
Unpublished report to Bureau of Land Management. Montana Natural Heritage Program, Helena. 1 16 pp. + app.
Vaness, B. M. and S. D. Wilson. 2007. Impact and management of crested wheatgrass (Agropyron cristatum) in the
northern Great Plains. Canadian Journal of Plant Science 87: 1023l1028.
Wilson, S.D. 1989. The suppression of native prairie by alien species introduced for
revegetation. Landscape and Urban Planning 17:113-119.
Wilson, S. D. and Pärtel, M. 2003. Extirpation or coexistence? Management of a persistent introduced grass in a
prairie restoration. Restoration Ecology 11: 410l416.
Wilson, S. D., J. D. Bakker, J. M. Christian, X. Li, L. G. Ambrose and J. Waddington. 2004. Semiarid Old-Field
Restoration: Is Neighbor Control Needed? Ecological Applications 14: 476-484.
Old fields dominated by crested wheatgrass are common throughout much of eastern Montana. Many of these fields are virtual monocultures, providing relatively little benefit to livestock or wildlife. Understanding secondary succession in these old fields in relation to soils can aid in planning reclamation activities. Our study addresses whether soil properties are related to the degree to which crested wheatgrass fields have been invaded by natives and the degree to which crested wheatgrass has invaded native vegetation. We also determine which native species are most successful at coexisting with crested wheatgrass and on which soils.
We conducted our study at 24 sites in central Montana where there were 75-year-old crested wheatgrass fields adjacent to unplowed native vegetation. Soil textures ranged from clay to loamy sand. We located three 0.8 ha macroplots at each site, one in the native vegetation and two in the crested wheatgrass field, one adjacent to native vegetation and one in the center of the field. We measured canopy cover of all species and density of native species in randomly located plots within each macroplot. We also sampled and analyzed soil in native and crested wheatgrass macroplots.
There was no difference in measured soil variables between native vegetation and crested wheatgrass fields. Greater canopy cover of crested wheatgrass was associated with heavier, especially siltier soils, but the associations were not strong (r20.30). Mean crested wheatgrass cover in native vegetation was low (2%) with the exception of one site with 56%. Total canopy cover of all native species at the edge of crested wheatgrass fields declined significantly as the canopy cover of crested wheatgrass increased, due primarily to the decline in total canopy cover of native grasses. Only 36 native perennials were common at the edge of crested wheatgrass fields. The majority (75%) of the native perennials had lower abundance in the center of crested wheatgrass fields compared to the edge, indicating that most species had limited dispersal.
Our results suggest that there is little reason to choose among crested wheatgrass fields for restoration based on soils. The general paucity of native species in crested wheatgrass fields more than seven decades after they were planted and the negative relationship between cover of native species and crested wheatgrass indicates that cover of crested wheatgrass will have to be diminished, at least temporarily, in order to establish significant number of native plants. We recommend using herbicide to weaken crested wheatgrass stands. This protocol will minimize disturbance and crested wheatgrass emergence and provide a low-competition window for native plant establishment.
Our study suggests that the majority of native species have a limited ability to disperse so native species will have to be seeded into crested wheatgrass fields after they have been treated with herbicide. Results of our study provide guidance on which species are able to successfully coexist with crested wheatgrass. Although there was broad overlap among the native species present in crested wheatgrass fields with either heavy or light soil, there were some species, especially forbs, that occurred preferentially on one of the two soil types. Our recommendations omit some species that were common in crested wheatgrass fields. These include prickly pear cactus and narrow-leaved sedge. The former has large flowers that are attractive to pollinators but is unpalatable to livestock and most wildlife. The latter produces only very small amounts of palatable forage. Forbs in recommended seed mixtures show a range of phenologies. Some species flower and produce fruit early in the growing season, while others are active throughout most of the summer. Having a mixture of forbs that are active at different times helps provide adequate habitat for birds such as sage grouse and generalist pollinators such as bumblebees.
Introduction
Much public land in eastern Montana is retired agricultural fields planted to crested wheatgrass 70-80 years ago. These old fields provide little benefit for livestock or wildlife. Many crested wheatgrass fields have resisted invasion by native species and are still virtual monocultures, even after more than seven decades (Christian and Wilson 1999). Monoculture crested wheatgrass fields lack the shrubs and forbs to support strong populations of antelope, sage grouse and Baird`s sparrow compared to native prairie or sagebrush steppe (Reynolds and Trost 1980, Urness 1986). Lloyd and Martin (2005) found that nestlings of chestnut-collared longspurs grew more slowly and had a 17% lower chance of surviving in crested wheatgrass fields compared to native prairie of northeast Montana. However, studies in Washington suggest that crested wheatgrass can provide suitable nesting habitat for sage grouse when it has been reinvaded by sagebrush (Schroeder and Vander Haegen 2006).
Some crested wheatgrass fields are returning to native composition, but most are not (McHenry and Newell 1947, Smoliak et al. 1967, Looman and Heinrichs 1973, Box 1986, Marlette and Anderson 1986, Wilson 1989). Crested wheatgrass has been successfully planted on a wide variety of soils (Knowles and Buglass 1980), and edaphic differences are one possible explanatory variable for differential rates of old-field secondary succession. Furthermore, the identity of native species able to invade crested wheatgrass fields may depend on soil texture. Understanding secondary succession in these old fields in relation to soils can aid in planning reclamation activities.
In some cases crested wheatgrass has invaded native vegetation from adjacent planted fields (Christian and Wilson 1999, Heidinga and Wilson 2002, Bakker and Wilson 2004, Henderson and Naeth 2005, Romo 2005, Vaness and Wilson 2007). Knowing which native habitats are more prone to invasion by crested wheatgrass will help managers prioritize old-field
restoration efforts.
Our study describes secondary succession of crested wheatgrass fields to native prairie in central Montana with an experimental design inadvertently begun ca. 75 years ago when crested wheatgrass fields were planted in a mosaic of native steppe. Our study addresses whether soil properties are related to the degree to which crested wheatgrass fields have been invaded by natives and crested wheatgrass has invaded native vegetation from nearby fields. Our study will also determine which native species are most successful at coexisting with crested wheatgrass and on which soils. These results have important implications for restoration of these fields to more diverse native composition.
Methods
Study area and the history of crested wheatgrass introductions
Crested wheatgrass was first introduced into the U.S. in 1898, and the first successful experimental plantings were at U.S. Department of Agriculture stations in 1907 to 1910 (Rogler and Lorenz 1983, Lorenz 1986). The first planting in Montana was in 1915. The crested wheatgrass cultivar, Fairway, was developed in Canada at the same time as a similar selection in
Montana. It is a diploid and can be referred to Agropyron cristatum var. pectinatum (M. Bieb.) Roshev. ex B. Fedtsch. (Dewey 1986). Fairway was the only cultivar in mass production at the time, and was used almost exclusively in reclaiming Montana LU lands in the late 1930`s.
Crested wheatgrass has been planted on 6 to 10 million ha in the northern Great Plains (Lesica & DeLuca 1996); much of this occurred in the first half of the last century as a response to abandoned marginal cropland (Lorenz 1986). The region was settled by homesteaders in the early 1900`s and millions of hectares of native prairie were plowed and planted to wheat. Wheat farming was profitable until the great drought of 1928-1937. During the drought there was a mass exodus from the region, and most of the farm fields were abandoned. The Agricultural Adjustment Act of 1933 and the Bankhead/Jones Act of 1937 authorized the U.S. Department of Agriculture to purchase these abandoned homesteads (Land Utilization or LU lands), and the Soil Conservation Service was given the responsibility of stabilizing the soil and returning the land to productivity (Lorenz 1986).
LU lands administered by the Bureau of Land Management in Montana are found mainly from the north-central portion of the state just north of the Missouri River through Fallon County in the southeast. Our study sites were all located on LU lands planted to crested wheatgrass ca. 75 years ago and located primarily in Fergus and Petroleum counties with three sites in Musselshell and one each in Chouteau and Yellowstone counties in the east-central part of the state.
Field Methods
The Lewistown District Office of BLM provided a list of 34 potential study sites. We randomly selected 16 sites for sampling in 2010 and four sites in 2011, stratified by soil type based on existing soil surveys. Since most of the study sites in Fergus and Petroleum counties were on heavy soils, we also sampled four sites in Musselshell, Choteau and Yellowstone cos.
that had light soil. Locations are provided in Table 1.
Table 1. Location of study sites and soil variables from the center of the crested wheatgrass fields.
We located three circular, 0.8 ha macroplots (50 m radius) at each site (Table 2). The
Native macroplot was in native vegetation at the edge of the crested wheatgrass field. The Edge
Site name County Location pH EC (S) Soil texture class Soil texture
Erie Fergus S 21 & 20, T20N., R19E 6.64 77 heavy Loam
Maruska Fergus S 28, T19N., R23E 7.56 330 heavy Clay
Chippewa Fergus S 11, T15N., R22E 6.62 132 heavy Clay-loam
Milwaukee Fergus S 8, T20N., R19E 6.95 340 heavy Clay-loam
GrassRange Fergus S 27, T15N., R23E 6.41 132 heavy Loam
Boxelder Petroleum S 23, 24 & 25, T17N., R25E. 7.06 85 heavy Loam
Hansom Petroleum S 25, T18N., R26E 7.72 184 heavy Clay-loam
East Christina Fergus S 25, T18N., R26E 6.81 46 heavy Clay-loam
GrassRangeE Fergus S 8, T15N., R24E 7.01 70 heavy Clay
Wolf Fergus Ss 9, 10, 11, 12, T18N., R24E 6.62 152 heavy Clay
Kosir Fergus S 8, T18N., R24E 8.05 120 heavy Clay
Moulton Fergus S 23, T15N., R23E. 6.91 158 heavy Clay
Duck Creek Fergus S 34 & 35, T17N., R24E 7.46 130 heavy Silty-clay
Windmill Petroleum S 5 & 6, T15N., R28E 7.55 290 light Loam
Maidenhead Fergus S 26 & 35, T16N., R21E 7.12 170 heavy Clay-loam
Manuel Petroleum S 3, T15N., R28E 7.94 160 heavy Loam
Carl Spring Fergus S 7, 8 & 17, T14N., R23E 7.56 227 light Loamy-sand
RifleRange Yellowstone S 8, T3N, R26E 8.13 185 light Loamy-sand
Loma Chouteau S 13, T25N, R9E 7.85 224 light Sandy-loam
West Bohemia Fergus S 7, T18N., R23E 7.32 445 heavy Sandy-clay-loam
Gallatin Fergus Ss 25, 36, 31 T 23 N, R 17,18 E 8.12 185 light Sandy-loam
Milton 1 Musselshell S 3, T9N, R26E 7.83 296 light Loamy-sand
Milton 2 Musselshell S 14, T9N, R26E 7.45 152 light Sandy-loam
Milton 3 Musselshell S 11, T9N R26E 8.2 199 light Sandy-loam
We located three circular, 0.8 ha macroplots (50 m radius) at each site (Table 2). The Native macroplot was in native vegetation at the edge of the crested wheatgrass field. The Edge macroplot was in the crested wheatgrass field near the edge of the native vegetation. In all but three cases the centers of these two macroplots were within 150 m of each other (mean=114m). The Center macroplot was either in the center of the crested wheatgrass field or as far from native vegetation as possible. In all but one case the center of the Center macroplot was greater than 150 m from the center of the Native macroplot (mean=314 m). The Edge and Center macroplots were almost always at least 100 m apart. In majority of cases the three macroplots were in the same grazing pasture.
Table 2. Distance (m) between Native and Edge and between Native and Center macroplots at the 24 study sites.
Site Native-Edge Native-Center Site Native-Edge Native-Center
Erie 116 227 Duck Creek 129 298
Maruska 100 435 Windmill 495 897
Chippewa 60 154 Maidenhead 119 339
Milwaukee 110 530 Manuel 117 286
Grass Range 120 271 Carl Spring 121 247
BoxElder 158 342 Rifle Range 147 373
HansomDam 77 106 Loma 94 223
EChristina 75 201 West Bohemian 645 554
GrassRangeE 95 276 Gallatin 110 327
Wolf Ind 147 315 Milton 1 100 168
Kosir 72 203 Milton 2 115 198
Moulton 232 395 Milton 3 103 162
We located four 0.02 ha sample plots (8 m radius) within each macroplot, one on each principal compass bearing at randomly-chosen distances from the macroplot center. We estimated canopy cover of all vascular plants observed in each sample plot using the following cover classes: T- = <0.05%, T=0.5%, T+=0.6-0.9%, 1%, 2%, 5%, 10%, 15%, etc. We located
three 1-m2 microplots in each sample plot 4 m from plot center, one along each of three evenly spaced radii. The number of individuals of all native perennial species was counted in each microplot. Individual stems or clumps of stems were counted for rhizomatous species.
We used a hand-held GPS device to record the latitude and longitude of the center of each macroplot at each site. We collected three soil samples from haphazardly-chosen locations near plot center in the Native and Center macroplots at each site. Macroplot soil samples were combined and analyzed for particle-size distribution by a modified Bouyoucos hydrometer method (Day 1965). pH and conductivity (a measure of salinity) were measured using hand-held meters. Nomenclature follows Dorn (1984). Wyoming big sagebrush (Artemisia tridentata ssp. wyomingensis) is hereafter referred to as sagebrush. We did not attempt to quantify Selaginella densa in density microplots because individual ramets and/orgenets cannot be distinguished, and the plant is inappropriate for restoration. We attempted to distinguish western wheatgrass (Elymus smithii) from thick-spike wheatgrass (Elymus lanceolatus), but the two species can be very similar (Lavin and Seibert 2011), and we may have erred occasionally.
Data analysis
We used the total canopy cover of perennial native species in Edge macroplots to represent the degree of native plant invasion into crested wheatgrass fields. We used the Edge Density Ratio (EDR = Mean DensityEdge/Mean DensityNative) as the measure of a species` success coexisting with crested wheatgrass. We assumed that dispersal limitation was minimal because
Edge macroplots were immediately adjacent to native vegetation. Density in the Edge macroplots alone is a poor measure of success in the crested wheatgrass field because it does not take into account how large the propagule source is nor the size or growth form of the species compared to other species. For example, a rhizomatous species such as Elymus smithii, could have a mean density of 50 stems/m2 in the Native macroplot and only 5 stems/m2 in the crested field, a 90% reduction. However, a bunchgrass like Stipa viridula might have 4 plants/m2 and 2 plants/m2 in Native and Edge macroplots respectively, only a 50% reduction, and yet have a much lower Edge density than the sod grass. Dividing Edge macroplot density by Native macroplot density takes the source population into account. Occasionally a species was recorded in an Edge microplot but was not recorded in Native microplots. For these species, if the mean Edge macroplot density was <1.0, the species was considered an accidental and dropped from analysis for that site. If mean Edge macroplot density was >1.0, mean Native macroplot density
was assigned the lowest possible value (0.083, 1 plant in 12 microplots). Only species occurring in density microplots in at least two Edge macroplots with a mean density 1.0 in one of them were used in analyses. This helps insure that species are appropriate for a wide variety of sites.
We quantified the ability of native species to disperse using Center Density Change, the mean density in the Edge macroplot subtracted from the mean density in the Center macroplot (CDC = Mean DensityCenter l Mean DensityEdge).
We used principal components analysis (PCA) to explore the relationship among soil variables and paired-sample t-tests to compare soil variables between native vegetation and adjacent crested wheatgrass fields across all sites and plant cover between Edge and Center macroplots. Statistical significance was assigned at P0.05; P-values were not adjusted for
multiple comparisons.
We use linear regression analysis to assess the relationship between canopy cover of crested wheatgrass in Edge and Native macroplots as well as soil variables and the cover of sagebrush and other native species. We used a two-sample t-test to assess the difference in crested wheatgrass canopy cover and cover of sagebrush between heavy and light soils. One-sample
t-test were used to determine which plants had greater density in edge compared to center macroplots. We employed analysis of variance (ANOVA) to test the difference in Edge macroplot sagebrush cover between heavy and light soil with Native macroplot cover as a covariate.
We used the nonparametric Mann-Whitney test to determine the effect of soil texture on the ability of native species to coexist with crested wheatgrass and sagebrush density between native and edge macroplots because the raw data did not meet the assumptions of parametric tests. If the test statistic had a P-value <0.20, the species was considered to have a preference for soil texture. Fisher`s exact test was used to assess the distribution of native species on the two soil texture classes. If the P-value was <0.20, the species was considered to preferentially occur on one soil texture.
Results
Soils
There was no difference between pairs of Native and Center crested wheatgrass macroplots across all sites (n=24) for pH (P=0.64), electrical conductivity (P=0.21), % sand (P=0.78), % silt (P=0.12), or % clay (P=0.34). There were strong correlations among all soil variables except salinity. The first principal component in the Center macroplot soil PCA explained 65% of the variation in the five soil variables and had strong positive loadings by % sand (0.97) and pH (0.75) and strong negative loadings by % silt (-0.93) and % clay (-0.85). The second principal component explained 20% of the variation in the soil variables and was dominated by positive loading of salinity (electrical conductivity, 0.89).
Percent sand in soils of crested wheatgrass fields was bimodally distributed with 50% the low point between the two modes of the distribution. Thus we considered soils with <50% sand to be "heavy," these include clays, clay loams and loams. Soils with >50% sand were considered "light" and included sandy loams and loamy sands.
Crested wheatgrass and soil
Canopy cover of crested wheatgrass was positively associated with % silt in Edge (r2=0.30, P=0.006) and Center macroplots (r2=0.13, P=0.08) but negatively associated with % sand in Edge (r2=0.18, P=0.039) and Center (r2=0.13, P=0.09) macroplots. Canopy cover of crested wheatgrass was also negatively associated with pH in Edge macroplots (r2=0.28, P=0.008) but not in Center macroplots (P=0.61). Salinity (electrical conductivity) and % clay were not strongly associated with crested wheatgrass cover (r<0.29). Mean canopy cover of crested wheatgrass was 63% in heavy soils and 49% in light soils in Edge macroplots across all sites, but this difference was not statistically significant (t=1.57, P=0.14). The difference in crested wheatgrass cover between texture classes was even smaller in Center macroplots (P=0.70).
Canopy cover of crested wheatgrass in Native macroplots varied from 0 to 14% (mean=2%) with one outlier (Milwaukee) having 56%. Canopy cover of crested wheatgrass in Native macroplots was positively associated with its cover in Edge macroplots (r2=0.17, P=0.046). Crested wheatgrass cover was greater than 5% in Native macroplots at only three sites(Milwaukee, Wolf, Grass Range), and these all had heavy soils. Mean crested wheatgrass cover was 5.5% in Native macroplots with heavy soils but only 0.5% in those with light soil, but this difference was not statistically significant (t=1.4, P=0.18). Canopy cover of crested wheatgrass was weakly positively associated with % silt (r2=0.15, P=0.058) in Native macroplots.
Sagebrush
Sagebrush occurred in microplots in either or both Edge and Center macroplots at ten sites. Mean microplot density of sagebrush at these ten sites was 0.61 and 0.31 in Edge and Center macroplots respectively and this difference was marginally significant (Mann-Whitney U=29, P=0.10). There was a marginally significant trend for sagebrush cover at the edge of
crested wheatgrass fields to increase as the cover increased in adjacent native vegetation (r2=0.21, P=0.059).
Canopy cover of sagebrush in Native macroplots varied from 0 to 29% across the study sites (mean = 13%) and was greater than 1% at 18 of 24 sites. Half of the six no-sagebrush sites were on heavy soil and half on light soil, and there was no difference in mean sagebrush canopy cover between the two types of soil in Native macroplots where sagebrush occurred (P=0.71). Correlations between sagebrush cover and soil variables in native vegetation were all weak (r0.15). Sagebrush canopy cover at the edge of the crested wheatgrass fields did not differ between heavy and light soils after correcting for adjacent native cover (P=0.25). There were no strong correlations between sagebrush canopy cover at the edge of crested wheatgrass fields and any crested wheatgrass field soil variables (r0.22).
Native species and crested wheatgrass
Total canopy cover of all native species in Edge macroplots declined significantly as the canopy cover of crested wheatgrass increased (r2= 0.35, P=0.002). This decline was due primarily to the decline in total canopy cover of native grasses with increasing canopy cover of crested wheatgrass (r2= 0.37, P=0.002). There was also a negative association between the cover of crested wheatgrass and cover of native forbs and native shrubs, but these relationships were not statistically significant (P>0.17).
We recorded 191 species of vascular plants (not including Selaginella densa) in the macroplots across all sites. Of these, 152 were native perennials. Forty-one species of native perennials occurred in microplots of Native macroplots at one or more sites, but only 36 of these occurred at more than one site with a mean density >1 in the Edge macroplot. We relativized the
Edge Density Ratio values within each site (EDR/sum of EDRs) to give each site equal weight in comparing the performance of native species in occupying the edge of crested wheatgrass fields across all sites (sum of relativized CI values equals 1.0 for each site). The sum of these relativized EDR values for an individual species is a good measure of how well that species was
able to coexist with crested wheatgrass because it integrates mean species abundance in Edge macroplots with frequency across these macroplots. The sum of EDR values and number of sites in which each species occurred in Edge macroplots is presented in the two tables below.
Table 3. The co-occurrence of native species with crested wheatgrass in Edge macroplots with heavy soil based on the sum of relativized EDR values across all sites and the number of sites (N) at which each species occurred.
Shrubs N Sum EDR Graminoids N Sum EDR Forbs N Sum EDR
Opuntia polyacantha 7 0.8592168 Koeleria macrantha 12 1.4411399 Vicia americana 14 3.1073337
Artemisia frigida 9 0.6551899 Poa secunda 13 1.0424688 Sphaeralcea coccinea 7 1.3962516
Artemisia tridentata 6 0.2709372 Carex stenophylla 4 0.9727545 Aster falcatus 3 0.9404662
Gutierrezia sarothrae 1 0.118008 Stipa comata 6 0.4873855 Gaura coccinea 1 0.903789
Artemisia cana 1 0.0012113 Stipa viridula 6 0.2323152 Iva axillaris 2 0.6961672
Rosa woodsii 1 0.0002884 Elymus smithii 8 0.1701284 Achillea millefolium 3 0.6681698
Bouteloua gracilis 4 0.1000402 Bahia opposittifolia 1 0.5454068
Aristida longiseta 1 0.0860646 Antennaria parvifolia 2 0.4070356
Sporobolus cryptandrus 1 0.0472512 Lomatium foeniculaceum 4 0.2837278
Elymus lanceolatus 3 0.006359 Allium textile 3 0.1386688
Potentilla pensylvanica 1 0.1299714
Psoralea argophylla 2 0.1132627
Phlox hoodii 1 0.0714286
Lygodesmia juncea 1 0.0531576
Comandra umbellata 2 0.0522589
Cerastium arvense 1 0.0020188
Table 4. The co-occurrence of native species with crested wheatgrass in Edge macroplots with light soil based on the sum of relativized EDR values across all sites and the number of sites (N) at which each species occurred.
Shrubs N Sum EDR Graminoids N Sum EDR Forbs N Sum EDR
Opuntia polyacantha 5 1.035327 Poa secunda 4 1.064049 Psoralea argophylla 1 0.726392
Gutierrezia sarothrae 2 0.415943 Sporobolus cryptandrus 2 0.742689 Sphaeralcea coccinea 6 0.492516
Artemisia tridentata 3 0.226351 Koeleria macrantha 4 0.422414 Achillea millefolium 1 0.400376
Artemisia frigida 4 0.216029 Bouteloua gracilis 5 0.293246 Astragalus missouriensis 1 0.283447
Stipa comata 5 0.212543 Bahia opposittifolia 1 0.23382
Carex filifolia 1 0.121065 Vicia americana 3 0.199408
Elymus smithii 6 0.111379 Astragalus gracilis 2 0.152555
Aristida longiseta 2 0.106539 Aster falcatus 1 0.149243
Carex stenophylla 3 0.086255 Allium textile 3 0.074317
Stipa viridula 2 0.065148 Cerastium arvense 1 0.073244
Lygodesmia juncea 1 0.065045
Cymopterus acaulis 1 0.025088
Gaura coccinea 1 0.005575
Several species that occurred in ten or more sites had higher Edge macroplot density ratios (EDR) in heavy compared to light soils. These were Gaura coccinea (U=23.5, P=0.063), Koeleria macrantha (U=79, P=0.06), Opuntia polyacantha (U=13, P=0.067), and Psoralea argophylla (U=24, P=0.073). Only Allium textile had a significantly higher EDR in light
compared to heavy soils (Mann-Whitney U = 5.0, P=0.025).
Dispersal ability of native plants
Mean total native plant cover was 11.4% in Edge macroplots and only 7.8% in Center macroplots, but this difference was not statistically significant (paired t=1.55, P=0.14). Canopy cover of all three growth forms (shrubs, graminoids, forbs) of native species were greater in Edge compared to Center macroplots, but none of the differences were statistically significant
(P>0.11). The canopy cover of crested wheatgrass did not differ between Edge and Center macroplots across all sites (paired t=1.24, P=0.23).
Of those species that occurred in Center macroplots more than twice, most declined in density across all sites (2=4.27, P=0.04), and the median change in plant density between Edge and Center macroplots was negative for 75% of the 32 species. All shrubs were less common in Center macroplots compared to Edge macroplots. Carex stenophylla and Poa secunda were two grasses that were more common in Center compared to Edge macroplots more often than not. Several forbs were more abundant in Center compared to Edge macroplots across all sites, including Allium textile, Bahia oppositifolia, Lygodesmia juncea, Psoralea argophylla and Sphaeralcea coccinea (see Table 5).
Table 5. Mean density difference and median percent difference in density between Edge and Center macroplots (DensityCenter l DensityEdge / DensityEdge x 100) for native species across N study sites. Species with a negative median value were less common in Center macroplots more often than they were more common. An asterisk (*) indicates P-value 0.10 by one-sample t-test.
Shrubs N Mean density change Median % change
Artemisia frigida 16 -0.656* -80
Artemisia tridentata 10 -0.300 -87
Gutierrezia sarothrae 6 -0.077 -100
Opuntia polyacantha 13 -0.160* -88
Graminoids
Elymus lanceolatus 3 -0.639 -100
Elymus smithii 18 -0.266 -47
Aristida longiseta 4 -0.177 -91
Bouteloua gracilis 16 -0.617 -79
Carex stenophylla 14 +0.607 75
Koeleria macrantha 19 -0.355 -41
Poa secunda 22 +1.821 17
Sporobolus cryptandrus 3 -0.042 -40
Stipa comata 16 -1.737* -77
Stipa viridula 13 -0.006 -80
Forbs N Mean density change Median % change
Achillea millefolium 8 -0.844* -100
Allium textile 10 -0.008 100
Antennaria parvifolia 5 -0.050 -50
Aster falcatus 9 -1.676 -85
Astragalus gracilis 4 -0.094 -75
Astragalus missouriensis 5 -0.267 -100
Bahia oppositifolia 4 +1.104 125
Comandra umbellata 3 -0.667 -100
Gaura coccinea 10 +0.185 -38
Iva axillaris 3 -0.583 -91
Liatris punctata 3 -0.361 -100
Lomatium foeniculaceum 6 -1.514 -90
Lygodesmia juncea 4 +0.667 121
Phlox hoodii 3 -0.194 -100
Psoralea argophylla 8 +0.028 150
Sphaeralcea coccinea 18 +1.704 87
Vicia americana 19 +0.014 0
14
Discussion
Crested wheatgrass and soil
There was no evidence that soils in the center of crested wheatgrass field differed in any
systematic way from those still supporting native vegetation (unplowed) at the edge of the fields.
This result suggests that the overall dearth of native species in crested wheatgrass fields was due
to either the presence of the crested wheatgrass or poor dispersal of the natives. There may have
been differences in soil organic matter between soils supporting native and crested wheatgrass,
but we did not measure this variable.
Our results suggest that crested wheatgrass can attain and maintain dense stands across
the range of soil textures encompassed by our study, clay to loamy sand. Within this range
crested wheatgrass canopy cover increased with the proportion of silt and showed a tendency to
decrease with increasing sand. These results are consistent with reports that crested wheatgrass
can be established on a wide range of soils (Holechek 1981, Rogler and Lorenz 1983, Johnson
1986, Krzic et al. 2000).
Crested wheatgrass invasion from planted fields into adjacent native vegetation averaged
2% canopy cover in spite of the long time period since introduction in the 1930`s. This is
surprising because crested wheatgrass is reported to be a pernicious invader in the prairie
provinces of Canada (Heidinga and Wilson 2002, Henderson and Naeth 2005, Hansen and
Wilson 2006). Our results suggest that heavier, especially siltier soils are more likely to support
greater invasion of crested wheatgrass into native prairie, but the relationship between soil
texture and invasiveness was not strong.
Native species and crested wheatgrass
The negative relationship between the canopy cover of crested wheatgrass and cover of
native vegetation at the margins of crested wheatgrass fields suggests that crested wheatgrass is
interfering with native species, especially grasses, establishing and/or persisting in the fields.
This assumes that dispersal limitation is unimportant because of the proximity of Edge
macroplots to native vegetation and the amount of time over which plants had the opportunity to
disperse (but see below). The paucity of native species in crested wheatgrass fields is a common
observation (Looman and Heinrichs 1973, Anderson and Marlette 1986, Wilson 1989, Heidinga
15
and Wilson 2002), and the ability of crested wheatgrass to deter the colonization of native
species has been demonstrated experimentally (Wilson and Pärtel 2003).
Canopy cover of sagebrush in native vegetation was variable and not strongly related to
soil texture, pH or salinity across our study sites. These results are at odds with previous reports
stating that Wyoming big sagebrush occured more often on fine-textured compared to sandy
soils in eastern Montana (Morris et al. 1976) and was found on sandy loams in southeast
Montana only if soil depths were shallow (Vanderhorst et al. 1998). In the northern Great Basin
Davies et al. (2007) found the cover of Wyoming big sagebrush was negatively associated with
the sand content of the upper 15 cm of the soil profile. We speculate that the presence of
sagebrush in native vegetation depends more on fire history (Cooper et al. 2007) and perhaps
grazing history than on edaphic differences. The positive relationship between the canopy cover
of sagebrush in native vegetation and Edge macroplots and the fact that sagebrush density
decreased between the edges and centers of crested wheatgrass fields suggests that sagebrush
invasion into crested wheatgrass fields is limited by poor dispersal (Beetle 1960) and/or an
inability to establish well from seed except under special conditions.
Thirty-six native species that occurred in microplots in crested wheatgrass fields were
ranked by the sum of Edge Density Ratios (EDR) across all sites (Table 3, 4). Sum EDR
integrates the abundance of a species in crested wheatgrass fields and the number of sites it
occurred in. Several of these species occurred in only one soil textural class (heavy or light) or
performed differently on the two types of soil. Consequently we ranked species by sum EDR
separately for the two soil types. Opuntia polyacantha, Artemisia frigida and A. tridentata were
the most successful shrubs and Koeleria macrantha, Poa secunda, Carex stenophylla, Stipa
comata and S. viridula were the most successful graminoids to invade crested wheatgrass fields
with heavy soil. Vicia americana, Sphaeralcea coccinea, Aster falcatus, Gaura coccinea, Iva
axillaris, Achillea millefolium, Bahia opposittifolia, Antennaria parvifolia, and Lomatium
foeniculaceum were common forbs invading crested wheatgrass fields on heavy soil. Many of
these mgood invadersn were the same for crested wheatgrass fields with light soil. However the
grasses Sporobolus cryptandrus and Bouteloua gracilis ranked higher in light soils as did the
forbs Psoralea argophylla, Astragalus missouriensis, and A. gracilis. Many of these same
species, especially Artemisia frigida, Koeleria macrantha, Stipa comata, Sphaeralcea coccinea,
Elymus smithii, Bouteloua gracilis and Poa secunda, are reported to occur in old crested
16
wheatgrass fields in southern Canada (Looman and Heinrichs 1973). Poa secunda is reported to
increase as crested wheatgrass increased (Heidinga and Wilson 2002).
Our results differ from previous research in one significant way. Studies conducted in
southern Saskatchewan (Looman and Heinrichs 1973, Heidinga and Wilson 2002, Bakker and
Wilson 2004) indicate that Elymus smithii and Bouteloua gracilis were common in old crested
wheatgrass fields. This is what would be expected based on competition/niche theory (Keddy
1989, Chasde and Leibold 2003) because crested wheatgrass is a caespitose, cool-season grass
while E. smithii is rhizomatous, and B. gracilis is a warm-season grass. Our results were
strikingly different; the cool-season bunchgrasses, Koeleria macrantha and Stipa comata were
much more abundant in the crested wheatgrass fields than either E. smithii or B. gracilis. Soils
in the Canadian studies were reported to be loam or clay loam (Wilson et al. 2004) which is
similar to soils in the majority of our sites. We have no explanation for this disparity in results
unless residual native vegetation of the study sites in Saskatchewan were strongly dominated by
E. smithii and B. gracilis and these two species provided the majority of native grass propagules.
Native species dispersal ability
Poor dispersal may also partly explain the paucity of native species in crested wheatgrass
fields. Macroplots in the center of crested wheatgrass fields had similar canopy cover of crested
wheatgrass but were, on average, three times farther from native vegetation than Edge
macroplots. The great majority of native species were less abundant in Center compared to Edge
macroplots, suggesting that dispersal is limiting reinvasion of many native species. Our results
suggest that this mlimited dispersal effectn is not strong because it did not result in a statistically
significant difference in overall native canopy cover between vegetation in the edge and centers
of the fields.
Our results also suggest that some native species are able to disperse better than others
(higher values in Table 5), and this effect has persisted even after many decades. Native species
showing little difference in mean microplot density between Edge and Center macroplots or with
greater density in Center macroplots across sites (mean density change >-0.10) are assumed to
have good dispersal abilities. Poa secunda, Carex stenophylla, Sporobolus cryptandrus and
Stipa viridula were grasses that showed good dispersal as did the shrub Gutierrezia sarothrae
(Table 5). The forbs Allium textile, Antennaria parvifolia, Astragalus gracilis, Bahia
17
oppositifolia, Gaura coccinea, Lygodesmia juncea, Psoralea argophylla, Sphaeralcea coccinea
and Vicia Americana appeared to be as common in Center macroplots as Edge macroplots.
There is no reason to believe that dispersal of these aforementioned species into crested
wheatgrass fields is limited.
Large differences between Center and Edge macroplot densities indicate poor dispersal
ability (mean density change -0.30). Such species include the shrubs, Artemisia frigida, A.
tridentata, and Opuntia polyacantha, the grasses, Elymus lanceolatus, Bouteloua gracilis,
Koeleria macrantha and Stipa comata, and the forbs, Achillea millefolium, Aster falcatus,
Comandra umbellate, Iva axillaris, Liatris punctata and Lomatium foeniculaceum. It is possible
that the abundance of native species in Edge macroplots may be limited by poor dispersal to
some extent even though Edge macroplots were adjacent to Native macroplots. Poor dispersal of
native species into crested wheatgrass fields has been reported for the Intermountain region of
Idaho (Marlette and Anderson 1986). There is accumulating evidence that range expansions of
many species is limited by dispersal (Primack and Miao 1992, Eriksson 1998, Marisco and
Hellmann 2009), and restorations will require active seeding (Bush et al. 2007).
Restoration recommendations
Crested wheatgrass was somewhat more common and more invasive on heavier and
especially siltier soils, but these differences were not great. Furthermore, crested wheatgrass
invasion of adjacent native vegetation was minimal, and there was little difference in invasion
between heavy and light soils. These results suggest that there is little reason to choose crested
wheatgrass fields for restoration based on soils.
The general paucity of native species in crested wheatgrass fields more than seven
decades after they were planted and the negative relationship between cover of native species
and crested wheatgrass indicates that cover of crested wheatgrass will have to be diminished, at
least temporarily, in order to establish significant number of native plants. It is unlikely that any
treatment outside of draconian measures will eliminate crested wheatgrass from these fields
(Ambrose and Wilson 2003, Hulet et al. 2009, Fansler and Mangold 2011). However, temporary
reduction of crested wheatgrass dominance through herbicide application may allow the early
establishment of natives (Bakker and Wilson 2004) that will allow conversion of fields from
virtual monocultures to more diverse grasslands that support increased native vertebrate and
18
invertebrate diversity. Crested wheatgrass cover can be diminished through plowing or herbicide
application. Plowing or disking is unlikely to be very successful because these methods distuirb
the soil, and older crested wheatgrass stands have a large seed bank and re-establish quickly from
seed (Anderson and Marlette 1986, Henderson and Naeth 2005). Herbicide application will not
eliminate crested wheatgrass (Ambrose and Wilson 2003). However, we recommend using
herbicide to weaken crested wheatgrass stands. This protocol will minimize disturbance and
crested wheatgrass emergence and provide a low-competition window for native plant
establishment.
Our study suggests that the majority of native species have a limited ability to disperse
even distances as little as 300 meters over a period of several decades. Native species will have
to be seeded into crested wheatgrass fields after they have been treated with herbicide. This will
be especially true for poor dispersers such as sagebrush, Artemisia frigida, Stipa comata, Aster
falcatus and Lomatium foeniculaceum (Table 5). Repeated seeding may be needed if weather
conditions are suboptimal to any great extent (Wilson et al. 2004).
Results of our study provide guidance on which species are able to successfully coexist
with crested wheatgrass. We recommend that species with high sum relativized EDR values
(Tables 3, 4) should be seeded into herbicide-treated fields in order to provide enhanced
biological diversity. Native grass species will be used by ungulates as well as livestock.
Increased abundance and diversity of forbs provide habitat for birds, including sage grouse
(Peterson 1970, Drut et al. 1994), as well as insects, including native pollinators (Dramstad and
Fry 1995). Although there was broad overlap among the native species present in crested
wheatgrass fields on different soils, there were some species, especially forbs, that occurred
preferentially on one of the two soil types. Our annotated species lists for light and heavy soils
are presented in Table 6.
Our recommendations omit some species that had high sum EDR values. Prickly
pear cactus (Opuntia polyacantha) was one of the most common woody plants in crested
wheatgrass fields, and it has large flowers that attract insects. However, planting cactus may not
be desirable in pastures that are being grazed by livestock. Narrow-leaved sedge (Carex
stenophylla) was one of the most common native species in crested wheatgrass fields. Although
it is palatable, it is small and provides little plant cover. Sandberg bluegrass (Poa secunda) is an
early, small bunchgrass that provides minimal forage and could be left out of seed mixtures.
19
However, it is competitive with introduced annual bromes, and several cultivars are
commercially available.
Forbs in both seed mixtures show a range of phenologies. Some species such as Vicia
americana and Astragalus missouriensis flower and produce fruit early in the growing season,
while others, such as Bahia oppositifolia and Aster falcatus are active throughout most of the
summer. Having a mixture of forbs that are active at different times helps provide adequate
habitat for birds such as sage grouse and generalist pollinators such as bumblebees.
Table 6. Species recommended for crested wheatgrass restoration seed mixes for heavy and light
soils (see Methods). Forbs are categorized according to when they flower: early (E), mid (M), or
late (L) phenology classes.
Heavy soil
Shrubs
Artemisia frigida
Artemisia tridentata
Gutierrezia sarothrae
Grasses
Koeleria macrantha
Poa secunda
Stipa comata
Stipa viridula
Elymus smithii
Forbs
Vicia americana (E-M)
Sphaeralcea coccinea (M)
Aster falcatus (L)
Gaura coccinea (M)
Iva axillaris (M)
Achillea millefolium (M)
Bahia oppositifolia (M-L)
Antennaria parvifolia (E)
Lomatium foeniculaceum (E)
Allium textile (M)
Potentilla pensylvanica (M)
Psoralea argophylla (M)
Light Soil
Shrubs
Gutierrezia sarothrae
Artemsisia tridentata
Artemisia frigida
Grasses
Koeleria macrantha
Poa secunda
Bouteloua gracilis
Stipa comata
Sporobolus cryptandrus
Elymus smithii
Forbs
Psoralea argophylla (M)
Sphaeralcea coccinea (M)
Achillea millefolium (M)
Astragalus missouriensis (E)
Bahia oppositifolia (M-L)
Vicia americana (E-M)
Astragalus gracilis (M)
Aster falcatus (L)
Allium textile (M)
Cerastium arvense (E)
Lygodesmia juncea (M)
20
Acknowledgements Adam Carr, Dan Brunkhorst, Mike Barrick and Dustin Crowe of BLM
identified potential study sites for us. Bill and Dana Milton allowed us to collect data on their
ranch. We are grateful to the many ranch families that allowed us to conduct our study on their
grazing allotments.
Literature Cited
Ambrose, L. G. and Wilson, S. D. 2003. Emergence of the introduced grass Agropyron cristatum and the native
grass Bouteloua gracilis in a mixed-grass prairie restoration. Restoration Ecology 11: 110l115.
Bakker, J. D. and Wilson, S. D. 2004. Using ecological restoration to constrain biological invasion. Journal of
Applied Ecology 41: 1058l1064.
Beetle, A. A. 1960. A study of sagebrush: the section Tridentatae of Artemisia. Wyoming Agricultural Experiment
Station Bulletin 368.
Box, T.W. 1986. Crested wheatgrass: Its values, problems and myths; where now? In:
K.L. Johnson (ed.), Crested wheatgrass: Its values, problems and myths. Symposium
proceedings. Utah State University, Logan. pp. 343-345.
Bush, R. T., T. R. Seastedt and D. Buckner. 2007. Plant community response to the decline of diffuse knapweed in
a Colorado grassland. Ecological Restoration 25: 169-174.
Chase, J. M. and M. A. Leibold. 2003. Ecological niches. University of Chicago Press, Chicago.
Christian, J. M., and S. D. Wilson. 1999. Long-term ecosystem impacts of an introduced grass in the northern Great
Plains. Ecology 80:2397l2407.
Cooper, S. V., P. Lesica and G. M. Kudray. 2007. Post-fire Recovery of Wyoming Big Sagebrush Shrub-steppe in
Central and Southeast Montana. Report to USDI Montana BLM, Montana Natural Heritage Program, Helena.
Davies, K. W., J. D. Bates and R. F. Miller. 2007. Environmental and vegetation relationships of the Artemisia
tridentata ssp. wyomingensis alliance. Journal of Arid Environments 70: 478-494.
Day, P. A. 1965. Particle fractionation and particle-size analysis. In, C. A. Black, editor-in-chief mMethods of soil
analysis, Part 1n American Society of Agronomy, Inc., Madison, WI.
Dewey, D. R. 1986. Taxonomy of the crested wheatgrasses. In: K.L. Johnson (ed.), Crested wheatgrass: Its values,
problems and myths. Symposium proceedings. Utah State University, Logan. pp. 31-44.
Dorn, R. D. 1984. Vascular Plants of Montana. Mountain West Publishing, Cheyenne, Wyoming.
Dramstad, W. and G. Fry. 1995. Foraging activity of bumblebees (Bombus) in relation to flower resources on
arable land. Agriculture, Ecosystems and Environment 53: 123-135.
Drut, M. S., W. H. Pyle and J. A. Crawford. 1994. Diets and food selection of sage grouse chicks in Oregon.
Journal of Range Management 47: 90-93.
Eriksson, A. 1998. Regional distribution of Thymus serpyllum: management history and dispersal limitation.
Ecography 21: 35-43.
Hansen, M. J. and Wilson, S. D. 2006. Is management of an invasive grass Agropyron cristatum contingent on
environmental variation? Journal of Applied Ecology 43: 269l280.
21
Heidinga, L. and S. D. Wilson. 2002. The impact of an invading alien grass (Agropyron cristatum) on species
turnover in native prairie. Diversity and Distributions 8: 249-258.
Henderson, D. C. and Naeth, M. A. 2005. Multi-scale impacts of crested wheatgrass invasion in mixed-grass prairie.
Biological Invasions 7: 639l650.
Holchek, J.L. 1981. Crested wheatgrass. Rangelands 3:237-250.
Johnson, K. L. 1986. The social values of crested wheatgrass: pros, cons and tradeoffs. In: K.L. Johnson (ed.),
Crested wheatgrass: Its values, problems and myths. Symposium proceedings. Utah State University, Logan. pp.
331-335.
Keddy, P. A. 1989. Competition. Chapman & Hall, London.
Knowles, R.P., and E. Buglass. 1980. Crested wheatgrass. Agriculture Canada
Publication 1295, Ottawa.
Maja Krzic, M., K. Broersma, D. J. Thompson and A. A. Bomke. 2000. Soil Properties and Species Diversity of
Grazed Crested Wheatgrass and Native Rangelands. Journal of Range Management 53: 353-358.
Lavin, M. and C. Seibert. 2011. Grasses of Montana. http://www.montana.edu/mlavin/herb/mtgrass.pdf.
Lesica, P. and T. H. DeLuca. 1996. Long-term harmful effects of crested wheatgrass on Great Plains grassland
ecosystems. Journal of Soil and Water Conservation 51: 408l411.
Lloyd, J. D. and T. E. Martin. 2005. Reproductive success of Chestnut-collared Longspurs in native and exotic
grasslands. Condor 107: 363-374.
Looman, J., and D.H. Hendrichs. 1973. Stability of crested wheatgrass pastures under
long-term pasture use. Canadian Journal of Plant Science 53: 501-506.
Lorenz, R.J. 1986. Introduction and early use of crested wheatgrass in the Northern Great
Plains. In: K.L. Johnson (ed.), Crested wheatgrass: Its values, problems and myths.
Symposium proceedings. Utah State University, Logan. pp. 9-19.
Majerus, M., Holzworth, L., Tilley, D., Ogle, D., Stannard, M. 2009. Plant Guide for Sandberg bluegrass (Poa
secunda J. Presl). USDA-Natural Resources Conservation Service, Idaho Plant Materials Center, Aberdeen, ID.
Marlette, G. M. and J. E. Anderson. 1986. Seed banks and propagule dispersal in crested wheatgrass stands.
Journal of Applied Ecology 23: 161-175.
McHenry, J.R., and L.C. Newell. 1947. Influence of some perennial grasses on the
organic matter content and structure of an eastern Nebraska fine-textured soil. Journal of the American Society of
Agronomy 39: 981-994.
Morris, M. S., R. G. Kelsey and D. Griggs. 1976. The geographic and ecological distribution of big sagebrush and
other woody Artemisias in Montana. Proceedings of the Montana Academy of Sciences 36: 56-79.
Peterson, J. G. 1970. Food habits and summer distribution of juvenile sage grouse in central Montana. Journal of
Wildlife Management 34: 147-154.
Primack, R. B. and Miao, S. L. 1992. Dispersal can limit local plant distribution. Conservation Biology 6: 513-519.
Reynolds, T. D. and C. H. Trost. 1980. The response of native vertebrate populations to crested wheatgrass
planting and grazing by sheep. Journal of Range Management 33: 122-125.
22
Rogler, G.A., and R.J. Lorenz. 1983. Crested wheat-grass-early history in the United States. Journal of Range
Management 36: 91-93.
Romo, J. T. 2005. Emergence and establishment of Agropyron desertorum Fisch. (crested wheatgrass) seedlings in
a sandhills prairie of central Saskatchewan. Natural Areas Journal 25: 26-35.
Schroeder, M. A. and W. M. Vander Haegen. 2006. Use of Conservation Reserve Program fields by greater sage
grouse and other shrubsteppe-associated wildlife in Washington state. Technical report prepared for USDA Farm
Service Agency. Washington Department of Fish and Wildlife, Olympia, WA.
Smoliak, S., A. Johnston, and L.E. Lutwick. 1967. Productivity and durability of crested
wheatgrass in southeastern Alberta. Canadian Journal of Plant Science 47: 539-548.
Urness, P. J. 1986. Value of crested wheatgrass for big game. Page 147-153 in K. L. Johnson (ed.), Crested
wheatgrass: its values, problems and myths. Utah State University, Logan, UT.
Vanderhorst, J., B. L. Heidel, and S. V. Cooper. 1998. Botanical and vegetation survey of Carter County. Montana.
Unpublished report to Bureau of Land Management. Montana Natural Heritage Program, Helena. 1 16 pp. + app.
Vaness, B. M. and S. D. Wilson. 2007. Impact and management of crested wheatgrass (Agropyron cristatum) in the
northern Great Plains. Canadian Journal of Plant Science 87: 1023l1028.
Wilson, S.D. 1989. The suppression of native prairie by alien species introduced for
revegetation. Landscape and Urban Planning 17:113-119.
Wilson, S. D. and Pärtel, M. 2003. Extirpation or coexistence? Management of a persistent introduced grass in a
prairie restoration. Restoration Ecology 11: 410l416.
Wilson, S. D., J. D. Bakker, J. M. Christian, X. Li, L. G. Ambrose and J. Waddington. 2004. Semiarid Old-Field
Restoration: Is Neighbor Control Needed? Ecological Applications 14: 476-484.