How grazing and soil quality affect native and exotic plant diversity in rocky mountain grasslands

We used multiscale plots to sample vascular plant diversity and soil charac teristics in and adjacent to 26 long-term grazing exclosure sites in Colora do, Wyoming, Montana, and South Dakota, USA. The exclosures were 7-60 yr ol d (31.2 +/- 2.5 yr, mean +/- 1 SE). Plots were also randomly placed in the broader landscape in open rangeland in the same vegetation type at each sit e to assess spatial variation in grazed landscapes. Consistent sampling in the nine National Parks, Wildlife Refuges, and other management units yield ed data from 78 1000-m(2) plots and 780 1-m(2) subplots. We hypothesized th at native species richness would be lower in the exclosures than in grazed sites, due to competitive exclusion in the absence of grazing. We also hypo thesized that grazed sites would have higher native and exotic species rich ness compared to ungrazed areas, due to disturbance (i.e., the intermediate -disturbance hypothesis) and the conventional wisdom that grazing may accel erate weed invasion. Both hypotheses were soundly rejected. Although native species richness in l-mz subplots was significantly higher (P < 0.05) in g razed sites, we found nearly identical native or exotic species richness in 1000-m(2) plots in exclosures (31.5 +/- 2.5 native and 3.1 +/- 0.5 exotic species), adjacent grazed plots (32.6 +/- 2.8 native and 3.2 +/- 0.6 exotic species), and randomly selected grazed plots (31.6 +/- 2.9 native and 3.2 +/- 0.6 exotic species). We found no significant differences in species div ersity (Hill's diversity indices, N1 and N2), evenness (Hill's ratio of eve nness, E5), cover of various life-forms (grasses, forbs, and shrubs), soil texture, or soil percentage of N and C between grazed and ungrazed sites at the 1000-m(2) plot scale. The species lists of the long-ungrazed and adjac ent grazed plots overlapped just 57.9 +/- 2.8%. This difference in species composition is commonly attributed solely to the difference in grazing regi mes. However, the species lists between pairs of grazed plots (adjacent and distant 1000-m(2) plots) in the same vegetation type overlapped just 48.6 +/- 3.6%, and the ungrazed plots and distant grazed plots overlapped 49.4 /- 3.6%. Differences in vegetation and soils between grazed and ungrazed si tes were minimal in most cases, but soil characteristics and elevation were strongly correlated with native and exotic plant diversity in the study re gion. For the 78 1000-m(2) plots, 59.4% of the variance in total species ri chness was explained by percentage of silt (coefficient = 0.647, t = 5.107, P < 0.001), elevation (coefficient = 0.012, t = 5.084, P < 0.001), and tot al foliar cover (coefficient = 0.110, t = 2.104, P < 0.039). Only 12.8% of the variance in exotic species cover (log(10)cover) was explained by percen tage of clay (coefficient = -0.011, t = -2.878, P < 0.005), native species richness (coefficient = -0.011, t = -2.156, P < 0.034), and log(10)N (coeff icient = 2.827, t = 1.860, P < 0.067). Native species cover and exotic spec ies richness and frequency were also significantly positively correlated wi th percentage of soil N at the 1000-m(2) plot scale. Our research led to five broad generalizations about current levels of graz ing in these Rocky Mountain grasslands: (1) grazing probably has little eff ect on native species richness at landscape scales; (2) grazing probably ha s little effect on the accelerated spread of most exotic plant species at l andscape scales; (3) grazing affects local plant species and life-form comp osition and cover, but spatial variation is considerable; (4) soil characte ristics, climate, and disturbances may have a greater effect on plant speci es diversity than do current levels of grazing; and (5) few plant species s how consistent, directional responses to grazing or cessation of grazing.