Background/Question/Methods Cheatgrass (Bromus tectorum) may be the greatest threat to the sagebrush steppe ecosystems of the Intermountain West. At present, cheatgrass invasions are more severe in warm, dry, low elevation sites than in colder and wetter high elevation sites. However, the temperature increases predicted for the 21st century could increase the success of cheatgrass at higher elevations, impacting additional native communities and further hindering restoration projects. We conducted a warming experiment across an elevation gradient in northern Utah to test the role of temperature limitation, competition from neighboring vegetation and seed provenance in limiting cheatgrass invasion at high elevations. Two sites were located at 4800ft of elevation, one at 5700ft. At each site, two levels of a warming treatment (control and warming) were crossed with a neighbor removal treatment with two levels (control and removal of all vegetation). Each plot contained three subplots which were planted with seeds collected at each one of the three sites. Warming was achieved installing Open Top Chambers and neighboring vegetation was removed with glyphosate. Results/Conclusions Warming had a significant positive effect on cheatgrass population growth at all sites, but at the lowest sites there was a significant Warming x Removal interaction. Removal had a significant positive effect at all three sites while seed provenance had a significant effect at only one site. The positive effect of warming at all sites suggests that cheatgrass is temperature limited at both high and low elevations. However, one of the two significant Warming x Removal interactions resulted from neighboring vegetation offsetting the positive effects of warming, indicating that cheatgrass response to warming can be mediated by competition in certain conditions. Last, the similar performance of different seed provenances might occur because cheatgrass populations evolved in response to limitations other than temperature, or because genetic differentiation among our populations was not large enough to produce detectable differences.
Available at: http://works.bepress.com/peter_adler/80/