Minimizing Risk of Cheatgrass Invasion and Dominance at the Idaho National Laboratory Site


Predicting plant community susceptibility to invasion by introduced species and determining mechanisms of resistance are fundamental concerns of ecology and ecosystem management. In the Great Basin, the invasive annual cheatgrass (Bromus tectorum) was introduced in the late 1800s and by the 1990s has grown to dominate more than 3 million acres, with another 14 million acres heavily infested and 60 million acres considered at risk for potential domination (Pellant and Hall 1994). However, the eastern portion of the Snake River Plain, including the INL Site, has largely escaped the cheatgrass dominance found in the western portions of the Snake River Plain and in northern and central Nevada.

There are several characteristics of the eastern Snake River Plain that might contribute to the relatively minor extent of cheatgrass invasion. The maintained cover of native species may make the vegetation of the INL Site resistant to invasion (Anderson and Inouye 2001). INL Site has a markedly different landscape disturbance history than more heavily cheatgrass invaded sites. Climate variables, such as colder winter temperatures and more late spring precipitation on the eastern Snake River Pains also differ from most cheatgrass dominated areas. The relatively minor extent of cheatgrass invasion at the INL Site in comparison with surrounding areas provides an exciting and unique opportunity to identify environmental conditions, community characteristics or management practices conferring ecosystem resistance to invasion.


The goal of this project is to use a combination of field surveys and mechanistic hypothesis driven greenhouse experiments to tease out the influences of environment, plant community, and land management on cheatgrass invasion success.

Comparative Surveys. We are conducting comparative surveys along a latitudinal climatic gradient from central Nevada, where cheatgrass dominated much of the landscape, to INL Site. We are establishing sampling plots at several hundred locations along this ‘mega-transect’ taking care to adequately sample sites with different types of disturbance legacies, management histories, vegetation composition, temperature and precipitation regimes. We will continue to sample intensively at the INL Site; at sites near INL Site which are climatically similar but with different land use and disturbance histories; and at sites in both northern and central Nevada with a range of disturbance, community composition and climatic variables. We are collecting information ranging in scale from microscopic (soil nutrients and microbes) to community (vegetation and animal) to landscape (climate and land use patterns) to parameterize a structural equation model (SEM) (Grace 2006) and specifically test hypotheses about how site characteristics affect invasion success of cheatgrass.

SEM is a powerful statistical way to infer causality: specifically, we are using it to determine why cheatgrass is more abundant in certain locations and less in others. An additional benefit of SEM is that we can include variables based on ‘expert opinion’ rather than relying on strictly empirical data. This means we can include a wealth of invaluable information that would not be otherwise useable in a more traditional quantitative model.

Controlled Greenhouse Studies. We are using controlled-environment experiments that involve individual species and constructed communities to establish a mechanistic understanding of competition between cheatgrass and native species. We are investigating competitive relationships, effects of diversity, density and disturbance and response to variation in water regime (timing and pulse size). Preliminary single-species trials indicate that cheatgrass and perennial species differ in their abilities to respond to water pulses depending on size and frequency of water events, and that moisture at the right time in the life cycles of cheatgrass could promote high competitive ability and possible invasion (K. Allcock, unpublished data). A mesocosm experiment is currently underway to test the interactions of precipitation timing and community composition in determining invasion success.

Accomplishments Through 2007

Comparative Surveys. The GIS data collected in 2006 was used to help identify potential sampling points. For our sites at INL Site, we selected areas with a diversity of vegetation type and fi re history. In June 2007, we visited INL Site and sampled our first 100 sites. We measured several plant community characteristics, signs of disturbance and physical environment variables. Soil samples were collected and analyzed for soil nutrients, texture, seed bank and soil food web dynamics. In October 2007, we returned to INL Site and inserted resin capsules into the soil. These capsules will collect soil nutrients over the winter. We will collect the resin capsules when we return to INL Site in spring 2008. It is our hope that these resin capsules will decrease the amount of lab work required to characterize soil nutrients as well as provide a time integrated measure of soil nutrient availability.

In November 2007, over 150 field sites in Nevada were identified, visited and resin capsules inserted. Our Nevada sites are in two areas, one is located outside Midas in northwestern Nevada and the other about 40 miles north of Austin in the central part of the state. These sites offer a huge variation in land use patterns, fire history, vegetation types and climate variables.
The data collected is being processed and used for model building and method refinement.

Controlled Greenhouse Studies. In late 2006 and early 2007, we established a series of two-species plant communities in 50-gallon barrels on the University of Nevada Reno. These communities were comprised of combinations of early-season native species (Poa secunda, Achnatherium hymenoides or Elymus elemoides), late-season native species (Pseudoroegneria spicata, Acnatherium thurberii or Hesperostipa comata), or one of each group. All plants were collected from the wild and transplanted to our constructed communities. One fourth of the barrels were not planted with any perennial species. All barrels were seeded with cheatgrass at a rate of 2000 seeds per m2. Each of these communities (early, late, mixed, or no perennials) was then subjected to either elevated total precipitation (150 percent normal precipitation for Reno, Nevada) or ambient total precipitation (equal to the amount of precipitation received through the growing season in Reno, Nevada). Finally, this ‘precipitation’ was either all distributed evenly through the course of the experiment (watered uniformly once per week) or 50 percent of the total precipitation amount was distributed evenly and the other 50 percent was applied in three randomly-timed ‘storm events’ in which barrels received 1/6 of the total allotted water volume for that treatment over the course of three days. We had six replicates of each community type, water amount, and water distribution combination, giving a total of 96 barrels.

Substantial mortality of transplanted perennials in the constructed communities in early 2007 meant that many plants had to be replaced at the beginning of the 2007 growing season (March-April 2007), so we delayed implementation of our experimental treatments until June 2007 in order to allow the replaced plants to establish. Watering treatments continued through November 2007, and final harvest occurred in December 2007. At the time of harvest we recorded density of cheatgrass, and clipped above-ground biomass, sorted by species. Samples are currently being oven-dried and weighed.


Comparative Survey. We only have data from 100 of the anticipated 500+ sites in the comparative survey and are still processing the samples and data. Thus, preliminary results are not yet available.

Controlled Greenhouse Studies. We are processing the above-ground biomass samples collected in December 2007. While the data are not yet ready to analyze, it appears that the ambient-amount, irregular-distribution watering regime caused some stress to both cheatgrass and perennial transplants, with fewer cheatgrass plants germinating and emerging, and several perennial transplants dying. The higher-precipitation treatments fared better. Emergence of cheatgrass in the high-precipitation, irregular-distribution treatment was initially low, but increased dramatically after the first ‘storm event’. There did not appear to be any obvious visual effect of the planted species on cheatgrass density or biomass. There was no effect of planted species on soil water content (as measured by time domain reflectometry, [TDR]) in the top 10 cm of soil, and minimal effect of the watering treatments on surface soil water content 24 hours after the water pulses were applied.

Plans for Continuation

This project will continue through 2010. We will continue collecting field data for the comparative survey at INL Site and our other field sites in 2008 and 2009. A select number of sites at INL Site will be followed year to year; however, most sites for the comparative survey will only be visited once. SEMs require a large number of data points in order for the algorithms used to identify reliable parameter values (Tanaka 1987), and we plan on sampling approximately 500 sites through the course of this study.

Publications, Theses, Reports, etc.

We anticipate several peer reviewed publication and conference proceedings on varied topics (such as, but not limited to: the effects of soil microbial community on cheatgrass success, the effects of soil surface morphology on cheatgrass germination and the effects of varied precipitation regime on cheatgrass competitive ability), in addition to the Ph.D. dissertation to be completed by Lora Perkins in 2009.

Investigators and Affiliations

Lora Perkins, PhD student, Department of Natural Resources and Environmental Science, University of Nevada Reno, Nevada

Robert S. Nowak, Professor, Department of Natural Resources and Environmental Science, University of Nevada, Reno, Nevada

Kimberly G. Allcock, Postdoctoral Associate, Department of Natural Resources and Environmental Science, University of Nevada, Reno, Nevada

Funding Sources

U.S. Department of Energy Idaho Operations Office

Nevada Arid Rangeland Initiative and the Nevada Agricultural Experiment Station


Anderson, J., and R. Inouye. 2001. Landscape scale changes in species abundance and biodiversity of a sagebrush steppe over 45 years. Ecological Monographs 71:531-556.

Grace, J.B. 2006. Structural Equation Modeling and Natural Systems. Cambridge University Press, NY.

Pellant, M. and C. Hall. 1994. Distribution of two exotic grasses on intermountain rangelands: status in 1992. p. 109-112 In: S.B. Monsen and S. G. Kitchen (compilers). Proceedings—ecology and management of annual rangelands. General Technical Report INT-GTR-313, Ogden, UT, USDA Forest Service, Intermountain Research Station.

Tanaka, J.S. 1987. How big is big enough? Sample size and goodness-of-fit in structural equation models with latent variables. Child Development 58: 134-146.


Home | Background | Surveillance |  Land Management | Education | Research | Risk Assessment | Publications |  Links | Feedback | Opportunities