This experiment evaluates the responses of plant species composition, net primary production (NPP), and ecosystem processes to the removal of dominant, foundation plant species. The experiment was initiated by Deb Peters (NMSU, JRN LTER) in 1995 to address questions about competitive interactions between foundation species and has since been maintained annually by the SEV LTER Field Crew. In addition to the original vision, we repurposed this experiment to address: How will changes in climate variance and mean affect ecosystem-specific biogeochemical processes and alter biophysical feedbacks?


Changes in climate mean and variance could alter the strength of biophysical feedbacks between plants and the abiotic environment in ways that accelerate or inhibit ecosystem transitions, dramatically altering predictions on future states. It remains unclear whether or when such changes will occur in drylands experiencing greater inter-annual climate variance. Plant species removals provide a powerful tool to assess feedbacks, by comparing process rates in the presence versus absence of a foundation plant species. These experiments provide a window on how biogeochemical and biophysical processes, as well as other community members, respond to the loss of foundation species. This experiment also enables the collection of new plant demographic data in the presence/absence of competitors to inform our predictions of the trajectories of ecosystem transitions, using our WAVE model (see Models Overview). Finally, plant removals have led to new tests of how biodiversity-function relationships change when foundation plant species are lost from ecosystems.


We selected five sites that were dominated by either blue grama (site 1), blue and black grama (site 2), black grama (site 3), black grama and creosote (site 4), or creosote (site 5). A sixth site was later added in the blue grama community along the foothills of the Los Pinos Mountains (site 6) near the Blue Grama Core Site (see Core Site Overview). At sites 1, 3, 5, and 6, five 3 m x 4 m plots had all plants of the dominant species removed initially and then allowed to recolonize; five 3 m x 4 m plots were controls. At site 2, 5 plots have blue grama removed annually, 5 plots have black grama removed annually, and 5 plots are controls. At site 4, 5 plots have black grama removed annually, 5 plots have creosote removed annually, and 5 plots are controls. Initial cover prior to removal was estimated by species for each plot. Grass was removed using a shovel to collect above-ground biomass and crowns just below the soil surface. Shrubs were removed using large clippers to collect above-ground biomass to the soil surface. All biomass removed was bagged, dried, and weighed. Plot maintenance or removal of the target dominant species is performed annually or as needed.


We estimate plant species cover for each live plant species within each 3 m x 4 m plot. Cover is estimated for all species as well as litter. Bare ground is calculated by subtracting the total cover of all species from 100%. Initial, pre-treatment measurements were made in 1996. Plant data are collected using standard SEV LTER protocols annually at peak fall biomass (Aug/Sept/Oct). Laboratory procedures: plant biomass removed from plots is dried, sorted by live and dead material, and weighed. Rain gauges were installed at each site and the corners of the areas containing each set of plots recorded in our ARCGIS database. Soil erosion bridges (1 m long) were installed in plots 1, 3 and 5 (removals and controls) at sites 1-5. We have recently measured soil biogeochemistry and texture.

Key Datasets:

Plant Removal Study: Recovery of Vegetation Following Disturbance at the Sevilleta National Wildlife Refuge, New Mexico. https://portal.edirepository.org/nis/mapbrowse?scope=knb-lter-sev&identifier=168

Featured Publications:

Peters, D. P. C., and J. Yao. 2012. Long-term experimental loss of foundation species: consequences for dynamics at ecotones across heterogeneous landscapes. Ecosphere 3:art27. https://doi.org/10.1890/ES11-00273.1

Peters, D. P. C. 2002. Plant species dominance at a grassland–shrubland ecotone: an individual-based gap dynamics model of herbaceous and woody species. Ecological Modelling 152:5–32. https://doi.org/10.1016/S0304-3800(01)00460-4

Peters, D. P. C. 2002. Recruitment potential of two perennial grasses with different growth forms at a semiarid-arid transition zone. American Journal of Botany 89:1616–1623. https://doi.org/10.3732/ajb.89.10.1616