Our understanding of ecological function in desert systems suggests that the physiological and morphological differences among species act as strong influences on resource capture and ecosystem modification; hence alteration of the composition or the relative abundances of species should alter availability and cycling of energy, water, and other resources.
We initiated a long-term experiment in which the diversity of plant "functional types" and of plant species has been manipulated by species removals. We predict that after plant biomass has recovered from the initial disturbance, ecosystem function will vary according to the diversity of plants and plant types within the plots. Methodology. Given the long lifespans and difficulty of establishment of desert perennials, establishment of desired communities from seed or transplant was impractical. Instead, we established our stands of varying plant diversity by making appropriate removals of species or species groups from replicate plots of a diverse Chihuahuan scrub site. Treatments were established in summer/fall 1995. Plants were removed by simple cutting at the soil surface (no soil disturbance or herbicide). For most species, this has been effective; recruits of perennial grasses are removed from the appropriate plots twice per year, and regrowing shrubs are clipped back as needed. After 2 years only one perennial grass and two shrubs require substantial effort for plot maintenance. The experimental area is a 250 m x 250 m area located on the NMSU Chihuahuan Desert Rangeland Research Center, immediately northeast of the intersection of the Summerford powerline road and the road running along the southern boundary of the Jornada Experimental Range and the CDRRC. The area was gridded into 25 m x 25 m plots; because of some existing environmental gradients (the area slopes to the east, and mesquite abundance varies from north to south) blocks were established and treatments were randomly assigned to plots within blocks. Plot boundaries are marked with rebar in the corners and simple steel wire stretched along boundaries; plots are identified by number and treatment on aluminum tags and on plastic rebar caps in the southwest corner. Design and hypotheses. The treatments are designed to distinguish the effects of plant biomass per se from those of plant functional groups and plant species richness within functional groups. We recorded the mass of plants removed from the plots when the treatments were imposed in 1995; the total biomass removed from a plot is an index of the disturbance associated with the start of the experiment. One hypothesis is that plant biomass, rather than the species or functional group composition of the plants, is the critical regulator of ecosystem function. If so, most response variables would be highly correlated with initial biomass removed in the early seasons of the experiment. If remaining species increase in biomass over time, to the point where all treatments support approximately equal biomass, that correlation (and the differences among treatments) should disappear. Another hypothesis is that the architecture and physiology of different groups of plants differ sufficiently that each functional group contributes in a unique way to ecosystem function. We therefore are testing the impact of removing various groups or growth forms from the plant community. Four treatments involve the removal of all individuals of all species of a given functional group of perennial plants: shrubs, subshrubs, perennial grasses, and succulents (both leaf and stem succulents). That is, one treatment involves the removal of all shrubs from a plot; another the removal of all perennial grasses, and so on. All other treatments (Control plus three other treatments) contain at least some species representative of all functional groups, so the contrast between the functional group removals and these other treatments should reflect the effect of functional group diversity. There are multiple species in each of the functional groups, so an additional hypothesis is that higher species richness within a functional group alters ecosystem function significantly. We have imposed a Simplified treatment, where only a single species (the most abundant species at the time of initiation) of each of the four groups remains, and all other species of those functional groups have been removed. That is, the Simplified plots contain a single shrub (Larrea tridentata), a single subshrub (Zinnia acerosa), a single succulent (Yucca baccata), and a single perennial grass (Muhlenbergia porteri). In contrast, we have also imposed two versions of a Reduced diversity treatment, where the dominant species have been removed and the subordinate species remain. (In one version of this treatment, Larrea is assumed to be the dominant shrub and was removed; in another version, Prosopis is assumed to be dominant and is removed. In both of these, the dominants of other growth forms are as identified above in the Simplified treatment.) Finally, in the Control, all species (dominant and subordinate) of all four functional groups remain. Hence the contrast among the Control, the Reduced, and the Simplified treatments should reflect the importance of species richness within functional groups. Response variables. Perennial plants serve as resources for many other organisms, and modify the abiotic environment as well; hence we are monitoring both biological and physical variables. Initial studies are characterizing the amount of sediment moving on the surface of the plots (indicator of erosion vulnerability),composition and abundance of grasshoppers and ants, herbaceous vegetation, and growth and demography of a shrub (Prosopis glandulosa) and a succulent (Opuntia violacea).