Legacy effects of an ecosystem stress experiment on a shortgrass-steppe consumer community

In a classic experiment of the US IBP Grassland Biome Studies, nitrogen and water were added singly and in combination over a 5-year period (1971-1975) to replicated 1-ha plots of shortgrass steppe in northern Colorado, with the aim of determining how removal of nutrient and water limitation altered primary production and plant community structure.  The resulting changes in plant biomass, litter and vegetation cover affected community patterns of nocturnal rodents, leading to invasion by two species, Microtus ochrogaster and Reithrodontomys megalotis, that were not usually found in upland habitats of shortgrass steppe.  We revisited these plots 20 and 30-36 years later to determine if these experimental effects persisted, and to assess the resilience of the resident nocturnal rodent community, which is dominated by Onychomys leucogaster and Peromyscus maniculatus.  We live-trapped rodents on each plot in September of each year, and used ordination analyses (MDS, ANOSIM procedures) on estimates of relative abundance to quantify and compare rodent diversity and community structure across treatments.  In 1995-96, rodent communities were grouped similarly to that observed during the original experiment (1970s), although treatment differences were not statistically significant.  Three new species (Chaetodipus hispidus, R. montanus, Mus musculus) were captured during this period.  By 2005-2011, rodent communities on experimental plots could not be distinguished from one another, although, collectively, manipulated plots tended to support greater total abundance and diversity than untreated control plots, suggesting some persistent treatment effects.  During this period, Dipodomys ordii became abundant on experimental plots, reflecting an increase in this species’ numbers regionally, a fifth new species (Perognathus flavus) was captured, and R. megalotis returned.  More than 30 years after the original experiment ended, these results reveal a highly dynamic rodent assemblage, one that stands in contrast to the relatively depauperate community that typifies upland shortgrass steppe.  Although it is tempting to envision complicated explanations involving community interactions for these long-term patterns, we attribute many of these differences to an overlooked aspect of the original experiment, the removal of cattle grazing by fencing in 1969.  Rodent communities on untreated control plots resembled grazed uplands in 1971 and 1972, but grazed uplands diverged from controls and treated plots over time, especially as vegetation inside the fenced pasture changed in response to climate variation.  The stark and consistent differences between the low-diversity, low-abundance communities of grazed uplands and those in ungrazed pastures underscores the effects of livestock grazing on plant community composition and vegetation structure, especially at our northern shortgrass steppe site, where taller cool-season grasses, forbs and shrubs increase in the absence of grazing.  The historical and cumulative consequences of the ongoing “press” experiment (grazing removal) apparently outweigh those of the relatively “pulsed” irrigation and fertilization treatments some 4 decades ago.