Bottom-up regulation of desert grassland and shrubland rodent communities: implications of species-specific reproductive potentials

We conducted an 11-year comparative study on temporal variation in rodent density, biomass, and species composition dynamics from adjacent grassland and shrubland environments in the Chihuahuan Desert of North America, in relation to rainfall and plant production. We found that rodent assemblages from those environments were only 14% similar in overall species composition, but consisted of different species in the same genera with similar ecological attributes. Each rodent community was numerically dominated by a different species of granivorous kangaroo rat, and the 2 rodent communities paralleled each other in body sizes and trophic structure. Rodent species compositions changed little over the 11-year period, despite considerable variation in rodent densities and biomass and in rainfall and plant production. Rodent abundance and biomass from both communities increased in relation to temporally variable rainfall and plant production, especially resulting from a series of El Niño and La Niña Southern Oscillation events. However, the grassland rodent community exhibited more rapid within-1-year lag-time responses to plant production, and prolonged high densities for 1 year before declining, whereas the shrubland rodent community exhibited primarily 1-year lag responses and immediate rapid decline in densities. Changes in rodent densities and biomass from both communities were significantly predicted by the production of annual grasses and forbs. Measured rodent reproductive activity was greater and happened sooner after rain and plant production events at the grass site than at the shrub site, and differences in the timing of rodent bottom-up responses between the grassland and shrubland habitats appeared to result from differences in the reproductive potentials of the 2 dominant rodent species. Dipodomys ordii, the dominant grassland rodent species, is known to produce more offspring than Dipodomys merriami, the dominant rodent in the shrubland community. We conclude that differences in the reproductive potentials of these 2 dominant rodent species likely accounted for the quicker and prolonged response of the grassland rodent community to bottom-up influences, rather than differences in the timing of plant production between the sites. Variation in reproductive potentials among rodent communities is likely a key factor affecting the timing of overall rodent community dynamics relative to changes in environmental resources.