The Effect of Native Lagomorph Herbivory and Predator-Prey Dynamics on Grass Establishment in a Shrub-Encroached, Semi-Arid Grassland

Abstract

Over the last two centuries, shrub encroachment into arid grasslands has been associated with reduced grass abundance, increased soil erosion, and local declines in biodiversity. Overgrazing by livestock and the associated reduction of fine fuels has been thought to be a primary driver of shrub encroachment in the southwestern United States, but shrublands have continued to persist despite livestock removal and grassland restoration efforts. The persistence of shrublands has been attributed, in part, to abiotic and biotic feedbacks that suppress grass reestablishment and reinforce shrub dominance. At a study site in southeastern Arizona, the removal of livestock grazing (cattle) more than 50 years ago and a follow-on shrub removal and grass seeding effort more than 30 years ago have not returned the area to its former grassland state. Because grass occurs within long-term herbivore exclosures, but remains generally absent anywhere else on the site, we hypothesized that grass establishment may be limited by herbivory from native herbivores. Additionally, we hypothesized that changes in the vegetative physiognomy of the site may have an impact on predator-prey dynamics, such as the dynamics between desert cottontails (Sylvilagus audubonii) and coyotes (Canis latrans). We established an herbivore exclusion experiment to test the effect of herbivory from native herbivores on grass establishment in an arid shrubland and used motion-activated trail cameras to document herbivore detection frequency. We created five treatment levels in our herbivore exclosures based on the relative body size of the herbivore permitted to enter, separating the effects of rodents, lagomorphs and mule deer. We also included two control treatments and replicated each treatment 10 times (n = 50). To measure the amount of biomass consumed in each exclosure, we introduced uniform divisions of fresh lawn sod (Cynodon dactylon) into each exclosure for 24-hour periods prior to and following the monsoon rains. Our motion-activated trail cameras operated for one year and recorded herbivore detections during and outside of sod trial periods. We used detections to examine how vegetative attributes (percent shrub cover, visibility, distance to nearest shrub) affected coyote and desert cottontail detections and how those attributes affect desert cottontail giving up density (GUD) through backward variable selection in linear models. We found that in the pre-monsoon trials, exclosures accessible to lagomorphs had less sod biomass relative to other treatments (p < 0.001), averaging only 44% ( 36%) and 29% ( 45%) after the 24-hour trial periods. We also observed desert cottontails visiting our exclosures significantly more frequently (p < 0.001) than any other herbivore. Visibility and percent shrub cover were positively correlated with desert cottontail detections in our species detection models and coyote detections were positively correlated with percent shrub cover and distance to nearest shrub. We also found that GUD was not correlated with any variable except for cottontail detection frequency, which suggests the threat of coyote predation on the site is minimal. Based on the results from our two studies, we concluded that desert cottontails are a significant consumer of grass in arid shrublands and may contribute to continued shrub dominance. We also concluded that ongoing shrub encroachment is likely to benefit desert cottontail populations, especially in areas with minimal predation risk.