Top-down and bottom-up consequences of unchecked ungulate browsing on plant and animal diversity in temperate forests: lessons from a deer introduction

Debate on the relative importance of competition for resources and trophic interactions in shaping the biological diversity of living communities remains unsettled after almost a century. Recently, dramatic increases in ungulate populations have provided a useful quasi-experiment on the effects of unrestrained ungulates on forest ecology. The islands of Haida Gwaii (Canada) offer a unique situation to investigate the potential of large herbivores to control temperate forest community structure and diversity. Black-tailed deer Odocoileus hemionus Merriam, native to adjacent mainland areas of British Columbia, were introduced in 1878 and spread to all but a few islands. Because deer were not native to the archipelago, islands that still lack deer provide a rare instance of temperate forest vegetation and fauna that developed in the absence of large herbivores. The colonisation of different islands at different times, and the absence of significant predation allow us to assess whether and how a large herbivore can exert "top-down" control on vegetation and its associated fauna. We studied plant communities in forest interior and shoreline, on seven small islands of varying browse history. Three islands were untouched by deer, deer had been resident for about 15 years on two, and on another two deer had been present for more than 50 years. Without deer, vegetation in the understorey and/or shrub layer was dense or very dense. Structure and composition varied markedly within and between shoreline and interior communities. Without deer, shoreline communities were dominated by species absent from islands with deer. Where deer had been present for less than 20 years most plant species characteristic of shorelines on islands without deer were already absent or scarce, but in the forest interior species richness was less affected and extensive shrub thickets remained. On islands where deer had been present for > 50 years vegetation below the browse line was extremely simplified, converging in both forest interior and shoreline towards an open assemblage of a few deer-tolerant species, basically two coniferous trees. This top down effect on the plant community reflected up the food chain so that understorey invertebrate and shrub-dependent songbird communities became simplified. In contrast, species densities of litter arthropods (especially weevils and millipedes) were highest where deer were present for > 50 years. Canopy birds were unaffected by deer presence. In the absence of predators, major climatic stress or other means to control the herbivore, deer browsing created greatly simplified plant and animal communities.