Deer exclusion unveils abiotic filtering in forest understory plant assemblages

International audienceBACKGROUND AND AIMS: The role of deer (family Cervidae) in ecosystem functioning has traditionally been neglected by forest ecologists due to the animal’s scarcity in most parts of the northern hemisphere. However, the dramatic rebound in deer populations throughout the 20 th century has brought deer browsing to the forefront of forest ecological questioning. Today there is ample evidence that deer affect tree regeneration, understory plant and animal diversity and even litter decomposition. However, the mechanisms underlying the effects of deer on forest ecosystems remain unclear. Among others, the relative role of abiotic factors versus biotic interactions (e.g. herbivory) in shaping plant assemblages remains largely unknown. METHODS: We used a large-scale experiment with exclosures distributed along abiotic gradients to understand the role of black-tailed deer (Odocoileus hemionus sitchensis) on forest understory on the Haida Gwaii archipelago (western Canada), a unique context where most of the key ecological effects of deer presence had already been intensively studied. KEY RESULTS: Our results demonstrate that 20 years of deer exclusion resulted in a clear increase in vascular plant richness, diversity and cover, and caused a decline in bryophyte cover. Exclusion also unveiled abiotic (i.e. soil water availability and fertility) filtering of plant assemblages that would otherwise have been masked by the impact of abundant deer populations. However, deer exclusion did not lead to an increase in beta diversity, probably because some remnant species had a competitive edge to regrow after decades of over-browsing. CONCLUSIONS: We demonstrated that long-term herbivory by deer can be a dominant factor structuring understory plant communities that overwhelms abiotic factors. However, while exclosures prove useful to assess overall effects of large herbivores, the results from our studies at broader scales on the archipelago, suggest that exclosure experiments should be used cautiously when inferring the mechanisms at work