Predicting forest damage using relative abundance of multiple deer species and national forest inventory data

Abstract

Deer, both native and non-native, can damage forest ecosystems when occurring at high densities, impacting biodiversity and ecological functioning at multiple levels. The ecological drivers of forest damage and the roles of single and multiple co-occurring deer species are not well understood due to a scarcity of simultaneous high resolution multi-species deer distribution and forest damage data. Here, we aimed to disentangle the relationship between forest damage, forest characteristics and the roles deer play in damaging forest ecosystems. To achieve this, we integrated novel high resolution deer distribution data for multiple deer species (native and non-native) with forest inventory data collected in 1,681 sampling stations across Ireland to (i) understand environmental drivers of damage, and (ii) demonstrate the utility of this approach in providing predicted damage scenarios based on the relative abundance of different co-occurring deer species. We found that forest characteristics played a key role in the severity and type of damage risk that deer posed. All damage types were more prevalent in forests with greater tree densities where deer are more likely to find refuge from human disturbance. Bark stripping damage was more prevalent in mature forests with high tree diversity and ground level flora (e.g., bryophytes, herbs, and shrubs). Similarly, browsing damage was more prevalent in forests with greater tree richness but with understorey vegetation dominated by grass and ferns. Fraying damage was more common in mixed woodlands with understories dominated by bryophytes and grass. Crucially, we found that type and severity of forest damage were shaped by the interaction of multiple deer species occurring simultaneously, particularly at high densities, suggesting subtle inter-species competition and exclusion/partition dynamics that require further investigation to understand the ecological mechanism. Finally, we developed predicted damage scenarios to precisely predict where damage is likely to occur based on relative densities of two co-occurring deer species. We predicted high levels of damage in sika and/or red deer hotspots, matching areas of highly concentrated deer distributions. Our study highlights the ecological drivers of deer-caused forest damage, the role that co-occurring native and non-native deer species may play on forest damage within a large spatial scale and describes an approach that can be replicated in other countries with similar datasets. By combining reliable species distribution models with the national forest inventory data, we can now provide a useful tool for practitioners to help alleviate and mitigate forest damage and human wildlife conflicts.