97th ESA Annual Meeting (August 5 -- 10, 2012)

COS 170-2 - Elevated levels of herbivory on an urban-rural gradient reveal species-specific signals of landscape degradation in novel ecosystems

Thursday, August 9, 2012: 1:50 PM
D135, Oregon Convention Center
Dieter F. Hochuli, School of Life and Environmental Sciences, The University of Sydney, Sydney, Australia
Dieter F. Hochuli, The University of Sydney

Background/Question/Methods

Loss of biodiversity is a major factor contributing to declines in the quality of remnant vegetation in urban landscapes.  We have previously shown that these novel ecosystems often support depauperate assemblages, with losses from higher trophic levels contributing to the decline of some ecosystem processes.  The aim of this work was to identify the extent to which levels of herbivory in urban remnants could be predicted from landscape and plant traits, ultimately identifying benchmarks for ecological function. I assessed levels of herbivory along an urban-rural gradient in Sydney, Australia, for four prominent plants, Banskia serrata, Angophora costata, Acacia longifolia and Pittosporum undulatum.  These plants represent different functional groups within remnant vegetation;  respectively a sub-canopy tree, dominant canopy tree, fast growing shrub and an invasive native sub-canopy tree.  I assessed 3 types of insect damage (chewing, mining and galling) from almost 5000 leaves obtained from 235 trees. I used generalised linear models to test for coarse and fine scale associations of landscape traits and plant traits (specific leaf area, water content and tree size) with herbivory.  

Results/Conclusions

The damage caused by different functional groups of insect herbivores varied significantly across plant species, with A. costata and B. serrata suffered the highest levels of chewing damage and  P. undulatum showing highest levels of galling and mining.  Damage on A. longifolia was highly variable among sites, suggesting that site level differences were central to predicting patterns across these landscapes.   There were no association between herbivory and specific leaf area and water content, although leaves from larger trees had higher levels of total leaf damage.  There was increased herbivory on A. costata and B. serrata in non-urban areas, while P. undulatum suffered reduced herbivory in urban areas.  There was no coherent signal of herbivory for all 4 plant species across the urban-rural gradient owing to the extensive variability among sites.  The results revealed idiosyncratic associations with degree of urbanization across plant species, with fine scale habitat traits consistently being more informative in predicting levels of herbivory than coarse habitat traits.  While these results showed that urban remnants supported vegetation suffering significant levels of herbivory, the goal of identifying benchmarks of ecological function in urban landscapes required species-specific approaches accommodating coarse and fine scale landscape traits.