Community and population-level responses of an Afromontane chameleon assemblage to forest fragmentation

Background/Question/Methods

Habitat modification in the form of fragmentation and loss is a leading cause of biodiversity decline. The basic predictions from island biogeography theory that species richness and population size decrease with declining area and increased isolation have received considerable support. However much of this research has focused on birds and mammals in temperate regions or the Neotropics, limiting our ability to generalize to other taxa and regions. Reptiles in particular are understudied and have not yet shown the clear response of other taxa. Here we examine the community and population-level responses of an Afromontane chameleon assemblage to forest fragmentation. The East Usambara Mountains of Tanzania have high rates of endemicity in a forested landscape fragmented by logging, tea plantations, and small-scale farming. Remaining fragments have been isolated for approximately 80 years, making it unlikely that short-lived species retain any extinction debt. Within this fragmented habitat are eight species of chameleon, many of which are of conservation concern. We used repeated distance-based sampling in a large forest block and 11 forest fragments in order to estimate species’ densities and overall richness. This allowed us to quantify the population and community-level responses to habitat fragmentation while accounting for differences in detectability.

Results/Conclusions

Chameleon richness decreased with both decreasing fragment size and with increasing isolation. The chameleon communities of 10 of the 11 forest fragments were subsets of the community of the largest block sampled, suggesting that smaller fragments contribute little to landscape-level biodiversity. The densities of Trioceros deremensis, Rhampholeon temporalis, and the combination of Kinyongia matschiei and K. vosseleri are better predicted by fragment area alone than by any model including isolation, but species’ responses differ. R. temporalis shows a strong decrease in density with decreasing fragment area, while T. deremensis’ density decreases only slightly with decreasing fragment area. Neither species was found in the two smallest forest fragments (<3.5 ha), suggesting that each has a similar fragment area threshold. Kinyongia spp., which are commonly found in disturbed habitat, show an increase in density as fragment area decreases. Due to the alteration of chameleon communities in smaller fragments, with declines in species richness and species-dependent changes in density, several small blocks are not equivalent to a single large block. A possible functional cause for this non-equivalency is altered vegetation characteristics in smaller fragments. In some cases the vegetation variables are more strongly correlated with species’ density than is fragment area.