The Impact of Oil Palm Conversion on Tropical Amphibians.

Abstract

Agriculturally-altered habitats, especially oil palm plantations, are rapidly dominating the Southeast Asian landscape. Although recent studies have shown reduced species diversity associated with this commodity, data on amphibian diversity are rare. The following thesis explores the impact of oil palm plantations on amphibians in Peninsular Malaysia based on (1) amphibian biodiversity, (2) quality and use of breeding sites, (3) habitat use and (4) parasitism. Contrary to expectation, not all metrics of biodiversity differed between oil palm plantations and secondary forest sites. Amphibian community composition, however, differed greatly between the two habitat types, with oil palm communities being dominated by species known to prosper in disturbed habitats, indicating that the community is currently of limited conservation value. Within plantations, temporary pools were found to serve as important breeding habitats for amphibians so a focused study on the characteristics of these pools was carried out. Although we found differences between pools, the proportion of occupied pools did not differ significantly between plantation and forest sites. I did observe evidence of breeding site preferences of least concern, plantation amphibians, along with habitat partitioning between species, a similar pattern also seen in forest communities. I compared parasite burdens between habitats by screening for the fungus Batrachochytridium dendrobatidis (Bd) and for nematode parasite load. Bd was not detected in any of my samples and there was no difference in nematode loads between habitat types. However, patterns of nematode prevalence was affected by host type, while nematode intensity was dependent on an interaction between host and body size. The final chapter indicates that in terms of parasite, the host environment is the most important. Additionally, differences in host-parasite patterns between habitats indicate a possible underlying problem that rapid biodiversity censuses would be unable to detect.