Ecology and conservation of large-bodied avian frugivores of Luzon, Philippines

Española, Carmela P. (2013) Ecology and conservation of large-bodied avian frugivores of Luzon, Philippines. Doctoral thesis (PhD), Manchester Metropolitan University.

Preview

Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (3MB) | Preview

Abstract

Avian frugivores across Southeast Asia, and in the Philippines in particular, are seriously threatened owing to massive loss of habitat and direct exploitation through hunting and the pet trade. Their declines may have dire consequences for wider ecological processes as many frugivores are also seed dispersers. Conservation programmes in the Philippines are crippled by a lack of knowledge on the status, abundance and ecology of frugivores which extend to other endemic species in the country. This PhD identified factors that influenced frugivore community composition as well as drivers of frugivore distribution across Luzon, the largest island in the Philippines. It also developed cost-effective methods for gathering baseline ecological data to inform conservation measures for frugivores and other little-known species over large geographical areas. This included a way of correcting for the bias caused by non-random transect placement in a study site which is often the case during bird surveys in the tropics. Twenty-five species of pigeons, parrots and hornbills were surveyed using distance sampling along nearly 500 km of line transects at 14 sites across the island of Luzon. I documented surprisingly few reliable disappearances of frugivores from individual forest patches – in fact this and other fieldwork since 2000 has increased the known extent of occurrence of several species. However, green racquet-tail Prioniturus luconensis may have suffered a real range contraction. More alarming was the absence of large parrots from most sites with apparently intact habitat surveyed. Even where present, large parrots exhibited lower densities than related species in similar habitat in Southeast Asia. For six species, including four of six parrots, the largest estimates of population in any reserve in Luzon numbered < 1000 individuals, and nearly one-third of all iii populations in reserves were < 100. At minimum viable population (MVPs) of 500, frugivore communities in all but 2–3 of the largest reserves are not expected to remain intact. Although seed dispersers may fare better than seed predators (large parrots), a major collapse of frugivore communities may occur across Luzon, with serious implications for ecosystem functioning. The Philippines comprise islands of different origins, climate and habitat, a situation which is expected to produce a biogeographically complex set of animal and plant communities, which themselves are influenced by anthropogenic actions. I explored similarities between frugivore communities across 24 sites in Luzon using non-metric multidimensional scaling and attempted to explain site differences in terms of a series of geographical, habitat, and disturbance predictors using Mantel tests. In both analyses using species presence/absence and densities, sites/species did not seem to ordinate simply according to region. Consistent outliers included three sites in West Luzon and two in Central Luzon, and, in terms of species, several large rare parrots and pigeons. The strongest correlates of site dissimilarity were differences in altitude and several human disturbance measures, including path width, canopy closure and a ‘human impact index’ (reflecting human pressures and forest management). While Luzon’s frugivore communities have been no doubt shaped by natural biogeographical processes, their effects have been largely obscured by anthropogenic environmental degradation. There is an urgent need to understand better the drivers of frugivore species distribution in order to develop appropriate conservation management strategies. To identify precise habitat associations of 18 avian frugivores, the presence/absence of each species along 400 m long segments of 213 transects was examined in relation to vegetation structure and composition, measured at 1227 plots, using generalised linear mixed models (the 24 sites were entered as a random factor). Individual frugivore species showed unique patterns of iv association with habitat variables but five species were high-altitude specialists while six preferred lowland sites. Another six species strongly preferred primary forest while one thrives in disturbed forest with the attendant increase in food availability. I then ran generalised additive mixed models (GAMMs) to identify any non-linearities in responses of species to habitat features. Relationships with habitat variables were on the whole simple linear or quadratic for the majority of species, suggesting that there were gains to be had in improving habitat right along the disturbance gradient. Precise and accurate estimates of wildlife population density and sizes are essential to evidence effective conservation programmes. Line transect distance sampling is a robust method in that variability in detectability due to distance from the transect line, but many conservation studies cannot, by necessity, be based on random transect placement, but instead use transects along existing trails. This study estimates the bias in abundance estimates due to non-random placement of transects along hunter trails (path width <100cm) and access roads (path width >100cm) as compared with random paths (especially cut transects). Path types were similar in altitude, but differed in terms of tree girths, slope, canopy covers, and presence of crops. Hunter trails yielded lowest densities and encounter rates for nine of 12 species and lowest effective strip width for seven of 12 species. Highest densities and encounter rates were along random paths for seven of 12 species. Differences in density across trail types were driven by differences in encounter rates rather than differences in detectability. Density estimates calculated from surveys which used non-random transects should be upwardly corrected by on average 90% (18-187%). In fragmented forests where random placement of transects is not always possible, this method of correction will allow species density estimates from sampling along hunter trails and access roads to be adjusted. v Top on the list of research and conservation priorities arising from this PhD would be to map the remaining populations of the Luzon-endemic Green Racquet-tail, Luzon Racquet-tail and Flame-breasted Fruit-dove and to formulate conservation intervention measures for these threatened/near-threatened species taking into account habitat preferences and threats to the species. Forest and reserve management programmes and policies in the Philippines and elsewhere in the tropics would greatly benefit from empirical data on species occurrence and accurate estimates of population abundance using methods described in the study. Sound ecological research by local biologists/ecologists must be encouraged to further our understanding of species requirements, species tolerance to disturbance, and viability of populations, especially of the many unique and/or threatened species in the Philippines and the wider SE Asia region.