Comparing the cost-effectiveness of methods for estimating population density for primates in the Amazon rainforest Peru

With increasingly extreme fluctuations in flood levels in the Amazon basin (Malhi et al. 2008, Marengo et al. 2012, Bodmer et al. 2014) the future of its' fauna is becoming more uncertain. It is essential therefore that effective monitoring is in place in order to detect drops in population before irreversible damage is done. In developing countries such as the ones situated in the Amazon basin funding for conservation is very limited (Danielsen et al. 2003), it is therefore vital that cost effective methods of monitoring the wildlife of the Amazon are found. Three surveying techniques for monitoring primates are compared in this thesis to find the most cost effective method of estimating population densities of primate species local to the Amazon basin; these are terrestrial transects, aquatic transects and audio-playback point counts. Data was collected in the Pacaya-Samiria National Reserve using these methods over a period of four months, from January to May 2014. For both terrestrial and aquatic transects, transect lines were traversed and data was recorded every time an individual or group of the 7 primates species were spotted. Audio-playback point counts were used to record data for red howler monkeys (Alouatta seniculus)and brown capuchin monkeys (Cebus apella). This was done by mimicking primate vocalisations at a point and recording any resultant responses or sightings of the species under observation. Each survey technique was compared with regards to three qualities; precision, ability to react to change and cost. On average over all 7 species of primate aquatic transects produced the most precise estimations of population density with a mean estimation CV% (percentage coefficient of variance) of 36.35% in comparison the 47.3% averaged by terrestrial transects. Both methods failed to produce precise results for the two rarest species present, the monk saki monkey (Pithecia monachus) and the white fronted capuchin monkey (Cebus albifrons). Aquatic transects were also shown to react to sudden change in population levels. For the two species Alouatta seniculus and Cebus apella aquatic transects once again on average gained the most precise results with a mean estimation CV% value of 20.05% in comparison to the 31.08% of terrestrial transects and 36.35% for audio-playback point counts. The estimates created using audio-playback point counts used considerably less time and resources than the other two methods for single species and was also shown to be the quickest to react to immediate changes in population densities. Thus it was concluded that audio-playback point counts can produce relatively precise estimates that react to population changes at low cost. However only one species can be observed at a time using audio-playback point counts; when observing multiple species at one time it was concluded that aquatic transects are by far the cheapest survey technique and the method that produces precise estimates more consistently. I would therefore recommend for a monitoring effort of several primate species at one given time in the Amazon basin, that aquatic transects be used as it is the most cost-effective overall. However if a single species is a monitoring target, perhaps to be used as an indicator species or because the primate is of most concern, then audio-playback point counts be used as it is possible to gain relatively precise results at a minimal cost. I would also like to suggest that the use of audio-playback point counts be tested on rarer primate species in future as neither terrestrial transects nor aquatic transects could produce a useful estimate in a combined effort of 104 half days. If audio-playback point counts could be used to get good estimates for rare primate species then monitoring strategies could be developed combining the use of audio- playback point counts and aquatic transects to gain precise density estimates for all primate species in an area whilst keeping costs low. A generic decision tree is presented at the end of this thesis as a guideline to cost-effective primate monitoring for any seasonally flooding rainforest study site.