A review of methods used to survey avian species and their potential for surveying ground-dwelling birds in Asia

Quantitative data on populations (i.e. density or size) of a given species in a speci!c area are very important for wildlife conservation in a range of di"erent contexts and for varying purposes. Density estimates, for example, provide baseline data on species abundance that can be used to derive population sizes and these can be monitored over time to assess the suitability of conservation strategies (Sutherland et al. 2004, Nijman & Menken 2005). Understanding these population changes is very important if policy and management plans are to be based on a sound understanding of species status. Reliable estimates of such population trends require careful design of the sampling strategy and !eld protocols that are standardised and repeatable with high detection probability and low observer variability (#ompson et al. 1998, Yoccoz et al. 2001, Pollock et al. 2002). #e demand for data to both assess the status of species of conservation concern and to inform the management of those with the highest probability of extinction is most acute in the tropics and subtropics where most threatened vertebrates are thought to occur (Laurance 2007), the knowledge of species ecology is severely limited, and habitat loss and illegal hunting present serious problems for many species (Wright & Muller-Landau 2006). #is is compounded by the di$culty in accessing terrain in areas that are mountainous or that have dense forest structure (e.g. tropical rainforest), making it a challenge to relate information about encounter rate (sighting or vocal detections) to populations (abundance, density or size). There is, therefore, a considerable need to encourage the collection and analysis of data that can be used to inform assessments of species status, wherever this is possible. Yet, for many who have an interest in avian !eldwork and who are keen to gather such data but lack detailed knowledge in survey techniques, determining which method to use and understanding how to use it properly is di$cult. #is is because of the complexity of di"erent contexts in which each method has been used, the purpose of the study and the various assumptions that have been made when developing the !eld protocols and in turning the !eld data into population estimates. Although there is a variety of texts available for surveying birds (e.g. Sutherland 1996, Bibby et al. 1998, 2000), a range of studies and developments have taken place since these were published that we seek to re%ect here. Furthermore, our aim is to review, brie%y, methods that have long been used or considered for birds in the tropics. We seek to present a summary of these methods in a form that will guide potential !rst-time users towards a method that will best suit their purpose and circumstances. Although we consider Asian birds in general, we place special emphasis on grounddwelling species such as pheasants, because they are o&en of high conservation concern (e.g. because of ground-snaring) and can be di$cult to detect. A range of methods have been used to collect data on terrestrial bird populations elsewhere, but their reliability and usefulness have not been reviewed critically and related to the data needs of species in Asia, especially where terrain and habitat are challenging. #ese methods include those that have been developed because of the way individuals in a species are detected (e.g. spot mapping, triangulation and now camera trapping) and those based on sampling theory (e.g. distance sampling using both point and line transect). All these methods have been used in order to generate population estimates of one sort or another, but there is limited guidance on what methods are suitable in which circumstances (e.g. the purpose of the survey, the ecology of the target species and the terrain and habitat being surveyed). Consequently, we seek to answer the question: what are the options and which should be considered for a particular purpose and context? #e aim of this synthesis is to provide a succinct review of the key features of four population estimation methods that have been used in South and South-East Asia, consider their utility for ground-dwelling birds in the region, and provide references to further information. We examine where each method has been used and for what purpose, the !eld protocol employed, the analytical methods required and their statistical assumptions. We also included gibbons in our review, because their vocalisations show striking parallels with songbirds (Clarke et al. 2006), and because they have also been the subject of methodological research in order to estimate population size from the detection of loud calls and sightings; there may therefore be lessons that can be learnt for avian population estimation in the region.