An Improved Technique for Indexing Abundance of Himalayan Thar

The Himalayan thar (Hemitragus jemlahicus Smith) is an introduced ungulate pest of the alpine zone in New Zealand’s Southern Alps. Considerable money is spent attempting to mitigate their effects on indigenous flora (Tustin, 1990). Current management focuses on maintaining thar densities below specified densities by a combination of recreational, commercial, and government-funded hunting (Department of Conservation, 1993). Management thus requires extensive monitoring of thar densities. Thar are typical of alpine ungulates in that they are difficult to survey accurately (Hutchins and Geist, 1987). Their montane habitat, crepuscular activity (Tustin and Parkes, 1988) and helicoptershyness (Tustin, 1990), coupled with the vagaries of alpine weather, makes censusing of thar populations difficult. However, the Himalayan Thar Control Plan (Department of Conservation, 1993) requires indices of thar densities if “unacceptable damage to conservation values” is to be avoided. The current method of surveying thar defines the location and size of thar groups on the basis of ‘the largest count and most confident classification’ (hereafter termed the ‘largest-count’ method) after at least two independent periods of observation by different observers (Challies, 1992 unpublished report; see also Tustin and Challies, 1978; Baddeley, 1985). Density is then calculated as the sum of the groups divided by the search area. Because thar occupy habitats that are difficult to search (e.g., subalpine scrub and rock bluffs) and observers have different searching abilities, this technique cannot be considered a true census (i.e., “the total enumeration of the animals in the area”; Caughley and Sinclair, 1994). Rather, it provides a more-or-less repeatable index of thar abundance. Although the largest-count method allows rate of increase over time (r) to be estimated, it does not enable statistical comparisons of changes in population density over time to be made. Clearly, a more rigorous method of sampling population density incorporating some measure of sampling error at each sampling time would be advantageous. We describe the application of a double survey estimate of population size (Magnusson, Caughley and Grigg, 1978) to a South Westland thar population that was controlled in 1994. The objective of the study was to develop an index (Caughley and Sinclair, 1994) of thar density such that the effectiveness of the control operation could be evaluated.