Line transect sampling in mountainous terrains

• Autor(a) principal: Afonso, Anabela
• Data de Publicação: 2011
• Outros Autores: Alpizar-Jara, Russell
• Tipo de documento: Artigo de conferência
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10174/3927
• topic: Line transect

Line transect sampling is an attractive method to estimate abundance in certain mountainous terrains, since it allows to cover a large region in a short period of time and there is no need to mark or capture animals. Abundance estimates are obtained with the number of detected animals adjusted by an estimate of the probability of detection. This probability is usually unknown and the estimation is based on measurements of perpendicular distances from a line to detected animals. In mountainous terrains, various measurements can be taken: real sighting (inclined), topographic and perpendicular distances. The elevation angle between the observer and the animal locations should also be measured. It is often recommended that all measured distances should be projected onto a horizontal plane in order to avoid potential bias in the conventional estimators of distance sampling. However, our results show that this procedure can lead to biased estimates, even with the inclusion of altitude as a covariate in the detection model. In uneven terrains, the range of sighting projected distances varies along the transect line. Thus, this bias can be due to violation of the uniformity assumption of projected perpendicular distances for the conventional estimator of detection probability. We propose corrections for this estimator based on the joint distribution of several independent uniform random variables. The performance of the corrected estimators is compared with the conventional estimator through a simulation study. In a highly varying slope surface, conventional estimator presents a poor performance, and a less biased and more precise estimates were achieved with the proposed estimator.