A fixed-wing sightability model for oryx in desert habitats

Abstract

We investigated factors affecting sightability of oryx (Oryx gazella gazella) during fixed-wing aerial surveys in desert grass-shrubland habitat types to develop a sightability-adjusted population estimator to aid in precise population management. Sightability of oryx was affected by group size, activity, and vegetation type. We used logistic regression to model all possible combinations of the three significant variables, and compared models with a variety of fit and information-theoretic statistics as well as by relative performance. Because no model was superior to the others, we used relative performance among all models and parsimony to select the preferred model. Our preferred model included variables for social group size and three levels of group activity (bedded, standing, moving). This model estimated oryx population size as 3917 (3534–4297) and 3312 (2999–3593) for two annual surveys, and showed a deviance of –0.007 and –0.03 from the grand mean of all models for these surveys, respectively. Full confidence interval widths from complete surveys (100% of area covered) were 19.0% and 17.9% of mean population estimates, allowing for precise estimation and consequently management of the oryx population, although confidence interval widths will vary with group size and behaviour. Oryx surveys should be conducted during periods when group sizes are largest locally, which was the summer in New Mexico, to maximize sightability of oryx and thus minimize variation in population estimates due to sighting error.