Assessment of methods for detecting an opportunistic and expanding mesocarnivore in southwestern Europe

Abstract

Knowing the distribution of expanding carnivore species is paramount for identifying and addressing potential human–wildlife conflicts. Occupancy models are useful tools to estimate the distribution and the probability of detection of wildlife species. In this study, we used these models with an empirical dataset to compare different survey methods and their combinations in order to optimize the estimated distribution in central Iberian Peninsula of the Egyptian mongoose (Herpestes ichneumon), the only Herpestidae species occurring in Europe. In particular, we aimed to identify the most cost‐effective (most accurate, with the lowest bias and cost) method or combination of methods. Sign surveys along transects on foot, hair‐traps (baited and unbaited) and camera‐traps were used as detection methods. We replicated these methods in 10 sampling units within four study zones in which the species was known to occur. We employed occupancy models to estimate the detection probability for each method in each zone, using covariates exclusively for detection probability, and made combinations of all methods. Camera‐trapping was the most precise and least biased single method, followed by transects on foot. In contrast, both baited and unbaited hair‐traps produced biased estimates of occupancy. However, camera‐traps was the most costly method, whereas single unbaited hair‐traps had the lowest cost. Our results demonstrate that the combination of several methods provides more precise and unbiased estimates of occupancy than those obtained from single methods. Even so, a biased method could contribute to improve the estimates if combined with other unbiased and precise methods. We recommend considering not only the precision and bias, but also the cost and effort required by each method to achieve the most cost‐effective results in distribution studies of carnivore species.

Update of the Egyptian mongoose (Herpestes ichneumon) distribution in Spain

“…We conducted field work in a few areas within CLM (18 municipalities) that were favourable for the mongoose according to favourability models (Díaz-Ruiz et al 2019) but where the species had not been previously recorded. In those municipalities, we employed three different methods to detect the mongoose: transects on foot searching for mongoose signs, camera-traps and hair-traps, which provide occurrence data at the point scale (Descalzo et al 2021a). The reliability of these methods was previously tested in locations with confirmed occurrence of the species (Descalzo et al 2021a).…”

“…In those municipalities, we employed three different methods to detect the mongoose: transects on foot searching for mongoose signs, camera-traps and hair-traps, which provide occurrence data at the point scale (Descalzo et al 2021a). The reliability of these methods was previously tested in locations with confirmed occurrence of the species (Descalzo et al 2021a). Transects on foot consisted on walking along edges of dense vegetation searching for mongoose scats and footprints (Descalzo et al 2021b).…”

“…Transects on foot consisted on walking along edges of dense vegetation searching for mongoose scats and footprints (Descalzo et al 2021b). The scats were inspected in the lab, being assigned to the mongoose only when its hairs were detected (Descalzo et al 2021a). Baited hair-traps consisted on two inverted U-shaped corrugated iron rods, deployed with the upper side at the height of mongoose shoulders with adhesive tape around each side and upper part of the arc (Fernández et al 2015, González-Broco et al 2017).…”

“…Transects on foot were walked only once. For further details about the field survey methods see Descalzo et al (2021a).…”

Monitoring methods influence native predator detectability and inferred occupancy responses to introduced carnivore management

“…Additionally, one monitoring method may have relatively high detectability, but may be biased if it only detects part of the population. Then integration with other methods can allow concurrent methods to better obtain unbiased estimates (Descalzo et al 2021). Using multiple methods is not without problems, including that different methods can achieve dissimilar detection of target species, or if one method is far better, it can produce inefficient monitoring (Nichols et al 2008;Mattfeldt et al 2009).…”