Genetic Monitoring for Managers


Type of GEM

Category 1a GeM Project Example: Estimating Vital Rates

Wolf survival and population trend using non-invasive capture–recapture techniques in the Western Alps

Marucco et al. 2009vitalrate phi wolves fig 1

Managers require reliable and precise estimates of population parameters such as abundance and survival rates, but these are hard to obtain for elusive, rare, and wide-ranging species (Boulanger et al. 2004, Stetz et al. 2010). Often, these are the very populations that require such information in order to make management decisions within a useful timeframe.

The Italian wolf (Canis lupus italicus) was recognized as a distinct subspecies in 1999. Despite being the national animal of Italy, wolves suffered from human persecution as have many wolf populations around the world. It was considered extirpated from the Alps in the early 20th century, and reached its lowest population levels in the 1970s. The population is believed to have been recovering since the 1980s, with current estimates of population growth of nearly 7%. Now a fully protected species in western Europe, the Italian wolf's range has expanded into Switzerland and southern France (Boitani 2003).


vitalrate phi wolves fig 2
Current estimated range of the Italian wolf.

Marucco et al. (2009) used open-model capture-mark-recapture (CMR) sampling with non-invasive individual identifications derived from fecal genotyping to estimate survival and growth rates for wolves in the Western Alps between 1999 and 2006.

In an effort to minimize bias and increase the precision of the estimates, they used a large-scale sampling strategy designed to reduce heterogeneity in individual detection rates.

Teams of 3–20 people on skis or snowshoes searched for wolf tracks a few days after a snowfall along transects that systematically monitored the entire study area. Tracks were followed as long as possible, with every scat along the track being collected. All transects were searched one to seven times during each of two winter sessions, plus fresh scats were collected when encountered during other field activities.


Further, their sampling design increases the probability of characterizing each individual based on differential marking behavior, especially young and dispersing wolves. With this design, they produced a dataset that met the assumptions of homogeneity in recapture rates, which is fundamental for CMR modeling.

They collected a total of 3382 scats from 3366 km of transects. From 1399 genotyped scats, 87 unique individuals were identified (39.1% F, 51.7% M, and 9.2% unknown sex).

Membership to one of seven documented packs was determined for each individual. Outside of packs, three solitary wolves with permanent territories and 18 dispersers were identified.

Using detection data from genotypes, the authors estimated apparent survival and recapture rates of wolves with open population, Cormack-Jolly-Seber (CJS) models. Their best models included age+season interactions for apparent survival rates, and effort (both field and lab) was significant for detection probabilities.

vitalrate phi wolves fig 4
The Western Alps study area showing the location of genotyped scats, 1999-2006.

vitalrate phi wolves fig3

Estimated apparent survival rates for young and adult wolves in the Western Alps, 1999–2006. Error bars represent 95% CI. Estimates were obtained by averaging the three best Cormack-Jolly-Seber (CJS) models.

Young wolves had lower apparent annual survival rates (0.24 ± 0.06) than adult wolves (0.82 ± 0.04). Survival rates were lower in the summer than in the winter for both young and adults.

The wolf population in the study area increased from 21 ± 9.6 wolves in 1999 to 47 ± 11.2 wolves in late winter 2005. This equates to a population growth rate of lambda=1.04 (± 0.27), which was lower than has recorded for other recolonizing wolf populations.

Regardless of the statistical method used (i.e., snow tracking index, CMR) they found a positive trend in wolf abundance. However, the abundance estimate based on snow-tracking was on average 36.2% (SD = 13.6%) lower than that from CMR modeling, likely because young dispersing wolves are likely to have lower sign detection rates in snow-tracks.

These are the first estimates for wolves in Italy and in the Alps and have important management implications.

The authors note that the successful application of these new sampling methods can be widely applied to broader spatial and temporal scales for other elusive and wide-ranging species in Europe and elsewhere. However, they suggest that a lower field effort than the present study could be adopted by avoiding multiple encounters of an individual within a sampling session.