Wolf habitat suitability in an area of Northern Apennines : a multimodel approach

The Italian wolf population was near to the extinction at the beginning of the seventies of the past century, owing to indiscriminate hunting and poisoning. In 1972 the wolf was named protected species and since this moment the recolonization of the Italian Peninsula has begun, starting from the remaining part of the range in Southern and Central Italy and reaching the Northern Apennines (N-Italy) during the eighties. This study was aimed to individuate suitable areas for wolves in the provinces of Pavia and Piacenza, that include a trait of the Apennines chain that links the more extended wolf areas in Emilia-Romagna and Tuscany to those of Liguria and Western Alps and where wolves suffer a high mortality due to illegal killings and accidents. From August 2011 to August 2012 we searched for wolf and wild ungulate signs of presence on 25 transect (total length 168 km) covered once a season and randomly located in an 860 km2 study area. We split the study area in 4-km2 cells by a grid spaced of 2 km and in each cell we measured the variables concerning altimetry, exposure, slope, and land use. We classified the cells crossed by transects as presence cells if wolf signs were found at least in a season and absence ones if no signs were found. We formulated habitat suitability models following an approach use vs. availability by Binary Logistic Regression Analysis (BLRA), Ecological Niche Factor Analysis (ENFA) and Maximum Entropy algorithm (MaxEnt) and by comparing presence cells with an equal number of randomly selected ones. Moreover we formulated a habitat suitability model following an approach presence vs. absence by BLRA, using only the cells crossed by transects and comparing presence cells with absence ones. For the first approach only the environmental variables were used, while for the second we used also the abundance of wild ungulates. We validated all the models by the K-Fold cross validation and ROC curves. The best model was that resulting from BLRA with presence vs. availability approach, and it was used to create a suitability map for the whole hill and mountainous area of the two provinces (2408 km2). Potential wolf range was 736 km2, i.e. 30.5% of available area. The suitability map can be used to the accomplishment of the actions to awaken public opinion towards the cohabitation with wolves and to plan the methods and interventions of damage prevention to the animal husbandry, so reducing the conflicts between human activities and wolf presence. Moreover on the basis of the habitat suitability map a network of protected areas can be designed to improve wolf protection and conservation policy