Habitat use by European wildcats (Felis silvestris) in central Spain: what is the relative importance of forest variables?

1. What they did

The European wildcat has long been described as a forest species, and conservation advice in Europe and Spain has focused on managing forests for them. Some studies, though, have found wildcats preferring scrubland and pasture. The author asked how important forests really are for wildcats in central Spain, and for the first time in wildcat habitat studies included agricultural steppe (open cereal farmland).

• Transects. Between autumn 2001 and spring 2002, the author surveyed 101 transects, each 1 km long, along paths and tracks 1 to 3 m wide. They covered five habitat types: Mediterranean vegetation in the plains (16 km), Mediterranean vegetation in the mountains (16 km), deciduous oak forests (21 km), mountain pine forests (25 km) and agricultural steppes (23 km). Transects were at least 1 km apart. Most followed the same routes as an earlier study in the same region (Lozano et al. 2003); the steppe transects were new.
• Wildcat abundance from scats. Each transect was split into five 200 m segments. The wildcat abundance index was the number of segments with a wildcat scat, divided by five.
• Rabbit abundance. In each 200 m segment, the author walked a short side transect (50 m long and 1 m wide) at right angles to the trail and counted rabbit latrines.
• Habitat structure. Every 200 m, the author estimated by eye tree cover, shrub cover, open ground, rock cover, tree height and a shelter score (1 to 5, based on cavities in rocks and trees and how dense and hidden the site was) within 25 m. Land-use maps were used to measure forest, pasture, cropland, scrub and urban cover, the number of watercourses, roughness and elevation in a 9 km² area around each transect, an area that covers most reported wildcat home ranges.
• Analysis. Abundance was compared between habitat types. The habitat and rabbit variables were combined into a few summary factors, which were then used to build a model of wildcat abundance.

2. What they found

Where wildcats were found

• Wildcat scats were found in all five habitat types, but on only 42 of the 101 transects.
• Abundance tended to be highest in the Mediterranean mountain vegetation, although this was marginally short of statistical significance.
• The only clear difference was between the Mediterranean mountains (highest) and the mountain pine forests (lowest).
• Agricultural steppes did not differ from the other habitats, although abundance there tended to be lower than in the Mediterranean mountains. Mountains and plains did not differ.

Habitat structure

The variables boiled down to three factors that together explained about 72% of the variation between transects:

• High, rough, dense forest versus flat cropland with many rabbits.
• Rough areas with a mosaic of scrub and pasture, rabbits and watercourses versus pure cropland.
• Scrubland with watercourses versus urban areas and pasture.

The model

Only the second factor entered the model. Wildcats were more numerous in scrub and pasture mosaics with rabbits and watercourses, and scarce where there were only crops. No forest variable was included. The model was highly significant but explained only about 11% of the variation in wildcat abundance.

3. What they concluded

• Wildcats can live in forests, but forest is not their preferred habitat where alternatives exist, especially scrubland in Mediterranean areas, so the species cannot be considered a true forest species.
• The “forest hypothesis” probably came from early studies in central Europe, where few other habitats were available, and survived partly because a key Scottish study showing a preference for open habitat was published as a technical report rather than in a scientific journal.
• Agricultural steppes were used as much as the other habitats, the first time this has been reported. Small scrub patches and good numbers of rodents and rabbits probably provide shelter and food there, much as forest fragments and riverside woodland do elsewhere.
• Intensive cropland, with little shelter, water or prey, was the worst habitat.
• Habitat structure alone explained little of where wildcats were, so other factors must matter too.

4. How tracking was or was not used

The whole study rests on scat surveys. The author walked trails and recorded wildcat scats in each 200 m segment, using the fact that, where the two species live together, only wildcats leave scats exposed along paths. Domestic cat scats were told apart from wildcat scats following the methods of the earlier study. Rabbit abundance was measured by counting rabbit latrines. No trailing was reported.

5. Tracker or observer certification

Reported but not validated. The author says confusion with scats of other carnivores “seems negligible in view of the author’s experience in cat excrement identification.” This experience was not tested in the study, and no genetic or other check of scat identification is reported.

6. Other technologies, skills or innovations

• A simple wildcat shelter index (1 to 5) based on cavities in rocks and trees and how far one can see.
• Landscape variables counted from land-use maps using a grid of 121 evenly spaced points in a 9 km² area around each transect.
• Habitat variables combined with principal component analysis (PCA) before modelling, to avoid problems from variables that are strongly related to each other.

7. Implications for conservation (from the paper)

• Wildcat populations should be protected wherever they are found, whatever the habitat type.
• In Mediterranean regions the key habitat element appears to be scrubland, and the author recommends preventing its destruction.
• Rural land management priorities, in the author’s view, should be keeping mosaic landscapes of open and closed habitat patches and recovering rabbit populations.
• The author considers that ending predator control in hunting areas is essential for long-term wildcat conservation and recovery, as it is currently one of the main problems the species faces in Spain.