Personal webpage

Habitat selection models for European wildcat conservation

Klar, N., Fernández, N., Kramer-Schadt, S., Herrmann, M., Trinzen, M., Büttner, I. & Niemitz, C.
Biological Conservation in press.

Populations of the European wildcat (Felis silvestris) are only slowly recovering in Central Europe after a severe decline in the last centuries and require specific conservation plans in many areas. However, detailed information on wildcat occurrence and habitat requirements is still scarce and controversial. We present a fine-scale habitat selection model for wildcats based on detailed species and land use information and evaluate its accuracy to predict habitat distribution in new areas. We analysed habitat use within home ranges using single locations of 12 radio-tracked individuals from south western Germany. Several competing models were fitted and compared using generalised linear mixed models (GLMM) and information-theoretic approaches. Radio-tracking data of 9 and 10 wildcats from two distant areas were used to evaluate the models. The selected model predicted habitat associated to close distance to forest, watercourses and meadows and a critical distance to villages, single houses and roads. To predict area suitable for home ranges we superimposed rules derived from home range attributes at a higher level of selection. Predictions from the combination of the fine-scale habitat model and home range rules matched well with more than 2 000 wildcat observations of south western Germany. We discuss the application of the model in wildcat conservation for finding potential reintroduction sites, identifying small isolated populations and aiding in the evaluation of the needs of mitigation and compensation within the scope of the European Habitats Directive.

Keywords: Carnivore conservation; Felis silvestris; Use-availability habitat model; Germany; GLMM; European Habitats Directive

Animal habitat quality and ecosystem functioning: exploring seasonal patterns using NDVI

Wiegand, T., Naves, J., Garbulski, M &Fernández, N.
Ecological Monographs in press.

Many animal species developed specific evolutionary adaptations to survive prolonged periods of low energy availability that characterize seasonal environments. The seasonal course of primary production, a major aspect of ecosystem functioning, should therefore be an important factor determining habitat quality of such species. We tested this hypothesis by analyzing the relationship between habitat quality and ecosystem functioning for brown bears (Ursus arctos), a species showing hyperphagia and hibernation as evolutionary adaptation to seasonal peaks and bottlenecks in ecosystem productivity, respectively. Our unique long-term data set comprised data from two populations in northern Spain on historical brown bear presence, current presence, and reproduction. The data were classified on a grid of 5km ´ 5km pixels into five classes: frequent reproduction, sporadic reproduction, frequent presence, sporadic presence, and recent extinction. We used the long-term average of the seasonal course of NDVI (normalized difference vegetation index) as proxy for ecosystem functioning and investigated the relation between habitat quality and ecosystem functioning with methods borrowed from statistical point pattern analysis. We found that brown bears selected indeed habitat with specific ecosystem functioning (i.e., the variance in all habitat classes was smaller than the landscape overall) and the relation between habitat quality and ecosystem functioning was ordered. First, the average distance in ecosystem functioning between two habitat classes was larger if the difference in habitat quality was larger. Second, habitat with the most excessive needs (i.e., breeding habitat) occupied the narrowest niche regarding ecosystem functioning and showed the most pronounced seasonality. Progressively poorer classes occupied wider niches which partly overlapped that of better classes. This indicated that non-breeding animals are less selective. Our methodology provided new insight into the relation between ecosystem functioning and habitat quality and could be widely applied for animal species living in seasonal environments. Since NDVI data are continuously collected our methodology allows for continuous monitoring of changes in habitat quality due to global change.

Keywords: Ecosystem functioning; endangered species; extinction; habitat quality; NDVI; remote sensing; Northern Spain; point-pattern analysis; seasonality; Ursus arctos.

Habitat-related heterogeneity in breeding in a metapopulation of the Iberian lynx

Fernández, N., Delibes, M. & Palomares, F. (2007).
Ecography 30: 431-439.

Identifying attributes associated with good breeding habitat is critical for understanding animal population dynamics. However, the association between environmental heterogeneity and breeding probability has been often overlooked in habitat analyses. We evaluated habitat quality in a metapopulation of the endangered Iberian lynx Lynx pardinus by analyzing spatiotemporal patterns in breeding records. Data summarizing successful production of litters after emergence from dens over four years within 13 lynx territories were examined. We designed a set of generalized linear mixed models representing different hypotheses regarding how patterns in breeding records relate to environmental heterogeneity. Environmental heterogeneity was described by two characteristics: 1) a landscape index measured in lynx territories indicative of time-averaged prey availability and 2) yearly variability in prey abundance not captured with this index. By including the random effect of the lynx territory we also accounted for other territory-specific effects on reproduction. We found significant differences in yearly prey density dynamics among lynx territories. However, temporal variation in prey density contributed poorly to explaining lynx breeding. The most parsimonious model included the landscape structure as the only effect explaining breeding patterns. A multinomial-model-representation of the landscape hypothesis explained nearly 50% of variability in breeding records. Results pointed to the existence of a habitat quality gradient associated with particular landscape structures influencing lynx habitat selection and breeding performance. Underlying this gradient was time-averaged prey availability. Probably as a result of longterm fitness strategies in long-lived territorial species, the short-term fluctuations in prey availability had a minor influence. Our results illustrate how habitat inferences can be enhanced by incorporating the link between spatiotemporal patterns in reproduction and environmental heterogeneity.