Original article Intercohort density dependence drives brown trout habitat selection Daniel Ayllón a , Graciela G. Nicola b , Irene Parra a , Benigno Elvira a , Ana Almodóvar a, * a Department of Zoology and Physical Anthropology, Faculty of Biology, Complutense University of Madrid, Ciudad Universitaria s/n, E-28040 Madrid, Spain b Department of Environmental Sciences, University of Castilla-La Mancha, E-45071 Toledo, Spain article info Article history: Received 24 April 2012 Accepted 22 October 2012 Available online Keywords: Resource selection functions Habitat modelling Population management Conservation Competition Salmonids abstract Habitat selection can be viewed as an emergent property of the quality and availability of habitat but also of the number of individuals and the way they compete for its use. Consequently, habitat selection can change across years due to fluctuating resources or to changes in population numbers. However, habitat selection predictive models often do not account for ecological dynamics, especially density dependent processes. In stage-structured population, the strength of density dependent interactions between individuals of different age classes can exert a profound influence on population trajectories and evolutionary processes. In this study, we aimed to assess the effects of fluctuating densities of both older and younger competing life stages on the habitat selection patterns (described as univariate and multivariate resource selection functions) of young-of-the-year, juvenile and adult brown trout Salmo trutta. We observed all age classes were selective in habitat choice but changed their selection patterns across years consistently with variations in the densities of older but not of younger age classes. Trout of an age increased selectivity for positions highly selected by older individuals when their density decreased, but this pattern did not hold when the density of younger age classes varied. It suggests that younger individuals are dominated by older ones but can expand their range of selected habitats when density of competitors decreases, while older trout do not seem to consider the density of younger individuals when distributing themselves even though they can negatively affect their final performance. Since these results may entail critical implications for conservation and management practices based on habitat selection models, further research should involve a wider range of river typologies and/or longer time frames to fully understand the patterns of and the mechanisms underlying the operation of density dependence on brown trout habitat selection. Ó 2012 Elsevier Masson SAS. All rights reserved. 1. Introduction Ecologists have long been interested in the consequences of habitat selection for predicting the distribution and abundance of animals (Morrison et al., 2006). Habitat selection is of great importance in ecological theory because this behaviour is a primary way that mobile organisms adapt to changing conditions (Morris, 2011; Railsback et al., 2003), which is turning an increasingly crit- ical matter in the light of current climate change. Habitat has been in fact the cornerstone for wildlife conservation and management, and so ecologists have developed sophisticated tools to charac- terize how species use space and resources (McLoughlin et al., 2010). These empirical models are often used to map habitat quality at different spatial scales and to inform managers on the future availability and use of habitats (Morris et al., 2008). At least three fundamental types of predictive models can be used to define habitat selection from correlative data: distributional or macro- habitat models, which predict the presence or absence of species at large spatial scales; capacity models, which predict density or population size when a taxon is present; and microhabitat models, which predict habitat associations at a fine spatial scale (Morrison et al., 2006; Rosenfeld, 2003). Within the available procedures that quantify relative use of habitat resources, habitat suitability models (HSMs; e.g., Hirzel and Le Lay, 2008), and related resource selection and resource selection probability functions (RSFs; Manly et al., 2002) are probably the most popular. These models are easily linked to geographic infor- mation systems (GIS), so the rapid development of this technology and the growing availability of digital landscape data have rendered HSM and RSF models powerful tools for wildlife management and the identification of conservation priority sites (Boyce et al., 2002; Braunisch et al., 2008). Likewise, HSMs or RSFs developed at either the micro or mesohabitat scale have been the biological basis to * Corresponding author. Tel.: þ34 91 3945135; fax: þ34 91 3944947. E-mail addresses: daniel.ayllon@bio.ucm.es (D. Ayllón), graciela.nicola@uclm.es (G.G. Nicola), irenepm@bio.ucm.es (I. Parra), belvira@bio.ucm.es (B. Elvira), aalmodovar@bio.ucm.es (A. Almodóvar). Contents lists available at SciVerse ScienceDirect Acta Oecologica journal homepage: www.elsevier.com/locate/actoec 1146-609X/$ e see front matter Ó 2012 Elsevier Masson SAS. All rights reserved. http://dx.doi.org/10.1016/j.actao.2012.10.007 Acta Oecologica 46 (2013) 1e9