Ecology, 92(4), 2011, pp. 859–870 Ó 2011 by the Ecological Society of America Interactive effects of disturbance and dispersal directionality on species richness and composition in metacommunities FLORIAN ALTERMATT, 1,3 SEBASTIAN SCHREIBER, 2 AND MARCEL HOLYOAK 1 1 Department of Environmental Science and Policy, University of California, 1 Shields Avenue, Davis, California 95616 USA 2 Department of Evolution and Ecology, University of California, 1 Shields Avenue, Davis, California 95616 USA Abstract. Dispersal among ecological communities is usually assumed to be random in direction, or to vary in distance or frequency among species. However, a variety of natural systems and types of organisms may experience dispersal that is biased by directional currents or by gravity on hillslopes. We developed a general model for competing species in metacommunities to evaluate the role of directionally biased dispersal on species diversity, abundance, and traits. In parallel, we tested the role of directionally biased dispersal on communities in a microcosm experiment with protists and rotifers. Both the model and experiment independently demonstrated that diversity in local communities was reduced by directionally biased dispersal, especially dispersal that was biased away from disturbed patches. Abundance of species (and composition) in local communities was a product of disturbance intensity but not dispersal directionality. High disturbance selected for species with high intrinsic growth rates and low competitive abilities. Overall, our conclusions about the key role of dispersal directionality in (meta)communities seem robust and general, since they were supported both by the model, which was set in a general framework and not parameterized to fit to a specific system, and by a specific experimental test with microcosms. Key words: asymmetric dispersal; density; directionally biased dispersal; ecological gradients; metacommunities; microcosms; species traits. INTRODUCTION Natural landscapes frequently contain variation in habitat patch quality and extent of disturbance among habitat areas. Individually the impact of these factors on species diversity has been extensively explored. This is exemplified by studies of fixed spatial habitat heteroge- neity (e.g., Davies et al. 2009), the metacommunity concepts of species sorting and mass effects (e.g., Kneitel and Chase 2004, Leibold et al. 2004, Urban 2004), and the intermediate disturbance hypothesis (IDH; Connell 1978, Shea et al. 2004, Cadotte 2007). Primarily such studies only consider the simplest form of dispersal among local communities. For instance, metacommun- ity models are usually either based on the spatially implicit Levins (1970) metapopulation model with equal dispersal among patches (Mouquet and Loreau 2002, Amarasekare et al. 2004), or dispersal is inversely related to inter-patch distance, without considering directionality (reviewed in Clobert et al. 2001, Hanski and Gaggiotti 2004). However, the ecology of dispersal is often much more complex than these simple model representations (e.g., Levin et al. 2003, Nathan et al. 2008). In many natural systems, landscape structure and physical flows may dictate dispersal of organisms and influence where and how far they disperse (Haddad 1999). Here, we consider the effects of disturbance on species richness when dispersal is directionally biased. Directionally biased dispersal means that most dispers- ers move predominantly in one spatial direction, and this type of dispersal is likely to be prevalent for species living along environmental gradients or subject to directionally biased environmental currents. There are both theoretical and empirical reasons for considering such forms of dispersal. Existing theoretical work shows that directionally biased dispersal can alter population viability and species coexistence (Vuilleumier and Possingham 2006, Salomon et al. 2010), and that it can be detrimental to metapopulation viability (Elkin and Possingham 2008). Certain types of systems, such as streams, oceanic current systems, or systems with relatively fixed prevailing wind direction, may predis- pose some organisms to a directional bias in dispersal. Surprisingly there have been few attempts to look at the effects of directionally biased dispersal on diversity in ecological communities, either empirically (but see for example MacIsaac 1996) or theoretically (Salomon et al. 2010). Another perspective on directionally biased dispersal is to consider the explicit spatial layout of a landscape. Manuscript received 30 May 2010; revised 20 September 2010; accepted 30 September 2010. Corresponding Editor: A. M. de Roos. 3 Present address: Department of Aquatic Ecology, Ea- wag: Swiss Federal Institute of Aquatic Science and Technology, U ¨ berlandstrasse 133, CH-8600 Du¨ bendorf, Switzerland. E-mail: faltermatt@bluewin.ch 859