The nature of ecological communities has been a longstanding question in ecology since the debate between F.E. Clements and H.A. Gleason (Ricklefs, 2008). While Clements (1936) viewed communities as closed structures that tend to persist through time, Gleason (1926) perceived them as dynamic entities resulting from the mere coincidence of species’ distributions in space and time. Taken to its extreme, the Clementsian view implies that the characteristics and ecological roles performed by the species that form a community determine which new species will be able to establish populations in the locality through either facilitation or competitive displacement (see Clements, 1936). The absence of large-scale processes from Clements’ ideas – which focus exclusively on local interactions – may create the false impression that community processes hardly scale up to geographical patterns. In contrast, Gleason’s view embraces processes acting at multiple scales (from local habitat selection to the species’ macroclimatic requirements), that determine the individualistic response of each species.
Reconciling both perspectives has proven difficult, perhaps partly due to the acrimonious debate on assembly rules held in the 1970s and 1980s (see Gotelli & Graves, 1996). As a result, community ecology and biogeography have developed as largely separate disciplines during the last decades, with the Gleasonian view being the dominant paradigm among biogeographers. Evidence about the importance of regional processes for the structure of local assemblages has led to the radical proposition of ‘disintegrating’ the concept of the ecological community (Ricklefs, 2008). This view neglects the impact of species’ interactions or metacommunity dynamics – typically studied by community ecologists – on the geographical distribution of diversity (but see Hortal et al., 2010). However, not all species from the pool of colonizers of a locality are present locally (Guisan & Rahbek, 2011). Understanding the processes determining the assembly of local communities remains a key objective of both community ecology and biogeography.
A bold proposal by Guisan & Rahbek (2011) may help unify analyses of local and regional processes into a single framework. It is well known that species richness predictions obtained by modelling directly the relationship between species numbers and the environment (macroecological models, MEMs) are different from those recovered by stacking the results of individual species distribution models (SDMs) (e.g. Aranda & Lobo, 2011). Guisan & Rahbek (2011) argue that such differences are the result of dispersal filtering and ecological assembly rules (EARs). Some of the species that could inhabit a given place according to their environmental requirements – which may be identified through species distribution modelling (SDM) – are prevented from establishing populations, first by historical events limiting dispersal within a region, and later by limitations to local coexistence – which can be partly measured through macroecological modelling (MEM). Feeding on this rationale, Guisan & Rahbek (2011) propose a hierarchical framework – named SESAM (spatially explicit species assemblage modelling) – to understand and predict species diversity patterns through a combination of SDM, MEM and the indirectly assumed roles of dispersal and biotic filters (i.e. EARs).
Long-term evolutionary processes build up regional biotas hosting a number of species with particular adaptations and, many times, a history of co-evolution (Ricklefs, 2008). The characteristics of such regional pools for a given biogeographical realm can be represented by the phylogeny and ecological traits of its species (e.g. Diniz-Filho et al., 2011; Pavoine & Bonsall, 2011). But within the realm, the source pools for different localities are determined by the spatio-temporal dynamics of species’ distribution ranges that constitute the core of the Gleasonian view. Only the species that could live in a particular landscape and manage to reach it and/or sustain populations will be able to be present in a given locality and thus form part of its local source pool. The dynamics of distribution ranges may be studied through new-generation SDMs accounting for dispersal and environmental filters (movement and abiotic factors sensuSoberón, 2010). However, the application of this particular part of SESAM may be hampered by the extreme difficulty of estimating the potential distribution of the species with reliability, because of both conceptual and data limitations of the SDM approach (see, e.g. Jiménez-Valverde et al., 2011). Notwithstanding that SDM may fail to recover the fundamental niche of the species in full, methods accounting for the lack of equilibrium of species’ distributions with the environment allow these local source pools to be studied and predicted (De Marco et al., 2008).
Only a few species from the local source pool are actually present in the local assemblage at a given moment of time. This is the scale at which the Gleasonian and Clementsian views collide, attributing the limits of local coexistence to different processes: habitat selection; or competition, facilitation and other biotic interactions, respectively. The effect of the latter processes on species’ distributions has been already accounted for by the inclusion of population-level processes that can eventually result in local extinctions due to bionomic effects (Soberón, 2010; see also Hortal et al., 2010). But SESAM provides a framework by which we may merge both views at the community level. Guisan & Rahbek (2011) argue that the total number of species that can inhabit a community can be estimated through common MEMs that relate factors such as available energy with the carrying capacity of the locality. This assumption is partly Gleasonian – for it denotes environment-driven community patterns – and partly Clementsian – because it attributes a major role to local competition. But, importantly, SESAM allows the distillation of most Clementsian processes into the separate filter of EARs, the final step that determines the identity and characteristics of the species that succeed in joining the local assemblage.
Given that the effects of dispersal processes and carrying capacity have already been included in previous levels of the SESAM framework, most EARs would correspond to Clementsian effects of different intensities. Here, the particular characteristics of each of the species present in the local source pool prevent or facilitate its establishment in the local community. These effects may be measured as the departures of different measures of phylogenetic and functional diversity from the null expectations derived from local richness and the composition of the source pool (see Pavoine & Bonsall, 2011). These two types of measures depict the evolutionarily structured and phenotypic – and thus functional for traits linked to ecosystem functioning – variation in the species present in the assemblage. Both sources of variation are highly correlated, because many morphological and functional variations are the result of long-term evolutionary processes. Their effects can, however, be partitioned into independent and shared fractions (Diniz-Filho et al., 2011). This may allow the separation of the overall effects of non-specific bionomic processes acting within ecological timeframes, such as competitive exclusion (which can be described by phylogenetically independent variation in functional diversity), from those arising from specific species interactions resulting from long-term co-evolutionary processes (which can be described by phylogenetically dependent variations in functional diversity).
Although integrating all the concepts and approaches discussed above needs further discussion, a comprehensive application of the SESAM framework has the potential to provide the basis for a synthesis of the joint effects of biogeographical and community-level processes on the diversity and structure of local communities. Future applications of SESAM in well-known systems, coupled with data on historical processes and life history traits, will allow a better understanding of the origin and dynamics of diversity patterns.