97th ESA Annual Meeting (August 5 -- 10, 2012)

COS 129-7 - Spatial interaction network and species traits predetermine metacommunity structure and its potential dynamics

Thursday, August 9, 2012: 10:10 AM
D137, Oregon Convention Center
Jurek Kolasa1, Julia Kim1, Tamara N. Romanuk2 and W. Mather A. Carscallen3, (1)Biology, McMaster University, Hamilton, ON, Canada, (2)Dalhousie University, (3)Biology, Dalhousie University, Halifax, NS, Canada
Background/Question/Methods . Quasi discrete communities are thought to form a metacommunity where species affect the dynamics of local communities by dispersal.  Dynamics of the entire metacommunity however is thought to reflect both the states of local communities as well as exchanges among them, with exchanges’ success being predicated on species interactions.  However, these interactions are only possible, if species occupy a location at the same time, which often is not the case.  We hypothesized that the potential for interactions will generally increase with the decreasing habitat specialization of species.  To test this we used network analysis where node size reflected the relative magnitude of special overlap of a species with all other species in the metacommunity of 49 coastal rock pools in Jamaica. We obtained nodes for 13 annual surveys and related them to species occupancy (number of sites occupied), Levins’ niche, and ecological range of each species that was based on a combination of physical attributes of rock pools and species abundance (Canonical Correspondence Analysis (CANOCO). 

Results/Conclusions . In the main, we found support for the idea that a decrease in a species’ specialization confers a greater potential to interact (network node size) with other members of metacommunity.  However, we discovered important secondary patterns.  Specifically, node size was affected differently by occupancy and niche breadth.   While the node size increased at a decelerating rate in response to occupancy of a species, it showed a decline after an initial and substantial rise for all the measures of niche breadth.  This seems to suggest that species with broadest niches are often decoupled from community majority.  On the opposite end of the specialization spectrum, specialists showed the greatest relative variation (Coefficient of Variation) of the network node size.  This implies that although specialists tend to have an appreciable node size (approximately 50% of that shown by the species in the middle of the specialization range), their potential for interaction with other species is highly unpredictable in space and time.  It is not yet entirely clear how the decoupling of generalists and esoteric behavior of specialists should be incorporated into metacommunity theory and conservation practice.  The results suggest however that dynamics of a metacommunity is affected by selective species sorting, partial independence of generalists (they can occur alone or with few other species), and a gradient of unpredictability inversely correlated with specialization.