Main content area

Patterns and processes in insular floras affected by hurricanes

Morrison, Lloyd W., Spiller, David A.
Journal of biogeography 2008 v.35 no.9 pp. 1701-1710
hurricanes, islands, linear models, nestedness, regression analysis, Bahamas
To investigate species compositions, rates of species turnover, species-area and species-distance relationships and patterns of nestedness in the floras of small Bahamian islands, by comparing two groups of islands that had been differentially affected by two hurricanes. Small islands occurring on either side of Great Exuma near Georgetown, Bahamas. We surveyed the plant species of 44 small islands over a 5-year period from 1998 to 2002. Hurricanes Lili and Michelle occurred in 1996 and 2001, respectively; both storms affected small islands on the more exposed south-west side of Great Exuma to a greater degree than small islands on the more protected north-east side. A set of 27 islands was surveyed in 1998 and 2002 to evaluate species turnover. Stepwise multiple linear regression analyses and an information-theoretic approach (the Akaike information criterion) were used to elucidate the importance of area and distance as predictors of plant species number. We compared a piecewise linear regression model with a simple linear regression of species number against area to determine whether a small island effect existed. Nestedness patterns were evaluated by Wilcoxon two-sample tests to analyse occurrence sequences. Species turnover was low in an absolute sense (overall = 0.74% year⁻¹), yet was over three times higher than that documented in a nearby archipelago in the absence of hurricanes. Both vegetated area and distance were important predictor variables for exposed islands but not for protected islands. Some support was found for a small island effect for the exposed islands based on a piecewise linear regression model. Both island groups revealed significant nestedness at the level of the assemblage (both P < 0.001). On exposed islands, 65-79% (depending upon the method of calculation) of all species were significantly nested, but only 47% of all species were significantly nested on protected islands. Overall, these insular floras seem highly resistant to hurricane-force disturbances. Species turnover was low (< 1% year⁻¹) in an absolute sense, particularly in comparison with rates for other taxa. Higher degrees of nestedness and significant species-area and species-distance relationships for exposed islands indicated stronger patterns of community assembly. It is likely that disturbance is a major structuring force for the exposed islands, although the type of disturbances that mediate these patterns may not be primarily hurricane-force storms.