Tags

Cayman Islands coral reef vertebrates food web

We have now constructed models of three Greater Antillean coral reef communities: the Cayman Islands, Jamaica and Cuba. Granted, the models/communities are defined by political entities, but there is no denying that physical contiguity of the reefs within these nations far exceeds the connections between localities. So we’ll consider them as separate communities within a regional ecosystem. The question that we’re interested in is, are there significant differences among these reef communities that cannot be explained by stochastic processes, such as the vagaries of immigration, colonization and population persistence.

Cuba

The overall vertebrate (fish and reptiles) regional diversity in the models, or gamma diversity, is 169 species (a total of 735 species in the food web, including primary producers, invertebrates, etc.). The food web alpha diversities are: Caymans – 155, Cuba – 134 and Jamaica – 147. These diversities are all very close to each other, as well as the gamma diversity. But are they unbiased samples of gamma diversity? As a first test of this, we asked whether the connectance of the local food webs could be derived from the regional pool in unbiased fashion. We therefore constructed 1,000 food webs, at each level of alpha diversity, by drawing species randomly from the regional pool, calculating their connectances, and then comparing the observed to the “expected”. Interestingly, the answer is a resounding NO!

• Cayman connectance (C) = 0.0594606; randomizations, $\bar{x}=0.0592537$, p=0.295.
• Cuba C = 0.0594165; $\bar{x}=0.0590932$, p = 0.298.
• Jamaica C = 0.0620494; $\bar{x}=0.0592085$, p<0.00001.

Jamaica

You should notice that connectance bears no relation to alpha diversity. In fact, the randomizations fail to reject the hypothesis that the Cayman and Cuban reefs are anything other than random draws from the regional pool. Jamaica, however, stands out. Connectance of the Jamaican web is significantly greater than expected. This implies that if the Jamaican web were indeed drawn from the regional pool, then there would be a bias in favour of link-rich species. Translated ecologically, Jamaica is relatively depauperate in trophically specialized species (few links).

What could generate that bias? We hypothesize the more degraded state of Jamaican reefs relative to the Caymans, a proportionally smaller area of protected reefs, and greater fishing pressure. Interestingly, Cuba has fewer species included in our model, yet shows no biases, as expected given our hypothesized factors.