Coral reef fish food webs

Reef fish food web, Greater Antilles

Here are a couple of renderings of the vertebrate-only component of the coral reef food web. Reminder: the food web is what we expect to see for a reef in the Greater Antilles of the Caribbean, based on data collected around the mid-20th century. The vertebrate component comprises all fish and sea turtle species. The upper figure is the expected food web, and includes 196 species and 995 trophic interactions. Species are arranged on the periphery of the diagram, with interaction represented by the lines crossing the interior. The very busy, or hub species are higher trophic level predators, mostly carcharhinid sharks.

Jamaica coral reef fish food web

The lower figure is what we observe today in Jamaica. (Note: Jamaica is of particular interest for me as a starting comparison, both because of the excellent documentation of those reefs, and my Jamaican heritage; not picking on Jamaica). The number of species, out of 196, observed there over the past 10 years is dramatically smaller. Perhaps more obvious is the loss of interactions. I won’t present the actual data yet, since we will eventually prepare a paper to report all this, but the differences between the two food webs are obvious. We are currently rendering the complete food web, including primary producers and invertebrates, which will be an update of the figures presented in earlier posts. But there are a lot of species in there, and the computers have been churning now for about 17 hours!

Jamaican coral reef I

Fig. 1 - Species-level trophic link distribution for entire coral reef.

We’ve examined records of fish occurrences on Jamaican reefs for the past 10 years, and compared it to our “master” food web. Of the 196 species in our food web, 136 have records in Jamaica. Many of these species are present in very low numbers, and some reefs are noticeably depauperate, recording less than 60 species. Nevertheless, to be conservative, we assume that we can integrate over all the reefs, thereby counting all 136 species as being present. We next expanded our metanetwork, or guild-level food web (in this case almost exactly the same as a trophic species-based web) to the species level, therefore accounting for all expected links in the food web. For the master or pristine web, this yields an overall connectance of 0.059. The trophic link distribution is shown in Fig. 1. Interestingly, this is clearly not a decay distribution (e.g. power law), but has a definite modality of about 25 links. One needs to question the extent to which under-sampling of natural food webs, and aggregation into trophic species, affects interpretation of link distributions.

The next step of course is to assess the state of the Jamaican reef system. Our initial analysis has been to simply remove the “missing” species (extirpated) from the web, and to re-calculate the statistics. Connectance declines to 0.055. Is this significant? Probably impossible to answer that question for network connectance. Also, it should be noted that hundreds of invertebrate species are included here, and they will dampen the impact of any fish removals or additions. Perhaps the next question regards the link properties of the extirpated species.