Coral Reefs in Crisis: The Reliability of Deep-Time Food Web Reconstructions as Analogs for the Present

Abstract

Ongoing anthropogenic alterations of the biosphere have shifted emphasis in conservation biology from individual species to entire ecosystems. Modern measures of ecosystem change, however, lack the extended temporal scales necessary to forecast future change under increasingly stressful environmental conditions. Accordingly, the assessment and reconstruction of ecosystem dynamics during previous intervals of environmental stress and climate change in deep time has garnered increasing attention. The nature of the fossil record, though, raises questions about the difficulty of reconstructing paleocommunity and paleoecosystem-level dynamics. In this study, we assess the reliability of such reconstructions by simulating the fossilization of a highly threatened and disturbed modern ecosystem, a Caribbean coral reef. Using a high-resolution coral reef food web from Jamaica, we compare system structures of the modern and simulated fossil reefs, including guild richness and evenness, trophic level distribution, predator dietary breadth, food chain lengths, and modularity. Results indicate that despite the loss of species, guilds, and trophospecies interactions, particularly zooplankton and other soft-bodied organisms, the overall guild diversity, structure, and modularity of the reef ecosystem remained intact. These results have important implications for the integrity of fossil food web studies and coral reef conservation, demonstrating that fossil reef communities can be used to understand reef community dynamics during past regimes of environmental change.

Appendices

Appendix 1

1.1 Hypergeometric Variance

The hypergeometric distribution describes the probability given a population with a featured subset of observations, and a sample from the population, that a random selection of individuals from the population will include a certain fraction of the sample comprising individuals with the featured characteristic. In this case, the population consists of 728 species (N), of which 433 are featured as fossilized (K). A sample is the richness of a trophospecies (n), of which a subset are observed as fossilized (k). The expected fossilization or hypergeometric mean value is

$$\displaystyle \begin{aligned} \mathtt{E}(\hat{k}) = n \frac{K}{N} \end{aligned} $$

(9)

The variance of the expectation is given as

$$\displaystyle \begin{aligned} \sigma^{2}(k) = n \frac{K}{N} \frac{N-K}{N} \frac{N-n}{N-1} \end{aligned} $$

(10)

This value grows initially as n because the number of unique ways in which k objects may be selected also grows. The value declines, however, as n → N (Fig. 12).

Fig. 12

Hypergeometric variance given a total population of 728 species, and trophospecies of increasing size

Appendix 2

See Table 1.