Databases and meta-analysis

The “CO2 science” web site has produced a database “on the response of marine organisms to ocean acidification as reported in the peer-reviewed scientific literature. Results are tabulated by response, including calcification, fertility, growth, metabolism and survival.”

This document is published by the Center for the Study of Carbon Dioxide and Global Change, a non-profit organization that is skeptical of the negative direct and indirect impacts of elevated CO2. It is not peer-reviewed. I strongly encourage readers to use peer-reviewed sources of ocean acidification data and meta-analysis such as those listed below.

Jean-Pierre Gattuso


Hendriks I. E. & Duarte C. M., 2010. Ocean acidification: separating evidence from judgment. A reply to Dupont et al. Estuarine, Coastal and Shelf Science 89:186-190. Article (subscription required).

Dupont S., Dorey N. & Thorndyke M., 2010. What meta-analysis can tell us about vulnerability of marine biodiversity to ocean acidification? Estuarine, Coastal and Shelf Science 89:182-185. Article (subscription required).

Nisumaa A.-M., Pesant S., Bellerby R. G. J., Middelburg J. J., Orr J. C., Riebesell U., Tyrrell T., Wolf-Gladrow D. & Gattuso J.-P., 2010. EPOCA/EUR-OCEANS data compilation on the biological and biogeochemical responses to ocean acidification. Earth System Science Data 2:167-175. Article.

Liu J., Weinbauer M. G., Maier C., Dai M. & Gattuso J.-P., in press. Effect of ocean acidification on microbial diversity, and on microbe-driven biogeochemistry and ecosystem functioning. Aquatic Microbial Ecology

Kroeker K., Kordas R. L., Crim R. N. & Singh G. G., in press. Meta-analysis reveals negative yet variable effects of ocean acidification on marine organisms. Ecology Letters. Article (subscription required).

1 Response to “Databases and meta-analysis”

  1. 1 Jon Havenhand 8 October 2010 at 09:40

    Excellent point Jean-Pierre.

    Just one comment about the use of different metrics in meta-analyses.

    Use of statistics such as means and 95% Confidence Intervals is only relevant for data that are normally distributed. The “Response Ratio” used by Hendriks et al (and to an extent Dupont et al) is extremely non-normally distributed, and therefore these metrics and the conclusions arising from them may be very misleading.

    Meta-analyses that use normalised metrics, or non-parametric estimates of means and confidence intervals (eg Kroeker et al) are to be highly recommended.


Comments are currently closed.

  • Reset


OA-ICC Highlights

%d bloggers like this: