Impacts of ocean acidification on marine organisms: quantifying sensitivities and interaction with warming

Abstract

Ocean acidification represents a threat to marine species worldwide, and forecasting the ecological impacts of acidification is a high priority for science, management, and policy. As research on the topic expands at an exponential rate, a comprehensive understanding of the variability in organisms' responses and corresponding levels of certainty is necessary to forecast the ecological effects. Here, we perform the most comprehensive meta-analysis to date by synthesizing the results of 228 studies examining biological responses to ocean acidification. The results reveal decreased survival, calcification, growth, development and abundance in response to acidification when the broad range of marine organisms is pooled together. However, the magnitude of these responses varies among taxonomic groups, suggesting there is some predictable trait-based variation in sensitivity, despite the investigation of approximately 100 new species in recent research. The results also reveal an enhanced sensitivity of mollusk larvae, but suggest that an enhanced sensitivity of early life history stages is not universal across all taxonomic groups. In addition, the variability in species' responses is enhanced when they are exposed to acidification in multi-species assemblages, suggesting that it is important to consider indirect effects and exercise caution when forecasting abundance patterns from single-species laboratory experiments. Furthermore, the results suggest that other factors, such as nutritional status or source population, could cause substantial variation in organisms' responses. Last, the results highlight a trend towards enhanced sensitivity to acidification when taxa are concurrently exposed to elevated seawater temperature.

Fig. 1

Mean effect of near future acidification on major response variables. Significance is determined when the 95% bootstrapped confidence interval does not cross zero. The number of experiments used to calculate the mean is included in parentheses. *denotes a significant effect.

Fig. 2

Variation in effect sizes among key taxonomic groups, divided by major response variables. Note there are different scales on the y-axes to highlight the variation among taxa. Means are from a weighted, random-effects model with bootstrapped bias-corrected 95% confidence intervals. The number of experiments used to calculate the means is given in parentheses. Not all response variables are considered in this analysis. *denotes a significant difference from zero.

Fig. 3

Variation in effects of acidification among taxa for development. Means are from weighted, random effects meta-analysis and are shown with bias-corrected bootstrapped 95% confidence intervals. The number of experiments used to calculate each mean is given in parentheses. *denotes a significant difference from zero.

Fig. 4

Summary of effects of acidification among key taxonomic groups. Effects are represented as either mean percent (+) increase or percent (−) decrease in a given response. Percent change estimates were back transformed from the mean LnRR, and represent geometric means, that are conservative of the arithmetic means.

Fig. 5

Significant variation in the effects of near-future ocean acidification among lifestages within taxonomic groups. Error bars represent bias-corrected bootstrapped 95% confidence intervals, and the number of experiments used to calculate the means is shown in parentheses. The * associated with mollusk survival and metabolism denotes a significant difference in effect size among life history stages (Significant Q_M).

Fig. 6

The effect of duration of experimental CO₂ enrichment on LnRR. The mean effect size and 95% CI (for all taxa pooled) is shown on the left of each figure (overall), while the individual LnRR estimates for each study are plotted against duration (days) on the right side of the figure for survival, calcification, growth, photosynthesis and development.

Fig. 7

Variation in effect of acidification treatment at ambient temperature and elevated temperature for different response variables. Means are from weighted, random effects categorical meta-analyses for each separate response variable. Error bars represent bias-corrected bootstrapped 95% confidence intervals, and the number of experiments used to calculate the means is shown in parentheses. *denotes a significant difference from zero.