Heteroscedasticity is your friend: diurnally fluctuating acidification increases variance in – but not mean of – growth rate in barnacles

Chair: Philip Munday

Jon N Havenhand (1)*, Louise Eriander (2), Anna-Lisa Wrange (1)

1 Department of Marine Sciences, University of Gothenburg, Tjärnö, 45296 Strömstad, Sweden
2 Department of Marine Sciences, University of Gothenburg, 40530 Gothenburg, Sweden

Background
Inshore marine habitats experience natural fluctuations in seawater chemistry far greater than those in the open oceans. In the shallow sublittoral, photosynthesis and respiration typically cause diurnal pH excursions of 0.2 – 0.4 units. Although these fluctuations have been known about for almost a century, their effects on organisms have been addressed only rarely.

Methods
We exposed multiple batches (different genotypes) of newly settled barnacles, Balanus improvisus, to constant “control” (pH 8.1), “acidified” (pH 7.7), or “fluctuating acidified” treatments (diurnally varying, pH 7.5 – 7.9) for 3 months. We measured multiple parameters of barnacle growth throughout the experiment, and shell mineralogy and shell hardness at the end of the experiment.

Findings
There was no effect of stable, or fluctuating, acidification on mean growth or shell mineralogy. However, variance in growth and mineralogy responses were ~20x greater in fluctuating acidification, such that some individuals grew up to 3x faster, (some also grew up to 3x more slowly), than in stable acidification. Both fluctuating and stable acidification caused significant reductions in shell hardness, and these reductions varied significantly among batches (genotypes).

Conclusions
Phenotypic variance is the raw material for natural selection. Our results show this variance can be far greater under diurnally fluctuating acidification than under stable acidification. As far as we’re aware, this has not been reported previously. We suggest that including simulated natural pH fluctuations in future experiments is likely to reveal ecologically and evolutionarily important responses that more traditional experimental designs would miss. Perhaps most importantly, these results show that rather than being a statistical headache that we might wish to transform away, unequal variances (“heteroscedasticity”) in our data can indicate important opportunities for selection and adaptation of responses to a future ocean.