Testing the predictability of morphological evolution in contrasting thermal environments

Gaining the ability to predict population responses to climate change is a
pressing concern. Using a ‘natural experiment’, we show that testing for
divergent evolution in wild populations from contrasting thermal
environments provides a powerful approach, and likely an enhanced
predictive power for responses to climate change. Specifically, we used a
unique study system in Iceland, where freshwater populations of threespine
sticklebacks (Gasterosteus aculeatus) are found in waters warmed by
geothermal activity, adjacent to populations in ambient-temperature water.
We focused on morphological traits across six pairs from warm and cold
habitats. We found that fish from warm habitats tended to have a deeper
mid-body, a sub-terminally orientated jaw, steeper craniofacial profile,
and deeper caudal region relative to fish from cold habitats. Our common
garden experiment showed that most of these differences were heritable.
Population age did not appear to influence the magnitude or type of
thermal divergence, but similar types of divergence between thermal
habitats were more prevalent across allopatric than sympatric population
pairs. These findings suggest that morphological divergence in response to
thermal habitat, despite being relatively complex and multivariate, are
predictable to a degree. Our data also suggests that the potential for
migration of individuals between different thermal habitats may enhance
non-parallel evolution and reduce our ability to predict responses to
climate change.