Does tolerance to fast warming predict tolerance to slow warming in fishes?

Emily Lechner, Erin Stewart, Alex Berry, Christian J Bihun, Sandra Ann Binning, Tamzin Blewett, Ivan Cadonic, Timothy D Clark, Zara-Louise Cowan, Jeremy De Bonville, Rasmus Ern, Maryane Gradito, Leon Green, Elizabeth Hoots, Luis Kuchenmüller, Vincent Mélançon, Moa Metz, Leon Pfeufer, Patrice Pottier, Graham D Raby, Dominique Roche, Josefin Sundin , Elena Tamarit-Castro, Mafalda Tomás, Juliane Vigneault, Zacharia Zouhair, Eirik Åsheim, Fredrik Jutfelt

1. Critical thermal maximum (CTmax) is widely used to quantify upper thermal limits in fishes, but its ecological relevance has been questioned because standard protocols use fast warming rates (0.3°C min-¹; 18°C h-¹) that rarely occur in nature, where thermal stress more typically unfolds over hours to days. Whether CTmax under fast warming predicts warming tolerance to slower, ecologically realistic warming remains a critical question for climate risk assessments and mechanistic species distribution models.
2. We measured thermal tolerance under fast (0.3°C min-¹) and slow (1°C h-¹) warming in the same individuals across seven temperate fish species. This paired design allowed us to estimate individual-level correlations between fast and slow warming tolerance and assess how this relationship varied among species.
3. We observed substantial interspecific variation in the direction of the fast–slow warming tolerance relationship. Slow CTmax was higher than fast CTmax in three-spined stickleback, lesser pipefish, corkwing wrasse, and European eel, while the opposite was true for European plaice, bluegill sunfish, and Atlantic cod.
4. Fast and slow CTmax were only weakly correlated at the individual level, and significant correlations were only detected in two of seven species. However, at the species level, fast and slow CTmax was significantly correlated (R² marginal = 0.873; between-species slope = 1.157).
5. Standard fast-warming CTmax does not consistently predict individual thermal tolerance under slower, more ecologically realistic warming rates and its usage may overlook biologically meaningful species-specific differences in thermal vulnerability. Fast-warming CTmax should be interpreted with caution in the context of ecological forecasting and climate risk assessments as it can over-estimate upper thermal tolerance and should be used at the appropriate biological scale.