UABDivulga - Barcelona recerca i innovació

Researches concerned with the ability of organisms to cope with rising temperatures routinely estimate their critical upper thermal limits placing individuals under acute constant stressful temperatures or subjecting them to more realistic conditions where temperature is gradually increased until individuals are knocked down (so-called ramping assays). The problem with ramping protocols is that results depend on the rate of warming, with higher heat tolerances usually observed when fast warming rates are used in experiments.

This has raised some concerns because natural warming rates are typically slow, and several authors claim that upper thermal tolerance limits may have been overestimated. An additional experimental observation is that the amount of heritable genetic variation for heat tolerance dramatically drops when “ecologically realistic” assays with slow ramping rates are performed. This has lead to the conclusion that natural populations exhibit low adaptive potential for upper thermal limits under relevant rates of temperature change.

However, the real problem is that the parameter CTmax researchers want to estimate is not constant but changes during a thermotolerance assay according to the experimental conditions. We have developed a physiological model that can accurately replicate the empirical observations. The model clearly shows that there are various confounding effects when estimating heat tolerance and that many experimental protocols, particularly those that use “ecologically realistic” slow ramping rates, produce highly unreliable underestimates.

In our simulations to artificially select for increasing heat tolerance we have combined the physiological model with genetic models for quantitative traits. The results show that the evolutionary response of the underlying “real” upper thermal limit (CTmax) is independent of the methodology used in the experiments. However, what researchers estimate and wrongly call “upper thermal limit” does not seem to appreciably increase after selection when slow ramping rates are used.

Our theoretical approach makes it clear that the apparent contradiction stems from a conceptual confusion between what we actually measure (“estimate”) and what we would like to measure (“parameter”). The problem gets worse and worse as warming rates slow down and approach to “ecologically realistic” scenarios. The take home message is clear: when dealing with physiological measurements as organisms’ tolerance to high temperatures, do not blindly trust what you see. Our results are good news because adaptive genetic responses for increasing upper thermal limits under the current scenario of global warming may be higher than acknowledged in recent studies.