Chronic heat tolerance reveals overestimated thermal safety margins and increased vulnerability in marine fish populations

Predicting vulnerability to global warming remains an elusive goal in thermal biology. In marine fishes, ongoing changes in distribution contrast with their apparent capacity to tolerate temperatures from 5 ºC up to 25 ºC higher than current conditions. Employing a dataset of 786 upper critical temperatures across 213 species and recent theoretical developments, we provide conclusive evidence that these so-called ‘thermal safety margins’ overestimate the resilience to warming and that most species inhabit thermal conditions approaching their physiological tolerance limit. This result holds across latitudes and based on historical records, several populations have encountered stressful temperatures in the recent past. While warming tolerance remains similar across geographic regions, behavioral responses are constrained at low latitudes as distribution shifts required to encounter cooler waters are disproportionally higher in the tropics. Overall, our results illustrate how thermotolerance measures can be extrapolated to the field and used to quantify vulnerability to warming.

Methods
We compiled heat tolerance data for marine fish species started by retrieving information in previous published reviews on the subject and combined the original sources from these reviews with more recent published references found with a Boolean search in ISI Web of Science. We used Engauge Digitizer software (http://markummitchell.github.io/engauge-digitizer) to retrieve data from published figures when the heat tolerance values were not available in the main text or as supplementary information. From the original references, we also compiled the geographic location where the populations were collected and details on the experimental protocol described in each study to estimate the duration of each assay. With geographical location, we estimated the maximum and minimum monthly temperatures experienced by each population/species from historical data grids of sea surface temperatures between 1891 and 2020 in the COBE SST dataset provided by the NOAA (https://www.psl.noaa.gov/data/gridded/data.cobe.html). We also estimate the maximum annual monthly temperature variation across all ocean. We also use the data base of Coutant (1972) for fish measured in static assays to standardize the values ​​of tolerance at comparable temporal windows (see methods of the main text).

Coutant, C. C. (1972). Water quality criteria. A report of the committee on water quality criteria. text and Appendix II-C, 410-419. Available at
https://www3.epa.gov/region1/npdes/merrimackstation/pdfs/ar/AR-166.pdf 
(accessed on 23 January 2024).