Acute heat stress experiment

A standardized temperature profile was used to measure heat tolerance in corals (e.g., (Voolstra et al. 2020; Cunning et al. 2021; Marzonie et al. 2022; Evensen et al. 2023)) with minor modification in temperature profiles. The climatological maximum monthly mean (MMM) of Heron Reef is 27.3°C (Weeks et al. 2008). A pilot experiment with all three genera indicated no difference in Fv/Fm between MMM, MMM+3°C and MMM+6°C, so the latter two treatments were increased to more accurately assess the decline in performance, such that the five treatments used here included ambient, MMM, MMM+4°C, MMM+6.5°C, and MMM+9°C. Generally, experiments began at ~12:00 with a 3-hr ramp to respective treatment temperatures (23°C, 27.3°C, 31.3°C, 33.8°C, 36.3°C), a 3-hr hold, and a 1-hr ramp down to MMM (Evensen et al. 2023) (Figure S1 of Brown et al., 2024). Lights were turned off at the onset of the 1-hr ramp down to correspond with sunset. Due to experimental constraints (space, equipment, and time), only two treatments (n = 1 tank) were performed per day and each site was done in isolation. Accordingly, a complete assay took two days per site, with treatments tested each day selected randomly (Table S1, Figure S1 of Brown et al., 2024). A fragment from each coral colony was randomly placed into each treatment, so that all colonies were present in each treatment. Temperatures were controlled using an Apex controller (Neptune Systems). Apex temperature probes were calibrated against a high-precision temperature probe (HANNA HI-98190; accuracy: ±0.4°C at 25°C; resolution: ± 0.10°C) at the onset of the experiment. Temperatures were also recorded using cross-calibrated temperature loggers (HOBO UA-001-64, accuracy: ± 0.29°C at 25°C). Photosynthetically active radiation (PAR) was static and controlled using aquarium lights (NICREW HyperReef LED, Shenzhen NiCai Technology Co.), averaging 250 µmol m-2 sec-1.

Physiological responses to acute heat stress

At the end of the ramp and after 1 h of darkness (~19:00), corals were assessed for dark-adapted photochemical yield (Fv/Fm) using a Diving-PAM (Walz GmbH) 5-mm diameter fiber-optic probe at a standardized distance (5 mm) above the coral tissue after Fo stabilized. Two random spots on either side of a single fragment were measured to obtain average measures of Fv/Fm. All readings with Fo values that were less than 110 were removed to avoid any false detections (Marzonie et al., 2022). The following morning at 07:00 corals were photographed with a color standard (WDKK Waterproof Color Chart, DGK Color Tools) to assess the effect of temperature on coral color, a proxy for relative chlorophyll density and bleaching severity(Winters et al. 2009; Voolstra et al. 2020).

Image analysis of coral color, a proxy for bleaching severity

Coral color was determined from each photograph in a semi-automated manner. Each photograph was first cropped to a standard size to remove excess background via a custom automated batch script in Adobe Photoshop (Version 21.1.2). Photographs were then loaded into ImageJ (v1.53c (Schneider et al. 2012) ), and the performance of 16 built-in segmentation models were tested on a sub-sample of coral images to remove the background of the cropped image, which was turned to black, thus leaving only the coral fragment. The segmentation model that best segmented all coral fragments from the background effectively with limited coral fragment cut off (Model Li) was then implemented on all images, which were batch processed using a custom image segmentation macro script modified from(Strock 2021). Once segmented, the script then extracted red pixel intensity of the fragment in RGB, HSB, and LAB color spaces. Finally, the mean red pixel intensity of the red color standard from the original (unsegmented) images was extracted from a region of interest drawn by hand in ImageJ. Pixel intensities of the coral and corresponding red standard were then converted to a ‘darkness’ score by subtracting the red channel 'brightness’ from the maximum value (255). The mean red channel darkness of each coral was then normalized by dividing by the mean red pixel darkness of the red color standard from the same photograph. These red-normalized color values were then used to calculate the changes in bleaching severity between species, sites and treatments. For visualization, the red-normalized color values were divided by the mean color under ambient (MMM) conditions for that species.

Organism identifiers (Genus, Lifesciences Identifier [LSID]):
Pocillopora, urn:lsid:marinespecies.org:taxname:206938
Porites, urn:lsid:marinespecies.org:taxname:206485
Acropora, urn:lsid:marinespecies.org:taxname:205469