Microalgal blooms in the skeletons of bleached corals during the 2020 bleaching event on Heron Island, Australia

Endolithic microbes inhabiting the coral skeleton are being increasingly recognised as an understudied but important component of the coral metaorganism (Fordyce et al. 2019; Ricci et al. 2019; Pernice et al. 2020). They have been identified as recycling waste products from the coral animal (Risk and Muller 1983) and directly parasitising live coral tissue (Le Campion-Alsumard et al. 1995). In addition, endolithic microbes may act as ‘secondary symbionts’ during coral bleaching caused by marine heatwaves (MHWs), when the symbiosis between corals and endosymbiotic algae in their tissue breaks down (del Campo et al. 2017). During coral bleaching, blooms of microalgae in the skeleton, caused by increased light penetration through the bleached tissue, have previously been associated with the diffusion of photoassimilated carbon into the coral tissue from the skeleton (Schlichter et al. 1995; Fine and Loya 2002; Sangsawang et al. 2017). Fine et al. (2005) previously hypothesised that the tolerance of the coral Montipora monasteriata to thermal stress may be due, in part, to blooms of endolithic algae and, in general, endolithic microalgal blooms have been suggested as promoting differential coral survival during MHWs (Fine and Loya 2002; Fine et al. 2006; Carilli et al. 2010; Fordyce et al. 2019; Ricci et al. 2019).

However, endolithic algae are also prolific bioeroders of coral reef substrates (Perry et al. 2014) and can rapidly degrade coral skeletons after mortality (Leggat et al. 2019). This reduces reef structural complexity following an MHW (Leggat et al. 2019), which can hinder the capacity of a reef to provide essential ecosystem services (Graham and Nash 2013). Through facultative symbiosis during coral bleaching and rapid bioerosion after mortality, endolithic microbes have the potential to effect acute changes in community structure and function that arise from a coral bleaching event (Hughes et al. 2018). As has been shown for other components of the coral microbiome, the role of endolithic microbes is expected to vary depending on coral species, morphology and size (Williams et al. 2015; Ziegler et al. 2019; Voolstra and Ziegler 2020). However, we lack baseline data on which species or morphologies these endolithic blooms occur most frequently during natural events. This is essential in forming predictions about how coral reef communities, and the services they provide, are likely to be affected by MHWs.

In 2020, the Great Barrier Reef (GBR) experienced its third mass coral bleaching event in 5 years, driven by anomalously high seawater temperatures (Great Barrier Reef Marine Park Authority 2020). Heron Island, in the southern Capricorn and Bunker Group of the GBR, has been a coral reef research hub for more than 70 years and, in 2020, was exposed to a Category I MHW (Hobday et al. 2016; http://www.marineheatwaves.org/tracker). This MHW developed over ~3 weeks and caused widespread bleaching on the Heron Island reef flat (Ainsworth et al., in press). During the bleaching event, macroscopically visible blooms of endolithic algae were evident in the skeletons of bleached corals from multiple different genera and morphologies (Fig. 1). To determine how common these blooms were and to provide baseline data for future studies, we quantified the prevalence of macroscopic endolithic algal blooms in bleached corals at three sites on the Heron Island reef flat.

Fig. 1.  Macroscopic endolithic algal blooms inside the skeletons of thermally bleached corals on the Heron Island reef flat. (a) Sub-massive Porites cylindrica; (b) branching Acropora aspera; (c) massive Platygyra sp.; (d) a large thicket of Acropora millepora; (e) encrusting Favites sp.; (f) foliose Montipora sp.; (g) tabular Acropora hyacinthus; (h) plating A. hyacinthus; and (i) sub-massive–branching Isopora palifera.

HOBO Pendant MX temperature loggers (Onset, Cape Cod, MA, USA) were used to continuously measure water temperatures (°C) in the lagoon at 15-min intervals. The loggers were deployed at three sites on the reef flat in shaded areas of the benthos to prevent internal heating of the logger resulting from exposure to direct sunlight (Bahr et al. 2016). Loggers were deployed for 2 months from 23 January to 23 March 2020, and the collected data were validated by comparing logger measurements to those made by a calibrated Hanna HI98196 multi-meter (Hanna Instruments, Woonsocket, RI, USA). Daily average wind speed (m s^–1) data were sourced from the continuous monitoring station administered by Heron Island Research Station (http://hirs-monitor02.hirs.science.uq.edu.au/) and tidal data were sourced from the Australian Bureau of Meteorology for the Gladstone region (http://www.bom.gov.au/oceanography/projects/ntc/qld_tide_tables.shtml). In situ data were converted to experimental degree heating weeks (eDHWs), a measure of accumulated heat stress. This metric is adapted from the formula used to calculate degree heating weeks (DHWs) from satellite-derived sea surface temperature (SST) data (Liu et al. 2014), the key difference being that the eDHW uses daily mean in situ temperatures instead of night-time remotely sensed SST. This means the eDHWs are inclusive of daytime peak temperature stress, which can induce physiological stress responses in a matter of hours (Ruiz-Jones and Palumbi 2017; Traylor-Knowles et al. 2017), whereas DHWs use night-time SSTs (recorded at 0900 and 1500 hours) as a proxy for heat accumulation throughout the water column. Heat accumulation measured by eDHWs is relative to 1°C above the maximum of the monthly means (i.e. MMM + 1 = 28.3°C; Liu et al. 2014).

To assess the prevalence of macroscopic endolithic algal blooms, three sites were surveyed within the Heron Island lagoon (Fig. 2): Site 1, the northern side of the reef rim with relatively high coral cover (40–60%; Joyce et al. 2013) and diversity; Site 2, the southern rim of the lagoon adjacent to the jetty, with high coral cover (40–60%; Joyce et al. 2013) of predominantly Acropora spp.; and Site 3, an area of shallow lagoon with lower coral cover (0–20%; Joyce et al. 2013), ~500 m south-east of Site 2. Surveys were first completed on 21 February 2020 (Site 2), 26 February 2020 (Site 3) and 28 February 2020 (Site 1), and then again on 21 March 2020 (all three sites). Quadrats were photographed (Olympus Tough TG-6 camera; Olympus, Tokyo, Japan) along six 100-m transects at 10-m intervals including at 0 and 100 m (n = 66 quadrats per site). For Site 1, transects extended from east-north-east to west-south-west along the reef rim and were conducted at high tide; using the transect tape as a reference object, the image was cropped to 1 m² using ImageJ (ver. 1.52s, National Institute of Health in the United States, see https://imagej.nih.gov/ij/index.html; Abràmoff et al. 2004). For Site 2, transects went from north-east to south-west towards the reef crest and were conducted by reef walking at low tide. For Site 3, transects were parallel to the shoreline (east-south-east to west-north-west) within the shallow lagoon and were surveyed at low tide. For Sites 2 and 3, a physical quadrat was used to demarcate a 1-m² area before taking a photograph.

Fig. 2.  Prevalence of macroscopic endolithic blooms at three sites and two time points on the Heron Island reef flat (left). The total number of surveyed bleached coral patches (n) is shown for each site × time point, with the percentage of surveyed corals that showed a partial greening of the their skeleton also provided. Stacked bar charts represent the proportion of all endolithic algal blooms seen in each coral genera.

We removed quadrats that had no hard coral, had hard coral but none of the patches or colonies were identifiable down to genus level and those that had identifiable hard corals but none showed signs of bleaching. As a result, only quadrats that had at least one bleached coral patch or colony identifiable to genus level were used to quantify the prevalence of endolithic algal blooms. After selecting this subset, each image was first assessed for the number and genera of bleached coral ‘patches’. The term ‘patches’ was used because it was not always possible to identify the boundaries for each coral colony, especially in larger Acropora thickets (e.g. Fig. 1f, h). ‘Patches’ therefore encompasses whole colonies, as well as partial colonies that were cropped by the quadrat and groups of colonies. The absolute number of bleached coral patches assessed (i.e. the site-specific sample size) varied between sites and across time (Fig. 2). In February, 202, 117 and 75 bleached patches were examined at sites 1, 2 and 3 respectively; in March, the samples sizes were 67, 45 and 58 at Sites 1, 2 and 3 respectively (Fig. 2). The lower samples sizes in March are probably a result of coral mortality between February and March, because dead corals that were overgrown by epilithic algae were excluded from our analysis. Coral recovery between the two surveys could have produced the same effect, but no significant coral recovery was recorded (Ainsworth et al., in press).

The criteria for identifying macroscopic endolithic algal blooms were a partial greening of the coral skeleton combined with an absence or very low abundance of epilithic overgrowth. This decision was made to reduce ambiguity in identifying endolithic blooms that occurred beneath epilithic overgrowth and to avoid overestimating total endolithic blooms. The prevalence of endolithic algal blooms, as assessed here, was expressed as the percentage of bleached corals in which a bloom was macroscopically visible (for each site, prevalence was calculated as the sum of all bleached coral patches with blooms divided by the sum of all bleached coral patches).

The daily mean water temperature in the reef lagoon of Heron Island at the onset of bleaching conditions in late January was 29.1°C (daytime maximum = 34.7°C; Fig. 3a), during neap tides and while wind speeds were <4 m s^–1 (Fig. 3b). The bulk of the heat stress accumulated in February when eDHWs increased from 0.59  on 31 January to 8.66  at the end of February. Daily mean and daytime maximum temperatures did not exceed 29 and 34°C respectively, except for a 6-day long extreme temperature pulse during which reef water reached daytime maxima exceeding 35°C. Similarly, acute, extreme heat pulses have been observed previously in this reef lagoon (MacKellar and McGowan 2010). Daily mean temperatures continued to exceed the MMM + 1 threshold (Fig. 3a) up until 10 March, after which temperatures decreased in association with increased wind speed (Fig. 3); eDHW accumulation reached a maximum of 9.78 DHW.

Fig. 3.  (a) In situ temperatures recorded from HOBO data loggers placed at three locations on the southern reef flat (blue line). Daily mean temperatures (black dashed line) were used to estimate accumulated heat stress expressed as experimental degree heating weeks (eDHWs; red line); daily averages exceeding the climatological baseline (MMM + 1°C; horizontal pink dashed line) are included in the eDHW calculation. Symbols represent sampling time points for each survey site: closed triangles, Site 2; crosses, Site 3; closed circles, Site 1. (b) Daily mean wind speed (black line), relative to a threshold for doldrum-like wind speeds (4 m s^–1; black dashed line), overlayed with changes in tidal height for Heron Reef (purple lines).

In February, macroscopic endolithic algal blooms were visible in 22.3, 29.1 and 20.0% of bleached corals at Sites 1, 2 and 3 respectively; corals from 9 genera experienced endolithic blooms, including encrusting, tabular, massive, submassive, branching and columnar morphologies (Fig. 1). Site 1 had the highest diversity of blooming corals (seven genera) and most were seen in either branching Acropora spp. (37.8%) or columnar–encrusting Isopora spp. (44.4%). Blooms were recorded in six and four genera at Sites 2 and 3 respectively; at both sites blooms were predominantly recorded in branching Acropora spp. corals (Site 2, 65.7%; Site 3, 66.7%). Surveys conducted again in late March showed a reduction in the proportion of bleached corals with endolithic algal blooms at Site 1 from 22.3 to 4.5%. Site 3 also showed a reduction in endolithic bloom prevalence from 20.0 to 15.5% in March; by contrast, the proportion at Site 2 increased from 29.1 to 42.2%. Across all three sites, eight coral genera experienced coral bleaching, but all the blooms were seen in Acropora spp., except for a single Isopora coral patch at each of the Sites 2 and 3. It is unclear why Site 2 alone had an increased prevalence of microalgal blooms in March. This may reflect the fact that 82% of the bleached corals surveyed at Site 2 were Acropora spp., compared with 70 and 72% for Sites 1 and 3 respectively. In addition, Site 2 may have had a different composition of Acropora species. This genus is highly variable in its susceptibility to thermal stress (Swain et al. 2016), and species with higher tolerance are expected to bleach later during an MHW. Therefore, the difference may be due to a delayed bleaching response by tolerant acroporids at this site. Identifying Acropora species accurately was beyond the scope of the present study, but the effect of coral thermal tolerance on endolithic microalgal blooms is an important and unanswered question.

These observations highlight that macroscopic endolithic blooms occur in a wide range of coral species and morphologies, including primary reef-building acroporids. At this location, branching Acropora corals are the major contributors to reef structural complexity (Bryson et al. 2017) and in our study at least 20% of them experienced endolithic algal blooms. This highlights the potential for endolithic blooms to affect coral reef structure acutely and significantly, by degrading the skeletons of primary reef builders. The macroscopic endolithic blooms were often observed adjacent to the coral–substrate interface (Fig. 1a, b, d, g, h, i), which suggests that the migration of microalgae from neighbouring substrates into the bleached skeletons may be contributing to the formation of blooms inside the skeletons. Diaz-Pulido and McCook (2002) have previously recorded rapid skeletal colonisation (<4 weeks) by Ostreobium spp. in severely bleached corals. Furthermore, the benthic organism adjacent to corals is likely to be an important factor; phototrophic endolithic communities can be both more abundant and diverse when living beneath turf algae compared with beneath Orbicella sp. corals (Gutiérrez-Isaza et al. 2015). Together, this suggests that a coral’s position within the benthos is a factor in determining the extent of endolithic algal blooms.

Here we have presented observations and data that demonstrate that macroscopic blooms of endolithic algae are widespread across coral taxa and morphologies bleached by an MHW. To the best of our knowledge, this is the first data to be presented on the frequency of these blooms during a coral bleaching event. Importantly, endolithic algae were found to be most common in primary reef-building corals that are not typically associated with having a high biomass of endolithic algae (Fordyce et al. 2021). Our observations suggest that bleached coral skeletons may be partly colonised by microalgae migrating from neighbouring substrates, which may explain this unexpected result. Whether such blooms are a regular feature of coral bleaching events or specific to this case study is unclear, but their prevalence here highlights an ongoing need to understand the roles that these microbes play in coral community responses to environmental stress.

The authors declare that they have no conflicts of interest.

This work was funded by an Australian Research Council Discovery Project grant (DP180103199) awarded to T. D. Ainsworth and W. Leggat.