The SIMP (30 °S, 153 °E, Figure 1), located on the east coast of northern NSW and LHIMP (31.5 °S, 159 °E, Figure 1), which lies approximately 600 km east of mainland Australia, have previously been recognized for their high scleractinian coral cover [23-26]. Criteria for site selection in this study included similar depth gradients and relatively homogenous cover of benthic assemblages with a clear dominance of scleractinian corals. Four island associated reefs (10–15 m depth) within the SIMP and four sheltered (2–5 m depth) and two exposed (10–15 m depth) sites at LHIMP, were randomly selected within areas of high coral cover to monitor coral health and bleaching response between 2005 and 2007 (refer to [27] for additional site descriptions).

Coral health and bleaching response assessments were completed in conjunction with Australian subtropical white syndrome (ASWS) evaluations within the SIMP between 2005 and 2007 and at LHIMP in 2005 [27]. Coral health surveys were conducted adjacent to four islands within the SIMP (Figure 1a) during March each year, which corresponded to the period of warmest seawater temperatures. Sampling at LHIMP was completed at four lagoon and two exposed sites (Figure 1b) during May, when seawater temperature ranged between 21–22 °C.

Five 20-metre fiberglass tapes were randomly placed on the benthos within 30 × 30 m permanent sites, (GIS referenced to enable relocation) and corals larger than 50 mm occurring within one meter either side of the tape were evaluated for signs of recent stress, which included stress caused by biotic (disease and predation) and abiotic (sediment abrasion and thermal bleaching) factors. Individual corals were recorded as bleached according to pigmentation color characteristics. Prior to assessment dives, all coral species observed at each study site were compared to the Coral Health Monitoring Chart (CHMC; 28) and normal pigment color was calibrated with the color chart codes. These codes were used as a reference during all subsequent coral health surveys. If a colony displayed white, patchy white or pale coloration, relative to normal pigmentation of that coral species at that location, then the individual coral was recorded as either moderately (1–50% tissue bleaching) or severely (51–100% tissue bleaching) bleached, according to either the percentage of colony discoloration and/or the level of pigment paling (consistent with [28,29]).

Hastings Tidbit Stowaway temperature loggers were deployed at North Solitary and South West Solitary Islands within the SIMP at a depth of 10 m. These loggers recorded in situ temperature data every 30 min between 2000 and 2006. These loggers were replaced with Dataflow Odyssey loggers in early 2006. The loggers were removed annually and temperature data downloaded before being re-calibrated and replaced at each site (see [30] for further description). During this period similar devices were not deployed at LHIMP; therefore as a proxy for in situ temperature, satellite derived sea surface temperature (SST) data were acquired from the National Oceanic Atmospheric Administration (NOAA) and descriptive statistics and Maximum Monthly Mean (MMM) climatology determined between 2001 and 2007 for LHIMP. NOAA reef watch program utilize bleaching thresholds 1 °C above MMM to predict bleaching events [31].

A total of 83,951 individual coral colonies were counted at the eight sites investigated between 2005–2007, of which 2,950 and 138 colonies were recorded as moderately and severely bleached, respectively. Overall, the proportion of total bleached coral colonies (moderate and severe categories combined) within each location varied between 4 to 16% during all summer surveys. The highest proportion of colonies noted as moderately bleached was recorded during summer 2005 (11% pooled from all sites); however severe bleaching was greatest during summer 2006 (<1% pooled from all sites). Total bleaching response was highest at Anemone Bay (NSI) during summer 2005, where 22% and 1% of coral colonies were observed with moderate and severe bleaching response, respectively. At the other sites, moderate and severe bleaching ranged between 6 and <1% (SSI-Coral Corner) and 12 and 1% (NSI-Trail Mooring), respectively. Total bleaching response during summer was highest at NSI compared to the more southern nearshore island reefs (Figure 2).

Five coral families (Acroporidae, Dendrophylliidae, Faviidae, Pocilloporidae and Poritidae) were observed with variable levels of pigmentation loss during this study. Common coral species from the families Agariciidae, Mussidae and Siderastreidae, showed no loss of pigmentation during this time. The mean proportion of total bleaching response was variable at the family, year and location levels (Figure 3). Pocilloporidae spp., notably Pocillopora damicornis and Stylophora pistillata, were frequently observed with tissue paling, particularly at the branch tips during summer surveys at all island sites. Total bleaching response (moderate and severe combined) was highest in this taxon during the 2005 summer surveys, during which time total bleaching ranged between 22.5 ± 3.1% (NSI) and 34.6 ± 5.8% (NWSI).

During this study, the bleaching susceptibility of dominant corals families (e.g., Dendrophylliidae, Faviidae, Pocilloporidae and Poritidae) within the SIMP was significantly higher than rarer families (e.g., Agariciidae, Mussidae and Siderastreidae) (see also [34]). This suggests that variation in bleaching susceptibility between locations may be associated with temperature and light history as well as ecological factors. This hypothesis is in accordance with Marshall and Baird [29] who indicated that environmental factors (e.g., temperature and light intensity) may explain large-scale patterns of bleaching occurrence, while small-scale patterns were determined by assemblage composition.

Poritidae colonies were regularly observed with decreased tissue color at all locations during summer observations, particularly at NSI, where bleaching reached 44.2 ± 8.2% in 2005 (Figure 3). This negative response to high summer seawater temperature may explain why Porites heronensis is common in subtropical localities but rare in the tropics [35]. However, other abiotic and biotic factors need to be considered when explaining distributions patterns. Massive Porites colonies have previously shown moderate bleaching susceptibility to thermal stress [29], while branching poritids displayed significant bleaching response and associated mortality on Kenyan reefs during the 1998 mass bleaching event [28].

A total of 7,042 individual colonies from eleven scleractinian families were recorded at the six sites evaluated, of which 396 (5.74%) were observed with varying degrees of pigmentation loss. The mean proportion of moderate bleaching was an order of magnitude higher than the mean proportion for the severe bleaching category. The total proportion of bleached colonies ranged between 4.68% at Comet's Hole and 6.36% at Erscott's Hole. Overall, five hard coral families (Acroporidae, Dendrophylliidae, Faviidae, Pocilloporidae, and Poritidae) were observed with reduced pigmentation (Figure 4). Bleaching was present in acroporids at all sites and ranged between 0.61 and 15.13%. Poritids and pocilloporids were observed with varying degrees of pigmentation loss at all sites. The highest bleaching response was within the pocilloporids at the exposed Malabar Reef, where 12.77 and 3.22% of this population were moderately and severely bleached, respectively.

In situ seawater temperature recorded between 2001 and 2008 within the SIMP tended to be highly variable over small temporal and spatial scales (Figure 5). Seawater temperature data at the offshore island (NSI) was generally 1–1.5 °C warmer than the midshelf island (SWSI), and this pattern was consistent throughout all seasons and years (Figure 5). Average daily maximum temperature was highest at NSI and ranged between 24.99 and 27.35 °C (Table 1). Maximum daily temperature was lowest during all years at SWSI, ranging between 24.32 and 25.94 °C (Table 1).

Weekly temperature fluctuations of up to 5 °C occurred consistently at the midshelf island reefs (SWSI and SSI), with an 8.4 °C differential observed at SWSI during a one-week period (Figure 5; 31 January 2008–7 February 2008). Daily temperature changes up to 4 °C day⁻¹ were common at all locations during summer months. Summer seawater temperatures were highest during 2006 at all locations, and average daily temperature exceeded 27 °C on three occasions at NSI. However, the number of days seawater temperature exceeded 25 °C was greatest during 2002, where 52 and 33 days exceeded this measure at NSI and SWSI, respectively (Table 1). Maximum monthly mean temperatures occurred during February and March during most years at all island locations (Table 1). At NSI this measure was highest in 2006, but was highest during February 2002 at the midshelf islands.

Scleractinian corals within the SIMP are exposed to a varied temperature regime throughout the year, with temperatures ranging between 16 and 28 °C. This high level of variation may generate evolutionary pressures to develop resilience mechanisms for the coral holobiont (zooxanthellae, bacteria, viruses and fungus associated with the coral host), enabling them to adjust to changing sea temperature within a short period of time. In contrast to coral reefs environments such as the GBR, seawater temperature varies considerably at both small and large time scales along the NSW coastline. Therefore, hard corals within the SIMP may require short-term physiological mechanisms that help to prevent thermal stress and damage. Homeostatic processes such as: increased production of mucus polysaccharides which contain microsporine amino acids (MAA) that absorb UVR; endosymbiont reshuffling to more heat or cool tolerant clades if multiple symbionts are associated with subtropical coral species [45]; and the production of fluorescence pigments [46], provide acclimatization and adaptive mechanisms that increase the ability of the holobiont to combat stress caused by temperature variations.

Thermal bleaching threshold values reported in the literature [46] displayed a distinct inverse trend with latitude (Figure 6). Results from the second-order polynomial analysis indicated a significant relationship (F_2,7 = 175.70, p < 0.0001) between bleaching thresholds and latitude; the model accounted for a high proportion of the variance in the data set (Figure 6, r² = 0.99). Applying the equation generated by the regression procedure indicated that the hypothetical thermal bleaching threshold for corals located at 30 °S (SIMP) and 31.5 °S (LHIMP) are 26.8 and 26.5 °C, respectively. The hypothetical bleaching threshold for the SIMP has previously been exceeded in 2002 and 2006, where average in situ daily seawater temperature at NSI exceeded this measure for a period of 3 and 6 days, respectively. However, this sea temperature threshold has not been exceeded at 10 m depth at the midshelf island (SWSI; Table 1).

By modeling the seawater temperature data and bleaching threshold indices from other reef locations, a prediction can be made that the summer bleaching threshold within the SIMP and LHIMP are 26.8 °C and 26.5 °C, respectively. However, as indicated above, this measure does not take into consideration seasonal differences in coral susceptibility to temperature-induced bleaching, as well as other variables such as light intensity that are known to cause bleaching in isolation or in combination with other parameters. It is hypothesized that thermally induced bleaching would occur throughout eastern Australia coral populations during periods when seawater temperature exceeds the predicted threshold for a period greater than four weeks (sensu [31,47]). This however, would only occur when the EAC exposes reef communities to higher than normal summer seawater temperatures for extended periods, with minimal mixing of cooler water.

Manzello et al. [48] tested the effectiveness of nine thermal indices to predict bleaching response at several locations along the Florida Keys Tract. They concluded that maximum monthly SST and the number of days above the predicted thermal threshold was the most accurate predictor of past bleaching patterns. They also reported similar trends when regressing bleaching curves with latitude and seasonal maximum SST. Other factors however, such as local small-scale temperature variations, differences in local seasonal climatic conditions, temperature history [48] and turbidity [49], also require interpretation and inclusion into the aforementioned bleaching model. Nevertheless, the high latitude bleaching model presented in this study provides a useful framework from which to advance studies on bleaching response at finer spatial scales, particularly in the context of local oceanographic conditions.

Environmental factors such as upwelling, strong currents, cloud cover, and light attenuation with increasing depth, have been reported to reduce the impact of thermal stress [32,50,51]. These extrinsic factors may be acting independently or synergistically to improve the resistance and resilience of eastern Australian subtropical coral communities to recent episodes of stress. This hypothesis is supported by Reigl and Piller's [50] study, which found that at two subtropical regions (Bahamas and South Africa), upwelled water provided a mechanism that minimized the effects of thermal stress on coral communities. They also proposed that deeper reefs at these locations were refuges for coral recruitment following bleaching events.

Localized upwelling and deep reefs are common features of benthic habitats along the east coast of northern NSW, particularly throughout the SIMP. Recent swath mapping of seabed habitats has identified reefs extending from mainland Australia to the continental shelf, and hard corals are a dominant component of midshelf exposed reefs between depths of 10–25 m at 30 °S [49]. Small-scale localized upwelling, which are a common feature along the eastern Australian coastline [52], enable vertical mixing of cooler water with warmer equatorial EAC driven waters. This study therefore supports the hypothesis that high latitude marginal reefs found adjacent to deeper reefs, where vertical mixing between cool deeper water and shallow heated waters occurs, may provide some relief from future predicted climate change driven bleaching events [53,54]. However, upwelled seawater may also be enriched in CO₂, which can exacerbate localized ocean acidification [55].

Other factors, such as limited genetic diversity in subtropical coral populations [37,56,57], reduced local reproductive capacity [58-61], sporadic long-distance recruitment [62], ocean acidification [55,63-65], and poor water quality [66] may also reduce the ability of subtropical coral communities to successfully recover from chronic or catastrophic disturbance that may result in a significant loss of coral diversity.

In mid December 2009, seawater temperatures within the LHIMP lagoon began to rise above the predicted thermal threshold level (26.5 °C; Figure 7). A bleaching warning was issued through the NOAA Coral Reef Watch Satellite Bleaching Alert System. During January 2010, a reduction in zooxanthellae pigmentation was observed in many coral species within the lagoon. Quantitative bleaching assessments in March 2010 showed that bleaching ranged between 50–98% within the coral assemblages at five sites within the lagoon [67]. As the bleaching model predicted, Pocilloporid corals (Pocillopora, Stylophora and Seriatopora) and Montipora spp. severely bleached with some bleaching associated mortality observed at several sites. Additionally, Isopora cuneata and Porites spp. displayed variable pigmentation loss [67]. Data collected during the early stages of the 2010 LHI thermal anomaly supports the predictions of the subtropical bleaching model. Further research will investigate the capacity of marginal coral communities to acclimate to future predicted thermal stress events, by quantifying the ability of bleached coral species to shuffle to more thermally tolerant zooxanthellae clades.

The high latitude bleaching model presented in this study predicts that at eastern Australian subtropical locations, some scleractinian coral genera from the families Acroporidae (e.g., Isopora and Montipora), Dendrophylliidae (e.g., Turbinaria), Pocilloporidae (e.g., Pocillopora and Stylophora) and Poritidae (e.g., Porites) will be more susceptible to thermal stress than coral genera from the families Faviidae (e.g., Goniastrea and Favites), Mussidae (e.g., Acanthastrea) and Siderastreidae (e.g., Psammocora). In light of predicted increases in seawater temperature (see [2]) we hypothesize that seawater temperatures that exceed 26.5 °C and 26.8 °C within the LHIMP and SIMP, respectively, for a period greater than four weeks (see [31,47]), will result in extensive coral bleaching, with associated mortality determined by the duration of the thermal anomaly and the intensity of insolation. It is predicted that future repeated bleaching events may lead to a decline in live hard coral cover of dominant coral species that have limited northern distributions, with a subsequent increase in the relative abundance of other rarer coral taxa. Coral species diversity at these locations may not increase; rather a shift in the coral assemblage is predicted to occur through time [69].