Optimal thermal conditions for corals extend poleward with oceanic warming

The capacity for poleward range expansions beyond the tropics in corals hinges on ecophysiological constraints and resulting responses to climatic variability. We aimed to determine how future warming will affect coral habitat suitability at the poleward range edges of these foundational species in the Northwest Pacific.

of tropical species unable to tolerate more frequent and extreme warming events and the poleward expansion of species into subtropical and temperate ranges as historical cold barriers are removed.In oceanic systems the distributional processes are overtly clear.High larval connectivity within marine systems allows for relatively rapid expansion into newly available (Nakabayashi et al., 2019) or recently disturbed (Edmunds et al., 2018) habitats so that ranges tend to expand to the edge of both upper and lower thermal limits (Sunday et al., 2012).This tendency for marine species to fill their entire thermal niche makes thermal suitability-based distribution modelling approaches particularly useful.
One of the most pronounced disturbances to marine ecosystems is the ongoing collapse of tropical coral reefs.Since the 1990's, rising sea surface temperatures have driven massive losses of coral cover with changes in community assembly and function that threaten these exceptionally high-value ecosystems (Hughes et al., 2018;Sully et al., 2022).Meanwhile, observations of corals expanding into subtropical and temperate environments have been reported worldwide (Baird et al., 2012;Precht & Aronson, 2004), including being well documented in the North Pacific (Nakabayashi et al., 2019;Yamano et al., 2011).Distribution changes for corals and other habitat forming, or foundational species, have an outsized role in determining global patterns of biodiversity making the forecasting of these species' distributions a critical component of sustainable longterm management strategies.
Thermal physiology is intrinsically linked to both survival (i.e.persistence) and performance (i.e.proliferation) within a particular environment.Relative to tropical species, subtropical populations must adapt their physiology not only to lower annual mean temperatures but also wider temperature variation (i.e.seasonality) and more intense temperature extremes.The physiological suitability of a species in a new environment will be dependent on both the acclimative capacity of individuals as well as the adaptive capacity of populations.The majority of corals are adapted to tropical environments where the selection pressure of frequent thermal stress is now embedded into their genome (Dixon et al., 2015;Torda et al., 2017).Conversely, corals from marginal environments, near the edges of coral ranges, have for centuries been shaped by strong seasonality and exemplify the physiological traits and adaptive capacity necessary for range expansions under climate change.Indeed, studies on subtropical corals make clear how drastically physiology can vary across seasons (Jurriaans & Hoogenboom, 2020) and latitudes (Jurriaans & Hoogenboom, 2019;Silbiger et al., 2019).In this regard, there has been a growing interest in how various thermal environments shape coral physiological performance (Aichelman et al., 2019;Gould et al., 2021;Jurriaans et al., 2021;McIlroy et al., 2019;Silbiger et al., 2019).In their mechanistic applicability, data from marginal reefs are ideal for estimating future spatial and temporal patterns under environmental change at the global scale, specifically at their poleward limits.
Recent predictions of coral range shifts under climate change have been conducted through correlative species distribution models (e.g.Adam et al., 2021;Doherty et al., 2021;Larkin et al., 2021;Yamakita et al., 2022).While correlative approaches deserve merit for including comprehensive analyses of the environmental factors driving habitat suitability (Sully et al., 2022), they lack a consideration for organismal physiological performance and adaptation (Kearney et al., 2010;Kearney & Porter, 2009;Peterson et al., 2016).
Species distribution models that adopt a mechanistic approach can therefore fill this gap and provide an estimate of the proportion of time an organism lives in conditions optimal for maintaining physiological function (Cavanaugh et al., 2015;Huang et al., 2020;Kearney et al., 2010;Kearney & Porter, 2009;Martínez et al., 2015).
In ectothermic species, thermal physiology is a relatively simple yet informative trait to project ecophysiological constraints, as well as potential shifts, under climate change because of its direct link to environmental temperature (Angilletta, 2009;Kearney & Porter, 2009;McIlroy et al., 2019;Perotti et al., 2018).From this perspective, model outputs yield estimates of thermal habitat suitability; a trait fundamental to the persistence and proliferation of marine ectotherms (Sunday et al., 2012).Despite the importance of physiology in dictating environmental suitability, thermal performance-based distribution modelling has yet to be comprehensively applied to corals.
Thermal performance approaches are used to describe the way that organismal performance changes with temperature and have been well utilised in defining the physiological strategies and limitations that support coral persistence within marginal habitats (Jurriaans et al., 2021;McIlroy et al., 2019;Silbiger et al., 2019).These data can be applied to infer how an organism's fitness may change as a consequence of changing thermal regimes and parameterise models of thermal suitability under various contexts including past, present and future environmental conditions.For example, the temperature range at which an organism maintains at least 80% of its peak performance, coined 'B 80 ' (Hertz et al., 1993), is a relatively simple yet ecologically relevant way to represent the temperature range within which an organism's performance is optimised (Landry Yuan et al., 2016;McIlroy et al., 2019;Perotti et al., 2018).Considering the influence of acclimation and adaptation on coral physiology, appropriate predictions of thermal habitat suitability shifts require long term data sets extending across seasons, natural temperature ramping rates (as opposed to acute stress) and coral populations that are previously adapted or acclimated to seasonal variation.A study of five species across four families of corals in Hong Kong assessed the thermal physiology of individuals across a year in which natural environmental temperatures ranged from 14 to 31°C (McIlroy et al., 2019).This thermal regime well reflects subtropical areas of ongoing coral range expansions or local emergence, particularly, the physiological challenges associated with both lower and more seasonally variable temperature regimes.That study revealed that, within this highly seasonal environment, a few consistent physiological patterns emerged across species.Using photosynthesis to respiration ratio values (P:R) as a proxy for physiological limits, (1) B 80 was maintained for approximately half of the days annually, and (2) a minimum P:R value of ~2.0, previously described as the minimum value to maintain functional symbiosis between corals and the photosynthetic algae that sustain them (Coles & Jokiel, 1977), was estimated just below the experienced annual temperature minimum.
While more precise metrics are necessary for quantifying nutritional energy budgets of coral symbioses and their response to stress (Grottoli et al., 2006;Muscatine et al., 1981), the ubiquity of these traits across species, measured within populations adapted to relevant temperature regimes, underlies their application as the main physiological metrics used herein.
In this mechanistic modelling approach, we applied the previously defined limits of thermal physiology derived from Hong Kong coral populations as a proxy for populations at the latitudinal range limits.Given the evolutionary and ecological context of Hong Kong corals inhabiting highly seasonal habitats, and a lack of comparative physiological data across these species' ranges, these populations of marginal corals are used to represent the limits of coral adaptation/ acclimation capacity appropriate for understanding high-latitude range edges specifically.Yet, these populations are poorly suited for the estimation of species' suitability within lower latitudes.By representing the maximum seasonal capability for these five species, we estimate the geographical extent of their thermal suitability within the Northwest Pacific under current and future climates with an exclusive focus on high-latitude habitats.We examined the performance of these models in projecting expansions of thermally suitable habitats with respect to pre-industrial to present conditions as well as under various climate change scenarios as modelled by the IPCC in the coming century.We sought not only to define thermal suitability changes in the poleward (leading) range limits but also to identify how climate change may affect the proliferation of reefs currently near their poleward limit.Our goal was to identify the magnitude and spatial extent of (1) thermal suitability increases from pre-industrial historical to present conditions, as well as (2) thermal suitability increases under future warming of seasonally adapted populations of these five corals.As a result of the Kuroshio current driving stark temperature regime differences along Japan's coastline, we further contextualise our analyses on a finer spatial scale by narrowing in on this region specifically.

| Thermal performance data
Thermal performance metrics for five coral species were derived from measurements in McIlroy et al. (2019).That study modelled thermal performance of coral holobionts within an existing, marginal habitat across a year in which in-situ temperature variation ranged from 14 to 31°C.These species, namely Acropora samoensis, Galaxea fascicularis, Montipora caliculata, Oulastrea crispata and Porites lobata, exhibit a variety of morphologies and each a broad latitudinal distribution range (Veron et al., 2022).Briefly, corals were collected from Tung Ping Chau (22°32′34.3″N, 114°26′19.6″E), and were reared and assessed in mesocosms at the Swire Institute of Marine Science where they were exposed to seawater pumped in from a nearby bay.As a proxy for holobiont productivity, we used the gross photosynthesis to respiration ratio (P:R) measured monthly from August 2015 to August 2016.Here, we extracted maximum and minimum B 80 limit values for three species found to yield unimodalshaped thermal performance curves (A.samoensis, G. fascicularis and M. caliculata).Thermal performance of two species (O.crispata and P. lobata) yielded positive linear regressions within the annual temperature range, and thus those B 80 limits were not defined or applied herein.Additionally, for all species, we then estimated the temperature minimum for functional symbiosis, defined as the lower bound wherein P:R = 2 (Coles & Jokiel, 1977; Table S1).A P:R value of at least 2 was consistently maintained through annual temperature variations in Hong Kong coral populations of all five species (McIlroy et al., 2019).

| Thermal suitability modelling
We calculated mean monthly sea surface temperatures at the regional scale in Eastern Asia for present and future conditions from a set of 11 and 12 general circulation models from the Coupled Model Intercomparison Project Phase 6 (CMIP6), respectively (Kwiatkowski et al., 2020;Rogers et al., 2019).We converted projections from all models to a standardised 1° × 1° spatial grid through bilinear interpolation (Sung et al., 2021), ranging from E 90° to 170° in longitude, and N 10° to 60° in latitude.We estimated historical and present conditions as 30-year averages spanning 1870-1900 and 1984-2014, respectively, and future conditions as a 30-year average spanning 2070-2100 under the SSP5-8.5 shared socioeconomic pathway, representing a high emissions scenario.
To assess and predict thermal habitat at a finer, local spatial scale, we used the future ocean regional projection (FORP, Nishikawa et al., 2021) dataset, which is the product of mechanistic downscaling of projections from the Coupled Model Intercomparison Project Phase 5 (CMIP5, Taylor et al., 2012) at a 2 km resolution around Japan.In order to allow for inter-model comparisons, we used monthly mean values derived from daily sea surface temperatures, using two models, MIROC5 and MRI-CGCM3, that are known to well represent the Kuroshio current.Here we projected conditions for present (1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005) and future (2086-2100) time periods.For future conditions we applied a severe Representative Concentration Pathway scenario (RCP 8.5) in both models and additionally applied a moderate scenario (RCP 2.6) for MIROC5.

To model shifts in thermal habitat suitability throughout Eastern
Asia at both regional and local scales, we estimated the proportion of time spent within thermally optimal conditions across the spatial grid for each species.For A. samoensis, G. fascicularis and M. caliculata, we calculated time spent within B 80 performance breadths, with the assumption that at least 50% of the time in this range is necessary for persistence (McIlroy et al., 2019).From these estimates we therefore determined the contour areas where species are within their B 80 ranges at 50% of time, as well as 75% of the time as a conservative estimate for comparison.For A. samoensis, G. fascicularis, M. caliculata, O. crispata and P. lobata, we calculated time spent at temperatures where P:R ≥ 2, which species should maintain 100% of the time for survival (Coles & Jokiel, 1977;McIlroy et al., 2019).
In our local scale analysis, we tabulated thermal habitat suitability values based on time spent above or below B 80 and P:R thresholds for marine protected areas in Japan (Figure 1).These areas include specially protected marine park areas that are situated within national parks and quasi national parks.For each protected area we obtained the minimum value of the grid intersecting the region.

| Regional-scale thermal suitability projections
Model projections based on proportion of time spent within the B 80 thermal range, where a minimum of 50% is required for survival, suggest northward shifts in all three species are possible (Figure 2).
We found latitudinal limits of thermally suitable habitat for A. samoensis, G. fascicularis and M. caliculata to have shifted northward from historical (1870-1900) to current  conditions, and predict further, more extensive shifting in this direction from current to future (2070-2100) conditions.Projected future shifts in suitable range for all three species cover approximately 2° in latitude at the northern edge.The thermal regime for corals at 20°-30° latitude is also changing markedly (Figure 4).Southern habitats show marked changes in time spent in B 80 , so that suitable spatial extents for historically seasonally adapted populations appears to be narrowing for all three species from historical to current conditions, a trend projected to continue under future conditions.
We assumed that corals are only able to inhabit ranges wherein coral productivity can be maintained at P:R ≥ 2, so that a monthly mean below this point for any species-model combination was sufficient to exclude corals, that is, represented their critical thermal minimum.Based on this criterion, our models predict poleward expansions in thermally suitable habitat along Northern Japan in all five species (Figure 3).From historical to current conditions, the latitudinal limits defined by their lower thermal limit have shifted for all species except for A. samoensis.We project this thermal habitat to further shift northward more extensively for all five species under future conditions, by approximately 6° in latitude for O. crispata, and 2°-3° for the other four species.

| Local fine-scale thermal suitability projections
Projections of time spent in the B 80 range at a local, fine resolution, where 50% of the time is the threshold, also suggest northward shifts in the extent of thermally suitable habitat for all three species (Figure 4).This shift is more extensive under the RCP 8.5 rather than the RCP 2.6 emissions scenarios for the MIROC model across species.RCP 8.5 projections from the MIROC model are also more extensive than those from MRI-CGCM3.Yet for both models, RCP 8.5 projections suggest changes in the thermal suitability of habitats at lower latitudes within proximity of Japan, which is not evident from the RCP 2.6 projections from MIROC.Averaging RCP 8.5 projections from both models, latitudinal shifts at the northern edge of the thermally suitable range approximate 1° for A. samoensis and G. fascicularis and 3° for M. caliculata (Table S2).

F I G U R E 1
Recorded distribution range of study species within surveyed national and quasi national marine park areas of Japan (numbered, with red marks indicating survey sites).Coloured dots represent the reported presence of each coral species in within each park area, as extracted from the Ministry of the Environment ( 2004).The species include Acropora samoensis (dark red), Galaxea fascicularis (pink), Montipora caliculata (yellow), Oulastrea crispata (teal) and Porites lobata (blue).Where coloured dots are absent, these coral species have not been recorded.
Local scale projections also depict a poleward extension of thermally suitable habitat according to the P:R ≥ 2 threshold (Figure 5).
The northern shift for P. lobata, A. samoensis and G. fascicularis follow similar patterns based on P:R ≥ 2 constraints, though with A. samoensis reaching the northernmost latitudes.However, projections suggest less extensive change for M. caliculata, and moderate shifts for O. crispata yet at more southern latitudes than the other four species.Overall, projected northern shifts in suitable ranges are more extensive under the RCP 8.5 emissions scenario from MIROC, while RCP 2.6 projections suggest less pronounced change, and are similar to RCP 8.5 projections from MRI-CGCM3.
Under current conditions, most of Japan's marine protected areas, including subtropical to mid-temperate regions (i.e., until No. 17 Minami Boso), are thermally suitable according to either B 80 or P:R ≥ 2 thresholds for all five coral species (Table 1).These include areas not covered by the known distribution of these species.Under a MIROC RCP 8.5 scenario, the southern latitudinal limits according to B 80 constraints for A. samoensis and G. fascicularis are below their known southern limits of distribution especially in subtropical areas where corals are currently abundant and conserved (such as No.1-5).
According to B 80 constraints, future projections suggest the three species will achieve sufficient thermal suitability in Japan's third (No. areas of real and ongoing poleward expansions (Veron et al., 2022;Yamano et al., 2011).Potential shifts in thermally suitable habitat are even more pronounced from current to future conditions.
Along the Japanese coast, the poleward expansion of thermally suitable habitats for these species could span, on average, approximately 185 km (1.67°; Figure 4; Table S2) under severe future warming and based on B 80 constraints.While these northern areas may soon support coral survival, constraints on time spent in B 80 , relative to tropical reefs, suggest that corals will be limited in their ability to proliferate, and may exist as species rich but patchy, non-reefal coral communities similar to those found in marginal reefs of today.
The application of our mechanistic model to current conditions is broadly consistent with the current range limits of high-latitude coral distributions both regionally and globally (Beger et al., 2014;Makino et al., 2014;Ministry of the Environment, 2004;Veron et al., 2022), supporting the usefulness of Hong Kong corals and their physiological limits as a proxy for understanding range edge populations.Indeed, while P:R values provide the framework for the use of mechanistic models herein, more species/population F I G U R E 3 Thermal suitability for seasonally adapted populations of five coral species (Oulastrea crispata, Porites lobata, Acropora samoensis, Galaxea fascicularis and Montipora caliculata) calculated as proportion of time spent at temperatures with their respective P:R ≥ 2, averaged across 30 years in Eastern Asia under historical (left), current (middle) and future (right) conditions.Grey dashed lines in each panel represent the northernmost latitude where the species maintains a P:R ≥ 2 for 100% of the time in a coastal region.specific estimates of coral resilience which incorporate both autotrophic and heterotrophic energy sources (Conti-Jerpe et al., 2020) may help refine expectations.Moreover, the P:R data used in this study were collected under naturally occurring temperatures (McIlroy et al., 2019).Further experiments under controlled conditions and including temperature extremes could enhance thermal performance estimates.Thermal performance can be measured easily for diverse biological functions, hence its applicability to many marine species (Hui et al., 2019;McIlroy et al., 2019), yet have mainly been explored in terrestrial systems (Hertz et al., 1993).In this regard, B 80 limits are a broadly applicable metric of thermal physiology for ectotherms (Angilletta, 2009).Such additional studies on the physiology of these corals would therefore aid in further developing the applicability of our modelling approach to actual species distributions under climate change.Assessing the thermal physiology of locally-adapted populations along the latitude gradient ranging from Hong Kong to Northern Japan, and incorporating more comprehensive assessments of physiology, would be especially informative.
The current actual distribution limits for the corals, with the exception of O. crispata discussed below, occur at slightly lower latitudes than those projected from thermal suitability (Table 1).
Therefore, the strong potential for expansion in the region from the perspective of thermal physiology stresses the importance of other biotic and abiotic factors dictating coral edge population survival (Maes & van Dyck, 2022).Within a conservation framework, the thermal thresholds we have applied in this study should therefore be interpreted as a bare minimum for survival.While 50% of the time spent in the B 80 range can serve as a fundamental threshold for coral survival, in this case, a 75% cutoff was a better match to the rough estimate of the realised thermal range of corals currently (Ministry of the Environment, 2004, Table S2).Thus, the requirements for proliferation, rather than survival, may require a higher proportion of time spent in B 80 to allow for sufficient growth to overcome ecological and physical constraints.Increases in time spent in B 80 may also contribute to increasing abundances of hard corals near range edges where they were previously present but rare (Kim & Kang, 2022).The largest mismatch was seen for M. caliculata which is currently only reported in the most southern regions of Japan (Figure 1).Where present, M. caliculata is relatively rare (Vroom et al., 2010), suggesting strong ecological structuring within its current range that may also limit its expansion.These thresholds provide insights on the capacity for current corals to persist generally, while also highlighting the challenges posed to populations that may recruit into habitats beyond the current range edges.
Yet there are differences in the extent to which the species' actual current distribution matches modelled thermal suitability predictions.The most striking is O. crispata, which currently extends to

TA B L E 1 (Continued)
at least ~36° N (Figure 1; Lien et al., 2013), well beyond the ~30° N threshold set by our models.These models assume that the obligate symbiosis between corals and photosynthetic Symbiodiniaceae cannot be sustained below a threshold of P:R = 2 (Coles & Jokiel, 1977).
Yet there is evidence that O. crispata is an exceptional example of a facultative host among hermatypic, reef building corals, which can survive long periods (>1 year) without symbionts (Denis et al., 2012).
This indicates that O. crispata is not bounded by the productivity threshold estimated for a functional symbiosis which was defined in our model (P:R > 2.0; Coles & Jokiel, 1977).Thus, while populations from Hong Kong conformed with the physiological traits of cooccurring coral species (McIlroy et al., 2019), this parameter appears poorly suited as a survival threshold for O. crispata more broadly.
It is critical to note that under future warming, environmental adaptation (Bairos-Novak et al., 2021) and population connectivity (Nakabayashi et al., 2019) will play key roles in determining species range expansions.Previous studies have also predicted northward extensions of coral populations, but have used statistical models mainly focused on ocean currents and/or larval connectivity to predict the impact of warming temperatures along coastal Japan (Kumagai et al., 2018;Makino et al., 2014;Nakabayashi et al., 2019).
Building on these studies, the application of mechanistic models herein, reveals the nuances of seasonality and its constraints on coral thermal performance.For example, we observed a stark dif- As climate change progresses, our models highlight that thermal regime changes will, as in the tropics, decrease the thermal suitability of coral populations adapted to their current location, mainly near southern latitudinal limits.This poses significant threats to long-lived, sessile, coral individuals which become physiologically mismatched to new conditions (Putnam, 2021).However, threats to these marginal reef ecosystems may be less pronounced than for equatorial corals, due to their high capacity to receive warm adapted genotypes from lower latitude populations (Nakabayashi et al., 2019;Yasuda et al., 2014).In this region unidirectional larval dispersal of corals from lower to higher latitude regions is facilitated by the Kuroshio Current, which flows northwest along the East China Sea and loosely follows Japan's Pacific facing coasts before turning East again at approximately N 35° (Gallagher et al., 2015).While dispersal of planktonic gametes and larvae is limited in span and directionality in this case it allows for the delivery of larvae into new subtropical and temperate environments (Fifer et al., 2022;Nakabayashi et al., 2019).
This is an important factor given that coral genetic diversity for range edge populations in this region can be limited by annual minimum temperatures (Fifer et al., 2022).However, if the rates of natural connectivity and local adaptation in this group of foundational species cannot keep pace with environmental change, disruptions to these marginal ecosystems could be severe (Selmoni et al., 2020;Sunday, 2020), despite shifts in thermally suitable regions.
Japan contains many high-value coastal marine ecosystems facing ongoing changes in species distributions, emphasising its potential to lead national-level conservation management strategies.
Nearly the entire coastline is managed to some degree, although often locally and prioritising fisheries (Makino et al., 2014;Yagi et al., 2010).While there are marine protected areas aimed at the protection of corals, mainly for bolstering tourism, resources invested in conservation strategies can be limited (Abe et al., 2022).
Our results suggest Japan's northern marine protected areas to become thermally suitable for subtropical corals under future warming (Figure 1; Table 1).Larger scale management strategies should therefore take into account the warming induced northward range expansions of tropical corals, the increased favourability of range edge habitats to coral productivity, ensuing ecological cascades (Kumagai et al., 2018;Vergés et al., 2014) and the potential vulnerability of pre-existing temperate ecosystems (Kim & Kang, 2022;Makino et al., 2014;Nakabayashi et al., 2019;Pinsky et al., 2020;Vergés et al., 2019).Better integrating genetic connectivity and adaptive potential into mechanistic models such as the one outlined herein is a critical next step for understanding and potentially leveraging range shifts for coral management and conservation.
This will be key in designating areas most in need of a marine protected area status (Selmoni et al., 2020;Yamakita et al., 2022).
The availability of spatially and temporally downscaled thermal prediction data is also a key precondition for a global understanding of coral species ranges under past, current and future conditions.
Ocean climate datasets available at a fine temporal resolution are typically projected at coarse spatial resolutions, thus failing to capture coastal habitats, which are of prime importance to coral reefs and other high-value ecosystems.Considering the constraints of seasonality on the physiology of corals (Keshavmurthy et al., 2021;McIlroy et al., 2019), climate data at a fine temporal resolution, e.g.monthly, are critical for mechanistically estimating the potential for future range shifts at a precision comparable to that displayed by correlative models (Adam et al., 2021;Doherty et al., 2021;Drury et al., 2022;Larkin et al., 2021).While such datasets are available for some regions, such as that for Japan applied in this study, these are limited in their geographic scope.
Overall, we utilised a relatively simple proxy for the physiological limits of coral populations living in a highly seasonal environment to project shifts in thermally suitable range edges.Our models estimate northward expansions of coral physiological limits across the Northwest Pacific and especially along Japan's coastline.Following 23) northernmost marine protected area.In this regard M. caliculata in particular is projected to achieve thermal suitability in Japan's second northernmost protected area(No.24).Yet some marine protected areas on the west coast of Japan are projected to become less suitable for corals presently acclimated in those regions according to B 80 constraints forA.samoensis and G. fascicularis (such as No.   11, 18, 21 and 22).According to P:R constraints, up to six additional marine protected areas become suitable under future conditions for M. caliculata.4 | DISCUSS IONOur results consistently show critical changes in the thermal regime of coastal marine habitats at the poleward range edges of Asian Pacific corals, including changes to both annual extremes and seasonal variation.Using the thermo-physiological constraints of coral populations adapted to extreme seasonality, we highlight areas constituting an expansion of thermally suitable habitats into higher latitudes across the Northwest Pacific under current and predicted ocean warming scenarios.While the exact present day range limits are poorly defined for these species, applying physiological constraints to pre-historical and current conditions in the region reveal shifts in thermally suitable habitat broadly matching F I G U R E 2 Thermal suitability for seasonally adapted populations of coral species (Acropora samoensis, Galaxea fascicularis and Montipora caliculata), calculated as proportion of time spent by in their respective B 80 thermal performance ranges, averaged across 30 years in Eastern Asia under historical (left), current (middle) and future (right) conditions.Grey dashed lines in each panel represent the northernmost latitude where the species is within its B 80 range for at least 50% of the time in a coastal region.

F
Two kilometres resolution analysis of thermal suitability for seasonally adapted populations of coral species (Acropora samoensis, Galaxea fascicularis and Montipora caliculata), calculated as proportion of time spent in their respective B 80 thermal performance ranges.Values are averages across 10 years (1996-2005) for current conditions, and 15 years (2086-2100) for future scenarios, including RCP 8.5 and RCP 2.6 under the MIROC5 model, and RCP8.5 under MRI-CGCM3.Light blue areas in each panel represent a contour where a species is within their B 80 range for 50% of the time.

F
Two kilometres resolution analysis of thermal suitability for seasonally adapted populations of five coral species (Oulastrea crispata, Porites lobata, Acropora samoensis, Galaxea fascicularis and Montipora caliculata) calculated as proportion of time spent at temperatures with their P:R ≥ 2. Values are averages across 10 years(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005) for current conditions, and 15 years (2086-2100) for future scenarios, including RCP 8.5 and RCP 2.6 under the MIROC5 model, and RCP8.5 under MRI-CGCM3.Light blue areas in each panel represent a contour where a species maintains P:R ≥ 2 for 100% of the time.TA B L E 1 Tabulated thermal habitat suitability values according to marine protected area in Japan.
ference between modelled time spent in B 80 for source corals, at approximately 50% of time annually within the marginal coral community of Hong Kong, and in applied models, at nearly 100% of the time annually in Japan.A direct assessment of local coral populations is necessary to determine their true B 80 , as well as their potential for thermal adaptation under climate change(Bairos-Novak et al., 2021).Yet the similarity in temperature mean and annual range between Hong Kong (https://www.hko.gov.hk/) and Southern coastal Japan (https://www.data.jma.go.jp/) likely poses similar limits on thermal performance in terms of the minimum and maximum tolerable temperatures.Conversely, the buffering capacity of the Kuroshio current to prolonged temperature extremes appears to facilitate the maintenance of near optimal rates of coral productivity for most of the year and may underpin the proliferation of these well-developed coral reefs.
Bold values denote parks within regions where the species have been reported, bold italicised values denote parks within regions where the species are reported as absent, italicised values are unknown (Ministry of the Environment Japan, 2004).