Introduction

Over 15 years have passed since a perspective was published in Coral Reefs, entitled ‘Coral reef environmental science: Truth versus the Cassandra’. In this piece, Grigg (1992) compared the alarm of some in the coral reef research community at the first major episodes of coral bleaching to the earlier doomsday predictions of global reef decline that accompanied population outbreaks of the corallivorous crown-of-thorns seastar (COTS) (Acanthaster planci), in the Pacific Ocean in the 1960s (e.g. Chesher 1969). Grigg (1992) suggested that things were going to be different this time around and that the predictions being made at that time would be rigorously tested by science:

This time the Cassandras will be tested by the truth of careful experimentation, long-term monitoring and objective interpretation. Coral reef science appears to have come of age.

The Cassandras of Grigg’s day forecast that the first major episodes of coral bleaching and mortality in the Caribbean and Eastern Pacific were the early warning signs of global warming. At that time, and as in the case of Cassandra, few were prepared to listen to the predictions being made, which, with the benefit of hindsight, are now regarded as prescient. Indeed, climate-induced coral bleaching has already contributed to widespread and accelerating coral loss. This has lead, when combined with prospective impacts from ocean acidification, to predictions that functioning reef ecosystems will vanish within decades (Hoegh-Guldberg 1999; Knowlton 2001; Sheppard 2003). Those making predictions now, however, cannot be compared to Cassandra because they have clearly been heard by an increasingly attentive global audience, as evidenced by the prominence given to reefs in IPCC (2007) and in the international media. The question is are they, like Cassandra, correct, or are these the false prophets that Grigg (1992) warned us of 15 years ago? Only time will tell, yet the premises behind some of the predictions can be examined, and we argue that many are either unsupported by existing data or have yet to be thoroughly tested.

Dire predictions may well be necessary to invoke much-needed change in policy and public perceptions; however, if these are based on unsupported or untested assumptions, there is a risk that future more well-supported predictions will be ignored. In truth, predictions regarding specific changes in coral assemblages are seriously compromised by critical knowledge gaps, highlighting a need for targeted future research. Then, once gaps have been addressed, defensible and credible scenarios can be put forward. This essay assesses several essentially untested assumptions that form the basis of widely popularised predictions regarding the future of reefs and their capacity to cope with climate change. First, we challenge the proposition that all coral species are living close to their upper thermal limits (Mayer 1918; Hoegh-Guldberg 1999). Second, we point out that the data to determine whether corals can acclimatize or adapt to accelerating rates of environmental change are not available (Donner et al. 2005). Third, we argue that the experimental data do not yet exist to support suggestions that physiological trade-offs needed to cope with ocean acidification will lead to reduced reproductive potential (Hoegh-Guldberg et al. 2007). Last, we question whether ongoing climate-induced coral loss will cause fisheries to collapse (McClanahan 2002; Hoegh-Guldberg et al. 2007).

All coral species are living close to their upper thermal limits

Predictions that reefs will disappear as a result of global warming are based, at least in part, on the assumption that corals are living close (within 1–2°C) to their maximum thermal limits (Hoegh-Guldberg 1999). While differences in the susceptibility of coral taxa to thermal stress have been demonstrated experimentally (Edmondson 1928), comparisons of susceptibility among taxa in speciose field assemblages have only recently become available (Marshall and Baird 2000). While most colonies bleach following temperatures extreme in magnitude, duration, and rate of increase, the severity of bleaching responses varies dramatically within and among taxa (McClanahan et al. 2009). Such variable bleaching susceptibility implies that there is a considerable variation in the extent to which coral species are adapted to local environmental conditions. Furthermore, geographic variation in bleaching thresholds within species, sometimes over scales <100 km, provides circumstantial evidence for ongoing evolution of temperature tolerance between both species and reefs (see, review in Coles and Brown 2003). As a consequence, even in the absence of an adaptive response, a change in the relative abundance of species is a far more likely outcome of climate change than the disappearance of reef corals (Loya et al. 2001; McClanahan 2002; Hughes et al. 2003). It will be impossible to shelter reefs from large-scale disturbance, like the more frequent and severe bleaching events likely to accompany global warming (IPCC 2007). Even so, communities with high percentages of taxa susceptible to bleaching, as well as bleaching tolerant taxa can be identified for some reef regions (Loya et al. 2001; McClanahan et al. 2007). For other regions, more research is required to determine the ranges of thermal tolerance present. Then, efforts made to reduce localised and controllable sources of stress would be both more informed and targeted (Goreau et al. 2000).

Most pertinently, little is known about the sensitivity of population growth to climate-induced changes in vital rates. This results from a limited availability of empirical data on the effects of temperature on even the most basic vital rates in corals, such as growth, mortality and fecundity. A large body of evidence, however, supports temperature tolerance varying among species, populations, communities, and reef regions (Marshall and Baird 2000; Coles and Brown 2003). In the absence of long-term demographic studies to detect temporal trends in life history traits, predicting rates of evolution and whether they will be exceeded by rates of environmental change, is impossible (Visser 2008).

Corals cannot adapt or acclimatize to projected rates of change

That corals lack the capacity to adapt to projected rates of environmental change is a hypothesis based on three key assumptions: (1) that generation times are too long to allow for adaptation over the required timeframes (Hoegh-Guldberg 1999); (2) that the scale of dispersal is too large to allow for adaptation to local conditions (Potts 1984); and, (3) that there is insufficient genetic diversity in existing symbionts and corals (Hoegh-Guldberg et al. 2007). Many features of coral life histories such as extended life-spans, delayed maturation, and colony fission should indeed result in long generation times (Hughes et al. 1992) and in coral species with these traits, generation times are between 33 and 37 years (Babcock 1991). However, other corals, in particular those species most susceptible to thermal stress and most prolific, such as many species of Acropora and Pocillopora, mature early, grow rapidly, and senesce (Rinkevich and Loya 1979). There is also strong evidence that coral populations are already locally adapted to specific environmental conditions, including temperature (see, review in Coles and Brown 2003). While most corals have larvae capable of spending many months in the plankton (Graham et al. 2008), the actual dispersal distances are often in the order of 10 to 100 km (Hughes et al. 2000). Furthermore, while coral bleaching can affect coral fecundity and egg quality, most surviving individuals of even highly susceptible species produce viable gametes following bleaching, ensuring thermal tolerance can pass between generations (Baird and Marshall 2002).

On reefs, genetic studies indicate that gene flow is actually quite restricted, as populations of many coral species are highly sub-divided (Hellberg 2007; Underwood et al. 2007). Recent research into contemporary evolution indicates that moderate gene flow does not impede adaptation to local conditions (Garant et al. 2007). The life histories of ecologically important and relatively abundant genera, limited dispersal, and restricted gene flow all suggest an as yet undefined and under-appreciated capacity to evolve to rapidly changing environments. Repeated bleaching episodes in the same coral assemblages and the increasing scale and frequency of coral bleaching are cited as evidence that corals have exhausted their capacity to evolve to rising sea surface temperatures (Hoegh-Guldberg 1999). However, few studies have compared the rates of mortality or thermal tolerance within populations in sequential bleaching events. Conceivably, mortality rates may be declining even while the spatial scale and frequency of bleaching is increasing. Indeed, a number of studies suggest that bleaching mortality rates have declined and thermal tolerance has increased in some regions. For example, mortality rates in the Eastern Pacific were significantly lower in 1998 when compared with 1982 and 1983 (Glynn et al. 2001) Similarly, Maynard et al. (2008) found thermal tolerance of three common coral genera on the Great Barrier Reef to be greater in 2002 than that expected from the relationship between temperature stress and bleaching severity observed in 1998. While there must be upper limits to rates of adaptation, these examples suggest that the evolutionary capacity of corals to respond to thermal stress is far from exhausted.

Acclimation through symbiont shuffling from less to more stress-resistant clades is another mechanism by which corals may increase the thermal tolerance of the holobiont. Similarly, shifts in the patterns of association between host and symbiont may occur between generations in species which do not transmit symbionts in the gametes (Baird et al. 2007). Indeed, there is growing evidence that such shuffling can increase thermal tolerance, at least in the short term (Berkelmans and van Oppen 2006). This is a promising area of research, but heralding symbiont shuffling as a panacea for reefs threatened by climate change is premature. Efforts must now be directed towards establishing whether the increase in thermal tolerance of the holobiont following symbiont shuffling persists and reduces mortality during subsequent thermal anomalies.

Trade-offs resulting from ocean acidification lead to reduced fecundity

Ocean acidification has been hypothesised to affect energy allocation within coral colonies. In particular, acidification may force corals to invest more energy in calcification to produce skeletons of similar strength (Hoegh-Guldberg et al. 2007), leaving less energy for allocation to reproduction. However, it has yet to be shown that corals are capable of expending more energy on producing skeletons under lower aragonite saturation states. On the contrary, if calcification rates decrease, more energy could become available for reproduction, maintenance, or storage—a slight net positive benefit, at least in the short term, for corals within low-energy environments. In addition, Anthony et al. (2007) present evidence that high initial energy levels reduce mortality risk from bleaching. Most importantly, more work is required to discern whether the biology of scleractinians allows for a variable rate of calcification and, if so, what, if any, physiological trade-offs will result.

Climate-induced coral loss leads to widespread fisheries collapse

Climate change has already and will continue to cause changes in the structure of coral reef communities (e.g. Graham et al. 2008). However, the extent to which these changes will affect ecosystem function, productivity and fisheries yields is far from certain. For example, extensive degradation of coral reef habitats is predicted to jeopardise current catches of coral reef fishes (McClanahan 2002), but no studies have actually shown that total catch, catch composition, or fisheries value have been significantly affected by severe mass bleaching (McClanahan et al. 2002; Grandcourt and Cesar 2003; Pratchett et al. 2008). Direct effects of coral loss on fishes appear to be primarily limited to the 10–12% of species that rely on coral for food or shelter and none of these species are important fisheries species (Wilson et al. 2006; Pratchett et al. 2008). The overall abundance and diversity of reef fishes are certainly sensitive to declines in ‘reef rugosity’, which may occur several years after extensive coral mortality (e.g. Graham et al. 2007). However, declines in reef rugosity do not always occur, and depend on relative contributions of contemporary coral growth versus erosion of the underlying reef framework (Pratchett et al. 2008). On fronts of relatively exposed reef systems (e.g. midshelf and offshore reefs of the Great Barrier Reef in Australia), the underlying reef matrix contributes greatly to reef rugosity, so the death and collapse of existing coral growth will have limited impact on the habitat structure or fish assemblages (Halford et al. 2004). The critical and currently unanswerable question is whether ocean acidification will further weaken the carbonate framework leading to more extensive and accelerated reef collapse and further reductions in topographic complexity.

Even if climate change is compounding upon current fishing activities to reduce stocks of major target species (e.g. Graham et al. 2006; Hoegh-Guldberg et al. 2007), it is likely that this will precipitate a change in fishing practices and target species, rather than widespread fisheries collapse (e.g. Jackson et al. 2001). Extensive research is still required to combine emerging information on species-specific vulnerabilities of individual fishes to climate change (e.g. Munday et al. 2008) with improved understanding of the motivations of fishers for targeting certain species, and their potential to adapt to changing resource availability. Effects of climate change on fisheries yields will have significant ramifications, especially in tropical countries that rely extensively on small-scale reef-based fisheries (Brander 2007). However, it is premature to suggest that widespread reef collapse is a certain consequence of ongoing bleaching, or that this will inevitably lead to fisheries collapses.

Future research and conclusions

In this article, we identify serious knowledge gaps that limit our ability to predict the future of reefs and their capacity to adjust to climate change. In particular:

  1. 1.

    In order to determine generation times and changes in vital rates associated with thermal anomalies, long-term ecological studies of species with contrasting susceptibilities to thermal stress are required;

  2. 2.

    Critical thermal maxima and acidification thresholds for corals need to be identified and the latitudinal variation in these thresholds needs to be explored. When combined with detailed ecological parameters, such as generation times, knowledge of thresholds can be used to make robust predictions about the potential of coral species to adapt to changing environments;

  3. 3.

    Comprehensive data should be gathered on the extent to which shuffling symbiont types results in a long-term increase in the thermal tolerance of the holobiont;

  4. 4.

    Experimental studies should be conducted on the effects of increasing temperature and changing water chemistry on fundamental biological processes, in particular, trade-offs among skeletal density, growth, and fecundity;

  5. 5.

    The effects of increasing temperature and acidification on the early life history stages of corals need to be determined; and,

  6. 6.

    Experimental studies are required to assess species-specific vulnerabilities of fishes (especially, important fisheries species) to projected environmental changes.

Although the six research areas detailed above do not encompass a comprehensive list of climate-change-related research priorities, the filling of these critical knowledge gaps would help provide resource managers and policy makers with well-supported predictions of likely changes to reefs. Significantly, well-supported predictions facilitate fine-tuning the allocation of what will always be limited management resources to lessen the suite of impacts on reefs, of which climate change is only one.

Regardless of future emissions and emission policies, existing inertia within the climate system means further climate change is inevitable (Donner et al. 2005; IPCC 2007). Consequently, predictions made today will not be forgotten quickly and if incorrect, might constrain the capacity of the scientific community to influence future policy. In the International Year of the Reef, and beyond, it seems prudent to focus on ensuring that reef managers maintain the will to maximise the capacity of reefs to withstand disturbances by implementing the sorts of on-the-ground actions that must accompany as yet unsuccessful agitation for policy change on emissions. Such an approach seems more productive than publicising what can and is easily interpreted by the media and some stakeholders as the pointlessness of making any effort at all.