Reef crises need to be separated from mass extinctions because they are manifested in reductions of reefal carbonate production rather than elevated extinction rates. The volume of preserved fossil reefs per unit time is perhaps the best accessible metric to assess reefal carbonate production rates in the geologic record. Although this metric is prone to biases introduced by weathering, burial, and sampling, it offers the possibility to analyze general connections between reef crises and mass extinctions. The biases can be partially corrected by looking at short-term variations and by utilizing independent proxies of sampling. Using a comprehensive database of ancient reefs and considering the generally high volatility in reefal carbonate production, we can identify five significant metazoan reef crises in the post-Cambrian Phanerozoic, only three of which correspond to traditional mass extinctions. Ancient reefs crises appear to be due to episodes of rapid CO2 release and warming, rather than cooling or meteorite impacts.

From the beginning of the Late Ordovician (Sandbian: 460.9myr) through end Devonian (Famennian: 359.2myr), coral-stromatoporoid sponge reefs formed a remarkable, evolving ecosystem that dominated sediment production on tropical carbonate platforms in a calcitic ocean. This was a time of maximal and unparalleled reef development in the Phanerozoic, with reef tracts vastly exceeding in size and biodiversity of those in the Holocene (e.g., the Great Barrier Reef). Within this circum-equatorial niche, the calcitic tabulate and rugose corals, and the aragonitic (or high Mg calcite) stromatoporoid sponges, were the primary Middle Paleozoic reef frame builders. These were supplemented ecologically and skeletally by now extinct groups of calcitic bryozoans, crinoids, brachiopods, and red algae, alongside aragonitic green algae, and enigmatic CaCO3 precipitating and binding calcimicrobes. This 100 myr long Middle Paleozoic reef consortium thrived under SST averages of 30°+, to latitudes as high as 45°–55°, under high atmospheric CO2 conditions of 6000+ ppm, and sealevels 150–200 m higher than today. This reef ecosystem was disrupted by several relatively short duration south polar glacial episodes, centered around northern Gondwana, defining the O/S boundary Mass Extinction Events (MEEs). Nearly all coral and stromatoporoid families survived this MEE: there were losses at the genus level. Reef-building stopped nearly everywhere, and during the ‘recovery’ interval, solitary rugose corals initially prevailed, and stromatoporoids were small. Full global re-establishment of the reef ecosystem, and biodiversity, took another 3–4 million years (not until the late Aeronian, Early Silurian). This was followed by a remarkable reef expansion in the Middle Silurian (Wenlock), then by declines in the latest Silurian (Ludlow-Pridoli), and earliest Devonian (Lochkovian) possibly due to sealevel lowstands, tectonic plate re-assembly, and ocean current re-direction. Maximal Phanerozoic reef success was during the Emsian-Givetian, when some 15 barrier reef tracts more than 1100 km long flourished in tropical shallow seas. Reef-building coral diversity exceeded 200 genera, and the calcifying stromatoporoids evolved 60+ genera, especially in the ‘Old World’ faunal province (Euramerica, Cathaysia, northern Australia). Near the end of the Middle Devonian (mid- to late Givetian), the primary reef dwellers declined sharply in diversity, marked generally by sealevel lowstand, followed in the Frasnian (Late Devonian) by shrinking latitudes for carbonate platforms, and reduced reef accommodation space. Sharp cooling, with the arrival of a global Icehouse climate, and aragonitic oceans, led to the second largest Phanerozoic Mass extinction around the Frasnian/Famennian boundary, with reef builder and reef inhabitant losses exceeding those of the O/S MEE. The global absence of coral-sponge reefs persisted for nearly all of the 16 myr long Famennian, as total CaCO3 production fell some 60–90%, as aragonitic oceans took over. Only small and scattered Famennian coral-stromatoporoid patch reefs are known, with the last of these in the late Famennian (Strunian), punctuated by total disappearance of the whole keystone reef-building order. Famennian and Strunian corals belonged to Carboniferous families. During the Famennian, calcimicrobes, the first calcifying foraminiferans, and select ‘lithistid’ calcareous sponges dominated a highly stressed reef ecosystem, lacking barrier reef tracts. Biodiversity and reef construction were decoupled under global climatic stress during the succeeding icehouse Late Paleozoic.

Photosymbiosis has been an important process in the evolution of ancient reef systems and in reef success today. Modern reefs and many of those in the geologic past inhabited nutrient-depleted settings. The complete collapse of some ancient reef ecosystems may be attributed to the breakdown of the ecologic and physiologic relationships between symbiont and host. Many algal groups developed symbioses with calcifying metazoans and protists and live with them, but the most common of these today are dinoflagellates in the genus Symbiodinium, sometimes called zooxanthellae. This photosymbiotic relationship conferred important metabolic advantages to both partners, allowing exploitation of tropical, shallow-water oligotrophic settings. In addition to improved metabolism, a by-product was rapid calcification which increased the growth of reefs and provided advantages to the hosts through larger and stronger skeletal support. Strong evolutionary pressures exerted by the symbiont-host relationship created bonds and favored longevity and adaptive novelty. Photosynthesis accounts for the remarkable reef growth and carbonate sedimentation in the tropics. Photosymbiosis gave reef organisms an adaptive edge to develop new life strategies that could not be developed by organisms which did not foster this relationship. Many living calcified organisms harbor many different photosymbionts and likely a variety of ancient calcified organisms did too (foraminifera, calcified sponges, corals, brachiopods and bivalve mollusks). Symbiodinium now a dominant symbiont apparently appeared in the Eocene and so was probably not utilized by earlier reef organisms, although the fossil record of dinoflagellates most closely related to Symbiodinium extends back to the Triassic. Today Symbiodinium inhabits a wide variety of unrelated host organisms from protists to mollusks. While the identity of more ancient photosymbionts is unclear, indirect evidence suggests photosymbiotic ecosystems existed as far back as the Proterozoic and possibly even earlier.

Assessment of photosymbiosis in ancient reef ecosystems requires recognition of specific characteristics possessed by the calcifying reef organisms. Since the symbionts do not fossilize, the presence of photosymbiosis in fossils is a working hypothesis based on modern symbioses and best confirmed by a set of specific morphologic adaptations and isotopic analyses. Important among these is the thin tissue syndrome—the modification to achieve the “solar panel” effect. Large size and unusual or complex morphology also may indicate photosymbiosis. In the case of colonial organisms such as corals, high levels of corallite integration, where corallites are modified for increasing cooperation, may assist because most colonial photosymbiotic organisms today, such as corals, are exclusively photosymbiotic.

Analysis of organisms and reefs through geologic time permits assessment of the strength of photosymbiosis as a driving force. Reef ecosystems revealing the strongest assessment for photosymbiosis are those of the mid-Paleozoic (Late Ordovician to Devonian), late Paleozoic, early Mesozoic and Neogene. The Early Cambrian archaeocyathan (sponge) reefs indicate photosymbiosis but perhaps with different ancient symbionts such as cyanobacteria, also contained in some modern sponges. Reef ecosystems of the late Paleozoic and early part of the Jurassic indicate the presence of some photosymbiosis. The extinction of many photosymbiotic reef ecosystems during critical intervals of mass extinctions may have been driven by the failure of the symbiosis or demise of the symbionts. Reef gaps in the geologic record reflect the absence of photosymbiosis. The present-day reef crisis involves disturbance of photosymbiosis, and study of future reef declines will benefit by application of data from the fossil record.

The history of Scleractinian corals, richly documented by the fossil record, is one of complex dynamics linked to the dynamics of coral reefs themselves. In spite of all the waxing and waning of marine biodiversity throughout the post-Paleozoic, scleractinians have remained remarkably resilient as a lineage and have traversed two mass extinctions and repeated episodes of global change before becoming the chief builders of modern coral reefs. Understanding this history becomes all the more relevant in face of the current human driven coral reef biodiversity crisis. The advent of molecular phylogenetics has changed our perspective of those dynamics because it has uncovered pervasive morphological convergence in traditionally used taxonomic characters, revealing that the current classification is highly artificial. Taxonomy not only obscures important patterns, but also introduces artifacts into estimates of origination and extinction obtained directly from the fossil record. I present a brief review of the impact of molecular phylogenetics on the current understanding of coral evolution, with emphasis on the recently uncovered phyletic link between photosymbiotic, reef dwelling and azooxanthellate, deepwater coral biota. Then, I discuss the role of molecular-based techniques in a future research agenda of the evolutionary dynamics of the order. The greatest challenge for the future is the re-assessment of morphological characters from a cladistic perspective so that extinct and extant species are integrated in a unified phylogenetic framework, allowing rigorous testing of hypotheses on the fascinating biodiversity dynamics of the order.

Statistical analyses of occurrence data based on collections made from scattered Caribbean sections over the past 20 years indicate that turnover occurred in the Caribbean reef coral fauna between the late Miocene and early Pleistocene. The collections have been identified using standardized procedures, and age-dates assigned using high-resolution chronostratigraphic methods. During turnover, ~80% of the > 100 species and 17 of the 41 genera that were living in the Caribbean during the early Pliocene became extinct, and > 60% of the species now living in the Caribbean originated. Turnover involved increased speciation beginning in the late Miocene and ended with a pulse of extinction in Plio-Pleistocene time. Turnover was preceded by faunal collapse during the late Oligocene to early Miocene, and it was followed by stasis during the late Pleistocene to Recent. During these preceding and succeeding intervals, reef development was at a maximum, although reef coral diversity was relatively low. As a consequence of origination preceding extinction during turnover, most modern Caribbean reef coral species originated before the Plio-Pleistocene peak of extinction, under quite different ecological conditions from those in which they have lived over the past million years. The unusual relationship between origination and extinction may have been caused by changes in productivity associated with emergence of the Central American Isthmus, followed by the onset of Northern Hemisphere glaciation.

During turnover, faunal change was stepwise or gradual. Local assemblages consisted of a mix of extinct and living species, which varied in composition but not in richness. Important reef dominants such as Acropora palmata and A. cervicornis arose in the southern Caribbean and appear to have migrated northward. Faunal change took place in shallow exposed environments, before it occurred in deep protected environments that served as refuges. Plio-Pleistocene extinction was selective for corals with small colonies, and resulted in a faunal shift to the large, fast-growing species that dominate Caribbean reefs today.

Coral reefs have experienced a profound shift in community structure in recent decades, a pattern that contrasts with the apparent constancy of Caribbean reef zonation over the past 2 million years. The abrupt decline in branching Acropora palmata and massive frame-builders like Montastrea annularis in the Caribbean is troubling, and similar patterns have been reported from virtually every ocean. As we ponder the future of coral reefs, we must be mindful that our best monitoring records span perhaps half a century – and those are exceedingly rare. “Pristine” reefs may not have existed since Columbus sailed for the new world, and anthropogenic impacts probably extend even farther back in time.

Despite the vagaries of evolutionary change, taphonomy and time averaging, the geologic record still represents a unique source of important information about the processes that have controlled community structure and reef building in the absence of human influences. The creation of rigid and elevated structures requires calcification rates that are capable of filling the accommodation space created by rising sea level. This has been complicated over the past three to four decades as accelerated sea-level rise has been joined by a suite of stresses that probably slow accretion. Explaining the recent reef decline and posing realistic models of future change will require an understanding of carbonate cycling in the past, the processes that have been involved and a quantitative assessment of how anthropogenic stresses are affecting both.

At the least a look back in time may help to constrain the thresholds at which change might be expected to occur in the future. At best, the context gained from examining the “recent” geological past may provide insights into which possible solutions are most consistent with observed patterns at larger spatial and temporal scales.

Coral reefs are the most biodiverse marine ecosystems on the planet, with at least one quarter of all marine species associated with reefs today. This diversity, which remains very poorly understood, is nevertheless extraordinary when one considers the small proportion of ocean area that is occupied by coral reefs. Networks of competitive and trophic linkages are also exceptionally complex and dense. Reefs have a long fossil record, although extensive reef building comes and goes. In the present, coral reefs sometimes respond dramatically to disturbances, and collapses are not always followed by recoveries. Today, much of this failure to recover appears to stem from the fact that most reefs are chronically stressed by human activities, judging by observations of recovery at exceptional locations where local human activity is minimal. How long reefs can continue to bounce back in the face of warming and acidification remains an open question. Another big uncertainty is how much loss of biodiversity will occur with the inevitable degradation of coral reefs that will continue in most places for the foreseeable future.

As shallow-water reefs decline worldwide, opportunities abound for researchers to expand understanding of carbonate depositional systems. Recognizing the myriad of paradoxes associated with reefs and carbonate research hopefully can stimulate new questions that will assist researchers in understanding paleoenvironmental changes and mass extinction events. Two often counter-intuitive concepts are discussed, first that coral reefs thrive in clear, nutrient-poor waters, except when they don't; and second, that aragonite is energetically efficient for reef-builders to precipitate in tropical waters, except when it isn't. Coordinated studies of carbonate geochemistry with photozoan physiology and calcification will contribute to understanding carbonate sedimentation under environmental change, both in the future and in the geologic record.

Future environmental conditions for coral reefs are rapidly approaching states outside the ranges reefs have experienced for thousands to millions of years. Coral reef ecosystems, once thought to be robust to climate change because of their ability to bounce back after large scale physical impacts, have proven to be sensitive to both temperature rise and ocean acidification. Predicting what coral reefs will look like in the future is not an easy task, and one that is likely to be proven flawed. The discussion presented here is a starting point for those predictions, mostly from the perspective of reef building and ocean acidification.

Little good and a lot of bad come from reef restoration. Reefs damaged by humans and nature are “restored” using unnatural materials as substrata quickly occupied by corals and fish. This is commonly considered good, for seemingly the reef has been returned to a “healthy” state; fish can be caught again and tourists return for the “beautiful reefs”. Restoration, the act of reestablishing a former state, has never been accomplished on a reef; rather reefs have been manipulated to conform to particular human values without regard for the entire reef—its ecology, trophodynamics, hydrodyamics, physical or chemical characteristics of pseudo-substrata, geochemistry, nutrient supply, and even reef aesthetics among a multitude of others. People seemingly cannot leave well enough alone when it comes to reefs that have been noticeably damaged. Yet, that is exactly what reefs need—time without interference. Careful analysis of the total consequences of various methods is required. Reefs evolved over millions of years in one of the harshest environments on earth—the air-water interface. They are well adapted to recover from physical damage of almost any sort. Reefs are not fragile. Thoughtful assistance would help, using materials occurring naturally within reef systems, by involving regional stakeholders in natural processes of restoration, and by stringent protection regulations and agreements. Opportunistic “restoration” by well-meaning, misguided or avaricious people without careful consideration of what really constitutes a reef is a major mistake that will eventually degrade reefs and need restoration itself. Protection of reefs is the best option, followed by letting natural restorative processes take place over long times.