Predicting and responding to the effects of climate change is critical if we wish to preserve marine life in the oceans. The global effects of climate change are becoming glaringly apparent on coral reefs, where repeated thermal anomalies are forcing temperatures beyond the physiological tolerance range of many coral species. These stress events are causing widespread coral bleaching and mortality (Loya et al., 2001; Baker et al., 2008). Yet, predicting where corals are less likely to bleach, and more likely to survive thermal-stress, is an important scientific endeavor that might redirect ocean conservation efforts at both local and regional scales (Cacciapaglia and van Woesik, 2015). Here we use an 11-year dataset (from 2005 to 2015) to examine the spatial relationships between coral bleaching, coral disease, and environmental conditions at 2,398 study sites along the Florida reef tract.

Coral bleaching is a combination of irradiance and temperature stress (Takahashi et al., 2004). High seasonal irradiance causes photoinhibition of the coral symbionts, from which they can recover, unless water temperatures are high. When irradiance and water temperatures are both high, the symbiont's photosystems are compromised, and nighttime recovery from daytime photoinhibition is minimal (Warner et al., 1999; Takahashi et al., 2004). Weeks of high-water temperatures leads to chronic photoinhibition, or bleaching, that can lead to the expulsion of coral symbionts. This bleaching can be temporary and non-fatal for some coral species (Goreau, 1964; Loya et al., 2001), or fatal for other coral species that rely heavily on their symbionts as a food source (Grottoli et al., 2006).

Coral bleaching can also lower the tolerance of corals to pathogens, and bleaching can lead to disease (Muller et al., 2008; Brandt and McManus, 2009; Randall et al., 2014). The causative agents of most coral diseases are still unknown, but laboratory and field studies have shown an influence from a variety of stressors in addition to water temperatures, including population density (Bruno et al., 2007), nutrient enrichment (Bruno et al., 2003; Voss and Richardson, 2006), and irradiance (Kuta and Richardson, 2002). It is also likely that physiological and environmental parameters interact to influence the virulence of the pathogens, and the susceptibility of coral hosts (Randall and van Woesik, 2015).

By contrast, reducing irradiance during temperature-stress events relieves stress on the symbiont's photosystem, buffers corals from temperature stress, and reduces the likelihood of coral bleaching (Warner et al., 1999; Takahashi et al., 2004). Such reductions in irradiance are common on deep reef slopes (Smith et al., 2014), or nearshore, where turbidity and chlorophyll-a concentrations are high (Wagner et al., 2010; van Woesik et al., 2012). Yet, reef corals generally prefer low-nutrient waters (Tomascik and Sander, 1987), and nearshore environments support high nutrient concentrations, which in combination with high temperatures are detrimental to reef corals (Wooldridge and Done, 2009; Wagner et al., 2010; Wiedenmann et al., 2013). Still, if shading in nearshore, turbid environments provides protection to corals under thermal-stress events, then the corals that can tolerate those nearshore conditions may be selected for when thermal-stress events become frequent. We hypothesize that although shading by high turbidity and high organics are less than optimal for reef-building corals, these conditions may be physiologically beneficial for corals under thermal stress. Indeed, the benefits of shading under extreme thermal stress may override the costs of living in these less suitable nearshore environments, with high thermal variation, low irradiance, and high nutrients. We question whether global warming is progressively driving reef corals away from clear, oligotrophic waters, to more nearshore, turbid habitats.

The present study uses an extensive dataset collected between 2005 and 2015 (http://www.frrp.org) at 2,398 sites along the Florida reef tract to examine the spatial distribution of scleractinian corals and the environmental variables that cause and relieve stress. Specifically, the objectives of this study were to: (i) quantify the water-quality characteristics of the study sites, particularly temperature and chlorophyll-a concentrations, (ii) determine the spatial extent of bleaching and coral disease, and (iii) examine interactions between coral bleaching, coral disease, and water quality.

The analysis of this extensive dataset from the Florida reef tract suggests that thermal-stress events are responsible for coral bleaching and associated coral disease, but water clarity also plays an important role. We found that throughout the decade of observation, the clearer the water the more likely it was that corals bleached. Corals in general, and the two major reef-building corals, Acropora cervicornis and Orbicella annularis, in particular, occurred in greater numbers in habitats with low chlorophyll-a concentrations. Yet coral colonies in these low chlorophyll-a habitats showed more extensive bleaching during thermal stress events than coral colonies in high chlorophyll-a habitats. It is most likely that the high concentrations of chlorophyll-a acted as a thermal refuge for coral colonies because of the shading they provided to the corals.

Physiological studies have confirmed the benefits of reducing irradiance during temperature-stress events. Shading reduces stress on the symbiont's photosystem, which in turn reduces the likelihood of coral bleaching (Warner et al., 1999; Takahashi et al., 2004). Notwithstanding the obviously adverse effects that poor-water quality, with high levels of pollutants and high nutrient concentrations, has on corals (Wooldridge and Done, 2009; Wagner et al., 2010; Wiedenmann et al., 2013), some shading provided by high-primary productivity and high turbidity can benefit corals during thermal-stress events (Cacciapaglia and van Woesik, 2016). Indeed, since coral bleaching is essentially extreme photoinhibition, and high temperatures make that photoinhibition worse (Warner et al., 1999; Takahashi et al., 2004), high productivity and high turbidity during high temperature events should effectively reduce the probability of photoinhibition and coral bleaching.

Similarly, the results showed that corals in habitats with high water-column productivity had a lower prevalence of disease than elsewhere. Previously, Lesser et al. (2007) suggested that reducing chronic photoinhibition and bleaching reduces the likelihood of coral disease. Subsequent field studies have validated these observations. For example, Muller et al. (2008) showed that bleached Acropora palmata colonies were more likely to suffer disease than coral colonies that did not bleach.

Still, reducing irradiance can be detrimental to the calcification process, causing reductions in coral growth rates (Tomascik and Sander, 1987). Moreover, high sedimentation environments can deplete coral resources and increase their susceptibility to disease (Sheridan et al., 2014). Therefore, reducing irradiance, through primary productivity in the water column and high turbidity, can have beneficial effects up to a point, beyond which the costs may override the benefits for reef corals. Indeed, high and persistent productivity and extremely high turbidity can reduce the capacity of corals to build reefs entirely, because the photic zone becomes so narrow that there is insufficient irradiance for photosynthesis and calcification (Tomascik et al., 1993; Kleypas, 1996; Toth et al., 2012). Previously, Kleypas et al. (1999) showed that reef building is unlikely in habitats with an average irradiance of <250 μE m⁻² s⁻¹ (at a depth of 3 m). Therefore, locating optimal niches, where corals can survive through thermal-stress events, should be a research priority, to not only detect natural refuges, but also for the rapidly burgeoning coral restoration endeavor that is expanding throughout Florida and the Caribbean.

Ocean temperatures are clearly increasing globally (Hansen et al., 2010; IPCC, 2013), and as the oceans continue to warm, thermal anomalies will most likely continue to cause coral bleaching and subsequent diseases (Hoegh-Guldberg, 1999; Harvell et al., 2002; Donner et al., 2005; Hoegh-Guldberg et al., 2007; Muller and van Woesik, 2012). Some researchers question the capacity of corals to adapt to rapid climate change (Hoegh-Guldberg, 2006; Frieler et al., 2013), whereas other researchers suggest that adaptive radiations and directional selection of thermally tolerant genotypes are likely (Thompson, 1998; Hoffmann and Sgro, 2011; Guest et al., 2012; Poloczanska et al., 2016). The present study suggests an additional, more nuanced, effect of ocean warming—causing geographical shifts of reef corals toward more turbid environments.

Adaptation, however, will occur only under persistent selective pressure. In the past, there may have been little selective pressure for corals to live in habitats with consistently high turbidity and high chlorophyll-a-concentrations, since reef-building corals grow best under moderate irradiance (Done, 1982). Most recently however that selective pressure may have increased. The almost annual reoccurrence of thermal-stress events and coral bleaching in the Florida reef tract, shown by this study, may be selecting for coral genotypes that can live in naturally shaded environments, with high primary productivity and turbidity. Corals in these shaded habitats are less likely to suffer thermal stress than elsewhere. Therefore, in a rapidly warming ocean, there is likely to be a fitness advantage for corals that can live in habitats with less than optimal irradiance, because those same environments shield corals from short-term temperature-stress events. This study suggests that directional selection in a warming ocean may favor corals that are able to tolerate inshore environments with high turbidity and productivity. Because of reduced thermal stress in those shaded environments, selection for colonies in these habitats may provide the genetic pool of corals needed to survive through climate change.

RvW designed the work, wrote the R code to conduct analysis, and wrote the manuscript. KM conducted the kriging analysis, wrote some of the R code, and edited the manuscript.