Emissions scenarios

Temperature projection

CO₂ concentration projection

DHM Constant

DHM Adaptation

DHM Saturation

Coral reefs have been shown to be sensitive to elevated sea temperatures that can trigger a breakdown of the symbiosis between corals and the dinoflagellate symbionts residing in coral tissue (Hoegh-Guldberg and Smith, 1989). Corals do not only derive most of their energy from this close symbiotic relationship with this special type of microalgae but also most of their famous color. Thus, when the vital symbiosis between corals and algae breaks down corals “bleach” or turn pale.

First mass coral bleaching and mortality events have been observed worldwide since the early 1980s and have affected reefs at regional scales (e.g. Eakin et al., 2010). While there are potential other localized triggers of bleaching, the observed large scale mass events have been related to unusual high temperatures (Hoegh-Guldberg, 1999). During 1997 and 1998, mass coral bleaching events affected coral reefs in almost every part of the world and caused mortality of approximately 16% of reef-building corals (Wilkinson, 2004). Although corals can re-establish themselves after mass bleaching events, in some cases it takes one to two decades for the ecosystem to return to the pre-bleaching state (Baker et al., 2007). An increase in the frequency and severity of mass coral bleaching could overwhelm the ability of coral reefs to recover between events. If this happens, coral reef ecosystems would shift towards systems that are dominated by other organisms such as cyanobacteria and algae.

Here, we use a well-established temperature indicator that basically measures the exposure of corals to particularly high temperatures outside the local ranges they are adjusted to (Hoegh-Guldberg, 1999). The indicator (Degree Heating Weeks, DHW) has been shown to allow for predictions of bleaching based on observed sea surface temperatures (http://coralreefwatch.noaa.gov) and has been adjusted for monthly mean temperatures as routinely provided by Global Climate Models (Degree Heating Month, Donner et al. 2007; Donner 2009).

In the default case a DHM value above a threshold of 2°C · month is considered as an indicator for bleaching. Assuming that coral reefs require at least 5 years to recover from such an event a reef cell is counted as “under risk of severe degradation” if the local frequency of projected bleaching events exceeds two per decade.

The coral bleaching model provides the fraction of coral reef locations that are under risk of severe degradation (bleaching events are more frequent than recovering time) in terms of global mean air temperature for a range of global climate models from the CMIP3 repository. Given this relationship between global mean temperature change and the fraction of affected reef cells impact projections are translated to other global emission scenarios than the ones covered by CMIP3 by using the associated global mean temperature pathways (and their uncertainty ranges) as provided by the simple climate model MAGICC6. Bleaching thresholds are varied according to three different scenarios: a) constant DHM bleaching threshold of 2°C · month, b) a thermal adaptation case (DHM bleaching threshold increases linearly from 2°C · month to 8°C · month by 2100) and c) a sensitivity case assuming that the bleaching threshold declines with ocean acidificationincreasing the threat to coral reef ecosystem. A more detailed description can be found in Frieler et al. (2012).

Note that the degree heating months are calculated until 4°C of warming leading to cut-offs for high temperature scenarios. It is important to note that our analysis examines the general effects of rising temperatures on coral reefs at different levels of global warming at the coarse resolution of global climate models. This modeling exercise is not intended or appropriate, for projecting the precise future of specific coral species or their extinction. The results show that, if coral reefs were to experience severe bleaching at DHM > 2°C · month, then the majority would be exposed to frequent events (which produce big impacts today) even at the modest increase in average global temperature of 1.5°C.

The spatial resolution of the temperature does not allow us to draw specific conclusions at the fine scale or about individual species. A number of physical factors, including irradiance (Takahashi et al., 2004; Brown et al., 2002), historical temperature variability (Carilli et al., 2012; Ainsworth et al., 2016), and local fine-scale temperature variability (e.g. deep reef refugia) can influence the response of specific corals or coral communities to heat stress but may be of minor importance for long term changes on global scale.