Anthropogenic climate change poses a significant threat to coral reefs, pushing them towards critical climate tipping points (CTPs). The Intergovernmental Panel on Climate Change (IPCC) and other syntheses suggest that limiting global warming to 1.5 °C is insufficient to protect most coral reefs, largely based on a limited subset of models using 'excess heat' threshold methodologies. Warm-water coral reefs are highly biodiverse ecosystems that provide crucial ecosystem services, supporting global food security, coastal protection, economic revenue, and livelihoods for up to one billion people. The near-synchronous nature of historical bleaching events underscores the vulnerability of these ecosystems to even modest levels of warming. Current CTP assessments, which identify a CTP of 1.5-2 °C for coral reefs, heavily rely on 'excess heat' models, despite the existence of alternative methodologies. This review addresses the need for a comprehensive evaluation of existing projections to better understand the uncertainty surrounding coral reef futures under climate change.

The review examined 79 studies published between 1999 and 2023 that modeled coral reef responses to climate change. These studies employed five main methodologies: 'excess heat' threshold models, population dynamic models, species distribution models (SDMs), ecological-evolutionary models, and projective meta-analyses. 'Excess heat' models, which use thermal stress thresholds to predict bleaching events, constituted 32% of the studies but attracted 68% of the citations, underscoring their disproportionate influence. Population dynamic models (23%), SDMs (23%), and ecological-evolutionary models (12%) provided more mechanistic insights but received fewer citations. Meta-analyses (5%) synthesized existing experimental data but were limited by the short-term nature of the underlying experiments. The review also highlighted several other emerging approaches, such as the IUCN Red List of Ecosystems.

The authors conducted a systematic review using the PRISMA guidelines. They searched the Web of Science database for articles projecting coral reef responses to climate change, applying specific inclusion criteria. Data extraction focused on methodology, geographic scale, model inputs, and key assumptions. The models were categorized into five groups: 'excess heat' threshold models, population dynamic models, species distribution models (SDMs), ecological-evolutionary models, and meta-analyses. The studies were further classified based on their approach to modeling heat stress: thermal threshold techniques or continuous variable techniques. An exploratory meta-analysis was performed on a subset of studies that provided comparable data on key coral reef metrics (percentage of reef cells at risk, percentage of habitat change, and percentage of coral cover change) under different warming scenarios. Hedges' g effect sizes and variances were calculated using the metafor package in R. The analysis considered the influence of different methodological approaches, including the choice of metric, model type, and parameterization, and geographic scale.

The review revealed substantial discrepancies in projected impacts on coral reefs across different methodologies. Studies using 'excess heat' threshold techniques tended to project more severe negative impacts than studies employing alternative approaches. More than half of the studies (53%) employed thermal threshold techniques, primarily degree heating weeks (DHWs) or months (DHMs), to model heat stress. However, the efficacy of these metrics varies, and they may not capture other relevant heatwave characteristics. The meta-analysis, though limited by data availability (only 39 scenarios from 8 studies), demonstrated that studies relying on 'excess heat' thresholds produced more severe effect sizes, especially in the 2-4 °C and >4 °C warming scenarios. This contrasts with the IPCC's AR6 projections, which relied heavily on studies using 'excess heat' models and predicted substantial coral reef declines even under 1.5 °C warming. The study further highlighted geographic biases in modeling efforts, with some regions receiving more attention than others. The majority of studies focused on warming alone or in combination with a single additional stressor, neglecting the complexities of interacting pressures on coral reefs. Finally, the lack of standardized reporting practices and common output metrics hinders the synthesis of existing projections. The study found considerable spatial heterogeneity, with regional and global models presenting discrepancies.

The findings highlight the critical need for improved coordination in modeling coral reef futures. The reliance on a limited subset of methodologies, particularly 'excess heat' models, may overestimate the severity of projected impacts. The authors advocate for a multi-model ensemble approach, similar to that used in climate science, to better quantify uncertainty and generate probabilistic projections. This requires improved coordination among modelers to establish common output metrics and emissions scenarios. The selection of ecologically relevant variables and spatial scales is crucial for informing effective conservation strategies. Future modeling efforts should consider integrating a wider range of stressors and incorporating management and intervention scenarios, particularly those related to reducing local stressors (pollution and fishing pressure) and restoration efforts. Improved reporting standards are needed to enhance the transparency and comparability of model outputs.

This review emphasizes the significant uncertainty surrounding projections of coral reef futures under climate change. The reliance on a small subset of modeling methodologies may overestimate the severity of impacts. A multi-model ensemble approach is proposed, requiring better coordination, standardized metrics, and the integration of multiple stressors and intervention scenarios. Improved reporting standards are essential for future syntheses of coral reef projections.

The exploratory meta-analysis was limited by the availability of comparable data across studies. The limited number of studies included in the meta-analysis restricts the generalizability of the findings. Geographic biases in modeling efforts were identified, with some regions receiving less attention than others. The study primarily focused on the impacts of warming, neglecting the full complexity of interacting pressures on coral reefs.