Mangrove forests mitigate coral bleaching under thermal stress from climate change

Anthropogenic marine heatwaves are progressively degrading coral reef ecosystems worldwide via the process of coral bleaching (the expulsion of photosynthetic endosymbionts which reveals the coral skeleton). Corals from mangrove lagoons are hypothesised to increase resistance and resilience to coral bleaching, highlighting these areas as potential natural refuges for corals. Our study, the first conducted at a global-scale, reveals that coral reefs associated with mangrove forests are less likely to bleach under thermal stress, and thus, under scenarios of climate warming. The onset of severe bleaching occurred after 3.6 Degree Heating Weeks (DHW) in mangrove-associated reefs, compared to 2.23 DHW for non-mangrove associated reefs. These findings highlight the critical role of mangrove forests for coral reef persistence under climate change. Accordingly, conservation actions targeting the protection of mangroves are expected to contribute to the resilience and resistance of reef corals from bleaching as marine heatwaves continue to become more common.


23
Coral reefs are continuing to decline globally, primarily driven by anthropogenic heating (1-24 3). Increased sea water temperatures and marine heatwaves are inciting coral bleaching, a 25 process whereby photosynthetic endosymbionts are expelled by the cnidarian host (4-8). 26 Continued coral bleaching is increasingly leading to mortality of coral colonies over entire 27 reefs (1, 9, 10), critically threatening the key ecosystem services provided by coral reefs 28 which host ~25% of marine biodiversity and support the livelihoods of over half a billion 29 people (11, 12). 30 Given the complex interactions between the environmental conditions at local scales with the 31 coral holobiont (i.e. the coral polyp, endosymbionts and microbiota) (13), the degrees of 32 bleaching from different coral taxa vary across different regions of the world, creating spatial 33 patterns of bleaching (14-16). In addition, coral bleaching can often be a sub-lethal process to 34 aid individuals for withstanding environmental stress (7). However, given the increased 35 frequency and intensity of marine heatwaves which induces bleaching over mass scales, 36 leading to high levels of mortality(1, 9, 10), functional and taxonomic homogenisation (i.e. 37 reef flattening) of entire coral reef ecosystems is occurring (2, 9, 10). Therefore, elucidating 38 patterns of spatial variation in coral susceptibility to bleaching, and the underlying 39 mechanisms responsible, is critical for the identification of factors and ecological settings 40 with the potential to mitigate the decline of coral reefs, which can in turn translate into the 41 design of conservation actions.

42
In particular, mangrove lagoons have been suggested to act as potential refuges for coral compositions allowing them to survive in extreme environments which may represent future 50 climate conditions (22, 23). Therefore, coral reefs associated with mangrove forests may be 51 more resilient or resistant to the synergistic effects of factors that trigger bleaching in exposed 52 marine areas as they have been preconditioned to environmental stress (18, 23). 53 Concurrently, these corals may proliferate further out onto coral reef environments, 54 displacing corals more susceptible to thermal stress (24). As rising global temperatures 55 continue to increase marine heatwaves (25) and thus bleaching intensity (1), taxonomic 56 displacement of corals with low thermal tolerance by preconditioned coral species from 57 chronic stress becomes more likely (16, 24, 26). 58 Despite emerging evidence of mangrove-associated corals showing higher resilience and 59 resistance to thermal stress at local-scales (18, 23), the generality of the impact of mangroves 60 as environments that buffer bleaching remains unknown as global-scale analyses are lacking.

61
In this study, we implement the first global-scale analyses ( Fig. 1)

83
Bleaching responses 84 The Bayesian hierarchical ordinal regression model shows that the interaction between DHW 85 and mangrove area reduce the probability of coral bleaching globally (Fig. 2). 86 Meanwhile, increasing DHW alone increased the probability of coral bleaching, as did an 87 increase in mangrove area alone. Therefore, under elevated thermal stress, coral bleaching is 88 less likely at coral reef sites associated with mangrove forests. When comparing the raw DHW values where the onset of bleaching occurred, our analyses 105 revealed no significant difference between mangrove and non-mangrove associated sites.

106
Furthermore, the onset of bleaching for the mild and moderate severity bleaching categories 107 did not occur at significantly different DHW thresholds between mangrove and non-108 mangrove sites. However, the onset of severe bleaching occurred after 3.6 DHWs at 109 mangrove sites, which was significantly higher (Pr(>Χ 2 ) < 0.05) than the 2.23 DHWs 110 required to induce severe bleaching at non-mangrove sites (Fig. 3). Therefore, mangrove sites 111 are more resistant to severe bleaching under elevated thermal stress from climate change.  Table S3. Our study provides the first global-scale analysis of the role that mangroves contribute to 123 mitigate the impact of anthropogenic warming on coral bleaching. The evidence we present 124 reveals a reduction in bleaching probability at mangrove associated sites when DHW are 125 elevated (Fig. 2). This reduction in bleaching probability is likely to be the result of the 126 significantly higher DHWs threshold which triggers the onset of severe bleaching at 127 mangrove sites compared to the non-mangrove sites (Fig. 3). Therefore, it is likely that corals The global standard to determine thermal anomalies are degree heating weeks (DHW) which 179 quantify the temperature exposure above a 1°C for the local mean climatic temperature at that 180 period in time over the last 12 weeks (2, 42, 43). The DHWs data were collected to assess the 181 interaction effect between high thermal stress and mangrove area on coral bleaching severity.

182
Weekly DHW values were obtained from the Coral Reef Temperature Anomaly Database subsequently achieved using the autojags function from the "R2jags" package (52).

233
Convergence of the model was determined by a Gelman-Rubin statistical value of 1.1 (53).

234
Model coefficients were considered to have a significant effect when the 89% credible 235 intervals (CI) did not pass zero (54), which is mathematically more stable than using the 236 frequentist 95% threshold when conducting Bayesian analysis with less than 10,000 samples 237 (55, 56). 238 We ran sensitivity analysis using the Bayesian model described above to determine which 239 distance showed the highest effect. Distance refers to the radius from which the entire area is 240 sampled to determine the total mangrove area around the Reef Check survey. We used a 241 doubling sequence from 1km to 4km (i.e., 1km x 2, 2km x 2) and determined the 4km 242 distance showed the strongest effect for reducing bleaching under thermal stress (Table 1).

243
Differences in bleaching severity thresholds for DHWs were statistically analysed using 244 likelihood ratio tests on the DHW gamma distributions (16).