Effect of SRS on modeled coral persistence
The effect of SRS on coral cover during warming within the Indo-Pacific metapopulation model (Matz, Treml, and Haller 2020) was surprisingly subtle (Fig. 2 A-C): diminishing coral cover was only observed at the most severe SRS setting (Fig. 2 D), where only 0.001-th fraction of all individuals were allowed to reproduce. The reason for this lack of effect is that SRS did not affect adaptive genetic diversity much (Fig. 2 E), likely because the model with overlapping generations allowed for accumulation of genetic diversity in populations via multiple rounds of recruitment. Yet, there was a clear effect of SRS on year-to-year variation in the coral cover: even mild SRS of 0.1 already amplifies the coral cover volatility compared to the base model (Fig. 2 F). This increase is most pronounced at smaller reefs (Fig. 2 G), which is not surprising since they should be more prone to demographic fluctuations of any kind. Since coral cover in the model depends on the match between genetic thermal optimum of the local population and the environmental setting, most likely SRS increased the chance of better-than-average match in some years and worse-than-average match in other years.
While SRS in nature can be well below 0.001 (Dennis Hedgecock and Pudovkin 2011), it should be noted that the sizes of our modeled subpopulations were much smaller than in nature: the smallest modeled subpopulations had the carrying capacity of only 100. We therefore believe that the emerging effect of severe SRS was predominantly due to the fact that many subpopulations, with carrying capacity near or below 1/SRS, did not reproduce at all except once in several generations. This would have led to their diminishing size due to yearly mortality and thus less reproduction overall than in the base model. This is a rather unrealistic situation since natural census sizes of coral populations are well above 1000, although there are notable exceptions such as the pillar coral Dendrogyra cylindrus in the Florida Keys, (Neely et al. 2021). Overall, SRS does not seem to hurt coral adaptive capacity too much, at least not within our model. That said, SRS might matter more in more complex models, for example, in models involving adaptation to multiple stressors or adaptation to local environmental factors that do not change during warming, which should be explored in the future.