The coral diversity recorded in this work represents most of the species present throughout the Caribbean. About 50 scleractinian coral species are identified in research realized in the wider Caribbean (Huang and Roy 2015; McWilliam et al. 2018). We found a substantial loss of coral diversity in the period analyzed. Both species richness and diversity considering abundance (Hill’s N1 and N2 of diversity) decreased in the MAR coral reefs. Additionally, we note that most of the species show a considerable reduction in the frequency of occurrence along the MAR, which could express a simplification of the coral composition in the region, leaving a lower diversity in most of the coral reefs. It is important to note that in this work we used an intercept point method, which may have a lower detection of coral species than other methods that consider a larger area, such as belt-transects. However, the accumulation curve of the analyzed periods shows that the diversity seems to be well represented at the MAR scale (Supplementary Fig. 4), but it is possible that the most cryptic species have not been correctly represented. We recommend that future work consider other monitoring methods that present a better identification, especially of the rarest species, to better capture the loss of diversity in these ecosystems. Losing diversity of reef-building corals has been observed in different regions of the world, such as reefs in the Indo-Pacific and Florida (Vega-Rodriguez et al. 2015; Sheppard et al. 2020). These losses have even led to local or regional extinction of several coral species (Glynn and De Weerdt 1991; Cramer et al. 2020), which can also reduce reef ecosystem functionality and processes (Bellwood et al. 2004; Brandl et al. 2019; Sheppard et al. 2020). Our results underline the importance of considering coral diversity as a potential indicator of the current degradation of coral reefs. Most reef assessments focus on aggregated coral cover as the main indicator of coral reef condition (Gardner et al. 2003; Wilkinson 2008; Souter et al. 2020). Still, the potential loss of diversity in the medium term and the repercussions this may have on the ecosystem are being overlooked.
The major drivers of the change in coral diversity were initial diversity and the number of annual bleaching risk events, which were related to a greater loss of diversity during the period analyzed. First, the most diverse coral reefs were those most likely to lose diversity. Different authors have suggested that diversity is a good indicator of resilience to disturbances with coral reefs, based on the premise that greater diversity tends to offer a wider variety of functional responses (Nyström et al. 2008; Roff and Mumby 2012; McClanahan et al. 2012; Mora et al. 2016). Initial diversity is fundamental in the trajectory of temporal change in diversity that a community may have because of exposure to different disturbances; different levels of initial diversity may even interact with different degrees of disturbance and cause different trajectories of change (Randall Hughes et al. 2007). It has been hypothesized that communities with a low recruitment rate, high exposure to disturbance, and high initial diversity are communities with a very high decline in diversity because of ecological disturbance events (Randall Hughes et al. 2007). Our results highlight the importance of protecting the most diverse reefs, as they may be potentially vulnerable to future disturbances, because in the case of the MAR, high coral diversity is not synonymous with resistance capacity.
Heat stress, as represented by the number of annual bleaching risk events (years with DHW > 4°C - weeks) was an important driver of coral diversity loss. Exposure to heat stress affects corals in diverse ways. First, this stressor causes bleaching events and often mass mortality in different reefs worldwide (Baker et al. 2008; Eakin et al. 2019). Heat stress causes the loss of coral cover when exposure is very high, for example, in the Caribbean when reefs are exposed to events of a magnitude greater than 8°C – weeks (Eakin et al. 2010). Besides, exposure to extreme heat has also been associated with loss of diversity in different reefs worldwide (Glynn and De Weerdt 1991; Vega-Rodriguez et al. 2015; Sheppard et al. 2020). It is worth mentioning that the MAR was exposed to high heat stress during the period analyzed. In fact, from 2015 to 2017 the reefs of this region were exposed to unprecedented heat stress events, emphasizing that much of the Mesoamerican reef system experienced its maximum exposure during this period (Muñiz-Castillo et al. 2019). The context of exposure to extreme oceanic heat events during the analyzed period initiated to be determinant in the change of coral diversity because the high frequency of these heat stress events may be one of the main drivers of degradation in the reefs of the MAR. This result highlights the importance of establishing management and conservation measures that consider mitigation strategies for the effects of climate change on coral diversity in the region.
On the other hand, wind recorded during hurricane events was positively associated with the change in diversity. Distinct reasons can explain this. The first is that the corals were not considerably exposed to hurricanes and storms during the analyzed period (2010 to 2018). Most coral reefs sampled presented between 0 to 3 hurricanes within a radius of 30 km. Only ten of the reefs monitored had a maximum wind speed during storms of over 60 knots, with most of the reefs exposed to intermediate levels of wind exposure during storms and hurricanes, recognized as non-severe degrees of exposure for the condition of coral reefs (Gardner et al. 2005). These reefs suffered from intermediate magnitude disturbances to hurricanes, and therefore no loss of diversity occurred. Connell’s classic intermediate disturbance theory (Connell 1978) and field research on coral reefs indicate that intermittent and intermediate magnitude exposure to hurricanes allows maintaining a greater diversity of corals (Huston 1985; Rogers 1993).
Recent analyses highlight the null negative effect of hurricanes and storms on Caribbean coral reefs during the last decade. This null impact is mainly because many of these reefs present a high state of degradation and only the most resistant corals to this type of extreme event are present in the region's reefs (Mudge and Bruno 2021). However, it is recognized that high magnitude (strong winds) of exposure to hurricanes can be an important driver in the degradation of Caribbean coral reefs (Mumby 1999; Gardner et al. 2005; Edmunds 2019; Reguero et al. 2019). Besides being an extreme event of great relevance in the socio-economic vulnerability of the region (Reguero et al. 2019). With MAR reefs, it is recognized that the effect of hurricanes may not be as severe as that of heat stress, although the combined effect of these two could have a considerable negative effect on these ecosystems (Edwards et al. 2011) as recorded in Belize due to the 1998 bleaching event and hurricane Mitch (McField 2001). Cumulative exposure to hurricanes and storms can cause different effects, ranging from direct removal and death of corals from direct wave action (Gardner et al. 2005; Madin and Connolly 2006; Edmunds 2019) to decreasing coral recruitment (Mumby 1999). Hurricanes can have a negative effect on corals that can be seen even in long-term periods, causing a decrease in coral cover even several years after the impact occurs (Gardner et al. 2005). These contrasting results suggest that if exposure to hurricanes is intermediate the effect on diversity may be positive (increase during the period considered). Yet, if exposure is remarkably high over a historical period this may affect reefs even in the long-term causing a decrease in coral diversity.
Other indicators linked to potential local impact because of sediments or nutrients were also highlighted as relevant drivers of change in diversity. With these indicators, contrasting results were observed. The water turbidity proxy (Kd-490) showed a positive relationship with coral diversity, with the more turbid reefs experiencing an increase in coral diversity. This result may seem contradictory, but the explanation for this pattern is that few reefs in the MAR have remarkably high turbidity. During the entire period analyzed, less than 25% of the reefs showed a value higher than 0.30 m− 1, which expresses relatively low exposure to turbidity in the reefs of the region (Chollett et al., 2012b; Rivera-Sosa et al., 2018; Geiger et al., 2021). Although water turbidity can be an important stressor on corals (Fabricius 2005), there is also some research indicating that “turbid water reefs” have been more protected from solar radiation and small scale heat stress (Sully and van Woesik 2020). We suggest it is also possible that these “toughened” reefs, normally exposed to fluvial runoff may have been conditioned to this stress, and that the more turbid surface layer runoff can act to insulate these corals from heat and solar radiative stress because the surface layer infrequently interacts with the seafloor. Several of the more turbid reefs (located in the coastal zone of Guatemala and Honduras) have some of the highest coral cover values in the MAR. (Mcfield et al. 2018; Rivera-Sosa et al. 2018). Therefore, this indicator does not correctly reflect the potential nutrient input because of human activities on the mainland. Considering metrics that represent nutrient, agrochemicals, or sediment proxies on reefs remains a critical area of opportunity in the MAR region. To date, very few regional-scale efforts have been able to model or represent the influence that land-based activities may have on the waters surrounding coral reefs, most notably the report by the World Resources Institute (Burke and Sugg 2006). Future efforts should focus on estimating or modeling the spatial distribution and the temporal variation of nutrients and chemical substances associated with human pollution, as this would be a great step forward in understanding the potential direct impact of human activities in the watershed or in coastal areas near coral reefs.
In this work, an indirect indicator of nutrient input was the change in macroalgae cover, observing a negative relationship with the shift in diversity (increase in macroalgae = decrease in diversity). It is recognized that with the MAR, the phase shift occurred in the coral reefs, in which the macroalgae cover is gaining ground over corals, is related to the constant input of nutrients because of coastal gentrification and human activities in the basin (Suchley et al. 2016; Martínez-Rendis et al. 2016; Arias-González et al. 2017). However, the change in macroalgal cover is also highly related to herbivory control, because if there is a loss of key herbivores it is possible that there is less top-down control (Bellwood et al. 2004; McManus and Polsenberg 2004; Brandl et al. 2019; Bruno et al. 2019), so we cannot be sure that the increase in macroalgae is only a cause of increased nutrient input.
In the results obtained in this work, the human population density and the change in the human population density were not selected as relevant drivers. Work on a global scale has shown that human population density near reefs is not directly related to the condition of coral reefs (Bruno and Valdivia 2016). In macroecological studies, indicators of human population density and human activities for potential local impact do not currently take nutrient inputs into account. The existing indicators are often only available at very large spatial scales, providing data at the country or regional level (Halpern et al. 2019; Cramer et al. 2020). This lack of consistent and frequent spatiotemporal data makes it difficult to analyze the relationship between local human activities and change in coral reef conditions or diversity, leaving a gap about the effect of direct impact because of nutrient inputs. Future research needs to develop more sensitive metrics that better represent nutrient inputs associated with human presence activities (Pawlik et al. 2016; Delevaux et al. 2018; Wolff et al. 2018; Bruno et al. 2019).
The results obtained showed a negative relationship between the size of the MPAs and the change in coral diversity. This negative relationship reflects that MPAs size is not an indicator of the corals' level of protection or conservation. Sometimes the size does not matter, as in the case of MPAs, the level of compliance and the time since the area was designated seem to be more critical (Bonaldo et al. 2017; Cortés-Useche et al. 2019). However, in some cases the size of the MPAs does not have a considerable effect or it is even possible to observe a positive effect in large MPAs (Halpern 2003). It is important to mention that, with this analysis, all management or conservation units were considered, including Ramsar areas and other designations. With the MAR, information is currently being generated and available with a better designation of the degree of protection by the Healthy Reefs Initiative, which could be incorporated into future work. This information would make it possible to evaluate the state of marine ecosystems in the region, as has been done in some regions, such as the Mexican Caribbean (Suchley and Alvarez-Filip 2018). Meanwhile, our results reflect that large MPAs appear to be ineffective in protecting coral diversity, especially because these large areas tend to have low operational capacity.