Patterns of Spatial Covariation in Herbivore Functions on Coral Reefs: Implications for Reef Resilience

Patterns of Spatial Covariation in Herbivore Functions on Coral Reefs: Implications for Reef Resilience | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Patterns of Spatial Covariation in Herbivore Functions on Coral Reefs: Implications for Reef Resilience Dana T. Cook, Sally J. Holbrook, Russell J. Schmitt This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5160728/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 07 Jan, 2025 Read the published version in Scientific Reports → Version 1 posted 12 You are reading this latest preprint version Abstract Persistent shifts to undesired ecological states, such as shifts from coral to macroalgae, are becoming more common. This highlights the need to understand processes that can help restore affected ecosystems. Herbivory on coral reefs is widely recognized as a key interaction that can keep macroalgae from outcompeting coral. Most attention has been on the role ‘grazing’ herbivores play in preventing the establishment of macroalgae, while less research has focused on the role of ‘browsers’ in extirpating macroalgae. Here we explored patterns, environmental correlates and state shift consequences of spatial co-variation in grazing and browsing functions of herbivorous fishes. Grazing and browsing rates were not highly correlated across 20 lagoon sites in Moorea, French Polynesia, but did cluster into 3 (of 4) combinations of high and low consumption rates (no site had low grazing but high browsing). Consumption rates were not correlated with grazer or browser fish biomass, but both were predicted by specific environmental variables. Experiments revealed that reversibility of a macroalgal state shift was strongly related to spatial variation in browsing intensity. Our findings provide insights and simple diagnostic tools regarding heterogeneity in top-down forcing that influences the vulnerability to and reversibility of shifts to macroalgae on coral reefs. Biological sciences/Ecology/Community ecology Earth and environmental sciences/Ecology/Community ecology herbivory reversal of state shifts grazing browsing community recovery vulnerability and reversibility Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction A major focus in ecology is to understand the complex dynamics of ecosystems that simultaneously are being buffeted by rapidly changing disturbance regimes and by more slowly changing environmental drivers. Of considerable interest are abrupt transitions between qualitatively different ecosystem states (Scheffer and Carpenter 2003 , Conversi et al. 2015 ), which can be challenging to anticipate and problematic to reverse (Mumby et al. 2013 , Schmitt et al. 2019 ). The consequences to society of such shifts can be profound if an alternative community state provides either fewer or qualitatively different ecosystem services (Suding and Hobbs 2009 , Mumby et al. 2013 ). This adds urgency to the need to better understand the processes that influence transitions between alternative, self-reinforcing states (regime shifts). In general, shifts in state can occur when the magnitude of a perturbation overwhelms the internal resilience of a system, which itself can be eroded when key biological processes are altered by human activities or natural events. Resolving links between state shifts, biological processes, and ecosystem resilience can inform management strategies intended to maintain desired ecosystem states during this era of unprecedented human impact on the environment (Mcleod et al. 2019 ). Coral reefs are being degraded worldwide by local and global stressors and acute disturbances (Hughes et al. 2007 , Sully et al. 2019 ). Declines in coral cover caused by storms, predator or disease outbreaks, or bleaching frequently have been accompanied by concomitant increases in macroalgae (coral-algae ‘phase shifts’) (Roff and Mumby 2012 , Adam et al. 2021 ). Macroalgae generally are superior competitors to coral and can preclude a return to a coral-dominated state even if sufficient coral propagules are present (Hughes et al. 2007 , Schmitt et al. 2022 ). Top-down control of macroalgae by herbivores is widely recognized as a crucial biological process that can influence shifts between coral and macroalgae (Hughes et al. 2007 , Ledlie et al. 2007 , Cheal et al. 2010 , Adam et al. 2011 , 2015 , Holbrook et al. 2016 , Schmitt et al. 2019 , 2022 , Tebbett et al. 2024 ). In this context, despite a broad range of feeding strategies, herbivorous fishes on coral reefs have two general functions (Bellwood et al. 2004 , Green and Bellwood 2009 , Burkepile and Hay 2010 , Cheal et al. 2010 , Han et al. 2016 , Burkepile et al. 2020 , Tebbett et al. 2024 ), which we operationally term ‘grazing’ and ‘browsing’ (e.g., Han et al. 2016 ). Following a coral mortality event, grazing herbivores that consume endolithic and filamentous algae prevent the establishment of macroalgae by consuming their early developmental stages on the disturbed substrate, thereby maintaining the reef surface in a condition (cropped turf) that can be colonized by coral (Birrell et al. 2005 , Cheal et al. 2010 , Graham et al. 2013 , Han et al. 2016 , Robinson et al. 2020 , Adam et al. 2022 , Schmitt et al. 2022 , Tebbett et al. 2024 ). However, if macroalgae do become established on the reef, they subsequently can be removed by species of browsing herbivores that consume mature macroalgae (Cheal et al. 2010 , Graham et al. 2013 , Han et al. 2016 , Topor et al. 2019 , Tebbett et al. 2024 ). Thus, browsers, but not grazers, play a critical role in reversing coral-to-macroalgae state shifts and enhancing the potential for reestablishment of coral (Bellwood et al. 2006 ). It has been posited that having robust populations of both grazing and browsing species contributes to coral reef resilience (Cheal et al. 2010 ), a notion that modeling of these systems has supported (Cook et al. 2024 ). Evidence suggests that grazers typically have higher species richness and biomass than browsers on tropical reefs, both at local reef tract (Adam et al. 2014 , Han et al. 2016 , Schmitt et al. 2022 ) and larger scales (Cheal et al. 2010 , Vérges et al. 2012, Heenan and Williams 2013 , Edwards et al. 2014 , Topor et al. 2019 , Rassweiler et al. 2020 ). This pattern can be exacerbated on reefs where local-scale fisheries heavily target browsing species (Cook et al. 2024 , Rassweiler et al. 2020 , 2022 ), potentially lessening the chance of a state shift from macroalgae back to coral. Indeed, persistent shifts to macroalgae on the Great Barrier Reef have been correlated with low abundances of browsing fishes (Cheal et al. 2010 ). Knowledge of the spatial pattern of covariation in herbivore grazing and browsing functions might provide insight into both the vulnerability of a reef to a coral-to-macroalgae state shift as well as the potential for such a shift to be reversed. There has been an emphasis in prior research and management approaches on the maintenance of coral dominance to prevent a shift to macroalgae, and less on the reversibility of such a shift to re-establish coral dominance (Bellwood et al. 2006 , Graham et al. 2013 ). Here we address that information deficit by exploring patterns and environmental correlates of spatial covariation in grazing and browsing functions of herbivorous fishes, and relating those patterns to both the vulnerability to and potential reversibility of a coral-macroalgae phase shift. The lagoons of Moorea, French Polynesia, provide an ideal model system to explore spatial variation in herbivory functions and the potential consequences of that spatial pattern to coral resilience at the local patch reef scale. Experimental studies suggest coral and macroalgae can be alternative basins of attraction (i.e., represent a state shift, Schmitt et al. 2019 , 2022 ), and time series data and other process studies show that herbivory has been essential to coral recovery following large mortality events (Adam et al. 2011 , Holbrook et al. 2016 , Schmitt et al. 2019 ). Here we quantified at a set of 20 sites in the north shore lagoons a range of biological and physical characteristics (rates of grazing and browsing, biomass and taxonomic composition of the herbivorous fish community, benthic community composition, productivity of algal turf and of macroalgae, nutrient enrichment and proximity to the reef crest and deep-water channels). These data were used to explore spatial patterns in herbivory within the lagoons. Further, ‘hot spots’ and ‘cold spots’ that had either high or low levels of herbivory respectively were identified and the environmental attributes of each type of site were characterized. In the coral reef literature, there is a paucity of experimental attempts that attempt to quantify the potential for reversibility of the macroalgae state. We explored this using a field experiment conducted along a natural gradient in browsing intensity to evaluate whether reversal potential was related to variation in strength of browsing. Our findings have considerable relevance to the development of spatially-explicit management actions to enhance resilience of coral reefs. Results Patterns of spatial variation in rates of grazing and browsing Assays to estimate the strength of grazing on algal turf and of browsing on macroalgae were conducted at 20 study sites in the lagoon. These revealed high spatial variation in herbivory rates across the lagoon reef system (Fig. 1 ). Of the sites examined, many had moderate to high levels of grazing, and only a few had low levels of grazing activity (Fig. 1 a, c). By contrast, many sites were characterized by low or moderate levels of browsing, with few having high browsing activity (Fig. 1 b, d). We first explored patterns in grazing and browsing along two spatial gradients: 1) alongshore (west to east along the north shore), and 2) cross-shore between two major reef habitats – fringing reefs and mid-lagoon reefs. No alongshore trend was apparent in either grazing or browsing ( Fig. 1 c, d ) . However, grazing – but not browsing – differed between habitats; grazing was higher on mid-lagoon reefs than fringing reefs (grazing: t (11) = 2.5, p < 0.05), browsing: Wilcoxon’s test W = 48, p = 0.9) ( SI Fig. 1 ). Grazing and browsing rates were not spatially correlated with each other across the twenty sites (Pearson’s r = 0.36, p = 0.12). Despite this lack of concordance, a striking spatial pattern was apparent (Fig. 2 ). Sites clustered into three of the four possible herbivory regimes (i.e., the orthogonal combinations of high and low rates of grazing and browsing), which were: (1) low grazing-low browsing, (2) high grazing-low browsing, and (3) high grazing-high browsing (Fig. 2 ). None of the sites were characterized by low grazing but high levels of browsing (Fig. 2 ). Relationships between spatial variation in herbivory, herbivore biomass, and environmental attributes Principal component analyses (PCA) explored the degree to which environmental factors [cover of macroalgae and of different types of turf (cropped turf, sedimented turf, cultivated turf within farmerfish Stegastes gardens), productivity of algal turf and of macroalgae, nutrient enrichment, and geographic location (proximity to the reef crest and deep-water channels)] could explain observed spatial patterns in browsing and grazing, specifically, the ‘hot spots’ (high grazing or browsing activity) and ‘cold spots’ (low activity) in the lagoon. A subset of sites that reflected the highest (hot spots) and lowest (cold spots) levels in each of the herbivory processes was selected for analysis. We found that different sets of variables explained the variance for each aspect of herbivory. Grazing ‘hot’ and ‘cold’ spots (i.e., high and low rates respectively) separated in ordination space along both axes PC1 and PC2, with cumulative proportion of variance explained = 0.71 (Fig. 3 a). Loading revealed that spatial variation in grazing was related to turf productivity, cover of the different categories of turf, and geographic location. Grazing hot spots were characterized by high productivity and cover of cropped turf as well as increased distance away from deep water drop-offs and closer proximity to the reef crest. Conversely, grazing cold spots were associated with turf containing high sediment loads, high cover of turf algae gardens defended by farmerfish ( Stegastes ), increased distance from the reef crest and closer proximity to deep water drop-offs. By comparison with grazing, the PCA for browsing revealed that hot and cold spots separated almost entirely along PC2, which was associated with geographic location (Fig. 3 b; cumulative proportion of variance explained for Axes 1 and 2 = 0.84). Loadings revealed that browsing hot spots tended to be further from the reef crest and closer to deep water drop-offs, whereas cold spots were generally the converse - at sites closer to the reef crest and farther from deep water. Patterns of grazing were associated with differences in the composition of turf communities between the mid-lagoon and fringing reef habitats. Mid-lagoon sites, which consistently had high grazing rates, were characterized by high cover of cropped turf that contained little or no sediment (Wilcoxon’s test = 21, p = 0.03, SI Fig. 2 a). Fringing reefs, which supported generally lower and spatially more variable levels of grazing, were characterized by higher cover of turf with high sediment loads ( t (3) = 2.7, p = 0.008, SI Fig. 2 b) and turf gardens defended by the farmerfish Stegastes ( t (3) = 2.5, p = 0.02, SI Fig. 2 c). Visual surveys of the herbivorous fishes conducted by divers at the 20 study sites revealed a total of 19 species. As expected, browsing species were less speciose in the assemblage (4 of 19 species) and they comprised only about three percent of the total biomass and of the total abundance ( SI Table 1 ). Browsers included species in 3 different genera ( Leptoscarus , Naso , Siganus ). Grazers were represented by 5 genera, with Acanthurus and Scarus predominant with 7 and 4 species respectively ( SI Table 1 ). We did not find the observed spatial variation in our measured rates of grazing or browsing were correlated with our visually-based estimates of fish herbivore biomass among the sites ( SI Fig. 3 ). Neither the among-site variation in the measured rate of grazing (Pearson’s r = 0.21, p = 0.37, SI Fig. 3 a) or browsing (Pearson’s r = -0.24, p = 0.3, SI Fig. 3 b) was related to the estimated local biomass of fish grazers or browsers, respectively. Predicting variation in herbivore biomass using environmental attributes Despite the fact that herbivore biomass and herbivory rates were not strongly correlated, spatial variation in the biomass of each herbivore group did map onto the same major environmental variables relating to the spatial patterns in grazing and browsing as revealed by our PCA analyses ( SI Table 2 ). An AIC-based model selection multiple linear regression indicated that variation in grazer biomass across the twenty sites was best predicted by turf productivity and, more weakly, cover of sedimented turf, which together explained 24% of the among-site variation in grazer biomass ( F (2, 7) = 3.96, p = 0.04). The best model for browsers revealed that variation in macroalgal productivity and distance from the reef crest were the best combination of predictors, which together explained 36% of variation in browser biomass ( F (2, 17) = 6.28, p = 0.009). Linking spatial variation in herbivory to the resilience of coral Because of a general lack of data that link browsing intensity and reversibility of state shifts to macroalgae, we explored this relationship using a field experiment. Macroalgae-dominated experimental units were exposed to herbivores along a natural gradient in browsing intensity (determined by herbivory assays) to evaluate whether reversal potential was related to variation in browsing. Spatial variation in reversibility was strongly related to browsing intensity; the capacity of browsing herbivores to remove late successional communities of macroalgae increased along a gradient in browsing (Fig. 4 a). After 10 days, the average density of adult Turbinaria within experimental communities in low browsing areas either did not change or decreased by only 6% (Fig. 4 a). This implies reefs in areas with low browsing capacity are vulnerable to remaining trapped in a macroalgae-dominated state (Fig. 4 b yellow and red quadrants). By contrast, herbivores in the high browsing site removed 84 ± 9% (mean ± SE) of Turbinaria and in some plots completely removed macroalgal communities, suggesting high potential for reversibility if a shift were to occur (Fig. 4 b tan and green quadrants). We found that reversibility significantly increased with closer proximity to deep water (Pearson’s r = 0.89, p = 0.044, SI Fig. 4 ), which paralleled the pattern observed for spatial variation in browsing intensity revealed by our short-term assays (Fig. 3 b). Discussion Grazers and browsers are critical for preventing and reversing state shifts on coral reefs, respectively, and there is concerning evidence that small-scale fisheries may disproportionately reduce browser biomass relative to grazers. Given the posited implications of such a disparity in functional impact of these two herbivore groups for coral-to-macroalgae state shifts (Cheal et al. 2010 , Schmitt et al. 2022 , Cook et al. 2024 ) it is crucial to better understand spatial patterns of grazing and browsing and how variation in these processes impacts the resilience of coral. Here, we investigated spatial variation in grazing and browsing, identified key environmental factors contributing to that variation, and assessed how covariation in these distinct herbivory processes may influence resilience of the coral state in a system where coral and macroalgae have been demonstrated experimentally to be alternative basins of attraction under certain environmental conditions (Schmitt et al. 2019 , 2022 ). Our study revealed high spatial variation in herbivory rates across lagoon reefs with different patterns emerging for grazing and browsing. Grazing hot spots were more common across lagoons than browsing hot spots, which likely reflects the disparity in species richness and abundance of grazers and browsers in Moorea ( SI Table 1 ). Although we did not detect alongshore or cross-shore gradients in browsing, we found grazing varied between fringing and mid-lagoon reef habitats that differ in distance from shore. Rates of grazing were consistently high on mid-lagoon reefs but variable on fringing reefs. This decline in grazing impact across a reef gradient has been observed in other coral reef systems (Fox and Bellwood 2007 ) and has been attributed to the effect of territorial fishes such as farming damselfish that defend turf gardens and exclude other grazers (Ceccarelli et al. 2005 ), increased rates of sedimentation in nearshore habitats that reduce quality of algal turfs (Goatley and Bellwood 2012 ), and increasing productivity of algae with distance from shore (Klumpp and McKinnon 1992 ). In our case all three hypotheses are supported; compared to mid-lagoon reefs, fringing reefs had higher cover of Stegastes turf and sedimented turf as well as lower turf productivity. Therefore, in our system fringing reefs appear to be less attractive foraging habitats for grazers due to the quality and accessibility of algal turfs. By contrast, characteristics related to the macroalgae community did not appear to explain variation in browsing. Instead, browsing was most influenced by distance from deep water. This relationship between browsing and deep water may not be universal on coral reefs, but in this case, it is likely driven by local fishing dynamics. In Moorea and generally in French Polynesia, browsers such as Naso (unicornfish) are highly prized for their taste and market value (Nassiri et al. 2021 ). Although Naso are typically difficult to catch and demonstrate fearful behavior of humans – further indicative of the high fishing pressure they face – fishers in Moorea show high selectivity for them (Rassweiler et al. 2020 ). Free-diving spearfishers report targeting browsers on reefs near deep water as they believe it provides a spatial refuge for the fish which are much less frequently encountered in the lagoon ( SI Table 1 ). We found that estimates of herbivore biomass derived from diver surveys alone were poor predictors of herbivory rate. This decoupling between herbivore biomass and functional impact has been noted previously (e.g., Hoey and Bellwood 2010 ), and can occur for various reasons. The grazers and browsers investigated in our study are mainly comprised of highly mobile herbivores, and therefore may not be counted in visual censuses where they feed (Fox and Bellwood 2014 ). Further, as mentioned previously, fearful behavior in response to human presence by some herbivorous fishes can result in low biomass estimates in diver surveys (Gotanda et al. 2009 ) like we observed in this study for browsers ( SI Table 1 ), which are highly prized by local spearfishers (Rassweiler et al. 2022 , Cook et al. 2024 ). Another potential explanation is related to the functional dilution of herbivores as the cover of algae increases, which may result in low herbivory estimates even in areas with high herbivore biomass. Consequently, caution should be exercised when linking measures of abundance to functional impact (Tebbett et al. 2024 ). Therefore, we propose that monitoring both herbivore biomass and rates of algal removal may be the most effective method to evaluate the status of vulnerable reefs, as grazing and browsing assays are relatively rapid and cost-effective, and appear to provide reliable assessments of key trophic processes associated with coral reef resilience (Chong-Seng et al. 2014 , Nash et al. 2016 ). Estimating the two types of herbivory processes, in addition to standing biomass of herbivores, provides a more direct measure of the potential influence of the complementary guilds of herbivores on ecosystem functioning and reef resilience (Goatley et al. 2016 , Cook et al. 2024 , Tebbett et al. 2024 ). Our results indicate grazing and browsing were spatially variable and not strongly correlated with each other across sites. Experimental work and time-series data demonstrate that reductions in grazing can lead to the proliferation of macroalgae on reefs (Holbrook et al. 2016 , Schmitt et al. 2019 , 2022 ), and that browsing may be linked to the likelihood of reversing a coral-to-macroalgae state shift (Bellwood et al. 2006 , Cheal et al. 2010 ). It follows that sites with low levels of grazing and browsing are likely vulnerable to shifting to and remaining trapped in a macroalgae-dominated state. By contrast, sites with high levels of both herbivory processes may be resilient due to high prevention and reversal potential. Sites characterized by high grazing and low browsing probably have moderate resilience: establishment of macroalgae is unlikely, but so is recovery of the coral state should a shift occur (i.e., a coral-macroalgae ‘phase shift’ rather than a regime shift). Interestingly, our surveys revealed that sites were not evenly distributed among these four possible herbivory regimes. There was a complete lack of low grazing-high browsing sites, which may not be surprising given the lower ratio of browsers to grazers generally in Moorea. Only 15% and 20% of sites fell into low grazing-low browsing and high grazing-high browsing regimes, respectively, whereas the majority of sites (65%) were characterized by high grazing and low browsing. The observed pattern of spatial covariation in grazing and browsing functions could be shaped in part by patterns and preferences of fishers in the small-scale fishery of Moorea (Rassweiler et al. 2020 , 2022 , Holbrook et al. 2022 ) that, coupled with different life histories that make browsers more susceptible to overfishing (Cook et al. 2024 ), can influence both overall biomass of herbivorous fishes and the disparity between grazers and browsers. In systems such as in Moorea where browsers are target species, the first component of resilience to be weakened through fishing is likely to be the reversibility of a coral-to-macroalgae state shift (i.e., system moves from green to yellow box in Fig. 4 b). As fishing intensity continues to increase, the ability of the herbivore community to prevent a shift to macroalgae will subsequently be eroded (i.e., system moves from the yellow to the red box in Fig. 4 b), and the system will remain trapped in the macroalgae state. This trajectory of reef degradation could represent the future for reefs worldwide that support small-scale fisheries in which herbivores are target species (Cheal et al. 2010 , Edwards et al. 2014 , Heenan et al. 2016 , Topor et al. 2019 ). However, knowledge of spatial covariation in herbivore functions could provide a useful template for spatially-explicit management actions tailored to local conditions. Our results indicate there likely will not be a single strategy to preserve grazing and browsing functions on reefs. Management goals will require different strategies based on key environmental factors and human pressures influencing each herbivory process. For example, managers may need to mitigate sedimentation caused by certain land-use practices to preserve grazing, while limiting fishing may be a better strategy to enhance browsing capacity. Managers may also need to apply strategies in different reef areas since grazing and browsing hot spots do not necessarily coincide in the same locations. One factor we found important in predicting the biomass of grazing and browsing herbivores alike was algal productivity, which itself can be influenced by bottom-up forcing (nutrient enrichment) (Adam et al. 2021 , Holbrook et al. 2022 ). Therefore, management could support grazers and browsers simultaneously by mitigating stressors or disturbances that alter algal productivity in ways that affect top-down control. Finally, management practices historically have focused on means to preserve grazing functions on coral reefs to prevent coral-to-macroalgae state shifts, which is a reasonable priority given the potential challenges in reversing such a shift. However, the frequency and intensity of disturbances to coral reefs that alone can trigger state tipping are predicted to increase. As such, greater attention should be given to enhance protection of browsers and their functional impact to reefs. To that end, our study suggests that understanding spatial covariation in grazing and browsing functions can help better target management actions to enhance resilience of coral. Methods Study site This study was conducted in the shallow lagoons of Moorea, French Polynesia (17°30’ S, 149°50’ W). A barrier reef ~ 1 km offshore protects the lagoons from the open ocean except for 2 to 4 breaks produced by deep reef passes on each of the 3 sides of the island. Back reef habitats shoreward of the barrier reef are characterized by a short band of contiguous reef substrate that then transitions into patch reefs (bommies) surrounded by sand in the mid-lagoon. Mid-lagoon reefs are highly variable with respect to cover of coral, macroalgae, cropped turf algae, and other taxa. Directly adjacent to shore are shallow fringing reefs that can be separated from the mid-lagoon by deep channels. Water circulation within the lagoon is driven by waves forcing water over the crest of the barrier reef, through the lagoons, and out the passes (Leichter et al. 2013 ). Hydrodynamic circulation patterns, along with terrestrial run-off, concentrate nutrient enrichment at passes, bays, and beneath major watersheds (Adam et al. 2021 ). Along Moorea’s north shore, we selected twenty sites spread across the four lagoons that are divided by three reef passes (Fig. 1 a, b). Ten sites were located on the fringing reef, and ten were in the mid-lagoon. At each site we quantified rates of grazing and browsing, as well as biomass and taxonomic composition of the herbivorous fish community, benthic community composition, productivity of algal turf and of macroalgae, and nutrient enrichment. To determine the distance of each site from geographic features of interest (e.g., the barrier reef crest, deep water channels), we used spatial data layers produced from a previous study (Holbrook et al. 2022 ) that mapped the inner edge of the reef crest and the coastline based on LiDAR-based digital elevation maps (Collin et al. 2018 ), as well as satellite imagery provided by Google Earth Pro (Version 7.3.6.9345). For each site, we calculated the minimum distance from the reef crest using the ‘sf’ package (Pebesma 2018 ) in R (version 4.1.1). Then, we used the ruler tool in Google Earth Pro to measure the minimum distance between a site and its nearest deep-water channel, which was visually classified by having deep blue color (as opposed to visible reef substrate or sand) and being located within the lagoon. Quantifying spatial patterns in herbivory We deployed assays to quantify levels of both grazing on algal turf and browsing on macroalgae at the 20 sites. For grazing, we exposed uniform pieces of turf-covered reef substrate to herbivores for 3 h between the hours of 10:00 and 16:00. Turf offerings consisted of 7 x 7 cm 2 pieces of dead coral rubble covered by highly palatable turf (e.g., Polysiphonia spp.) that were collected from gardens in the lagoon that were cultivated by farming damselfish ( Stegastes spp.). Each of these was fastened to a rack constructed of PCV-coated galvanized wire mesh that was affixed to open substrate (i.e., lacking coral or macroalgae) on the tops of patch reefs (N = 5 replicates per site). Racks at each site were placed at least 5 m apart. The percent of turf that was consumed was estimated after 3 h by the same observer (see Davis 2018 , Schmitt et al. 2022 ). Two trials of the grazing assay were conducted two weeks apart and all replicates at each site were pooled (N = 10 replicates per site) and then averaged to obtain the percent turf consumed, which we used as an estimate of grazing intensity. Simultaneously, we conducted browsing assays using macroalgae offerings. We attached two 15 cm-long pieces of the palatable brown macroalga, Sargassum pacificum , to a rack using clothes pins to hold them upright and deployed them as described above (N = 5 replicates per site). Lengths of Sargassum pieces were measured after 24 h (see Davis 2018 ) to determine the percent consumed; values for the two fronds deployed together were averaged for each replicate. Three trials of browsing assays were conducted two to three weeks apart and all replicates were pooled (N = 15 replicates per site) and averaged for each site to calculate the percent consumption, our estimate of browsing intensity. To explore spatial patterns of herbivory we created maps of grazing and browsing estimates from the twenty sites using the ‘sf’ package (Pebesma 2018 ) in R (version 4.1.1). Using the R statistical computing software (version 4.3.2; R Core Team 2023 ), covariation between grazing and browsing across sites was determined using Pearson’s correlation coefficient. Quantifying herbivorous fish communities and benthic composition The assemblage of herbivorous fishes at each site was characterized using visual surveys in which an observer swam a timed transect, counting and visually estimating total lengths (TLs) of mobile herbivorous fishes ≥ 10 cm TL in a 5 m wide swath. The observer towed an inflatable float behind them with a Garmin GPSMap 78 handheld GPS (Olathe, Kansas, USA) to geo-reference fish counts and provide an estimate of the area covered by the survey (Miller et al. 2023 ). TLs were used to estimate biomass using published species-specific relationships, and species were assigned to functional groups (Moorea Coral Reef LTER and Brooks 2023). Based on the fish biomass estimates and the area covered by the survey, we calculated biomass per unit area, expressed as g/m 2 of grazers and browsers. Benthic composition was quantified along 50 m x 1 m long transects (N = 3 transects per site), in which the substratum was recorded at every 0.5 m (101 points per transect, 303 points per site). Substratum categories consisted of live scleractinian corals and macroalgae identified to genus, ‘other sessile invertebrates’ (mainly giant clams and sea cucumbers), rubble, sand, and three categories of turf: (1) turf inundated with sediment (hereafter ‘sedimented turf’), (2) turf growing within farmerfish ( Stegastes nigricans ) gardens (hereafter ‘ Stegastes turf’), and (3) closely cropped turf without sediment and located outside of a Stegastes garden (hereafter ‘cropped turf’). The number of points for each substratum category was divided by the total 303 points per site and multiplied by 100 to estimate percent cover. Quantifying the nutrient environment and algal productivity The waters surrounding Moorea are highly oligotrophic, so we used the nitrogen tissue content of a common macroalga, Turbinaria ornata , as a proxy for the local nutrient environment during a period of up to three months prior to Turbinaria collection (Adam et al. 2021 , Holbrook et al. 2022 ). Turbinaria responds to N pulses by storing surplus N (Schaffelke 1999 ) and N tissue content can provide a time-integrated measure of N availability (Atkinson and Smith 1983 , Fong et al. 1994 , Shantz et al. 2015 ). In June of 2017, a total of 10 Turbinaria stipes were randomly collected at each site and transported damp to the laboratory, where 10 florets were removed from each stipe 5 cm from the tip. Samples were dried in a drying for 4 days at 60° C. Total N content was determined via elemental analysis using a CHN Carlo-Erba elemental analyzer (NA1500) at the University of Georgia Center for Applied Isotope Studies. The productivity of turf was measured by allowing turf to colonize and grow for 21 days on 2.5 cm 2 unglazed terra cotta tiles in herbivore-exclusion cages (N = 8 caged tiles per site). Herbivore exclusion cages were 10 x 10 x10 cm galvanized mesh, with mesh size of 2.5cm. After the deployment period, tiles were brought to the lab, and turf was removed and processed to obtain ash-free dry weight (AFDW). The mean value of AFDW for each site provided an estimate of turf productivity (g accumulated 3 wk − 1 ). Macroalgal productivity was quantified by allowing juvenile Turbinaria , and other colonizing macroalgae, to grow protected from herbivores for 8 weeks. We collected reef substratum with attached juvenile Turbinaria from the mid-lagoon and removed all other algae so only juvenile Turbinaria (≤ 3 cm length) of uniform size and density remained. The substrates containing Turbinaria were attached to the bottom of herbivore-exclusion cages at each site (N = 8 per caged replicates per site). After two months, Turbinaria and any other colonizing macroalgae were removed and damp weighed. Weights of replicates were pooled at a given site to provide an estimate of macroalgal productivity (g accumulated 8 wk − 1 ). Relationships between spatial variation in herbivory, herbivore biomass, and environmental attributes Covariation between grazing rate and the biomass of grazing fishes from our fish counts across the twenty sites was determined using Pearson’s correlation coefficient (version 4.3.2; R Core Team 2023 ). The same analysis was conducted to assess the relationship between browsing rate and the biomass of browsing fishes across sites. We then utilized principal component analysis (PCA) in R (version 4.3.2; R Core Team 2023 ) to explore the degree to which environmental factors could explain observed spatial patterns in browsing and grazing, specifically, the ‘hot spots’ (high grazing or browsing activity) and ‘cold spots’ (low activity) in the lagoon. A subset of sites that reflected the highest (hot spots) and lowest (cold spots) levels in one of the herbivory processes was selected for this analysis, utilizing an additional criterion that the level of the second herbivory process was similar among the group. To explore variation in grazing we chose sites that exhibited the highest or lowest grazing, but that had similar (low) levels of browsing. This selection process resulted in five high grazing sites (i.e., grazing hot spots) and three low grazing sites (i.e., grazing cold spots) (Fig. 2 bottom right and left corners). The full range of browsing only occurred among sites with moderate to high levels of grazing. Four high browsing sites (i.e., browsing hot spots) and five low browsing sites (i.e., browsing cold spots), all of which had high levels of grazing were chosen for analysis (Fig. 2 right top and bottom corners). Explanatory variables for the PCA analyses included both spatial factors as well as variables that reflected the amount and productivity of the food resources of the fishes and therefore could reflect the intensity of browsing or grazing. Location in the lagoon can be important predictors of fish spatial patterns of abundance and activity, and each site’s distance from the reef crest and from deep water were included as predictor variables. The amounts of algal cover and productivity impact the distribution and feeding behavior of herbivorous fishes (Tootell and Steele 2016 ). For example, turf can vary in palatability or accessibility to grazing herbivores due to sediment load (Bellwood and Fulton 2008 , Goatley and Bellwood 2012 ) or whether it is guarded by territorial farming damselfish ( Stegastes spp.), with sedimented turf and turf within farmerfish gardens comprising less preferred feeding substrates than unsedimented and undefended turf. Therefore, turf productivity and cover of different types of turf (turf, sedimented turf, turf inside Stegastes gardens) were included in the PCA for grazing, and cover and productivity of macroalgae for browsing. The analysis was based on a correlation matrix of these data. Predicting variation in herbivore biomass using environmental attributes To explore whether environmental attributes associated with hot spots and cold spots in herbivory also predict variation in the biomass of herbivores, we utilized the best subset selection approach, which is an exploratory model building regression analysis. Using predictor variables from the grazing PCA analysis, we utilized the R package “leaps” (Version 3.1) to test all possible combinations of predictors in a multiple linear regression to explain variation in grazer biomass, and then selected the best model according to lowest AIC score with the R package “AICcmodavg” (Version 2.3). The same procedure was done for browsers using predictors from the browsing PCA. Linking spatial variation in herbivory to the resilience of coral Because of a general lack of data that link browsing intensity and reversibility of state shifts to macroalgae, we explored this relationship using a field experiment. Macroalgae-dominated communities were exposed to herbivores along a natural gradient in browsing intensity to evaluate whether reversal potential was related to variation in browsing. We selected five sites that represented a range in browsing intensity based on short-term herbivory assays described above. At each site, we deployed macroalgae communities that mimicked patch reefs that had undergone a shift to macroalgal dominance. Each community was constructed by chiseling off reef substrate with attached Turbinaria , which was then assembled into a community that reflected the size and density of macroalgae-dominated patch reefs within the lagoon. This assemblage was then epoxied to a cinderblock and the density of adult Turbinaria stipes recorded. The cinderblock macroalgal community (N = 6 replicates per site) was then deployed at a depth within 2–5 m and left exposed to herbivores for 10 days. At the end of the experiment, final adult Turbinaria density was recorded. The percent change in density of adult Turbinaria was calculated for each replicate using the equation (Final – Initial) / Initial x 100 . Replicate values were then averaged for each site and used as a proxy for reversibility of a coral-to-macroalgae state shift. Lastly, we were interested in whether environmental features associated with variation in browsing were also good predictors of reversibility. Covariation between reversibility and distance from deep water (log-transformed) for the five sites was determined using Pearson’s correlation coefficient (version 4.3.2; R Core Team 2023 ). Declarations Additional Information Competing Interests Statement : The author(s) declare no competing interests. Author Contribution Dana T. Cook (DTC), Sally J. Holbrook (SJH) and Russell J. Schmitt (RJS) collectively designed the study; laboratory and field work was done primarily by DTC with contributions by SJH and RJS; DTC performed the statistical analyses and data visualizations; DTC, SJH and RJS co-wrote the paper with DTC producing the initial draft; SJH and RJS secured the funding. Acknowledgement We thank Andrew Rassweiler, Sarah Lester, Matthew Lauer and Jean Wencélius for fruitful discussions, Hillary Krumbholz for Information Management assistance, and Lauren Enright, Olivia Isbell, Taylor Traxler and Kai Kopecky for assistance in the field. We also thank the staff of the University of California Gump Research Station for logistic support. This work is a contribution of the NSF Moorea Coral Reef Long Term Ecological Research Site and was supported by National Science Foundation grants OCE-1637396, BCS-1714704, and OCE-2224354. Field research was completed under permits issued by the Territorial Government of French Polynesia (Délégation à la Recherche) and the Haut-commissariat de la République en Polynésie Française (DTRT) (Protocole d’Accueil 2006-2022); we thank them for their continued support. With respect to the spelling of Moorea, we followed the Raapoto transcription system that is adhered to by a large segment of the Tahitian community, but also recognize other community members follow the Te Fare Vanā’a transcription system where the island name is spelled with an ’eta (Mo’orea). Data Availability All data sets and code utilized for this research are publicly available on GitHub [https://github.com/danatcook/herbivory_hotspots]. Final versions of data and code files will be permanently deposited with the Environmental Data Initiative (EDI) upon manuscript acceptance. References Adam, T. C. et al . How will coral reef fish communities respond to climate-driven disturbances? Insight from landscape-scale perturbations. Oecologia 176 ,285-296 (2014). Adam, T. C. et al . Landscape-scale patterns of nutrient enrichment in a coral reef ecosystem: implications for coral to algae phase shifts. Ecol. Appl . 31 , e02227 (2021). Adam, T. C., Burkepile, D. E., Ruttenberg, B. I. & Paddack, M. J. Herbivory and the resilience of Caribbean coral reefs: knowledge gaps and implications for management. Mar . Ecol . Prog . Ser . 520 , 1-20 (2015). Adam, T. C. et al . Priority effects in coral–macroalgae interactions can drive alternate community paths in the absence of top-down control. Ecology 103 ,e3831 (2022). Adam, T. C. et al . Herbivory, connectivity, and ecosystem resilience: response of a coral reef to a large-scale perturbation. PLoS ONE 6 , e23717 (2011). Atkinson, M. J. & Smith, S. V. C:N:P ratios of benthic marine plants. Limnol. Oceanogr. 28 , 568-574 (1983). Bellwood, D. R. & Fulton, C. J. Sediment-mediated suppression of herbivory on coral reefs: decreasing resilience to rising sea levels and climate change? Limnol. Oceanogr. 53 , 2695-2701 (2008). Bellwood, D. R., Hughes, T. P., Folke, C. & Nyström, M. Confronting the coral reef crisis. Nature 429 , 827-833 (2004). Bellwood, D. R., Hughes, T. P. & Hoey, A. S. Sleeping functional group drives coral-reef recovery. Curr. Biol. 16 ,2434-2439 (2006). Birrell, C. L., McCook, L. J. & Willis, B. L. Effects of algal turfs and sediment on coral settlement. Mar . Pollut . Bull . 51 , 408-414 (2005). Burkepile, D. E. & Hay, M. E. Impact of herbivore identity on algal succession and coral growth on a Caribbean reef. PLoS ONE 5 , e8963 (2010). Burkepile, D. E., Schmitt, M. H., Stears, K., Donovan, M. K. & Thompson, D. I. Shared insights across the ecology of coral reefs and African savannas: are parrotfish wet wildebeest? BioScience 70 , 647-658 (2020). Ceccarelli, D. M., Jones, G. P. & McCook, L. J. Foragers versus farmers: contrasting effects of two behavioural groups of herbivores on coral reefs. Oecologia 145 , 445-453 (2005). Cheal, A. J. et al . Coral – macroalgal phase shifts or reef resilience: links with diversity and functional roles of herbivorous fishes on the Great Barrier Reef. Coral Reefs 29 , 1005-1015 (2010). Chong-Seng, K. M., Nash, K. L.,Bellwood, D. R. & Graham, N. A. J. Macroalgal herbivory on recovering versus degrading coral reefs. Coral Reefs 33 , 409-419 (2014). Collin, A. et al . Very high resolution mapping of coral reef state using airborne bathymetric LiDAR surface-intensity and drone imagery. Int . J . Remote Sens . 39 , 5676-5688 (2018). Conversi, A. et al . A holistic view of marine regime shifts. Phil. Trans. R. Soc. B: Biol. Sci. 370 , 20130279 (2015). Cook, D. T., Schmitt, R. J., Holbrook, S. J. & Moeller, H. V. Modeling the effects of selectively fishing key functional groups of herbivores on coral resilience. Ecosphere 15 (1) , e4749 (2024). Davis, S. L. Associational refuge facilitates phase shifts to macroalgae in a coral reef ecosystem. Ecosphere 9 , e02272 (2018). Edwards, C. B. et al . Global assessment of the status of coral reef herbivorous fishes: evidence for fishing effects. Proc. R. Soc. B. 281 , 20131835 (2014). Fong, P., Donohoe, R. M. & Zedler, J. B. Nutrient concentration in tissue of the macroalga Enteromorpha as a function of nutrient history: an experimental evaluation using field microcosms. Mar. Ecol. Prog. Ser. 106 , 273-281 (1994). Fox, R. J. & Bellwood, D. R. Quantifying herbivory across a coral reef depth gradient. Mar. Ecol. Prog. Ser. 339 , 49-59 (2007). Fox, R. J. & Bellwood, D. R. Herbivores in a small world: network theory highlights vulnerability in the function of herbivory on coral reefs. Func . Ecol . 28 , 642-651 (2014). Goatley, C. H. R. & Bellwood, D. R. Sediment suppresses herbivory across a coral reef depth gradient. Biol . Lett . 8 , 1016-1018 (2012). Goatley, C. H. R., Bonaldo, R. M., Fox, R. J. & Bellwood, D. R. Sediments and herbivory as sensitive indicators of coral reef degradation. Ecol. Soc . 21(1) , 29 (2016). Gotanda, K. M., Turgeon, K. & Kramer, D. L. Body size and reserve protection affect flight initiation distance in parrotfishes. Behav. Ecol. Sociobiol . 63 , 1563-1572 (2009). Graham, N. A. et al . Managing resilience to reverse phase shifts in coral reefs. Front . Ecol . Environ . 11 , 541-548 (2013). Green, A. L., & Bellwood, D. R. Monitoring Functional Groups of Herbivorous Reef Fishes as Indicators of Coral Reef Resilience: A Practical Guide for Coral Reef Managers in the Asia Pacific Region (IUCN, Gland, Switzerland, 2009). Han, X., Adam, T. C., Schmitt, R. J., Brooks, A. J. & and Holbrook, S. J. Response of herbivore functional groups to sequential perturbations in Moorea, French Polynesia. Coral Reefs 35 , 999-1009 (2016). Heenan, A. & Williams, I. D. Monitoring herbivorous fishes as indicators of coral reef resilience in American Samoa. PLoS ONE 8(11) , e79604 (2013). Heenan, A., Hoey, A. S., Williams, G. J. & Williams, I. D. Natural bounds on herbivorous coral reef fishes. Proc. R. Soc. B. 283 , 20161716 (2016). Hoey, A. S. & Bellwood, D. R. Cross-shelf variation in browsing intensity on the Great Barrier Reef. Coral Reefs 29 , 499-508 (2010). Holbrook, S. J., Schmitt, R. J., Adam, T. C. & Brooks, A. J. Coral reef resilience, tipping points and the strength of herbivory. Sci. Rep . 6 , 35817 (2016). Holbrook, S. J. et al . Spatial covariation in nutrient enrichment and fishing of herbivores in an oceanic coral reef ecosystem. Ecol . Appl . 32, e2515 (2022). Hughes, T. P. et al . Phase shifts, herbivory, and the resilience of coral reefs to climate change. Curr. Biol . 17 , 360-365 (2007). Klumpp, D. W. & McKinnon, A. D. Community structure, biomass and productivity of epilithic algal communities on the Great Barrier Reef: dynamics at different spatial scales. Mar . Ecol . Prog . Ser . 86 , 77-89 (1992). Ledlie, M. H. et al . Phase shifts and the role of herbivory in the resilience of coral reefs. Coral Reefs 26 , 641-653 (2007). Leichter, J. et al . Biological and physical interactions on a tropical island coral reef: transport and retention processes on Moorea, French Polynesia. Oceanography 26 , 52-63 (2013). Mcleod, E. et al . The future of resilience-based management in coral reef ecosystems. J . Environ . Manage . 233 , 291-301 (2019). Miller, S. D. et al . Using machine learning to achieve simultaneous, georeferenced surveys of fish and benthic communities on shallow coral reefs. Limnol. Oceanogr. Methods 21 , 451-466 (2023). Moorea Coral Reef LTER & Brooks, A. MCR LTER: Coral Reef: Long-term population and community dynamics: fishes, ongoing since 2005 ver 62. Environmental Data Initiative. https://doi.org/10.6073/pasta/75644add7e7f90c568bf5045264d359a (2023). Mumby, P. J., Steneck, R. S. & Hastings, A. Evidence for and against the existence of alternate attractors on coral reefs. Oikos 122 , 481-491 (2013). Nash, K. L., Abesamis, R. A., Graham, N. A. J., McClure, E. C. & Moland, E. Drivers of herbivory on coral reefs: species, habitat and management effects. Mar . Ecol . Prog . Ser . 554 , 129-140 (2016). Nassiri, A. et al . Hedonic evaluation of coral reef fish prices on a direct sale market. Mar . Policy 129 , 104525 (2021). Pebesma, E. Simple features for R: Standardized support for spatial vector data. The R Journal 10 , 439-446 (2018). R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/ (2023). Rassweiler, A. et al . Perceptions and responses of Pacific Island fishers to changing coral reefs. Ambio 49 , 130-143 (2020). Rassweiler, A. et al . How do fisher responses to macroalgal overgrowth influence the resilience of coral reefs? Limnol . Oceanogr . 67 , S365-S377 (2022). Robinson, J. P. W. et al . Habitat and fishing control grazing potential on coral reefs. Func . Ecol . 34 , 240-251 (2020). Roff, G. & Mumby, P. J. Global disparity in the resilience of coral reefs. Trends Ecol. Evol . 27 , 404-413 (2012). Schaffelke, B. Short-term nutrient pulses as tools to assess responses of coral reef macroalgae to enhanced nutrient availability. Mar. Ecol. Prog. Ser . 182 , 305-310 (1999). Scheffer, M. & Carpenter, S. R. Catastrophic regime shifts in ecosystems: linking theory to observation. Trends Ecol. Evol . 18 , 648-656 (2003). Schmitt, R. J., Holbrook, S. J., Brooks, A. J. & Adam, T. C. Evaluating the precariousness of coral recovery when coral and macroalgae are alternative basins of attraction. Limnol . Oceanogr . 67 , S285-S297 (2022). Schmitt, R. J., Holbrook, S. J., Davis, S. L., Brooks, A. J. & Adam, T. C. Experimental support for alternative attractors on coral reefs. Proc. Nat. Acad. Sci . 116 , 4372-4381 (2019). Shantz, A. A., Ladd, M. C., Schrack, E. & Burkepile, D. E. Fish-derived nutrient hotspots shape coral reef benthic communities. Ecol. Appl . 25 , 2142-2152 (2015). Suding, K. N. & Hobbs, R. J. Threshold models in restoration and conservation: a developing framework. Trends Ecol. Evol . 24 , 271-279 (2009). Sully, S., Burkepile, D. E., Donovan, M. K., Hodgson, G. & Van Woesik, R. A global analysis of coral bleaching over the past two decades. Nat. Commun . 10 , 1264 (2019). Tebbett, S. B., Bennett, S. & Bellwood, D. R. A functional perspective on the meaning of the term ‘herbivore’: patterns versus process in coral reef fishes. Coral Reefs 43 , 219-232 (2024). Tootell, J. S. & Steele, M. A. Distribution, behavior, and condition of herbivorous fishes on coral reefs track algal resources. Oecologia 181 , 13-24 (2016). Topor, Z. M., Rasher, D. B., Duffy, J. E. & Brandl, S. J. Marine protected areas enhance coral reef functioning by promoting fish biodiversity. Conserv. Lett . 12 , e12638 (2019). Vergés A., Bennett, S. & Bellwood, D. R. Diversity among macroalgae-consuming fishes on coral reefs: a transcontinental comparison. PLoS ONE 7(9) , e45543 (2012). Additional Declarations No competing interests reported. Supplementary Files SupplementaryInformation.pdf Cite Share Download PDF Status: Published Journal Publication published 07 Jan, 2025 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 30 Oct, 2024 Reviews received at journal 28 Oct, 2024 Reviews received at journal 24 Oct, 2024 Reviewers agreed at journal 21 Oct, 2024 Reviewers agreed at journal 14 Oct, 2024 Reviewers agreed at journal 13 Oct, 2024 Reviewers agreed at journal 13 Oct, 2024 Reviewers invited by journal 11 Oct, 2024 Editor assigned by journal 10 Oct, 2024 Editor invited by journal 10 Oct, 2024 Submission checks completed at journal 10 Oct, 2024 First submitted to journal 26 Sep, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5160728","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":372273525,"identity":"4f1b7118-56fa-4134-9b8a-b9f01e4c4661","order_by":0,"name":"Dana T. Cook","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0UlEQVRIiWNgGAWjYDCCA1Bagr0BSBpYEKkFhCV4QKSBBClaJBLAlhHWwXf87MHPH2ps5CRnPr+64UeBBAN/e3cCXi2SZ/KSJQ4cSzOWls4pu9kDdJjEmbMb8GoxOJBjIHGw4XDiPOmctBs8QC0GErkEtJx/Y/wDrEXyTNrNP0RpuZFjBrZltgT7sdtE2SJ5442ZxRmgXyR7cthuyxhI8BD0C9/5HOMbFcAQkzh+/NnNN39s5Pjbe/FrQQI8BmCSWOUgwP6AFNWjYBSMglEwggAAkXRLtIaX398AAAAASUVORK5CYII=","orcid":"","institution":"University of California Santa Barbara","correspondingAuthor":true,"prefix":"","firstName":"Dana","middleName":"T.","lastName":"Cook","suffix":""},{"id":372273528,"identity":"e9e06710-f068-41fa-8800-64451bc4fd98","order_by":1,"name":"Sally J. Holbrook","email":"","orcid":"","institution":"University of California Santa Barbara","correspondingAuthor":false,"prefix":"","firstName":"Sally","middleName":"J.","lastName":"Holbrook","suffix":""},{"id":372273531,"identity":"b92d7ecd-00fb-44c9-8aa6-36841639f8f2","order_by":2,"name":"Russell J. Schmitt","email":"","orcid":"","institution":"University of California Santa Barbara","correspondingAuthor":false,"prefix":"","firstName":"Russell","middleName":"J.","lastName":"Schmitt","suffix":""}],"badges":[],"createdAt":"2024-09-26 19:23:06","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5160728/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5160728/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-024-83672-1","type":"published","date":"2025-01-07T15:58:04+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":70532953,"identity":"24c5e9c8-4612-49dc-b3de-e3083d848489","added_by":"auto","created_at":"2024-12-04 06:18:21","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1151643,"visible":true,"origin":"","legend":"\u003cp\u003eMaps showing percent consumption (mean ± SE) of \u003cstrong\u003e(a) \u003c/strong\u003eturf assays after 3 hr (N = 10 replicates per site) (i.e., grazing rate) and \u003cstrong\u003e(b)\u003c/strong\u003e macroalgae assays after 24 hr (N = 15 replicates per site) (i.e., browsing rate) across the twenty sites. Yellow denotes high rates of herbivory and blue indicates low rates. Bar plots show the same rates of\u003cstrong\u003e (c)\u003c/strong\u003e grazing and \u003cstrong\u003e(d)\u003c/strong\u003e browsing at the twenty sites ordered from west to east. Bars are colored according to reef habitat.\u003c/p\u003e","description":"","filename":"Figure1.tif.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5160728/v1/2b03b5e6f9b0f652b28e077b.jpg"},{"id":70532621,"identity":"bad92530-9daf-411c-abc0-4ac82d92f89e","added_by":"auto","created_at":"2024-12-04 06:10:21","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":195489,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial covariation between grazing and browsing rates across the twenty sites. Values are the site averages shown in Fig. 1. Grazing rate is mean percent consumption of turf assays after 3 hr and browsing rate is mean percent consumption of macroalgae assays after 24 hr.\u003c/p\u003e","description":"","filename":"Figure2.tif.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5160728/v1/12f511e35d57b7b4ea0638af.jpg"},{"id":70533988,"identity":"5917b8cc-5461-463f-8414-8f4751827812","added_by":"auto","created_at":"2024-12-04 06:26:21","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":479131,"visible":true,"origin":"","legend":"\u003cp\u003ePrincipal component analysis showing the relationships between spatial variation in \u003cstrong\u003e(a)\u003c/strong\u003e grazing and \u003cstrong\u003e(b)\u003c/strong\u003e browsing with environmental attributes at hot and cold spots (i.e., sites with high or low herbivory rates). Yellow circles: hot spots in grazing (N = 5 hot spots) or browsing (N = 4 hot spots). Blue circles: cold spots in grazing (N = 3 cold spots) or browsing (N = 5 cold spots).\u003c/p\u003e","description":"","filename":"Figure3.tiff.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5160728/v1/b9e95d377566ed76bcf492e9.jpg"},{"id":70532625,"identity":"daa368d2-8192-4fcb-9525-620788e92fb3","added_by":"auto","created_at":"2024-12-04 06:10:22","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1094471,"visible":true,"origin":"","legend":"\u003cp\u003e(\u003cstrong\u003ea)\u003c/strong\u003e Percent change (mean ± SE) in adult \u003cem\u003eTurbinaria\u003c/em\u003e density within macroalgal communities (N = 6 plots per site) along a gradient in ambient browsing (based on short-term browsing assays; sites ordered from lowest to highest levels). Three of the sites were chosen from the original set of twenty sites. Their location on panel b is noted using the same color. \u003cstrong\u003e(b) \u003c/strong\u003eConceptual framework linking variation in herbivory to vulnerability to and reversibility of coral-to-macroalgae state shifts. Vulnerability to shifting to macroalgae is tied to grazing level; sites with low grazing are highly vulnerable to shifting to macroalgae (tan and red quadrants) and those with high grazing will likely remain in a cropped turf state (green and yellow quadrants). Reversibility of a coral-to-macroalgae shift is tied to browsing level; if macroalgae establish, sites with low browsing are likely to remain in a macroalgae state (i.e., low reversibility –yellow and red quadrants) and sites with high browsing could return to the coral state (i.e., high reversibility – tan and green quadrants).\u003c/p\u003e","description":"","filename":"Figure4.tif.jpg","url":"https://assets-eu.researchsquare.com/files/rs-5160728/v1/1c7447631be885e28cec688a.jpg"},{"id":73694131,"identity":"15ea8471-506f-48ba-bfa5-402404fed3e2","added_by":"auto","created_at":"2025-01-13 16:11:20","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3949768,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5160728/v1/bdb1fe12-27a2-486c-a2f0-bef3ae3c3a6a.pdf"},{"id":70532628,"identity":"74fc8fde-8f9a-45f7-be27-741fd95d4737","added_by":"auto","created_at":"2024-12-04 06:10:22","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1739212,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryInformation.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5160728/v1/595c3e719da20545754e94e0.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Patterns of Spatial Covariation in Herbivore Functions on Coral Reefs: Implications for Reef Resilience","fulltext":[{"header":"Introduction","content":"\u003cp\u003eA major focus in ecology is to understand the complex dynamics of ecosystems that simultaneously are being buffeted by rapidly changing disturbance regimes and by more slowly changing environmental drivers. Of considerable interest are abrupt transitions between qualitatively different ecosystem states (Scheffer and Carpenter \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2003\u003c/span\u003e, Conversi et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2015\u003c/span\u003e), which can be challenging to anticipate and problematic to reverse (Mumby et al. \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2013\u003c/span\u003e, Schmitt et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). The consequences to society of such shifts can be profound if an alternative community state provides either fewer or qualitatively different ecosystem services (Suding and Hobbs \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2009\u003c/span\u003e, Mumby et al. \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). This adds urgency to the need to better understand the processes that influence transitions between alternative, self-reinforcing states (regime shifts). In general, shifts in state can occur when the magnitude of a perturbation overwhelms the internal resilience of a system, which itself can be eroded when key biological processes are altered by human activities or natural events. Resolving links between state shifts, biological processes, and ecosystem resilience can inform management strategies intended to maintain desired ecosystem states during this era of unprecedented human impact on the environment (Mcleod et al. \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eCoral reefs are being degraded worldwide by local and global stressors and acute disturbances (Hughes et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2007\u003c/span\u003e, Sully et al. \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Declines in coral cover caused by storms, predator or disease outbreaks, or bleaching frequently have been accompanied by concomitant increases in macroalgae (coral-algae \u0026lsquo;phase shifts\u0026rsquo;) (Roff and Mumby \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2012\u003c/span\u003e, Adam et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Macroalgae generally are superior competitors to coral and can preclude a return to a coral-dominated state even if sufficient coral propagules are present (Hughes et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2007\u003c/span\u003e, Schmitt et al. \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Top-down control of macroalgae by herbivores is widely recognized as a crucial biological process that can influence shifts between coral and macroalgae (Hughes et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2007\u003c/span\u003e, Ledlie et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2007\u003c/span\u003e, Cheal et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, Adam et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2011\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2015\u003c/span\u003e, Holbrook et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Schmitt et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e, Tebbett et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). In this context, despite a broad range of feeding strategies, herbivorous fishes on coral reefs have two general functions (Bellwood et al. \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2004\u003c/span\u003e, Green and Bellwood \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2009\u003c/span\u003e, Burkepile and Hay \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, Cheal et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, Han et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Burkepile et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Tebbett et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), which we operationally term \u0026lsquo;grazing\u0026rsquo; and \u0026lsquo;browsing\u0026rsquo; (e.g., Han et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Following a coral mortality event, grazing herbivores that consume endolithic and filamentous algae \u003cem\u003eprevent\u003c/em\u003e the establishment of macroalgae by consuming their early developmental stages on the disturbed substrate, thereby maintaining the reef surface in a condition (cropped turf) that can be colonized by coral (Birrell et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2005\u003c/span\u003e, Cheal et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, Graham et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2013\u003c/span\u003e, Han et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Robinson et al. \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Adam et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2022\u003c/span\u003e, Schmitt et al. \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e, Tebbett et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). However, if macroalgae do become established on the reef, they subsequently can be removed by species of browsing herbivores that consume mature macroalgae (Cheal et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, Graham et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2013\u003c/span\u003e, Han et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Topor et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, Tebbett et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Thus, browsers, but not grazers, play a critical role in \u003cem\u003ereversing\u003c/em\u003e coral-to-macroalgae state shifts and enhancing the potential for reestablishment of coral (Bellwood et al. \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2006\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIt has been posited that having robust populations of both grazing and browsing species contributes to coral reef resilience (Cheal et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), a notion that modeling of these systems has supported (Cook et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Evidence suggests that grazers typically have higher species richness and biomass than browsers on tropical reefs, both at local reef tract (Adam et al. \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2014\u003c/span\u003e, Han et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Schmitt et al. \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) and larger scales (Cheal et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, V\u0026eacute;rges et al. 2012, Heenan and Williams \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2013\u003c/span\u003e, Edwards et al. \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2014\u003c/span\u003e, Topor et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, Rassweiler et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). This pattern can be exacerbated on reefs where local-scale fisheries heavily target browsing species (Cook et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2024\u003c/span\u003e, Rassweiler et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), potentially lessening the chance of a state shift from macroalgae back to coral. Indeed, persistent shifts to macroalgae on the Great Barrier Reef have been correlated with low abundances of browsing fishes (Cheal et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2010\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eKnowledge of the spatial pattern of covariation in herbivore grazing and browsing functions might provide insight into both the vulnerability of a reef to a coral-to-macroalgae state shift as well as the potential for such a shift to be reversed. There has been an emphasis in prior research and management approaches on the maintenance of coral dominance to prevent a shift to macroalgae, and less on the reversibility of such a shift to re-establish coral dominance (Bellwood et al. \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2006\u003c/span\u003e, Graham et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Here we address that information deficit by exploring patterns and environmental correlates of spatial covariation in grazing and browsing functions of herbivorous fishes, and relating those patterns to both the vulnerability to and potential reversibility of a coral-macroalgae phase shift. The lagoons of Moorea, French Polynesia, provide an ideal model system to explore spatial variation in herbivory functions and the potential consequences of that spatial pattern to coral resilience at the local patch reef scale. Experimental studies suggest coral and macroalgae can be alternative basins of attraction (i.e., represent a state shift, Schmitt et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), and time series data and other process studies show that herbivory has been essential to coral recovery following large mortality events (Adam et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2011\u003c/span\u003e, Holbrook et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Schmitt et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Here we quantified at a set of 20 sites in the north shore lagoons a range of biological and physical characteristics (rates of grazing and browsing, biomass and taxonomic composition of the herbivorous fish community, benthic community composition, productivity of algal turf and of macroalgae, nutrient enrichment and proximity to the reef crest and deep-water channels). These data were used to explore spatial patterns in herbivory within the lagoons. Further, \u0026lsquo;hot spots\u0026rsquo; and \u0026lsquo;cold spots\u0026rsquo; that had either high or low levels of herbivory respectively were identified and the environmental attributes of each type of site were characterized. In the coral reef literature, there is a paucity of experimental attempts that attempt to quantify the potential for reversibility of the macroalgae state. We explored this using a field experiment conducted along a natural gradient in browsing intensity to evaluate whether reversal potential was related to variation in strength of browsing. Our findings have considerable relevance to the development of spatially-explicit management actions to enhance resilience of coral reefs.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePatterns of spatial variation in rates of grazing and browsing\u003c/h2\u003e \u003cp\u003eAssays to estimate the strength of grazing on algal turf and of browsing on macroalgae were conducted at 20 study sites in the lagoon. These revealed high spatial variation in herbivory rates across the lagoon reef system (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Of the sites examined, many had moderate to high levels of grazing, and only a few had low levels of grazing activity (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea, c). By contrast, many sites were characterized by low or moderate levels of browsing, with few having high browsing activity (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb, d). We first explored patterns in grazing and browsing along two spatial gradients: 1) alongshore (west to east along the north shore), and 2) cross-shore between two major reef habitats \u0026ndash; fringing reefs and mid-lagoon reefs. No alongshore trend was apparent in either grazing or browsing \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec, d\u003cb\u003e)\u003c/b\u003e. However, grazing \u0026ndash; but not browsing \u0026ndash; differed between habitats; grazing was higher on mid-lagoon reefs than fringing reefs (grazing: \u003cem\u003et\u003c/em\u003e\u003csub\u003e(11)\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;2.5, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), browsing: Wilcoxon\u0026rsquo;s test W\u0026thinsp;=\u0026thinsp;48, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.9) (\u003cb\u003eSI\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eGrazing and browsing rates were not spatially correlated with each other across the twenty sites (Pearson\u0026rsquo;s \u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.36, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.12). Despite this lack of concordance, a striking spatial pattern was apparent (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Sites clustered into three of the four possible herbivory regimes (i.e., the orthogonal combinations of high and low rates of grazing and browsing), which were: (1) low grazing-low browsing, (2) high grazing-low browsing, and (3) high grazing-high browsing (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). None of the sites were characterized by low grazing but high levels of browsing (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eRelationships between spatial variation in herbivory, herbivore biomass, and environmental attributes\u003c/h3\u003e\n\u003cp\u003ePrincipal component analyses (PCA) explored the degree to which environmental factors [cover of macroalgae and of different types of turf (cropped turf, sedimented turf, cultivated turf within farmerfish \u003cem\u003eStegastes\u003c/em\u003e gardens), productivity of algal turf and of macroalgae, nutrient enrichment, and geographic location (proximity to the reef crest and deep-water channels)] could explain observed spatial patterns in browsing and grazing, specifically, the \u0026lsquo;hot spots\u0026rsquo; (high grazing or browsing activity) and \u0026lsquo;cold spots\u0026rsquo; (low activity) in the lagoon. A subset of sites that reflected the highest (hot spots) and lowest (cold spots) levels in each of the herbivory processes was selected for analysis. We found that different sets of variables explained the variance for each aspect of herbivory. Grazing \u0026lsquo;hot\u0026rsquo; and \u0026lsquo;cold\u0026rsquo; spots (i.e., high and low rates respectively) separated in ordination space along both axes PC1 and PC2, with cumulative proportion of variance explained\u0026thinsp;=\u0026thinsp;0.71 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea). Loading revealed that spatial variation in grazing was related to turf productivity, cover of the different categories of turf, and geographic location. Grazing hot spots were characterized by high productivity and cover of cropped turf as well as increased distance away from deep water drop-offs and closer proximity to the reef crest. Conversely, grazing cold spots were associated with turf containing high sediment loads, high cover of turf algae gardens defended by farmerfish (\u003cem\u003eStegastes\u003c/em\u003e), increased distance from the reef crest and closer proximity to deep water drop-offs.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eBy comparison with grazing, the PCA for browsing revealed that hot and cold spots separated almost entirely along PC2, which was associated with geographic location (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb; cumulative proportion of variance explained for Axes 1 and 2\u0026thinsp;=\u0026thinsp;0.84). Loadings revealed that browsing hot spots tended to be further from the reef crest and closer to deep water drop-offs, whereas cold spots were generally the converse - at sites closer to the reef crest and farther from deep water.\u003c/p\u003e \u003cp\u003ePatterns of grazing were associated with differences in the composition of turf communities between the mid-lagoon and fringing reef habitats. Mid-lagoon sites, which consistently had high grazing rates, were characterized by high cover of cropped turf that contained little or no sediment (Wilcoxon\u0026rsquo;s test\u0026thinsp;=\u0026thinsp;21, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.03, \u003cb\u003eSI\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). Fringing reefs, which supported generally lower and spatially more variable levels of grazing, were characterized by higher cover of turf with high sediment loads (\u003cem\u003et\u003c/em\u003e\u003csub\u003e(3)\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;2.7, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.008, \u003cb\u003eSI\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb) and turf gardens defended by the farmerfish \u003cem\u003eStegastes\u003c/em\u003e (\u003cem\u003et\u003c/em\u003e\u003csub\u003e(3)\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;2.5, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.02, \u003cb\u003eSI\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec).\u003c/p\u003e \u003cp\u003eVisual surveys of the herbivorous fishes conducted by divers at the 20 study sites revealed a total of 19 species. As expected, browsing species were less speciose in the assemblage (4 of 19 species) and they comprised only about three percent of the total biomass and of the total abundance (\u003cb\u003eSI Table\u0026nbsp;1\u003c/b\u003e). Browsers included species in 3 different genera (\u003cem\u003eLeptoscarus\u003c/em\u003e, \u003cem\u003eNaso\u003c/em\u003e, \u003cem\u003eSiganus\u003c/em\u003e). Grazers were represented by 5 genera, with \u003cem\u003eAcanthurus\u003c/em\u003e and \u003cem\u003eScarus\u003c/em\u003e predominant with 7 and 4 species respectively (\u003cb\u003eSI Table\u0026nbsp;1\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eWe did not find the observed spatial variation in our measured rates of grazing or browsing were correlated with our visually-based estimates of fish herbivore biomass among the sites (\u003cb\u003eSI\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Neither the among-site variation in the measured rate of grazing (Pearson\u0026rsquo;s \u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.21, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.37, \u003cb\u003eSI\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea) or browsing (Pearson\u0026rsquo;s \u003cem\u003er\u003c/em\u003e = -0.24, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.3, \u003cb\u003eSI\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb) was related to the estimated local biomass of fish grazers or browsers, respectively.\u003c/p\u003e\n\u003ch3\u003ePredicting variation in herbivore biomass using environmental attributes\u003c/h3\u003e\n\u003cp\u003eDespite the fact that herbivore biomass and herbivory rates were not strongly correlated, spatial variation in the biomass of each herbivore group did map onto the same major environmental variables relating to the spatial patterns in grazing and browsing as revealed by our PCA analyses (\u003cb\u003eSI Table\u0026nbsp;2\u003c/b\u003e). An AIC-based model selection multiple linear regression indicated that variation in grazer biomass across the twenty sites was best predicted by turf productivity and, more weakly, cover of sedimented turf, which together explained 24% of the among-site variation in grazer biomass (\u003cem\u003eF\u003c/em\u003e\u003csub\u003e(2, 7)\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;3.96, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.04). The best model for browsers revealed that variation in macroalgal productivity and distance from the reef crest were the best combination of predictors, which together explained 36% of variation in browser biomass (\u003cem\u003eF\u003c/em\u003e\u003csub\u003e(2, 17)\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;6.28, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.009).\u003c/p\u003e\n\u003ch3\u003eLinking spatial variation in herbivory to the resilience of coral\u003c/h3\u003e\n\u003cp\u003eBecause of a general lack of data that link browsing intensity and reversibility of state shifts to macroalgae, we explored this relationship using a field experiment. Macroalgae-dominated experimental units were exposed to herbivores along a natural gradient in browsing intensity (determined by herbivory assays) to evaluate whether reversal potential was related to variation in browsing. Spatial variation in reversibility was strongly related to browsing intensity; the capacity of browsing herbivores to remove late successional communities of macroalgae increased along a gradient in browsing (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). After 10 days, the average density of adult \u003cem\u003eTurbinaria\u003c/em\u003e within experimental communities in low browsing areas either did not change or decreased by only 6% (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). This implies reefs in areas with low browsing capacity are vulnerable to remaining trapped in a macroalgae-dominated state (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb yellow and red quadrants). By contrast, herbivores in the high browsing site removed 84\u0026thinsp;\u0026plusmn;\u0026thinsp;9% (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SE) of \u003cem\u003eTurbinaria\u003c/em\u003e and in some plots completely removed macroalgal communities, suggesting high potential for reversibility if a shift were to occur (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb tan and green quadrants). We found that reversibility significantly increased with closer proximity to deep water (Pearson\u0026rsquo;s \u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.89, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.044, \u003cb\u003eSI\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), which paralleled the pattern observed for spatial variation in browsing intensity revealed by our short-term assays (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eGrazers and browsers are critical for preventing and reversing state shifts on coral reefs, respectively, and there is concerning evidence that small-scale fisheries may disproportionately reduce browser biomass relative to grazers. Given the posited implications of such a disparity in functional impact of these two herbivore groups for coral-to-macroalgae state shifts (Cheal et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, Schmitt et al. \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e, Cook et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) it is crucial to better understand spatial patterns of grazing and browsing and how variation in these processes impacts the resilience of coral. Here, we investigated spatial variation in grazing and browsing, identified key environmental factors contributing to that variation, and assessed how covariation in these distinct herbivory processes may influence resilience of the coral state in a system where coral and macroalgae have been demonstrated experimentally to be alternative basins of attraction under certain environmental conditions (Schmitt et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eOur study revealed high spatial variation in herbivory rates across lagoon reefs with different patterns emerging for grazing and browsing. Grazing hot spots were more common across lagoons than browsing hot spots, which likely reflects the disparity in species richness and abundance of grazers and browsers in Moorea (\u003cb\u003eSI Table\u0026nbsp;1\u003c/b\u003e). Although we did not detect alongshore or cross-shore gradients in browsing, we found grazing varied between fringing and mid-lagoon reef habitats that differ in distance from shore. Rates of grazing were consistently high on mid-lagoon reefs but variable on fringing reefs. This decline in grazing impact across a reef gradient has been observed in other coral reef systems (Fox and Bellwood \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2007\u003c/span\u003e) and has been attributed to the effect of territorial fishes such as farming damselfish that defend turf gardens and exclude other grazers (Ceccarelli et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2005\u003c/span\u003e), increased rates of sedimentation in nearshore habitats that reduce quality of algal turfs (Goatley and Bellwood \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), and increasing productivity of algae with distance from shore (Klumpp and McKinnon \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e1992\u003c/span\u003e). In our case all three hypotheses are supported; compared to mid-lagoon reefs, fringing reefs had higher cover of \u003cem\u003eStegastes\u003c/em\u003e turf and sedimented turf as well as lower turf productivity. Therefore, in our system fringing reefs appear to be less attractive foraging habitats for grazers due to the quality and accessibility of algal turfs.\u003c/p\u003e \u003cp\u003eBy contrast, characteristics related to the macroalgae community did not appear to explain variation in browsing. Instead, browsing was most influenced by distance from deep water. This relationship between browsing and deep water may not be universal on coral reefs, but in this case, it is likely driven by local fishing dynamics. In Moorea and generally in French Polynesia, browsers such as \u003cem\u003eNaso\u003c/em\u003e (unicornfish) are highly prized for their taste and market value (Nassiri et al. \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Although \u003cem\u003eNaso\u003c/em\u003e are typically difficult to catch and demonstrate fearful behavior of humans – further indicative of the high fishing pressure they face – fishers in Moorea show high selectivity for them (Rassweiler et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Free-diving spearfishers report targeting browsers on reefs near deep water as they believe it provides a spatial refuge for the fish which are much less frequently encountered in the lagoon (\u003cb\u003eSI Table\u0026nbsp;1\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eWe found that estimates of herbivore biomass derived from diver surveys alone were poor predictors of herbivory rate. This decoupling between herbivore biomass and functional impact has been noted previously (e.g., Hoey and Bellwood \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2010\u003c/span\u003e), and can occur for various reasons. The grazers and browsers investigated in our study are mainly comprised of highly mobile herbivores, and therefore may not be counted in visual censuses where they feed (Fox and Bellwood \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Further, as mentioned previously, fearful behavior in response to human presence by some herbivorous fishes can result in low biomass estimates in diver surveys (Gotanda et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2009\u003c/span\u003e) like we observed in this study for browsers (\u003cb\u003eSI Table\u0026nbsp;1\u003c/b\u003e), which are highly prized by local spearfishers (Rassweiler et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2022\u003c/span\u003e, Cook et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Another potential explanation is related to the functional dilution of herbivores as the cover of algae increases, which may result in low herbivory estimates even in areas with high herbivore biomass. Consequently, caution should be exercised when linking measures of abundance to functional impact (Tebbett et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Therefore, we propose that monitoring both herbivore biomass \u003cem\u003eand\u003c/em\u003e rates of algal removal may be the most effective method to evaluate the status of vulnerable reefs, as grazing and browsing assays are relatively rapid and cost-effective, and appear to provide reliable assessments of key trophic processes associated with coral reef resilience (Chong-Seng et al. \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2014\u003c/span\u003e, Nash et al. \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Estimating the two types of herbivory processes, in addition to standing biomass of herbivores, provides a more direct measure of the potential influence of the complementary guilds of herbivores on ecosystem functioning and reef resilience (Goatley et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Cook et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2024\u003c/span\u003e, Tebbett et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eOur results indicate grazing and browsing were spatially variable and not strongly correlated with each other across sites. Experimental work and time-series data demonstrate that reductions in grazing can lead to the proliferation of macroalgae on reefs (Holbrook et al. \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Schmitt et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2019\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), and that browsing may be linked to the likelihood of reversing a coral-to-macroalgae state shift (Bellwood et al. \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2006\u003c/span\u003e, Cheal et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). It follows that sites with low levels of grazing and browsing are likely vulnerable to shifting to and remaining trapped in a macroalgae-dominated state. By contrast, sites with high levels of both herbivory processes may be resilient due to high prevention and reversal potential. Sites characterized by high grazing and low browsing probably have moderate resilience: establishment of macroalgae is unlikely, but so is recovery of the coral state should a shift occur (i.e., a coral-macroalgae ‘phase shift’ rather than a regime shift). Interestingly, our surveys revealed that sites were not evenly distributed among these four possible herbivory regimes. There was a complete lack of low grazing-high browsing sites, which may not be surprising given the lower ratio of browsers to grazers generally in Moorea. Only 15% and 20% of sites fell into low grazing-low browsing and high grazing-high browsing regimes, respectively, whereas the majority of sites (65%) were characterized by high grazing and low browsing.\u003c/p\u003e \u003cp\u003eThe observed pattern of spatial covariation in grazing and browsing functions could be shaped in part by patterns and preferences of fishers in the small-scale fishery of Moorea (Rassweiler et al. \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2022\u003c/span\u003e, Holbrook et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) that, coupled with different life histories that make browsers more susceptible to overfishing (Cook et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), can influence both overall biomass of herbivorous fishes and the disparity between grazers and browsers. In systems such as in Moorea where browsers are target species, the first component of resilience to be weakened through fishing is likely to be the reversibility of a coral-to-macroalgae state shift (i.e., system moves from green to yellow box in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb). As fishing intensity continues to increase, the ability of the herbivore community to prevent a shift to macroalgae will subsequently be eroded (i.e., system moves from the yellow to the red box in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb), and the system will remain trapped in the macroalgae state. This trajectory of reef degradation could represent the future for reefs worldwide that support small-scale fisheries in which herbivores are target species (Cheal et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2010\u003c/span\u003e, Edwards et al. \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2014\u003c/span\u003e, Heenan et al. \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Topor et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). However, knowledge of spatial covariation in herbivore functions could provide a useful template for spatially-explicit management actions tailored to local conditions.\u003c/p\u003e \u003cp\u003eOur results indicate there likely will not be a single strategy to preserve grazing and browsing functions on reefs. Management goals will require different strategies based on key environmental factors and human pressures influencing each herbivory process. For example, managers may need to mitigate sedimentation caused by certain land-use practices to preserve grazing, while limiting fishing may be a better strategy to enhance browsing capacity. Managers may also need to apply strategies in different reef areas since grazing and browsing hot spots do not necessarily coincide in the same locations. One factor we found important in predicting the biomass of grazing and browsing herbivores alike was algal productivity, which itself can be influenced by bottom-up forcing (nutrient enrichment) (Adam et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2021\u003c/span\u003e, Holbrook et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Therefore, management could support grazers and browsers simultaneously by mitigating stressors or disturbances that alter algal productivity in ways that affect top-down control. Finally, management practices historically have focused on means to preserve grazing functions on coral reefs to prevent coral-to-macroalgae state shifts, which is a reasonable priority given the potential challenges in reversing such a shift. However, the frequency and intensity of disturbances to coral reefs that alone can trigger state tipping are predicted to increase. As such, greater attention should be given to enhance protection of browsers and their functional impact to reefs. To that end, our study suggests that understanding spatial covariation in grazing and browsing functions can help better target management actions to enhance resilience of coral.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003c/div\u003e \u003c/div\u003e\n\n "},{"header":"Methods","content":"\u003ch2\u003eStudy site\u003c/h2\u003e\u003cp\u003eThis study was conducted in the shallow lagoons of Moorea, French Polynesia (17°30’ S, 149°50’ W). A barrier reef ~ 1 km offshore protects the lagoons from the open ocean except for 2 to 4 breaks produced by deep reef passes on each of the 3 sides of the island. Back reef habitats shoreward of the barrier reef are characterized by a short band of contiguous reef substrate that then transitions into patch reefs (bommies) surrounded by sand in the mid-lagoon. Mid-lagoon reefs are highly variable with respect to cover of coral, macroalgae, cropped turf algae, and other taxa. Directly adjacent to shore are shallow fringing reefs that can be separated from the mid-lagoon by deep channels. Water circulation within the lagoon is driven by waves forcing water over the crest of the barrier reef, through the lagoons, and out the passes (Leichter et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Hydrodynamic circulation patterns, along with terrestrial run-off, concentrate nutrient enrichment at passes, bays, and beneath major watersheds (Adam et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eAlong Moorea’s north shore, we selected twenty sites spread across the four lagoons that are divided by three reef passes (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea, b). Ten sites were located on the fringing reef, and ten were in the mid-lagoon. At each site we quantified rates of grazing and browsing, as well as biomass and taxonomic composition of the herbivorous fish community, benthic community composition, productivity of algal turf and of macroalgae, and nutrient enrichment. To determine the distance of each site from geographic features of interest (e.g., the barrier reef crest, deep water channels), we used spatial data layers produced from a previous study (Holbrook et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) that mapped the inner edge of the reef crest and the coastline based on LiDAR-based digital elevation maps (Collin et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), as well as satellite imagery provided by Google Earth Pro (Version 7.3.6.9345). For each site, we calculated the minimum distance from the reef crest using the ‘sf’ package (Pebesma \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) in R (version 4.1.1). Then, we used the ruler tool in Google Earth Pro to measure the minimum distance between a site and its nearest deep-water channel, which was visually classified by having deep blue color (as opposed to visible reef substrate or sand) and being located within the lagoon.\u003c/p\u003e\u003ch3\u003eQuantifying spatial patterns in herbivory\u003c/h3\u003e\u003cp\u003eWe deployed assays to quantify levels of both grazing on algal turf and browsing on macroalgae at the 20 sites. For grazing, we exposed uniform pieces of turf-covered reef substrate to herbivores for 3 h between the hours of 10:00 and 16:00. Turf offerings consisted of 7 x 7 cm\u003csup\u003e2\u003c/sup\u003e pieces of dead coral rubble covered by highly palatable turf (e.g., \u003cem\u003ePolysiphonia\u003c/em\u003e spp.) that were collected from gardens in the lagoon that were cultivated by farming damselfish (\u003cem\u003eStegastes\u003c/em\u003e spp.). Each of these was fastened to a rack constructed of PCV-coated galvanized wire mesh that was affixed to open substrate (i.e., lacking coral or macroalgae) on the tops of patch reefs (N = 5 replicates per site). Racks at each site were placed at least 5 m apart. The percent of turf that was consumed was estimated after 3 h by the same observer (see Davis \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2018\u003c/span\u003e, Schmitt et al. \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Two trials of the grazing assay were conducted two weeks apart and all replicates at each site were pooled (N = 10 replicates per site) and then averaged to obtain the percent turf consumed, which we used as an estimate of grazing intensity. Simultaneously, we conducted browsing assays using macroalgae offerings. We attached two 15 cm-long pieces of the palatable brown macroalga, \u003cem\u003eSargassum pacificum\u003c/em\u003e, to a rack using clothes pins to hold them upright and deployed them as described above (N = 5 replicates per site). Lengths of \u003cem\u003eSargassum\u003c/em\u003e pieces were measured after 24 h (see Davis \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) to determine the percent consumed; values for the two fronds deployed together were averaged for each replicate. Three trials of browsing assays were conducted two to three weeks apart and all replicates were pooled (N = 15 replicates per site) and averaged for each site to calculate the percent consumption, our estimate of browsing intensity. To explore spatial patterns of herbivory we created maps of grazing and browsing estimates from the twenty sites using the ‘sf’ package (Pebesma \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) in R (version 4.1.1). Using the R statistical computing software (version 4.3.2; R Core Team \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), covariation between grazing and browsing across sites was determined using Pearson’s correlation coefficient.\u003c/p\u003e\u003ch2\u003eQuantifying herbivorous fish communities and benthic composition\u003c/h2\u003e\u003cp\u003eThe assemblage of herbivorous fishes at each site was characterized using visual surveys in which an observer swam a timed transect, counting and visually estimating total lengths (TLs) of mobile herbivorous fishes ≥ 10 cm TL in a 5 m wide swath. The observer towed an inflatable float behind them with a Garmin GPSMap 78 handheld GPS (Olathe, Kansas, USA) to geo-reference fish counts and provide an estimate of the area covered by the survey (Miller et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). TLs were used to estimate biomass using published species-specific relationships, and species were assigned to functional groups (Moorea Coral Reef LTER and Brooks 2023). Based on the fish biomass estimates and the area covered by the survey, we calculated biomass per unit area, expressed as g/m\u003csup\u003e2\u003c/sup\u003e of grazers and browsers.\u003c/p\u003e\u003cp\u003eBenthic composition was quantified along 50 m x 1 m long transects (N = 3 transects per site), in which the substratum was recorded at every 0.5 m (101 points per transect, 303 points per site). Substratum categories consisted of live scleractinian corals and macroalgae identified to genus, ‘other sessile invertebrates’ (mainly giant clams and sea cucumbers), rubble, sand, and three categories of turf: (1) turf inundated with sediment (hereafter ‘sedimented turf’), (2) turf growing within farmerfish (\u003cem\u003eStegastes nigricans\u003c/em\u003e) gardens (hereafter ‘\u003cem\u003eStegastes\u003c/em\u003e turf’), and (3) closely cropped turf without sediment and located outside of a \u003cem\u003eStegastes\u003c/em\u003e garden (hereafter ‘cropped turf’). The number of points for each substratum category was divided by the total 303 points per site and multiplied by 100 to estimate percent cover.\u003c/p\u003e\u003ch2\u003eQuantifying the nutrient environment and algal productivity\u003c/h2\u003e\u003cp\u003eThe waters surrounding Moorea are highly oligotrophic, so we used the nitrogen tissue content of a common macroalga, \u003cem\u003eTurbinaria ornata\u003c/em\u003e, as a proxy for the local nutrient environment during a period of up to three months prior to \u003cem\u003eTurbinaria\u003c/em\u003e collection (Adam et al. \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2021\u003c/span\u003e, Holbrook et al. \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). \u003cem\u003eTurbinaria\u003c/em\u003e responds to N pulses by storing surplus N (Schaffelke \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e1999\u003c/span\u003e) and N tissue content can provide a time-integrated measure of N availability (Atkinson and Smith \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e1983\u003c/span\u003e, Fong et al. \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e1994\u003c/span\u003e, Shantz et al. \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). In June of 2017, a total of 10 \u003cem\u003eTurbinaria\u003c/em\u003e stipes were randomly collected at each site and transported damp to the laboratory, where 10 florets were removed from each stipe 5 cm from the tip. Samples were dried in a drying for 4 days at 60° C. Total N content was determined via elemental analysis using a CHN Carlo-Erba elemental analyzer (NA1500) at the University of Georgia Center for Applied Isotope Studies.\u003c/p\u003e\u003cp\u003eThe productivity of turf was measured by allowing turf to colonize and grow for 21 days on 2.5 cm\u003csup\u003e2\u003c/sup\u003e unglazed terra cotta tiles in herbivore-exclusion cages (N = 8 caged tiles per site). Herbivore exclusion cages were 10 x 10 x10 cm galvanized mesh, with mesh size of 2.5cm. After the deployment period, tiles were brought to the lab, and turf was removed and processed to obtain ash-free dry weight (AFDW). The mean value of AFDW for each site provided an estimate of turf productivity (g accumulated 3 wk\u003csup\u003e− 1\u003c/sup\u003e). Macroalgal productivity was quantified by allowing juvenile \u003cem\u003eTurbinaria\u003c/em\u003e, and other colonizing macroalgae, to grow protected from herbivores for 8 weeks. We collected reef substratum with attached juvenile \u003cem\u003eTurbinaria\u003c/em\u003e from the mid-lagoon and removed all other algae so only juvenile \u003cem\u003eTurbinaria\u003c/em\u003e (≤ 3 cm length) of uniform size and density remained. The substrates containing \u003cem\u003eTurbinaria\u003c/em\u003e were attached to the bottom of herbivore-exclusion cages at each site (N = 8 per caged replicates per site). After two months, \u003cem\u003eTurbinaria\u003c/em\u003e and any other colonizing macroalgae were removed and damp weighed. Weights of replicates were pooled at a given site to provide an estimate of macroalgal productivity (g accumulated 8 wk\u003csup\u003e− 1\u003c/sup\u003e).\u003c/p\u003e\u003ch2\u003eRelationships between spatial variation in herbivory, herbivore biomass, and environmental attributes\u003c/h2\u003e\u003cp\u003eCovariation between grazing rate and the biomass of grazing fishes from our fish counts across the twenty sites was determined using Pearson’s correlation coefficient (version 4.3.2; R Core Team \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The same analysis was conducted to assess the relationship between browsing rate and the biomass of browsing fishes across sites. We then utilized principal component analysis (PCA) in R (version 4.3.2; R Core Team \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) to explore the degree to which environmental factors could explain observed spatial patterns in browsing and grazing, specifically, the ‘hot spots’ (high grazing or browsing activity) and ‘cold spots’ (low activity) in the lagoon. A subset of sites that reflected the highest (hot spots) and lowest (cold spots) levels in one of the herbivory processes was selected for this analysis, utilizing an additional criterion that the level of the second herbivory process was similar among the group. To explore variation in grazing we chose sites that exhibited the highest or lowest grazing, but that had similar (low) levels of browsing. This selection process resulted in five high grazing sites (i.e., grazing hot spots) and three low grazing sites (i.e., grazing cold spots) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e bottom right and left corners). The full range of browsing only occurred among sites with moderate to high levels of grazing. Four high browsing sites (i.e., browsing hot spots) and five low browsing sites (i.e., browsing cold spots), all of which had high levels of grazing were chosen for analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e right top and bottom corners).\u003c/p\u003e\u003cp\u003eExplanatory variables for the PCA analyses included both spatial factors as well as variables that reflected the amount and productivity of the food resources of the fishes and therefore could reflect the intensity of browsing or grazing. Location in the lagoon can be important predictors of fish spatial patterns of abundance and activity, and each site’s distance from the reef crest and from deep water were included as predictor variables. The amounts of algal cover and productivity impact the distribution and feeding behavior of herbivorous fishes (Tootell and Steele \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). For example, turf can vary in palatability or accessibility to grazing herbivores due to sediment load (Bellwood and Fulton \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2008\u003c/span\u003e, Goatley and Bellwood \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2012\u003c/span\u003e) or whether it is guarded by territorial farming damselfish (\u003cem\u003eStegastes\u003c/em\u003e spp.), with sedimented turf and turf within farmerfish gardens comprising less preferred feeding substrates than unsedimented and undefended turf. Therefore, turf productivity and cover of different types of turf (turf, sedimented turf, turf inside \u003cem\u003eStegastes\u003c/em\u003e gardens) were included in the PCA for grazing, and cover and productivity of macroalgae for browsing. The analysis was based on a correlation matrix of these data.\u003c/p\u003e\u003ch2\u003ePredicting variation in herbivore biomass using environmental attributes\u003c/h2\u003e\u003cp\u003eTo explore whether environmental attributes associated with hot spots and cold spots in herbivory also predict variation in the biomass of herbivores, we utilized the best subset selection approach, which is an exploratory model building regression analysis. Using predictor variables from the grazing PCA analysis, we utilized the R package “leaps” (Version 3.1) to test all possible combinations of predictors in a multiple linear regression to explain variation in grazer biomass, and then selected the best model according to lowest AIC score with the R package “AICcmodavg” (Version 2.3). The same procedure was done for browsers using predictors from the browsing PCA.\u003c/p\u003e\u003ch2\u003eLinking spatial variation in herbivory to the resilience of coral\u003c/h2\u003e\u003cp\u003eBecause of a general lack of data that link browsing intensity and reversibility of state shifts to macroalgae, we explored this relationship using a field experiment. Macroalgae-dominated communities were exposed to herbivores along a natural gradient in browsing intensity to evaluate whether reversal potential was related to variation in browsing. We selected five sites that represented a range in browsing intensity based on short-term herbivory assays described above. At each site, we deployed macroalgae communities that mimicked patch reefs that had undergone a shift to macroalgal dominance. Each community was constructed by chiseling off reef substrate with attached \u003cem\u003eTurbinaria\u003c/em\u003e, which was then assembled into a community that reflected the size and density of macroalgae-dominated patch reefs within the lagoon. This assemblage was then epoxied to a cinderblock and the density of adult \u003cem\u003eTurbinaria\u003c/em\u003e stipes recorded. The cinderblock macroalgal community (N = 6 replicates per site) was then deployed at a depth within 2–5 m and left exposed to herbivores for 10 days. At the end of the experiment, final adult \u003cem\u003eTurbinaria\u003c/em\u003e density was recorded. The percent change in density of adult \u003cem\u003eTurbinaria\u003c/em\u003e was calculated for each replicate using the equation \u003cem\u003e(Final – Initial) / Initial x 100\u003c/em\u003e. Replicate values were then averaged for each site and used as a proxy for reversibility of a coral-to-macroalgae state shift.\u003c/p\u003e\u003cp\u003eLastly, we were interested in whether environmental features associated with variation in browsing were also good predictors of reversibility. Covariation between reversibility and distance from deep water (log-transformed) for the five sites was determined using Pearson’s correlation coefficient (version 4.3.2; R Core Team \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003e \u003cb\u003eAdditional Information\u003c/b\u003e \u003c/h2\u003e \u003cp\u003e \u003cstrong\u003eCompeting Interests Statement\u003c/strong\u003e: The author(s) declare no competing interests.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eDana T. Cook (DTC), Sally J. Holbrook (SJH) and Russell J. Schmitt (RJS) collectively designed the study; laboratory and field work was done primarily by DTC with contributions by SJH and RJS; DTC performed the statistical analyses and data visualizations; DTC, SJH and RJS co-wrote the paper with DTC producing the initial draft; SJH and RJS secured the funding.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eWe thank Andrew Rassweiler, Sarah Lester, Matthew Lauer and Jean Wenc\u0026eacute;lius for fruitful discussions, Hillary Krumbholz for Information Management assistance, and Lauren Enright, Olivia Isbell, Taylor Traxler and Kai Kopecky for assistance in the field. We also thank the staff of the University of California Gump Research Station for logistic support. This work is a contribution of the NSF Moorea Coral Reef Long Term Ecological Research Site and was supported by National Science Foundation grants OCE-1637396, BCS-1714704, and OCE-2224354. Field research was completed under permits issued by the Territorial Government of French Polynesia (D\u0026eacute;l\u0026eacute;gation \u0026agrave; la Recherche) and the Haut-commissariat de la R\u0026eacute;publique en Polyn\u0026eacute;sie Fran\u0026ccedil;aise (DTRT) (Protocole d\u0026rsquo;Accueil 2006-2022); we thank them for their continued support. With respect to the spelling of Moorea, we followed the Raapoto transcription system that is adhered to by a large segment of the Tahitian community, but also recognize other community members follow the Te Fare Vanā\u0026rsquo;a transcription system where the island name is spelled with an \u0026rsquo;eta (Mo\u0026rsquo;orea).\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eAll data sets and code utilized for this research are publicly available on GitHub [https://github.com/danatcook/herbivory_hotspots]. Final versions of data and code files will be permanently deposited with the Environmental Data Initiative (EDI) upon manuscript acceptance.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAdam, T. C. \u003cem\u003eet al\u003c/em\u003e. How will coral reef fish communities respond to climate-driven disturbances? Insight from landscape-scale perturbations. \u003cem\u003eOecologia\u003c/em\u003e \u003cstrong\u003e176\u003c/strong\u003e,285-296 (2014).\u003c/li\u003e\n\u003cli\u003eAdam, T. C. \u003cem\u003eet al\u003c/em\u003e. Landscape-scale patterns of nutrient enrichment in a coral reef ecosystem: implications for coral to algae phase shifts. \u003cem\u003eEcol. Appl\u003c/em\u003e. \u003cstrong\u003e31\u003c/strong\u003e, e02227 (2021).\u003c/li\u003e\n\u003cli\u003eAdam, T. C., Burkepile, D. E., Ruttenberg, B. I. \u0026amp; Paddack, M. J. Herbivory and the resilience of Caribbean coral reefs: knowledge gaps and implications for management. \u003cem\u003eMar\u003c/em\u003e. \u003cem\u003eEcol\u003c/em\u003e. \u003cem\u003eProg\u003c/em\u003e. \u003cem\u003eSer\u003c/em\u003e. \u003cstrong\u003e520\u003c/strong\u003e, 1-20 (2015).\u003c/li\u003e\n\u003cli\u003eAdam, T. C. \u003cem\u003eet al\u003c/em\u003e. Priority effects in coral\u0026ndash;macroalgae interactions can drive alternate community paths in the absence of top-down control. \u003cem\u003eEcology\u003c/em\u003e \u003cstrong\u003e103\u003c/strong\u003e,e3831 (2022).\u003c/li\u003e\n\u003cli\u003eAdam, T. C. \u003cem\u003eet al\u003c/em\u003e. Herbivory, connectivity, and ecosystem resilience: response of a coral reef to a large-scale perturbation. \u003cem\u003ePLoS ONE\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, e23717 (2011).\u003c/li\u003e\n\u003cli\u003eAtkinson, M. J. \u0026amp; Smith, S. V. C:N:P ratios of benthic marine plants. \u003cem\u003eLimnol. Oceanogr.\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, 568-574 (1983).\u003c/li\u003e\n\u003cli\u003eBellwood, D. R. \u0026amp; Fulton, C. J. Sediment-mediated suppression of herbivory on coral reefs: decreasing resilience to rising sea levels and climate change? \u003cem\u003eLimnol. Oceanogr.\u003c/em\u003e \u003cstrong\u003e53\u003c/strong\u003e, 2695-2701 (2008).\u003c/li\u003e\n\u003cli\u003eBellwood, D. R., Hughes, T. P., Folke, C. \u0026amp; Nystr\u0026ouml;m, M. Confronting the coral reef crisis. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e429\u003c/strong\u003e, 827-833 (2004).\u003c/li\u003e\n\u003cli\u003eBellwood, D. R., Hughes, T. P. \u0026amp; Hoey, A. S. Sleeping functional group drives coral-reef recovery. \u003cem\u003eCurr. Biol.\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e,2434-2439 (2006).\u003c/li\u003e\n\u003cli\u003eBirrell, C. L., McCook, L. J. \u0026amp; Willis, B. L. Effects of algal turfs and sediment on coral settlement. \u003cem\u003eMar\u003c/em\u003e. \u003cem\u003ePollut\u003c/em\u003e. \u003cem\u003eBull\u003c/em\u003e. \u003cstrong\u003e51\u003c/strong\u003e, 408-414 (2005).\u003c/li\u003e\n\u003cli\u003eBurkepile, D. E. \u0026amp; Hay, M. E. Impact of herbivore identity on algal succession and coral growth on a Caribbean reef. \u003cem\u003ePLoS ONE\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, e8963 (2010).\u003c/li\u003e\n\u003cli\u003eBurkepile, D. E., Schmitt, M. H., Stears, K., Donovan, M. K. \u0026amp; Thompson, D. I. Shared insights across the ecology of coral reefs and African savannas: are parrotfish wet wildebeest? \u003cem\u003eBioScience\u003c/em\u003e \u003cstrong\u003e70\u003c/strong\u003e, 647-658 (2020).\u003c/li\u003e\n\u003cli\u003eCeccarelli, D. M., Jones, G. P. \u0026amp; McCook, L. J. Foragers versus farmers: contrasting effects of two behavioural groups of herbivores on coral reefs. \u003cem\u003eOecologia\u003c/em\u003e \u003cstrong\u003e145\u003c/strong\u003e, 445-453 (2005).\u003c/li\u003e\n\u003cli\u003eCheal, A. J. \u003cem\u003eet al\u003c/em\u003e. Coral \u0026ndash; macroalgal phase shifts or reef resilience: links with diversity and functional roles of herbivorous fishes on the Great Barrier Reef. \u003cem\u003eCoral Reefs\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 1005-1015 (2010).\u003c/li\u003e\n\u003cli\u003eChong-Seng, K. M., Nash, K. L.,Bellwood, D. R. \u0026amp; Graham, N. A. J. Macroalgal herbivory on recovering versus degrading coral reefs. \u003cem\u003eCoral Reefs\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 409-419 (2014).\u003c/li\u003e\n\u003cli\u003eCollin, A. \u003cem\u003eet al\u003c/em\u003e. Very high resolution mapping of coral reef state using airborne bathymetric LiDAR surface-intensity and drone imagery. \u003cem\u003eInt\u003c/em\u003e. \u003cem\u003eJ\u003c/em\u003e. \u003cem\u003eRemote\u003c/em\u003e \u003cem\u003eSens\u003c/em\u003e. \u003cstrong\u003e39\u003c/strong\u003e, 5676-5688 (2018).\u003c/li\u003e\n\u003cli\u003eConversi, A. \u003cem\u003eet al\u003c/em\u003e. A holistic view of marine regime shifts. \u003cem\u003ePhil. Trans. R. Soc. B: Biol. Sci.\u003c/em\u003e \u003cstrong\u003e370\u003c/strong\u003e, 20130279 (2015).\u003c/li\u003e\n\u003cli\u003eCook, D. T., Schmitt, R. J., Holbrook, S. J. \u0026amp; Moeller, H. V. Modeling the effects of selectively fishing key functional groups of herbivores on coral resilience. \u003cem\u003eEcosphere\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e\u003cstrong\u003e(1)\u003c/strong\u003e, e4749 (2024). \u003c/li\u003e\n\u003cli\u003eDavis, S. L. Associational refuge facilitates phase shifts to macroalgae in a coral reef ecosystem. \u003cem\u003eEcosphere\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, e02272 (2018).\u003c/li\u003e\n\u003cli\u003eEdwards, C. B. \u003cem\u003eet al\u003c/em\u003e. Global assessment of the status of coral reef herbivorous fishes: evidence for fishing effects. \u003cem\u003eProc. R. Soc. B.\u003c/em\u003e \u003cstrong\u003e281\u003c/strong\u003e, 20131835 (2014).\u003c/li\u003e\n\u003cli\u003eFong, P., Donohoe, R. M. \u0026amp; Zedler, J. B. Nutrient concentration in tissue of the macroalga \u003cem\u003eEnteromorpha\u003c/em\u003e as a function of nutrient history: an experimental evaluation using field microcosms. \u003cem\u003eMar. Ecol. Prog. Ser.\u003c/em\u003e \u003cstrong\u003e106\u003c/strong\u003e, 273-281 (1994).\u003c/li\u003e\n\u003cli\u003eFox, R. J. \u0026amp; Bellwood, D. R. Quantifying herbivory across a coral reef depth gradient. \u003cem\u003eMar. Ecol. Prog. Ser.\u003c/em\u003e \u003cstrong\u003e339\u003c/strong\u003e, 49-59 (2007).\u003c/li\u003e\n\u003cli\u003eFox, R. J. \u0026amp; Bellwood, D. R. Herbivores in a small world: network theory highlights vulnerability in the function of herbivory on coral reefs. \u003cem\u003eFunc\u003c/em\u003e. \u003cem\u003eEcol\u003c/em\u003e. \u003cstrong\u003e28\u003c/strong\u003e, 642-651 (2014).\u003c/li\u003e\n\u003cli\u003eGoatley, C. H. R. \u0026amp; Bellwood, D. R. Sediment suppresses herbivory across a coral reef depth gradient. \u003cem\u003eBiol\u003c/em\u003e. \u003cem\u003eLett\u003c/em\u003e. \u003cstrong\u003e8\u003c/strong\u003e, 1016-1018 (2012).\u003c/li\u003e\n\u003cli\u003eGoatley, C. H. R., Bonaldo, R. M., Fox, R. J. \u0026amp; Bellwood, D. R. Sediments and herbivory as sensitive indicators of coral reef degradation. \u003cem\u003eEcol. Soc\u003c/em\u003e. \u003cstrong\u003e21(1)\u003c/strong\u003e, 29 (2016).\u003c/li\u003e\n\u003cli\u003eGotanda, K. M., Turgeon, K. \u0026amp; Kramer, D. L. Body size and reserve protection affect flight initiation distance in parrotfishes. \u003cem\u003eBehav. Ecol. Sociobiol\u003c/em\u003e. \u003cstrong\u003e63\u003c/strong\u003e, 1563-1572 (2009).\u003c/li\u003e\n\u003cli\u003eGraham, N. A. \u003cem\u003eet al\u003c/em\u003e. Managing resilience to reverse phase shifts in coral reefs. \u003cem\u003eFront\u003c/em\u003e. \u003cem\u003eEcol\u003c/em\u003e. \u003cem\u003eEnviron\u003c/em\u003e. \u003cstrong\u003e11\u003c/strong\u003e, 541-548 (2013).\u003c/li\u003e\n\u003cli\u003eGreen, A. L., \u0026amp; Bellwood, D. R. \u003cem\u003eMonitoring Functional Groups of Herbivorous Reef Fishes as Indicators of Coral Reef Resilience: A Practical Guide for Coral Reef Managers in the Asia Pacific Region\u003c/em\u003e (IUCN, Gland, Switzerland, 2009).\u003c/li\u003e\n\u003cli\u003eHan, X., Adam, T. C., Schmitt, R. J., Brooks, A. J. \u0026amp; and Holbrook, S. J. Response of herbivore functional groups to sequential perturbations in Moorea, French Polynesia. \u003cem\u003eCoral Reefs\u003c/em\u003e \u003cstrong\u003e35\u003c/strong\u003e, 999-1009 (2016).\u003c/li\u003e\n\u003cli\u003eHeenan, A. \u0026amp; Williams, I. D. Monitoring herbivorous fishes as indicators of coral reef resilience in American Samoa. \u003cem\u003ePLoS ONE\u003c/em\u003e \u003cstrong\u003e8(11)\u003c/strong\u003e, e79604 (2013).\u003c/li\u003e\n\u003cli\u003eHeenan, A., Hoey, A. S., Williams, G. J. \u0026amp; Williams, I. D. Natural bounds on herbivorous coral reef fishes. \u003cem\u003eProc. R. Soc. B.\u003c/em\u003e \u003cstrong\u003e283\u003c/strong\u003e, 20161716 (2016).\u003c/li\u003e\n\u003cli\u003eHoey, A. S. \u0026amp; Bellwood, D. R. Cross-shelf variation in browsing intensity on the Great Barrier Reef. \u003cem\u003eCoral Reefs\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 499-508 (2010).\u003c/li\u003e\n\u003cli\u003eHolbrook, S. J., Schmitt, R. J., Adam, T. C. \u0026amp; Brooks, A. J. Coral reef resilience, tipping points and the strength of herbivory. \u003cem\u003eSci. Rep\u003c/em\u003e. \u003cstrong\u003e6\u003c/strong\u003e, 35817 (2016).\u003c/li\u003e\n\u003cli\u003eHolbrook, S. J. \u003cem\u003eet al\u003c/em\u003e. Spatial covariation in nutrient enrichment and fishing of herbivores in an oceanic coral reef ecosystem. \u003cem\u003eEcol\u003c/em\u003e. \u003cem\u003eAppl\u003c/em\u003e. 32, e2515 (2022).\u003c/li\u003e\n\u003cli\u003eHughes, T. P. \u003cem\u003eet al\u003c/em\u003e. Phase shifts, herbivory, and the resilience of coral reefs to climate change. \u003cem\u003eCurr. Biol\u003c/em\u003e. \u003cstrong\u003e17\u003c/strong\u003e, 360-365 (2007).\u003c/li\u003e\n\u003cli\u003eKlumpp, D. W. \u0026amp; McKinnon, A. D. Community structure, biomass and productivity of epilithic algal communities on the Great Barrier Reef: dynamics at different spatial scales. \u003cem\u003eMar\u003c/em\u003e. \u003cem\u003eEcol\u003c/em\u003e. \u003cem\u003eProg\u003c/em\u003e. \u003cem\u003eSer\u003c/em\u003e. \u003cstrong\u003e86\u003c/strong\u003e, 77-89 (1992).\u003c/li\u003e\n\u003cli\u003eLedlie, M. H. \u003cem\u003eet al\u003c/em\u003e. Phase shifts and the role of herbivory in the resilience of coral reefs. \u003cem\u003eCoral\u003c/em\u003e \u003cem\u003eReefs\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, 641-653 (2007).\u003c/li\u003e\n\u003cli\u003eLeichter, J. \u003cem\u003eet al\u003c/em\u003e. Biological and physical interactions on a tropical island coral reef: transport and retention processes on Moorea, French Polynesia. \u003cem\u003eOceanography\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, 52-63 (2013).\u003c/li\u003e\n\u003cli\u003eMcleod, E. \u003cem\u003eet al\u003c/em\u003e. The future of resilience-based management in coral reef ecosystems. \u003cem\u003eJ\u003c/em\u003e. \u003cem\u003eEnviron\u003c/em\u003e. \u003cem\u003eManage\u003c/em\u003e. \u003cstrong\u003e233\u003c/strong\u003e, 291-301 (2019).\u003c/li\u003e\n\u003cli\u003eMiller, S. D. \u003cem\u003eet al\u003c/em\u003e. Using machine learning to achieve simultaneous, georeferenced surveys of fish and benthic communities on shallow coral reefs. \u003cem\u003eLimnol. Oceanogr.\u003c/em\u003e \u003cem\u003eMethods\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 451-466 (2023).\u003c/li\u003e\n\u003cli\u003eMoorea Coral Reef LTER \u0026amp; Brooks, A. MCR LTER: Coral Reef: Long-term population and community dynamics: fishes, ongoing since 2005 ver 62. Environmental Data Initiative. https://doi.org/10.6073/pasta/75644add7e7f90c568bf5045264d359a (2023).\u003c/li\u003e\n\u003cli\u003eMumby, P. J., Steneck, R. S. \u0026amp; Hastings, A. Evidence for and against the existence of alternate attractors on coral reefs. \u003cem\u003eOikos\u003c/em\u003e \u003cstrong\u003e122\u003c/strong\u003e, 481-491 (2013).\u003c/li\u003e\n\u003cli\u003eNash, K. L., Abesamis, R. A., Graham, N. A. J., McClure, E. C. \u0026amp; Moland, E. Drivers of herbivory on coral reefs: species, habitat and management effects. \u003cem\u003eMar\u003c/em\u003e. \u003cem\u003eEcol\u003c/em\u003e. \u003cem\u003eProg\u003c/em\u003e. \u003cem\u003eSer\u003c/em\u003e. \u003cstrong\u003e554\u003c/strong\u003e, 129-140 (2016).\u003c/li\u003e\n\u003cli\u003eNassiri, A. \u003cem\u003eet al\u003c/em\u003e. Hedonic evaluation of coral reef fish prices on a direct sale market. \u003cem\u003eMar\u003c/em\u003e. \u003cem\u003ePolicy \u003c/em\u003e\u003cstrong\u003e129\u003c/strong\u003e, 104525 (2021).\u003c/li\u003e\n\u003cli\u003ePebesma, E. Simple features for R: Standardized support for spatial vector data. \u003cem\u003eThe R Journal\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 439-446 (2018). \u003c/li\u003e\n\u003cli\u003eR Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/ (2023).\u003c/li\u003e\n\u003cli\u003eRassweiler, A. \u003cem\u003eet al\u003c/em\u003e. Perceptions and responses of Pacific Island fishers to changing coral reefs. \u003cem\u003eAmbio\u003c/em\u003e \u003cstrong\u003e49\u003c/strong\u003e, 130-143 (2020).\u003c/li\u003e\n\u003cli\u003eRassweiler, A. \u003cem\u003eet al\u003c/em\u003e. How do fisher responses to macroalgal overgrowth influence the resilience of coral reefs? \u003cem\u003eLimnol\u003c/em\u003e. \u003cem\u003eOceanogr\u003c/em\u003e. \u003cstrong\u003e67\u003c/strong\u003e, S365-S377 (2022).\u003c/li\u003e\n\u003cli\u003eRobinson, J. P. W. \u003cem\u003eet al\u003c/em\u003e. Habitat and fishing control grazing potential on coral reefs. \u003cem\u003eFunc\u003c/em\u003e. \u003cem\u003eEcol\u003c/em\u003e. \u003cstrong\u003e34\u003c/strong\u003e, 240-251 (2020).\u003c/li\u003e\n\u003cli\u003eRoff, G. \u0026amp; Mumby, P. J. Global disparity in the resilience of coral reefs. \u003cem\u003eTrends Ecol. Evol\u003c/em\u003e. \u003cstrong\u003e27\u003c/strong\u003e, 404-413 (2012).\u003c/li\u003e\n\u003cli\u003eSchaffelke, B. Short-term nutrient pulses as tools to assess responses of coral reef macroalgae to enhanced nutrient availability. \u003cem\u003eMar. Ecol. Prog. Ser\u003c/em\u003e. \u003cstrong\u003e182\u003c/strong\u003e, 305-310 (1999).\u003c/li\u003e\n\u003cli\u003eScheffer, M. \u0026amp; Carpenter, S. R. Catastrophic regime shifts in ecosystems: linking theory to observation. \u003cem\u003eTrends Ecol. Evol\u003c/em\u003e. \u003cstrong\u003e18\u003c/strong\u003e, 648-656 (2003).\u003c/li\u003e\n\u003cli\u003eSchmitt, R. J., Holbrook, S. J., Brooks, A. J. \u0026amp; Adam, T. C. Evaluating the precariousness of coral recovery when coral and macroalgae are alternative basins of attraction. \u003cem\u003eLimnol\u003c/em\u003e. \u003cem\u003eOceanogr\u003c/em\u003e. \u003cstrong\u003e67\u003c/strong\u003e, S285-S297 (2022).\u003c/li\u003e\n\u003cli\u003eSchmitt, R. J., Holbrook, S. J., Davis, S. L., Brooks, A. J. \u0026amp; Adam, T. C. Experimental support for alternative attractors on coral reefs. \u003cem\u003eProc. Nat. Acad. Sci\u003c/em\u003e. \u003cstrong\u003e116\u003c/strong\u003e, 4372-4381 (2019).\u003c/li\u003e\n\u003cli\u003eShantz, A. A., Ladd, M. C., Schrack, E. \u0026amp; Burkepile, D. E. Fish-derived nutrient hotspots shape coral reef benthic communities. \u003cem\u003eEcol. Appl\u003c/em\u003e. \u003cstrong\u003e25\u003c/strong\u003e, 2142-2152 (2015).\u003c/li\u003e\n\u003cli\u003eSuding, K. N. \u0026amp; Hobbs, R. J. Threshold models in restoration and conservation: a developing framework. \u003cem\u003eTrends Ecol. Evol\u003c/em\u003e. \u003cstrong\u003e24\u003c/strong\u003e, 271-279 (2009).\u003c/li\u003e\n\u003cli\u003eSully, S., Burkepile, D. E., Donovan, M. K., Hodgson, G. \u0026amp; Van Woesik, R. A global analysis of coral bleaching over the past two decades. \u003cem\u003eNat. Commun\u003c/em\u003e. \u003cstrong\u003e10\u003c/strong\u003e, 1264 (2019).\u003c/li\u003e\n\u003cli\u003eTebbett, S. B., Bennett, S. \u0026amp; Bellwood, D. R. A functional perspective on the meaning of the term \u0026lsquo;herbivore\u0026rsquo;: patterns versus process in coral reef fishes. \u003cem\u003eCoral Reefs\u003c/em\u003e \u003cstrong\u003e43\u003c/strong\u003e, 219-232 (2024).\u003c/li\u003e\n\u003cli\u003eTootell, J. S. \u0026amp; Steele, M. A. Distribution, behavior, and condition of herbivorous fishes on coral reefs track algal resources. \u003cem\u003eOecologia\u003c/em\u003e \u003cstrong\u003e181\u003c/strong\u003e, 13-24 (2016).\u003c/li\u003e\n\u003cli\u003eTopor, Z. M., Rasher, D. B., Duffy, J. E. \u0026amp; Brandl, S. J. Marine protected areas enhance coral reef functioning by promoting fish biodiversity. \u003cem\u003eConserv. Lett\u003c/em\u003e. \u003cstrong\u003e12\u003c/strong\u003e, e12638 (2019).\u003c/li\u003e\n\u003cli\u003eVerg\u0026eacute;s A., Bennett, S. \u0026amp; Bellwood, D. R. Diversity among macroalgae-consuming fishes on coral reefs: a transcontinental comparison. \u003cem\u003ePLoS ONE\u003c/em\u003e \u003cstrong\u003e7(9)\u003c/strong\u003e, e45543 (2012).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"herbivory, reversal of state shifts, grazing, browsing, community recovery, vulnerability and reversibility","lastPublishedDoi":"10.21203/rs.3.rs-5160728/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5160728/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003ePersistent shifts to undesired ecological states, such as shifts from coral to macroalgae, are becoming more common. This highlights the need to understand processes that can help restore affected ecosystems. Herbivory on coral reefs is widely recognized as a key interaction that can keep macroalgae from outcompeting coral. Most attention has been on the role \u0026lsquo;grazing\u0026rsquo; herbivores play in preventing the establishment of macroalgae, while less research has focused on the role of \u0026lsquo;browsers\u0026rsquo; in extirpating macroalgae. Here we explored patterns, environmental correlates and state shift consequences of spatial co-variation in grazing and browsing functions of herbivorous fishes. Grazing and browsing rates were not highly correlated across 20 lagoon sites in Moorea, French Polynesia, but did cluster into 3 (of 4) combinations of high and low consumption rates (no site had low grazing but high browsing). Consumption rates were not correlated with grazer or browser fish biomass, but both were predicted by specific environmental variables. Experiments revealed that reversibility of a macroalgal state shift was strongly related to spatial variation in browsing intensity. Our findings provide insights and simple diagnostic tools regarding heterogeneity in top-down forcing that influences the vulnerability to and reversibility of shifts to macroalgae on coral reefs.\u003c/p\u003e","manuscriptTitle":"Patterns of Spatial Covariation in Herbivore Functions on Coral Reefs: Implications for Reef Resilience","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-12-04 06:10:14","doi":"10.21203/rs.3.rs-5160728/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-10-30T14:24:50+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-29T00:24:23+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-25T03:27:31+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"182787742694120777410948802679740922437","date":"2024-10-21T15:12:28+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"334578315280145657443151896921210710467","date":"2024-10-14T14:37:48+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"172452441504831074875684733828368565046","date":"2024-10-13T22:08:22+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"122510942172545838841461692747001372496","date":"2024-10-13T21:20:10+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-10-11T09:07:25+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-10-10T09:05:54+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2024-10-10T08:36:22+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-10-10T08:34:48+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2024-09-26T19:08:24+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"6a1a5e79-0c3d-47ea-be57-46bef72493b8","owner":[],"postedDate":"December 4th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":39622109,"name":"Biological sciences/Ecology/Community ecology"},{"id":39622110,"name":"Earth and environmental sciences/Ecology/Community ecology"}],"tags":[],"updatedAt":"2025-01-13T16:04:45+00:00","versionOfRecord":{"articleIdentity":"rs-5160728","link":"https://doi.org/10.1038/s41598-024-83672-1","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2025-01-07 15:58:04","publishedOnDateReadable":"January 7th, 2025"},"versionCreatedAt":"2024-12-04 06:10:14","video":"","vorDoi":"10.1038/s41598-024-83672-1","vorDoiUrl":"https://doi.org/10.1038/s41598-024-83672-1","workflowStages":[]},"version":"v1","identity":"rs-5160728","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5160728","identity":"rs-5160728","version":["v1"]},"buildId":"_2-kVJe1T_tPrBINL-cwx","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}