The Potential for Alternative Stable States in Food Webs Depends on Feedback Mechanism and Trait Diversity.

AbstractAlternative stable ecosystem states are possible under the same environmental conditions in models of two or three interacting species and an array of feedback loops. However, multispecies food webs might weaken the feedbacks loops that can create alternative stable states. To test how this potential depends on food web properties, we develop a many-species model where consumer Allee effects emerge from consumer-resource interactions. We evaluate the interactive effects of food web connectance, interspecific trait diversity, and two classes of feedbacks: specialized feedbacks, where consumption of individual resources declines at high resource abundance (e.g., from schooling or reaching size refugia), and aggregate feedbacks, where overall resource abundance reduces consumer recruitment (e.g., from resources enhancing competition or mortality experienced by recruits). We find that aggregate feedbacks maintain, and specialized feedbacks reduce, the potential for alternative states. Interspecific trait diversity decreases the prevalence of alternative stable states more for specialized than for aggregate feedbacks. Increasing food web connectance increases the potential for alternative stable states for aggregated feedbacks but decreases it for specialized feedbacks, where losing vulnerable consumers can cascade into food web collapses. Altogether, multispecies food webs can limit the set of processes that create alternative stable states and impede consumer recovery from disturbance.

How far to build it before they come? Analyzing the use of the Field of Dreams hypothesis in bull kelp restoration.

In restoration ecology, the Field of Dreams hypothesis posits that restoration efforts that create a suitable environment could lead to the eventual recovery of the remaining aspects of the ecosystem through natural processes. Natural processes following partial restoration has led to ecosystem recovery in both terrestrial and aquatic systems. However, understanding the efficacy of a "Field of Dreams" approach requires a comparison of different approaches to partial restoration in terms of spatial, temporal, and ecological scale with what would happen given more comprehensive restoration efforts. We explore the relative effect of partial restoration and ongoing recovery on restoration efficacy with a dynamical model based on temperate rocky reefs in Northern California. We analyze our model for both the ability and rate of bull kelp forest recovery under different restoration strategies. We compare the efficacy of a partial restoration approach with a more comprehensive restoration effort by exploring how kelp recovery likelihood and rate change with varying intensities of urchin removal and kelp outplanting over different time periods and spatial scales. We find that, in the case of bull kelp forests, setting more favorable initial conditions for kelp recovery by implementing both urchin harvesting and kelp outplanting at the start of the restoration project has a bigger impact on the kelp recovery rate than applying restoration efforts through a longer period of time. Therefore, partial restoration efforts, in terms of spatial and temporal scale, can be significantly more effective when applied across multiple ecological scales in terms of both the capacity and rate for achieving the target outcomes.

Fast behavioral feedbacks make ecosystems sensitive to pace and not just magnitude of anthropogenic environmental change.

Anthropogenic environmental change is altering the behavior of animals in ecosystems around the world. Although behavior typically occurs on much faster timescales than demography, it can nevertheless influence demographic processes. Here, we use detailed data on behavior and empirical estimates of demography from a coral reef ecosystem to develop a coupled behavioral-demographic ecosystem model. Analysis of the model reveals that behavior and demography feed back on one another to determine how the ecosystem responds to anthropogenic forcing. In particular, an empirically observed feedback between the density and foraging behavior of herbivorous fish leads to alternative stable ecosystem states of coral population persistence or collapse (and complete algal dominance). This feedback makes the ecosystem more prone to coral collapse under fishing pressure but also more prone to recovery as fishing is reduced. Moreover, because of the behavioral feedback, the response of the ecosystem to changes in fishing pressure depends not only on the magnitude of changes in fishing but also on the pace at which changes are imposed. For example, quickly increasing fishing to a given level can collapse an ecosystem that would persist under more gradual change. Our results reveal conditions under which the pace and not just the magnitude of external forcing can dictate the response of ecosystems to environmental change. More generally, our multiscale behavioral-demographic framework demonstrates how high-resolution behavioral data can be incorporated into ecological models to better understand how ecosystems will respond to perturbations.

Characterizing Long Transients in Consumer-Resource Systems With Group Defense and Discrete Reproductive Pulses.

During recent years, the study of long transients has been expanded in ecological theory to account for shifts in long-term behavior of ecological systems. These long transients may lead to regime shifts between alternative states that resemble the dynamics of alternative stable states for a prolonged period of time. One dynamic that potentially leads to long transients is the group defense of a resource in a consumer-resource interaction. Furthermore, time lags in the population caused by discrete reproductive pulses have the potential to produce long transients, either independently or in conjunction to the transients caused by the group defense. In this work, we analyze the potential for long transients in a model for a consumer-resource system in which the resource exhibits group defense and reproduces in discrete reproductive pulses. This system exhibits crawl-by transients near the extinction and carrying capacity states of resource, and a transcritical bifurcation, under which a ghost limit cycle appears. We estimate the transient time of our system from these transients using perturbation theory. This work advances an understanding of how systems shift between alternate states and their duration of staying in a given regime and what ecological dynamics may lead to long transients.

Grazer behaviour can regulate large-scale patterning of community states.

Ecosystem patterning can arise from environmental heterogeneity, biological feedbacks that produce multiple persistent ecological states, or their interaction. One source of feedbacks is density-dependent changes in behaviour that regulate species interactions. By fitting state-space models to large-scale (~500 km) surveys on temperate rocky reefs, we find that behavioural feedbacks best explain why kelp and urchin barrens form either reef-wide patches or local mosaics. Best-supported models in California include feedbacks where starvation intensifies grazing across entire reefs create reef-scale, alternatively stable kelp- and urchin-dominated states (32% of reefs). Best-fitting models in New Zealand include the feedback of urchins avoiding dense kelp stands that can increase abrasion and predation risk, which drives a transition from shallower urchin-dominated to deeper kelp-dominated zones, with patchiness at 3-8 m depths with intermediate wave stress. Connecting locally studied processes with region-wide data, we highlight how behaviour can explain community patterning and why some systems exhibit community-wide alternative stable states.

Transient dynamics during kelp forest recovery from fishing across multiple trophic levels.

Outcomes of management efforts to recover or restore populations of harvested species can be highly dependent on environmental and community context. Predator-prey interactions can alter recovery trajectories, and the timing of management actions within multi-trophic level harvest scenarios may influence the dynamics of recovery and lead to management trade-offs. Recent work using a generalist predator-prey model suggests that management promoting synchronized recovery of predators and prey leads to faster and less variable recovery trajectories than sequential recovery (predator or prey first). However, more complex communities may require different management actions to minimize recovery time and variability. Here, we use a tri-trophic level rocky reef community dynamics model with size-structure and fisheries at multiple trophic levels to investigate the importance of three ecological processes to recovery of fished communities: (1) size-structured predation, (2) non-consumptive effects of predators on prey behavior, and (3) varying levels of recruitment. We also test the effects of initiating recovery from community states associated with varying degrees of fishery-induced degradation and develop a simulation in which the basal resource (kelp) is harvested. In this system, a predator-first closure generally leads to the least volatile and quickest recovery, whether from a kelp forest, urchin barren, or intermediate community state. The benefits gained by selecting this strategy are magnified when recovering from the degraded community, the urchin barren, because initial conditions in the degraded state lead to lengthy recovery times. However, the shape of the size-structured predation relationship can strongly affect recovery volatility, where the differences between alternate management strategies are negated with size-independent predation. External recruitment reduces return times by bolstering the predatory lobster population. These results show that in a tightly linked tri-trophic level food web with top-down control, a predator-first fishery closure can be the most effective strategy to reduce volatility and shorten recovery, particularly when the system is starting from the degraded community state. Given the ubiquity of top predator loss across many ecosystems, we highlight the value of incorporating insights from community ecology into ecosystem management.

Quantifying the statistical power of monitoring programs for marine protected areas.

Marine Protected Areas (MPAs) are increasingly established globally as a spatial management tool to aid in conservation and fisheries management objectives. Assessing whether MPAs are having the desired effects on populations requires effective monitoring programs. A cornerstone of an effective monitoring program is an assessment of the statistical power of sampling designs to detect changes when they occur. We present a novel approach to power assessment that combines spatial point process models, integral projection models (IPMs) and sampling simulations to assess the power of different sample designs across a network of MPAs. We focus on the use of remotely operated vehicle (ROV) video cameras as the sampling method, though the results could be extended to other sampling methods. We use empirical data from baseline surveys of an example indicator fish species across three MPAs in California, USA as a case study. Spatial models simulated time series of spatial distributions across sites that accounted for the effects of environmental covariates, while IPMs simulated expected trends over time in abundances and sizes of fish. We tested the power of different levels of sampling effort (i.e., the number of 500-m ROV transects) and temporal replication (every 1-3 yr) to detect expected post-MPA changes in fish abundance and biomass. We found that changes in biomass are detectable earlier than changes in abundance. We also found that detectability of MPA effects was higher in sites with higher initial densities. Increasing the sampling effort had a greater effect than increasing sampling frequency on the time taken to achieve high power. High power was best achieved by combining data from multiple sites. Our approach provides a powerful tool to explore the interaction between sampling effort, spatial distributions, population dynamics, and metrics for detecting change in previously fished populations.

Identifying robust strategies for assisted migration in a competitive stochastic metacommunity.

Assisted migration (AM) is the translocation of species beyond their historical range to locations that are expected to be more suitable under future climate change. However, a relocated population may fail to establish in its donor community if there is high uncertainty in decision-making, climate, and interactions with the recipient ecological community. To quantify the benefit to persistence and risk of establishment failure of AM under different management scenarios (e.g., choosing target species, proportion of population to relocate, and optimal location to relocate), we built a stochastic metacommunity model to simulate several species reproducing, dispersing, and competing on a temperature gradient as temperature increases over time. Without AM, the species were vulnerable to climate change when they had low population sizes, short dispersal, and strong poleward competition. When relocating species that exemplified these traits, AM increased the long-term persistence of the species most when relocating a fraction of the donor population, even if the remaining population was very small or rapidly declining. This suggests that leaving behind a fraction of the population could be a robust approach, allowing managers to repeat AM in case they move the species to the wrong place and at the wrong time, especially when it is difficult to identify a species' optimal climate. We found that AM most benefitted species with low dispersal ability and least benefited species with narrow thermal tolerances, for which AM increased extinction risk on average. Although relocation did not affect the persistence of nontarget species in our simple competitive model, researchers will need to consider a more complete set of community interactions to comprehensively understand invasion potential.

Identificación de Estrategias Sólidas para la Migración Asistida en una Metacomunidad Estocástica Competitiva Resumen La migración asistida es la translocación de especies más allá de su extensión histórica a localidades que se espera sean más adecuadas bajo el cambio climático futuro. Sin embargo, una población reubicada puede no establecerse en su comunidad donante si existe una mucha incertidumbre en cuanto a la toma de decisiones, el clima y las interacciones con la comunidad ecológica receptora. Para cuantificar el beneficio para la persistencia y el riesgo de fallas en el establecimiento de la migración asistida bajo diferentes escenarios de manejo (p. ej.: elección de especies objetivo, proporción de la población a reubicar y localidad óptima para la reubicación) construimos un modelo de metacomunidad estocástica para simular la reproducción, dispersión y competencia de varias especies a lo largo de un gradiente de temperatura conforme la temperatura incrementa con el tiempo. Sin la migración asistida, las especies presentaron vulnerabilidad ante el cambio climático cuando presentaron un tamaño poblacional menor, una dispersión reducida y una competencia fuerte hacia los extremos. Cuando se reubicó a especies con estas características, la migración asistida incrementó más la persistencia a largo plazo de las especies cuando se reubicó una fracción de la población donante, incluso si la población remanente era muy pequeña o se encontraba en una rápida declinación. Esto sugiere que dejar una fracción de la población podría ser una estrategia sólida que permite a los gestores repetir la migración asistida en caso de que muden a la especie al lugar equivocado en el momento equivocado, especialmente cuando es difícil identificar el clima óptimo de la especie. Descubrimos que la migración asistida benefició más a las especies con una baja habilidad de dispersión y tuvo menos beneficios para las especies con una tolerancia térmica reducida, para las que la migración asistida aumentó en promedio el riesgo de extinción. Aunque la reubicación no afectó la persistencia de las especies que no consideramos como objetivo en nuestro modelo competitivo simple, los investigadores necesitarán considerar un conjunto más completo de interacciones comunitarias para entender por completo el potencial de invasión.

Economic value of ecological information in ecosystem-based natural resource management depends on exploitation history.

Ecosystem approaches to natural resource management are seen as a way to provide better outcomes for ecosystems and for people, yet the nature and strength of interactions among ecosystem components is usually unknown. Here we characterize the economic benefits of ecological knowledge through a simple model of fisheries that target a predator (piscivore) and its prey. We solve for the management (harvest) trajectory that maximizes net present value (NPV) for different ecological interactions and initial conditions that represent different levels of exploitation history. Optimal management trajectories generally approached similar harvest levels, but the pathways toward those levels varied considerably by ecological scenario. Application of the wrong harvest trajectory, which would happen if one type of ecological interaction were assumed but in fact another were occurring, generally led to only modest reductions in NPV. However, the risks were not equal across fleets: risks of incurring large losses of NPV and missing management targets were much higher in the fishery targeting piscivores, especially when piscivores were heavily depleted. Our findings suggest that the ecosystem approach might provide the greatest benefits when used to identify system states where management performs poorly with imperfect knowledge of system linkages so that management strategies can be adopted to avoid those states.

Protected areas buffer against harvest selection and rebuild phenotypic complexity.

Harvest mortality typically truncates the harvested species' size structure, thereby reducing phenotypic complexity, which can lead to reduced population productivity, increased population variability, and selection on an array of life history traits that can further alter these demographic processes. Marine protected areas (MPAs) are a potential tool to protect older, larger individuals and therefore mitigate such ecological and evolutionary effects of harvest, depending on the degree of connectivity among areas. Such MPA protection relies on a shift in size-dependent mortality, the measurement of which can therefore serve as an early indicator of whether MPAs might achieve the desired longer-term ecological and evolutionary responses. We directly measured MPA effects on size-selective mortality and associated size structure using mark-recapture data on European lobster (Homarus gammarus) collected at three MPA-control area pairs in southern Norway during one decade (n = 5,943). Mark-recapture modeling, accounting for variation in recapture probabilities, revealed (1) that annual mean survival was higher inside MPAs (0.592) vs. control areas (0.298) and (2) that significant negative relationships between survival and body size occurred at the control areas but not in the MPAs, where the effect of body size was predominantly positive. Additionally, we found (3) that mean and maximum body size increased over time inside MPAs but not in control areas. Overall, our results suggest that MPAs can rebuild phenotypic complexity (i.e., size structure) and provide protection from harvest selection.

Partitioning colony size variation into growth and partial mortality.

Body size is a trait that broadly influences the demography and ecology of organisms. In unitary organisms, body size tends to increase with age. In modular organisms, body size can either increase or decrease with age, with size changes being the net difference between modules added through growth and modules lost through partial mortality. Rates of colony extension are independent of body size, but net growth is allometric, suggesting a significant role of size-dependent mortality. In this study, we develop a generalizable model of partitioned growth and partial mortality and apply it to data from 11 species of reef-building coral. We show that corals generally grow at constant radial increments that are size independent, and that partial mortality acts more strongly on small colonies. We also show a clear life-history trade-off between growth and partial mortality that is governed by growth form. This decomposition of net growth can provide mechanistic insights into the relative demographic effects of the intrinsic factors (e.g. acquisition of food and life-history strategy), which tend to affect growth, and extrinsic factors (e.g. physical damage, and predation), which tend to affect mortality.

Life history and temporal variability of escape events interactively determine the fitness consequences of aquaculture escapees on wild populations.

Domesticated individuals are likely to be maladaptive in the wild due to adaptation to captivity. Escaped aquaculture fish can cause unintended fitness and demographic consequences for their wild conspecifics through interbreeding and competition. Escape events from different sources exhibit great heterogeneity in their frequencies and magnitudes, ranging from rare but large spillover during a storm, to continuous low-level leakage caused by operational errors. The timescale of escape events determines the distribution of gene flow from aquaculture to the wild. The evolutionary consequences of this variation in timescale will depend on the degree of generation overlap and the focal species' life history attributes, especially those under selection in aquaculture (e.g., growth rate, which can influence additional demographically important traits such as age at maturity). To evaluate the effects of variable escape both within and across generations, we construct an age-structured model of coupled genetic and demographic dynamics and parameterize it for species with contrasting life history characteristics (Salmo salar and Gadus morhua). Our results are consistent with earlier discrete-generation models that constant, low-level spillover can have a greater impact than rare, large pulses of leakage, even after accounting for the averaging effects of overlapping generations. The age-structured model also allows detailed evaluation of the role of different life history traits, which reveals that species with longer generation times might experience greater fitness consequences of aquaculture spillover but are less sensitive to variability in spillover. Additionally, environment-induced earlier maturity of escapees can increase the fitness effects on wild fish, especially those with shorter generation times. Our results suggest that effective management to minimize the unintended fitness consequences of aquaculture releases might require extensive monitoring efforts on constant, low-level spillover and assessment of the focal species' life history characteristics.

Setting expected timelines of fished population recovery for the adaptive management of a marine protected area network.

Adaptive management of marine protected areas (MPAs) requires developing methods to evaluate whether monitoring data indicate that they are performing as expected. Modeling the expected responses of targeted species to an MPA network, with a clear timeline for those expectations, can aid in the development of a monitoring program that efficiently evaluates expectations over appropriate time frames. Here, we describe the expected trajectories in abundance and biomass following MPA implementation for populations of 19 nearshore fishery species in California. To capture the process of filling in the age structure truncated by fishing, we used age-structured population models with stochastic larval recruitment to predict responses to MPA implementation. We implemented both demographically open (high larval immigration) and closed (high self-recruitment) populations to model the range of possible trajectories as they depend on recruitment dynamics. From these simulations, we quantified the time scales over which anticipated increases in abundance and biomass inside MPAs would become statistically detectable. Predicted population biomass responses range from little change, for species with low fishing rates, to increasing by a factor of nearly seven, for species with high fishing rates before MPA establishment. Increases in biomass following MPA implementation are usually greater in both magnitude and statistical detectability than increases in abundance. For most species, increases in abundance would not begin to become detectable for at least 10 years after implementation. Overall, these results inform potential indicator metrics (biomass), potential indicator species (those with a high fishing : natural mortality ratio), and time frame (>10 yr) for MPA monitoring assessment as part of the adaptive management process.

Integrating Genetic and Demographic Effects of Connectivity on Population Stability: The Case of Hatchery Trucking in Salmon.

Connectivity among populations can have counteracting effects on population stability. Demographically, connectivity can rescue local populations but increase the synchrony across populations. Genetically, connectivity can counteract drift locally but homogenize genotypes across populations. Population independence and diversity underlies system-level buffering against environmental variability, termed the portfolio effect. The portfolio effect has declined in California fall-run Chinook salmon, possibly in part because of the trucking of juvenile hatchery-reared fish for downstream release, which reduces juvenile mortality but increases the connectivity between rivers. We use a dynamical population model to test whether this increased connectivity can explain the loss of the portfolio effect and quantify the relative demographic and genetic contributions to portfolio effect erosion. In the model, populations experience different within-population environmental conditions and the same time-variable ocean conditions, the response to which can depend on a quantitative genetic trait. We find that increased trucking for one population's hatchery can lead to a loss of the portfolio effect, with a system-level trade-off between increased average abundance and increased variability in abundance. This trade-off is much stronger when we include the effects of genetic homogenization than when we consider demographic synchronization alone. Therefore, genetic homogenization can outweigh demographic synchrony in determining the system-level effect of connectivity.

Interactive effects of predator and prey harvest on ecological resilience of rocky reefs.

A major goal of ecosystem-based fisheries management is to prevent fishery-induced shifts in community states. This requires an understanding of ecological resilience: the ability of an ecosystem to return to the same state following a perturbation, which can strongly depend on species interactions across trophic levels. We use a structured model of a temperate rocky reef to explore how multi-trophic level fisheries impact ecological resilience. Increasing fishing mortality of prey (urchins) has a minor effect on equilibrium biomass of kelp, urchins, and spiny lobster predators, but increases resilience by reducing the range of predator harvest rates at which alternative stable states are possible. Size-structured predation on urchins acts as the feedback maintaining each state. Our results demonstrate that the resilience of ecosystems strongly depends on the interactive effects of predator and prey harvest in multi-trophic level fisheries, which are common in marine ecosystems but are unaccounted for by traditional management.

Predicting evolutionary rescue via evolving plasticity in stochastic environments.

Phenotypic plasticity and its evolution may help evolutionary rescue in a novel and stressful environment, especially if environmental novelty reveals cryptic genetic variation that enables the evolution of increased plasticity. However, the environmental stochasticity ubiquitous in natural systems may alter these predictions, because high plasticity may amplify phenotype-environment mismatches. Although previous studies have highlighted this potential detrimental effect of plasticity in stochastic environments, they have not investigated how it affects extinction risk in the context of evolutionary rescue and with evolving plasticity. We investigate this question here by integrating stochastic demography with quantitative genetic theory in a model with simultaneous change in the mean and predictability (temporal autocorrelation) of the environment. We develop an approximate prediction of long-term persistence under the new pattern of environmental fluctuations, and compare it with numerical simulations for short- and long-term extinction risk. We find that reduced predictability increases extinction risk and reduces persistence because it increases stochastic load during rescue. This understanding of how stochastic demography, phenotypic plasticity, and evolution interact when evolution acts on cryptic genetic variation revealed in a novel environment can inform expectations for invasions, extinctions, or the emergence of chemical resistance in pests.

Exploring the effect of the spatial scale of fishery management.

For any spatially explicit management, determining the appropriate spatial scale of management decisions is critical to success at achieving a given management goal. Specifically, managers must decide how much to subdivide a given managed region: from implementing a uniform approach across the region to considering a unique approach in each of one hundred patches and everything in between. Spatially explicit approaches, such as the implementation of marine spatial planning and marine reserves, are increasingly used in fishery management. Using a spatially explicit bioeconomic model, we quantify how the management scale affects optimal fishery profit, biomass, fishery effort, and the fraction of habitat in marine reserves. We find that, if habitats are randomly distributed, the fishery profit increases almost linearly with the number of segments. However, if habitats are positively autocorrelated, then the fishery profit increases with diminishing returns. Therefore, the true optimum in management scale given cost to subdivision depends on the habitat distribution pattern.

Fitting state-space integral projection models to size-structured time series data to estimate unknown parameters.

Integral projection models (IPMs) have a number of advantages over matrix-model approaches for analyzing size-structured population dynamics, because the latter require parameter estimates for each age or stage transition. However, IPMs still require appropriate data. Typically they are parameterized using individual-scale relationships between body size and demographic rates, but these are not always available. We present an alternative approach for estimating demographic parameters from time series of size-structured survey data using a Bayesian state-space IPM (SSIPM). By fitting an IPM in a state-space framework, we estimate unknown parameters and explicitly account for process and measurement error in a dataset to estimate the underlying process model dynamics. We tested our method by fitting SSIPMs to simulated data; the model fit the simulated size distributions well and estimated unknown demographic parameters accurately. We then illustrated our method using nine years of annual surveys of the density and size distribution of two fish species (blue rockfish, Sebastes mystinus, and gopher rockfish, S. carnatus) at seven kelp forest sites in California. The SSIPM produced reasonable fits to the data, and estimated fishing rates for both species that were higher than our Bayesian prior estimates based on coast-wide stock assessment estimates of harvest. That improvement reinforces the value of being able to estimate demographic parameters from local-scale monitoring data. We highlight a number of key decision points in SSIPM development (e.g., open vs. closed demography, number of particles in the state-space filter) so that users can apply the method to their own datasets.

Marine reserves can enhance ecological resilience.

The goals of ecosystem-based management (EBM) include protecting ecological resilience, the magnitude of a perturbation that a community can withstand and remain in a given state. As a tool to achieve this goal, no-take marine reserves may enhance resilience by protecting source populations or reduce it by concentrating fishing in harvested areas. Here, we test whether spatial management with marine reserves can increase ecological resilience compared to non-spatial (conventional) management using a dynamic model of a simplified fish community with structured predation and competition that causes alternative stable states. Relative to non-spatial management, reserves increase the resilience of the desired (predator-dominated) equilibrium state in both stochastic and deterministic environments, especially under intensive fishing. As a result, spatial management also increases the feasibility of restoring degraded (competitor-dominated) systems, particularly if combined with culling of competitors or stock enhancement of adult predators.

The differential effects of increasing frequency and magnitude of extreme events on coral populations.

Extreme events, which have profound ecological consequences, are changing in both frequency and magnitude with climate change. Because extreme temperatures induce coral bleaching, we can explore the relative impacts of changes in frequency and magnitude of high temperature events on coral reefs. Here, we combined climate projections and a dynamic population model to determine how changing bleaching regimes influence coral persistence. We additionally explored how coral traits and competition with macroalgae mediate changes in bleaching regimes. Our results predict that severe bleaching events reduce coral persistence more than frequent bleaching. Corals with low adult mortality and high growth rates are successful when bleaching is mild, but bleaching resistance is necessary to persist when bleaching is severe, regardless of frequency. The existence of macroalgae-dominated stable states reduces coral persistence and changes the relative importance of coral traits. Building on previous studies, our results predict that management efforts may need to prioritize protection of "weaker" corals with high adult mortality when bleaching is mild, and protection of "stronger" corals with high bleaching resistance when bleaching is severe. In summary, future reef projections and conservation targets depend on both local bleaching regimes and biodiversity.