Factors influencing the detectability of early warning signals of population collapse

Early warning indicators capture catastrophic transitions driven by explicit rates of environmental change

In response to external changes, ecosystems can undergo catastrophic transitions. Early warning indicators aim to predict such transitions based on the phenomenon of critical slowing down at bifurcation points found under a constant environment. When an explicit rate of environmental change is considered, catastrophic transitions can become distinct phenomena from bifurcations, and result from a delayed response to noncatastrophic bifurcations. We use a trophic metacommunity model where transitions in time series and bifurcations of the system are distinct phenomena. We calculate early warning indicators from the time series of the continually changing system and show that they predict not the bifurcation of the underlying system but the actual catastrophic transition driven by the explicit rate of change. Predictions based on the bifurcation structure could miss catastrophic transitions that can still be captured by early warning signals calculated from time series. Our results expand the repertoire of mechanistic models used to anticipate catastrophic transitions to nonequilibrium ecological systems exposed to a constant rate of environmental change.

A predictive timeline of wildlife population collapse

Contemporary rates of biodiversity decline emphasize the need for reliable ecological forecasting, but current methods vary in their ability to predict the declines of real-world populations. Acknowledging that stressor effects start at the individual level, and that it is the sum of these individual-level effects that drives populations to collapse, shifts the focus of predictive ecology away from using predominantly abundance data. Doing so opens new opportunities to develop predictive frameworks that utilize increasingly available multi-dimensional data, which have previously been overlooked for ecological forecasting. Here, we propose that stressed populations will exhibit a predictable sequence of observable changes through time: changes in individuals' behaviour will occur as the first sign of increasing stress, followed by changes in fitness-related morphological traits, shifts in the dynamics (for example, birth rates) of populations and finally abundance declines. We discuss how monitoring the sequential appearance of these signals may allow us to discern whether a population is increasingly at risk of collapse, or is adapting in the face of environmental change, providing a conceptual framework to develop new forecasting methods that combine multi-dimensional (for example, behaviour, morphology, life history and abundance) data.

Early warnings for multi-stage transitions in dynamics on networks

Successfully anticipating sudden major changes in complex systems is a practical concern. Such complex systems often form a heterogeneous network, which may show multi-stage transitions in which some nodes experience a regime shift earlier than others as an environment gradually changes. Here we investigate early warning signals for networked systems undergoing a multi-stage transition. We found that knowledge of both the ongoing multi-stage transition and network structure enables us to calculate effective early warning signals for multi-stage transitions. Furthermore, we found that small subsets of nodes could anticipate transitions as well as or even better than using all the nodes. Even if we fix the network and dynamical system, no single best subset of nodes provides good early warning signals, and a good choice of sentinel nodes depends on the tipping direction and the current stage of the dynamics within a multi-stage transition, which we systematically characterize.

Quantifying resilience and the risk of regime shifts under strong correlated noise

Early warning indicators often suffer from the shortness and coarse-graining of real-world time series. Furthermore, the typically strong and correlated noise contributions in real applications are severe drawbacks for statistical measures. Even under favourable simulation conditions the measures are of limited capacity due to their qualitative nature and sometimes ambiguous trend-to-noise ratio. In order to solve these shortcomings, we analyze the stability of the system via the slope of the deterministic term of a Langevin equation, which is hypothesized to underlie the system dynamics close to the fixed point. The open-source available method is applied to a previously studied seasonal ecological model under noise levels and correlation scenarios commonly observed in real world data. We compare the results to autocorrelation, standard deviation, skewness, and kurtosis as leading indicator candidates by a Bayesian model comparison with a linear and a constant model. We show that the slope of the deterministic term is a promising alternative due to its quantitative nature and high robustness against noise levels and types. The commonly computed indicators apart from the autocorrelation with deseasonalization fail to provide reliable insights into the stability of the system in contrast to a previously performed study in which the standard deviation was found to perform best. In addition, we discuss the significant influence of the seasonal nature of the data to the robust computation of the various indicators, before we determine approximately the minimal amount of data per time window that leads to significant trends for the drift slope estimations.

Community structure determines the predictability of population collapse

Early warning signals (EWS) are phenomenological tools that have been proposed as predictors of the collapse of biological systems. Although a growing body of work has shown the utility of EWS based on either statistics derived from abundance data or shifts in phenotypic traits such as body size, so far this work has largely focused on single species populations.

However, to predict reliably the future state of ecological systems, which inherently could consist of multiple species, understanding how reliable such signals are in a community context is critical.

Here, reconciling quantitative trait evolution and Lotka–Volterra equations, which allow us to track both abundance and mean traits, we simulate the collapse of populations embedded in mutualistic and multi‐trophic predator–prey communities. Using these simulations and warning signals derived from both population‐ and community‐level data, we showed the utility of abundance‐based EWS, as well as metrics derived from stability‐landscape theory (e.g. width and depth of the basin of attraction), were fundamentally linked. Thus, the depth and width of such stability‐landscape curves could be used to identify which species should exhibit the strongest EWS of collapse.

The probability a species displays both trait and abundance‐based EWS was dependent on its position in a community, with some species able to act as indicator species. In addition, our results also demonstrated that in general trait‐based EWS were less reliable in comparison with abundance‐based EWS in forecasting species collapses in our simulated communities. Furthermore, community‐level abundance‐based EWS were fairly reliable in comparison with their species‐level counterparts in forecasting species‐level collapses.

Our study suggests a holistic framework that combines abundance‐based EWS and metrics derived from stability‐landscape theory that may help in forecasting species loss in a community context.

Avoiding critical thresholds through effective monitoring

A major challenge in sustainability science is identifying targets that maximize ecosystem benefits to humanity while minimizing the risk of crossing critical system thresholds. One critical threshold is the biomass at which populations become so depleted that their population growth rates become negative-depensation. Here, we evaluate how the value of monitoring information increases as a natural resource spends more time near the critical threshold. This benefit emerges because higher monitoring precision promotes higher yield and a greater capacity to recover from overharvest. We show that precautionary buffers that trigger increased monitoring precision as resource levels decline may offer a way to minimize monitoring costs and maximize profits. In a world of finite resources, improving our understanding of the trade-off between precision in estimates of population status and the costs of mismanagement will benefit stakeholders that shoulder the burden of these economic and social costs.

Machine learning methods trained on simple models can predict critical transitions in complex natural systems

Forecasting sudden changes in complex systems is a critical but challenging task, with previously developed methods varying widely in their reliability. Here we develop a novel detection method, using simple theoretical models to train a deep neural network to detect critical transitions-the Early Warning Signal Network (EWSNet). We then demonstrate that this network, trained on simulated data, can reliably predict observed real-world transitions in systems ranging from rapid climatic change to the collapse of ecological populations. Importantly, our model appears to capture latent properties in time series missed by previous warning signals approaches, allowing us to not only detect if a transition is approaching, but critically whether the collapse will be catastrophic or non-catastrophic. These novel properties mean EWSNet has the potential to serve as an indicator of transitions across a broad spectrum of complex systems, without requiring information on the structure of the system being monitored. Our work highlights the practicality of deep learning for addressing further questions pertaining to ecosystem collapse and has much broader management implications.

Early warning signal reliability varies with COVID-19 waves

Early warning signals (EWSs) aim to predict changes in complex systems from phenomenological signals in time series data. These signals have recently been shown to precede the emergence of disease outbreaks, offering hope that policymakers can make predictive rather than reactive management decisions. Here, using a novel, sequential analysis in combination with daily COVID-19 case data across 24 countries, we suggest that composite EWSs consisting of variance, autocorrelation and skewness can predict nonlinear case increases, but that the predictive ability of these tools varies between waves based upon the degree of critical slowing down present. Our work suggests that in highly monitored disease time series such as COVID-19, EWSs offer the opportunity for policymakers to improve the accuracy of urgent intervention decisions but best characterize hypothesized critical transitions.

Probabilistic early warning signals

Ecological communities and other complex systems can undergo abrupt and long-lasting reorganization, a regime shift, when deterministic or stochastic factors bring them to the vicinity of a tipping point between alternative states. Such changes can be large and often arise unexpectedly. However, theoretical and experimental analyses have shown that changes in correlation structure, variance, and other standard indicators of biomass, abundance, or other descriptive variables are often observed prior to a state shift, providing early warnings of an anticipated transition. Natural systems manifest unknown mixtures of ecological and environmental processes, hampered by noise and limited observations. As data quality often cannot be improved, it is important to choose the best modeling tools available for the analysis.We investigate three autoregressive models and analyze their theoretical differences and practical performance. We formulate a novel probabilistic method for early warning signal detection and demonstrate performance improvements compared to nonprobabilistic alternatives based on simulation and publicly available experimental time series.The probabilistic formulation provides a novel approach to early warning signal detection and analysis, with enhanced robustness and treatment of uncertainties. In real experimental time series, the new probabilistic method produces results that are consistent with previously reported findings.Robustness to uncertainties is instrumental in the common scenario where mechanistic understanding of the complex system dynamics is not available. The probabilistic approach provides a new family of robust methods for early warning signal detection that can be naturally extended to incorporate variable modeling assumptions and prior knowledge.

Density dependence and the spread of invasive big-headed ants (Pheidole megacephala) in an East African savanna

Supercolonial ants are among the largest cooperative units in nature, attaining extremely high densities. How these densities feed back into their population growth rates and how abundance and extrinsic factors interact to affect their population dynamics remain open questions. We studied how local worker abundance and extrinsic factors (rain, tree density) affect population growth rate and spread in the invasive big-headed ant, which is disrupting a keystone mutualism between acacia trees and native ants in parts of East Africa. We measured temporal changes in big-headed ant (BHA) abundance and rates of spread over 20 months along eight transects, extending from areas behind the front with high BHA abundances to areas at the invasion front with low BHA abundances. We used models that account for negative density dependence and incorporated extrinsic factors to determine what variables best explain variation in local population growth rates. Population growth rates declined with abundance, however, the strength of density dependence decreased with abundance. We suggest that weaker density dependence at higher ant abundances may be due to the beneficial effect of cooperative behavior that partially counteracts resource limitation. Rainfall and tree density had minor effects on ant population dynamics. BHA spread near 50 m/year, more than previous studies reported and comparable to rates of spread of other supercolonial ants. Although we did not detect declines in abundance in areas invaded a long time ago (> 10 years), continued monitoring of abundance at invaded sites may help to better understand the widespread collapse of many invasive ants.

Prospects for detecting early warning signals in discrete event sequence data: Application to epidemiological incidence data

Early warning signals (EWS) identify systems approaching a critical transition, where the system undergoes a sudden change in state. For example, monitoring changes in variance or autocorrelation offers a computationally inexpensive method which can be used in real-time to assess when an infectious disease transitions to elimination. EWS have a promising potential to not only be used to monitor infectious diseases, but also to inform control policies to aid disease elimination. Previously, potential EWS have been identified for prevalence data, however the prevalence of a disease is often not known directly. In this work we identify EWS for incidence data, the standard data type collected by the Centers for Disease Control and Prevention (CDC) or World Health Organization (WHO). We show, through several examples, that EWS calculated on simulated incidence time series data exhibit vastly different behaviours to those previously studied on prevalence data. In particular, the variance displays a decreasing trend on the approach to disease elimination, contrary to that expected from critical slowing down theory; this could lead to unreliable indicators of elimination when calculated on real-world data. We derive analytical predictions which can be generalised for many epidemiological systems, and we support our theory with simulated studies of disease incidence. Additionally, we explore EWS calculated on the rate of incidence over time, a property which can be extracted directly from incidence data. We find that although incidence might not exhibit typical critical slowing down properties before a critical transition, the rate of incidence does, presenting a promising new data type for the application of statistical indicators.

Effect of time series length and resolution on abundance- and trait-based early warning signals of population declines

Natural populations are increasingly threatened with collapse at the hands of anthropogenic effects. Predicting population collapse with the help of generic early warning signals (EWS) may provide a prospective tool for identifying species or populations at highest risk. However, pattern-to-process methods such as EWS have a multitude of challenges to overcome to be useful, including the low signal-to-noise ratio of ecological systems and the need for high quality time series data. The inclusion of trait dynamics with EWS has been proposed as a more robust tool to predict population collapse. However, the length and resolution of available time series are highly variable from one system to another, especially when generation time is considered. As yet, it remains unknown how this variability with regards to generation time will alter the efficacy of EWS. Here we take both a simulation- and experimental-based approach to assess the impacts of relative time series length and resolution on the forecasting ability of EWS. We show that EWS' performance decreases with decreasing time-series length. However, there was no evident decrease in EWS performance as resolution decreased. Our simulations suggest a relative time series length between 10 and five generations as a minimum requirement for accurate forecasting by abundance-based EWS. However, when trait information is included alongside abundance-based EWS, we find positive signals at lengths one-half of what was required without them. We suggest that, in systems where specific traits are known to affect demography, trait data should be monitored and included alongside abundance data to improve forecasting reliability.

Imperfect detection alters the outcome of management strategies for protected areas

Designing protected area configurations to maximise biodiversity is a critical conservation goal. The configuration of protected areas can significantly impact the richness and identity of the species found there; one large patch supports larger populations but can facilitate competitive exclusion. Conversely, many small habitats spreads risk but may exclude predators that typically require large home ranges. Identifying how best to design protected areas is further complicated by monitoring programs failing to detect species. Here we test the consequences of different protected area configurations using multi-trophic level experimental microcosms. We demonstrate that for a given total size, many small patches generate higher species richness, are more likely to contain predators, and have fewer extinctions compared to single large patches. However, the relationship between the size, number of patches, and species richness was greatly affected by insufficient monitoring, and could lead to incorrect conservation decisions, especially for higher trophic levels.

Eco-evolutionary processes underlying early warning signals of population declines

Environmental change can impact the stability of ecological systems and cause rapid declines in populations. Abundance-based early warning signals have been shown to precede such declines, but detection prior to wild population collapses has had limited success, leading to the development of warning signals based on shifts in distribution of fitness-related traits such as body size. The dynamics of population abundances and traits in response to external environmental perturbations are controlled by a range of underlying factors such as reproductive rate, genetic variation and plasticity. However, it remains unknown how such ecological and evolutionary factors affect the stability landscape of populations and the detectability of abundance and trait-based early warning signals. Here, we apply a trait-based demographic approach and investigate both trait and population dynamics in response to gradual and increasing changes in the environment. We explore a range of ecological and evolutionary constraints under which stability of a population may be affected. We show both analytically and with simulations that strength of abundance- and trait-based warning signals are affected by ecological and evolutionary factors. Finally, we show that combining trait- and abundance-based information improves our ability to predict population declines. Our study suggests that the inclusion of trait dynamic information alongside generic warning signals should provide more accurate forecasts of the future state of biological systems.

Inferring critical thresholds of ecosystem transitions from spatial data

Ecosystems can undergo abrupt transitions between alternative stable states when the driver crosses a critical threshold. Dynamical systems theory shows that when ecosystems approach the point of loss of stability associated with these transitions, they take a long time to recover from perturbations, a phenomenon known as critical slowing down. This generic feature of dynamical systems can offer early warning signals of abrupt transitions. However, these signals are qualitative and cannot quantify the thresholds of drivers at which transition may occur. Here, we propose a method to estimate critical thresholds from spatial data. We show that two spatial metrics, spatial variance and autocorrelation of ecosystem state variable, computed along driver gradients can be used to estimate critical thresholds. First, we investigate cellular-automaton models of ecosystem dynamics that show a transition from a high-density state to a bare state. Our models show that critical thresholds can be estimated as the ecosystem state and the driver values at which spatial variance and spatial autocorrelation of the ecosystem state are maximum. Next, to demonstrate the application of the method, we choose remotely sensed vegetation data (Enhanced Vegetation Index, EVI) from regions in central Africa and northeast Australia that exhibit alternative states in woody cover. We draw transects (8 × 90 km) that span alternative stable states along rainfall gradients. Our analyses of spatial variance and autocorrelation of EVI along transects yield estimates of critical thresholds. These estimates match reasonably well with those obtained by an independent method that uses large-scale (250 × 200 km) spatial data sets. Given the generality of the principles that underlie our method, our method can be applied to a variety of ecosystems that exhibit alternative stable states.

Early warning signals of recovery in complex systems

Early warning signals (EWSs) offer the hope that patterns observed in data can predict the future states of ecological systems. While a large body of research identifies such signals prior to the collapse of populations, the prediction that such signals should also be present before a system’s recovery has thus far been overlooked. We assess whether EWSs are present prior to the recovery of overexploited marine systems using a trait-based ecological model and analysis of real-world fisheries data. We show that both abundance and trait-based signals are independently detectable prior to the recovery of stocks, but that combining these two signals provides the best predictions of recovery. This work suggests that the efficacy of conservation interventions aimed at restoring systems which have collapsed may be predicted prior to the recovery of the system, with direct relevance for conservation planning and policy.

While several studies have documented early warning signals of population collapse, the use of such signals as indicators of population recovery has not been investigated. Here the authors use models and empirical fisheries data to show that there are statistical indicators preceding recovery of cod populations.

Global assessment of early warning signs that temperature could undergo regime shifts

Climate change metrics have been used to quantify the exposure of geographic areas to different facets of change and relate these facets to different threats and opportunities for biodiversity at a global scale. In parallel, a suite of indicators have been developed to detect approaching transitions between alternative stable states in ecological systems at a local scale. Here, we explore whether particular geographic areas over the world display evidence for upcoming critical transitions in the temperature regime using five Early Warning Indicators (EWIs) commonly used in the literature. Although all EWIs revealed strong spatial variations regarding the likelihood of approaching transitions we found differences regarding the strength and the distribution of trends across the world, suggesting either that different mechanisms might be at play or that EWIs differ in their ability to detect approaching transitions. Nonetheless, a composite EWI, constructed from individual EWIs, showed congruent trends in several areas and highlighted variations across latitudes, between marine and terrestrial systems and among ecoregions within systems. Although the underlying mechanisms are unclear, our results suggest that some areas over the world might change toward an alternative temperature regime in the future with potential implications for the organisms inhabiting these areas.

Indicators of transitions in biological systems

In the face of global biodiversity declines, predicting the fate of biological systems is a key goal in ecology. One popular approach is the search for early warning signals (EWSs) based on alternative stable states theory. In this review, we cover the theory behind nonlinearity in dynamic systems and techniques to detect the loss of resilience that can indicate state transitions. We describe the research done on generic abundance-based signals of instability that are derived from the phenomenon of critical slowing down, which represent the genesis of EWSs research. We highlight some of the issues facing the detection of such signals in biological systems - which are inherently complex and show low signal-to-noise ratios. We then document research on alternative signals of instability, including measuring shifts in spatial autocorrelation and trait dynamics, and discuss potential future directions for EWSs research based on detailed demographic and phenotypic data. We set EWSs research in the greater field of predictive ecology and weigh up the costs and benefits of simplicity vs. complexity in predictive models, and how the available data should steer the development of future methods. Finally, we identify some key unanswered questions that, if solved, could improve the applicability of these methods.

No evidence of critical slowing down in two endangered Hawaiian honeycreepers

There is debate about the current population trends and predicted short-term fates of the endangered forest birds, Hawai`i Creeper (Loxops mana) and Hawai`i `Ākepa (L. coccineus). Using long-term population size estimates, some studies report forest bird populations as stable or increasing, while other studies report signs of population decline or impending extinction associated with introduced Japanese White-eye (Zosterops japonicus) increase. Reliable predictors of impending population collapse, well before the collapse begins, have been reported in simulations and microcosm experiments. In these studies, statistical indicators of critical slowing down, a phenomenon characterized by longer recovery rates after population size perturbation, are reported to be early warning signals of an impending regime shift observable prior to the tipping point. While the conservation applications of these metrics are commonly discussed, early warning signal detection methods are rarely applied to population size data from natural populations, so their efficacy and utility in species management remain unclear. We evaluated two time series of state-space abundance estimates (1987-2012) from Hakalau Forest National Wildlife Refuge, Hawai`i to test for evidence of early warning signals of impending population collapse for the Hawai`i Creeper and Hawai`i `Ākepa. We looked for signals throughout the time series, and prior to 2000, when white-eye abundance began increasing. We found no evidence for either species of increasing variance, autocorrelation, or skewness, which are commonly reported early warning signals. We calculated linear rather than ordinary skewness because the latter is biased, particularly for small sample sizes. Furthermore, we identified break-points in trends over time for both endangered species, indicating shifts in slopes away from strongly increasing trends, but they were only weakly supported by Bayesian change-point analyses (i.e., no step-wise changes in abundance). The break-point and change-point test results, in addition to the early warning signal analyses, support that the two populations do not appear to show signs of critical slowing down or decline.

Quantifying the expected value of uncertain management choices for over-abundant Greylag Geese

In many parts of the world, conservation successes or global anthropogenic changes have led to increasing native species populations that then compete with human resource use. In the Orkney Islands, Scotland, a 60-fold increase in Greylag Goose Anser anser numbers over 24 years has led to agricultural damages and culling attempts that have failed to prevent population increase. To address uncertainty about why populations have increased, we combined empirical modelling of possible drivers of Greylag Goose population change with expert-elicited benefits of alternative management actions to identify whether to learn versus act immediately to reduce damages by geese. We built linear mixed-effects models relating annual goose densities on farms to land-use and environmental covariates and estimated AICc model weights to indicate relative support for six hypotheses of change. We elicited from experts the expected likelihood that one of six actions would achieve an objective of halting goose population growth, given each hypothesis for population change. Model weights and expected effects of actions were combined in Value of Information analysis (VoI) to quantify the utility of resolving uncertainty in each hypothesis through adaptive management and monitoring. The action with the highest expected value under existing uncertainty was to increase the extent of low quality habitats, whereas assuming equal hypothesis weights changed the best action to culling. VoI analysis showed that the value of learning to resolve uncertainty in any individual hypothesis for goose population change was low, due to high support for a single hypothesis of change. Our study demonstrates a two-step framework that learns about the most likely drivers of change for an over-abundant species, and uses this knowledge to weight the utility of alternative management actions. Our approach helps inform which strategies might best be implemented to resolve uncertainty when there are competing hypotheses for change and competing management choices.

Body size shifts and early warning signals precede the historic collapse of whale stocks

Predicting population declines is a key challenge in the face of global environmental change. Abundance-based early warning signals have been shown to precede population collapses; however, such signals are sensitive to the low reliability of abundance estimates. Here, using historical data on whales harvested during the 20th century, we demonstrate that early warning signals can be present not only in the abundance data, but also in the more reliable body size data of wild populations. We show that during the period of commercial whaling, the mean body size of caught whales declined dramatically (by up to 4 m over a 70-year period), leading to early warning signals being detectable up to 40 years before the global collapse of whale stocks. Combining abundance and body size data can reduce the length of the time series required to predict collapse, and decrease the chances of false positive early warning signals.

Elevated nonlinearity as an indicator of shifts in the dynamics of populations under stress

Populations occasionally experience abrupt changes, such as local extinctions, strong declines in abundance or transitions from stable dynamics to strongly irregular fluctuations. Although most of these changes have important ecological and at times economic implications, they remain notoriously difficult to detect in advance. Here, we study changes in the stability of populations under stress across a variety of transitions. Using a Ricker-type model, we simulate shifts from stable point equilibrium dynamics to cyclic and irregular boom-bust oscillations as well as abrupt shifts between alternative attractors. Our aim is to infer the loss of population stability before such shifts based on changes in nonlinearity of population dynamics. We measure nonlinearity by comparing forecast performance between linear and nonlinear models fitted on reconstructed attractors directly from observed time series. We compare nonlinearity to other suggested leading indicators of instability (variance and autocorrelation). We find that nonlinearity and variance increase in a similar way prior to the shifts. By contrast, autocorrelation is strongly affected by oscillations. Finally, we test these theoretical patterns in datasets of fisheries populations. Our results suggest that elevated nonlinearity could be used as an additional indicator to infer changes in the dynamics of populations under stress.

Evaluating early-warning indicators of critical transitions in natural aquatic ecosystems

Ecosystems can show sudden and persistent changes in state despite only incremental changes in drivers. Such critical transitions are difficult to predict, because the state of the system often shows little change before the transition. Early-warning indicators (EWIs) are hypothesized to signal the loss of system resilience and have been shown to precede critical transitions in theoretical models, paleo-climate time series, and in laboratory as well as whole lake experiments. The generalizability of EWIs for detecting critical transitions in empirical time series of natural aquatic ecosystems remains largely untested, however. Here we assessed four commonly used EWIs on long-term datasets of five freshwater ecosystems that have experienced sudden, persistent transitions and for which the relevant ecological mechanisms and drivers are well understood. These case studies were categorized by three mechanisms that can generate critical transitions between alternative states: competition, trophic cascade, and intraguild predation. Although EWIs could be detected in most of the case studies, agreement among the four indicators was low. In some cases, EWIs were detected considerably ahead of the transition. Nonetheless, our results show that at present, EWIs do not provide reliable and consistent signals of impending critical transitions despite using some of the best routinely monitored freshwater ecosystems. Our analysis strongly suggests that a priori knowledge of the underlying mechanisms driving ecosystem transitions is necessary to identify relevant state variables for successfully monitoring EWIs.

Rate of forcing and the forecastability of critical transitions

Critical transitions are qualitative changes of state that occur when a stochastic dynamical system is forced through a critical point. Many critical transitions are preceded by characteristic fluctuations that may serve as model‐independent early warning signals, implying that these events may be predictable in applications ranging from physics to biology. In nonbiological systems, the strength of such early warning signals has been shown partly to be determined by the speed at which the transition occurs. It is currently unknown whether biological systems, which are inherently high dimensional and typically display low signal‐to‐noise ratios, also exhibit this property, which would have important implications for how ecosystems are managed, particularly where the forces exerted on a system are anthropogenic. We examine whether the rate of forcing can alter the strength of early warning signals in (1) a model exhibiting a fold bifurcation where a state shift is driven by the harvesting of individuals, and (2) a model exhibiting a transcritical bifurcation where a state shift is driven by increased grazing pressure. These models predict that the rate of forcing can alter the detectability of early warning signals regardless of the underlying bifurcation the system exhibits, but that this result may be more pronounced in fold bifurcations. These findings have important implications for the management of biological populations, particularly harvested systems such as fisheries, and suggest that knowing the class of bifurcations a system will manifest may help discriminate between true‐positive and false‐positive signals.

Including trait-based early warning signals helps predict population collapse

Foreseeing population collapse is an on-going target in ecology, and this has led to the development of early warning signals based on expected changes in leading indicators before a bifurcation. Such signals have been sought for in abundance time-series data on a population of interest, with varying degrees of success. Here we move beyond these established methods by including parallel time-series data of abundance and fitness-related trait dynamics. Using data from a microcosm experiment, we show that including information on the dynamics of phenotypic traits such as body size into composite early warning indices can produce more accurate inferences of whether a population is approaching a critical transition than using abundance time-series alone. By including fitness-related trait information alongside traditional abundance-based early warning signals in a single metric of risk, our generalizable approach provides a powerful new way to assess what populations may be on the verge of collapse.

Predicting population collapse by monitoring key early warning signals in time-series data may highlight when interventions are needed. Here, the authors show that including information on phenotypic traits like body size can more accurately predict critical transitions than abundance data alone.