Research work · 2016
Quantifying resilience
David G. Angeler, Craig R. Allen
Journal of Applied Ecology
Abstract
The biosphere is under unprecedented pressure, reflected in rapid changes in our global ecological, social, technological and economic systems. In many cases, ecological and social systems can adapt to these changes over time, but when a critical threshold is surpassed, a system under stress can undergo catastrophic change and reorganize into a different state. The concept of resilience, introduced more than 40 years ago in the ecological sciences, captures the behaviour of systems that can occur in alternative states. The original definition of resilience forwarded by Holling (1973) is still the most useful. It defines resilience as the amount of disturbance that a system can withstand before it shifts into an alternative stable state. The idea of alternative stable states has clear and profound implications for ecological management. Coral reefs, for example, are high-diversity systems that provide key ecosystem services such as fisheries and coastal protection. Human impacts are causing significant, ongoing reef degradation, and many reefs have shifted from coral- to algal-dominated states in response to anthropogenic pressures such as elevated water temperatures and overfishing. Understanding and differentiating between the factors that help maintain reefs in coral-dominated states vs. those that facilitate a shift to an undesired algal-dominated state is a critical step towards sound management and conservation of these, and other, important social–ecological systems. Resilience has gained popularity among both academicians and laypeople, as a term meant to describe a systems' ability to withstand disturbance. Resilience has become a buzzword in the last decade, as shown by its increasing appearance in calls for research proposals and scientific citation data bases. The term resilience has in many cases lost the clarity of the original definition and in fact is frequently used in a manner in direct opposition to the original definition. Many current uses of the concept are loose and incorrect. The term is becoming increasingly used in a normative sense (Brand & Jax 2007), as if resilience were a desirable quality of systems. However, even systems in highly undesirable states, such as macro-algae dominated reefs, or city cores in poverty traps, may be highly resilient, which is to say they withstand attempts to transform them into different (desirable) states. Operationalizing the concept of resilience for application and management has been difficult. Misuse of the term can have significant negative impacts, because resilience is being used to help guide responses to natural disasters and to assess the sustainability of ecosystems and urban systems and has been driving international research priorities. Resilience has been argued to be a basic emergent property of systems, a process or a rate. We focus on the original concept as described by Holling, which is that of an emergent system property; when a system is in a desirable state and managers wish to enhance resilience, or when the system is in an undesirable state and managers wish to erode resilience and foster a transformation to an alternative state. Fostering or eroding resilience is a process. When a system is perturbed but resilience is not exceeded, then the recovery can be measured as a rate. Several frameworks to operationalize resilience have been proposed. A decade ago, a special feature focused on quantifying resilience was published in the journal Ecosystems (Carpenter, Westley & Turner 2005). The approach there was towards identifying surrogates of resilience, but few of the papers proposed quantifiable metrics. Consequently, many ecological resilience frameworks remain vague and difficult to quantify, a problem that this special feature aims to address. However, considerable progress has been made during the last decade (e.g. Pope, Allen & Angeler 2014). Although some argue that resilience is best kept as an unquantifiable, vague concept (Quinlan et al. 2016), to be useful for managers, there must be concrete guidance regarding how and what to manage and how to measure success (Garmestani, Allen & Benson 2013; Spears et al. 2015). Ideas such as 'resilience thinking' have utility in helping stakeholders conceptualize their systems, but provide little guidance on how to make resilience useful for ecosystem management, other than suggesting an ambiguous, Goldilocks approach of being just right (e.g. diverse, but not too diverse; connected, but not too connected). Here, we clarify some prominent resilience terms and concepts, introduce and synthesize the papers in this special feature on quantifying resilience and identify core unanswered questions related to resilience. To progress towards quantification of resilience, it is necessary to have a common vocabulary. Prior to introducing the broad approaches that have been developed to assess and quantify resilience and discussing the contributions of this special feature, we provide common definitions of terms commonly used in resilience science and theory. Ecological resilience is a measure of the amount of change needed to change an ecosystem from one set of processes and structures to a different set of processes and structures (Holling 1973). Resilience is an emergent property of ecosystems and other complex systems and recognizes that systems operate in multiple basins of attraction. From a human perspective, it is crucial because it implies a predictable, although variable, delivery of expected ecosystem services. Understanding the dynamics of resilience is critical to achieving sustainable human interactions with their supporting ecosystems. An ecosystem with high ecological resilience requires a substantial amount of energy to transition to an alternative state, whereas a low resilience system would transition with a relatively small amount of energy. Ecological resilience as a systemic phenomenon can be measured through an assessment of mutually non-exclusive attributes, including scales, alternative states, feedbacks and thresholds. Cross-scale resilience posits that resilience in ecological systems is enhanced when functional diversity and traits are diverse within scales and reinforced across scales. This model explicitly considers the compartmentalization of ecological patterns and processes by spatial and temporal scales. The cross-scale resilience model was established for operationalizing and quantifying ecological resilience (Peterson, Allen & Holling 1998), and especially for understanding the relationship between biodiversity, scale and resilience. Cross-scale resilience can be assessed by evaluating the distribution of functional traits of species within and across scales. Several approaches, including discontinuity analysis and time-series or spatial modelling tools, are available for assessing cross-scale resilience (Angeler et al. 2014). Engineering resilience focuses on the return of structural and functional attributes of systems to predisturbance conditions following a disturbance. Rapid return times are interpreted as reflecting high engineering resilience (Pimm 1991). The unit of measurement is time of recovery. This definition assumes that systems are characterized by a single equilibrium and therefore fails to account for the potential for alternative states of the same system. In a management context, engineering resilience incorrectly assumes that an ecosystem always recovers from a degraded state to a previous or more desirable state, and therefore, the only measure of interest is return time. This term is analogous to engineering resilience or 'bounce back' and refers explicitly to the capacity of a system to return to its initial state following disturbance. Although this definition does not explicitly exclude the existence of thresholds (e.g. the physical property of a material to return to its original shape or position without exceeding its elastic limit), multiple equilibria are not considered in this definition. More clear terms for this behaviour include return time, recovery and engineering resilience. Based on current empirical and theoretical knowledge of spatial resilience, a tractable 'shorthand' definition of spatial resilience is: the contribution of spatial attributes to the feedbacks that generate resilience in ecosystems and other complex systems and vice versa. This definition allows for the operationalization of spatial resilience in management and is consistent with the foundational aspects of spatial resilience described by Nyström & Folke (2001). It builds upon the spatially relevant aspects of complexity, including asymmetries, networks, connectivity, dispersal and information processing and how these facets change over time. A general and generic property of systems, the broad ability of a system to cope with disturbances without changing state. It does not define the part of the system that might cross a threshold and the kinds of shocks the system needs to deal with, and it copes with uncertainty in all ways (Folke et al. 2010). Managing for general resilience is complex given that multiple patterns and processes operating at distinct spatiotemporal scales need to be accounted for. However, this complexity may create opportunities for re-evaluating present and learning from past situations, boosting novelty and innovation and triggering social and policy change. That is, understanding general resilience may create new possibilities for adaptive or transformative change to swiftly changing social–ecological baselines. Resilience of a system or component of a system to a known, or anticipated disturbance. The resilience of what, to what (Carpenter et al. 2001). A management focus on specified resilience can become problematic because increasing resilience of particular parts of a system to specific disturbances may cause the system to lose resilience in other ways. For instance, international travel in Europe became increasingly focused on developing air travel, while international ground and water transportation was deemphasized. The volcano eruption on Iceland in 2010 uncovered the low resilience of this transportation system to the extensive cloud of ash in the air that interfered with the operation of aircrafts. This term has been used within ecological stability research to characterize the property of communities or populations to remain unchanged when subject to disturbance. The inverse of resistance is sensitivity (Grimm & Wissel 1997). As is the case with most ecological stability definitions, relevant patterns and processes are considered from a single equilibrium point of view and do not account for the existence of thresholds. Walker et al. (2004) broadened the use of this concept to consider resistance to be a component of ecological resilience. Stability is a complex and multifaceted concept, including components such as variability, resistance, resilience, persistence and robustness (Donohue et al. 2013). Despite this complexity, concepts related to stability are reductionist in a sense that they are less integrative of the dynamic and complex system components from which resilience emanates. Simply, stability refers to the ability of a system to remain unchanged in the face of perturbation, and to return to the initial state quickly when perturbation alters system parameters. That is, stability concepts generally ignore that ecological patterns and processes are compartmentalized by distinct spatial and temporal scales or operate in different attractor domains. From a management perspective, stability is often desirable while resilience, on the other hand, considers variability to be a desirable system property, and one that through adaption and evolution infers greater ecological resilience. Functional redundancy refers to the existence of more than one species or process delivering the same ecological function, or trait. Functional redundancy is often studied without explicitly accounting for scaling relationships in ecological systems (Truchy et al. 2015). In terms of cross-scale resilience, redundancy is considered in two forms: (i) Redundancy existing within a scale, whereby identical functions would be redundant in the strict sense. (ii) Redundancy that occurs in functionally similar species that exploit their environment at different scales. In this case, the more appropriate term is cross-scale reinforcement because it accounts for the fact that species and processes operating at other scales can maintain a function impacted by a disturbance at a single scale. Rather than focusing on the redundancy of a specific functional trait across scales, this concept emphasizes the variation in responses to environmental change by species within a functional group within scales (Elmqvist et al. 2003). Response diversity considers the functional make-up of a species accounting for multiple traits that modulate species responses to disturbances through, for instance, distinct colonization, growth, competition and dispersal abilities. Variability in combinations of multiple traits increases the adaptive capacity to cope with and respond to disturbances, maintaining functional patterns and processes in ecosystems. The behavioural or evolutionary response of species or systems to a particular stress. In ecology, adaptation has a single-species focus. Adaptation is perhaps best understood in a Darwinian sense, whereby species constantly evolve to adjust to changing environmental conditions, developing a set of traits that confer the ability to reproduce, grow and survive in habitats. With global warming, there is an increasing mismatch between such traits (e.g., thermal tolerances of cold-stenothermic species) and changing temperature regimes, leading to maladaptation and potential extinction. A measure of the capability of a system to adapt and change behaviourally or evolutionarily to unspecified future stresses. The concept extends the single-species adaptation focus, emphasizing a system-level property that focuses on learning and the constant adjustment of ecosystem properties, including community composition, function and processes to anticipate and respond to natural and anthropogenic disturbances, maintaining a system within a specific regime. It is frequently confused with adaptation or adaptive response to specific stresses, for example climate change. Adaptive capacity has been used qualitatively in the social and ecological sciences and thus varies between different contexts and systems (Adger et al. 2007). Its current use makes it either indistinguishable from, or a subset of, resilience, highlighting the need for operationalization and quantitative approaches. Thresholds indicate that ecosystems can undergo a shift between alternative states when critical disturbance levels are surpassed (Suding & Hobbs 2009). Thresholds are equivalent to tipping points and may be detected as discontinuities or bifurcation points in complex systems. Thresholds emphasize that an ecosystem and other complex systems undergo a fundamental reorganization in structure, functions, processes and feedbacks, and need to be distinguished from statistically detectable curvilinear patterns within ecosystem states. A potential alternative configuration in terms of abundance and composition, function and process, of a system. Alternative states are explicit in ecological resilience. A stable state is defined by stable structures, functions, processes and feedbacks, separated by thresholds (Lewontin 1969). Patterns, processes and feedbacks differ between alternative states. These attributes characterize the basin of attraction. The terms state and regime are often used interchangeably. However, regime specifically refers to the processes and feedbacks that confer dynamic structure to a given state of a system. A change in the process regime of a system necessarily results in a state change. In ecological systems feedbacks arise from the set of interactions between processes, abiotic structure and species. Feedbacks control an effect by influencing and being influenced by the process which gave rise to it. A positive feedback enhances or amplifies these processes, while negative feedbacks have the opposite effects. Positive and negative feedbacks generally do not imply any judgement of value regarding the desirability of the effects or outcomes. Positive feedbacks are of most interest in resilience theory; these help maintain ecological structure, function and processes in specific alternative states or regimes. A shift in regime is a persistent change in the structure, function and mutually reinforced processes or feedbacks of an ecosystem (Scheffer et al. 2001). The change of regimes, or the shift, usually occurs when a smooth change in an internal process (feedback) or a single disturbance (external shocks) triggers a completely different system behaviour. Shifts can be abrupt (years, decades) or transitory (centuries, millennia), depending on the time-scale of observation (Spanbauer et al. 2014), and affect ecosystems locally, regionally or globally (Hughes et al. 2013). Regime shifts have gained importance in ecology because they can substantially affect the flow of ecosystem service provisioning to human societies. Refers to situations where 'the path out is not the same as the path in', or to the difficulty to reverse a regime change. Hysteresis operates when the observed equilibrium of a system cannot be predicted solely based on environmental variables, but also requires knowledge of the system's past history. For instance, when a system has undergone a regime change, for example a lake, if subjected to excessive nutrient loading, will unlikely return to a previous state through mitigation of nutrients alone. Feedbacks in the alternative state are stabilizing the system, and to break hysteresis, a series of interventions (nutrient elimination, food web manipulations) are required. Discontinuity theory recognizes the hierarchical organization of which is in patterns of structure, function and Discontinuity theory and its application to identify spatiotemporal scaling patterns in ecosystems have been distribution of (Angeler et al. of similar in the environment and therefore operate at a specific scale. of based on their use at different scales, and the patterns of species can therefore be used to distinct scales in ecosystems et al. 2014). This also to other complex systems such as urban systems and et al. 2014). has different but in ecology refers to the spatial and temporal of a specific set of processes or These processes differ across different scales but are in and time processes operate at small spatial scales, while processes occur over broad spatial That is, ecological scaling has a that it both the spatial and temporal of processes and structures & 2013). Ecological are often out in a temporal or spatial without both necessarily In these the of or needs to be made of management as to enhance learning and uncertainty et al. Adaptive management, often as a and management It is appropriate where both uncertainty and in the system are that is, where there are unanswered questions regarding system dynamics and response to but where is also Adaptive management was developed in with resilience as a to systems without causing a state change. Adaptive is an research for the social, economic and ecological of that are necessary for resilience to cope with the by global change and its effects on complex adaptive systems of and Adaptive adaptive management. The of adaptive management, adaptive ecosystem management and of and management approaches in the of adaptive et al. 2015). to quantifying resilience have developed and are diverse, but can be Rapid assessment approaches have been developed that focus on and knowledge of the systems they et al. 2013). This of approach is more than but which can be used to assess resilience among similar systems, and assess among social, economic and ecological components of complex systems. approaches include the Resilience which stakeholders in an although relatively of the of resilience in the systems in This approach increases of resilience and its but and complexity such as the structure in complex systems. approaches have developed more and focus on relationships among spatial attributes of systems. of functional diversity and response diversity have been and have to for assessing resilience (Angeler et al. 2014). discontinuity approaches, often with functional specifically assess and cross-scale structure et al. 2014). Resilience assessment also focus on the of thresholds et al. 2014), including that at identifying regime shifts et al. The papers in this special feature and on current quantitative approaches in resilience the ecological but also the social sciences and social–ecological systems. The special feature is to aspects of quantifying ecological resilience in a of systems and assessing leading of regime between ecological complexity management and The of system is including reefs, coastal and a social–ecological system with The feature with papers that focus on assessing the resilience of particular contexts and systems, and Allen et al. resilience in a spatial context, an definition of spatial resilience and an of how attributes of spatial resilience, including internal and system can be An important from this is that spatial resilience of dynamic cannot be understood without a temporal to This view is in the spatial resilience by & The use analysis of data to assess the social–ecological resilience of the that a following an of This an example of engineering resilience in a social–ecological system, whereby the of learning and adaptation in the a system state that might be to future et al. the of discontinuity theory to management The studied how the functional attributes of such as behaviour and reef recovery a approach uses as a for the spatial scale at which from the reefs where the to the were made of a of and thus that over a of spatial scales, the reefs were more to to coral-dominated states the disturbance. This that cross-scale redundancy managers with a of resilience in reefs, and in other ecosystems. The use of discontinuity approaches has little application in that or which has the potential of the approach for some and systems. to this et al. published in of how resilience can be in systems. the use of the of variability concept to characterize and the behaviour of and in and the basin of of a system given its in the of ecological resilience. The also on engineering resilience, how the and of recovery in to of variability of ecosystem can as an to measure the resilience of ecosystems and for regime The of regime shifts et al. and their application to management has become a focus of research et al. et al. how current can be with a based on that the dynamic of ecological and other complex systems. has the potential to abrupt changes and between ecological (Spanbauer et al. 2014), helping to guide management based on adaptive or transformative approaches. et al. the potential of and to indicate regime shifts across multiple levels in and habitats. that these to regime changes across all levels in the food web in and changes that not occur were often The the need to a of and time series into an model to with empirical and modelling approaches that process et al. the and in both assessing and resilience, in a social–ecological and social systems' The the difficulty to assess and quantify resilience as an emergent complex systems emphasize the need to approaches in resilience assessment and measurement and to on and key that have been for resilience. Many of these approaches have in ecology but have not the social and social–ecological system For instance, Angeler et al. how discontinuity theory can current management approaches, a for ecological management and how discontinuity approaches can help managers measure and thus the resilience of ecosystems. also how discontinuity theory can be used in other of ecological management, including the of critical scales of and time in ecological systems, the of regime change, and understanding and and Resilience theory has developed across multiple scientific However, the of has our ability to and between the ecological and social sciences and policy to our understanding and management of social–ecological systems. resilience has to have two ecological which are often but with important A between social and ecological resilience research is the explicit application of quantitative for resilience. have been made in years in ecology, with many approaches, including discontinuity and other modelling and spatial for quantifying attributes of resilience. of the social resilience research and the of quantitative approaches is because of the of to make complex system dynamics tractable et al. existing knowledge is the for scientific and the made in ecology help questions in social resilience Despite in the of resilience as quantitative approaches to quantify resilience are often and to the scale of ecosystems. These approaches often focus on specific which assessing specific resilience that might not be of other or the ecosystem at multiple resilience measurement approaches across including the assessment of functional traits and processes, might in a understanding of the general resilience of ecosystems. approaches in ecosystems will also need to be with to into relationships in resilience. resilience research based on approaches have from the analysis of However, assessment of the resilience of is by the of data that the relevant scales of and time. This the need for significant for which are generally In to these theoretical remain that quantitative resilience For instance, there is need to operationalize aspects of resilience. Adaptive capacity is increasing as a to cope with natural and anthropogenic disturbances, but the concept defined in multiple in distinct contexts and thus This concept is in need of operationalization to make it for quantification in approaches to between the of species that are by vs. processes have shown results to assess facets of adaptive capacity et al. 2014). application of these approaches is to assess how the structural and functional attributes of these species builds adaptive capacity in ecosystems. the of and structures a ground for basic and understanding that can affect how we manage for resilience of complex systems into the future & Allen 2014). environmental and is often too and therefore accounts for the of ecosystems. Resilience theory and measurement have potential to environmental and into current policy will for a more management of ecosystems and their services. are and to quantitative resilience research and to research that different ecological ecological complexity, and environmental economic and social sciences, and of the non-exclusive components of resilience and of them into and assessment for resilience is assessment that account for complexity current approaches and may management and policy that and the in ecosystems and other complex systems under changing ecological baselines. This special feature was by the of the of research from the and and the of feedback on this The and is by a among the the and the of the and and the use of or is for only and does not imply by the have not been because this does not