Refocusing on nature: Holistic assessment of ecosystem services



The benefits people obtain from ecosystems vary from direct benefits that are easily monetized (e.g., timber) to indirect benefits that are not easily monetized (e.g., maintenance of water quality). Commonly, there is wide variation among individuals in the values placed on ecosystem benefits or services. The lack of consensus both in identifying ecosystem services and in valuing them with respect to other services poses a great challenge to those charged with evaluating changes in the provision of ecosystem service after, for example, a natural disaster. Natural resource economics provides some tools, but economics alone will not ensure a balanced, holistic assessment. An inherent complexity in valuing services is often associated with the interrelationships between services and the background and expertise of those leading the assessment. We argue that a holistic evaluation of ecosystems founded on solid expertise in ecosystem dynamics is essential for the accurate assessment of ecosystem services. A reductionist approach to ecosystem service valuation often fails to capture ecological dynamics that are vital to the functioning and ultimate provision of services. In this article, we present case studies of ecosystem services valuation for forest fires, dam removal, and chemical contamination of sediment to explore the complexity of ecosystem service valuation. Additionally, we offer assessment strategies for recognizing the importance of holistic assessment of ecosystem services. Integr Environ Assess Manag 2012; 8: 401–411. © 2012 SETAC


People vary in their views of nature and in their understanding of ecosystems and the ecological services they provide. Evidence suggests that individuals will champion services that are in their near-term self interest to the detriment of services afforded to others or to future generations. This challenge has been historically expressed as the “tragedy of the commons” in which those who rely on ecological services act independently to extract those services, ultimately depleting the resource for others or for future generations (Hardin 1968). Such conflicts pervade the management of natural systems such as old-growth forests, and these conflicts highlight the need to understand the interaction and interdependence of multiple species-specific ecological services (Spies et al. 2006). Land, water, fire, and chemical management are rich with examples where efforts to maximize certain services of one kind or in one geographical area can alter other services either in the same area, other locations, or in the future. Management actions planned and taken to deal with the 2011 Mississippi River flood (USA) offer a recent example of the types of conflicts and tradeoffs that occur within natural resource management. Wetlands and farmland served as a relief valve for floodwaters in the interest of safeguarding the public in more populated locations. This article provides 3 examples of these external effects of management choices on natural resources.

Ecosystem services are the benefits people obtain from ecosystems (Millennium Ecosystem Assessment 2005) and the nature and flows of these benefits are frequently discussed in terms of currencies and service markets (Kroeger and Casey 2007; Achterman and Mauger 2010). Accounting for these services can be complex and challenging (NRC 2005). Although the National Research Council (NRC) acknowledges that values of ecosystems and their services include nonanthropocentric components, the NRC focuses on anthropocentric components and, in particular, on economic valuation expressed in terms of traditional currencies. We also note that there is increasing momentum among ecosystem services professionals to develop nonmonetary tools and metrics for quantifying ecosystem service tradeoffs in environmental decision-making (BSR 2011). We argue that developing a basic understanding of the interconnections and complexity of ecosystems and the compilation and tradeoffs of services in time and space are probably the most challenging aspects when estimating either monetary or nonmonetary value.

Because monetary valuations of services are often based on stated preference metrics, such as “willingness to pay,” valuation of ecosystems is limited by people's understanding of how ecosystem products or functions provide direct or indirect benefits to humans, and this understanding is often poor or misguided. Although a wide variety of methods are available for valuation including revealed preference approaches and combinations of stated and revealed preference approaches, the methods share the common challenge of properly understanding ecosystems, the varied tradeoffs in ecosystem services and, ultimately, the valuation of the services. This challenge applies equally to nonmonetary quantification of services, because these approaches also rely on eliciting preferences from stakeholders regarding ecological services.

Because of the complexity of ecosystems and of the methods used to value their services, it is easy (either intentionally or unintentionally) for biases or lack of knowledge to affect the valuation process. This challenge is clear when we consider that the valuation and forecasting of ecosystem services is likely much more complex than more traditional economic projections and analyses. The recent financial crisis shows the implications of ignorance among the populace, politicians, and even experts concerning a currency system we presumably understand. The crisis also revealed how some individuals and entities were able to game the system to extract profits at the expense of others. Understanding natural resource economics is more challenging, and therefore, the potential pitfalls are more troubling. How can we guard against the lack of understanding of ecological concepts and processes and consequent bias for valuations of ecosystems and their services? We argue that there is a need for a more holistic and coordinated approach that must be informed by a scientific as well as sociological and economic understanding of nature. In short, we argue that nature should be put back into the assessment of natural resources and that quantification not simply be left in the hands of individuals familiar with assessment tools but not ecosystem dynamics. This requires an understanding of interactions and tradeoffs.


Ecosystems are recognized to be rich in interactions in place and time. This is reflected in the earliest and simplest definitions of an ecosystem: a system composed of biotic communities and their abiotic environment interacting with each other (Odum 1953). Over the last several decades, ecologists have continued to explore the nature and evolution of ecosystems. Over time, researchers have shifted from a focus on classical ecosystem equilibrium to the recognition of the spatial and temporal complexity of ecosystems through, for example, the hierarchical patch dynamics paradigm (HPDP) (Wu and Loucks 1995; Wu and David 2002). The HPDP recognizes the importance of linkages among temporal and spatial scales, the influence of ecosystem dynamics on different patches, and the resulting dynamic equilibrium (or “nonequilibrium”) at certain scales and quasi-equilibrium at other scales (Wu and Loucks 1995). Recent writings continue to note the importance of ecosystem linkages. Raffaelli and Frid (2010) point out that unlike traditional reductionist approaches, a holistic perspective toward ecosystem services provides insights into the many unexpected consequences of human activity.

A holistic approach for recognizing and identifying the provisioning of ecosystem services and interactions among those services will require not only basic biophysical knowledge of such systems but also a consideration of temporal and spatial scales, along with a framework for explicit consideration of countervailing services. Knowledge of terrestrial and aquatic ecosystems comes largely from the scientific domain and this can involve a considerable amount of knowledge as well as uncertainty. The experts' meeting after the 1988 forest fires in Yellowstone National Park (Greenlee 1996) highlighted the lack of knowledge about forest fire ecology among the scientific experts. In a summary of the workshop, Knight (1996) noted that “the papers presented at the conference indicate once again that ecosystems are highly variable from place to place and from one year to the next. For example, lodgepole pine was an early invader in many cases, as predicted, but not everywhere. Also, erosion was accelerated in some areas, but the amount of soil loss and subsequent sediment deposition in streams varied greatly from place to place and in most cases was within the normal variation observed before the fires.” Some plants and animals benefited from the fire, some remained the same, and in the areas of especially hot fire, the systems appeared to suffer. Overall, the ecosystem responded differently to the fire depending on a broad range of other conditions. The work dispelled various popular hypotheses that existed before 1988. As we discuss later, insights into forest fire ecology have led to alterations in the approach to forest management and fire prevention and control. Knowledge gained has also resulted in greater recognition that fire can enhance certain ecosystem services and that frequent smaller fires can offset some service losses resulting from larger uncontrolled fires.

Ecologists are becoming increasingly engaged in delineating how ecosystem functions link to ecosystem services. The explicit consideration of ecosystem services is receiving increased attention in the environmental regulatory/policy arena (e.g., the National Ecosystem Services Partnership,, within the business community (e.g., Businesses for Social Responsibility,; World Business Council for Sustainable Development), nongovernmental organizations (e.g., Resources for the Future, World Resources Institute, and The Nature Conservancy), among ecological economists, and within scientific professional communities (Ecological Society of America and Society of Environmental Toxicology and Chemistry). Because of the complexity of ecosystems, ecologists will need to play a critical role in guarding against oversimplifications and the application of incorrect or incomplete ecological paradigms, including the common misconception that all ecosystems undergo ecological succession leading toward some balanced, stable, and harmonic “end state” or “climax community.”

Leopold (1989) wrote “A thing is right when it tends to preserve the integrity, stability and beauty of the biotic community; it is wrong when it tends otherwise.” Over the past 2 decades this perspective of nature, ecosystems, and the notion of “what is right” has been much debated and has been replaced to varying degrees with a very different perspective—that is, that ecosystems are reflections of dynamic change, disturbance, and nonequilibrium conditions (Pickett et al. 1992; Sagoff 1997; Soulé 1995; Paine 2002; Calderón-Contreras 2010; Dornelas 2010; Johnson and Miyanishi 2007). This shift is apparent when one attempts to establish a baseline condition for an ecosystem that is subject to periodic but still catastrophic influences. No longer do ecologists focus on one “ideal” or “undisturbed” state, but instead the baseline is evaluated over time, capturing natural cycles and long-term patterns. We discuss the disturbance-based baseline further in the forest fire case study that follows. The shift in the ecosystem and/or nature paradigm raises challenging questions about how ecosystems and associated ecosystem services should be identified, quantified, and ultimately valued.

The changing scientific perspective leads to the need for greater emphasis on clear communication with the public about a dynamic and changing nature. The public is more familiar with snapshot views and simple classifications of ecosystems, and they often do not recognize underlying dynamic processes and change. Therefore, when the picture moves outside the frame, it may not look “right.” Simus (2008) has considered the aesthetic implications of the new paradigm in ecology for the average person. He argues that this involves learning how to appreciate natural beauty in a way that is more aligned with the underlying current scientific thinking. Therefore, he states that if the new paradigm in ecological science tells us that nature is constantly changing, then aesthetic appreciation must adapt to constant change in natural systems in contrast to the view that pristine nature has only positive aesthetic qualities, such as balance, order, and harmony. Simus (2008) makes 2 recommendations in this regard: 1) aesthetic appreciation should be directed toward natural processes, not only natural objects; and 2) aesthetic qualities, such as imbalance, disorder, and disharmony, should be considered positive aesthetic qualities. He completes his recommendations by noting that these 2 suggestions imply that beauty in nature can be as chaotic as it is orderly. It follows that preference-based approaches for assessing ecosystem values need to incorporate more holistic representations of the nature of the system being valued. This requires informing people about dynamic processes and the nature of changes over relevant timescales. This also implies that it is misleading to structure surveys of preferences around ecosystem snapshots that imply that these views are “right” from an ecosystem perspective.

The foregoing serves as an important illustration of our point that incomplete or inaccurate understanding of ecological processes and functions leads to poorly founded judgments regarding value. Because many people, including natural resource economists, may view ecosystems in terms of the old paradigm, there can be an inconsistency between valuations of ecosystem services and the dynamic nature of ecological processes. We believe that the uncertainties associated with valuations and predictions of ecosystem services are likely greatest for systems that experience regular disturbances. Examples include periodic fire threats to forest ecosystems and major storm impacts on coastal marine ecosystems.

Influences and changes that occur at varying timescales are important but can be easily missed. In the case of forests, Holmes et al. (2008) point out that some economically important forest disturbance processes, such as pest outbreaks and fires, result from the interaction of variables across short and long timescales, and that policy-relevant economic models need to recognize the impacts of long-term ecosystem dynamics on the short-term behavior of economic variables. They note that because movement in a long-term (or slowly evolving) ecosystem component (e.g., forest foliage, fuel accumulation) can induce a sudden, catastrophic eruption in a short-term stressor (e.g., area infested by pests, area burned) that are linked, in turn, to various economic variables (e.g., pest eradication costs, fire suppression costs, economic damages), simple comparative static analyses may provide uninformed predictions of changes in economic variables.

Although the work of ecologists is essential for proper recognition of ecosystem services and their attendant uncertainties, value is derived by nonecologists. The challenge, therefore, is to narrow the knowledge gaps between biological and social scientists, the public, and the policy makers and regulators. This cannot be readily accomplished through simple survey tools, because there can be a substantial lack of knowledge on the part of the survey team and on the part of the public. Based on the experience from the Yellowstone fires of 1988, it is evident that surveys of the public carried out immediately before or after the fire would have yielded erroneous information even if they were well informed by the current scientific thinking. Although there are limits to scientific knowledge of ecosystems, there is still a need to rely on the most current understanding along with explicit consideration of the uncertainties associated with that knowledge. However, there is also a need to communicate this information in a nonbiased way to the public, regulators, and policy makers so that values can be properly assessed and management decisions made. We argue that this will require effective stakeholder engagement among all key stakeholders. Based on our experience, the lessons learned from watershed management programs that have engaged all stakeholders would be an appropriate model for this shared discourse. One-sided communications with scientists simply publishing studies or with the public responding to questionnaires concerning their values will not adequately bridge the knowledge gaps that we know to be present. As Simus (2008) observed, there is likely a need for explicit consideration of the dynamic and chaotic nature of the ecosystem, including inherent disturbance features. Any service quantification method that relies on simple snapshots of nature, without considering natural or human-induced variability, will fail to accurately reflect stakeholder preferences and, thus, the ultimate “true” value of an ecosystem service.


The assessment of ecosystem services needs to consider relevant spatial and temporal scales. Watershed and regional assessments are common examples of efforts to consider appropriate spatial scales. Ecologists recognize temporal scales that can range from less than a day to decades and even centuries. The study of disturbance ecology focuses on long timescales. For example, disturbance ecology recognizes that over large enough time periods, the growth and regrowth of some forests reflects various cycles of fire, drought, and insect infestations. These periodic events can remove older and weaker members of the population and allow for new growth and rejuvenation. Forest fires, for example, occur naturally at intervals that are specific to the type of forest and periods of relative dryness. Knight (1996) reports that for Yellowstone National Park, forest fires occur at intervals of 20–50 years for Douglas fir-dominated forests, and 200–300 years for lodgepole pine, Engelmann spruce, and sub alpine fir forests. The record for Yellowstone shows that fire frequency was higher for Yellowstone when the weather was drier. This certainly shows the potential interaction between warming of the earth surface and the frequency and probability of fires. In the Sierra Nevada Mountains, the interval between fires was 20 years before European settlement; however, after settlement, management decisions to suppress fire and harvest timber increased the interval between fires and contributed to the conditions that result in catastrophic fires (McKelvey et al. 1996). McKelvey et al. (1996) report that the reduction in fire activity, coupled with the selective harvest of many large pines, produced denser forests with generally smaller trees, a higher proportion of white fir and incense cedar than were present historically. They believe that these changes have increased the amount of fuel on the forest floor as well as “ladder fuels”—small trees and brush that carry the fire into the forest canopy.

The natural structure and function of ecosystems varies considerably over spatial and temporal scales. It is important to recognize natural spatial and temporal variability when evaluating ecosystem services, because the provision of services will vary in response to changes in ecosystem structure and function. Most people will recognize the change in a commonly identified service (e.g., abundance of large mammals), but other services (e.g., availability of refuge habitat for woodland birds) can be easily overlooked in a typical service assessment.


The assessment of ecosystem services needs to guard against “ecosystem service myopia,” which occurs when one group chooses to focus on one or a few services over others. This can happen either because the services are considered most valuable to the group constructing the analysis or are judged to be easier to quantify. A narrowly focused assessment of ecosystem services can miss important tradeoffs among services as well as fluctuation in the provisioning of services over space and time. This is reminiscent of the “risk versus risk” arguments put forward by Graham and Wiener (1995). The main thesis of that work is that there are risk tradeoffs in protecting health and the environment and that in addition to evaluating the target risk, it is important to recognize countervailing risks if the goal is to maximize overall health and environmental conditions. The authors identify 4 types of risk tradeoffs that we have considered in terms of ecosystem services: offsets, transfers, substitutions, and transformation. We suggest that an assessment of ecosystem services include an explicit consideration of all ecosystem service changes, whether positive or negative, easy or challenging to account. Also, the assessment of ecosystem services recognizes that there are close linkages among ecosystem services that result in at least 1 of these 4 tradeoffs.

  • Service offset occurs when a management action results in a reduction of a specific stressor on a population or ecosystem but allows for the increase in another stressor. For example, the introduction of a biological control species to reduce a pest can result in the population growth of other species that the pest had been controlling.

  • Service substitution occurs when actions designed to increase the provisioning of one service result in the enhancement of a negative service or suppression of another service. For example, a management decision to convert agricultural land to open space to enhance large mammal populations may result in an increased incidence in deer ticks and Lyme's disease (a “negative service”) or the suppression of habitat for neotropical migratory birds (as a result of grazing).

  • Service transfer occurs when a service enhanced in one population or ecosystem results in the suppression of similar or different services in another population or ecosystem. This can happen when environmental management decisions shift the geographical location of stressors. Media-specific pollutant management is an example of service transfer. Services gained in one ecosystem from the reduction of pollutants could result in a loss of services at another location. For example, use of water scrubbers to reduce air pollutants shifts the pollution burden from air to surface waters. Thus, services provided by clean air are enhanced, potentially at the expense of services provided by clean water.

  • Service transformations occur when changes in one type of service result in a different types of changes in services for another population or ecosystem. Many examples of service transformations can be found with respect to management of water flows and the use of water. Dams and levees, for example, provide certain types of recreational and safety services but can strongly influence other services elsewhere. We provide an example of this in relation to dams in terms of services provided (hydropower, recreation, and aesthetics) and services diminished (sediment replenishment for downstream areas, disruption of anadromous fish reproductive capacity).

Holistic assessments of ecosystem services and associated tradeoffs could be carried out and presented in a variety of ways. A recent example involved quantifying relative service losses and gains associated with various alternatives for restoring oysters in the Chesapeake Bay (USACE 2009). Relative service gains and losses were considered in relation to changes in populations and functions. For comparative purposes, these were displayed as bar charts indicating which populations and functions were increasing and which were decreasing with respect to time and space. The Relative Risk Model (RRM) (Landis and Wiegers 2005; Landis et al. 2004) was adapted for this purpose; a common scale was developed for the services that reflected the nature of the services (i.e., oyster hard bottom development, changes in water quality, influences on various fish and wildlife). This visual representation enabled environmental managers to examine the relative ecosystem services gains and losses over time. Thus, it provided insight into the overall net environmental benefits of each restoration alternative. The managers in this case framed their goals not in terms of present value but at a future point in time (a decade). This was an explicit recognition that the desired ecosystem services could not be attained in the near term. Instead, the analysis of ecosystem services flows needed to be applied over a time frame commensurate with the time dimension for the provisioning of the services themselves. Time horizons on the order of decades are probably appropriate for judging alternatives for major restoration activities involving contaminated sediments in rivers.

We provide a generic summary of changes in ecosystem services resulting from a hypothetical action or event. Figure 1 provides an example of a tally of ecosystem service values provided by a stream under current conditions compared to service values after a dam is constructed. It also provides an example of a comparison between 2 different spatial locations, upstream and downstream of the dam, and the resulting change in ecosystem service values. It is important to note that this illustrative representation of benefits and costs ignores many subtleties regarding tradeoffs listed above. For example, although creation of a reservoir by construction of a dam may enhance one type of recreation (boating and warmwater fisheries) this may be at the expense of other recreational services such as white-water rafting and coldwater fisheries. These tradeoffs are explored in more detail below.

Figure 1.

Conceptual example of ecosystem service values provided by a stream under different development scenarios (dam versus no dam) and spatial locations (upstream versus downstream).

To demonstrate the concepts and assessment strategies described above, we provide 3 case studies. Each focuses on a unique ecosystem disturbance—forest fires, dam construction, and sediment contamination—and explores ecosystem service changes and the challenges of designing a defensible ecosystem service assessment.


The impact of fires to wildlife can be complex because of indirect effects on habitat. Although fire is detrimental to some species, it is beneficial to others. Interestingly, many of the animals the public would place a high value on are often favored by fires. For example, many game animals, such as deer and elk, are favored because fires tend to result in a long-term increase of the amount of meadows. Nongame animals that provide recreational services, such as wildlife viewing, are also favored after a fire. Eagles, owls, and hawks hunt open areas flush with smaller animals. Insectivorous birds, such as woodpeckers, may benefit from increased food from insect outbreaks that may occur postfire and standing dead or damaged trees that provide cavity-nesting opportunities. Even grizzly bear can benefit from fires. According to Contreras and Evans (1986), grizzly bear populations decline in areas where fire suppression is practiced. Berry-producing shrubs, a food of the bears, thrive after fires as a result of the opening of the forest canopy. The corresponding increase in the ease of hunting and the availability of game also helps. Increased production of forb foliage and tuberous roots after the 1988 Yellowstone fires were judged to benefit grizzly bears (Blanchard et al. 1996).

Management of forest fires in our national parks and in the drier areas of the West provides good examples of where decisions influence transfers, substitutions, and transformations of ecosystems services. Paramount ecological, economic, and political issues for these systems have involved various fire suppression and controlled fire strategies. As we discuss, these various strategies necessarily involve many tradeoffs in ecosystem services. In particular, whereas fire suppression provides certain near-term ecosystem benefits valued by tourists and residents, these practices alter other services and can set up the inevitable catastrophic fire. Thus, practices designed to yield services in the present result in loss of these same services at a later time, in addition to influencing ecosystem services that may not have even been considered in the original decision-making process.

The history of episodic surface fires in the giant sequoia groves in Yosemite National Park has been documented by Swetnam et al. (1990) dating back more than 1400 years. The authors noted that “nearly all the largest and oldest sequoias have huge basal fire scars that bear witness to these ancient flames.” However, in about the mid-1800s, a practice of natural fire suppression prevailed as settlers moved into the surrounding areas. Swetnam et al. (1990) concluded that until that time, fires historically occurred at intervals ranging from 1 to 15 years in the park, producing an open forest structure with a grass and forb understory, scattered tree regeneration, and little accumulation of debris on the ground (Kilgore 1981; Weaver 1974). This natural condition maintained the dominance of pines and giant sequoias. However, after nearly a century of fire suppression, the undergrowth of shade tolerant vegetation was out-competing the sequoia seedlings and saplings. Instead of an understory of manzanita, deerbrush, wedgeleaf ceanothus, and bitter cherry (Wright and Bailey 1982), less fire-resistant species, such as white fir and incense cedar, chocked the forest floor. According to Harvey et al. (1980), the natural fire process created mineral soils best suited for germination and establishment of the sequoia seeds. However, the lack of fire that resulted from intentional fire suppression had altered the nutrient dynamics and successional patterns of the forest. The lack of ecosystem knowledge concerning the natural role of fires was evident in early policies but still occurs today.

As the role of fire in sustaining forests became better recognized, management policies evolved that included controlled burns at more frequent intervals. Controlled burns in Yosemite National Forest's Mariposa Grove have shown positive results for giant sequoia seedlings and saplings. Yet, managing a forest using fire generally conflicts with the public's aesthetics for a forested area. To the public, viewing charred bark, scorched foliage, and cleared underbrush may not provide the visual services or reflect values tied to notions of “a balanced ecosystem.” Managing a forest with controlled burns will still not match the mixed- and high-severity forest fire regimes that historically dominated in the West and evoke even stronger negative reaction from the public (Baker 2009). The public's resistance to the seeing the fire-impacted landscape, as well as the interest by public and timber company land managers to control and minimize fire are at direct odds with the health of natural communities that require intense fire to thrive. This represents an inconsistency between public perceptions of ecosystem services and the natural ecosystem processes of a forest. It is this inconsistency that Simus (2008) tells us should be addressed by developing a greater appreciation among the public for the chaotic processes of ecosystems. This will be a challenge and will require innovative educational approaches given that tourists and residents alike tend to value an apparently unblemished forest. However, there is now plenty of evidence that choosing to maintain a forest in the unnatural condition of fire suppression carries serious adverse and long-term consequences. Fire exclusion in a forest yields increased fuels, such as dead wood, unhealthy trees, and thick layers of leaves, which build up on the forest floor, creating conditions conducive to large fires.

In addition to the tradeoff between smaller, more frequent controlled fires (with frequent but smaller associated aesthetic impacts) and larger, catastrophic fires (with less frequent but larger aesthetic impacts), the public may also need to appreciate other fire-related ecosystem services related to forest nourishment, forest reproduction, and the influence of fires on ecosystem services afforded by wildlife species. Burning off the leaf litter increases N and other nutrients in the soil and allows seeds to come in contact with the soils, while the heat of the fire allows seeded cones to open for regeneration. Lodgepole and jack pines demonstrate these adaptations. The cones, which contain hundreds of seeds, remain closed while on the tree. In the presence of heat from a fire, the resin melts and the cones open to release their seeds onto the cleared ash bed. In forests where fires occur with reasonable frequency, tree species exhibit adaptive characteristics (e.g., the ability to heal fire scars, development of thicker bark, seed adaptations) to survive.

Public perceptions that fires result in losses of wildlife and associated ecosystem services are not consistent with what actually occurs. In reality, there is a complex interplay as different species and ecosystem processes respond to postfire conditions and to understand the implications of the response, the time horizon must extend beyond the immediate postfire conditions. According to Gleason and Gillette (2009), evidence suggests that most fires do not negatively affect wildlife populations; wildlife respond and adapt to fire in a variety of ways, and many wildlife species benefit from fires, directly and indirectly. Species such as the Black-backed woodpecker, for example, are so specialized in their habitat needs that they require severe fire to persist (Hutto 2008). Many animals, although not requiring fire per se, are capable of tolerating a broader range of habitats, and fire changes may have little or no adverse effects on their populations. Raptors, for example, can thrive immediately after a fire because the meadows make hunting easier for them. The grasses, seeds, and insects growing postfire make foods more abundant for small mammals, reptiles, invertebrates, and amphibians for a period of time. In turn, birds of prey benefit from the change. Postfire meadows provide herbivores with an abundance of plants to feed on. A long-standing forest can be too dense, creating intense competition among the trees, thus limiting their growth and the overall health of the timber. The benefits and costs to any given species after a fire will vary over time and will be driven not only by changes in the habitat postfire, but also by the resulting shifts in abundance that influence predator-prey relationships. What may be a positive benefit for an eagle immediately after a fire may be less beneficial or even detrimental after multiple years of forest regeneration postfire.

Many factors, including secondary disturbances, will influence the successional process in a forest, and these successional states have attendant ecosystem services. Thus, the whole process of fire and transition reflects a natural progression of ecosystem services consisting of a continuum of substitutions and transformations. The landscape of a burned area is a mosaic of recovery, with no recovery being the same in any given location. The intensity of the disturbance and the closeness to surviving stands are key factors in ecological recovery. In wildland fires, the intensity of the burn and the random burn patterns of the fire, which are influenced by weather patterns, landscape, and other forces, create the mosaic of surviving stands or relatively undisturbed islands. These less impacted or unburned areas provide the abiotic and biotic start for rapid successional processes. Large areas that are untouched by fire serve as a source of propagules for recolonization. Indigenous species from the regional pool will be the first groups to recolonize across an affected landscape. Under this scenario, there is value in dead biotic structures, as they provide habitat for regional species and erosion control. Snags, logs, and stumps are common after a severe fire and their decay provides a primary source of organic matter to the soils, thus improving long-term productivity. These patches of biotic and abiotic elements immediately provide complex habitats for survivors and may promote the establishment of many additional species.

Transformations of ecosystem services after a fire will depend on the landscape. The more heterogeneous a disturbed landscape is, the greater the acceleration and pattern of ecological recovery. As observed in the Yellowstone fires, the burned areas were not uniformly burned; rather, they were heterogeneous with varying patches of burned and unburned areas. Turner et al. (2003) stated that the majority of severely burned areas were within 5 to 200 meters of unburned or lightly burned areas, suggesting that few burned sites were very distant from the potential source of offsite propagules. Although we know that patches of unburned landscape provide immediate habitat and a source of propagules for recovery, not all fire recovery occurs in this manner. In the Yellowstone Park fires, Turner et al. (1997) found that most postfire colonization occurred from plant parts that survived within the burned areas and then produced seed, rather than by dispersal from unburned landscapes. Even in zones with severe heat and canopy fires, the roots and rhizomes from many plants lived. The following year, they reported that many plants quickly refilled the burned forests.

The wildfire case study points to a challenge when assessing natural and human-impacted changes in ecosystem services: the baseline itself naturally changes over time. Clearly, baselines for dynamic and changing systems are not captured in snapshots of a particular ecological state such as a “climax community.” We use the example of forests that naturally experience wildfire to examine how to consider a dynamic and changing baseline. Forests depend on a periodic fire to “reset” ecological processes and maintain a healthy, balanced tree population. Because the role of fire in forest ecology has an associated temporal component and periodicity, the baselines for this ecosystem are also temporally variable and characterized by transitions rather than any single state. Therefore, before and after comparisons of ecosystem services after a fire require more than contrasting snapshots. When the baseline is evaluated over ecologically-relevant timescales, the periodicity in service values becomes evident and can be overlaid on the service curves after a fire.

Figure 2 contrasts baselines for a natural forest ecosystem for which fire is a natural component and a managed forest ecosystem for which fire has been suppressed. Figure 2 illustrates the complexity inherent in evaluating changes in environmental services after a catastrophic event and the necessity for considering change over time. There are 2 types of service baselines for forest systems. A natural baseline service level results from a system that is not managed in any way, but is instead subject to natural processes and events. In contrast, a managed baseline service level is representative of conditions resulting from careful implementation of a management plan developed to achieve predetermined goals. For example, forests are often managed to maximize timber production. The result of the management activities is a forest that may be more or less susceptible to specific events such as fires. Regardless of externals forces acting on them, forests will display natural variability over time. For some factors there can be an obvious cyclical pattern, for example, by season. For other factors, the pattern may be less predictable. The 3 scenarios in Figure 2 represent 3 of any number of outcomes that might occur after a catastrophic event such as a fire. As illustrated in Figure 2, the change compared to baseline will be very different depending on when one assesses ecological services (see Time A versus Time B in Figure 2). Clearly, to complete a defensible and comprehensive service analysis, multiple time steps must be assessed.

Figure 2.

Example conceptualization of forest service level differences (natural versus managed) and the potential variability in postfire service levels under different recovery scenarios.

A defensible service analysis with respect to the potential impacts of a catastrophic forest fire depends on the quality and clarity of development of an appropriate baseline. One of the 2 baseline types discussed previously—managed versus natural baselines—is required to evaluate service level changes. Guidance is available for developing a baseline for specific services. For example, the Climate Action Reserve ( requires a baseline of forest carbon stocks for credit registration (CAR 2010). This guidance recognizes the importance of time in tracking changes in baseline and requires retrospective estimates, as well as updates moving forward to evaluate changes based on modifications to the system. The document provides a specific protocol for tallying all of the different carbon stocks within a forest (CAR 2010).

Acknowledging the importance of a clearly developed baseline that captures natural ecological, meteorological, and geological variability over time, one might ask what the components of a forest baseline are. The components of a forest baseline extend beyond tree metrics to include wildlife condition, soil condition, geological conditions, and nontimber condition. Driving the baselines of each component are major forces such as catastrophic events, human activity, and meteorology. In addition, each component of the baseline influences the other components and, to some extent, the major events. Figure 3 provides an illustration of the many linkages and factors that underlie a baseline. To arrive at an appropriate accounting of the changes in ecological services, one must clearly track changes in all of the components that define the service.

Figure 3.

Illustration of the interconnections between ecological services and stressors that act on those services.

Additionally, attempting to determine a background condition in a dynamic, transforming forest ecosystem is problematic because conditions may vary regionally and locally. It would seem reasonable to declare that the naturally occurring background condition is the environment that existed before the fire. However, the naturally occurring background condition is not found in an area in which active forest management is occurring to suppress fires. Because the condition of a forest at any given time is the result of multiple driving forces, careful analysis is required to select an appropriate comparative background condition. Ecological conditions in fire-dependent forest systems are historically reset by fire (or transformative event such as storm damage or long-term invasive species stress) and the path to recovery and the ecological services changes with each fire. The degree of service losses (i.e., degrees of injury) and the recovery path are likely to be different for each fire. Every fire is unique, in part, because of its intensity or severity, the weather conditions, and the various landscapes.

The selection of an appropriate baseline condition by which to evaluate service level changes after a severe fire is based on the following guidelines:

  • The baseline condition should be consistent with the management plan for the forest such that a forest that has a history of direct management treatments should not be compared to a baseline condition that assumes no management treatment.

  • Because service changes occur through seasons and across long time periods, the selected baseline condition should cover a similar analytical time frame to capture variability that is independent of the subject fire.

  • An acceptable baseline condition may vary by the type of service under evaluation with respect to the time period and other spatial extent of the analysis.

The components of a baseline may include both quantitative and qualitative features. Reliance solely on features that are easily quantified may miss important features that can be sufficiently characterized qualitatively.


Construction of dams provides another example where decision makers might often fail to consider the full range of ecosystem services that are gained and/or lost. Dams provide a range of ecosystem services including recreation, flood control, debris retention, water supply and irrigation, hydropower, inland navigation, and wildlife habitat (ICOLD 1999; FEMA 2010). Often, however, there are also losses of other ecosystem services throughout the watershed. This is a good example of the transformation aspect of countervailing services, i.e., the services are of different types and they can affect different populations or parts of the ecosystem. A significant transformation is that the impoundment of water (and the associated benefits that can arise from that) can lead to “sediment starvation” and “hungry water” in the watershed below the dam (Kondolf 1997; WCD 2000; McCully 2001). Although water released from a dam is less turbid, and from that one perspective would be perceived as improved quality, the downstream reaches may be starved for sediments. The result can be that these downstream portions are more susceptible to erosion and loss of fish spawning beds as well as alterations of other habitats that are nourished by sediments (Kondolf 1997; WCD 2000; American Rivers 2002). One of the arguments for removal of dams involves transforming the services gained and lost because of dams back to the pre-dam state (MDNR 1995; American Rivers 2002). A digest of work in this arena involves transformations among services such as recreation, hydropower, flood control, disruption of flood pulses, habitats, fisheries, migratory behavior (reproductive, ontogenic, foraging), wildlife, the integrity and stability of river channels and banks, and the sustainability of deltas. The current debate over the removal of dams tends to involve discussions of the relative benefits associated with the dams in contrast with their removal and with alternatives for management of flows and sediments.

The removal project for the San Clemente Dam on the Carmel River is a good example of an effort to transform ecosystem services. The California State Coastal Conservancy (2010) quotes John Steinbeck, who wrote in 1945, “The Carmel is a lovely little river. It isn't very long but in its course it has everything a river should have.” The dam was built in 1921 to supply water for the Monterey Peninsula and perhaps also to provide a measure of flood control, given the flashiness of the river. However, the impoundment behind the dam accumulated sediment, seriously reducing its capacity (Kondolf 1997), and there were consequential changes in the ecosystem services afforded by downstream areas. The services that the dam was constructed to provide have declined to negligible levels. The reservoir capacity has been reduced from approximately 1425 acre-feet to approximately 125 acre-feet. The removal project is intended to result in an increase in different types of ecosystem services in the downstream segments of the river and delta. These include: permanently removing a public safety risk posed by San Clemente Dam, which now threatens 1500 homes and other buildings; aiding in the recovery of central coast steelhead trout, a threatened species, by providing unimpaired access to more than 25 miles of spawning and rearing habitat; expanding public recreation by preserving more than 900 acres of coastal watershed lands, resulting in more than 5400 acres of contiguous regional park land; restoring the natural sediment regime, reducing channel incision and improving habitat for steelhead trout; reducing beach erosion that now contributes to destabilization of homes, roads and infrastructure; re-establishing a healthy connection between the lower Carmel River and the watershed above San Clemente Dam; and improving habitat for threatened California red-legged frogs (California State Coastal Conservancy 2010).

Generally, the original purposes of dams were to improve human quality of life by providing drinking water and to support economic growth by diverting water for power, navigation, flood control, and irrigation. In many ways, dams have succeeded with those goals. However, many large dams, such as the Aswan High Dam in Egypt, were constructed before required environmental impact and human health studies, so the focus was only on the beneficial services it would provide. The Aswan High Dam cost about $1 billion to build (Encyclopedia Britannica 2011) and benefited Egypt by controlling annual floods on the Nile River and preventing the damage that used to occur along the floodplain. The Aswan High Dam provides about a half of Egypt's power supply, it has improved navigation along the river by keeping the water flow consistent, and increased agricultural production and flood protection. It also helped the area avoid significant impacts after 9 years of drought (Encyclopedia Britannica 2011).

However, the construction of the dam changed the hydraulic regime of the downstream river and, thus, the ecosystem and services it provided. Prior to this dam being completed, the Nile River flooded yearly, bringing high water and naturally occurring nutrients and minerals to enrich the soils in the delta and floodplain. Since the completion of the dam, the downstream river is little more than a canal to the ocean. Farmers have been forced to use about a million tons of artificial fertilizer to offset the natural fertilization from yearly flood events. Sediments were captured within the dam, reducing sediment loads to the downstream delta, and causing the number of islands in the river to decline from 150 before the dam to 36 at present (Biswas 2002). Furthermore, the lack of sediments accelerated erosion in the coastal region. The coastline has moved inland a mile or more at some points, bringing salt water into the delta and compromising farmers' use of delta water to irrigate their fields.

The construction of the Aswan High Dam in Egypt provides a dramatic example of unanticipated tradeoffs in ecosystem services. Along with the known ecological costs and benefits was a significant increase in the prevalence of the parasitic disease schistosomiasis (a debilitating human disease of organs in the abdomen). Favorable conditions for snails, which are the aquatic host for the parasite, expanded as a result of increased water in the Delta region of the Nile. Before completion of the dam, irrigation canals and drains were closed and dried up for a few months each year in the winter. After construction, they were continually full. The parasite's life cycle is completed when an active swimming form of the parasite leaves the snail and penetrates a human's skin while in contact with water. The larvae mature in the veins of the urinary bladder plexus or the veins of the intestines. The dam became a center for human activities and evidently for transmission of this parasitic disease, and the irrigation canals and drains now harbor undisturbed and stable populations of the snail throughout the year. In the area above the dam, prevalence was very low before its construction; after completion in 1972, 76% of the fishermen examined in the impounded area were infected (Schmidt and Roberts 1981).

Dams change the chemical, physical, and biological processes of river ecosystems. They alter free-flowing systems by reducing river levels, blocking the flow of nutrients, changing water temperature and O2 levels, and impeding or preventing fish and wildlife migration. These changes can be beneficial or tragic, depending on your perspective. The results of not focusing on longer environmental and public health considerations can be catastrophic and incur monumental costs to deal with the problems and in its toll on lives. The current literature on dam construction and removal projects worldwide is rich with examples of the consideration of a multitude of ecosystem services that are increased and/or decreased. Particular attention is being given to the construction of dams in developing countries, as these may well experience some of the transformations in ecosystem services that have been observed in the developed countries where dams have been used to manage water and provide power. Explicit consideration of ecosystem services including all relevant countervailing services would be valuable for assessing alternatives for water management and power generation in these areas.


Our third example concerns a hazardous waste site in New England where a claim has been brought for natural resource damages (NRD). The site involves a storm water detention and management area that was contaminated by an industrial operation. A remedial investigation and feasibility study was carried out with respect to the contaminated sediments, and a decision was made to remove sediments from the management area. The management area served to buffer flows during high precipitation events in the watershed, thus dampening the potential for floods. In addition, the management area trapped sediments flushed into the system. Excessive sediment loads to downstream portions of the watershed were considered a potential impairment to benthic macroinvertebrate communities, and, therefore, the sediment retention by the storm water system served to reduce this impairment. Management of floods and sediment loads in urbanized watersheds has generally been considered a way to protect and even increase ecosystem services, and there are many examples of best management practices related to the control of excessive flows and sediments. Many states have developed guidance on the use of flood control and sediment detention basins; among the requirement for many of these systems is periodic removal and management of sediments (NCDEM 1994; WMI 1997; Donavan 2000; Sanger et al. 2010).

If storm water detention basins and storm water management areas become shallow as a result of sedimentation, they begin to lose their capacity and thus their service value. In addition, they can support submerged and emergent vegetation. The resultant plant beds offer services in their own right (e.g., as food for wildlife) but those services can accrue to the detriment of other services such as flood control. In our case example, a decision was made to remove the sediments including in areas where plants had developed to remove the contamination present in the sediments. The NRD claim assessed the loss of the services associated with the removal of habitat and considered that loss over the period of time it would take for the plant habitat to recover. We consider this an example of a narrow assessment of service loss. The NRD view of the system selected a set of services that arise only as the primary service of the system diminishes. The removal action for contaminated sediments for this case could also have been viewed as an action that increased the ecosystem services provided by the system, i.e., the original intent to have the system buffer floods and promote sediment deposition. This example of countervailing services would likely need to be resolved by considering priorities for the system, but would certainly have to consider the offsets and substitutions in services.


Consideration of ecosystem services in environmental decision making is often based on incomplete or erroneous understanding of ecosystem processes and functions, and frequently focuses on easily recognizable and popular services such as a view of “pristine” wilderness, or the loss of a fishing hole. The reality of ecosystem service assessment is much more complex. Ecosystem services, as we have explored, are the benefits humans derive from ecosystems. As with any service, the value of ecosystem services may change in response to different stressors, and these changes may occur over long time periods, in response to a catastrophic event, or as a result of a cascade of small changes that interact and subsequently result in subtle shifts. Although the typical approach to evaluating the change in ecosystem services after an event such as a forest fire is to focus on the most obvious and easily valued service changes, this approach ignores the complexity and interconnection of ecosystems. Numerous services are provided by ecosystems and any disturbance might increase or decrease the value of each service and its temporal and spatial distribution. If not accounted for holistically, important service changes may be missed and the system may not be restored to provide the desired or “optimal” suite of services. With a thorough understanding of ecosystems and careful assessment focusing on all components of the ecosystem, an ecosystem service accounting can be completed successfully.


This article reflects discussions among the authors and was not supported by external funding. The authors would also like to acknowledge and thank the anonymous reviewers who provided valuable suggestions for improving the manuscript.