The maturation of ecosystem services: Social and policy research expands, but whither biophysically informed valuation?
Abstract
- The concept of ecosystem services (ES) has risen to prominence based on its promise to vastly improve environmental decision-making and to represent nature's many benefits to people. Yet the field has continued to be plagued by fundamental concerns, leading some to believe that the field of ES must mature or be replaced.
- In this paper, we quantitatively survey a stratified random sample of more than 1,000 articles addressing ES across three decades of scholarship. Our purpose is to examine the field's attention to common critiques regarding insufficient credible valuations of realistic changes to services; an unjustified preoccupation with monetary valuation; and too little social and policy research (e.g. questions of access to and demand for services).
- We found that very little of the ES literature includes valuation of biophysical change (2.4%), despite many biophysical studies of services (24%). An initially small but substantially rising number of papers address crucial policy (14%) and social dimensions, including access, demand and the social consequences of change (5.8%). As well, recent years have seen a significant increase in non-monetary valuation (from 0% to 2.5%).
- Ecosystem service research has, we summarize, evolved in meaningful ways. But some of its goals remain unmet, despite the promise to improve environmental decisions, in part because of a continued pre-occupation with numerical valuation often without appropriate biophysical grounding. Here we call for a next generation of research: Integrative biophysical-social research that characterizes ES change, and is coupled with multi-metric and qualitative valuation, and context-appropriate decision-making.
A free Plain Language Summary can be found within the Supporting Information of this article.
1 INTRODUCTION
The concept of ecosystem services (ES)—the processes by which nature renders benefits for humans (Levine & Chan, 2011)—has risen to prominence remarkably quickly. From a virtual birth in 1997 with the publication of both Nature's Services (Daily, 1997) and an audacious bid to value the world's total ES (Costanza et al., 1997), the field has blossomed as reported in 2012 to be larger than environmental history, environmental ethics or ecological economics (Costanza & Kubiszewski, 2012). Since then, ES became a primary focus of a new intergovernmental UN body, the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES). It is thus now central to many science-policy efforts to describe and address the current ecological crisis (Secretariat of the Convention on Biological Diversity, 2010b; Chaplin-Kramer et al., 2019; Díaz et al., 2019; IPBES, 2019a, 2019b). And yet, the field has been plagued by controversy, from its birth (Nature, 1998) and throughout (De Lucia, 2018; Gómez-Baggethun, de Groot, Lomas, & Montes, 2010; Robertson, 2012). Most recently, controversy surrounds the ES name itself, as opposed to the umbrella term ‘nature's contributions to people’ (Díaz et al., 2018; Pascual et al., 2017). Misunderstanding or disbelieving the argument that ‘nature's contributions to people’ includes ES, some have even reported that the same UN body with ES in its name is distancing itself from both the term and the broader concept (Masood, 2018). Clearly, it is time for a taking stock of the field's intent as well as several substantial critiques. Is there evident progress? That is, is ES as a field delivering the science needed to fulfil its promise, including the knowledge and policy gaps long deemed critical?
There are innumerable decisions at stake, often highly consequential, wherein the role of ES could be decisive and for which ES was designed. Human activities and the infrastructures they introduce often involve large-scale impacts on ecosystems and people's use and enjoyment of them. Consider high-profile oil and gas pipelines (e.g. Keystone in Canada and the United States), hydroelectric megadams (e.g. Three Gorges in China), proposed/new roads (e.g. Trans-African Highway), or compensation for accidents such as oil spills (e.g. Deepwater Horizon in the US) and dam collapses (e.g. Mariana and Brumadinho mining dams in Brazil). Both democratic principles and legislation in almost every nation require the linking of ecosystem impacts and public interest or well-being. For instance, environmental assessment legislation asks whether a project's purported net benefits merit approval given a suite of impacts to ecosystems and ES. IPBES marks the latest effort to support such public policy decisions by leveraging research on biodiversity and ecosystem services (BES) as central to global governance. Antecedents include the Convention on Biological Diversity (CBD), of which most countries are signatories.
Such policy initiatives have recognized implicitly that better societal decisions will follow from a combination of (a) a deeper understanding of value knowledge and reflection: what people care about and how this relates to decisions at hand; and (b) decision support: decision-making that links biophysical and social considerations. The CBD's 2010 Aichi targets commit governments to making people aware of the ‘values of biodiversity’ (i.e. value knowledge and reflection), and to integrate ‘biodiversity values’ (including the importance of ES) into planning and policy (i.e. decision support; Secretariat of the Convention on Biological Diversity, 2010b). This is equally so in IPBES, which also explicitly addresses the importance of diverse conceptions of values and multiple knowledge traditions (Pascual et al., 2017). Note that the above uses of ‘values’ cover several different meanings, which we seek to distinguish from here on: we use ‘value’ in the singular (e.g. ‘value of’) or specify ‘monetary’ or ‘non-monetary’ to signal ‘value as a magnitude of preference’; otherwise we generally use values ‘as notions of appropriate or desired relationships between people and nature’ (Tadaki, Sinner, & Chan, 2017). When ES research is not designed optimally to facilitate knowledge and reflection about the values associated with nature, or to aid environmental decisions, it offers only compromised support in responding to the current ecological crisis. ES research in this form is thus of limited utility to efforts to shift societal trajectories and decisions that undermine nature and its widespread and equitable benefits (Chan et al., 2020; IPBES, 2019a).
Within academia, the legitimacy of ES as a research field has also been called into question. Longstanding debates concern a purported over-reliance on economic approaches to valuation, and the field's resistance to critical and interpretive scholarship in social sciences and humanities (Adams, 2014; Fish, 2011; Gómez-Baggethun & Ruiz-Pérez, 2011; McCauley, 2006). Recently, IPBES has become a lightning rod for these concerns (Díaz et al., 2015, 2018).
To fulfil the needs of CBD, IPBES, and multiple organizations and levels of government world-wide, what matters is that a broad expanse of ES research fulfils both value-knowledge and decision-support goals. Although key ES publications have long pursued these two objectives as primary (e.g. Costanza et al., 1997; Daily, 1997; Daily et al., 2009; TEEB, 2009), they have not always done so in recognition of the literatures on how to achieve value-knowledge and decision-support. This paper first delineates five substantial areas of critique and potential growth for the ES field, five calls for better or ‘best practices’ and for addressing the needs of specific applied contexts. From these five areas, which were inspired by the above concepts of value knowledge and reflection and decision support, we derive commensurate hypotheses regarding how this interdisciplinary field is maturing (or not). We then use these to analyse a randomly selected representative set of ES papers from a bibliometric list of all those published between 1990 and 2017 (inclusive; see Methods). Others have sought to answer whether the field of ES already had elements that inspired the term ‘nature's contributions to people’ (Kadykalo et al., 2019); here we ask, how responsive was the field of ES across years to a broader set of fundamental critiques?
1.1 Critiques and associated best practice
1.1.1 Valuation at the margin, informed by biophysical science about change
Most experts agree that economic values are only useful for aiding decisions when they express or reflect differences in human well-being across a set of options; and in an ES context, such options must include assessments of realistic changes in ecological (biophysical) function or outputs (Bockstael, Freeman, Kopp, Portney, & Smith, 2000; Ricketts et al., 2016; Wainger & Boyd, 2009). If the choice at hand is whether to manage an upstream watershed for water quality services, the biophysical science needed would express the difference in water quality associated with management. The marginal change is thus the difference between status quo future and watershed management options.
This need for biophysically informed valuation ‘at the margin’—that is, valuation of realistic ecological changes—might seem obvious, but it was not always so. Early ES research valued ES of whole systems in temporally static dollar equivalents, as would be appropriate if the decision at hand was possible biome obliteration. Most famously, Costanza et al. (1997) calculated the average value of the world's ES as $33 trillion per year. The continued interest in ‘total value’ (Costanza et al., 2014) estimates associated with the wholesale loss of a habitat, ecosystem or even biome—despite their lack of decision-aiding—can be explained by their utility as a pragmatically useful signal for raising awareness of natural capital itself (Nature, 1998). This paradox of apparent utility was best expressed by the juxtaposition of Michael Toman's critique of the $33T figure as ‘a serious underestimate of infinity’ (without ecosystems, there is no economy; Toman, 1998), and Trudy Cameron's assessment that the effort was ‘a recklessly heroic attempt to do something that's futile … [but] … very useful’ (Nature, 1998).
Responding to these critiques, newer ‘production-function’ approaches have advanced the characterization of changes in ES and how their ‘marginal’ valuation varies across alternative resource management decisions (Daily et al., 2009; Fisher et al., 2008; Kareiva, Tallis, Ricketts, Daily, & Polasky, 2011; Nelson et al., 2009). The point to be characterized ecologically, the biophysical culmination of the ES (e.g. water quality at the point of consumption), is the ‘ecological endpoint’ and so the quantity change that can be valued (Wainger & Boyd, 2009).
Given the resounding critiques and the shift towards this more technical, decision-support role for ES research, we should expect that most valuation research includes some characterization of biophysical differences across scenarios. We might also expect an increase in this proportion of biophysically informed valuation across time (Expectation 1).
1.1.2 An important, but limited, role for monetary valuation
Also prominent but less prevalent in the ES literature is recognition that monetary valuation is necessary, but also limited in applicability. Monetization is a key mechanism by which trade-offs can be weighed, a fundamental objective of evidence-based policymaking. Few would deny that, if an ecosystem change affects production of a market commodity, such effects should be valued in monetary terms (but see Melathopoulos, Cutler, & Tyedmers, 2015). For example, coffee production is enhanced by native habitat via pollination by native bees (Ricketts, Daily, Ehrlich, & Michener, 2004), and by pest control by native birds (Karp et al., 2013). Yet few are fully comfortable with monetary valuation of an endangered species. For instance, after Canada's environment ministry commissioned an economic valuation of polar bears, the proponent himself (Minister Kent) acknowledged the appropriateness of public scepticism, and admitted himself that polar bears are ‘priceless’ (Canadian Press, 2011). These limitations mean that monetary values in isolation are of limited utility to either value knowledge and reflection or decision making. A balance of monetary and non-monetary valuation would seem important for both overarching purposes.
While a thread of the ES literature continues to value all manner of things in monetary terms, another thread has emerged to challenge such applications (Saner & Bordt, 2016; Turnhout, Neves, & De Lijster, 2014). Costanza et al.'s aforementioned valuation of all the world's ES relied heavily on studies of stated willingness to pay, which were countered by Sagoff as inappropriately conflating consumer and citizen preferences (what each of us wants vs. what we believe is right for society; Sagoff, 1998). More recently, we and colleagues have argued that valuation methods should match the values at hand (the reason something matters; Chan et al., 2011)—for example, monetary valuation is effective for characterizing consumer but not citizen preferences. As one dimension, one can distinguish consumer versus citizen preferences (as expressions of value), adding also virtues (expressions of morals as standards or values). In addition, there are at least seven other dimensions by which we might categorize a given preference, principle or virtue (market-mediated vs. non; self- vs. other-oriented; individual vs. holistic/group; physical vs. metaphysical; final vs. supporting; transformative vs. non; anthropocentric vs. bio-/ecocentric; Chan, Satterfield, & Goldstein, 2012). Across this multidimensional values ‘space’, only one component is amenable to economic valuation (other components conflict with its assumptions). Importantly, these critiques apply not only to stated willingness to pay (e.g. McFadden & Train, 2017) but they also question the ability of revealed willingness-to-pay methods to reflect how much a thing matters or is worth. In essence, using monetary value to measure the worth of biodiversity is akin to measuring a person's worth based on their salary.
The ES literature has thus called for experimentation with a greater suite of valuation methods (Cheng, Van Damme, Li, & Uyttenhove, 2019; Scholte, van Teeffelen, & Verburg, 2015), alongside the recommendation that the focus of new methods be those services too complicated or contentious to value economically. Examples of trial methodological innovations have followed: group-discursive methods (Kenter et al., 2016; Wilson & Howarth, 2002), deliberative monetary valuation (Lliso, Mariel, Pascual, & Engel, 2020; Lo & Spash, 2013; Spash, 2008), use of exploratory interviews to identify benefits and values at stake connected to ES (Gould et al., 2015; Klain & Chan, 2012; Klain, Satterfield, & Chan, 2014; Walz, Grêt-Regamey, & Lavorel, 2016), photo elicitation of values (Sherren, Fischer, & Price, 2010; Southon, Jorgensen, Dunnett, Hoyle, & Evans, 2017), articulation and quantification of values using newly ‘constructed’ metrics fit for purpose (Failing, Gregory, & Higgins, 2013; Satterfield, Gregory, Klain, Roberts, & Chan, 2013) and more. A further critique is that many services include things for which monetary valuation would be controversial (e.g. biodiversity), and so only biophysical metrics are provided (Bateman et al., 2013; Guerry et al., 2012; Nelson et al., 2009; Ruckelshaus et al., 2015). And, lastly, many contend that mainstream ES research had avoided whole classes of ‘intangible’ or ‘messy’ benefits—those less amenable to monetary valuation—which led to the frequent representation of cultural ES as recreation and aesthetic value only (Daniel et al., 2012; Langemeyer, Calcagni, & Baró, 2018; Milcu, Hanspach, Abson, & Fischer, 2013).
Given these critiques, we expected that the longer tradition of monetary valuation will prevail, but that—as the field has matured—we might also see greater balance of monetary and non-monetary valuation. That is, accepting the longer tradition of monetary valuation, we should only see an increase in the proportion of non-monetary valuation over time (Expectation 2).
1.1.3 Without demand and access, there is no value
If people do not want or need a thing (including existence value), or if they cannot or do not obtain it, there is no valued benefit (Wieland, Ravensbergen, Gregr, Satterfield, & Chan, 2016). Conversely, shifts in demand and access have devastating implications for some ES (Erdozain et al., 2019). Thus understanding access and demand is crucial to both value knowledge and reflection and decision-support. When the demand for coffee on the global market made coffee-growing less feasible economically, large tracts of Costa Rican coffee plantations turned to pineapples, which wholly eliminated demand for insect pollination (Chan et al., 2007). And when First Nations people in BC, Canada were forcibly moved to residential schools, harvest of edible seaweed and the concomitant transmission of the knowledge and practice of harvest was largely lost (Turner & Turner, 2008). Access and so demand for marine foods has also curbed as a function of perceived contamination near industrial sites and increased price of boat fuel necessary for reaching harvest sites (Chan et al., 2011).
In recognition of the critical nature of both demand and access, other fields have scrutinized these aspects explicitly. Demand, for example, has been shown to be a product of much more than personal preferences or individual behaviour; it is also largely a function of the social-technical context in which decisions take place (Wilhite, Shove, Lutzenhiser, & Kempton, 2000). Thus energy demand is a function of infrastructure (e.g. the transmission grid), devices owned and social norms about their use (Shove & Walker, 2010; Wilson & Dowlatabadi, 2007). Surely the same is true of demand for many ES (e.g. demand for water-purifying ES depends on presence of filtration plants, private filters, habits and norms about water use).
The social sciences have also placed great emphasis on access—focusing on the ability to actually derive benefit from a thing, determined not just by a ‘bundle of rights’ but more broadly by ‘a bundle of powers’ (Ribot & Peluso, 2003). Such a broad notion of access inspires investigation of not only rules, but also skills, social networks and capital, access to markets and labour, tenure security, etc. (ibid). Accordingly, if we are to understand the value (and justice) of ES in various future scenarios, we need also to understand changes in these aspects of demand and access.
The ES field has paid little explicit attention to the dynamics of demand and access (Villamagna, Angermeier, & Bennett, 2013; Wieland et al., 2016), with some exceptions. This may stem from the origin of ES as a field to help communicate the benefits and value of nature (Daily, 1997; Gómez-Baggethun et al., 2010; Mooney & Ehrlich, 1997). Consequently, the focus has often conflated the role of changing ecosystems as changes in ES, while ignoring the role of changing social systems (including demand and access; Figure 1). This is so despite researchers explicitly identifying the importance of accounting for spatial variation in ES access and demand, especially for spatial prioritization (Bateman et al., 2013; Chan, Shaw, Cameron, Underwood, & Daily, 2006; Fisher et al., 2011; González-García, Palomo, González, López, & Montes, 2020; Luck, Chan, & Fay, 2009; Luck, Chan, & Klein, 2012; Naidoo & Ricketts, 2006; van Jaarsveld et al., 2005; Verhagen, Kukkala, Moilanen, van Teeffelen, & Verburg, 2017; Wolff, Schulp, Kastner, & Verburg, 2017). Similarly, and borrowing from the field of vulnerability studies, there is much evidence to suggest that social drivers such as poverty and inequality obstruct access to ES (Berbés-Blázquez, González, & Pascual, 2016; Hicks & Cinner, 2014; Nesbitt, Meitner, Girling, Sheppard, & Lu, 2019; Robards, Schoon, Meek, & Engle, 2011; Turner & Turner, 2008; Wieland et al., 2016).
Given the ecological roots of ES, we expected less examination of demand and access or other social factors than that of biophysical dimensions. Interpreted generously to include how valuations of ES vary due to social factors across space, time or social group, we thus anticipate only a small number of studies examining the social drivers of value, with an increase over time (Expectation 3).
1.1.4 ES for tailored communication
A key part of the mandate of ES research from the beginning was that it be useful to policymakers and lay people (in both value knowledge and reflection and decision-support); that is, to be specific, relevant and timely. An example is the critical difference between an agronomist touting the necessity of native prairie vegetation in hopes of changing farmer decision-making around field buffers, versus demonstrating that modest buffers of native prairie could improve ES such as soil retention and moisture, reduce runoff and also increase biodiversity on farms, even within a year (Liebman, Helmers, Schulte, & Chase, 2013). In this example of useful research, researchers calculated the costs of such on-farm changes, and demonstrated their economic feasibility (Tyndall, Schulte, Liebman, & Helmers, 2013). The combination of this tailored information and an appropriate participatory engagement strategy has led to widespread interest in this new farm management practice (Jordan et al., 2013). The general knowledge of prairie vegetation's ability to provide ES was widespread but alone was insufficient to garner such changes. Further optimal examples include demonstrating to foresters how alternative forestry business practices can supply ES and a sustainable stream of revenue (Goldstein et al., 2012), to regional planners and environmental regulators how marine plans can improve water quality, recreation, renewable energy, coastal protection and more (Guerry et al., 2012), and to policymakers and philanthropists how renewed support for ES programs could enhance carbon sequestration and reduce dust export (Liu, Li, Ouyang, Tam, & Chen, 2008).
While some ES studies have paid great attention to specificity and relevance of this kind, many have called for more and better demonstration in concrete terms. For example, how might biodiversity enhance needed services or human well-being more broadly. ES research should, ideally, change discussions and decisions (Balvanera et al., 2001; Daily et al., 2000; de Groot, Alkemade, Braat, Hein, & Willemen, 2010; Kumar, 2010; Posner, McKenzie, & Ricketts, 2016). The very design of the Natural Capital Project's InVEST tool was to meet the needs of decision-makers, supplying relevant information even without sophisticated data (Daily et al., 2009; Tallis & Polasky, 2011a, 2011b). And lessons from applications of this tool reveal that such simple information is useful, especially when linked with an iterative science-policy process and on-ground capacity building (Ruckelshaus et al., 2015).
Given this strong intellectual foundation as a solution-oriented science, we expected to see a large number of studies demonstrating how the biophysical underpinnings of ES are affected by particular decisions, scenarios or other factors—with or without valuation. We also expected a smaller number of studies directly addressing the effectiveness of policies (but not necessarily characterizing the biophysical underpinnings of ES; ‘policy efficacy’). Furthermore, we expected an increase in both of these aspects over time (Expectation 4).
1.1.5 A social turn to ES?
Three other social considerations have been central to ES critiques. First, at the heart of alternative framings like ‘nature's contributions to people’, as adopted by IPBES, is the contention that the language of ES may alienate a variety of audiences, including Indigenous people as well as scholars from the more interpretive social sciences (qualitative research, e.g. anthropology and critical human geography; Sikor, 2013), as opposed to the more predictive disciplines of psychology and economics (Díaz et al., 2015, 2018; Satz et al., 2013). Certainly, several reviews have found limited engagement with the social sciences in ES research (Chaudhary, McGregor, Houston, & Chettri, 2015; Fagerholm, Torralba, Burgess, & Plieninger, 2016; Haase et al., 2014; Liquete et al., 2013; Luederitz et al., 2015; Nieto-Romero, Oteros-Rozas, González, & Martín-López, 2014). Second, issues of justice, equity and the social implications of ES programs and policies have long been prominent in critiques of ES (Corbera, Kosoy, & Martínez Tuna, 2007; Daw, Brown, Rosendo, & Pomeroy, 2011; Gould et al., 2020; Pascual et al., 2014; Sikor, 2013). Finally, the recent rise in research on relational values (Chan et al., 2016; Chan, Gould, & Pascual, 2018; Pascual et al., 2017) stems in part from the recognition that values are not only the output of valuation exercises, but are also reciprocal responsibilities, relations and structured priorities that can drive changes in ES and social–ecological systems (Chan, Satterfield, et al., 2012; Comberti, Thornton, Wylliede Echeverria, & Patterson, 2015; Fish, Church, & Winter, 2016; Gould, Pai, Muraca, & Chan, 2019; Tadaki et al., 2017). By better including scholars from the qualitative social sciences and humanities, better attending to the social implications of programs and policies, and better reflecting diverse conceptions of values, ES research could improve in both value knowledge and reflection and decision-support.
Given this, we expected to see a relatively small engagement from the qualitative social sciences and humanities, investigating the social dimensions of ES policies and programs (policy equity), or investigating values as drivers of change (Figure 1). Believing ES to be a field responsive to critique, however, we also expected to see a rise in all three of these aspects over time (Expectation 5).
2 METHODS
We tested the above predictions via a stratified random sample (n = 1,047) of all eligible ES studies in ISI Web of Knowledge as our sample frame. The sample frame comprised 16,833 studies, which had to include the term ‘ecosystem service*’ (in the title, abstract or keywords), which was too many to be evaluated comprehensively (hence the random sample). We separated the frame into five time periods: incipient (1990–2001—242 studies, n = 64 included), early (2002–2008—1,465 studies, n = 70), early-middle (2009–2011—2,989 studies, n = 172), late-middle (2012–2015—5,487 studies, n = 335) and recent (2015–2017—6,650 studies, n = 406). Recognizing that any time division is somewhat arbitrary, we thought of the first period as the era of Nature's Services (Daily, 1997) and Costanza et al. (1997), the second as the era of the Millennium Ecosystem Assessment (2005), and the last being the era of IPBES (with the third and fourth periods evenly splitting the intervening years). Because of the near-exponential growth of publications (Figure 1), we employed wider temporal windows for the first two periods (assuming that change might occur more rapidly later) and oversampled from the first period (adding 50 to the initial 14). Contributions of the oversampled early period were proportionally weighted in the all-time-period averages of the 1,047 studies. This sample was sufficient for 95% confidence intervals with a margin of error of ±0.04.
Based on reviews of the abstracts (and where necessary, article text), each study was coded in binary terms for five qualities that directly addressed the hypotheses above (not mutually exclusive; as per Coding explanation below). These are summarized in Figure 2 and include whether the sampled study reported any (a) valuation results (total or marginal values), including any reporting of valuation in non-monetary terms; (b) biophysical changes pertinent to ES; (c) social changes or differences pertinent to ES (access, demand or differential valuations across time, space or social group—‘use’), and any contribution from qualitative social sciences and humanities; (d) policy research, either about efficacy in achieving desired changes in biophysical or ES outcomes, or about equity and other social dimensions; and (e) values as drivers of change. Because any given coding for any paper involves a degree of subjective judgment and substantial grey areas that do not neatly fit into classifications, we strived first and foremost for consistency, but assume that absolute proportions matter much less than the trends.
We interpreted these coded qualities (above) as tests of our five expectations: (Expectation 1, E1) the proportion of valuation studies also reporting relevant biophysical changes as the fraction of biophysically grounded marginal valuation; (E2) the proportion of valuation studies reporting non-monetary metrics; (E3) the number of studies investigating social aspects; (E4) the proportion of studies characterizing biophysical underpinnings of ES, and those investigating policy efficacy; and (E5) the proportions of studies using approaches from the qualitative social sciences and humanities, addressing policy equity, and values as drivers of change. Figure 2 summarizes these expectations.
More detail about coding appears below. All data generated or analysed during this study are included in this published article (and Supplementary Data).
2.1 Coding
Codes were intended to enable quantitative testing of the five hypotheses. Each code was binary and independent (so non-mutually exclusive, e.g. papers could be scored as biophysical and social). Papers could in theory be scored as ‘1’ for many codes, or ‘0’ for all codes (e.g. if the paper mentioned ES in a valorizing, that is as a pragmatically useful signal, without engaging substantially in biophysical or social classification, etc.). Codes were developed by Chan and Satterfield and operationalized by a team of research assistants. Papers were assigned codes based largely on abstracts, but where abstracts were unavailable or ambiguous, we consulted the full article. Discrepancies were noted and resolved, by iteratively coding selected papers (amounting to >30%) and resolving discrepancies in outcomes and fine-tuning criteria. Chan second-coded all papers where difficult or discrepant codes persisted.
2.1.1 Biophysical
Articles were scored as biophysical characterizations of ES if they investigated biological, physical or biophysical patterns or processes that pertained directly or explicitly to ES (which could directly inform valuation). For example, a study investigating the crop-pollinating behaviour of bees and its response to landscape characteristics would qualify as biophysical. In contrast, a study that merely notes that bats are ES providers while investigating bat mate preference (without any stated or apparent link between mate preference and ES) would not qualify. Articles can qualify as biophysical even if not gathering primary field or remote-sensing data, for example, by combining existing data to illuminate a fuller characterization of ecosystems’ contributions to benefits for people.
2.1.2 Valuation
Articles were scored as valuation if they explicitly characterized the importance or value of ES to people by one or more quantitative metrics (in practice, most used monetary metrics such as dollars). Articles were counted as valuation if they reported on a primary study of this nature or if they employed the results of other studies to characterize values associated with one or more ES or their changes; articles were not counted if they simply reviewed or mapped valuation data reported elsewhere. However, including such applications of valuation data had little effect on results. Articles assessing emergy were not counted as valuation due to the lack of an explicit link to any benefits or beneficiaries. One set of subcodes included monetary and non-monetary, where monetary included any expression of willingness to pay (in any currency), and non-monetary included any other subjective expression of importance or priority including, for example, indices of risk. A second set of subcodes included marginal versus total valuation, where marginal pertained to valuation of a realistic change and total pertained to the value of the whole of something (e.g. a whole ecosystem, a whole class of ES).
2.1.3 Social
Articles were scored as social characterizations of ES if they investigated social, cultural or psychological patterns or processes that pertained directly or explicitly to ES. Articles were counted as social if they investigated beneficiaries’ perceptions of an ES or even how valuation data varied across groups; articles were not counted if they investigated perceptions of an intervention (e.g. a park—but see policy, below) without specific reference to particular ES or if they only discussed the role of social dynamics without new data or analysis.
2.1.4 Policy
Articles were scored as policy if they investigated programs or policies for ES (e.g. payments for ES programs), or the effects of a broader range of programs and policies on ES (e.g. forest policy). Articles were counted as policy only if they investigated the effects and design of specified programs and policies, not—for example—potential biophysical effects assuming changes on landscapes resulting from purely hypothetical policies or programs via scenario modelling. Policy articles were subcoded into those that investigated the efficacy (effectiveness, efficiency) of a policy or program effecting a change in BES outcomes, and equity or other social side effects of those programs.
2.1.5 Valuation and biophysical
Because of the core promise of ES research to provide valuation data to relate the implications of biophysical changes for human well-being, we tracked the number of articles that did just that. This scoring is simply the union of valuation and biophysical above.
2.1.6 Qualitative social sciences and humanities
Articles were scored as qualitative social sciences and humanities if they adhered to the methods and conventions that emphasized the qualitative research that also engages with co-production of knowledge, traditional knowledge, meanings of nature as well as contextual and historical explanations of inequity and policy outcomes (e.g. as might a political ecologist using ethnographic methods). Articles could be scored as qualitative social sciences and humanities even if they also engaged in some quantitative research, but not if the prevailing approach was one generally associated with economics, psychology and other predictive social sciences.
2.2 Analysis
We calculated SEs for the proportions for each time period following standard protocol. For the total (across all time periods), we accounted for oversampling of 1990–2001 by normalizing the pertinent numbers (by multiplying by the original n = 14/the oversampled n = 64). Accordingly, error bars reflect a total n = 997, not n = 1,047. For all proportions, however, we applied the finite population correction (Isserlis, 1918).
3 RESULTS
Our analysis of the ES literature depicted in Figure 3 indicates relatively little valuation research, and very little valuation informed by needed biophysical or social changes or differences. There is however a marked increase in studies examining the social dimensions of ES-pertinent policy, and in the proportion of studies attending to social dynamics underpinning ES.
, 15%–26%); Policy (
, 5%–17%); Valuation (
, 5%–9%); social (
, 2%–9%); and Valuation and Biophysical (
, 0%–4%—a subset, biophysically grounded valuation). Error bars represent SEs. Proportions do not add to 1, as codes were independent (except valuation and biophysical) and some papers were coded as 0 for all categories depicted here. Colours conform to Figure 1Specifically, some key best practices for ES research are not yet manifest in substantial changes in ES scholarship: we see little biophysically grounded valuation of realistic changes (see 1 below), and little inclusion of non-monetary metrics alongside monetary ones. While it is unreasonable to expect any one study to ‘do it all’ (original biophysical, social and valuation research), our analysis did not require biophysical, social or valuation research to be fully original research—they could equally have integrated data or findings reported elsewhere. Furthermore, although in theory biophysical research could be published separately and subsequently integrated into valuation, until the final period (2015–2017), most ecological research in our sample was so removed from the processes or components of value that it could not even be integrated with valuation post hoc (that is, it did not qualify as biophysical ES research by our coding).
In contrast, the field shows clear trends towards improvement in a variety of respects, with impressive increases in many social aspects (including social aspects of policy; Figures 3 and 4), as described below (2–5 below). The following sections report on the findings in relation to the five expectations identified earlier (Figure 2).
, 0%–12%) and efficacy (
, 4%–5%); also shown are Qualitative social sciences and humanities (
, 0%–7%), non-monetary valuation (
, 0%–2%); Values as drivers (
, 0%–1.5%). Proportions do not add to 1, as codes were largely independent (note policy equity and policy efficacy are subcodes of Policy in Figure 3, and non-monetary valuation is a subcode of Valuation), and some papers were coded as 0 for all categories depicted here. Error bars represent SEs. Colours conform to Figure 13.1 Whither the biophysically grounded ‘marginal’ valuation?
Expectation 1: In contrast to optimistic expectations, a tiny fraction of the representative sample (n = 1,047) conducted biophysically grounded valuation (2.4 ± 0.5%, Figure 3), and this fraction seemed invariant across time.
Many of the qualifying papers were from agricultural contexts, where pest control or pollination services associated with field edges, forest patches or floristic diversity were quantified monetarily through enhanced yields (Geslin et al., 2017; Woodcock et al., 2016; Yang, Elbakidze, Marsh, & McIntosh, 2016). These contexts, with clear marketed commodities, might be atypical of other ES—low-hanging fruit, as it were.
Despite repeated calls for ‘marginal’ valuation of realistic biophysical changes, the ES literature continues to contain a large share of studies taking a ‘total’ (not marginal) valuation approach. All parties seemed to agree that marginal valuation is more relevant (Bockstael et al., 2000; Costanza et al., 1997; Fisher, Naidoo, & Ricketts, 2015; Nature, 1998). Yet the ease of publishing such total valuation results appears to continue to tempt many authors to report these results, often borrowing from existing global values, sometimes with revisions.
Of course, total values have some purpose. The continued presence of total valuation studies even in the last period is partly due to the rise of non-monetary valuation, most of which was not measuring the importance of realistic changes but of whole categories of benefits (one example of a ‘total’ valuation approach). If the ES field is to contribute meaningfully to on-ground decision-making, much greater use and examination of marginal valuation will be needed, including with non-monetary metrics.
3.2 Whither the balanced multi-metric valuation?
Expectation 2: In contrast to expectations, only a small fraction included valuation results (7.6 ± 0.8%). This number includes studies re-using others' results but not straightforward re-mapping exercises, as in several studies applying values from Costanza et al. (1997) at finer scales. There was less balance than expected between monetary and non-monetary valuation, and the rise of non-monetary came later than predicted. Only 1 of 641 papers in the sample until mid-2015 assessed people's values using non-monetary approaches (0.2 ± 0.2%; in contrast, 40 provided monetary values in that sample), but this proportion changed significantly and positively in the final period (to 2.5 ± 0.7%—10 non-monetary vs. 27 monetary across the 406 sampled; Figure 4). Only one paper in our sample appeared to integrate monetary and non-monetary valuation (Villegas-Palacio, Berrouet, López, Ruiz, & Upegui, 2016). Interestingly, the rise in non-monetary valuation appears to have worked against a rise in marginal valuation, as most non-monetary valuation studies used a total valuation approach (all but three).
Despite prominent calls for non-monetary valuation (EPA, 2009; IPBES, 2015; Value of Nature to Canadians Study Taskforce, 2017), the vast majority of studies report only monetary values (or simple biophysical quantities, which does not signal relative importance or need). The lack of attention to non-monetary valuation in the ES literature until 2015 was wholly out-of-sync with the balance of attention to monetary and non-monetary valuation in best-practice guides (approximately equal; EPA, 2009; IPBES, 2015; Value of Nature to Canadians Study Taskforce, 2017). This pattern did change dramatically in the 2015–2017 period, but monetary valuation continues to dominate—perhaps owing to a common notion that policymakers and the media only care about monetary values (a notion contradicted by documented engagements with decision-makers—Ruckelshaus et al., 2015).
Both monetary and non-monetary values have important contributions to make. To reflect the diverse ways in which nature matters to people, we need more of both kinds of valuation (Chan, Satterfield, et al., 2012; IPBES, 2015), more integration across the two (Jacobs et al., 2016; Kenter et al., 2016; Pascual et al., 2017) and more marginal non-monetary valuation to valuate realistic changes in ES.
3.3 Is ES research responsive to social drivers of change?
Expectation 3: As expected, characterization of relevant social patterns, changes or processes (5.8 ± 0.7%; Figure 3) was rare, particularly in the early periods. Until the final period, very few of these social studies of ES examined changing demand or access to ES—most examined only spatial or social variation in management practices or willingness to pay. The increase in the social dimensions of ES studies was marked (1.6%–8.6%; Figure 3).
The contribution of the social sciences to interdisciplinary ES science has thus far been largely to assess which ES matter, rather than to understand underlying access and demand. Our review also encountered a paucity of studies explicitly addressing social dimensions of ES, particularly pre-2012, even by the most generous scoring (Figure 3). Of course, there are whole fields studying access to and distribution of environmental goods and bads (e.g. environmental justice, Schlosberg, 2007), but these fields are having limited effect within the ES literature as we operationalized it (with noteworthy exceptions, e.g. Sikor, 2013). Thus, most studies qualifying as ‘social’ investigated only how valuation data varied across social groups, or made a link between a biophysical or policy change and loss of access to some groups. Very few studies in our sample explicitly studied social dynamics affecting access and demand for ES, or the social consequences of ES change, but these more sophisticated social studies were increasingly common. This dearth is perhaps unsurprising as even best-practice guides have largely overlooked these aspects.
Social sciences have much more to offer ES than the measurement of value. To understand how changing ES will affect people, we must include the studies of access, demand, exposure and co-production, among other dimensions (Chan & Satterfield, 2016; Fish et al., 2016; Palomo, Felipe-Lucia, Bennett, Martín-López, & Pascual, 2016; Soga, Gaston, Koyanagi, Kurisu, & Hanaki, 2016; Soga, Gaston, Yamaura, Kurisu, & Hanaki, 2016; Wolff, Schulp, & Verburg, 2015). We find optimism, however, in the increase in social dimensions studies since 2012 (Figure 3), and the qualitatively heightened sophistication of these studies with time.
3.4 Is ES research strong in targeted biophysical and policy studies?
Expectation 4: Approximately one quarter of the representative sample (n = 1,047) of studies included characterizations of biophysical processes relevant for ES (24.1 ± 1.3%; Figure 3). Thus, biophysical was the most common code depicted in Figure 1, but it was perhaps still too rare given that biophysical change is fundamental to ES and original research was not required in coding. Biophysical studies adopted a range of approaches including field experiments (especially in agricultural and forest contexts) and modelling using ES software and available data (e.g. for many hydrological and carbon storage studies). A smaller but rising number of studies addressed policies directly (Figure 3), but further differentiation revealed that the growth in policy studies was not true of the subset investigating policy efficacy (Figure 4).
Although initially a large number of studies referenced ES in a perfunctory ‘buzzword’ (valorizing) manner (as found by Abson et al., 2014), this proportion decreased substantially over time, in part due to the substantial rise in studies characterizing the biophysical underpinnings of ES (Figure 3). There was also a steady proportion of studies addressing the efficacy of policies (Figure 4).
3.5 Has ES made better use of the social science of values?
Expectation 5: As predicted, results suggest a substantial turn towards influences from a broader representation of the social sciences and humanities in the last or last two periods (2012–2015, 2015–2017) after near invisibility pre-2012, in three respects. First, there has been a substantial increase in the proportion of studies addressing the key social considerations of ES-driven policies or programs, including equity—the starkest rise we observed (policy equity: 0%–12%; Figure 4). Second, there is an evident rise in studies from the qualitative social sciences and humanities (0%–7%; Figure 4). Third, following a near complete domination of thinking about values as merely the outputs of valuation, since 2012 studies considering values as structured priorities or relations have increased, with the potential to drive change in social-ecological systems research (values as drivers: 0%–1.5%, but only eight studies of 1,047; Figure 4).
4 DISCUSSION
The partial maturation of ES research detailed here suggests that what initially began as a framework for ecological-economic integration has attracted other analytical approaches but resisted the long-sought integration. The rise of non-monetary valuation, qualitative social assessment, social studies of access, demand and equity reveals that economics is clearly no longer the only social science deeply engaged with ES. And yet the rarity of integrative studies—for example, that both characterize relevant biophysical changes underpinning services and the implications of those changes via valuation (Gregr et al., 2020)—suggests that the interdisciplinary integration that excited so many remains only partly fulfilled.
4.1 The answer? Not a number
Much of the ES literature we reviewed still appears to imagine (naïvely) that research is most effective if we can distil the value of nature into a number (monetary or otherwise), and then communicate that number broadly. In keeping with this additive and linear view of ‘valuation-plus-dissemination’, policy agents at the highest levels are developing methods for BES valuation (EPA, 2009; IPBES, 2015; Value of Nature to Canadians Study Taskforce, 2017) to supply indices—economic or otherwise—for planning and policymaking. We might hope that the communication of such unspecific values to ‘the public’ and policymakers will foster ‘changes in personal consumption and behaviour’ (Secretariat of the Convention on Biological Diversity, 2010a) and integration into management and policymaking, but this may be largely wishful thinking (Allum, Sturgis, Tabourazi, & Brunton-Smith, 2008; Singh et al., 2020). Indeed, abundant research on environmental values (i.e. as statements of economic and non-economic importance) reveals that the valuation-dissemination approach conveys only a detached and one-dimensional representation of why BES matters to people (Chan et al., 2016; Turnhout, Waterton, Neves, & Buizer, 2013). As such, this approach will likely fail to realize the key goals of (a) value knowledge and reflection and (b) decision-support, for two kinds of reasons: analytical and practical.
4.2 The analytical shortcomings of numerical valuation
Any approach that characterizes societal value primarily through an analytical aggregation of individual valuations (‘numerical valuation’)—as did most of the valuation papers in our sample—will fail to properly account for what matters and why (Gregory, Lichtenstein, & Slovic, 1993; Spash, 2008). This includes numbers derived from life-satisfaction surveys, which avoid some of the limitations of monetary valuation and offer solutions to some coarse-scale policy contexts (e.g. green-accounting measures as indicators of national well-being). The problem is that values of biodiversity and ES are diverse (Chan et al., 2011; Mace, Norris, & Fitter, 2011) and can be characterized by multiple qualities that defy appropriate representation via simple aggregation across individual valuations (Lliso, Mariel, et al., 2020). The reasons for this are well-articulated in the literature, including the following.
Many BES values are of a non-material (intangible) or extra-material nature (linked to material objects; e.g. place attachment, inspiration), and not amenable to simple numerical valuation (Chan, Satterfield, et al., 2012; Turner, Gregory, Brooks, Failing, & Satterfield, 2008). Some values are so culturally or socially sensitive as to defy analytical resolutions of trade-offs with other values, but often trade-offs can be resolved through discussion and negotiation (Chan, Satterfield, et al., 2012; Daw et al., 2015; Gregory, Failing, Harstone, Long, & McDaniels, 2012). One cannot determine group values—which are often crucial for BES—through individual polling alone (by group values we mean what a group wants for itself as a social whole; Kenter et al., 2016; Wilson & Howarth, 2002). Abundant evidence demonstrates that preferences—about non-market goods in particular—are malleable and dependent on framing (Fischhoff, 1991; Gregory et al., 1993). Some have argued persuasively that therefore values should be systematically and consistently ‘constructed’ in structured participatory decision-making approaches that reflect shared values (Gregory et al., 1993; Keeney & Gregory, 2005). Further, values can pertain not merely to outcomes, but also to rights and to processes of decision-making and implementation (Chan, Satterfield, et al., 2012; Spash, 2006). To ignore such values about how decisions are made or management is conducted seems folly: how something is done matters as much to many constituents as the outcomes produced. These issues can be addressed by using such methods as narrative approaches (Gould et al., 2015), negotiation (Vignola, McDaniels, & Scholz, 2012), group deliberation and participatory decision-making mechanisms (Gregory et al., 2012; Spash, 2008).
4.3 In practice, engagement is the problem and the solution
Despite the pre-occupation of ES studies with numerical valuation, numbers alone are an ineffective tool for crucial task of promoting value-knowledge and reflection (Spash, 2008). For this purpose, of helping people to truly understand the contributions of nature to their lives, we cannot expect a complete solution from the most space-efficient representations of importance. Perhaps we need to focus more on stories (including the one that speaks for many—Slovic, Zionts, Woods, Goodman, & Jinks, 2011), quotes, images and videos that viscerally express value, and more directly engage audiences (Echeverri et al., 2020; Echeverri, Naidoo, Karp, Chan, & Zhao, 2019; Klain, Satterfield, Sinner, Ellis, & Chan, 2018). By this logic, constituents would only understand and reflect on the importance of nature by engaging directly with it, or engaging with it through another's eyes.
Although valuation and engagement steps have often been thought of as sequential, it might be that the most appropriate methods for eliciting the intangible values of nature are simultaneously most effective for communicating them. Indeed, several interviewees of a novel value-elicitation protocol thanked interviewers for the opportunity to reflect on their connections to the natural world (Gould et al., 2015; Klain et al., 2014).
In addition to deliberative approaches, widespread interventions might enable large numbers of people to effectively express values that they already hold—for example, by making it much easier to live a low-impact lifestyle. Such ‘nudge’-like approaches have manifest thus far mainly through efforts to encourage particular favoured behaviours (e.g. saving for retirement; John, Smith, & Stoker, 2009; Thaler & Sunstein, 2008), rather than a conscious effort to transform broader systems and norms towards sustainability (Abson et al., 2017). One kind of effort that might help fill this gap is to investigate how people are currently engaging with environmental programs and policies (e.g. incentive programs, Chapman, Satterfield, & Chan, 2019), and how those programs and policies might be designed to better engage those people—and the full spectrum of populations—towards broad norm change for sustainability (Chan et al., 2019; Chan, Olmsted, Bennett, Klain, & Williams, 2017; Chapman et al., 2017; Fischer & Riechers, 2019; Raymond, Weldon, Kelly, Arriaga, & Clark, 2014).
4.4 Guidance for research
Such progress in practice rests upon a major expansion of existing ES research to characterize the linked biophysical and social dynamics underpinning ES. The stark difference we observed between such prominent calls and actual practice demonstrates the utility of systematic examinations of the literature; practice does change, with time and effort. Social research on access and demand will likely remain on the sidelines until it is explicitly recognized as critical.
Many benefits might also accrue from a diverse set of approaches to ES valuation and decision-making. Research could address real-world participatory valuation and decision-making and its potential to improve decisions (Beierle, 2002), empower stakeholders and facilitate acceptance (Irvin & Stansbury, 2004), and foster deeper understanding of the values of biodiversity and ES (Gregory et al., 1993). A particular area for growth will be to investigate the non-monetary values associated with marginal, realistic changes in ES (Chan, Guerry, et al., 2012).
4.5 Guidance for policymaking
There is ample opportunity to better account for the complex values of nature in policymaking, planning and assessment. There is impetus in CBD targets and the post-2020 Biodiversity Framework, where nations and researchers are fostering BES valuation to achieve ‘communication, education and awareness raising’ regarding ‘the value of biodiversity’ and to foster reflection of those values ‘within economic systems and markets’ (Secretariat of the Convention on Biological Diversity, 2010a).
- Cost–benefit analysis (CBA), environmental impact assessment and trade-off analysis more broadly (e.g. multi-criteria decision-making) might be adapted to assess a range of BES concerns. This might include scoring some benefits using appropriately constructed alternate metrics (Keeney & Gregory, 2005; Satterfield et al., 2013) or in reference to indicators of well-being (Biedenweg et al., 2014; Hicks et al., 2016). CBA results might be displayed on one axis and other metrics on other axes (DeFries, Foley, & Asner, 2004; Polasky, Nelson, Lonsdorf, Fackler, & Starfield, 2005). Participants and experts can together identify creative alternatives, and citizen juries can assist with decision-making.
- Planning/zoning can and must account for non-monetary benefits and values (Klain & Chan, 2012; Tew, Simmons, & Sutherland, 2019). This might involve structured decision-making that is value-focused (expresses what matters to people), analytic (addresses trade-offs and alternatives) and participatory (Gregory et al., 2012). Thus values can drive the development of broad alternative scenarios, and spatial analyses of benefits and costs can illuminate particular spatial configurations.
- New environmental policy and management mechanisms like payments for ES might identify appropriate rules and incentives through negotiation analysis (Vignola et al., 2012), rather than, for example, market valuation and CBA. Such innovations would more appropriately deal with equity concerns and the ambiguity of property rights and duties (‘polluter’ vs. ‘beneficiary’), and might also foster creative solutions (Chan, Anderson, Chapman, Jespersen, & Olmsted, 2017).
In sum, policymakers should think of values as not only the outputs of valuation, but also the preferences, principles and virtues that people have about relationships involving nature (relational values, Chan et al., 2016). In particular, policymaking would ideally involve considering how policies articulate these relational values (including especially notions of equity, justice and responsibility), and how they might do so more appropriately and strategically (Chapman et al., 2019; Chapman, Satterfield, Wittman, & Chan, 2020; Kaltenborn, Linnell, & Gómez-Baggethun, 2020; Lliso, Pascual, Engel, & Mariel, 2020; Schröter et al., 2020).
5 CONCLUSION
In ES research and other fields at the intersection of people and nature (Echeverri, Karp, Naidoo, Zhao, & Chan, 2018; Gaston et al., 2019) lies the makings of BES research that is grounded in ecological and social systems and research, with a diverse suite of appropriate valuation metrics and novel involvement of stakeholders and constituents. In multiple respects, the field is maturing in response to five fundamental critiques, with more progress on some fronts than others, and more progress needed on all fronts. Such advances could position new initiatives including IPBES to greatly improve environmental decision-making and its uptake, and to foster a deep and meaningful engagement of people with the ecosystems on which they depend. More broadly, although we argued that numerical valuation is rarely appropriate as a sole description of what matters, we should also question whether the use of ecosystem service approaches is standardizing knowledge and particular forms of biodiversity governance, with negative effects (Turnhout et al., 2014), such as crowding out governance approaches more conducive to genuine partnerships with Indigenous people and local communities (Gavin et al., 2018; Sterling et al., 2017).
ACKNOWLEDGEMENTS
We gratefully acknowledge Adrian Semmelink and Caitlin Millar for their help coding the papers, and both—along with Mollie Chapman and Cameron Bullen—for helping to format various drafts of this manuscript. We also benefited from comments and suggestions from Mollie Chapman, Alejandra Echeverri, Sarah Klain, Jordan Levine, Matthew Mitchell, Robin Naidoo, Gerald Singh and Jordan Tam. In review, this paper was improved by suggestions from Rob Fish, Unai Pascual, Becky Chaplin-Kramer, Esther Turnhout and an anonymous reviewer. This work was enabled by a Canadian Natural Sciences and Engineering Research Council Discovery Grant (RGPIN-2015-05105) and two Social Sciences and Humanities Research Council Insight Grants (#435-2013-2017 and #435-2017-1071).
CONFLICT OF INTEREST
Kai M. A. Chan is a Lead Editor at People and Nature. He had no involvement with the assessment of this article. The authors have no competing financial interests.
AUTHORS' CONTRIBUTIONS
K.M.A.C. and T.S. designed the analysis, wrote the first draft of the paper and revised multiple versions of the paper; K.M.A.C. coded and analysed the data along with two research assistants, including refinements to the analysis.
Open Research
DATA AVAILABILITY STATEMENT
This manuscript does not include data beyond the 1,047 references that were reviewed, which are included below under ‘Data Sources’.