1 Capacity, Pressure, Demand, and Flow: 2 A conceptual framework for analyzing ecosystem service provision and delivery 3 Amy M. Villamagnaa*, Paul L. Angermeierb, Elena M. Bennettc 4 5 6 7 a Department of Fish and Wildlife Conservation, Virginia Tech, Blacksburg, VA 24061- 8 0321, USA *Corresponding author, amv@vt.edu; b U.S. Geological Survey, Virginia 9 Cooperative Fish and Wildlife Research Unit1, Virginia Tech, Blacksburg, VA 24061- 10 0321, USA biota@vt.edu; c Department of Natural Resource Sciences and McGill School 11 of Environment, McGill University, Montreal, Quebec, CANADA 12 elena.bennett@mcgill.ca 1 13 ABSTRACT 14 15 Ecosystem services provide an instinctive way to understand the trade-offs associated with 16 natural resource management. However, despite their apparent usefulness, several hurdles have 17 prevented ecosystem services from becoming deeply embedded in environmental decision- 18 making. Ecosystem service studies vary widely in focal services, geographic extent, and in 19 methods for defining and measuring services. Dissent among scientists on basic terminology and 20 approaches to evaluating ecosystem services create difficulties for those trying to incorporate 21 ecosystem services into decision-making. To facilitate clearer comparison among recent studies, 22 we provide a synthesis of common terminology and explain a rationale and framework for 23 distinguishing among the components of ecosystem service delivery, including: an ecosystem’s 24 capacity to produce services; ecological pressures that interfere with an ecosystem’s ability to 25 provide the service; societal demand for the service; and flow of the service to people. We 26 discuss how interpretation and measurement of these four components can differ among 27 provisioning, regulating, and cultural services. Our flexible framework treats service capacity, 28 ecological pressure, demand, and flow as separate but interactive entities to improve our ability 29 to evaluate the sustainability of service provision and to help guide management decisions. We 30 consider ecosystem service provision to be sustainable when demand is met without decreasing 31 capacity for future provision of that service or causing undesirable declines in other services. 32 When ecosystem service demand exceeds ecosystem capacity to provide services, society can 33 choose to enhance natural capacity, decrease demand and/or ecological pressure, or invest in a 34 technological substitute. Because regulating services are frequently overlooked in environmental 35 assessments, we provide a more detailed examination of regulating services and propose a novel 2 36 method for quantifying the flow of regulating services based on estimates of ecological work. 37 We anticipate that our synthesis and framework will reduce inconsistency and facilitate 38 coherence across analyses of ecosystem services, thereby increasing their utility in 39 environmental decision-making. 40 41 KEYWORDS: ecological pressure, ecosystem services, inventory and assessment, regulating 42 services, service capacity, service demand, service flow 43 3 44 45 1. Introduction Ecosystem services (ES) have great potential to influence environmental decisions 46 because they link ecosystem functions and conditions to anthropocentric interests that resonate 47 with a broad spectrum of people. ES provide new currencies, often not represented in markets, 48 for understanding the tradeoffs associated with natural resource management (Raudseppe-Hearne 49 et al., 2010; Chan et al., 2012). Because of this, efforts to assess and inventory ES have been 50 extensive (Peterson et al., 2003; MA, 2005; Tallis and Polasky, 2011; Burkhard et al., 2012); 51 however, several hurdles have prevented ES from becoming deeply embedded in environmental 52 decision-making (Daily et al., 2009; de Groot et al., 2010). However, a fundamental hurdle in 53 using ES in decision-making is the inconsistency with which scientists have conceptualized 54 delivery of ES to society (Tallis et al., 2012). Recent strides towards greater consideration of ES 55 have been made in the European Union (TEEB, 2010; European Commission, 2011; Hauck et 56 al., 2013); however, use of the ES concepts in policy-making remains limited (Fisher et al., 57 2009) and many questions persist over how ES relate to each other, how ecosystems produce 58 services, how to consistently quantify ES flows, and how changes in landscapes are likely to 59 affect future delivery of ES (Chan et al., 2006; Carpenter et al., 2009; de Groot et al., 2010; 60 Hauck et al., 2013). 61 Despite real differences, few researchers distinguish among the capacity of an ecosystem 62 to produce a service, actual production or use of that service, societal demand for that service, 63 and the natural and anthropogenic pressures on the service (Burkhard et al., 2012; Nedkov and 64 Burkhard, 2012; van Oudenhoven et al., 2012). For example, the capacity of an ecosystem to 65 generate services differs from the actual services delivered to society. A farm may produce less 66 food than it could under different management choices, or a wetland may have greater capacity 4 67 to filter nitrogen than is ultimately needed in the system. The benefits actually delivered depend 68 not only on an ecosystem’s capacity to provide services, but also on demand for these services, 69 which is, in turn, driven by biophysical setting, population size, cultural preferences, and the 70 perceived value of the service. Demand for an ES can change independently of capacity, and 71 vice versa. Thus, measurements of any one component of ES delivery cannot capture the full ES 72 dynamic from production to benefit (Fig. 1). Despite this, studies that measure only one or two 73 components of ES provision are common (Tallis et al., 2012). 74 Frameworks for conceptualizing and analyzing ES are rapidly evolving (Boyd and 75 Banzhaf, 2007; Wallace, 2007; Costanza, 2008; de Groot et al., 2010; Nedkov and Burkhard, 76 2012; van Oudenhoven et al., 2012), with little consensus on which framework or analytical 77 products are most useful for environmental decision-makers. Some recent conceptual 78 frameworks distinguish components of ES delivery (e.g. demand; Tallis et al., 2012), but 79 definitions of components and relations among them differ widely across authors. For example, 80 capacity has also been referred to as potential supply (Burkhard et al., 2012), ecosystem potential 81 (van Oudenhoven et al., 2012), stocks of nature, and ES per se (Norgaard, 2009; Layke, 2009), 82 yet the basic concept behind each term is the same. In contrast, there seems to be weaker 83 consensus on how service flows, the benefits actually delivered to people, are measured or 84 defined (Fig. 1). Terminology often differs along an ecology-economics continuum, ranging 85 from economic concepts such as benefits (Wallace, 2007; Balmford et al., 2008) or supply (Hein 86 et al., 2006) to ecological concepts like performance indicators (de Groot et al., 2010) or flow 87 (Beier et al., 2009; Layke 2009). Moreover, some studies focus on the mechanics of ES delivery 88 (Bagstad et al., 2012) while others emphasize the ecosystem properties and processes that 89 influence the service production (de Groot et al., 2010; van Oudenhoven et al., 2012). While the 5 90 breadth of approaches has surely furthered the exploration of services and has likely enhanced 91 our ability to evaluate services, the disparate terminology and subtle differences among 92 frameworks can inhibit managers and decision-makers from choosing an approach appropriate to 93 their needs. To enhance our ability to quantify, map, and ultimately make ES information more 94 accessible to decision-makers, we must acknowledge the inherent differences among ES types 95 (Table 1), the dynamic process by which ES are produced (Fig. 1), and how ES benefit people 96 (Carpenter et al., 2009; Bagstad et al., 2012; Chan et al., 2012). The key is finding a flexible and 97 adaptive approach that still allows consistency while avoiding rigid, one-size-fits-all frameworks. 98 Ecosystem services are categorized in multiple ways, with different categories being 99 amenable to different analytical approaches and providing distinctive societal benefits (MA, 100 2005). However, incorporating differences among service categories in ES assessments while 101 acknowledging their interconnectedness has been difficult. Some researchers group ES based on 102 their contribution to human well-being: services that directly benefit people (e.g. water supply) 103 are considered final or end services while many regulating and supporting services that 104 contribute to provision of final services are considered to be intermediate services. Intermediate 105 services are often excluded from economic valuations to avoid double counting (Boyd and 106 Banzhaf, 2007; Fisher and Turner, 2008; Wallace, 2008), but, in some cases, changes in 107 intermediate service provision are central to the potential societal trade-offs associated with 108 environmental management decisions. Limiting ES assessments to final services precludes 109 considering environmental and economic trade-offs, often resulting in the undervaluation of 110 services across the board (Keeler et al., 2012). Improving ES assessments requires development 111 of methods for quantifying intermediate or regulating service capacity, demand, and flow in 112 biophysical terms (Layke, 2009; Chan et al., 2012; Keeler et al., 2012). 6 113 To help advance a common language associated with ES assessments and further the 114 application of ES frameworks, we reviewed and synthesized the literature on basic components 115 of ES delivery. From this synthesis we promote a framework in which an ES delivery model 116 comprises four distinct components: capacity (i.e. the potential to provide a service), ecological 117 pressures (i.e. anthropogenic and natural stressors on ES provision), demand (i.e. the amount of 118 service required or desired by society), and flow (i.e. the actual production of a service 119 experienced by people). Second, we discuss how the interpretation and measurement of these 120 components of ES differ among provisioning, regulating, and cultural ES and how measures of 121 each can be used to evaluate sustainability. Third, we discuss a new approach to evaluate 122 capacity, ecological pressure, demand, and flow specifically for regulating services (RS), which 123 are often left out of ES assessments due to complexities associated with quantifying them. To 124 strengthen the methodology for assessing RS, we describe how to quantify RS flow using 125 estimates of ecological pressure and environmental quality. 126 127 2. Distinguishing service capacity, pressure, demand, and flow 128 129 Separately measuring the components of ES delivery adds clarity to ES analyses and can 130 enhance integration into environmental planning and development. By distinguishing among ES 131 capacity, demand, ecological pressures, and flow we can 1) assess the current and future 132 biophysical capacity of an area to produce ES, 2) evaluate the sustainability of ES use under 133 different scenarios of ES demand, pressure, and capacity, and 3) examine how ES demand and 134 ecological pressures influence biophysical capacity via feedback loops in which pressure may 7 135 exceed ecological thresholds (Carpenter et al., 2009). By comparing measures of current and 136 future capacity, ecological pressures, demand, and flow planners can evaluate whether a) the 137 needs of people can be met by existing ecosystem properties and processes, b) technological 138 substitutes are needed to supplement service production, c) ES flows will be equitable, and d) the 139 flow of services is sustainable (i.e. doesn’t degrade ES capacity). 140 141 142 2.1. Service capacity Service capacity is an ecosystem’s potential to deliver services based on biophysical 143 properties, social conditions, and ecological functions (Cairns, 1997; Chan et al., 2006; 2011; 144 Egoh et al., 2008; Daily et al., 2009; van Oudenhoven et al., 2012). ES capacity is site- and time- 145 specific, but not static; capacity responds to natural or anthropogenic changes over time and 146 space. Land use and human population changes have an acute effect on ES capacity as well as 147 ES demand, ecological pressures and ES flows (Fig. 2; also Burkhard et al., 2012; van 148 Oudenhoven et al., 2012). Capacity can be measured and mapped by integrating the natural and 149 anthropogenic factors that influence the ecological properties and functions that provide services 150 regardless of how many people use or benefit from the services in question (Table 1). 151 Provisioning service capacity is typically measured directly by ecosystem properties (e.g., 152 volume of water supply). Although more difficult to measure, cultural service capacity depends 153 on a mix of biophysical (e.g. climate, topography, presence of key species) as well as 154 anthropogenic conditions (e.g. accessibility by humans, site management actions; Villamagna et 155 al., in review). Capacity of an ecosystem to provide regulating services is also challenging to 156 measure. Regulating service capacity tends to comprise several interconnected ecosystem 8 157 processes that each rely on a suite of ecosystem properties (Fig. 3). Thus measuring regulating 158 service capacity requires extensive knowledge of ecological processes, understanding of 159 ecological and hydrologic processes, process-based models and their limitations (e.g. Revised 160 Soil Loss Equation), and/or extensive field data . 161 162 2.2. Ecological pressure Ecological pressures are biophysical influences that interfere with an ecosystem’s ability 163 to provide the service. They do so by increasing the work (i.e. effort) needed to provide the 164 service or by reducing an ecosystem’s capacity to deliver service (MA, 2005; WRI, 2012). 165 Ecological work includes the processes that generate the service and are discussed further in 166 sections 2.4 and 4.2. Pressures make it more difficult for an ecosystem to meet societal demand 167 for that service (see section 2.3) and sustained or extreme pressures can alter the future capacity 168 of an ecosystem to deliver services (Carpenter et al., 2009). Pressures on ES can be natural, like 169 periodic weather fluctuations, or anthropogenic, like increases in impervious surfaces. The 170 source of the pressure can be related to overuse, like overfishing or crowding in recreation areas 171 (Fig. 1), or it can be a by-product of ES trade-offs, like aquatic nutrient inputs from agricultural 172 production. The World Resources Institute (WRI) manages an online database on ES indicators 173 including direct and indirect drivers and service pressures (WRI, 2012). Our use of “pressure” 174 varies slightly from that of the WRI in that we include direct drivers as service pressures if they 175 are measured in the same units as the flow of the service (e.g. nutrient inputs conveyed through 176 fertilizers and changes in land cover). 177 178 2.3. Service demand 9 179 Demand for ES, the amount of service desired by society, has been measured by a variety 180 of indicators (Table 1). Human population density combined with average consumption rates is a 181 common indicator (Burkhard et al., 2012; Nedkov and Burkhard, 2012), especially for services 182 that directly impact human well-being, such as water supply or crop production. For many 183 provisioning services, demand is concisely represented by market prices. For experience-based 184 cultural services, the number of people wanting to experience the ES (e.g. visitors to a park) can 185 indicate demand. Since RS produce or maintain desirable environmental conditions, societal 186 demand should be expressed as the amount of regulation needed to meet a desired end condition 187 (e.g. the percentage reduction needed to meet numeric criteria for a pollutant). Estimating RS 188 demand is inherently challenging because it requires information about desired end conditions as 189 well as the ecological pressures or inputs needing regulation. To date, few assessments have 190 quantified RS demand biophysically. Instead, RS demand has been measured in terms of human 191 population (Burkhard et al., 2012; Nedkov and Burkhard, 2012), which is weakly related to the 192 amount of regulation actually occurring. 193 For all ES, demand -- an outcome of socio-cultural preferences -- can exceed capacity, 194 but capacity ultimately sets the limit on long-term service provision. Burkhard et al. (2012) and 195 Nedkov and Burkhard (2012) found that demand for services as measured by population density 196 of beneficiaries largely exceeds service capacity in urban areas, whereas the opposite is true in 197 less-populated rural areas. While demand for provisioning and cultural services can be met by 198 moving resources or people, demand for RS must often be met locally. Sometimes this demand 199 can be met by a technological substitute, but often the substitute meets a single demand rather 200 than the suite of demands that might be met by natural systems. For example, ecosystems with 201 high capacity to purify water provide clean drinking water, healthy aquatic habitats, and sources 10 202 of aquatic recreation (Keeler et al., 2012), but water treatment plants may only address the 203 drinking water demand. 204 2.4. Service flow 205 We consider service flow to be the service actually received by people, which can be 206 measured directly as the amount of a service delivered, or indirectly as the number of 207 beneficiaries served. Total service flow can be quantified as the service delivery per beneficiary 208 multiplied by the number of beneficiaries (Table 1). Like other components of the ES delivery 209 process, we suggest incorporating differences among ES types into measurements of service 210 flow (Table 1). For provisioning services, the conventional metric of service flow is the 211 equivalent of the end good (e.g. timber production). Cultural services are similar to provisioning 212 in that the flow of cultural service is conventionally measured in terms of the duration and 213 quality of the experience with nature. Although inherently challenging to analyze because they 214 are individualistic, difficult to aggregate, and sometimes influenced by social or moral factors 215 (Chan et al., 2012), many cultural service flows are estimated using market and non-market 216 techniques. In contrast, regulating services lack a clear end product that is tractable or 217 commonly represented in markets. Instead, environmental quality has been adopted as a 218 convenient metric of service flow and ecosystem state (Dale and Polasky, 2007; Martinez et al., 219 2009). However, simply measuring environmental quality does not necessarily convey the 220 amount of ecological work or regulation that has occurred because the amount of ecological 221 pressure on the ecosystem itself and the capacity to regulate also play a role. Environmental 222 quality is the result of multiple services, regulating and provisioning, working against ecological 223 pressures. Instead, we propose that the flow of a regulating service be measured in terms of the 11 224 ecological work required to mitigate pressures and deliver the service demanded. We further 225 discuss ecological work in measures of regulating service flow in section 4.2. 226 While service capacity is site-specific, service flow is not limited to the site of 227 production. Consider downstream benefits of clean water from upstream soil or nutrient 228 retention. Where benefits can be experienced, given natural and anthropogenic pathways, is the 229 benefit zone (Bagstad et al., 2012) and the people within the benefit zone are potential 230 beneficiaries (Hein et al., 2006; Boyd and Banzhaf, 2007; Johnston and Russell, 2011; Martin- 231 López et al., 2011). The proximity and capacity of ES sources and pathways defines the potential 232 benefit zone, but natural and anthropogenic connectivity across landscapes influences 233 spatiotemporal patterns of ES flow (Fig. 4; also Fisher et al., 2008; Bagstad et al., 2012). 234 Moreover, some services are passively delivered to beneficiaries (e.g. clean air), while others 235 require additional capital inputs on the part of the beneficiary (e.g. financial or physical capacity 236 to access recreation services). Sometimes long-distance ES flows are fundamentally 237 asymmetrical, creating social inequity in terms of the human well-being derived from ES 238 management (Carpenter et al., 2009; Tallis et al., 2012). 239 The terms and methods used to describe ES flow are especially wide-ranging relative to 240 other components of ES delivery (Fig. 1). Although service flow represents the actual delivery of 241 services and capacity represents the potential production of services, these concepts are 242 sometimes used interchangeably (Layke, 2009), which can lead to misinterpretations of ES 243 condition that affect decision-making. Service flow is an important measure of current ES 244 delivery, whereas capacity provides a measure of the potential of the system. Flow and capacity, 245 and their measures, must be consistently distinguished in order to accurately evaluate changes in 246 service delivery over time and to identify areas of potential ES production in the future. 12 247 Recognizing the differences between ES capacity and ES flow is an important step towards 248 better understanding how changes in policy and management can affect ES values accruing to 249 beneficiaries. 250 251 3. Service delivery and sustainability 252 253 The benefit of the conceptual framework we have laid out here is that distinguishing 254 among measures of ES capacity, demand, pressure, and flow enables assessment of ecological 255 sustainability and identification of key trade-offs (McDonald, 2009). Given that areas of high ES 256 capacity and flow are often spatially mis-matched and that ES demand is influenced by many 257 factors extraneous to service production (e.g. technological substitutes for ES, cultural values, 258 and behavioral norms), quantifying ES components separately is an important step towards 259 enhancing the ability of ES assessments to inform environmental decision-making. Spatially 260 explicit ES budgets, the comparison of ES demand and capacity, can identify areas where 261 technological substitutes or additional capital inputs will be needed to meet demand, and, 262 likewise, areas where greater development and ES flow can be supported. ES flow is sustainable 263 when demand is met by flow without decreasing capacity for future provision of that service 264 (Fig. 1). ES flow is not sustainable when demand cannot be met by current capacity or when 265 meeting demand causes undesirable declines in other services or in the future provision of the 266 same service. For example, the flow of water purification services from a watershed would be 267 considered unsustainable if the quality of the water produced consistently failed to meet stated 268 criteria or the only way to meet those criteria was to significantly reduce food production. 13 269 Trade-off analyses (Rodríguez et al., 2006; Daily et al., 2009; Raudsepp-Hearne et al., 270 2010) can help assess landscape-level ES sustainability. Prolonged periods of excess ecological 271 pressure or overuse may shift ecosystem functions in ways that permanently alter ES capacity 272 and delivery (Scheffer and Carpenter, 2003; MA, 2005). For example, protracted over- 273 exploitation of tree, fish, and game populations decreases stocks and regenerative capacity of 274 provisioning services (Hilborn et al., 1995; Larkin, 2000). When ES demand exceeds ecosystem 275 capacity to provide services, society generally has three choices to avoid environmental damage 276 and decreases in human well-being. First, people can enhance the system’s natural capacity to 277 provide the services demanded, for example by applying fertilizers to increase food production. 278 Second, people can recognize that supply is limited and reduce their demand appropriately. 279 Third, people can invest in technology to help avoid the outcomes of diminished services. For 280 example, we build dams, levees, and seawalls to reduce flood damage when landscapes cannot 281 adequately modulate flood magnitude and frequency. While some technological solutions 282 address a single service (e.g. water treatment plant) and fail to restore all potential benefits from 283 a non-degraded system (e.g. habitat provision), others create novel ecosystems that enhance 284 multiple services (e.g. a reservoir provides water supply, flood regulation, and recreation). In 285 contrast, management choices can negatively impact the capacity of other services (Bennett et 286 al., 2009) and a change in the flow of one service can greatly influence the ecological pressures 287 on another service (Rodríguez et al., 2006; Barbier, 2009). Given the complex interactions 288 among services, understanding ES trade-offs based on analyses that quantify capacity, demand, 289 pressure, and flow are potentially valuable contributions of ES science to environmental 290 decision-making. 291 14 292 4. Moving forward with regulating services 293 294 We suggest that distinguishing among the four components of ES delivery will provide 295 planners with better information for decision-making. To successfully integrate this multi- 296 component framework into ES assessments, we must enhance our understanding of how RS 297 function and develop stronger methods for quantifying the demand for and flow of RS. 298 Regulating services are integral to the delivery of provisioning and cultural services, yet RS are 299 declining globally (MA, 2005; Carpenter et al., 2009). Regulating services are process-driven 300 and, unfortunately, the data needed to directly measure their condition are often unavailable at 301 scales large enough to support policy-making (Layke, 2009). Below, we review how RS differ 302 from other service types and how this impacts the way we should quantify the components of 303 RS. 304 305 306 4.1. Regulating services are inherently different Regulating services are distinct in that they often exert significant influence on the 307 capacity to provide other services (de Groot et al., 2002; Boyd and Banzhaf ,2007), but direct 308 impacts of RS on human well-being can be difficult to measure (Keeler et al., 2012). Even 309 though RS provide important benefits for humans (e.g. water and air purification, drought or 310 flood control, and regulation of disease), they tend to change slowly and are thus less amenable 311 to typical scientific study. Few comprehensive and reliable ecological indicators are monitored 312 for RS (Layke, 2009), which makes their value difficult to express in biophysical or monetary 15 313 units. In addition, without market prices as indicators of their supply and demand, changes in 314 capacity may go largely unnoticed (Cairns, 1997). Often, the value of RS is not apparent until the 315 declines cause problems for other, more commonly measured services (e.g. floods or droughts 316 affecting agricultural production). Together, the lack of economic and biophysical evaluations 317 leads to the general undervaluation of RS and their absence in many planning decisions. 318 319 320 4.2. Measuring demand for and flow of regulating services Regulating services help maintain environmental quality within socially desirable 321 ranges. The amount of RS delivered will vary among ecosystems depending on ecological and 322 social pressures and capacities. Based on our four component framework, measuring regulating 323 service flow requires information about both ecological pressures on the ecosystem providing the 324 service and societal demand for the service itself. However, environmental quality (e.g. water 325 quality) is often used as an indicator of regulating service flow (Dale and Polasky, 2007; 326 Martinez et al., 2009; Shibu, 2009). Key strengths of using environmental quality as a proxy for 327 some RS is that it is readily measured, meaningful to society, and changes in quality can be 328 expressed in economic terms through market and non-market valuation (Farber et al., 2006). 329 However, environmental quality is not equal to the capacity, pressure, demand, or flow of the 330 service; instead it depends on the service capacity, relative pressure on the ecosystem and service 331 processes, and for some, the flow of other services (e.g. nitrogen regulation is affected by 332 stormwater regulation). 333 In real landscapes, high environmental quality may be the result of high capacity to 334 regulate anthropogenic stress or weak ecological pressures. High-capacity systems are capable of 16 335 greater ecological work to regulate pressures, resulting in slower or less change in environmental 336 quality (i.e. more regulation). A system with no (or very low) capacity to regulate (Fig. 5A) 337 experiences quick decline in environmental quality (y axis) with increases in ecological pressure 338 (x axis), while systems with higher capacity (Fig. 4B) can maintain acceptable environmental 339 quality under great ecological pressure. Similarly, systems with identical capacity can differ in 340 environmental quality due to differences in ecological pressures. Consider two watersheds in 341 which water quality is equal and meets societal standards (i.e. demand), but differ in contaminant 342 loading. One receives heavy nutrient loading as it flows through a mixed crop-forest landscape 343 with fertilizer inputs while the other flows through a similar landscape mosaic without fertilizer. 344 Although downstream nutrient concentrations are similar, ecological pressures differ markedly 345 and the ecological work occurring is greater in the fertilized watershed. Simply using ambient 346 water quality as a surrogate for RS flow cannot distinguish these two systems since it not only 347 ignores the relationship between ES capacity and pressure, but does not differentiate among the 348 multiple processes that affect water quality (e.g. filtration, sedimentation, volatilization, plant 349 uptake; Fig.3). 350 Instead of using environmental quality as an indicator of RS flow, we propose estimating 351 the ecological work performed as the difference between ecological pressures and measured 352 environmental quality. For example, the flow of sediment filtration services can be estimated by 353 calculating the difference between ambient sediment concentrations (e.g. total suspended solids) 354 and cumulative sediment loading throughout the watershed. Likewise, the flow of carbon 355 sequestration should be measured as the amount of carbon taken up and stored in vegetation, 356 rather than the amount of atmospheric carbon. Ecological work provides a measure of RS flow 357 that cannot be deduced from environmental quality measures alone. Ecological pressures, like 17 358 sediment loading, can be quantified in several ways, including direct field monitoring or 359 estimated by widely accepted models, like the Revised Universal Soil Loss Equation (RUSLE) 360 or Soil and Water Assessment Tool (Sahu and Gu, 2009). Since absolute values may not exist for 361 all ecological pressures or for environmental quality in all areas, relative measures can be used. 362 Where neither relative measures nor appropriate models exist, expert judgment is an alternative 363 (MA, 2005; Burkhard et al., 2012; Nedkov and Burkhard, 2012). The analytical goal is to 364 incorporate spatiotemporal variability in ecological pressures to better evaluate RS flow. By 365 evaluating ecological pressures in conjunction with environmental quality, we get a more reliable 366 estimate of RS flow which can be compared to estimates of capacity to assess the state of the 367 ecosystem, the condition of the service, and sustainability of current land practices. 368 Our approach to estimating RS flow is designed to provide information to avoid 369 ecological degradation, but may also be helpful in developing mitigation strategies to reduce 370 existing degradation. Once areas of high ecological pressure and low capacity are known, 371 degradation can be avoided by reducing pressures, increasing capacity (e.g. via restoration or 372 best management practices), or enhancing the capacity of other services that influence pressures. 373 Identifying which ecosystem properties and processes contribute to RS capacity and which land 374 use practices influence ecological pressure (Fig. 2) can help managers develop strategies to plan 375 for or mitigate changes in environmental quality. 376 5. Conclusions 377 Our approach to assessing ES, using separate measurements of ES capacity, pressure, 378 demand, and flow, is useful and innovative in that it quantifies the components of ES delivery 379 rather than merely measuring final services or environmental quality. By so doing, we can more 18 380 accurately characterize service delivery, sustainability, and ES trade-offs across space and 381 through time. Using information about all four aspects of ES, planners can more effectively 382 evaluate whether the needs of people can be met sustainably (i.e. without degradation) by 383 existing capacity or if alternative measures are needed (e.g. restoration or technological 384 substitutes). This multi-component ES approach also enables scientists to assess regulating 385 services more accurately by measuring RS flow as the regulation of ecological pressures, rather 386 than measures of environmental quality. Measuring the actual flow of services provides a metric 387 for assessing ES equity, while capacity measures inform decisions about future development and 388 management. Collectively, our multi-component framework offers a more comprehensive 389 assessment of ES delivery, sustainability, and the trade-offs associated with land use. Our 390 approach also accounts for temporal variability in all components of ES provision, especially 391 ecological pressures and societal demand, which are likely to change through time. 392 To facilitate widespread use of ES knowledge in environmental management and 393 conservation planning, we need a more flexible, coherent, and informative framework that 394 accounts for spatiotemporal differences in how ES are produced and delivered (de Groot et al., 395 2010; Chan et al., 2012). This framework should distinguish between potential service 396 production and the flow of services and be applicable across a wide range of ecosystems and 397 services (de Groot et al., 2010; Tallis et al., 2012). Our approach for analyzing ES represents 398 significant steps toward meeting these needs as this ES framework can be applied at virtually any 399 spatial resolution or extent for which ES components are measured separately. Furthermore, the 400 framework can be easily incorporated into scenario analyses (MA, 2005; Troy and Wilson, 2006) 401 to produce more objective and accurate assessments of service capacity, ecological pressure, 19 402 expected demand, and service flow which can better guide land management decisions (van 403 Oudenhoven et al., 2012). 404 405 406 Acknowledgments 407 408 This work was funded by the U.S. Geological Survey’s National Aquatic Gap Analysis 409 Program. We thank D. Beard, C. Beier, E. Frimpong, K. Limburg, B. Mogollon and anonymous 410 reviewers for their valuable input and feedback throughout the development of this and related 411 studies. The Virginia Cooperative Fish and Wildlife Research Unit is jointly sponsored by the 412 U.S. Geological Survey, Virginia Polytechnic Institute and State University, Virginia 413 Department of Game and Inland Fisheries, and Wildlife Management Institute. Use of trade 414 names or commercial products does not imply endorsement by the U.S. government. 415 416 References 417 418 Bagstad, K.J., Johnson, G.W., Voigt, B., Villa, F. 2012. Spatial dynamics of ecosystem service 419 flows: A comprehensive approach to quantifying actual services. Ecosyst. Serv. In press. 420 Barbier, E. B. 2009. Ecosystem service trade-offs. In: Mcleod, K., Leslie, H. (Eds.), Ecosystem- 421 422 423 based management for the oceans. Island Press, New York, pp. 129-144. Bennett, E. M., Peterson, G. D., and Gordon, L. J. 2009. Understanding relationships among multiple ecosystem services. Ecol. Letters 12, 1394-1404. 20 424 425 426 427 428 429 430 431 432 433 434 435 436 437 Beier, C.M., Patterson, T.M., Chapin, F.S. 2008. Ecosystem services and emergent vulnerability in managed ecosystems, a geospatial decision-support tool. Ecosystems 11, 923-938. Boyd, J., Banzhaf, S. 2007. What are ecosystem services? The need for standardized environmental accounting units. Ecol. Econ. 63, 616-626. Burkhard, B., Kroll, F. Nedkov, S., Muller, F. 2012. Mapping ecosystem service supply, demand and budgets. Ecol. Indicators 21, 17-29. Carpenter, S., Defries, R., Dietz, T., Mooney, H.A., Polasky, S., Reid, W.V., Scholes, R.J. 2006. Millennium Ecosystem Assessment: Research Needs. Science 314, 257-258. Chan K., Shaw, M.R., Cameron, D.R., Underwood, E.C., Daily, G.C .2006. Conservation planning for ecosystem services. PLoS Biol. 4, 2138-2152. Chan, K., Satterfield, T., Goldstein, J. 2012. Rethinking ecosystem services to better address and navigate cultural values. Ecol. Econ. 74, 8-18. Costanza, R., 2008. Ecosystem services: Multiple classification systems are needed. Biol. Conservation 141, 350-352. 438 Daily, G. C., Polasky, S., Goldstein, J., Kareiva, P. M., Mooney, H. A., Pejchar, L., Ricketts, T. 439 H., Salzman, J., Shallenberger, R., 2009. Ecosystem services in decision making: time to 440 deliver. Frontiers Ecol. Environ. 7, 21-28. 441 Daily, G., Kareiva, P., Polasky, S., Ricketts, T., Tallis, H., 2011. Mainstreaming natural capital 442 into decisions, In: Kareiva, P., Tallis,H., Ricketts, T., Daily, G., Polasky, S. (Eds.), 443 Natural Capital: Theory and Practice of Mapping Ecosystem Services. Oxford University 444 Press. Oxford. pp. 3-12 445 446 Dale, V. H., Polasky, S., 2007. Measures of the effects of agricultural practices on ecosystem services. Ecol. Econ. 64, 286-296. 21 447 de Groot, R.S., Alkemade, R., Braat, L., Hein, L., Willemen, L., 2010. Challenges in integrating 448 the concept of ecosystem services and values in landscape planning, management and 449 decision making. Ecol. Complexity 7, 260-272. 450 Egoh, B., Reyers, B., Rouget, M., Richardson, D. M., Le Maitre, D. C. , van Jaarsveld, A. S., 451 2008. Mapping ecosystem services for planning and management. Agric. Ecosyst. 452 Environ. 127, 135-140. 453 Eigenbrod, F., Armsworth, P. R., Anderson,B. J., Heinemeyer, A., Gillings, S., Roy, D. B., 454 Thomas C. D., Gaston, K. J., 2010. The impact of proxy-based methods on mapping the 455 distribution of ecosystem services. J. Appl. Ecol. 47, 377-385. 456 European Commission. 2011. Communication from the Commission to the European Parliament, 457 the Council, the Economic and Social Committee and the Committee of the Regions. Our 458 life Insurance, Our Natural Capital: An EU Biodiversity Strategy to 2020. COM(2011). 459 Farber, S., Costanza, R., Childers, D. L., Erickson, J., Gross, K., Grove, M., Hopkinson, C. S., 460 Kahn, J., Pincetl, S., Troy, A., Warren, P., Wilson, M., 2006. Linking ecology and 461 economics for ecosystem management: a services-based approach with illustrations from 462 LTER sites. Biosci. 56, 117-129. 463 464 465 Fisher, B., Turner, K., 2008. Ecosystem services, Classification for valuation. Biol. Conservation 141, 1167-1169. Fisher, B., Turner, K., Zylstra, M., Brouwer, R., de Groot, R., Farber, S., Ferraro, P., Green, R., 466 Hadley, D., Harlow, J., Jefferiss, P., Kirkby, C., Morling, P., Mowatt, S., Naidoo, R., 467 Paavola, J., Strassburg, B., Yu, D., Balmford, A., 2008. Ecosystem services and 468 economic theory: integration for policy-relevant research. Ecol. Appl. 18, 2050–2067. 22 469 Haines-Young, R., Potschin, M.,. 2010. The links between biodiversity, ecosystem services and 470 human well-being. In: Raffaelli, D. Frid, C., (Eds.), Ecosystem Ecology, a new synthesis. 471 BES Ecological Reviews Series, CUP. Cambridge. pp 1-31. 472 Hauck, J., C. Görg, R. Varjopuro, O. Ratamäki, K. Jax, 2013. Benefits and limitations of the 473 ecosystem services concept in environmental policy and decision making: Some 474 stakeholder perspectives, Environ. Sci. Pol. 25, 13-213. 475 476 477 478 479 480 481 Hein, L., van Koppen, K., de Groot, R. S., van Ierland, E. C., 2006. Spatial scales, stakeholders and the valuation of ecosystem services. Ecol. Econ. 57, 209-228. Hilborn, H., Walters, C. J., Ludwig, D., 1995. Sustainable exploitation of renewable resources. Annu. Rev. Ecol. Syst. 26, 45-67. Johnston, R. J., Russell, M., 2011. An operational structure for clarity in ecosystem service values. Ecol. Econ. 70, 2243-2249. Keeler, B. L., Polasky, S., Brauman, K. A. Johnson, K. A. Finlay, J. C., O’Neille, A., Kovacsf, 482 K., Dalzellg, B., 2012. Linking water quality and well-being for improved assessment 483 and valuation of ecosystem services. Proc. Natl. Acad. Sci. USA 109, 18619-18624. 484 485 486 487 Larkin, P.A. 2000. Toward Sustainable Development: An Ecological Economics Approach. CRC Press. Boca Raton, Florida. Layke, C. Measuring Nature’s Benefits: A Preliminary Roadmap for Improving Ecosystem Service Indicators. World Resources Institute, Working Paper. 36pp 488 Martin-Lopez, B., Iniesta-Arandia, I., Garcia-Llorente,M., Palomo, I., Casado-Arzuaga, I., Del 489 Amo, D. G., Gomez-Baggethun, E., Oteros-Rozas, E., Palacios-Agundez, I., Willaarts, 490 B., Gonzalez, J. A., Santos-Martin, F., Onaindia, M., Lopez-Santiago, C., Montes, C., 491 2012. Uncovering Ecosystem Service Bundles through Social Preferences. Plos One 7. 23 492 Martinez, M. L., Perez-Maqueo, O., Vazquez, G., Castillo-Campos, G., Garcia-Franco, J., 493 Mehltreter, K., Equihua, M., Landgrave, R., 2009. Effects of land use change on 494 biodiversity and ecosystem services in tropical montane cloud forests of Mexico. For. 495 Ecol. Manage. 258, 1856-1863. 496 McMichael, A, Scholes, R., Hefny, M., Pereira, E., Palm, C., Foale, S., 2005. Linking Ecosystem 497 services and human well-being, In: Millennium Ecosystem Assessment, (Ed.), 498 Ecosystems and human well-being Multiscale Assessments, Volume 4. Island Press, 499 Washington, D.C., USA. pp. 43-60. 500 501 502 MA, 2005.Ecosystems and Human Well-Being, Our Human Planet, Summary for Decision Makers. Island Press, Washington, DC. Naidoo, R., Balmford, A., Costanza, R., Fisher, B., Green, R.E., Lehner, B., Malcolm, T.R., 503 Ricketts, T.H., 2008. Global mapping of ecosystem services and conservation priorities. 504 P Natl Acad Sci USA 105, 9495-9500. 505 Nedkov, S. Burkhard, B., 2012. Flood regulating ecosystem services-Mapping supply and 506 demand, in the Etropole municipality, Bulgaria. Ecol. Indicators 21, 67-79. 507 508 509 Norgaard, R. B. 2010. Ecosystem services: From eye-opening metaphor to complexity blinder. Ecol. Econ. 69, 1219-1227. Peterson, G. D., Beard, T. D., Beisner, B. E., Bennett, E. M., Carpenter, S. R., Cumming, G. S., 510 Dent, C. L., Havlicek T. D., 2003. Assessing future ecosystem services a case study of 511 the Northern Highlands Lake District, Wisconsin. Conservation Ecol. 7, 1-24. 512 Raudsepp-Hearne, C., Peterson, G., Tengö, M., Bennett, E., Holland, T., Benessaiah, K., 513 MacDonald, G., Pfeifer, L., 2010. Untangling the Environmentalist's Paradox: Why is 514 Human Well-Being Increasing as Ecosystem Services Degrade? Biosci. 60, 576-589 24 515 Rodriguez, J. P., T. D. Beard, E. M. Bennett, G. S. Cumming, S. J. Cork, J. Agard, A. P. Dobson, 516 and G. D. Peterson. 2006. Trade-offs across space, time, and ecosystem services. Ecol. 517 Soc. 11. 518 Rounsevell, M.D.A., Dawson, T.P., Harrison, P.A., 2010. A conceptual framework to assess the 519 effects of environmental change on ecosystem services. Biodivers Conserv 19, 2823- 520 2842. 521 522 523 524 525 Sahu, M. Gu, R., 2009. Modeling the effects of riparian buffer zone and contour strips on stream water quality. Ecol. Eng. 35, 1167-1177. Scheffer, M. Carpenter, S., 2003. Catastrophic regime shifts in ecosystems: linking theory to observation. Trends Ecol. Evol. 18, 648–656. Summers, J. K., Smith, L. M., Case, J., Linthurst, R. A., 2012. A Review of the Elements of 526 Human Well-Being with an Emphasis on the Contribution of Ecosystem Services. 527 AMBIO. Royal Swedish Acad. Sci. 41, 327-340. 528 Tallis, H., Kareiva, P., Marvier M., Chang, A., 2008. An ecosystem services framework to 529 support both practical conservation and economic development. Proc. Natl. Acad. of Sci. 530 USA 105, 9457-9464. 531 Tallis, H., Polasky, S., 2011. Assessing multiple ecosystem services: an integrated tool for the 532 real world. In: Daily, G.C., Kareiva, P., Tallis, H. Ricketts, T., Polasky, S., (Eds.) 533 Natural Capital: Theory and Practice of Mapping Ecosystem Services. Oxford University 534 Press, Oxford. pp. 34-48. 535 Tallis, H., Mooney, H., Andelman, S., Balvanera, P., Cramer, W., Karp, D., Polasky, S., Reyers, 536 B., Ricketts, T., Running, S., Thonicke, K., Tietjen, B., Walz, A., 2012. A Global System 537 for Monitoring Ecosystem Service Change. Bioscience 62, 977-986. 25 538 TEEB. 2010. The Economics of Ecosystems and Biodiversity: Mainstreaming the Economics of 539 Nature: A Synthesis of the Approach, Conclusions and Recommendations of TEEB. 540 September, 15, http://www.teebweb.org. 541 542 543 Troy, A. Wilson, M. A., 2006. Mapping ecosystem services: Practical challenges and opportunities in linking GIS and value transfer. Ecol. Econ. 60, 435-449. van Oudenhoven, A., Petz, K., Alkemade, R., Hein, L., de Groot, R., 2012. Framework for 544 systematic indicator selection to assess effects of land management on ecosystem 545 services. Ecol. Indicators 21, 110-122. 546 547 548 549 Wallace, K. 2008. Ecosystem services: Multiple classifications or confusion? Biol. Conservation 141, 353-354. WRI [World Resources Institute]. 2012. Ecosystem Service Indicators Database. Accessed 10 November 2012. http://www.esindicators.org/ 550 26 551 Table 1 Components of ES Delivery ECOSYSTEM SERVICE CATEGORIES Provisioning Regulating Cultural ECOSYSTEM Biophysical Biophysical Biophysical and social SERVICE CAPACITY: capacity; feature- capacity; capacity; feature- and An ecosystem’s potential based process-based process-based to deliver services based (e.g. modeled water (e.g. modeled (e.g. model potential on biophysical and social supply) carbon to provide experience) properties and functions1 sequestration) ECOLOGICAL Events that reduce Environmental Events that reduce PRESSURES: stock and/or disturbances that stock, regenerative, or Anthropogenic and regenerative increase the amount assimilative capacity natural stressors that capacity (e.g. of ecological work of a system; affect capacity or flow of overharvest; water required to meet commonly related to benefits; often attributed impoundments) societal demands overuse to overuse or feedback (e.g. pollution, (e.g. soil compaction, from land management impervious erosion) decision to enhance other surfaces) service capacities2 ECOSYSTEM Amount of service Amount of Desired total use (if SERVICE DEMAND: desired per unit regulation needed rival service) or The amount of a service space and time to meet pre- individual use (if non- required or desired by multiplied by the determined rival) 27 society3 number of potential condition (e.g. total visitor-days users (rival service) (e.g. % nitrogen from year prior; (e.g. liters of water reduction; TMDL) individual visitation per person) rates) ECOSYSTEM Quantity harvested, Ecological work = SERVICE FLOW: The consumed, or used; ecological pressures used measured in actual production or use number of people minus units of time and/or of the service; served; number of environmental space incorporates biophysical industries served quality (same units) (e.g. total visitor-days and beneficiary (e.g. nitrogen from current year; components4 inputs-in-stream individual visitation load) rates) 1 Amount of service Cairns (1997); Chan et al. (2006); (2011); Egoh et al. (2008); Daily et al. (2009); van Oudenhoven et al. (2012). 2Beier et al. (2008); Rounsevell et al. (2010); van Oudenhoven et al. (2012). 3McDonald (2009); Nedkov and Burkhard( 2012). 4Beier et al. (2008); Layke (2009); de Groot et al. (2010); Oudenhoven et al. (2012). 552 28 553 554 Table captions 555 Table 1: Ecosystem service delivery process comprises four distinct components which differ 556 among three ecosystem service categories. A general definition and examples are provided for 557 each category-component combination. 558 Figure captions. 559 Fig. 1: The main components of the ecosystem service delivery process (boxes) are 560 interconnected such that a change in one affects the others (arrows). A wide array of terms has 561 been used interchangeably throughout the literature. For each main component (box), we cite 562 authors who have adopted that term and provide alternative terminology cited in the literature. 563 Ecological pressures (pink box) have a direct effect on the capacity of an ecosystem to provide a 564 service and can affect the flow of the services (black box). Likewise, societal demand (red box) 565 can influence ecological pressures and the flow of services from ecosystems to beneficiaries 566 (purple box) and the needs and preferences of beneficiaries influence societal demand. 567 Fig. 2: Conceptual models illustrating the effects of land use (middle) and human population 568 (right) changes on regulating service (RS) capacity, ecological pressures, societal demand for 569 regulating services, and the flow of services in a watershed in which upstream areas are largely 570 forested e.g. 80-year rotation timber production) and downstream areas are predominantly 571 agricultural and rural-suburban development left). The middle panel illustrates how a clear-cut, 572 for coal extraction or a housing development, in the upper forested area would decrease the 573 landscape’s water retention capacity, which would increase runoff and ecological pressure on 574 flood regulation downstream. Similarly, the loss of forested cover would likely decrease the 575 sediment retention capacity upstream, thereby increasing ecological pressure on sediment 29 576 regulation in the lower basin. In this case, service flow increases because of the additional work 577 the ecosystem must perform to maintain desired environmental quality. Service flow can also 578 increase because of additional beneficiaries. The right panel illustrates how an increase in human 579 population density downstream would increase the societal demand for the regulating service. 580 Increases in ecological pressure or societal demand will increase service flow, given that the 581 system has sufficient capacity to produce the service. 582 Fig. 3: Conceptual model illustrating water quality regulation. the movement of water across the 583 landscape (surface and subsurface), and the major components of the ecosystem service delivery 584 process, including capacity (green boxes), ecological pressures (pink ovals), demand (red 585 arrows), and service flow (black arrows). Beneficiaries (purple ovals) are shown as the source of 586 demand and the recipients of regulating service flow. As water is introduced to the ecosystem, by 587 means of precipitation or upland flow, a series of processes can act to regulate water quality. 588 High capacity of horizontal and vertical retention reduces the ecological pressures on surface 589 filtration and deposition. 590 Fig. 4: The flow of ecosystem services (ES) can vary greatly depending on area of service 591 production, its natural flow paths, as well as anthropogenic flow corridors. For many freshwater- 592 related services the flow path is naturally hydrologic where the capacity to produce a service 593 upstream affects the flow of benefits downstream top). Alternatively, the benefit zone can be 594 extended by anthropogenic corridors like roads, canals, or exportation bottom). 595 Fig. 5: Differences in service delivery and the effects of ecological pressure on environmental 596 quality and ecological work within ecosystems with little to no capacity A) compared to that of a 597 system with high regulating capacity B). Environmental quality is a function of regulating 30 598 capacity and ecological pressure. In systems with little to no capacity A), environmental quality 599 is quickly degraded in response to increasing ecological pressure. Systems with higher capacity 600 can maintain better environmental quality under greater ecological pressure B). Ecological 601 thresholds are determined by the ecosystem’s capacity to provide a service. Once this threshold 602 of ecological pressure is exceeded, environmental quality will degrade. The shaded polygon B) 603 illustrates the amount of ecological work performed i.e. regulating service flow), which 604 represents the difference between environmental quality and ecological pressure. 31