Mapping and Valuing Ecosystem Services as an Approach for...

THE YEAR IN ECOLOGY AND CONSERVATION BIOLOGY, 2009

Mapping and Valuing Ecosystem Servicesas an Approach for Conservation and

Natural-Resource ManagementHeather Tallisa and Stephen Polaskyb

aThe Natural Capital Project, Woods Institute for the Environment, Stanford University,Stanford, California, USA

bDepartment of Applied Economics, Department of Ecology, Evolution & Behavior,University of Minnesota, St. Paul, Minnesota, USA

Current approaches to conservation and natural-resource management often focus onsingle objectives, resulting in many unintended consequences. These outcomes oftenaffect society through unaccounted-for ecosystem services. A major challenge in mov-ing to a more ecosystem-based approach to management that would avoid such societaldamages is the creation of practical tools that bring a scientifically sound, productionfunction-based approach to natural-resource decision making. A new set of computer-based models is presented, the Integrated Valuation of Ecosystem Services and Trade-offs tool (InVEST) that has been designed to inform such decisions. Several of the keyfeatures of these models are discussed, including the ability to visualize relationshipsamong multiple ecosystem services and biodiversity, the ability to focus on ecosystemservices rather than biophysical processes, the ability to project service levels and val-ues in space, sensitivity to manager-designed scenarios, and flexibility to deal with dataand knowledge limitations. Sample outputs of InVEST are shown for two case applica-tions; the Willamette Basin in Oregon and the Amazon Basin. Future challenges relatingto the incorporation of social data, the projection of social distributional effects, andthe design of effective policy mechanisms are discussed.

Key words: production function; benefit transfer; valuation; conservation; win-win;poverty; trade-off

The Problem with CurrentDecision Making

Conservation and natural-resource manage-ment have been dominated by approaches thattend to focus on a single sector and a narrowset of objectives. These approaches often failto include a wider set of consequences of de-cision making. For example, maximizing profitfrom industrial production leads to negative im-pacts on air quality and human health. Max-imizing agricultural production often leads topoor water quality and in some cases losses of

Address for correspondence: Heather Tallis, The Natural CapitalProject, Woods Institute for the Environment, Stanford University, 371Serra Mall, Stanford, CA 94305-5020. [email protected]

productivity in downstream fisheries. Maximiz-ing biodiversity conservation can come at thecost of local jobs, food, or other sources ofincome.

Many of these consequences are the resultof management decisions that overlook a wideset of ecosystem services, by which we meanthe goods and services ecosystems produce thatare important for human well-being. Theseservices include climate regulation, carbon se-questration, soil fertility, pollination, filtrationof pollutants, provision of clean water, floodcontrol, recreation, and aesthetic and spiritualvalues (Daily 1997). In most cases, and for mostservices, there is little incentive for decisionmakers, whether they are government officials,business managers, or local landowners, to ac-count for the continued provision of ecosystem

The Year in Ecology and Conservation Biology, 2009: Ann. N.Y. Acad. Sci. 1162: 265–283 (2009).doi: 10.1111/j.1749-6632.2009.04152.x C© 2009 New York Academy of Sciences.

265

266 Annals of the New York Academy of Sciences

services in their decision making. For example,in tropical coastal ecosystems, mangroves havebeen cleared in many areas and the resultantopen land used for shrimp aquaculture. Thoseclearing the mangroves receive high marketprices for the shrimp they produce, but theydo not bear the full costs associated with theloss of habitat for coastal fisheries, loss of pro-tection from storm surges, the loss of pollutionfiltration, and other ecosystem services pro-vided by intact mangroves. A more completeaccounting shows that maintaining mangrovesgenerally provides more benefits for society(Sathirathai and Barbier 2001; Barbier 2007).Some policies exist that provide incentives for awider set of ecosystem services beyond just whatis provided for by selling marketed commodi-ties. For example, the Conservation ReserveProgram in the United States pays landown-ers to adopt conservation-oriented land uses,and similar green payments programs existin other countries. These programs, however,do not exist in all sectors and incentives forecosystem services remain lacking in manyareas.

A single-sector approach that ignores themultitude of connections among componentsof natural and social systems generally fails toprovide as high a value to society from the bun-dle of services that the system is capable of pro-ducing as would management that accountedfor the complete range of services. The connec-tions among services and the links in ecosystemprocesses are, in the long run, often critical forthe maintenance of ecosystem health, humanwell-being, and the sector of interest itself (Mil-lennium Ecosystem Assessment 2005; Guerry2005).

Reforming Decision Making withEcosystem-Based Management

Growing recognition that single-sector man-agement leads to socially and ecologicallyharmful “unintended consequences” has mo-tivated the development of ecosystem-based

management (EBM), an approach capable ofincorporating effects of management on multi-ple ecosystem services in an integrated systemsapproach (Christensen et al. 1996; National Re-search Council 1999; Food and Agriculture Or-ganization [FAO] 2003; Pikitch et al. 2004; U.S.Commission on Ocean Policy 2004). The EBMparadigm recognizes that human and ecolog-ical well-being are tightly coupled so that sus-tainability only occurs when it is pursued inboth arenas simultaneously (FAO 2003). Morespecifically, EBM:

• emphasizes the protection of ecosystemstructure, functioning, and key processes;

• is place-based in focusing on a specificecosystem and the range of activities af-fecting it;

• explicitly accounts for the interconnected-ness within systems, recognizing the im-portance of interactions between manytarget species or key services and othernontarget species and services;

• acknowledges interconnectedness amongsystems, such as between air, land, and sea;

• integrates ecological, social, economic,and institutional perspectives, recognizingtheir strong interdependences (McLeodet al. 2005).

The EBM approach is meant to provide aframework to enable managers to broaden theirperspectives and consider the multiple linkedconsequences of their decisions. What makesfor a good EBM approach, in theory, is nowwell understood (Christensen et al. 1996). Fur-ther, the Millennium Ecosystem Assessment(MA) contributed substantially to our under-standing of an EBM framework applied ata global scale (Millennium Ecosystem Assess-ment 2005). Within a year of its completion,findings from the MA were incorporated intothe Convention on Biological Diversity, theRAMSAR Convention on Wetlands, and theConvention to Combat Desertification. Thereare also a small number of examples of suc-cessful application of EBM in specific terres-trial, aquatic, estuarine, and marine settings

Tallis & Polasky: Models for Ecosystem Service Decisions 267

(Imperial 1999). These successes are evidencethat the EBM integrated approach can be prac-tical and operational.

Despite the overall success of the MA at theglobal scale, we are still left with the grandchallenge of bringing useful models and in-formation to bear at local, regional, and na-tional scales where most decisions are made. Al-though we have several cases where people haveattempted EBM at subglobal scales (see Impe-rial 1999 for a review, and MA subglobal assess-ments), there are no systematic tools that canbe applied in a general, consistent way acrossmany sites at the spatial scales and time framesrelevant to subglobal decisions. One of the mostchallenging aspects of creating such tools is theapplication of sound ecological models and un-derstanding to develop “ecological productionfunctions” that define how the structure andfunction of ecosystems ultimately leads to re-sulting levels of ecosystem services provided.This challenge is particularly acute with ecosys-tem functions and ecosystem services that actacross ecosystem boundaries (such as nutrienttransport from land to sea) and across scales(Engel et al. 2008).

A second major challenge of applying EBMto specific ecosystems centers on generating es-timates of the value of ecosystem services, a taskthat requires linking ecological models with so-cial and economic models to reveal the val-ues people hold for different ecological services(National Research Council 2005). This taskis easier for many provisioning services thatare traded in markets with observable prices.The same task is especially hard for the manyecosystem services that generate public goods,such as climate regulation or existence value ofspecies, for which there is no market price orother readily available signals of value. Overthe past 40 years or so, economists have de-veloped a number of methods and tools fornon-market valuation that can be applied toestimate the value of ecosystem services (Free-man 2003; National Research Council 2005).Whether for market or nonmarket values, ap-

propriately linking social and economic valu-ation with ecological production functions isnecessary to ensure that values reflect underly-ing ecological conditions.

Ecological Production Functionsand Economic Valuation

for Mapping and ValuingEcosystem Services

In economics, a production function speci-fies the feasible output of goods and servicesthat can be produced with a given set of in-puts (labor, machinery, natural resources, etc.).The level of technology determines the eco-nomic production function. A technologicalimprovement will mean that more goods andservices can be produced from a given set of in-puts. An ecological production function spec-ifies the feasible output of ecosystem servicesthat are provided (“produced”) by an ecosys-tem. In an ecological production function itis ecosystem processes that determine the fea-sible output of services. Changes in ecosystemconditions from natural disturbances or humanmodification alter ecological production func-tions shifting the amount of various servicesthat can be provided. In the twentieth cen-tury, human alteration of ecosystems on a largescale, such as the conversion of native ecosys-tems to monoculture agriculture, led to an in-crease in some provisioning services (e.g., foodproduction) at the expense of many regulat-ing, supporting, and cultural services (Vitouseket al. 1997; Millennium Ecosystem Assessment2005).

As with efforts to apply the concept of EBM,most applications of an ecological productionfunction modeling approach have been doneat small scales or for a single service. Thereare a growing number of such studies (e.g., El-lis and Fisher 1987; Barbier and Strand 1998;Wilson and Carpenter 1999; Barbier 2000;Kaiser and Roumasset 2002; Ricketts et al.

2004), and useful overviews and summaries


have been compiled (Barbier 2007; NationalResearch Council 2005; Pagiola et al. 2004).One of the most challenging tenets of EBM ingeneral, and production functions specifically,is to integrate modeling across multiple services.The essential next step toward informing deci-sion making is a systematic approach that com-bines the rigor of the small-scale studies withthe breadth of broad-scale assessments. Recentwork has taken strides in this vein (e.g., Boodyet al. 2005; Jackson et al. 2005; Antle and Stoor-vogel 2006; Naidoo and Ricketts 2006; Nelsonet al. 2008; Nelson et al. 2009).

There are some cases where understandingecological production functions alone is suffi-cient for the successful application of EBM.For example, many government agencies makedecisions about what activities will be allowedbased on how they affect the ability of anecosystem to provide water that passes a qualitystandard. Their decision is not based on howmuch it would cost to treat that water for con-sumption or the value of access to clean drink-ing water, but rather on the expected changein contaminant levels. In these cases, simplyknowing how ecosystem services will changein biophysical terms is very informative anduseful.

Many other decisions are more tied to eco-nomic costs and benefits, and many decisionmakers are used to analyzing policy alterna-tives in terms of the net benefits measuredin monetary terms. A concern is that ecosys-tem services that are not measured in mon-etary terms may not be given full weightin decision making or may be ignored alto-gether (Daily 1997). In these cases, it may bevery useful to combine ecological productionfunctions with economic valuation methods toestimate and report the monetary values ofecosystem services. Some ecological productionfunction approaches have been able to use ap-propriate market prices and nonmarket valua-tion methods to estimate economic value, andshow how the dollar value of services changewith different environmental conditions (e.g.,

Swallow 1994; Naidoo and Ricketts 2006; Bar-bier 2007). At present, with the possible ex-ception of Naidoo and Ricketts (2006), welack comprehensive studies that tie togethereconomic valuation methods with ecologicalproduction functions to estimate the value ofecosystem services for a significant range ofecosystem services provided by an ecosystem(National Research Council 2005).

Doing this kind of comprehensive integratedstudy of ecosystem services requires detailedecological and economic understanding anddata, which is lacking for many systems. Gener-ating new data on each system studied can beexpensive and time-consuming. Benefit trans-fer is one approach to overcoming the lack ofsystem-specific information relatively cheaplyand quickly. Benefit transfer is a method inwhich research results on the value of ecosystemservices generated in one setting is used (trans-ferred) to value ecosystem services in anothersetting. If care is taken to closely match thesettings and methods are used to ensure con-sistency with basic economic principles, benefittransfer can be a useful approach to estimatethe value of ecosystem services (Smith et al.2002). Combining ecological production func-tions that predict biophysical changes alongwith benefit transfer for economic valuationcan be used to generate estimates of the valueof ecosystem services for systems that lack eco-nomic valuation studies.

Use of benefit transfer in valuing ecosys-tem services has garnered significant interna-tional attention, especially following the studyby Costanza et al. (1997) that estimated thevalue of ecosystem services for the entire planet.Costanza et al. (1997) used estimates of thevalue of services per hectare for an ecosys-tem type from specific locations and appliedthese estimates to value all hectares of thatecosystem type across all regions. Other pa-pers have also used this approach (e.g., Ingra-ham and Foster 2008; Troy and Wilson 2006;Turner et al. 2007). This approach, however, hassignificant disadvantages that limit its social,


economic, and ecological realism. Assumingthat every hectare of a given habitat type isof equal value ignores factors of rarity, spatialconfiguration, size, quality of habitat, numberof nearby people or their social practices, andpreferences that are often crucial in determin-ing the value of services. In most cases the ap-proach provides an estimate of total economicvalue rather than estimates of value for individ-ual services. When limited to total economicvalue estimates, we cannot analyze how theprovision and value of each individual servicewill change under new conditions. If a wetlandis converted to agricultural land, how does thissubsequently affect the provision of clean drink-ing water, floods downstream, climate regula-tion, or soil fertility? Without service-specificinformation, it is impossible to design effec-tive policies or payment programs that ensurethe continued provision of ecosystem services.For these reasons, we do not believe that ap-plication of benefit transfer based on value perhectare by habitat type is a promising directionto pursue. We direct attention instead towarddevelopment of ecological production func-tions linked to economic and social valuationmethods.

A major advance that would allow theapplication of EBM to diverse decisions at di-verse scales is to develop general tools capa-ble of bringing an ecological production func-tion approach to bear in ecosystems across theglobe that incorporate a broad suite of im-portant ecosystem services. At present, thereare no systematic tools that can be appliedin a general, consistent way across many sitesat the spatial scales and time frames relevantto subglobal decisions. In other words, withexisting methods and information, we haveto start over each time we want to bring anecosystem approach to bear on subglobal de-cisions, a time-consuming process often outof step with the pace of decision making. Inthe next section, we describe the developmentof a new general tool aimed at filling thisgap.

Mind the Gap: A New Toolfor Linking Ecological

Production Functions and EconomicValuation to Map and Value

Ecosystem Services

The Natural Capital Project (www.naturalcapitalproject.org) has developed a newtool designed to address the principles of EBM,bringing together credible, useful modelsbased on ecological production functionsand economic valuation methods, with theintention of bringing biophysical and economicinformation about ecosystem services to bearon conservation and natural-resource decisionsat an appropriate scale. The tool is called theIntegrated Valuation of Ecosystem Servicesand Tradeoffs tool (InVEST). We have built inseveral key features that make this a flexible,yet scientifically grounded tool. InVEST is aset of computer-based models that

• clearly reveals relationships among multi-ple services;

• focuses on ecosystem services rather thanbiophysical processes;

• is spatially explicit;• provides output in both biophysical and

economic terms;• is scenario driven;• has a tiered approach to deal with data

availability and the state of system knowl-edge. Several of these features are dis-cussed in greater detail later in the chapter.

A Multiple Ecosystem-ServiceApproach

Managers are often forced to make trade-offs among sectors and goals. A fundamentalmathematical truth is that one cannot simul-taneously maximize many different objectives.A fundamental socioeconomic truth is thata manager cannot simultaneously maximizereturns for all sectors of society at once. As


the old saying goes, we cannot have our cakeand eat it too. Despite the ubiquity of situa-tions involving trade-offs, managers frequentlylack a set of tools to inform them about thetrade-offs they face. Often, mental assessmentsof the existence and magnitude of trade-offsare wrong and lead to decisions that result inpoor outcomes. Although management actionsthat strike a balance among goals may exist,providing acceptable outcomes in multiple di-mensions, these actions may be hard to iden-tify in a highly charged political environmentwhere arguments are based on qualitative as-sumptions and there is no way to bring evidenceto the table in a systematic and clearly under-stood way. A modeling framework that can re-veal the likely relationship among services canhelp dispel incorrect assumptions and identifymanagement options that minimize trade-offs.There is growing evidence that decision mak-ers are ready for this kind of information, ifonly they had tools to help them move forward(Box 1).

Box 1. Decision Makers Readyfor Tools of Change

Colombia Ministry of theEnvironment, Housing, and

Territorial Development

Colombia’s national environmental regulationis changing the face of environmental permittingand mitigation. The Ministry of the Environment,Housing and Territorial Development is investedin a 7-year project with several universities, re-search institutes, and nongovernment organiza-tions to create a set of tools that will allow themto take a more comprehensive approach to permit-ting and offsetting. The Ministry regulates all majoreconomic sectors (agriculture, mining, transporta-tion, etc.). They aim to assess all major projectsproposed in each sector over the next 5 years, avoidpermitting projects in areas identified as priority ar-eas for conservation of biodiversity and ecosystemservices, set levels of compensation and mitigationfor damages that cannot be avoided, and target off-setting in areas of high biodiversity and ecosystem-

service provision. InVEST will be one of severaltools used in this new permitting process.

BC Hydro

British Columbia’s major power provider hasset an ambitious environmental goal for the next20 years. They plan to achieve “no net environ-mental impact” by identifying how their infras-tructure and activities affect air, land, and waterquality, and then designing programs that avoidand reduce all impacts as much as possible. Theywill then use mitigation to offset any remaining un-avoidable changes to the environment. To achievethis goal, they need to be able to project how muchenvironmental impact different activities incur andhow effective alternative programs might be. BCHydro is currently considering ways to conductsuch assessments, including the use of InVEST.

Kamehameha Schools

With a mission to provide economic, cultural,educational, environmental, and community val-ues from its land holdings, Hawaii’s largest pri-vate landowner has a major challenge in assessingtheir land assets. How does each parcel measureup across these five diverse objectives? As many oftheir long-term land leases are coming up for re-newal, they have the opportunity to change the faceof much of Hawaii’s landscape. What set of man-agement practices would optimize returns acrossall objectives in the future? Kamehameha Schoolsis exploring different ways to reveal the level ofeach set of values provided by their lands. Theycurrently apply a rigorous scoring rubric similar tothat used to assess art, and they are exploring theuse of InVEST.

There are several reasons why current man-agement decisions lead to trade-offs amongsectors, or among ecosystem services. One rea-son is that not all services are positively corre-lated. Using data from the Willamette Basin inOregon, Nelson et al. (2008) found that tar-geting policies to provide carbon sequestra-tion, by limiting enrollment to landowners whowould grow forests on their land, was effectiveat increasing carbon storage, but not effective


for species conservation. Alternatively, target-ing policies to meet species-conservation ob-jectives, by limiting enrollment to landownerswho would restore rare habitat types (e.g., oaksavannah and prairie), was effective at increas-ing species conservation but not effective forcarbon sequestration. More generally, the Mil-lennium Ecosystem Assessment (2005) foundpervasive trade-offs between provisioning ser-vices (e.g., food and timber production) andother types of services (regulating, supportingand cultural services). A modeling frameworkthat allows assessments of biodiversity and mul-tiple ecosystem services can identify policies orgeographies that would lead to a win-win out-come, where all objectives can be increasedrelative to the status quo, and those situa-tions where outcomes necessarily would leadto trade-offs.

We address the need to reveal synergiesand trade-offs by providing models for asuite of ecosystem services in InVEST. Theservices we currently model are climate regula-tion through carbon sequestration, water sup-ply for hydropower or irrigation, erosion con-trol for infrastructure (reservoir) maintenance,water quality control for regulatory compli-ance (phosphorous loading, not contaminantsrelated to drinking water such as fecal coliformbacteria), storm peak flow mitigation, recre-ation and tourism, provision of native pollina-tion for commercial agricultural crops, agricul-ture production, timber production, nontimberforest product production, provision of culturalvalues and nonuse values. We also provide mod-els for terrestrial biodiversity as an attributeof natural systems that underpins the deliv-ery of ecosystem services. All of these modelshave been described in detail elsewhere (Nel-son et al. 2008; Nelson et al. 2009; Kareiva et al.in press).

Additional ecosystem services should beconsidered in any natural-resource decision-making process besides those just listed. Wehave focused on this subset because of its im-portance, relevance to major decisions beingmade currently, and proximity of many of these

services to markets. As the modeling effort pro-gresses, the aim is to include other importantservices that likely provide large value to soci-ety, such as forage production and the regula-tion of pests and diseases. InVEST would alsobe greatly improved by the ability to modelfreshwater biodiversity and marine ecosystemservices. The potential for inclusion of ma-rine ecosystem services is embodied in theecosystem-based management literature, butpractical tools are still lacking. We are early inour development of marine models, but someinitial steps have been taken by mangers andscientists in the Puget Sound region of Wash-ington State (Anne Guerry, personal commu-nication, 2008).

Biophysical Processesvs. Ecosystem Services:An Important Distinction

Biophysical processes are essential for theprovision of ecosystem services, but processesare not synonymous with services. Until thereis some person somewhere who is benefitingfrom a given process it is only a process and nota service. This is a critical distinction to makeas we look at what science and tools are cur-rently available for decision makers. Extensiveresearch has been applied to the modeling andmeasurement of biophysical processes, and itis tempting to simply apply those to ecosystemservice–related decisions. However, biophysicalprocesses only tell us about the supply side ofthe ecosystem service–provision equation. It isequally critical to include the demand side. Inother words, where are the people who use ser-vices, and how much do they use? To trulymodel ecosystem services, we need to adjust thesupply of ecosystem services (biophysical pro-cesses) based on the location, type, and intensityof use of each service.

Water-related services provide good exam-ples for how to think about this distinction.Consider the provision of clean water for drink-ing. Many useful biophysical models exist that


can help us predict the concentration of con-taminants in waterways (e.g., SWAT, Gassmanet al. 2007; AGNPS, Yuan et al. 2006). How-ever, the presence of clean water that could beconsumed is not a service unless there is some-one there to drink it. This does not mean thata natural system providing clean water in a re-mote area with no people does not provide anyservices. Clean water in remote areas can main-tain biodiversity existence values and may alsobe a value as clean water for drinking in thefuture. But if no one currently makes use of thewater for drinking, then there is no provision ofthat ecosystem service in that particular placeat that particular time. Ecosystem services haveto be connected to beneficiaries (demand) togain an accurate picture of what the provisionof the service provided by the landscape (sup-ply) is worth.

Consider another example with the provi-sion of pollination to agricultural crops. Manyagricultural crops require insect pollination(e.g., almonds, strawberries), but many othercrops do not (e.g., rice, corn). A patch of na-tive habitat in an agricultural landscape mayhouse healthy native bee populations, but ifthere are no agricultural fields within forag-ing distance with a crop in need of pollination,that native habitat patch does not provide pol-lination benefits. In this case, a model of nativepollinator metapopulation dynamics could giveus a very clear sense of how much pollinationservice could be supplied by patches of nativehabitat in an agricultural landscape. Until thatinformation is paired with information on theidentity of crops grown, their distribution inthe area, and the crop-specific yield benefitsof native pollination, we cannot estimate theamount or value of pollination service beingprovided.

InVEST deals with this challenge by firstmodeling the biophysical processes that consti-tute the supply side of the ecosystem-serviceequation. These models draw heavily fromexisting knowledge. For example, our water-related service models start out with thesame fundamental hydrologic processes in-

cluded in models such as SWAT (Gassman et al.

2007). Our sediment retention for infrastruc-ture maintenance model draws heavily fromthe universal soil-loss equation (USLE) (Brookset al. 1982). Our biodiversity model is based onspecies–area relationships (Connor and McCoy1979; Pereira and Daily 2006). To determinethe value of the ecosystem services provided wecombine supply-side models about provisionwith the likely level of ecosystem-service de-mand. For example, in our model of water sup-ply for irrigation, we consider how much wateris demanded by crops grown in the region that isnot met by rainfall. The value of increased wa-ter for irrigation equals the increase in the valueof crops that can be grown with an increase inavailable water input. This increase in valuefrom an increase in available irrigation watercan arise because more land can be irrigated orbecause existing cropland can receive more wa-ter, which could increase yields of crops or allowmore water-intensive but higher value crops tobe grown. To the extent that vegetation andmanagement practices in an ecosystem leadto an increase in water supply, such practicesprovide a valuable ecosystem service in placeswhere there is demand for water (for irrigationor other purposes). Consider another exam-ple. The role that vegetation and managementpractices play in keeping sediment out of wa-terways can provide services to society, includ-ing avoided infrastructure maintenance costs(as reservoirs silt in and need to be dredged)and avoided flood risk (as rivers or reservoirs siltin and lose their capacity to control or bufferfloods). To estimate the value of avoided silta-tion in reservoirs, we ask how much of the ero-sion control predicted by the USLE is upstreamof a reservoir, and consider the characteris-tics of that reservoir to estimate the level andvalue of service provided by avoiding dredgingand other maintenance costs. If there are noreservoirs, then this specific service is not beingprovided.

The links between ecosystem processes,ecosystem services, and the value of ecosystemservices for the ecosystem services currently


Figure 1. Links between the biophysical supply, ecosystem services (which combine supply and demandfor services), and the value of ecosystem services for ecosystem services currently modeled in InVEST. Land-useand management decisions impact on ecosystems and ecological processes (far left). Ecological processesgive rise to the biophysical supply of ecosystem services (first column). Ecosystem-service supply combinedwith demand generates realized ecosystem services (second column). Applying economic or social valuationmethods yield estimates of the value of ecosystem services (third column). Important links among ecologicalprocesses or ecosystem services are indicated with arrows.

incorporated into InVEST is illustrated inFigure 1. The specifics of how supply anddemand are combined in specific models forthese and other services can be found in greaterdetail elsewhere (Nelson et al. 2008; Nelson et al.

2009; Kareiva et al. in press).

Spatially Explicit

The value of ecosystem services is deter-mined both by the location of ecological pro-cesses that create the provision of services (sup-

ply) and the location of people who derive ben-efits from the services (demand). Thus, anyecosystem-service modeling effort should bespatially explicit.

We consider two key elements of space inthe application of InVEST: the role of spatialpattern and heterogeneity in the landscape incontrolling the provision of services, and thescale across which different services act. In re-gard to the first element, decision makers oftenwant to know where to invest or how to targetprograms to get the greatest return from theirinvestment. For instance, where should


protected areas be located to gain the largestbiodiversity and climate-regulation benefits?Should a new agricultural subsidy program tocontrol water quality be targeted at riparian ar-eas in headwaters or further downstream? Willa tree planting program in a poor district helpwith flood control? All of these questions have aspatial element, but many existing biophysicalprocess models are aspatial and do not allowanalysis of landscape locations that are best forinvestment. All of the models in InVEST fo-cus on identifying how much each parcel (orpixel) on the landscape controls, or contributesto, each service.

In addressing the contribution of a parcelto the provision of services, we must considerthe scale across which services are provided.Some services, such as pollination and somewater-related services are provided at a very lo-cal scale, while other services such as climateregulation are provided at a global scale. Treesfixing carbon in the Amazon forest are provid-ing a benefit to you as you read this chapter,no matter where you are in the world. Eachmodel in InVEST looks across the appropri-ate scale for the service of interest. For ex-ample, the pollination model uses the forag-ing distance of native pollinators as the scopeof assessment, while the carbon-sequestrationmodel assumes that tree growth on any parcelprovides a benefit since the global atmosphere iswell-mixed.

Considerations of scale raise two importantissues, one related to modeling and one relatedto policy. It may be more difficult to apply mod-els for local services since input data on land useand cover patterns need to be at a high enoughresolution to capture important featuresof the service. One would not learn much fromthe pollination model if native pollinators atthe site of interest foraged 1.5 km, but land useand cover data were only available at 50 km2

resolution. When modeling multiple services,the scale of data resolution needed correspondsto the most finely detailed ecosystem servicemodel.

In terms of policy, the scale and location ofthe provision of ecosystem services by naturalsystems and the scale and location of benefi-ciaries of the services in society are often dis-connected. As we have mentioned, trees fixingcarbon in the Amazon rainforest are providinga global benefit enjoyed by people far removedfrom the Amazon. For other services, such aspollination or provision of clean water, benefitsare fairly local. But even where provision andbenefits of services are local, supply and de-mand may be disconnected in space. BecauseInVEST is a spatially explicit model, it canhighlight the locations on the landscape impor-tant for the supply of services as well as the lo-cation of people who benefit from services thatshow patterns of overlap or disconnect betweensupply and demand. A prime example of a dis-connect is where upstream landowners divertwater or increase nutrient loading that harmsdownstream water users. Such spatial discon-nects between provision and benefits have im-portant implications for policy. With spatial dis-connects explicit policies will be needed to giveincentives to decision makers who control theprovision of services so that they can recognizethe benefits that their provision provides to oth-ers, a point that we discuss further in the finalsection of the chapter. Such policies can be ex-plored through the development of scenarios.

Scenario-Driven Modeling: Makinga Decision-Relevant Tool

To be effective in a decision-making or pol-icy arena, analyses should be relevant to theneeds and questions of managers and decisionmakers. To apply InVEST in such situations,we envision embedding the modeling withina stakeholder engagement process that allowsmanagers to identify the choices of interest tothem (Box 2). InVEST is designed to respondto many different kinds of scenarios derivedthrough many different types of stakeholder en-gagement processes.


Box 2. InVEST in a Stakeholder Process

InVEST is meant to be used as part of a stakeholder engagement process with decision makers who give inputat every stage of the modeling process. First, scenarios, or spatially explicit maps of how the future could look,are created based on specific choices being considered. These landscapes are then fed into a set of biophysicalmodels that project the level of ecosystem services that will be provided by that landscape. InVEST modelsseveral ecosystem services and biodiversity.

If economic valuation is of interest, there is a second set of models that can be used to derive the value of eachof the ecosystem services. Biodiversity, as an attribute of the system, is never directly valued, rather we reveal itsvalue through the multiple services it supports. Finally, the outputs of either the biophysical or economic modelscan be viewed as maps of the landscape, trade-off curves, or balance sheets. After seeing the implications oftheir scenarios, stakeholders may choose to create new scenarios and use the process iteratively.

InVEST can take input from stakeholdersto set what land-use and land-managementscenarios to analyze. InVEST is quite flexi-ble in that it can consider a wide range ofland-use and resource-management alterna-tives. Each ecosystem-service model uses land-use and land-cover (LULC) patterns as inputsto predict the biodiversity and the productionof ecosystem-services outputs. The inclusion ofboth land use and land cover means that wecan consider choices that affect the type of land

cover (urban, wetland, closed-canopy, decidu-ous forest, etc.) and choices that keep land coverthe same but alter management practices onany particular part of the landscape (change inrelease pattern from an existing dam, changein crop type planted in existing agricultural ar-eas, change in fertilizer type or amount used,change in rotation time in existing plantationforests, etc.). Most natural-resource manage-ment decisions will affect land-use or land-cover patterns, either directly or indirectly, so


sensitivity of the models to LULC patternstranslates into broad sensitivity to managementchoices.

To represent the choices that managers orother stakeholders want to consider in a waythat InVEST can recognize and assess them,we need to translate those choices into likelyfuture LULC patterns. Just as there are manydifferent types of choices that managers con-sider, there are many different methods forturning choices into LULC patterns. When de-cision makers hold full control over the area ofinterest, their own planning processes usuallyinclude the development of scenarios. For ex-ample, we conducted a very simple scenario-development workshop with a large privatelandowner in Hawaii. Kamehameha Schools,the owner of 10% of Hawaii’s land, is consider-ing future management options for a relativelylarge parcel of land (26,000 acres) and have afew alternative options in mind. Since they arethe sole owners of the entire parcel, they wereable to draw simple maps of these options in ashort workshop that we then turned into sce-narios that were assessed by InVEST for likelyecosystem-service provision.

When the area of interest is a more com-plex landscape with multiple ownership andmultiple drivers of change, more complex sce-nario generation options are available. Onecan add predictions of how landowners willreact given various institutions and incentivesthat they face. For example, we can use mod-els of landowner decision making to predicthow landowners would react to changes incrop prices or government policies to gener-ate scenarios of LULC that can be input intoInVEST models. We used this kind of scenario-generation approach to analyze the effect ofalternative policies with some of the early In-VEST models (Nelson et al. 2008). In moredemonstrative applications, users may want toexplore what is possible on a particular land-scape given the fundamental ecosystem-servicerelationships in place. In these cases, landscapeoptimization modeling can be used to createscenarios of how the landscape could look un-

der optimal conditions. We have also used thisapproach to investigate optimal land-use pat-terns using early InVEST models (Polasky et al.

2008).We have emphasized management decisions

as drivers of landscape change, but there are ob-viously other factors at work. Climate changeand human population growth will undeniablychange LULC patterns and climate conditionsin the future. Scenarios can be built to includethese drivers of change in addition to manage-ment practices. Many efforts have now down-scaled global climate models and used regionalpredictions to evolve vegetation patterns, giv-ing us maps of likely future land cover and cli-mate. Similarly, several projections of humanpopulation growth have been made, and somegroups are in the process of turning these nu-meric projections into spatially explicit humanpopulation–density estimates or urban/rural-area extents (Salvatore et al. 2005). When mod-els or maps of these drivers are available, theycan be combined with any approaches thatproject management impacts, giving scenariosthat represent all three major drivers of futurechange.

Tiered Modeling: Flexibilityfor a Data-Limited World

There is always a trade-off in modeling be-tween making a model more complex anddetailed and keeping it simple. Simple mod-els require fewer data, are often less prone toparameter estimation errors and subsequenterror propagation, and can be easier to ex-plain and understand. Complex models re-quire more information, but they often inspiregreater confidence because they more faithfullydepict the details and underlying intricacies ofprocesses. Different applications and differentusers will have specific needs for either compli-cated or simple models of ecosystem servicesand valuation. For this reason, we have de-veloped a tiered system of models in InVEST(Fig. 2).


Figure 2. A tiered approach to modeling ecosys-tem services. Given the difficulty of matching modelswith the desired level of complexity with often sparsedata, we created InVEST with different types of mod-els. Tier 1 models are simple, require few data, andare best used for planning and priority setting. Tier2 models are more complex, require more data, andare necessary for setting payment levels or informingmitigation programs. Tier 3 models add even greaterconfidence and will often be site-specific models cre-ated by other research groups.

Tier 1 models are the simplest models. Wedeveloped these models to require few data, beeasy to understand and explain to others, andyet retain sufficient credibility to guide somemanagement decisions. Their distinguishingfeature is a reliance on readily available datathat are generally accessible everywhere inthe world. Tier 1 models can draw informa-tion from the published literature, global datasets, site-specific data sources, local traditionalknowledge, or expert opinion. Because of theirsimplicity, these models will be most appropri-ately applied in general scoping and planningactivities where the purpose is to understandthe general lay of the land.

In many applications the predictions of Tier1 models may be too crude or too prone to er-rors of “averaging” or “aggregating” to meetthe needs of decision makers. This will likelybe the case when it is critical to get quantita-tive estimates of ecosystem services for decisionsrelated to making actual payments for ecosys-tem services or for setting the levels of mit-igation required. For these cases, we providemore detailed Tier 2 models. These modelsrequire more data, have more parameters, andare more time-consuming and difficult to apply.

However, Tier 2 models are generally “better”in the sense of addressing more ecological com-plexity, allowing for greater spatial and tempo-ral heterogeneity, and allowing more refined es-timates of everything from carbon productionto species richness to pollination services. In-stead of representing a world of “average trees”and the “average pollinator” or habitat typessans species lists, Tier 2 models disaggregategroups to include age structure of trees, a vari-ety of pollinator guilds or species, and monthlyprecipitation patterns.

InVEST provides a general frameworkwithin which one can mix and match Tier1 and Tier 2 models, depending on differingdata availability or need for precision amongservices. By allowing modular mixing of tiers,InVEST allows users to customize its applica-tion to specific problems. While we have devel-oped only Tier 1 and Tier 2 models, it is im-portant to realize that it is possible to also useInVEST with what we call Tier 3 models—research level, state-of-the art models (e.g., theCENTURY carbon model, Parton et al. 1994).

InVESTing in Real Decisions

InVEST has been applied in decision-making processes at several sites including theWillamette Basin (Oregon), Oahu (Hawaii), thestate of California, Puget Sound (WashingtonState), the Eastern Arc Mountains (Tanzania),the Upper Yangtze Basin (China), the AmazonBasin and Northern Andes (South America)and at a national level in Ecuador and Colom-bia. Here, we highlight some of the outputsthe InVEST models can provide in a decision-making context by presenting the findingsfrom the Amazon Basin and the WillametteBasin.

Projecting Short-Term Ecosystem-ServiceLoss in the Amazon Basin

The United Kingdom’s Natural Environ-ment Research Council (NERC), Department


for International Development (DFID) andthe Economic and Social Research Council(ESRC) launched the Ecosystem Services andPoverty Alleviation Program in 2007. The goalof the program is to promote research and ca-pacity building to achieve sustainable ecosys-tem management and well-being in develop-ing countries (Porro et al. 2008). One of theirfocal regions is the Amazon Basin, a nearly10 million km2 area with some of the globe’shighest diversity and endemism, and largest re-maining tracts of intact rainforest. This is a re-gion of relatively low rural population density,but high poverty and extremely rapid rates offorest loss and degradation (Porro et al. 2008).The agencies funding the work are interestedin what is at stake to be lost in the Basin, whois likely to lose, and how losses can be avoided.

Through a collaboration with The NatureConservancy, InVEST was used as one of sev-eral tools to predict likely future changes inbiodiversity and ecosystem services in the re-gion. We considered the year 2000 as the basecase, and the future scenario created predictshow the basin will look in 2020 after likelylevels of deforestation (projected by the Insti-tuto de Pesquisa Ambiental da Amazonia) androad development (projected by the Initiativefor Integration of Regional Infrastructure inthe South) (Porro et al. 2008). Tier 1 modelsfor habitat quality and several forest products(timber, fruits and nuts, fiber, medicinal plants,and bushmeat) were applied to the whole basinat a 1-km2 resolution. Forest products are essen-tial staples for some indigenous groups in theregion. One study estimated that people con-sume 148.2 tons of bushmeat per year in theAmazon Basin (Fa et al. 2002).

Areas along roadways and near populationcenters are projected to lose the most overall interms of forest product provision (e.g., Fig. 3;Porro et al. 2008). These areas are mostly out-side of protected areas and indigenous territo-ries. However, in montane forests, people liv-ing in indigenous territories will experience thegreatest losses in both marketed and subsistenceproducts (Fig. 4; Porro et al. 2008). The same

Figure 3. Projected change in the provision offood (fruits and nuts) for market sale between 2000and 2020. The pink area shows the extent of analy-sis. The measure of food provision is a relative indexwhere the parcel with the highest initial abundance offruit and nut species for market sale in the region re-ceived score of 1.0 in the year 2000. This map showsthe difference between the index score for each par-cel in 2020 and 2000.

Figure 4. Loss in marketed forest products byhabitat type in the Amazon Basin between 2000 and2020. The highest initial provision of forest products(in 2000) had a score of 5.0.

pattern was projected for losses of marketedproducts in savanna and varzea ecosystems.Further work is underway to apply higher-resolution poverty data to ask whether the poorin the Basin are at especially high risk of losingaccess to ecosystem services in the next 20 years(Kareiva et al. in press). For now, the messageto policymakers in Brazil and elsewhere whoinfluence decisions affecting the Amazon Basin


is the importance of stemming the loss anddegradation of habitat and ecosystem servicesin the Basin while levels are still relativelyhigh. The Amazon is a unique system given itsrelatively large remaining intact habitat and rel-atively low rural population density. The Ama-zon has lost ∼84 million hectares of native for-est in the last few decades. Other regions thatfailed to stop such drastic losses now strugglein an extremely costly and often futile effort torebuild ecosystem services in severely degradedareas with overwhelming numbers of rural poor(Porro et al. 2008).

Willamette Basin Example

The Willamette Basin in Oregon (USA) hasbeen the site of intense debates over land andwater management in recent decades. In thelate 1990s, there were conflicts between en-vironmental groups and the timber industryabout the impacts of logging on the spottedowl, a species dependent on old-growth forestand listed under the U.S. Endangered SpeciesAct. In an effort to resolve the debate PresidentClinton met with various stakeholders at thePacific Northwest Forest Conference in 1993.Clinton directed federal agencies to develop aPacific Northwest Forest Plan to protect thespotted owl while addressing other economicand social concerns. Other conflicts in the re-gion revolve around efforts to save imperiledpopulations of salmon hurt by the cumulativeimpact of timber harvests, hydroelectric powerdams, and land use that increased siltation andwater temperature. Because of these variousconflicts and efforts to find solutions, much ef-fort has gone into collecting data on species,resources, and land use, making the WillametteBasin a perfect laboratory on which to developand test models of ecosystem services.

Nelson et al. (2009) applied InVESTto model multiple ecosystem services andbiodiversity conservation in the WillametteBasin under three alternative land-use sce-narios (plan trend, development, and con-servation) that track land-use trajectories

from 1990 to 2050. The scenarios were de-veloped by the Pacific Northwest Ecosys-tem Research Consortium, composed ofrepresentatives from government agencies,nongovernment organizations, and universi-ties (see details at http://www.fsl.orst.edu/pnwerc/wrb/access.html). The scenarios wereused as input into InVEST models that thentracked changes in water quality, storm peakmitigation, erosion control, carbon sequestra-tion, biodiversity conservation, and market re-turns for agriculture, timber, and housing de-velopment. Nelson et al. (2009) found that allecosystem services (water quality, storm peakmitigation, erosion control, and carbon seques-tration) along with biodiversity conservationwere highest in the conservation scenario, whilemarket returns were higher in the plan trendand development scenarios. According to theseresults, programs designed to enhance the pro-vision of ecosystem services would likely en-hance biodiversity conservation. However, let-ting landowners pursue their own self-interestwould likely harm both ecosystem services pro-vision and biodiversity conservation. If, how-ever, payments for ecosystem services couldbe arranged so that landowners received somebenefit from the provision of ecosystem services,then the apparent conflict between landownerinterests and ecosystem services disappears.Paying for just one ecosystem service, carbonsequestration, using a price of $43 per ton ofcarbon (Tol 2005), made the conservation sce-nario the most valuable outcome rather thanthe least valuable outcome in terms of mar-ket value of goods and services produced (seeFig. 5).

The Next Frontiers

InVEST provides a means to map and valuemultiple ecosystem services that can be usedto inform conservation and natural-resourcemanagement, but it is not a panacea. Manyof the components of InVEST are relativelynew and untested. Predicting how ecosystem


Figure 5. Trade-off between biodiversity andcommodity production in the Willamette Basin in2050 under different scenarios. The Conservationscenario (gray circle) has the highest biodiversityscore, but lowest commodity value, given current mar-kets. This scenario would have the highest commod-ity and biodiversity values if the carbon market wasformalized (gray triangle). SAR is species–area rela-tionship.

services will change as ecosystems are alteredand what the value of those changes will be invarious socioeconomic systems present a num-ber of novel challenges. Increasing our work-ing understanding of these processes will bean on-going task. The InVEST tools in theircurrent form are sufficient for providing infor-mation about priority areas on landscapes forthe provision of various services, but in mostcases have not been tested sufficiently to pro-vide quantitative estimates to underpin pay-ments for ecosystem services or other regula-tory mechanisms. Developing methods to assessthe validity and reliability of model predictionsat multiple scales are needed. The models willassuredly undergo modification and updatingas experience and applications increase.

A large unmet challenge in ecosystem-basedmanagement is to understand the distributionof ecosystem service among different groupsin society, and how this distribution will likelychange as a consequence of management de-cisions. While it is important to know the to-tal amount of ecosystem services provided andtheir overall value to society, it is also impor-tant to know who benefits from the provision

of services and their social and economic sta-tus. Without information about the distributionof benefits from ecosystem services, manage-ment decisions can lead to serious unintendedconsequences for equity and well-being. Thisconcern is especially strong for managementactions that negatively affect underprivilegedparts of society. For example, in developingcountries, establishing a new national park orconservation area has in some cases resultedin removing and separating people and theirwork from ecosystems they rely on (Kareivaand Marvier 2007), leading to a decline in theirlivelihood or well-being.

Another example of distributional issues atplay occurs under “wetland banking” programsallowed under the U.S. Clean Water Act. A pol-icy objective in the United States is to have “nonet loss” of wetlands. Under wetland banking,a developer who destroys wetlands in one placecan offset this loss by buying credits from awetland bank that has gained credits by restor-ing wetlands elsewhere. One of the uninten-tional consequences of this program has beenthe reallocation of wetlands from urban to ru-ral areas, with a corresponding shift in wetlandservice provision away from the poor in urbanareas to areas with significantly lower popula-tion densities (Ruhl and Salzman 2006). Thistype of regulatory effect is often overlooked inthe decision-making process, but it is an impor-tant outcome of the functioning of the programand needs to be addressed (BenDor et al. 2008).

Whether a particular social group wins, loses,or remains unaffected by a decision is deter-mined by several factors. Of utmost impor-tance is access. Underprivileged members ofsociety will not benefit or lose from changesin ecosystem services if they cannot accessthem. Access has two critical components inthis context, physical overlap in space and le-gal rights. InVEST can show clearly whereservices will be provided on a landscape andhow their provision is likely to change in space.This ability can give insights into the spatial-overlap part of access. Rapid advances arebeing made in the mapping of social indicators


of poverty (CIESIN 2006; World Resources In-stitute 2007), and we can draw from these ap-proaches to ask where ecosystem services andthe poor overlap on the landscape. However,we currently do not have a standardized way tobring information about institutions and theirlevel of enforcement into a mapping and val-uation context. Developing ways to representand predict the interactions among ecosystemservices, people, and institutions will be criticalto the assessment of the distributional effects ofmanagement decisions.

A related challenge in ecosystem services is toalter the financial and institutional infrastruc-ture to give incentives to maintain and enhancethe provision of ecosystem services. Often thereis a spatial or temporal mismatch between thosewho control the provision of ecosystem servicesand those who benefit from the services. With-out the ability to connect the demand for ser-vices from beneficiaries with the supply fromthose who control the supply, there will beinsufficient incentive for suppliers to protectecosystems to maintain the supply of ecosys-tem services. InVEST can help to highlightthe spatial patterns of both supply and demandand illustrate and the spatial mismatch betweenthem. Addressing the mismatch and provid-ing proper incentives for provision of ecosys-tem services will require changes in policies oflocal and national governments and in inter-national agreements. Some progress on thesefronts can be seen, as with the recent emer-gence of carbon markets, international policydiscussions on reduced emissions from defor-estation in developing countries (REDD), andexpansion of programs of payments for ecosys-tem services (PES). Successfully linking the sci-ence of mapping and valuing ecosystem ser-vices with proper institutions and policies willlikely remain a major challenge for decades tocome.

Conflicts of Interest

The authors declare no conflicts of interest.

References

Antle, J.M. & J.J. Stoorvogel. 2006. Incorporating systemsdynamics and spatial heterogeneity in integrated as-sessment of agricultural production systems. Environ.

Dev. Econ. 11: 39–58.Barbier, E.B. 2007. Valuing ecosystem services as produc-

tive inputs. Econ. Policy 22: 177–229.Barbier, E.B. 2000. Valuing the environment as input:

applications to mangrove-fishery linkages. Ecol. Econ.

35: 47–61.Barbier, E.B. & I. Strand. 1998. Valuing mangrove-fishery

linkages: a case study of Campeche, Mexico. Environ.

Resour. Econ. 12: 151–166.BenDor, T., N. Brozovic & V.G. Pallathucheril. 2008.

The social impacts of wetland mitigation poli-cies in the United States. J. Plann. Lit. 22:10.1177/0885412207314011.

Boody, G., B. Vondracek, D.A. Andow, et al. 2005. Multi-functional agriculture in the United States. Bioscience

55: 27–38.Brooks, K.N., H.M. Gregersen, E.R. Berglund & M.

Tayaa. 1982. Economic evaluation of watershedprojects—An overview methodology and applica-tion. Water Res. Bull. 18: 245.

Christensen, N.L., A.M. Bartuska, J.H. Brown, et al. 1996.The Report of the Ecological Society of AmericaCommittee on the Scientific Basis for EcosystemManagement. Ecol. Appl. 6(3): 665–691.

CIESIN (Center for International Earth Science Infor-mation Network). 2006. Where the poor are: an atlas of

poverty. Columbia University, Palisades, NY. Avail-able at www.ciesin.columbia.edu/povmap/. [Ac-cessed February 13, 2009]

Connor, E.F. & E.D. McCoy. 1979. The statistics and bi-ology of the species-area relationship. Am. Nat. 113:791–833.

Costanza, R., R. d’Arge, R. de Groot, et al. 1997. Thevalue of the world’s ecosystem services and naturalcapital. Nature 387: 253–2260.

Daily, G.C. 1997. Nature’s Services. Island Press.Washington, DC.

Ecosystem Principles Advisory Panel. 1999. Ecosystem-

based fishery management: A report to Congress by the Ecosys-

tem Principles Advisory Panel. U.S. National MarineFisheries Service. Washington, DC.

Ellis, G.M. & A.C. Fisher. 1987. Valuing the environmentas an input. J. Environ. Manage. 25: 149–156.

Engel, S., S. Pagiola & S. Wunder. 2008. Designing pay-ments for environmental services in theory and prac-tice: an overview of the issues. Ecol. Econ. 65: 663–674.

Fa, J.E., C.A. Peres & J. Meewig. 2002. Bushmeatexploitation in tropical forests: an intercontinentalcomparison. Conserv. Biol. 16: 232–237.


Food and Agriculture Organization. 2003. FAO Fisheries

Department Technical Report No. 4 (Supply 2). Food andAgriculture Organization. Rome, Italy.

Freeman, A.M., III. 2003. The Measurement of Environmental

and Resource Values: Theory and Methods. Resources forthe Future. Washington, DC.

Gassman, P.W., M.R. Reyes, C.H. Green & J.G. Arnold.2007. The soil and water assessment tool: historicaldevelopment, applications and future research direc-tions. Trans. Am. Soc. Agric. Biol. Eng. 50: 1211–1250.

Guerry, A. 2005. Icarus and Daedalus: conceptual andtactical lessons for marine ecosystem-based manage-ment. Front. Ecol. Environ. 3: 202.

Imperial, M.T. 1999. Institutional analysis and ecosystem-based management: the institutional analysis and de-velopment framework. Environ. Manage. 24: 449.

Ingraham, M.W. & S.G. Foster. 2008. The value of ecosys-tem services provided by the U.S. National WildlifeRefuge System in the contiguous U.S. Ecol. Econ. 67:608–618.

Jackson, R.B., E.G. Jobbagy, R. Avissar, et al. 2005. Trad-ing water for carbon with biological sequestration.Science 310: 1944–1947.

Kaiser, B. & J. Roumasset. 2002. Valuing indirect ecosys-tem services: the case of tropical watersheds. Environ.

Dev. Econ. 7: 701–714.Kareiva, P., G. Daily, T. Ricketts, et al. 2009. The Theory

and Practice of Ecosystem Service Valuation in Conservation.Oxford University Press. Oxford. In press.

Kareiva, P. & M. Marvier. 2007. Conversation for thepeople—Pitting nature and biodiversity against peo-ple makes little sense. Many conservationists nowargue that human health and well-being should becentral to conservation efforts. Sci. Am. 297: 50–57.

McLeod, K.L., J. Lubchenco, S.R. Palumbi & A.A.Rosenberg. 2005. Scientific consensus statement on

marine ecosystem-based management. CommunicationPartnership for Science and the Sea. Available atwww.compassonline.org/pdf_files/EBM_Consensus_Statement_v12.pdf [Accessed February 13, 2009]

Millennium Ecosystem Assessment. 2005. Ecosystems and

Human Well-Being: Synthesis. Island Press. Washington.DC.

Naidoo, R. & T.H. Ricketts. 2006. Mapping the eco-nomic costs and benefits of conservation. PLoS Biol.

4: 2153–2164.National Research Council. 1999. Sustaining marine fisheries.

National Academy Press. Washington, DC.National Research Council. 2005. Valuing Ecosystem Ser-

vices: Toward Better Environmental Decision-Making. Na-tional Academy Press. Washington, DC.

Nelson, E., G. Mendoza, J. Regetz, et al. 2009. Modelingmultiple ecosystems services, biodiversity conserva-tion, commodity production and tradeoffs at land-scape scales. Front. Ecol. Environ. 7(1): 4–11.

Nelson, E., S. Polasky, D.J. Lewis, et al. 2008. Efficiencyof incentives to jointly increase carbon sequestrationand species conservation on a landscape. Proc. Nat.

Acad. Sci. 105: 9471–9476.Pagiola, S., K. von Ritter & J. Bishop. 2004. How Much

is an Ecosystem Worth? Assessing the Economic Value of

Conservation. The World Bank. Washington, DC.Parton, W.J., D.S. Schimel, D.S. Ojima & C.V. Cole. 1994.

Quantitative Modeling of Soil Forming Processes. In Special

Publication. Soil Science Society of America. Madi-son, WI.

Pereira, H.M. & G.C. Daily. 2006. Modeling biodiver-sity dynamics in countryside landscapes. Ecology 87:1877–1885.

Pikitch, E.K., C. Santora, E.A. Babcock, et al. 2004.Ecosystem-based fisheries management. Science 305:346–347.

Polasky, S., E. Nelson, J. Camm, et al. 2008. Where toput things? Spatial land management to sustain bio-diversity and economic returns. Biol. Conserv. 141:1505–1524.

Porro, R., J. Borner, A. Jarvis, et al. 2008. Challenges to Man-

aging Ecosystems Sustainably for Poverty Alleviation: Securing

Well-Being in the Andes/Amazon. Final Report, May2008. Situation Analysis prepared for the ESPA Pro-gram. Amazon Initiative Consortium, Belem, Brazil.Available at www.ecosystemsandpoverty.org/wp-content/uploads/2008/05/espa-aa-final-report-small-version_.pdf [Accessed February 13, 2009].

Ricketts, H.T., G.C. Daily, P.R. Ehrlich & C.D. Michener.2004. Economic value of tropical forest to coffeeproduction. Proc. Nat. Acad. Sci. 101: 12579–12582.

Ruhl, J.B. & J. Salzman. 2006. The effects of wetlandmitigation banking on people. Nat. Wetl. News 28: 1,9–14.

Salvatore, M., F. Pozzi, E. Ataman, et al. 2005. Mapping

global urban and rural population distributions (Environ-ment and Natural Resources Report No. 24). Foodand Agriculture Organization. Rome.

Sathirathai, S. & E.B. Barbier. 2001. Valuing mangroveconservation in southern Thailand. Contemp. Econ.

Policy 19: 109.Smith, V.K., G. Van Houtven & S.K. Pattanayak. 2002. A

benefit transfer via preference calibration: ‘pruden-tial algebra’ for policy. Land Econ. 78: 132–152.

Swallow, S.K. 1994. Renewable and nonrenewable re-source theory applied to coastal agriculture, forest,wetland and fishery linkages. Mar. Resour. Econ. 9:291–310.

Tol, R.S.J. 2005. The marginal damage costs of carbondioxide emissions: an assessment of the uncertainties.Energy Policy 33: 2064–2074.

Troy, A. & M.A. Wilson. 2006. Mapping ecosystem ser-vices: Practical challenges and opportunities in link-ing GIS and value transfer. Ecol. Econ. 60: 435–449.


Turner, W.R., K. Brandon, T.M. Brooks, et al. 2007.Global conservation of biodiversity and ecosystemservices. Bioscience 57: 868–873.

U.S. Commission on Ocean Policy. 2004. Preliminary Report

of the U.S. Commission on Ocean Policy, Governor’s Draft.U.S. Commission on Ocean Policy. Washington,DC.

Vitousek, P.M., J.D. Aber, R.W. Howarth, et al. 1997.Human alteration of the global nitrogen cycle:sources and consequences. Ecol. Appl. 7: 737–750.

Wilson, M.A. & S.R. Carpenter. 1999. Economic val-uation of freshwater ecosystem services in theUnited States: 1971–1997. Ecol. Apps. 9: 772–783.

World Resources Institute. 2007. Nature’s Benefits in Kenya:

An Atlas of Ecosystem Services and Human Well-being.World Resources Institute. Washington, DC.

Yuan, Y., R.L. Bingner & J. Boydstun. 2006. Develop-ment of TMDL watershed implementation plan us-ing annualized AGNPS. Land Use Wat. Resour. Res. 6:2.1–2.8.

Mapping and Valuing Ecosystem Services as an Approach for...

Documents

Transcript of Mapping and Valuing Ecosystem Services as an Approach for...