Modeling multiple ecosystem services, biodiversity conservation, commodity production, and tradeoffs at landscape scales

Foo

d4

wwwwww..ffrroonnttiieerrssiinneeccoollooggyy..oorrgg ©© The Ecological Society of America

Ecosystems generate a range of goods and servicesimportant for human well-being, collectively called

ecosystem services. Over the past decade, progress hasbeen made in understanding how ecosystems provide ser-vices and how service provision translates into economicvalue (Daily 1997; MA 2005; NRC 2005). Yet, it hasproven difficult to move from general pronouncementsabout the tremendous benefits nature provides to peopleto credible, quantitative estimates of ecosystem servicevalues. Spatially explicit values of services across land-scapes that might inform land-use and management deci-sions are still lacking (Balmford et al. 2002; MA 2005).

Without quantitative assessments, and some incentivesfor landowners to provide them, these services tend to beignored by those making land-use and land-managementdecisions. Currently, there are two paradigms for generat-ing ecosystem service assessments that are meant to influ-ence policy decisions. Under the first paradigm,researchers use broad-scale assessments of multiple ser-vices to extrapolate a few estimates of values, based onhabitat types, to entire regions or the entire planet (egCostanza et al. 1997; Troy and Wilson 2006; Turner et al.2007). Although simple, this “benefits transfer” approach

incorrectly assumes that every hectare of a given habitattype is of equal value – regardless of its quality, rarity, spa-tial configuration, size, proximity to population centers,or the prevailing social practices and values.Furthermore, this approach does not allow for analyses ofservice provision and changes in value under new condi-tions. For example, if a wetland is converted to agricul-tural land, how will this affect the provision of cleandrinking water, downstream flooding, climate regulation,and soil fertility? Without information on the impacts ofland-use management practices on ecosystem servicesproduction, it is impossible to design policies or paymentprograms that will provide the desired ecosystem services.

In contrast, under the second paradigm for generatingpolicy-relevant ecosystem service assessments, researcherscarefully model the production of a single service in a smallarea with an “ecological production function” – how pro-vision of that service depends on local ecological variables(eg Kaiser and Roumasset 2002; Ricketts et al. 2004).Some of these production function approaches also usemarket prices and non-market valuation methods to esti-mate the economic value of the service and how that valuechanges under different ecological conditions. Althoughthese methods are superior to the habitat assessment bene-fits transfer approach, these studies lack both the scope(number of services) and scale (geographic and temporal)to be relevant for most policy questions.

What is needed are approaches that combine the rigorof the small-scale studies with the breadth of broad-scaleassessments (see Boody et al. 2005; Jackson et al. 2005;

ECOSYSTEM SERVICES ECOSYSTEM SERVICES ECOSYSTEM SERVICES

Modeling multiple ecosystem services,biodiversity conservation, commodityproduction, and tradeoffs at landscape scalesEErriikk NNeellssoonn11**,, GGuuiilllleerrmmoo MMeennddoozzaa11,, JJaammeess RReeggeettzz22,, SStteepphheenn PPoollaasskkyy33,, HHeeaatthheerr TTaalllliiss11,, DD RRiicchhaarrdd CCaammeerroonn44,,KKaaii MMAA CChhaann55,, GGrreettcchheenn CC DDaaiillyy66,, JJoosshhuuaa GGoollddsstteeiinn77,, PPeetteerr MM KKaarreeiivvaa88,, EErriicc LLoonnssddoorrff 99,, RRoobbiinn NNaaiiddoooo1100,,TTaayylloorr HH RRiicckkeettttss1100,, aanndd MM RReebbeeccccaa SShhaaww44

Nature provides a wide range of benefits to people. There is increasing consensus about the importance of incor-porating these “ecosystem services” into resource management decisions, but quantifying the levels and values ofthese services has proven difficult. We use a spatially explicit modeling tool, Integrated Valuation of EcosystemServices and Tradeoffs (InVEST), to predict changes in ecosystem services, biodiversity conservation, and com-modity production levels. We apply InVEST to stakeholder-defined scenarios of land-use/land-cover change in theWillamette Basin, Oregon. We found that scenarios that received high scores for a variety of ecosystem servicesalso had high scores for biodiversity, suggesting there is little tradeoff between biodiversity conservation andecosystem services. Scenarios involving more development had higher commodity production values, but lowerlevels of biodiversity conservation and ecosystem services. However, including payments for carbon sequestrationalleviates this tradeoff. Quantifying ecosystem services in a spatially explicit manner, and analyzing tradeoffsbetween them, can help to make natural resource decisions more effective, efficient, and defensible.

Front Ecol Environ 2009; 7(1): 4–11, doi:10.1890/080023

1Natural Capital Project, Stanford University, Stanford, CA*([email protected]); 2National Center for Ecological Analysis andSynthesis, University of California–Santa Barbara, Santa Barbara,CA; 3Department of Applied Economics and Department of Ecology,Evolution, and Behavior, University of Minnesota, St Paul, MN; 4TheNature Conservancy, San Francisco, CA; (continued on p11)

Pollin

ation

E Nelson et al. Modeling the tradeoffs between ecosystem services and biodiversity

Antle and Stoorvogel 2006; Chan et al.2006; Naidoo and Ricketts 2006; Egoh etal. 2008; and Nelson et al. 2008 for someinitial attempts). Here, we present resultsfrom the application of a new, spatiallyexplicit modeling tool, based on ecologi-cal production functions and economicvaluation methods, called IntegratedValuation of Ecosystem Services andTradeoffs (InVEST). We apply InVESTto three plausible land-use/land-cover(LU/LC) change scenarios in theWillamette Basin, Oregon (Figure 1). Weshow how these different scenarios affecthydrological services (water quality andstorm peak mitigation), soil conserva-tion, carbon sequestration, biodiversityconservation, and the value of severalmarketed commodities (agricultural cropproducts, timber harvest, and rural–resi-dential housing). We also explore thespatial patterns of ecosystem service pro-vision across the landscape under thesethree scenarios, highlighting synergiesand tradeoffs between multiple ecosystemservices, biodiversity conservation, andmarket returns to landowners.

�Methods

InVEST consists of a suite of modelsthat use LU/LC patterns to estimate lev-els and economic values of ecosystemservices, biodiversity conservation, andthe market value of commodities pro-vided by the landscape. Examples ofecosystem services and commodity pro-duction that InVEST can model includewater quality, water provision for irriga-tion and hydropower, storm peak mitiga-tion, soil conservation, carbon seques-tration, pollination, cultural andspiritual values, recreation and tourism,timber and non-timber forest products,agricultural products, and residentialproperty values. InVEST can be run at different levels ofcomplexity, making it sensitive to data availability and anunderstanding of system dynamics. Results can bereported in either biophysical or monetary terms, depend-ing on the needs of decision makers and the availabilityof data. However, biodiversity conservation results arereported in biophysical terms only.

In this paper, we use a subset of the simpler InVESTmodels and focus largely on reporting ecosystem ser-vices in biophysical terms. We run InVEST across threedifferent projections of LU/LC change in theWillamette Basin. Below, we briefly describe the major

features and data inputs for the ecosystem services, bio-diversity conservation, and commodity productionvalue models. For greater detail, please refer to thispaper’s appendix, at www.naturalcapitalproject.org/pubs/NelsonetalFrontiersAppendix.pdf.

� Land-use/land-cover projections in the WillametteBasin

The base map in this study was a 1990 LU/LC map for theWillamette Basin (29 728 km2) developed by the PacificNorthwest Ecosystem Research Consortium, a multi-stake-

FFiigguurree 11.. Maps of the Willamette Basin and the land-use/land-cover (LU/LC)patterns for 1990 and under the three LU/LC change scenarios for 2050. A 500-hahexagon is the spatial unit used in the LU/LC pattern maps. Each hexagon cancontain more than one LU/LC. However, for illustrative purposes, we only show ahexagon’s most dominant LU/LC. The light brown lines delineate the three ecoregionsthat intersect the Basin (Omernik 1987); from west to east, the ecoregions are theCoast Range, the Willamette Valley, and the Cascades Range. The Coast Range is alow mountain range (122–762 m) that runs the entire Oregon coast, with three of thetallest conifer species in the world supported by high annual rainfall and intensive fogduring the summer. The Willamette Valley incorporates terraces and the floodplain ofthe Willamette River system, and most of the agricultural and urban land use in theBasin. The Cascades Range is large, steep, and high (up to 3424 m).

2050 Plan Trend

Orchard/vineyard Young conifer

Grass seed Other forestPasture/hayfield Old conifer/other naturalRow crops Dense development/Rural–residential bare ground

2050 Development1990

165 km

OregonPortland

Salem

AlbanyCorvalis

Eugene

270

km

Cas

cad

es M

oun

tain

Ran

ge

eco

reg

ion

Val

ley

floo

r ec

ore

gio

n

Co

ast

Mo

unta

in R

ang

e ec

ore

gio

n

5

© The Ecological Society of America wwwwww..ffrroonnttiieerrssiinneeccoollooggyy..oorrgg

2050 Conservation

Wa

ter

filt

rati

onModeling the tradeoffs between ecosystem services and biodiversity E Nelson et al.

6

wwwwww..ffrroonnttiieerrssiinneeccoollooggyy..oorrgg ©© The Ecological Society of America

holder alliance between government agencies, non-govern-mental organizations, and universities (Hulse et al. 2002;US EPA 2002; Baker et al. 2004; www.fsl.orst.edu/pnwerc/wrb/access.html). This alliance facilitated the cre-ation of three stakeholder-defined scenarios of LU/LCchange, from 1990 to 2050 (Baker et al. 2004). Each sce-nario includes a set of spatially explicit raster grid LU/LCmaps (30 m x 30 m grid cells) of the Basin at 10-yearintervals, from 2000 to 2050 (Figures 1 and 2). The threescenarios are:

• Plan Trend: “the expected future landscape, should cur-rent policies be implemented as written and recenttrends continue” (US EPA 2002).

• Development: “a loosening of current policies, to allowfreer rein to market forces across all components of thelandscape, but still within the range of what stakehold-ers considered plausible” (US EPA 2002).

• Conservation: placed greater emphasis on ecosystem pro-tection and restoration; however, as with theDevelopment scenario, the model still reflects “a plausiblebalance among ecological, social, and economic consid-erations, as defined by stakeholders” (US EPA 2002).

The three scenarios assume that human population inthe Willamette Basin will increase from 2.0 million in1990 to 3.9 million people in 2050 (Hulse et al. 2002).

�Models

Ecosystem services, biodiversity conservation, andcommodity production values are a function of landcharacteristics and the LU/LC pattern. Models wererun using the 30 m x 30 m resolution data. For report-ing and display purposes, we aggregated results to 500-ha hexagon units (model results are given in Figures 3,4, and 5). In general, InVEST can be run on spatialunits of any resolution.

Water service models: water qualityand storm peak mitigation

In this application, we used the dis-charge of dissolved phosphorus into thelocal watershed to measure water pollu-tion. Although this single measureignores many other sources of water pol-lution, it provides a proxy for non-point-source pollution. Slope, soildepth, and surface permeability wereused to define potential runoff by loca-tion. Areas with a greater potentialrunoff, less downhill natural vegetationfor filtering, greater hydraulic connec-tivity to water bodies, and LU/LC asso-ciated with the export of phosphorous(ie agricultural land) have greater ratesof phosphorus discharge. Areas thathave the highest water quality scores

export relatively little phosphorous to waterways. The storm peak mitigation model highlights the areas

on the landscape that contribute most to potential flood-ing after a uniform rainfall event. The model estimatesthe volume and timing of water flow from an area to itscatchment’s outlet on the Willamette River. Both thevolume and timing of water flow across the landscape areaffected by water retention on the land. Water retentionin an area is greater if its LU/LC has a rougher surface orprovides opportunities for water infiltration. In general,as water retention rates increase in a catchment, themore that flood risk at the catchment’s outlet decreases.Areas in a catchment that contribute less to the stormpeak at the catchment’s outlet – because they export littlewater, deliver water at off-peak times, or both – have thehighest storm peak mitigation scores.

Soil conservation

The soil conservation model uses the Universal Soil LossEquation (Wischmeier and Smith 1978) to predict theaverage annual rate of soil erosion in a particular area (usu-ally reported in tons acre–1 yr–1; in Figure 4 we map the rel-ative change in erosion rates across space and time). Therate of soil erosion is a function of the area’s LU/LC, soiltype, rainfall intensity, and topography. For this study, weassumed that rainfall intensity was homogenous across theentire landscape. In general, the model predicts greater soillosses in agricultural areas and sites with steeper slopes, andlower soil losses in forested and paved areas. regions withlower potential soil losses received higher scores.

Carbon sequestration

We tracked the carbon stored in above- and belowgroundbiomass, soil, and harvested wood products (HWP) usingstandard carbon accounting methods (Adams et al. 1999;Plantinga et al. 1999; Feng 2005; Lubowski et al. 2006;

FFiigguurree 22.. Distribution of land area under each LU/LC category for 1990 and 2050under the three LU/LC change scenarios (see Figure 1).

Orc

hard

/vin

eyar

d

Gra

ss s

eed

Pas

ture

/hay

field

Ro

w c

rop

s

Rur

al–r

esid

entia

l

Youn

g c

oni

fer

Oth

er f

ore

st

Old

co

nife

r/o

ther

nat

ural

Den

se d

evel

op

men

t/b

are

gro

und

Tho

usan

ds

of

hect

ares

1990 LU/LC Plan Trend Development Conservation

1200

900

600

300

Recrea

tionE Nelson et al. Modeling the tradeoffs between ecosystem services and biodiversity

Smith et al. 2006; Kirby and Potvin 2007; Nelson et al.2008). To determine how much carbon was stored in anarea, we estimated above- and belowground biomass andsoil carbon pools as a function of the area’s distribution ofpresent and historic LU/LC and biomass age. We alsoestimated how much timber was removed from the areain previous time periods to determine the carbon thatremained stored in HWP. The amount of carbonsequestered in an area across a particular time period isdetermined by subtracting the carbon stored in the areaat the beginning of the time period from that stored inthe area at the end of the time period.

In this study, we also estimated the social value of car-bon sequestration (all sequestration, not just the portionof sequestration that would be eligible for trading in a car-bon offset market; see Watson et al. 2000). We assumed asocial value of $43 per Mg of carbon, which is the meanvalue of the social cost of carbon from Tol’s (2005) surveyof peer-reviewed literature. The social cost of carbon isequal to the marginal damage associated with the releaseof an additional metric ton of carbon into the atmosphere– or, in this case, the monetary benefit of an additionalsequestered metric ton. Payments beyond 1990 were dis-counted to reflect the decrease in monetary value overtime. We used the US Office of Management and Budgetrecommended rate of 7% per annum as the discount rate(US OMB 1992). In addition, we adopted the conserva-tive assumption that the social value of carbon sequestra-tion will decline over time (ie in the future, the socialcost of carbon will decline at a rate of 5% per annum).Whether the social value of carbon will decrease,increase, or remain constant in the future is uncertain.

Biodiversity conservation

We used a countryside species–area relationship (SAR;Sala et al. 2005; Pereira and Daily 2006) to determine thecapacity of each LU/LC map to support a suite of 24 ver-tebrate species that previous analysis found to be sensitiveto LU/LC change in the Willamette Basin (Polasky et al.2008). The score for each species on a given LU/LC mapdepended on the amount of actual and potential habitatarea provided for a species. Actual habitat area for aspecies was equal to the amount of LU/LC in the species’geographic range that was compatible with its breedingand feeding requirements. Potential habitat area wasgiven by a species’ total mapped geographic range withinthe Willamette Basin. The countryside SAR score foreach species was equal to the ratio of actual habitat areato potential habitat area raised to the power z (0 < z < 1).Lower z values imply less of a penalty for losing small por-

tions of habitat and large penalties for losing the last fewunits of habitat. In this application, we used a z value of0.25 for each species. We averaged across the countrysideSAR scores of each species to calculate an aggregate scorefor each scenario.

In order to allocate biodiversity scores spatially acrossthe landscape, we calculated a second biodiversity metric

FFiigguurree 33.. Trends in normalized landscape-level ecosystem ser-vices, biodiversity conservation, and market value of commodityproduction for the three LU/LC change scenarios. All scores arenormalized by their 1990 levels. Carbon sequestration andcommodity production values are not discounted in this figure.

1.25

1.15

1.05

0.95

0.85

0.751.25

1.15

1.05

0.95

0.85

0.751.25

1.15

1.05

0.95

0.85

0.751.25

1.15

1.05

0.95

0.85

0.75

1.25

1.15

1.05

0.95

0.85

0.75

1.25

1.15

1.05

0.95

0.85

0.751990 2000 2010 2020 2030 2040 2050

Years

Plan Trend Development Conservation

Wat

er q

ualit

y

Rel

ativ

e re

duct

ion

inan

nual

dis

char

ge o

fdi

ssol

ved

phos

phor

ous

Uni

tless

So

ilco

nser

vatio

n

Red

uctio

n in

ave

rage

annu

al ra

te o

f soi

ler

osio

n in

sho

rt to

ns

Sto

rm p

eak

man

agem

ent

Uni

tless

Car

bo

nse

que

stra

tion

Met

ric to

ns

Bio

div

ersi

tyco

nser

vatio

n

Cou

ntry

side

spe

cies

–ar

ea re

latio

nshi

p (S

AR

)

Mar

ket

valu

e o

fco

mm

od

ityp

rod

uctio

n

Con

stan

t yea

rU

S$20

00

7

© The Ecological Society of America wwwwww..ffrroonnttiieerrssiinneeccoollooggyy..oorrgg

Wa

ve

att

enu

ati

onModeling the tradeoffs between ecosystem services and biodiversity E Nelson et al.

8

wwwwww..ffrroonnttiieerrssiinneeccoollooggyy..oorrgg ©© The Ecological Society of America

that could be applied to distinct areas on the landscape(countryside SAR applies only at the landscape level).This metric estimated an area’s relative contribution tothe sustainability of each species. The marginal biodiver-sity value (MBV) of an area measures the value of habitatin the area for all species under consideration, relative tothe composite value of habitat available to all speciesacross the whole landscape. We then calculated the rela-tive MBV (the RMBV), a modified version of MBV, tomeasure the change in an area’s value over time, andreported the ratio of this number to the area’s MBV valueon the 1990 LU/LC map.

Commodity production value

In addition to the ecosystem services and biodiversityconservation, we also estimated the market value of com-

modities provided by an area. Themarket value is equal to the aggregatenet present value of commodities(agricultural crops, timber, andrural–residential housing) producedin the area. The market value modelswere taken from Polasky et al. (2008).We lacked a model to value urbanland use. To make fairer comparisonsacross scenarios, we excluded thevalue of commodities produced onland that was developed for urbanland uses in any scenario.

The net present value of agricul-tural crop production in an areadepends on crop type, soil productiv-ity, irrigation, crop prices, and pro-duction costs. The net present valueof timber production depends on themix of tree species, soil productivity,forestry rotation time, timber price,and harvest cost. We used price andproduction cost estimates from 2000for both agriculture and forestry. Thenet present value of housing in anarea is a function of its proximity tourban areas (Kline et al. 2001) andthe area’s county, mean elevation,slope, lot size, and existing buildingdensity. We assumed that the annualper-hectare net return for rural resi-dential housing in the Basindecreased by 0.75% for each 1%increase in rural residential land usein the Basin (ie elasticity of demandfor rural residential housing is–0.75%) and that the value of ruralresidential land-use increased 2% perannum. We used a discount rate of7% per annum to compute the net

present values of commodity production across time.

� Results

Of the three LU/LC change scenarios, the Conservationscenario produced the largest gains (or the smallest losses)in ecosystem services and biodiversity conservation (Figure3). Under the Conservation scenario, carbon sequestration,water quality, and soil conservation scores increased sub-stantially. Carbon sequestration also increased under thePlan Trend and Development scenarios, although lesssteeply, mainly because of sequestration losses in the lowerelevations of the Cascade Mountains as a result of rural res-idential development and timber production (Figure 4).Water quality and potential soil conservation changedonly slightly in the Plan Trend and Development scenarios,but improved under the Conservation scenario, because of

FFiigguurree 44.. Maps of change in ecosystem services, biodiversity conservation, and marketvalue of commodity production from 1990 to 2050 for the three LU/LC changescenarios. Carbon sequestration and commodity production values are not discounted.

Water qualityRelative reduction

in annual dischargeof dissolvedphosphorousper hexagon

SoilconservationReduction in avgannual rate of soil

erosion in short tonsper hexagon

Storm peakmanagement

Unitless

Carbonsequestration

Metric tonsper hexagon

Biodiversityconservation

2050 relative marginalbiodiversity valueratio of a hexagon

Market valueof commodity

productionConstant year 2000US$ per hexagon

Pla

n Tr

end

Dev

elo

pm

ent

Co

nser

vatio

nD

eclin

e in

serv

ice/

attr

ibut

epr

oduc

tion

Impr

ovem

ent i

nse

rvic

e/at

trib

ute

prod

uctio

n

Greatestdecline

Least declineNo changeLeastimprovement

Greatestimprovement

Greatestdecline

Least declineNo changeLeastimprovement

Greatestimprovement

Greatestdecline

Least declineNo changeLeastimprovement

Greatestimprovement

Greatestdecline

Least declineNo changeLeastimprovement

Greatestimprovement

Greatestdecline

Least declineNo changeLeastimprovement

Greatestimprovement

Greatestdecline

Least declineNo changeLeastimprovement

Greatestimprovement

Food

E Nelson et al. Modeling the tradeoffs between ecosystem services and biodiversity

replacement of agricultural land withforests, prairies, and other land uses onthe Basin floor (Figure 1).

Storm peak mitigation scores declinedslightly under all three scenarios (Figure3), but the Conservation scenario exhib-ited the smallest reduction. Reductionsin hexagon storm peak managementscores (indicative of increased flood riskat the hexagon’s catchment outlet onthe Willamette River, all else beingequal) were greatest under theDevelopment scenario, which had thelargest increase in impervious surfacearea of any of the scenarios. Outside ofdeveloping areas on the Basin floor,storm peak management scores werelargely unchanged (Figure 4).

Landscape-level biodiversity conser-vation scores also showed only smallchanges through time under each of thethree scenarios. The 24-species coun-tryside SAR showed a small increaseunder the Conservation scenario, butdeclined slightly under both the PlanTrend and Development scenarios(Figure 3). The areas immediately sur-rounding urban areas saw the greatest biodiversity losses, asmeasured by RMBV ratios. Some of the greatest increases inRMBV ratios occurred in the Coast Mountain Range andtoward the southern end of the valley floor (Figure 4).Despite widespread declines in RMBV ratios across thelandscape in the Plan Trend and Development scenarios, thedeclines were not great enough to greatly reduce the 24-species countryside SAR score under either scenario. Theuse of a higher z value in the countryside SAR calculationwould result in greater biodiversity conservation scoredeclines in the Plan Trend and Development scenarios.

The aggregate market value of commodities produced onthe landscape was the only measure where the Conservationscenario did not outperform the Plan Trend andDevelopment scenarios (Figure 3). The market value ofcommodity production increased in many areas under thePlan Trend and Development scenarios, as a result of bothincreased residential development and more intensive tim-ber harvesting (Baker et al. 2004; Figure 4). Although themarket value of commodity production declined in amajority of areas under the Conservation scenario (4143 outof 6214 hexagons), aggregate market value of commodityproduction summed over the whole region increased,because the high value of rural residential developmentnear cities more than compensated for losses elsewhere.

Given the emerging interest in carbon markets, we cal-culated the aggregate market value of carbon sequestra-tion under the three scenarios. We assumed the marketvalue of carbon sequestration was equal to its social valueof $43 Mg–1 of sequestered carbon (this may be an under-

estimate, since carbon prices on the European carbonmarket were $133–162 Mg–1 of sequestered carbon, at anexchange rate of US$1.58–€1 in July 2008, and$88–112 Mg–1 of sequestered carbon, at an exchange rateof US$1.33–€1 in October 2008). The total present valueof carbon sequestration on the landscape from 1990 to2050 was $1.6 billion, $0.9 billion, and $0.8 billion,under the Conservation, Plan Trend, and Development sce-narios, respectively (and $1.5 billion, $0.8 billion and$0.7 billion, respectively, if we only applied a marketvalue to 50% of HWP carbon sequestration on the land-scape). If these carbon sequestration values are added toaggregate market value of commodities for each scenario,then Conservation generates more monetary value thanPlan Trend and Development ($16.38 versus $16.16 or$16.07 billion [Figure 5]; or $16.27 versus $16.05 or$15.96 billion, if we only applied a market value to 50%of HWP carbon sequestration on the landscape). If pay-ments were made for the other ecosystem services, thevalue of the Conservation scenario would increase evenfurther relative to the other two scenarios.

� Discussion

We applied the InVEST model to predict the provision ofecosystem services, biodiversity conservation, and themarket value of commodities across space and time forthree contrasting scenarios of future LU/LC change. Thisresearch contributes to an emerging literature thatattempts to quantify the value of multiple ecosystem ser-

FFiigguurree 55.. Tradeoffs between market values of commodity production and biodiversityconservation on the landscape between 1990 and 2050, excluding (circles) andincluding (triangles) the market value of carbon sequestration (we assume that thesocial value of carbon is equal to the market value of sequestered carbon). The x axismeasures the total discounted value of commodities, whereas the y axis measures thebiodiversity (ie countryside SAR) score for 2050.

Agricultural, timber,and rural–residentialcommodities

All commoditiesand carbonsequestration

Conservation

DevelopmentPlan Trend

0.60

0.59

0.58

0.57

0.56

0.5514.5 15.0 15.5 16.0 16.5

Net present market value of 1990–2050

commodity production (billions of US$)

Co

untr

ysid

e S

AR

sco

re in

205

0

9

© The Ecological Society of America wwwwww..ffrroonnttiieerrssiinneeccoollooggyy..oorrgg

Rec

rea

tion

Modeling the tradeoffs between ecosystem services and biodiversity E Nelson et al.

10

wwwwww..ffrroonnttiieerrssiinneeccoollooggyy..oorrgg ©© The Ecological Society of America

vices at a broad scale (geographic and temporal) by wayof ecological production functions and economic valua-tion methods. Analyzing how ecosystem service provisionand value change under alternative realistic scenarios dis-tinguishes our approach from the well known maps of“total” value (ie benefits transfer) that can be producedfor a site (Troy and Wilson 2006), a state (Costanza et al.2006), or the world (Costanza et al. 1997).

Combining multiple outputs under different LU/LC sce-narios demonstrates the extent of the synergies or trade-offs among these outputs. In the Willamette Basin appli-cation, we found little evidence of tradeoffs betweenecosystem services and biodiversity conservation: scenar-ios that enhance biodiversity conservation also enhancethe production of ecosystem services. Fears that a focus onecosystem services will fail to help us achieve biodiversityconservation goals (eg Terborgh 1999; McCauley 2006)were not borne out in this case. A negative correlationbetween commodity production values and (1) ecosystemservices and (2) biodiversity conservation is the one cleartradeoff we found. These results indicate that whenlandowner decisions are based solely on market returns(without payments for ecosystem services), they will tendto generate LU/LC patterns with lower provision ofecosystem services and biodiversity conservation.

Even this tradeoff, however, can be modified by policyinterventions. If markets for carbon sequestrationemerge, payments for sequestered carbon may make itmore profitable for landowners to choose LU/LC favoringconservation. In this application, payments for carbonsequestration cause the aggregate market value of theConservation scenario to be greater than the aggregatemarket value of the Development and Plan Trend scenarios(Figure 5). This result doesn’t necessarily mean that theConservation scenario would emerge if payments for car-bon sequestration were made. The actual LU/LC patternthat emerges under a carbon market will depend on theprices paid for sequestration, which carbon pools are eli-gible for payment, and the individual preferences oflandowners. However, it is more likely that land-usechoices with carbon payments, especially in rural areas,would generate a spatial pattern more like theConservation scenario than those of the Development andPlan Trend scenarios. Payments for water quality, soil con-servation, and storm peak mitigation would strengthenthe likelihood that LU/LC patterns similar to thosedescribed in the Conservation scenario would emerge.

Before payments for these ecosystem services are insti-tuted, however, clear links need to be made between theirbiophysical provision and their ultimate use by people.Other than carbon sequestration, we have only modeledbiophysical production of ecosystem services. The crucialsecond step is to determine how much of this productionis actually of value to people and where that value is cap-tured. In this study, we have done this with carbon seques-tration (we assumed that all sequestration provides valueto all people in the world). For other services, use values

will be determined by local patterns of land use and popu-lation density. For example, in a flooding-prone watershedin which few people or farms occur, flood mitigation ser-vices will provide relatively little benefit to people.

Another important caveat to our analysis is that we didnot include the market value of commodities generatedin urbanized areas in any scenario (this was done to keepthe base land area in the market value model equal acrossall scenarios). Because market returns on urban land tendto be higher than returns for other land uses, we probablyunderestimated the aggregate value of marketed com-modities for scenarios that experience greater urbaniza-tion (ie the Development scenario). In general, for land-use decisions involving a choice between intensive urbandevelopment and conservation, development valuesmight very well overwhelm the ecosystem services valuesthat could be generated by conserving the land. Weshould not expect existing markets or market valuation ofecosystem services inevitably to favor conservation, espe-cially in high-value urban areas. The kinds of analyses weshow here, however, make transparent the tradeoffsbetween ecosystem services, biodiversity conservation,and market returns, and that transparency alone is desir-able in engaging stakeholders and decision makers.

Another intriguing outcome of our analyses was thatthe scenarios did not produce more marked differences inthe provision of ecosystem services and biodiversity con-servation. This may be a reflection of the relatively mod-est LU/LC change under the scenarios considered here:“The stakeholder advisory group, which oversaw designof the future scenarios, did not consider…drastic land-scape alterations plausible, given Oregon’s history ofresource protection, social behaviors, and land-ownershippatterns” (Baker et al. 2004). Indeed, using more complexhabitat–species relationship data, Schumaker et al.(2004) also found little change in a biodiversity statusmeasure (essentially a countryside SAR score with 279species and a z value of 1) from 1990 to 2050 across thethree scenarios. The Willamette Basin has large tracts ofcontiguous forests in the Cascade and Coastal MountainRanges that remained relatively unchanged cross allthree scenarios. Most of these areas are not suitable foragriculture or urban development. They probably act as abuffer for maintaining provision of ecosystem servicesand biodiversity, no matter how great the changes on theBasin floor (Figures 1, 2, and 4). We expect the modelingand valuation approach illustrated here to reveal morestriking tradeoffs between conservation and developmentin rapidly developing regions.

Although the structure of the models presented herecan, in principle, include drivers besides land-use change(eg climate change), we have not included these in theanalysis to date. Furthermore, there may be importantfeedback effects, such as the amenity value of conservedland, that increases development pressure on land nearconserved areas. Including changes in climate, technol-ogy, market prices, human population, and feedback

Pollin

ation

E Nelson et al. Modeling the tradeoffs between ecosystem services and biodiversity

effects – all of which are likely to drive the ecological,social, and economic relationships that determine thevalue of ecosystem services in the future – is an essentialnext step in the application of InVEST.

� Acknowledgements

The authors thank D White, J Lawler, J Kagan, S Wolny,N Sandhu, S White, A Balmford, N Burgess, and MRouget for help in developing, testing, running, and pro-viding data for the InVEST models, as well as the conser-vation staffs of The Nature Conservancy and WorldWildlife Fund for comments on model design. In addi-tion, the National Center for Ecological Analysis andSynthesis, The Nature Conservancy, P Bing, H Bing, VSant, R Sant, B Hammett, and the Packard and WinslowFoundations are recognized for their generosity in sup-porting the Natural Capital Project.

� ReferencesAdams DM, Alig RJ, McCarl BA, et al. 1999. Minimum cost strate-

gies for sequestering carbon in forests. Land Econ 7755: 360–74.Antle JM and Stoorvogel JJ. 2006. Predicting the supply of ecosystem

services from agriculture. Am J Agr Econ 8888: 1174–80.Baker JP, Hulse DW, Gregory SV, et al. 2004. Alternative futures for

the Willamette River Basin, Oregon. Ecol Appl 1144: 313–24.Balmford A, Bruner A, Cooper P, et al. 2002. Economic reasons for

conserving wild nature. Science 229977: 950–53.Boody GB, Vondracek D, Andow M, et al. 2005. Multifunctional agri-

culture in the United States. BioScience 5555: 27–38.Chan KMA, Shaw MR, Cameron DR, et al. 2006. Conservation

planning for ecosystem services. PLoS Biology 44: 2138–52.Costanza R, Wilson M, Troy A, et al. 2006. The value of New Jersey’s

ecosystem services and natural capital. Burlington, VT: GundInstitute for Ecological Economics, University of Vermont.www.state.nj.us/dep/dsr/naturalcap/nat-cap-2.pdf. Viewed 14Nov 2008.

Costanza R, d’Arge R, de Groot R, et al. 1997. The value of theworld’s ecosystem services and natural capital. Nature 338877:253–60.

Daily GC (Ed). 1997. Nature’s services: societal dependence on nat-ural ecosystems. Washington, DC: Island Press.

Egoh B, Reyers B, Rouget M, et al. 2008. Mapping ecosystem servicesfor planning and management. Agri Ecosyst Environ 112277: 135–40.

Feng H. 2005. The dynamics of carbon sequestration and alternativecarbon accounting, with an application to the Upper MississippiRiver Basin. Ecol Econ 5544: 23–35.

Hulse D, Gregory S, and Baker J (Eds). 2002. Willamette River Basinplanning atlas: trajectories of environmental and ecologicalchange. Corvallis, OR: Oregon State University Press.

Jackson RB, Jobbagy EG, Avissar R, et al. 2005. Trading water for car-bon with biological carbon sequestration. Science 331100: 19–47.

Kaiser B and Roumasset J. 2002. Valuing indirect ecosystem services:the case of tropical watersheds. Environ Dev Econ 77: 701–14.

Kirby KR and Potvin C. 2007. Variation in carbon storage among treespecies: implications for the management of a small-scale carbonsink project. Forest Ecol Manag 224466: 208–21.

Kline JD, Moses A, and Alig RJ. 2001. Integrating urbanization intolandscape-level ecological assessments. Ecosystems 44: 3–18.

Lubowski RN, Plantinga AJ, and Stavins RN. 2006. Land-use changeand carbon sinks: econometric estimation of the carbon seques-tration supply function. J Environ Econ Manag 5511: 135–52.

MA (Millennium Ecosystem Assessment). 2005. Ecosystems andhuman well-being. Synthesis. Washington, DC: Island Press.

McCauley D. 2006. Selling out on nature. Nature 444433: 26–27.Naidoo R and Ricketts TH. 2006. Mapping the economic costs and

benefits of conservation. PLoS Biol 44: 2153–64.NRC (National Research Council). 2005. Valuing ecosystem service:

towards better environmental decision-making. Washington,DC: National Academies Press.

Nelson E, Polasky S, Lewis D, et al. 2008. Efficiency of incentives tojointly increase carbon sequestration and species conservation ona landscape. P Natl Acad Sci USA 110055: 9471-9476.

Omernik JM 1987. 2008. Ecoregions of the conterminous UnitedStates. Map (scale 1:7 500 000). Ann Assoc Am Geogr 7777:118–25.

Pereira HM and Daily GC. 2006. Modeling biodiversity dynamics incountryside landscapes. Ecology 8877: 1877–85.

Plantinga AJ, Mauldin T, and Miller DJ. 1999. An econometricanalysis of the costs of sequestering carbon in forests. Am J AgrEcon 8811: 812–24.

Polasky S, Nelson E, Camm J, et al. 2008. Where to put things: spatialland management to sustain biodiversity and economic returns.Biol Conserv 114411: 1505–24.

Ricketts, TH, Daily GC, Ehrlich PR, et al. 2004. Economic value oftropical forest to coffee production. P Natl Acad Sci USA 110011:12579–82.

Sala OE, Vuuren DV, Pereira HM, et al. 2005. Biodiversity across sce-narios. In: Carpenter SR, Pingali PL, Bennett EM, et al. (Eds).Ecosystems and human well-being: vol 2. Scenarios: findings ofthe Scenarios Working Group. Washington, DC: Island Press.

Schumaker NH, Ernst T, White D, et al. 2004. Projecting wildliferesponses to alternative future landscapes in Oregon’s WillametteBasin. Ecol Appl 1144: 381–400.

Smith JE, Heath LS, Skog KE, et al. 2006. Methods for calculatingforest ecosystem and harvested carbon with standard estimatesfor forest types of the United States. Newtown Square, PA: USDepartment of Agriculture, Forest Service, NortheasternResearch Station.

Terborgh J. 1999. Requiem for nature. Covelo, CA: Island Press. Tol RSJ. 2005. The marginal damage costs of carbon dioxide emis-

sions: an assessment of the uncertainties. Energ Policy 3333:2064–74.

Troy A and Wilson MA. 2006. Mapping ecosystem services: practicalchallenges and opportunities in linking GIS and value transfer.Ecol Econ 6600: 435–49.

Turner WR, Brandon T, Brooks M, et al. 2007. Global conservation ofbiodiversity and ecosystem services. BioScience 5577: 868–73.

US EPA (US Environmental Protection Agency). 2002. WillametteBasin alternative futures analysis. Environmental assessmentapproach that facilitates consensus building. Washington, DC:US Environmental Protection Agency, Office of Research andDevelopment.

US OMB (US Office of Management and Budget). 1992. Guidelinesand discount rates for benefit–cost analysis of Federal programs.Washington, DC: US Office of Management and Budget.

Watson RT, Noble IR, Bolin B, et al. 2000. Project-based activities.In: Watson RT, Noble IR, Bolin B, et al. (Eds). IPCC specialreport on land use, land-use change and forestry. Cambridge, UK:Cambridge University Press.

Wischmeier WH and Smith DD. 1978. Predicting rainfall erosionlosses – a guide to conservation planning. Washington, DC: USDepartment of Agriculture.

5Institute for Resources, Environment & Sustainability, University ofBritish Columbia, Vancouver, Canada; 6Department of Biology,Stanford University, Stanford, CA; 7Department of Human Dimensionsof Natural Resources, Colorado State University, Fort Collins, CO;8The Nature Conservancy, Arlington, VA; 9Conservation and ScienceDepartment, Lincoln Park Zoo, Chicago, IL; 10Conservation ScienceProgram, World Wildlife Fund-US, Washington, DC

11

© The Ecological Society of America wwwwww..ffrroonnttiieerrssiinneeccoollooggyy..oorrgg