Transcript

Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at SciVerse ScienceDirect

Ecosystem Services

2212-04http://d

n CorrE-m

neville.c1 Te

Pleassupp

journal homepage: www.elsevier.com/locate/ecoser

Impact of multiple interacting financial incentives on land use changeand the supply of ecosystem services

Brett Anthony Bryan 1, Neville David Crossman n

CSIRO Ecosystem Sciences and Sustainable Agriculture Flagship, Waite Campus, SA 5064, Australia

a r t i c l e i n f o

Article history:Received 5 October 2012Received in revised form6 February 2013Accepted 28 March 2013

Keywords:Agri-environment schemeLand use changeMarket-based instrumentsSpatial modeling

16/$ - see front matter Crown Copyright & 20x.doi.org/10.1016/j.ecoser.2013.03.004

esponding author. Tel.: +61 8 8303 8663.ail addresses: [email protected] (B.A. [email protected] (N.D. Crossman).l.: +61 8 8303 8581.

e cite this article as: Bryan, B.A., Croly of ecosystem services. Ecosystem

a b s t r a c t

Multiple financial incentives are increasingly common for managing agro-ecosystems. We explored theimpact of incentive interactions across multiple ecosystem services through their influence on land usechange potential. Taking a spatial approach, we quantified the economic potential for land use changefrom agriculture to carbon monocultures and environmental plantings. We assessed 1875 scenarios—exhaustive combinations of five incentive price levels for four services (food and fiber, fresh water,carbon sequestration and habitat), and three cost settings. Incentive interactions had complex effects—characterized by synergies and tensions, non-linearity, dependencies, and thresholds. Tensions occurredbetween commodity price and carbon price in supplying food and fiber, carbon sequestration, freshwater, and indirectly, habitat services. Water price displayed synergies with commodity price, andtensions with carbon price in supplying fresh water services. For the supply of habitat services, abiodiversity price depended on either high carbon prices or low commodity prices. Interaction effectsmay reduce policy efficiency wherever multiple incentives encourage the supply of services from agro-ecosystems.

Crown Copyright & 2013 Published by Elsevier B.V. All rights reserved.

1. Introduction

In agro-ecosystems, financial incentives commonly occurwhich affect the supply of ecosystem services through theirinfluence on land use and management (Antle and Stoorvogel,2006; Bryan, 2013; Lubowski et al., 2008; Metzger et al., 2006).Broadly, financial incentives for ecosystem services may be gen-erated by institutions ranging from global commodity trade inmarketed goods and services (e.g. cereals), to regionally- orlocally-implemented market-based schemes designed to encou-rage the production of non-marketed ecosystem services (e.g.habitat). The latter include a range of policy instruments such asdirect payments/rewards, tax incentives, cap and trade markets,voluntary markets, auctions, and certification programs (Farleyand Costanza, 2010; Tallis et al., 2008). However, little is known ofthe potential for interaction between financial incentives and theresulting impacts on policy efficiency and ecosystem services(Zhang and Pagiola, 2011). Exploring these incentive interactioneffects is the focus of this paper.

Policy interventions such as financial incentives often haveunanticipated consequences which may be positive (co-benefits),

13 Published by Elsevier B.V. All ri

),

ssman, N.D., Impact of mulServices (2013), http://dx.d

negative (trade-offs), or even perverse (the opposite of what wasintended) (Merton, 1936). In agro-ecosystems, commodity marketsare a prime example which have increased agricultural productionof marketed services like food and bioenergy, but at the expense ofnon-marketed services like habitat and water quality (Power,2010). The potential for achieving co-benefits has been demon-strated, particularly through the spatial targeting of paymentswhich prioritize cost-effectiveness across multiple services andrecognize spatial heterogeneity in service provision (Crossmanand Bryan, 2009; Crossman et al., 2010). Recent studies havesought to harness these co-benefits through the bundling ofmultiple ecosystem services (Connor et al., 2008; Deal et al.,2012; Wainger et al., 2010; Wendland et al., 2010). However, thepredominance of trade-offs between ecosystem services overspace and time (Raudsepp-Hearne et al., 2010; Rodriguez et al.,2006; Tallis et al., 2008) means that the failure to consider thebroader impacts of financial incentives across multiple ecosystemservices often leads to negative outcomes (Bateman, 2009).

In many agro-ecosystems, multiple financial incentives co-existfor the supply of ecosystem services. These incentives may inter-act, with consequences for land use and ecosystem services(Bryan, 2013). The interaction of financial incentives occurs aslandholders change land use and management in response to thetotality of economic opportunities and risks (Just and Antle, 1990).Some financial incentives act synergistically, working togethertoward achieving a policy objective. Other incentives may be

ghts reserved.

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i

B.A. Bryan, N.D. Crossman / Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎e2

divergent or antagonistic, creating tensions. Zhang and Pagiola(2011) found potential for synergies between a watershed con-servation payment scheme and a forest conservation paymentscheme for achieving watershed, biodiversity, and developmentobjectives in Costa Rica. In South Australia, Crossman et al. (2011b)found that a biodiversity payment could be used to augment acarbon price to enhance biodiversity conservation. Examples oftensions between financial incentives were evident in the US asthe federal Conservation Reserve Program paid people to retireenvironmentally-sensitive land from agriculture whilst other fed-eral farm subsidies encouraged continued agricultural production(Lubowski et al., 2008). Similarly, accounting for the costs of thewater used by reforested areas was found to reduce the effective-ness of a carbon price incentive in motivating reforestation inSouth Africa (Chisholm, 2010).

The influence of financial incentives on land use, and in turn,the influence of land use on ecosystem services, involve complexmany-to-many relationships (Bryan, 2013). Each financial incentivecan influence multiple land uses, and each land use can affectmultiple ecosystem services. These influences can be positive ornegative. Hence, the aggregate impact of multiple incentivesacross multiple ecosystem services through their influence onland use is difficult to predict (Bryan, 2013). Although seldomexplored, understanding these interaction effects is necessary toensure the efficiency of financial incentives for ecosystem servicesin agro-ecosystems including capturing synergies and avoidingtensions (White et al., 2012). Here, we present the first quantita-tive, integrated exploration of the interaction of multiple financialincentives and their impacts across multiple ecosystem services.

We assessed the impact of financial incentives on ecosystemservices through their effect on land use profitability—a key driverof land use change (Irwin and Geoghegan, 2001; Lubowski et al.,2008). We took a spatial approach in identifying areas with landuse change potential (areas where an economic opportunity existsfor land use change) under a range of financial incentive (price)scenarios. We then assessed the impact of this potential change onmultiple ecosystem services. This type of approach has beencompared to more sophisticated land use change forecasts andshown to provide timely insight at a level of detail sufficient for

Fig. 1. Location map and major land uses in the stu

Please cite this article as: Bryan, B.A., Crossman, N.D., Impact of mulsupply of ecosystem services. Ecosystem Services (2013), http://dx.d

informing policy decisions (Antle and Stoorvogel, 2006; Antle andValdivia, 2006). Similar approaches have been widely used toassess the impact of financial incentives on the supply of servicesfrom agro-ecosystems for land uses such as bioenergy feedstock(Bryan et al., 2008, 2010a, 2010b) and reforestation (Dymond et al.,2012; Flugge and Abadi, 2006; Harper et al., 2007; Hunt, 2008;Paterson and Bryan, 2012; Townsend et al., 2012).

Focussing on the 15 million ha agricultural region of SouthAustralia, we quantified the supply of four ecosystem services(food and fiber production, carbon sequestration, fresh waterprovision, habitat for local native species) from three land uses(existing agriculture, carbon monocultures (single species Euca-lyptus plantations), environmental plantings (ecological restora-tion of mixed native species)) using a range of biophysical processmodels. We calculated the net economic returns from each landuse over 40 years from 2010 to 2050 in net present value (NPV)terms, given the presence of exogenously-determined incentiveprices for the supply of these services. We assessed 1875 scenariosincluding all combinations of the five prices for each of the fourservices, and assessed model sensitivity using three economic costparameter settings (high, median and low). For each scenario, weidentified areas with land use change potential based on neteconomic returns and quantified the impact of these changes onthe four ecosystem services. We quantified the effect of incentivesusing Spearman's rank correlation analysis and visualized theinteractions between influential incentives on each ecosystemservice. The implications of incentive interactions on policyefficiency in agro-ecosystems are discussed.

2. Methods

2.1. Study area

Land use in the study area is dominated by mixed cropping/grazing systems, interspersed with patches of remnant naturalland (Fig. 1). Climate is Mediterranean in the south grading tosemi-arid in the north, and soils are nutrient-deficient. The regionis responsible for around 18% of Australia's cereal production and

dy area—South Australia's agricultural regions.

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i

B.A. Bryan, N.D. Crossman / Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎ e3

13% of its' sheep production. In addition to established agriculturalcommodity markets, several financial incentives for the supply ofecosystem services are existing or planned in the region includinga water market (cap and trade) implemented through bothnational and state-based legislation (Young and McColl, 2009),biodiversity payments under a range of programs (Connor et al.,2008; DCCEE, 2011), and a national carbon trading scheme (DCCEE,2011).

2.2. Modeling and mapping ecosystem services

We employed a range of spatial biophysical models to estimatethe production of food and fiber, carbon sequestration, fresh water,and habitat services. All spatial analysis was raster-based at 1 hagrid cell resolution. Each spatial layer consisted of 6715rows�8561 columns and 9,190,508 valid grid cells. A GeographicInformation Systemwas used to assemble and map the data layers.

For food and fiber, we used a crop growth and agriculturalsystems model (the Agricultural Production Systems sIMulator orAPSIM) (Keating et al., 2003) to estimate mean annual productionof wheat and legumes (field peas) for 80 unique soil-climate zonesunder historical mean climate (supporting information text). Wederived livestock production data from the 2006 agriculturalcensus (supporting information text).

For carbon sequestration, we used a forest growth model(3-PG) (Landsberg and Waring, 1997) to estimate sequestrationrates from reforestation from 2010 to 2050 under historical meanclimate (supporting information text). For carbon monocultureswe stratified the study area into high, medium, and low rainfallzones and simulated growth of a climatically-adapted, fast-growing Eucalyptus species in each. Environmental plantings weresimulated as a suite of mixed local native tree and understoreyspecies.

For fresh water, a landscape hydrology model (the AustralianWater Resources Assessment—Landscape model or AWRA-L) (vanDijk and Renzullo, 2011) was used to quantify marginal changes inwater resource availability induced by reforestation (ML/ha/yr) ata 0.051 grid cell resolution. We specified that reforestationimpacted fresh water services in water supply catchments only(supporting information text).

For habitat, we used a systematic landscape restoration modelto identify agricultural areas of high priority for ecological restora-tion (Bryan and Crossman, 2008; Crossman et al., 2007, 2011b).High priority areas were those closer to remnant vegetation

Table 1Incentive prices analyzed in this study.

Incentives Units Price levels

Agricultural commodities multiplier 0.5 1.0 1.5 2.0 2.5Carbon $/tCO2

−e 10 20 30 40 50Water $/ML 0 500 1000 1500 2000Biodiversity $/ha/yr 0 50 100 150 200

Table 2Economic parameters comprising the three economic scenarios used in sensitivity anal

Economic cost setting(S)

Discount rate(r)

Establishment costs ($/ha)

Carbon monocultures(ECf)

Environmental plantin(ECf)

High cost 0.11 4000 5000Median cost 0.07 3000 3750Low cost 0.03 2000 2500

Please cite this article as: Bryan, B.A., Crossman, N.D., Impact of mulsupply of ecosystem services. Ecosystem Services (2013), http://dx.d

(thereby reducing fragmentation and enhancing connectivity)such that at least 30% of the area (with a minimum area of1000 ha) of each of 12 major pre-European vegetation commu-nities, 15 climate zones, 18 soil types, and 26 biogeographicsubregions was covered by vegetation (restored and remnant)(supporting information text)).

2.3. Incentive price and cost scenarios

For each service, five constant, exogenously-determined incen-tive price levels were specified which covered the range ofplausible values given recent trends and future projections(Table 1). Commodity prices were taken as the average over theperiod 2002–2007 (Australia Bureau of Agricultural and ResourceEconomics, 2010) and price multipliers selected to cover agricul-tural commodity price movements over this period and morerecently, including the 2008 and 2011 food price spikes. Carbonprices were selected to center around the 23 $/tCO2

−e startingprice for Australia's carbon market but also include a price rangerepresentative of price variation experienced within local volun-tary markets and the European market, and price trajectoriesmodeled by the Australian Government (Australian Government,2011b; Lawson et al., 2008). Water prices were specified based onprice movements of high security, high reliability water entitle-ments, and prescribed water resources in the South Australian partof the southern Murray basin (Australian Government, 2011a).Biodiversity prices were specified based on prices paid for on-ground works in auctions for conservation contracts in southernAustralia (Bryan and Kandulu, 2009; Connor et al., 2008; Crossmanet al., 2011a).

Economic parameters such as discount rate, establishmentcosts, and ongoing costs such as maintenance and transactionscosts are uncertain and strongly influence the profitability ofreforestation land uses (Bryan et al., 2008). To assess the sensitivityof the model to these costs we specified three cost parametersettings—high, median, and low (Table 2). Cost parametervalues were derived from a range of previous studies and fieldexperience (Bryan and Kandulu, 2009; Bryan et al., 2008, 2010b;Crossman et al., 2011b; Polglase et al., 2008), and field experi-ence (e.g. River Murray Forest Program, Greening Australia,professional reforestation contractors). In total, we analyzed1875 combinations of agricultural commodity price, carbonprice, water price, biodiversity price (Table 1), and three costparameters settings (Table 2).

2.4. Economic valuation

Economic returns to agriculture, carbon monocultures, andenvironmental plantings were then calculated for the price andcost scenarios in net present value (NPV) terms in Australiandollars. Economic returns from food and fiber (wheat, field peas,sheep, and beef cattle) were calculated as profit at full equity (i.e.economic return to land, capital, and management, exclusive offinancial debt) based on mean production estimates and commod-ity price, minus fixed and variable costs (Bryan et al., 2011b). Long

ysis.

Ongoing costs ($/ha/yr) Carbon sequestration riskbuffer (B)

gs Maintenance costs(MC)

Transaction costs(TC)

100 100 0.480 80 0.260 60 0.0

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i

B.A. Bryan, N.D. Crossman / Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎e4

term expected returns were calculated in NPV terms for eachcommodity price scenario considering farming system rotations(Bryan et al., 2011b).

We calculated profit at full equity by adapting a profit functionthat has been widely used (Bryan et al., 2008, 2009, 2010a, 2011a;Hajkowicz and Young, 2005) and validated (Bryan et al., 2011b). Aprofit layer πðk; pAÞ was calculated for each agricultural commodityk in the set of agricultural commodities K wheat; field peas;beef and cattle sheep and for each agricultural commodity pricemultiplier pA in PA (Table 1) as revenue minus fixed and variablecosts such that

πðk;pAÞ ¼ ðP1kpAQ1kTRNk þ P2kpAQ2kQ1kÞ−ðQCkQ1k þ ACk þ FOCk þ FDCk þ FLCkÞ ð1Þ

The profit function incorporates multiple products from asingle enterprise. Whilst the primary product for livestock (Q1kfor k∈{sheep, beef cattle}) is meat, sheep also produce wool whichis represented as a secondary product (Q2k for k∈{sheep}) in theprofit function. Production data for crops (Q1k for k∈{wheat, fieldpeas}) was derived from the APSIM modeling, whilst productiondata for livestock (Q1k, Q2k, TRNk for k∈{beef cattle, sheep}) wasderived from agricultural census data (see above). Price data forboth primary and secondary products (P1k, P2k) were specified asthe average commodity prices for the period 2002–2007 (i.e.before the commodity price spike of mid-2008) from AustraliaBureau of Agricultural and Resource Economics (2010). Variableand fixed cost data was derived from state government extensioninformation (Rural Solutions, 2008) which provides indicativefigures. Profit function parameters and their values are summar-ized in Table 3.

Agricultural systems in the study area typically consist ofannual rotations of crops and livestock and these vary geographi-cally as a function of soils, climate, and management. To capturethe economic rent from the farming system rotation we incorpo-rated rotation frequency in a measure of long term expectedannual returns (Bryan et al., 2011b). To do this we calculated theareal proportion of each type of crop and livestock within a 10 kmradius of each grid cell using a moving-window approach based onstate government land use mapping (Department of Water Landand Biodiversity Conservation, 2008). We assumed the arealproportion within this spatial neighborhood reflects the temporalrotation frequency of each enterprise within each grid cell. Theoutput was a frequency layer wk for each land use k in K where∑k∈kwk ¼ 1 for each grid cell. We then incorporated the profitlayers with the farming system rotation frequency layers tocalculate a single layer of long term expected annual economicreturns from agriculture πðAg; pAÞ (Bryan et al., 2011b)

πðAg;pAÞ ¼ ∑k∈K

wkπðk; pAÞ ð2Þ

Table 3Description of profit function parameters, their units, and value ranges across the study

Parameter Description Units

P1k Price of primary product (grain/meat) $/t or $/DSEP2k Price of secondary product (wool only) $/kgQ1k Yield of primary product t/ha or DSE/haQ2k Yield of secondary product (wool only) kg/DSETRNk Proportion of herd sold 0≤TRN≤1 for liveQCk Quantity dependent variable costs $/t or $/DSEACk Area dependent variable costs $/haFOCk Fixed operating costs $/haFDCk Fixed depreciation costs $/haFLCk Fixed labor costs $/ha

Please cite this article as: Bryan, B.A., Crossman, N.D., Impact of mulsupply of ecosystem services. Ecosystem Services (2013), http://dx.d

Long term expected annual economic returns from agriculturefor each agricultural commodity price multiplier πðAg;pAÞ (Table 1)was then converted into NPV terms for each economic cost settings in S (Table 2) using the specific discount rate rs for the costsetting

πNPV ðAg; pAÞ 1−1

ð1−rsÞT� �

=rs ð3Þ

We then calculated the economic returns from the two refor-estation land uses f in F{carbon monocultures, environmentalplantings} from 2010 to 2050 across the study area in NPV terms.We assumed both reforestation land uses were not harvested.Economic returns were calculated for each combination of carbonprice pC in PC, water price pW in PW, biodiversity price pB in PB(Table 1) and economic cost setting s in S (Table 2) such that

πNPV ðf ; pC ; pB; pW ; sÞ ¼ ∑T

t ¼ 0

ðpCCf tð1−γsÞ þ pBθÞ−ðMCS þ TCsÞð1þ rsÞt

� �

−ðECf s þ pWφδÞ ð4ÞT is the time horizon of 40 years. Cft is a spatial layer

quantifying the amount of carbon sequestered in each year t foreach tree-based land use f in F as modeled using 3-PG. γs is thecarbon sequestration risk buffer and (1−γs) is the proportion ofcarbon sequestration expected under each economic cost setting sin S. θ is a binary layer describing the spatial distribution ofbiodiversity priority areas, such that θ¼1 in priority areas whenf¼environmental plantings, and 0 otherwise. Thus, pBθ is thebiodiversity payment for environmental plantings in high prioritybiodiversity areas. The maintenance cost MCs and transaction costTCs parameters, and the discount rate rs, for each economic costsetting s in S (Table 2) are the same for both carbon monoculturesand environmental plantings and are uniform over the study area.The establishment costs ECfs vary for each f in F and s in S. φ is abinary layer describing the spatial distribution of water supplycatchment areas, such that φ¼1 in catchment areas, 0 otherwise. δis the spatial layer quantifying water use by trees as modeled byAWRA-L. Thus, pWφδ represents the upfront cost of water entitle-ments for both carbon monocultures and environmental plantingsfor each water price pW in PW. Economic returns from carbonmonocultures and environmental plantings were calculated from2010 to 2050 for each of the 1875 incentive price and costscenarios.

2.5. Modeling land use change

The spatial distribution of land use change potential wasidentified under each scenario as the land use with the maximumeconomic return for each grid cell. We used a spatial approach,assumed uncorrelated incentive prices and held them constant in

area.

Values

Wheat Field Peas Beef Sheep

216 274 97 25–35n.a. n.a. n.a. 3.19–4.270–3.62 0–4.84 0–9.7 0–9.7n.a. n.a. n.a. 5

stock n.a. n.a. 0–0.46 0–0.9616–32 16–32 1.63–3.28 3–7122–311 153–197 6–13 1–1322–32 22–32 5–21 1–4610–29 10–29 1–7 1–1522–36 22–36 1–20 1–25

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i

Fig. 2. Long term expected economic returns (net present value) from agricultural systems in the study area under the five commodity price multiplier scenarios calculatedunder the median cost scenario discount rate (7% pa).

B.A. Bryan, N.D. Crossman / Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎ e5

real terms over the modeling horizon. This enabled us to investi-gate interaction effects on ecosystem services across the entireincentive space.

We excluded natural land and applied to model to agriculturalland use areas only. We assessed two alternative land uses (i.e.carbon monocultures and environmental plantings) which maybecome financially attractive in the presence of price incentives forecosystem services. All three land uses vary in the amount of eachecosystem service produced and, hence, economic returns. Thesealso vary spatially and under different price and cost scenarios.

To calculate land use change potential we used an integerprogramming approach. We specified three binary land use vari-ables x0, x1 and x2. The variable x0¼1 when the cell is agriculture,

Please cite this article as: Bryan, B.A., Crossman, N.D., Impact of mulsupply of ecosystem services. Ecosystem Services (2013), http://dx.d

0 otherwise; x1¼1 when the cell is carbon monocultures, 0 other-wise, and; x2¼1 when the cell is environmental plantings,0 otherwise. To create a spatial layer of land use change potentialfor each price and cost scenarios we solved the following integerprogram for each grid cell:

maximize πNPVðAg; pA; sÞx0 þ πNPVðCM; pC ; pB; pW ; sÞx1þ πNVPðEP; pC ; pB; pW ; sÞx2 ð5Þ

subject to x0 þ x1 þ x2 ≤1 ð6ÞHere we use CM to explicitly represent carbon monocultures

and EP to represent environmental plantings rather than the fterm used above.

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i

B.A. Bryan, N.D. Crossman / Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎e6

The model has a high spatial, temporal, and scenario resolutionwhich exerts a high computational demand. To reduce thecomputational load, we aggregated calculations to 5-year timeintervals t in T{2.5, 7.5, 12.5, 17.5, 22.5, 27.5, 32.5, 37.5} years andused the carbon sequestration averaged over the 5-year periodsurrounding each time step t. For example, for t¼2.5 we averagedthe carbon sequestration at years 2010 and 2015, for t¼7.5 weaveraged the carbon sequestration at years 2015 and 2020, and soon. To speed processing, modeling was undertaken with Python(van Rossum and the Python community, 2010) and NumPy (Joneset al., 2001). Simulations were computed in parallel using Python'smultiprocessing package. All 1875 simulations took 3–4 h to

Fig. 3. Example of the net present value of economic returns from carbon sequestraeconomic cost settings, a 0 $/ML water price, and a 0 $/ha/yr biodiversity price.

Please cite this article as: Bryan, B.A., Crossman, N.D., Impact of mulsupply of ecosystem services. Ecosystem Services (2013), http://dx.d

complete on a 32-core server with 512GB of memory. We usedthe to visualize land use change potential.

2.6. Impact of land use on ecosystem services

We then summarized the impact of land use change potentialon the supply of ecosystem services. To calculate the impact ofland use change potential on ecosystem services for each marketprice and scenario combination, we intersected each land use withits respective layers of ecosystem service provision for the areaswhere each land use was most profitable in the land use change

tion by carbon monocultures under the five carbon price scenarios assuming median

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i

B.A. Bryan, N.D. Crossman / Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎ e7

potential map. Total impact for each ecosystem service was thenaggregated over all land uses in the land use change potential mapand summarized.

2.7. Analysis of incentive interactions

Spearman rank correlation tests were used to quantify theinfluence of each financial incentive on each ecosystem serviceusing R version 2.12.2 (R Development Core Team, 2011). Correla-tions were calculated across the full set of 1875 scenarios.Correlations were then graphed using ggplot2 (Wickham, 2009)to provide an overall picture of the strength of synergies andtensions in incentive interactions affecting each ecosystem service.We then visualized the interaction effects between pairs of

Fig. 4. Example of the net present value of economic returns from carbon sequestrationeconomic cost settings, a 0 $/Ml water price, and a 0 $/ha/yr biodiversity price.

Please cite this article as: Bryan, B.A., Crossman, N.D., Impact of mulsupply of ecosystem services. Ecosystem Services (2013), http://dx.d

influential financial incentives on each of the four ecosystemservices using contour graphs created using Matplotlib (Hunter,2007).

3. Results

The spatial distribution of ecosystem service production ispresented in the supporting information. This includes the agri-cultural production of food and fiber (wheat, legumes, sheep,wool, and beef cattle) (Fig. S1), carbon sequestration from carbonmonocultures (Fig. S2) and environmental plantings (Fig. S3),impact on fresh water provision by reforestation through theincreased evapotranspiration by trees (Figs. S4 and S5), and the

by environmental plantings under the five carbon price scenarios, assuming median

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i

B.A. Bryan, N.D. Crossman / Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎e8

habitat services provided by environmental plantings (Fig. S6).Economic returns to agriculture, carbon monocultures, and envir-onmental plantings showed substantial spatial variation andvariation with price and cost multiplier (Figs. 2–4).

Incentive prices, particularly agricultural commodity and car-bon prices, had a significant effect on the spatial distribution ofland use change potential. Land use change potential increasedwith carbon price and decreased with commodity price (Fig. 5).Reforestation land uses became more profitable first in the wetterareas dominated by less-profitable grazing enterprises. Environ-mental plantings became profitable where these wetter, grazing-dominated areas were also of high biodiversity priority. High cost

Fig. 5. Land use change potential under the median cost settings for the

Fig. 6. Influence of financial incentives on ecosystem service

Please cite this article as: Bryan, B.A., Crossman, N.D., Impact of mulsupply of ecosystem services. Ecosystem Services (2013), http://dx.d

parameter values greatly reduced land use change potential (Fig. S7)whilst low cost parameters greatly increased it (Fig. S8).

We found both synergies and tensions between incentives intheir effect on ecosystem service supply (Fig. 6). Food and fiberproduction was positively influenced by commodity price butnegatively influenced by carbon price and, to a lesser extent,biodiversity price. Carbon sequestration was positively influencedby carbon price and negatively influenced by commodity price.Fresh water supply was positively influenced by both commodityprice and water price but negatively influenced by carbon price.Habitat was positively influenced by both carbon price andbiodiversity price but negatively influenced by commodity price.

median water price (1000 $/ML) and biodiversity price (100 $/ha/yr).

s including significance at po0.05 (*) and po0.01 (**).

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i

Fig. 7. Tension between agricultural commodity price and carbon price incentives and their impact on food and fiber production (median values presented, negative valuesrepresent a decrease in service provision whilst positive values represent an increase).

Fig. 8. Tension between agricultural commodity price and carbon price incentives and their impact on carbon sequestration.

B.A. Bryan, N.D. Crossman / Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎ e9

A tension between agricultural commodity and carbon priceincentives was reflected by losses of food and fiber productiondecreasing with commodity price and increasing with carbonprice (Fig. 7). The interactions between agricultural commodityand carbon price incentives was not linear, with most of the effecton food and fiber occurring at higher carbon price and lowercommodity price, under the median cost settings. Much fewerlosses of food and fiber production occurred under the high costscenario, with greater losses experienced under the low costscenario.

A tension between agricultural commodity and carbon priceincentives also affected carbon sequestrationwhich increased withcarbon price and decreased with commodity price (Fig. 8). Theeffects of interactions between agricultural commodity and carbonprice incentives on carbon sequestration displayed a similar butopposite pattern to food and fiber, with most influence occurringat higher carbon price, and lower commodity price for the mediancost settings. Very little carbon sequestration occurred under thehigh cost scenario with greater sequestration achieved under thelow cost scenario.

A tension between agricultural commodity and carbon priceincentives affecting fresh water, where losses increased sharplywith carbon price and decreased with commodity price (Fig. 9).Water price and carbon price incentives also displayed a tension,whereas water price and commodity price incentives were syner-gistic in the supply of fresh water services. The impact on freshwater services had a steep gradient, occurring through narrowband of carbon, commodity, and water prices. Negligible loss of

Please cite this article as: Bryan, B.A., Crossman, N.D., Impact of mulsupply of ecosystem services. Ecosystem Services (2013), http://dx.d

fresh water occurred at high cost settings, with a substantial lossat low cost settings.

Habitat was influenced in complex, non-linear ways by inter-acting commodity, carbon, and biodiversity prices (Fig. 10). Inter-actions between biodiversity price and carbon price incentivewere characterized by dependencies and threshold effects suchthat biodiversity incentives strongly increased habitat, but only athigh carbon prices. A similar relationship was found with com-modity price incentives whereby biodiversity incentives stronglyincreased habitat but only at low commodity prices. Interactioneffects between commodity price and carbon price incentiveswere indirect and secondary to these previous two interactioneffects (Fig. 10). Low levels of habitat services were supplied underhigh cost settings with much more supplied under low costs. Costparameter settings also altered the pattern of the interaction effects.

4. Discussion

We have presented the first quantitative, integrated assessmentof the interactions between multiple financial incentives and theimpacts across multiple ecosystem services, through their influ-ence on land use potential. Some incentives worked together toprovide economic opportunities for changing land use to carbonmonocultures and/or environmental plantings. Other incentiveshad the opposite effect. Tensions occurred between commodityprice and carbon price incentives in the supply of food and fiber,carbon sequestration, fresh water, and indirectly, habitat services.

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i

Fig. 9. Incentive interactions and their impact fresh water. Tensions occur between agricultural commodity price and carbon price incentives, and between water price andcarbon price incentives. A synergy occurs between commodity price and water price incentives. The steep gradient in service supply over a narrow incentive price rangeillustrates a threshold effect.

B.A. Bryan, N.D. Crossman / Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎e10

Water price exhibited synergies with commodity price and ten-sions with carbon price incentives in encouraging supply of freshwater services. To influence habitat services, a biodiversity pricewas dependent on either a high carbon price or low commodityprice.

The nature of the relationships between incentives and theirimpact on ecosystem services varied and it is instructive to exploretheir causation. Some incentive interactions were more linear withthe impact on ecosystem service supply occurring over a broadrange of prices (e.g. the impact of interacting carbon price andcommodity price on food and fiber Fig. 7 and carbon sequestrationFig. 8). The flatter frequency distributions and uncorrelated eco-nomic returns to competing land uses (agriculture and reforesta-tion) result in land use change occurring gradually with changes inincentive price. Other interactions were more complex. An exam-ple was the 3-way dependency interaction effect found betweenbiodiversity, carbon, and commodity prices in influencing habitatservices (Fig. 10). Here, biodiversity prices generated habitatservices, but required a higher carbon price and lower commodityprice to be effective. Higher carbon prices and lower commodityprices were required to make reforestation land uses competitivewith agriculture, with the biodiversity price covering the differ-ential between the higher economic returns possible through

Please cite this article as: Bryan, B.A., Crossman, N.D., Impact of mulsupply of ecosystem services. Ecosystem Services (2013), http://dx.d

carbon monocultures compared to environmental plantings(Crossman et al., 2011b). Several incentive interactions were alsocharacterized by non-linearity and threshold effects. Examplesinclude the sharp gradient in fresh water service provisionoccurring with small changes in incentive prices (Fig. 9). Freshwater services were restricted to only a few small water catchmentareas with relatively similar economic returns to agriculture andreforestation land uses. Thresholds reflect non-linear land usechange behavior driven by a reservation price—as incentive pricesreach a point where the economic returns from reforestation landuses exceed those from agriculture, land use change occurs rapidlyacross these small areas. The above analysis suggests that thecomplex nature of the interaction effects emerge from the specificgeographic and economic situation of the study area. The general-izability of the shape of these relationships to other study areas isunclear and warrants further research.

Incentive interactions have significant policy implications.Awareness of interactions is essential to inform the design ofpolicy and institutional arrangements which avoid competitionbetween incentives and antagonistic inflation of the price ofecosystem services (Lubowski et al., 2008; Stavins and Jaffe, 1990).Although there is limited scope to influence global markets foragricultural commodities, other policy institutions supplementing

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i

Fig. 10. Incentive interactions and their impact habitat. A tension occurred between agricultural commodity price and carbon price incentives under the median costscenario which changes to a synergy under the low cost scenario. A synergy occurred between biodiversity price and carbon price incentives characterized by dependenciesand threshold effects. A tension is shown between biodiversity price and commodity price incentives also characterized by dependencies and threshold effects.

B.A. Bryan, N.D. Crossman / Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎ e11

agricultural commodity markets (e.g. agricultural subsidies in Europeand the US, exceptional circumstances payments in Australia) can bemodified if they compete with carbon and biodiversity incentives.Also, competition between water and carbon incentives could inflatethe price of both fresh water and carbon sequestration services.Incentive interactions also indicate that biodiversity incentives aloneare not likely to supply much habitat—either high carbon prices orlow commodity prices are also necessary. An understanding of theseinteractions can help guide institutional design to enhance policyefficiency. For example, regulatory controls can be conceived such asprecluding reforestation in prime farmlands and water supply catch-ments, making carbon payments for environmental plantings only toachieve both carbon and biodiversity co-benefits, and limitingagricultural subsidies to prime farmland and, in marginal agriculturalareas, re-focussing subsidies towards land retirement and reforesta-tion for ecosystem service provision. Institutional design is requiredto support financial incentives to capture the synergies and mitigatetensions between them.

Here, we made no assumption about the relationships betweenincentives but rather provide an exhaustive exploration of the fullincentive interaction space. However, economic theory suggests

Please cite this article as: Bryan, B.A., Crossman, N.D., Impact of mulsupply of ecosystem services. Ecosystem Services (2013), http://dx.d

that well-designed markets are dynamic, reacting to changes insupply and demand to efficiently supply ecosystem services. Thus,for example, as reforestation motivated by a local carbon incentivereduces the supply of agriculture commodities, the price of thesecommodities should rise. Reality, however, is very different. Manyfactors conspire to reduce the efficiency of ecosystem servicesmarkets including institutional design, divergent scales of opera-tion (e.g. local payments vs. global markets), imperfect informa-tion, mis-specified property rights, and other factors contributingto market friction. Our results can guide more fully-specifiedpartial or general equilibrium models for detailed analyses ofthese dynamic interactions between diverse incentives for ecosys-tem services. Detailed market analysis can better inform thedesign of more efficient market-based policy and institutionalarrangements.

While our approach identifying land use change potentialbased on economic profitability has been widely used (Antle andStoorvogel, 2006; Antle and Valdivia, 2006; Bryan et al., 2008,2010a, 2010b; Dymond et al., 2012; Hunt, 2008), it has limitations.Although economic returns are a major driver of change in landuse and management (Lubowski et al., 2008), a range of other

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i

B.A. Bryan, N.D. Crossman / Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎e12

individual and societal factors are also important including con-straints and barriers to adoption (e.g. labor, capital, materials andknow-how), unmodelled costs and benefits (Lubowski et al.,2008), and policy and institutional arrangements. Even thoughthe assessment of land use change potential based on economicprofitability has been effective for the broad assessment of policyon ecosystem services (Antle and Stoorvogel, 2006; Briassoulis,2008) a more complete specification of behavioral response (e.g.Radeloff et al., 2012) would improve the realism of land usechange scenarios.

5. Conclusion

Interactions between financial incentives impacted multipleecosystem services, and this has significant implications for policy.We found both synergies and tensions between incentives in thesupply of ecosystem services. While some incentive interactionswere linear, others were more complex—displaying non-linearity,dependencies, and threshold effects. Failure to understand theseinteractions between incentives and their effect across multipleecosystem services can result in inefficient policy outcomes suchas unintended negative impacts and the inflated costs of ecosys-tem services. Whilst the nature of incentive interactions may differfrom place to place, given the ubiquity of financial incentives forecosystem services globally, we suspect that these inefficienciesare commonplace. We conclude that the quantitative, integratedassessment of the impact of incentive interactions on ecosystemservices is urgently required to guide more detailed analysis ofdynamic interactions, and inform institutional design thatachieves synergies and minimizes tensions.

Acknowledgments

We are grateful for the support of CSIROs Sustainable Agricul-ture Flagship and Integrated Carbon Pathways initiative, and theSA Government. The assistance of A. van Dijk, D. Summers, and M.Nolan is appreciated. Comments from S. Hatfield-Dodds and otheranonymous reviewers greatly improved the manuscript.

Appendix A. Supporting information

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.ecoser.2013.03.004.

References

Antle, J.M., Stoorvogel, J.J., 2006. Predicting the supply of ecosystem services fromagriculture. American Journal of Agricultural Economics 88, 1174–1180.

Antle, J.M., Valdivia, R.O., 2006. Modeling the supply of ecosystem services fromagriculture: a minimum-data approach. Australian Journal of Agricultural andResource Economics 50, 1–15.

Australia Bureau of Agricultural and Resource Economics, 2010. Australian Com-modities, Commonwealth of Australia, Canberra.

Australian Government, 2011a. Market price information for Murray—Darling BasinWater Entitlements.

Australian Government, 2011b. Strong Growth, Low Pollution: Modeling a CarbonPrice, Canberra, Australia.

Bateman, I.J., 2009. Bringing the real world into economic analyses of land usevalue: incorporating spatial complexity. Land Use Policy 26, S30–S42.

Briassoulis, H., 2008. Land-use policy and planning, theorizing, and modeling: lostin translation, found in complexity? Environment and Planning B 35, 16–33.

Bryan, B.A., 2013. Incentives, land use, and ecosystem services: synthesizingcomplex linkages. Environmental Science and Policy 27, 124–134.

Bryan, B.A., Crossman, N.D., 2008. Systematic regional planning for multipleobjective natural resource management. Journal of Environmental Manage-ment 88, 1175–1189.

Please cite this article as: Bryan, B.A., Crossman, N.D., Impact of mulsupply of ecosystem services. Ecosystem Services (2013), http://dx.d

Bryan, B.A., Crossman, N.D., King, D., Meyer, W.S., 2011a. Landscape futuresanalysis: assessing the impacts of environmental targets under alternativespatial policy options and future scenarios. Environmental Modelling andSoftware 26, 83–91.

Bryan, B.A., Hajkowicz, S., Marvanek, S., Young, M.D., 2009. Mapping economicreturns to agriculture for informing environmental policy in the Murray DarlingBasin, Australia. Environmental Modeling and Assessment 14, 375–390.

Bryan, B.A., Kandulu, J.M., 2009. Cost-effective alternatives for mitigating Cryptos-poridium risk in drinking water and enhancing ecosystem services. WaterResources Research, 45.

Bryan, B.A., King, D., Wang, E., 2010a. Biofuels agriculture: landscape-scale trade-offs between fuel, economics, carbon, energy, food, and fiber. GCB Bioenergy 2,330–345.

Bryan, B.A., King, D., Wang, E.L., 2010b. Potential of woody biomass production formotivating widespread natural resource management under climate change.Land Use Policy 27, 713–725.

Bryan, B.A., King, D., Ward, J.R., 2011b. Modelling and mapping agriculturalopportunity costs to guide landscape planning for natural resource manage-ment. Ecological Indicators 11, 199–208.

Bryan, B.A., Ward, J., Hobbs, T., 2008. An assessment of the economic and.environmental potential of biomass production in an agricultural region. LandUse Policy 25, 533–549.

Chisholm, R.A., 2010. Trade-offs between ecosystem services: water and carbon in abiodiversity hotspot. Ecological Economics 69, 1973–1987.

Connor, J.D., Ward, J.R., Bryan, B., 2008. Exploring the cost effectiveness of landconservation auctions and payment policies. Australian Journal of Agriculturaland Resource Economics 52, 303–319.

Crossman, N.D., Bryan, B.A., 2009. Identifying cost-effective hotspots for restoringnatural capital and enhancing landscape multifunctionality. Ecological Eco-nomics 68, 654–668.

Crossman, N.D., Bryan, B.A., King, D., 2011a. Contribution of site assessment towardprioritising investment in natural capital. Environmental Modelling and Soft-ware 26, 30–37.

Crossman, N.D., Bryan, B.A., Ostendorf, B., Collins, S., 2007. Systematic landscaperestoration in the rural-urban fringe: meeting conservation planning and policygoals. Biodiversity and Conservation 16, 3781–3802.

Crossman, N.D., Bryan, B.A., Summers, D.M., 2011b. Carbon payments and low-costconservation. Conservation Biology 25, 835–845.

Crossman, N.D., Connor, J.D., Bryan, B.A., Summers, D.M., Ginnivan, J., 2010.Reconfiguring an irrigation landscape to improve provision of ecosystemservices. Ecological Economics 69, 1031–1042.

DCCEE, 2011. Securing a Clean Energy Future: the Australian Government's ClimateChange Plan. Department of Climate Change and Energy Efficiency, Common-wealth of Australia, Canberra, Australia.

Deal, R.L., Cochran, B., LaRocco, G., 2012. Bundling of ecosystem services to increaseforestland value and enhance sustainable forest management. Forest Policy andEconomics 17, 69–76.

Department of Water Land and Biodiversity Conservation, 2008. Land Use Mappingfor South Australia. Soil and Land Program, Department of Water, Land andBiodiversity Conservation, Adelaide.

Dymond, J.R., Ausseil, A.-G.E., Ekanayake, J.C., Kirschbaum, M.U.F., 2012. Tradeoffsbetween soil, water, and carbon—a national scale analysis from New Zealand.Journal of Environmental Management 95, 124–131.

Farley, J., Costanza, R., 2010. Payments for ecosystem services: from local to global.Ecological Economics 69, 2060–2068.

Flugge, F., Abadi, A., 2006. Farming carbon: an economic analysis of agroforestry forcarbon sequestration and dryland salinity reduction in Western Australia.Agroforestry Systems 68, 181–192.

Hajkowicz, S., Young, M., 2005. Costing yield loss from acidity, sodicity and drylandsalinity to Australian agriculture. Land Degradation and Development 16, 417–433.

Harper, R.J., Beck, A.C., Ritson, P., Hill, M.J., Mitchell, C.D., Barrett, D.J., Smettem, K.R.J.,Mann, S.S., 2007. The potential of greenhouse sinks to underwrite improved landmanagement. Ecological Economics 29, 329–341.

Hunt, C., 2008. Economy and ecology of emerging markets and credits for bio-sequestered carbon on private land in tropical Australia. Ecological Economics66, 309–318.

Hunter, J.D., 2007. Matplotlib: a 2D graphics environment. Computing in Scienceand Engineering 9, 90–95.

Irwin, E.G., Geoghegan, J., 2001. Theory, data, methods: developing spatially expliciteconomic models of land use change. Agriculture, Ecosystems and Environment85, 7–23.

Jones, E., Oliphant, T., Peterson, P., SciPy community, 2001. SciPy: Open SourceScientific Tools for Python. ⟨http://www.scipy.org.⟩.

Just, R.E., Antle, J.M., 1990. Interactions between agricultural and environmentalpolicies—a conceptual-framework. American Economic Review 80, 197–202.

Keating, B.A., Carberry, P.S., Hammer, G.L., Probert, M.E., Robertson, M.J., Holzworth, D.,Huth, N.I., Hargreaves, J.N.G., Meinke, H., Hochman, Z., McLean, G., Verburg, K.,Snow, V., Dimes, J.P., Silburn, M., Wang, E., Brown, S., Bristow, K.L., Asseng, S.,Chapman, S., McCown, R.L., Freebairn, D.M., Smith, C.J., 2003. An overview ofAPSIM, a model designed for farming systems simulation. European Journal ofAgronomy 18, 267–288.

Landsberg, J.J., Waring, R.H., 1997. A generalised model of forest productivity usingsimplified concepts of radiation-use efficiency, carbon balance and partitioning.Forest Ecology and Management 95, 209–228.

Lawson, K., Burns, K., Low, K., Heyhoe, E., Ahammad, H., 2008. Analysing theEconomic Potential of Forestry for Carbon Sequestration Under Alternative

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i

B.A. Bryan, N.D. Crossman / Ecosystem Services ∎ (∎∎∎∎) ∎∎∎–∎∎∎ e13

Carbon Price Paths. Australian Bureau of Agricultural and Resource Economics,Canberra.

Lubowski, R.N., Plantinga, A.J., Stavins, R.N., 2008. What drives land-use change inthe United States? A national analysis of landowner decisions. Land Economics84, 529–550.

Merton, R., 1936. The unanticipated consequences of purposive social action.American Sociological Review 1, 894–904.

Metzger, M.J., Rounsevell, M.D.A., Acosta-Michlik, L., Leemans, R., Schrotere, D.,2006. The vulnerability of ecosystem services to land use change. Agriculture,Ecosystems and Environment 114, 69–85.

Paterson, S.E., Bryan, B.A., 2012. Food-carbon trade-offs between agriculture andreforestation and the efficiency of market-based policies. Ecological Economics17, 21.

Polglase, P., Paul, K., Hawkins, C., Siggins, A., Turner, J., Booth, T., Crawford, D.,Jovanovic, T., Hobbs, T., Opie, K., Almeida, A., Carter, J., 2008. RegionalOpportunities for Agroforestry Systems in Australia. Rural Industries Researchand Development Corporation, Canberra.

Power, A.G., 2010. Ecosystem services and agriculture: tradeoffs and synergies.Philosophical Transactions of the Royal Society B 365, 2959–2971.

R Development Core Team, 2011. R: A language and environment for statisticalcomputing. R Foundation for Statistical Computing, Vienna, Austria.⟨http://www.R-project.org/⟩.

Radeloff, V.C., Nelson, E., Plantinga, A.J., Lewis, D.J., Helmers, D., Lawler, J.J., Withey, J.C.,Beaudry, F., Martinuzzi, S., Butsic, V., Lonsdorf, E., White, D., Polasky, S., 2012.Economic-based projections of future land use in the conterminous United Statesunder alternative policy scenarios. Ecological Applications 22, 1036–1049.

Raudsepp-Hearne, C., Peterson, G.D., Bennett, E.M., 2010. Ecosystem servicebundles for analyzing tradeoffs in diverse landscapes. Proceedings of theNational Academy of Sciences of the United States of America 107, 5242–5247.

Rodriguez, J.P., Beard, T.D., Bennett, E.M., Cumming, G.S., Cork, S.J., Agard, J., Dobson,A.P., Peterson, G.D., 2006. Trade-offs across space, time, and ecosystem services.Ecology and Society, 11.

Please cite this article as: Bryan, B.A., Crossman, N.D., Impact of mulsupply of ecosystem services. Ecosystem Services (2013), http://dx.d

Rural Solutions, 2008. South Australian Farm Gross Margin Guide 2008. RuralSolutions, Adelaide.

Stavins, R.N., Jaffe, A.B., 1990. Unintended impacts of public-investments on privatedecisions—the depletion of forested wetlands. American Economic Review 80,337–352.

Tallis, H., Kareiva, P., Marvier, M., Chang, A., 2008. An ecosystem services frameworkto support both practical conservation and economic development. Proceed-ings of the National Academy of Sciences of the United States of America 105,9457–9464.

Townsend, P.V., Harper, R.J., Brennan, P.D., Dean, C., Wu, S., Smettem, K.R.J., Cook, S.E.,2012. Multiple environmental services as an opportunity for watershed restora-tion. Forest Policy and Economics 17, 45–58.

van Dijk, A.I.J.M., Renzullo, L.J., 2011. Water resource monitoring systems and therole of satellite observations. Hydrology and Earth System Sciences 15, 39–55.

van Rossum, G., the Python community, 2010. The Python Programming Language:Version 2.7.1. The Python Software Foundation.

Wainger, L.A., King, D.M., Mack, R.N., Price, E.W., Maslin, T., 2010. Can the concept ofecosystem services be practically applied to improve natural resource manage-ment decisions? Ecological Economics 69, 978–987.

Wendland, K.J., Honzak, M., Portela, R., Vitale, B., Rubinoff, S., Randrianarisoa, J.,2010. Targeting and implementing payments for ecosystem services: opportu-nities for bundling biodiversity conservation with carbon and water services inMadagascar. Ecological Economics 69, 2093–2107.

White, C., Costello, C., Kendall, B.E., Brown, C.J., 2012. The value of coordinatedmanagement of interacting ecosystem services. Ecology Letters 15, 509–519.

Wickham, H., 2009. ggplot2: Elegant Graphics for Data Analysis. Springer, NewYork.

Young, M.D., McColl, J.C., 2009. Double trouble: the importance of accounting forand defining water entitlements consistent with hydrological realities. Austra-lian Journal of Agricultural and Resource Economics 53, 19–35.

Zhang, W., Pagiola, S., 2011. Assessing the potential for synergies in the imple-mentation of payments for environmental services programmes: an empiricalanalysis of Costa Rica. Environmental Conservation 38, 406–416.

tiple interacting financial incentives on land use change and theoi.org/10.1016/j.ecoser.2013.03.004i


Top Related