Spatial valuation of forests' enviromental assets: an application to Andalusian Silvopastoral farms

23 Pages·2017·0.65 MB·English

Paola Ovando, Alejandro Caparrós, Luis Díaz-Balteiro,
María Pasalodos, Santiago Beguería, José L. Oviedo,
Gregorio Montero and Pablo Campos
Spatial valuation of forests' enviromental
assets: an application to Andalusian
Silvopastoral farms

Article (Published version)
(Refereed)

Original citation:
Ovando, Paola, Caparrós, Alejandro, Díaz-Balteiro, Luis, Pasalodos, María, Beguería, Santiago,
Oviedo, José L., Montero, Gregorio and Campos, Pablo (2016) Spatial valuation of forests'
enviromental assets: an application to Andalusian Silvopastoral farms. Land Economics . 93, (1) pp. 87-
108. ISSN 0023-7639

Reuse of this item is permitted through licensing under the Creative Commons:

© 2017 The University of Wisconsin Press © CC BY-NC-ND 4.0

This version available at: http://eprints.lse.ac.uk/66172/
Available in LSE Research Online: September 2017

LSE has developed LSE Research Online so that users may access research output of the School.
Copyright © and Moral Rights for the papers on this site are retained by the individual authors and/or
other copyright owners. You may freely distribute the URL (http://eprints.lse.ac.uk) of the LSE Research
Online website.
Spatial Valuation of Forests’ Environmental Assets:
An Application to Andalusian Silvopastoral Farms
Paola Ovando, Alejandro Caparro´s, Luis Diaz-Balteiro,
Mar´ıa Pasalodos, Santiago Beguer´ıa, Jose L. Oviedo,
Gregorio Montero, and Pablo Campos
ABSTRACT.Wedevelopamodelthatestimatesspa- translatedintothevaluationofenvironmental
tiallyallocatedenvironmentalassetvaluesforthesi- assets (EAs)1 in a way that is meaningful for
multaneousprovisionofsevenecosystemservices.We decision makers(Fenicheland Abbott 2014).
examinetheeffectofheterogeneousspatialandeco- Forest ecosystems are spatially heteroge-
nomicfactorsonassetfigures,andidentifypotential
neous areas in which the provision of ESs is
forestryabandonmentproblemswhencontinuingwith
not distributed uniformly, either in space or
forestry activity becomes unprofitable for the land-
over time (Ha¨yha¨ et al. 2015; Lawler et al.
owner.Ourresultsshowarelevantspatialvariability
2014; Schaafsma et al. 2014; Yuan et al.
accordingtoforestspeciesdistributionandstructure.
Weexaminepotentialtrade-offsamongsilvopastoral 2012).Thus,movingfromEStoEAvaluesis
provisioning services, water, and carbon sequestra- especially pertinent in these ecosystems, as
tion services. Results forecast the abandonment of tree growth, forest depletion, and forestry
forestryactivityandquantifythesignificantimpactof operations might affect the dynamics of ES
discountratesandpricesonassetvalues.(JELQ23,
Q51)
1Hereweusetheterm“environmentalasset,”as“nat-
ural capital” is considered to be a broader measure that
I.INTRODUCTION
would include the stock of all EAs, embracingecosystem
assetsandmineralandenergyresources(UNetal.2014b).
Recent initiatives for moving toward a
green economy triggered the interest in de-
Theauthorsare,respectively,postdoctoralresearcher,
velopingenvironmentalaccountingtoanalyze
DepartmentofEnvironmentalSocialSciences,Swiss
and track the stateofecosystemsandtheser- FederalInstituteofAquaticScienceandTechnology,
vices they provide (Millennium Ecosystem Du¨bendorf, Switzerland; director and senior re-
Assessment Board 2005; UN et al. 2014a, searcher,InstituteofPublicGoodsandPolicies,Con-
2014b). In recent years, there has been a no- sejo Superior de Investigaciones Cient´ıficas(CSIC),
ticeableefforttoconsiderexplicitlythespatial Madrid, Spain; professor, School of Forestry Engi-
configurationoftheprovisionofvariouseco- neeringandNaturalResources,TechnicalUniversity
ofMadrid(UPM),Madrid,Spain;researcher,Forest
systemservices(ESs)(seeWolff,Schulp,and
Research Centre, National Institute for Agricultural
Verburg2015forareview)andnaturalstocks.
andFoodResearchandTechnology(INIA),Madrid,
Likewise, there has been an appreciable pro-
Spain; researcher, Estacio´n Experimental de Aula
gress in the integration of biophysical and
Dei,ConsejoSuperiordeInvestigacionesCient´ıficas
economiclandusemodelstosimulatethespa- (CSIC),Madrid,Spain;researcher,InstituteofPublic
tial and temporalpatternsofprovisionofdif- Goods and Policies, Consejo Superior de Investiga-
ferentESsatrelevantspatialscales(Bateman cionesCient´ıficas(CSIC),Madrid,Spain;researcher,
et al. 2013; Lawler et al. 2014). Nonetheless, Forest Research Centre, National Institute for Agri-
and despite recent attempts at ES quantifica- culturalandFoodResearchandTechnology(INIA),
tion and mapping, these have rarely been Madrid, Spain; and professor, Institute of Public
Goods and Policies, Consejo Superior de Investiga-
cionesCient´ıficas(CSIC),Madrid,Spain.
Land Economics (cid:129) February 2017 (cid:129) 93 (1): 87–108 Thisopenaccessarticleisdistributedundertheterms
ISSN0023-7639;E-ISSN1543-8325 of the CC-BY-NC-ND license (http://creative
(cid:2)2017bytheBoardofRegentsofthe commons.org/licenses/by-nc-nd/4.0) and is freely
UniversityofWisconsinSystem availableonlineat:http://le.uwpress.org.
88 LandEconomics February2017
supply (Biber et al. 2015; Ovando, Oviedo, al. 2014b) and to provide a practical model
andCampos2016).ThisstudyfocusesonAn- for its spatial valuation.
dalusia, a region in the south of Spain whose Our model simultaneously computes for
forests are mainly of the Mediterraneantype. five silvopastoral provisioning services, in-
This type of forest forms a unique mosaic of cluding timber, cork, firewood, pinenuts,
terrestrialecosystemsshapedbydivergingcli- grazingresources,andtheprovisionofwater2
matic(oftenextreme),geomorphological,and and one climate regulating service through
anthropogenic factors, and that is frequently carbon dioxide (CO2) sequestration (carbon
characterized by its multifunctionality (Scar- hereinafter).ThemodelestimatestheEAval-
ascia-Mugnozza et al. 2000) and high levels ues derived from the provision of the afore-
of biodiversity (Myers et al. 2000). The An- said ESs in a group of 567 private silvopas-
dalusian case is a good example for illustrat- toral farms that are distributed across
ing the spatial variation in the intertemporal Andalusia.3 The application integrates spa-
provision of ESs and the potential trade-offs tially explicit biophysical and economic data
involved. atthefarmlevelforthemainforestspeciesin
Thebenefitsassociatedwithmarket-priced this region (Quercus ilex, Q. suber, Pinus pi-
nea,P.halepensis,P.pinaster,andEucalyptus
andnonmarketforestproducts,suchasprivate
globulusandE.camadulensis,jointlyreferred
amenities, biodiversity-scenic values, public
to as Eucalyptus sp.), as well as for treeless
recreation, and carbon sequestration, have
shrubland and grassland.
been estimated for different Mediterranean
TheEAmodelestimatestheexpectedtem-
forests, mainly at the forest case study level
poral pattern of benefits and costs linked to
(Bernues et al. 2014; Campos and Caparro´s
silvopastoral, carbon, and water production
2006; Caparro´s, Campos, and Montero2003;
functions by forest species and farm. Those
Caparro´s et al. 2010; Ovando et al. 2010;
benefits and costs are time-varying figures
Ovando, Oviedo, and Campos 2016). Those
that fluctuate with the assumptions on price
benefits have also been estimated for larger
levels and discounting rates, aswell asinac-
spatial scales such as regions and countries,
cordance with the expected forest manage-
although,inaveryaggregatedmanner(Merlo
ment practices and tree growth and with ex-
and Croitoru 2005). Both case study and re-
plicit spatial attributes such as the slope
gionalapproachesshowthevariouswaysfor-
gradient, existing tree and shrub inventories,
estscontributetohumanandeconomicactiv-
the quality of the sites for growing timber or
ities, but do not delve deeply into the spatial
cork, soil structure, and precipitation levels.
and temporal distributionsof benefitsandas-
Our benefit and cost estimations consider the
setvaluesassociatedwiththeprovisionoffor-
spatial-explicit age class distribution of pres-
est ESs.
ent forests and two alternative management
In this study we develop an EA valuation
options once current forest rotations come to
model that extends the System of Environ-
an end: carrying out forest regeneration in-
mental and Economic Accounts Central
vestment or, alternatively, forestry activity
Framework (SEEA-CF) criteria, in terms of
abandonment,thechoiceofwhichdependson
its production function boundaries. The
the profitability of those optionsfortheland-
SEEA-CF offers an internationally accepted owner.
statistical standard for environmental ac- Ourresultsrevealanoticeablespatialvari-
counting, and provides the guidelines neces- abilityinEAvaluesandindicatethepotential
sary to develop EA accounts for individual trade-offs associated with silvopastoral pro-
naturalresourcessuchastimberorwater.Our
approach,incontrasttotheSEEA-CF,consid-
erstheforestasafunctionalunitthatsupplies 2Ecosystemsregulatetheflowandpurificationofwater,
multiple products, entailing trade-offsamong while forests influence the quantity of water available lo-
theprovisionfunctionsofasingleES.Inthis cally; in this sense water is considered as a provisioning
service(Haines-YoungandPotschin2013;TEEB2010).
manner,weaimtocontributetothescientific
3SeeFigureA1intheonlinesupplement,availableat
debateonecosystemassetsaccounting(UNet http://le.uwpress.org.
93(1) Ovandoetal.:SpatialValuationofEnvironmentalAssets 89
visioningservices,carbon,andwater.EAval- in Andalusia) that is distributed across 193
uesarehighlydependentonfutureforestevo- municipalities. The farms have an average
lution and management. Therefore, an size of 525 ha (standard deviation 849 ha).
additional outcome of our model is that it Thesevenforestspeciesincludedinthisstudy
identifies potential forestry abandonment at represent67%ofthefarms’area,whileshrubs
the site level, as a result of an expected un- and grasslands make up 19%. Other forest
profitable forest regeneration investment. species (4%) and crops (10%) occupy the re-
Likewise, the model allows for the explora- maining area.
tionoftheeffectofpaymentsforESsonforest
investmentdecisions.Finally,ourresultsalso PricingESsandEAs
highlight the significant effect of economic
assumptions regarding discount rates and From an environmental accounting stand-
prices, on both the EA values and the extent point, an EA is defined as the naturally oc-
of anticipatedforestryabandonment. curring biotic (whether natural, seminatural,
ormodified)andnonbioticcomponentsofthe
II.MATERIALSANDMETHODS Earth that provide a flow of ESs, which, in
combinationwithlaborandmanufacturedas-
SilvopastoralFarmsCaseStudiesandArea sets, contribute to the generation of products
ofStudy used in human and economic activities (UN
et al. 2014a, 2014b; Obst and Vardon 2014).
Andalusiaisaverydiverseregion,withal- MarketsforEAsandtheservicestheyprovide
titudes ranging from sea level up to 3,400 m areoftenincompleteormissing,especiallyfor
and with climatic conditions that vary from stocks and goods with weak exclusion, such
the rainiest point in the Iberian Peninsula to as public products (Fenichel and Abbott
the desert of Almer´ıa. This region covers 2014).
84,023 km2, which is similar in size to Aus- The EA and ES values are, however, not
tria.Morethan50%ofthisterritoryiscovered directlyobservableevenforstocksandgoods
by Mediterranean forests, consisting mainly withstrongexclusion,sincethoseareembed-
of a mix of native slow-growing oaks, pine ded in the market price for assets and prod-
species, shrubs, and grasses (CMA 2010). ucts, respectively. Market asset prices would
These are complex ecosystemsinwhichtree, internalize the value of ESs associated with
shrub, and herbaceous vegetation have been forest products, as landownersholdtheprop-
traditionally managed jointly to obtain raw erty rights on them. The challenge for eco-
materials such as cork, timber, firewood, and nomic valuation is to split up the asset value
pinenuts and to provide hunting and grazing intothesinglecontributionofeachforestben-
resources, which ascribes them as silvopas- efit and its associated ES. Hedonic pricing
toral systems.4 Private ownership dominates models might be useful to estimate the land
(73%) the area covered by silvopastoral sys- assetvalueassociatedwithdifferentcommer-
tems in Andalusia (Campos 2015). cial forest benefits (Zhang, Meng, and Poly-
The 567 silvopastoral farms included in akov 2013) when statistical information on
thisstudyaredistributedacrossAndalusiaand forest properties’ sales and their attributes is
weretakenfromasurveyof765forestowners available. This isnotthecasewithlandprice
whose properties were randomly selected in statistics for forest properties in Andalusia
this region (see Oviedo et al. [2015] and the (Campos et al. 2009), which require alterna-
online supplement5). They jointly occupy an tiveassetvaluationmethods,aswedetaillater
areaof2,975km2(9.3%oftotalprivatefarms on. In any case, the hedonic price approach
wouldnotbeabletocapturepublicnonmarket
values,asthemarketdoesnotassigntheprop-
4Thesesystemscompriseadeliberategrowingofwoody ertyrightsovertheseproductstolandowners.
perennialsonthesameunitoflandaslivestockininteracting
Land leasing and forest products’ prices
combinations to obtain multiple products from the same
embedthevalueofprovisioningservicessuch
managementunit(Nair1993).
5Availableathttp://le.uwpress.org. as grazing resources, cork, or timber. Like-
90 LandEconomics February2017
wise,thereisusuallyaquantifiablehumanin- Ontheotherhand,thisstudyconsidersthe
put in terms of both labor and manufactured mostcommonforestrypracticesinAndalusia,
assets, which is combined with the relevant assumingthecontinuationofthebusiness-as-
ES to produce benefits to humans (UN et al. usualforestmanagement.Forestryoperations
2014b).Thedifferencebetweenmarketprices includeshrubclearing,pruning,thinning,and
and the unit labor, manufactured input, and commercial harvesting of pinenuts, cork, or
fullcapitalcostswouldrendertheunitnatural timber, with probabilities of occurrence de-
resource rent (UN et al. 2014a, 2014b), and fined by silvicultural models (Montero et al.
thisunitpriceisusedtovaluetheprovisioning 2015). Forestry costs depend on specificspa-
services considered in this study. For those tialattributesoftheforests,inparticulartheir
ESs whose property rights are not attributed structure (species, density, and age class dis-
tothelandowner,suchaswaterandcarbonin tribution)andtheslopegradient,andtheyac-
theforestsofthestudyarea,weusesurrogate countforunitwageandinputpricesobserved
marketprices.Wefurtherassumethatcarbon in Andalusia in 2010.6
andwaterarejointbenefitsofforestmanage- Our cost and benefit projections consider
ment, thus no labor and manufactured costs that the relative prices (output/costs) will re-
main constant in the future.Weacknowledge
are attributedto their production functions.
thatthismightbeastrongassumptioninview
Forest carbon is not included in the Euro-
of the price tendencies observedoverthelast
pean Union Emissions Trading Scheme (EU-
decade,7 but, on the other hand, there is no
ETS).Nonetheless,theEU-ETSistheclosest
robust evidence to back the idea that those
market available for forestry carbon in An-
price tendencies will persist over time, espe-
dalusia, and its emission allowance (EUA)
cially so since our model accounts for slow-
pricescanbeusedandarepreferabletoprices
growing and long-rotation forest species.
obtained from completely simulatedmarkets.
Therefore, we opt for a more conservative
We use a single regional environmentalprice
scenarioinwhichtheunitpricesforsilvopas-
toestimatetheeconomicvalueofwaterflows.
toralproducts,carbon,andwaterandtheunit
ThispricecorrespondstotheunitEApriceof
productioncostsareconstantovertime.How-
water estimated by (Berbel and Mesa 2007,
ever, we further check the sensitivity of EA
141)usingahedonicpricemodelforirrigated
values to increases and decreases up to 50%
agricultural lands in Andalusia. This model
inthenetbenefit8obtainedfromsilvopastoral
useslandpricestatisticsthat,inAndalusia,are
products, carbon sequestration, and water as
available only for agricultural lands (CAP
theresultofchangesinthepricelevelofthose
2011) and not for forestlands. The EA price
outputs with respect to the baseline 2010
ofwater(Pw),updatedtoyear2010,attainsa
prices,whileproductioncostsareassumedto
value of 4.04 €/m3, and the water ES price
remain constant.
(pw) is estimated using real discounting rates
EAvaluesarequantifiedasthediscounted
(r) ranging from 2% to 6%: pw=Pwr.
net presentvalue(NPV)ofthestreamofESs
Output prices and forestry operation costs
(estimated as resource rent) that a foresteco-
includedinthisstudydonotaccountforsub-
sidies and taxes on production. Our benefit
and cost projections assume constant unit
6Seetheonlinesupplementfordetails,availableathttp:/
prices for output and forestry operations, as
/le.uwpress.org.
well as that the returns to scale are constant. 7The prices of forestry products have experienced a
The baseline prices correspond to those ob- markeddecrease,withacompoundannualgrowth(CAG)
served in Andalusia and in the EU-ETS mar- rateof–3.3%,overthelastdecade(2005–2014).Bycon-
trast, agricultural basic input prices and wages have in-
kets for silvopastoral provisioning services
creasedataCAGrateof3.3%and2.1%,respectively,over
and carbon, respectively, in year 2010. Tim- the same time period (SGAPC 2014; MAGRAMA 2014;
ber,cork,andpinenutyieldsarevaluedusing MARM2009).
average stumpage prices observed in Anda- 8Estimated as the difference between the benefits ac-
cruedfromsalesofforestryproductsandleasingtheland
lusia in the period 2008–2010 (updated to
outforgrazing,theimputedvalueofnetcarbonsequestra-
2010)byspecies,product,andqualityclasses. tionandeconomicwater,minusforestryproductioncosts.
93(1) Ovandoetal.:SpatialValuationofEnvironmentalAssets 91
system is expected to yield in the futurecon- (2003) to price timber stock, to multiperiodi-
sideringaninfinitetimehorizon.TheNPVap- caloutputssuchascork,pinenuts,orfirewood
proach is the standard rule for pricing assets andcarbonsequestrationduetotreegrowth.11
in a deterministic case (Dixit and Pindyck EAs account for both the present forest rota-
1994; Fenichel and Abbot 2014) and follows tion (EA ) and for the expected stream of
T1
the SEEA-CF recommendations (UN et al. ESsafterthisrotation.Ourassetvaluationap-
2014a). The SEEA-CF recommends estimat- proachcouldbeappliedtobotheven-andun-
ing EA values by capitalizing the flow of re- even-agedforests,regardlessoftheinitialfor-
source rents over the life of assets. This re- eststructure,speciesdistribution,androtation
source rent represents the economic rent age, as detailed later in this section.
accrued in relation to EAsandshouldideally The EAT1 is estimatedas follows:
accountbothfortheremunerationtothoseas-
setsas productionfactorsandfortheirdeple- EAT1=pp′⋅Q,
tion (UN et al. 2014a). p′=(p1,p2,...,pd,...,pn),
p p p p p
The ES monetary value we estimate rep-
resents the returns to EAs after covering all being
the operating and full manufactured capital
costs. The operating costs include labor, in- pd=(cid:2)T [(pd−pd)⋅γ ⋅β ⋅δ(t−d)]
p j=s f m dt dt
termediate manufactured inputs (raw materi-
for each d={1,2,...,T},
als and services), and the depletion of manu- q
t
factured assets involved in the production γ = , [1]
dt
process of different forest products; while qd
capitalcostsembraceanormalreturntoman-
where p′⋅ is a vector of unit resource rent
ufactured assets used in this production pro- p
(eurospercubicmeterorpermetricton).This
cess9 (Ovando, Oviedo, and Campos 2016).
pricevectorincludesforitsTrowsthestand-
Ourmodelimplicitlycomputesfor(1)thepo-
ing price (pd) of the product and the cost of
tential EA depletion, by anticipating existing f
forestrytreatments(pd),12comprisingtheop-
treeinventorywithdrawalsduetoforestfires, m
portunity cost of manufactured capital. β
naturalmortality,ormanagement,and(2)im- dt
represents the conditional probability that a
provements (entries) due to tree growth and
treeofanagedisloggedatanytageclassto
recruitment (described later in this section).
be reached (d≤t). Q is a vector that records
The expected ES values of timber, cork, and
theexistingstockofforestproductsorcarbon
firewoodarequantifiedunderasimplifiedap-
foreachageclassattheinitialperiod(2010).
proach,10 based on the value of expected ex-
γ is a vector of expansion/contraction fac-
tractions minus operating and full capital dt
tors that relate the unit stock of a tree at age
costs.
class d (q ) and the unit stock of that same
d
treeattheageclasst(q).Finally,δrepresents
ProvisioningandRegulatingServicesDependingon t
the discount function δ=(1+r)−1.
TreeGrowth
The provision of ESs after the present ro-
Weextendtheassetvaluationapproachap- tationdependsontheprobabilityofforestre-
plied by Caparro´s, Campos, and Montero generation investment (φ), which equals 1 if
current forestry activity continues in the fu-
tureandequals0ifthisactivityisabandoned.
9Inourapplicationthisnormalreturntomanufactured
assetsequals3%forthemainscenarioandvarieswithinthe Thedecisionaboutforestregenerationinvest-
discountrateapplied.
10AmorecomprehensiveapproachforestimatingESs
as an environmental income (Ovando et al. 2015) would
considernaturaltimber/cork/firewoodgrowthasanoutput 11Carbonsequestrationduetotreegrowthisestimated
of each period, the standing value of the woody products asafunctionoftreediameter(Montero,Ruiz-Peinado,and
thatareharvestedasanintermediatecost(input)intheform Mun˜oz2006).
ofwork-in-progressused,andtherevaluationofthose(hold- 12Forestry treatments refer to those operations sched-
ing gains) woody products along the accounting period uledfortheyearsthatareleftbeforereachingtherotation
(Campos2015;Ovando,Oviedo,andCampos2016). ofaparticularforestspecies.
92 LandEconomics February2017
ment is simulated at each forest unit as trees The EA associated with the provision of
reach their theoretical rotation age (Montero silvopastoralproductswouldtakeazerovalue
et al. 2015), which varies spatially according in the event that the NPV of net benefits as-
totheinitialageclassdistributionoftheforest sociatedwiththeproductionofasilvopastoral
unit.13 We assume that the forestry activity product is negative (UN et al. 2014a, 158).
willcontinueiftheNPVoftheexpectedbene- The negative net benefits are then redistrib-
fits of the new rotation surpasses the NPV of uted as returns to manufactured investment,
its costs. This probability changes within the with no return to the EA. The ESs related to
pricelevelanddiscountingratesimulatedsce- carbon are estimated each period as the dif-
narios. ference betweengrossCO sequestrationand
2
We assume thatforestregenerationinvest- release, and as we assume that carbon se-
mentwillresultinanewforestrotationofthe questration does not involve any manufac-
samespeciesateachforestunit.Forestregen- tured assets or labor, a negative EA value
erationinvestmentincludesoperationssuchas would indicate loss in carbon environmental
shelterwood cutting to promote seedling and stock value.
recruitment of new individuals, weeding out, For the main (business-as-usual) scenario
a grazing set-aside period of up to 20 years, weassumetherewillbenorelevanttechnical
andtheclear-cuttingofmaturetreesafterthis innovations in forest management that in-
set-aside period (Ovando et al.2010).Onthe creasethenetbenefitsassociatedwithforestry
other hand, we consider that forestry aban- activityandtreecarbonsequestration.Thisas-
donment would lead to shrub encroachment sumption implies that the growth and yields
and would change the present distribution of of the new forest rotation (if φ=1) will be
forestspecies.14Grazingresources,carbonse- similar to those of the former one. This as-
questration, and water will be the only ESs sumption may not be unrealistic in Mediter-
delivered by this land use. raneanforests,whicharecharacterizedbylow
WithaninfinitetimehorizontheEAisthen commercial profitability and productivity
estimated as
rates(CamposandCaparro´s2006;Camposet
al.2008),andalowadoptionoftechnological
(cid:3)φ⋅(δT+1−s⋅(1−δT)−1⋅EA )+
EA=EA + T2 [2] innovations such as the use of genetically
T1 (1−φ)⋅(δT+1−s⋅(1−δ)−1⋅ylt), modified trees (Montero et al. 2015). Never-
theless,throughthesensitivityanalysisweex-
where s is the age of the trees at the starting amine the effect of relative increases in the
valuation period and T their rotation age, value of forestry benefits with respect to the
which varies among forest species and silvi- cost,whichcouldbeduetoanincreaseeither
culturalmodels.EAT2representstheEAvalue inpricesorinproductivity(asresultsoftech-
associated with the rotation that follows the nological innovations). Alternatively, we ex-
present one if there are no economic restric- amine the effect of relative decreases in for-
tions to tree regeneration. The measurement estry benefits, which in turn may represent a
ofEAT2issimilartothatofEAT1usingequa- decline in forest productivity due to adverse
tion [1], although in that case, the model ac- climatic conditions or a decline in prices of
counts for the complete forestry rotation outputs with respectto the production costs.
(from year 1 to T), assuming that the second
rotationisfollowedbyaninfinitesequenceof
GrazingResources
identical rotations. The variable yl represents
t
the annual ES of the alternative land use l in Grazingresourcesincludeacorns(onlyfor
case of forestryabandonment. Q. ilex) and grass (swards, browses, and
fruits) produced in forest, shrublands, and
grasslands and that are consumed by live-
stock,game,andotherwildspecies.Theeco-
13Seedefinitionlaterinthissection. nomic value of grazing resources depends
14Theabandonmentofforestsandruralareasisacom-
upon the market opportunity cost of leasing
montrendinnorthernMediterraneancountriesandcanin-
creasetheriskofwildfires(Allardetal.2013). the land out for livestock grazing and the
93(1) Ovandoetal.:SpatialValuationofEnvironmentalAssets 93
numberofforageunitsobtainedbydominant The third term of equation [3] represents the
vegetation (e) and province (j) in Andalusia. transition from qk to qlj considering aperiod
g g
Thepricesandquantitiesusedaretakenfrom (τ) of 50 years after forestry abandonment,
a survey of 765 agroforestry farm owners in and that this transition is linear. Finally, we
this region that includes the sample of 567 consider that after the period, qlj and plj
g g
farmsconsideredinthisstudy(Oviedo,Cam- would remain constant over time.
pos, and Caparro´s2015). TheEAi estimationadditionallyconsiders,
g
We estimate the EA value associated with assubtrahend,theoperatingandcapitalman-
grazing resources (EAi) at the farm level us- ufacturedcostsinvolvedinthesupplyofgraz-
g
ing the average land leasing price (pej) per ing resources (cmj), as well as an additional
g g
forageunitofadominantvegetation15andthe correction factor defined by ωk. This factor
total forage units (qk) produced by each k indicates the probability of the farm k being
g
farm (Ovando et al. 2015). We assume that usedforlivestockgrazinginthefuture(hence,
pej and qk wouldremainconstantatthefarm 0≤ωk ≤1).16
g g
level over the forest rotation and would
change only in the event of forestry activity CarbonSequestrationinShrubBiomass
abandonment in a forestunit i:
Carbon sequestration in shrub biomass is
δS−s estimatedusingPasalodos-Tatoetal.’s(2015)
EAi =(cid:2)T δt⋅Yek+φ⋅ ⋅Yek+(1−φ) functions that relate shrub biomassgrowthto
g t=s 1−δ thefractionofshrubcanopycoverandtheav-
( erage height of shrub formations. Net carbon
⋅ δS−s⋅(cid:2)tU=−Ss−s((1−αt)⋅Yek−αt⋅Ylk) sequestrationbyshrubgrowthfurtherconsid-
ers potential CO withdrawals due to forest
δU−s ) 2
+ ⋅Ylk , [3] fires and shrub clearing. The spatialinforma-
1−δ tiononthevariablesusedtoestimatenetcar-
bon sequestration in shrub biomass is taken
where from D´ıaz-Balteiro et al. (2015), for both
shrubformationsunderthetreelayerandtree-
S=T+1;U=S+1+τ;αt=t/U: less shrublands. It is assumed that the shrub
Yek=ωk⋅(pej⋅qk−cmj);Ylk=ωk⋅(plj⋅qlj−cmj). vegetation would maintain its current carbon
g g g g g g
stock and growth ability at each site in the
Thefirsttermofequation[3]referstotheas- future,exceptintheeventofforestryactivity
set value of grazing resources for the present abandonment.
inventory until the trees reach their rotation Forestry abandonment would imply, in
age. The second term of equation [3] repre- most cases, changing the present fraction of
sents the grazing resources value for an infi- shrubcanopycover.Ourestimationsconsider
nite sequence of forest rotations of the same a set of scenarios concerning forest species
speciesandsilviculturalmodelinaforestunit and silvicultures that define the maximum
i, if the regeneration investment takes place fraction of land covered by shrub in a transi-
(φ=1). We expect that theforestryabandon- tion period τ after forestry abandonment.17
ment scenario (φ=0) would lead to changes Weassumealineartransitionforshrubcarbon
in the provision of grazing resources units.
The variables qlj and plj define, respectively,
g g
the quantityandpriceofgrazingresourcesin
thatscenario,whichweassumetobeequalto 16Inthecasethatgrazingresourcesarecurrentlycon-
those observed in farms dominatedbyshrub- sumedbylivestockonafarm,ωk wouldtakeavalueof1.
Alternatively,thisprobabilitywouldrepresenttheaverage
landineachoneoftheAndalusianprovinces.
share of farms that are currently being used for livestock
grazingaccordingtotheirdominantvegetationandprovince
(see the online supplement for more details, available at
15TheclassificationoffarmsforestimatinggrazingEA http://le.uwpress.org).
valueconsidersthevegetationthatoccupiesthelargestpart 17See the online supplement, available at http://le.
ofthefarm. uwpress.org.
94 LandEconomics February2017
growth from the present situation to the one aryconditions)andwouldchangeonlyinthe
expected50yearsafterforestryabandonment. event of forestryabandonment.
The estimation of the associated EAvalue The abandonment scenario would imply
followsequation[3],butwereplacetheprice variationsintheestimatedforestwaterdueto
variablesofthisequationwithasinglecarbon changes in the forest species distribution and
price,p .Likewise,wesubstitutetheequation thefractionoftreecanopycover.Weestimate
c
[3] quantity variables with qk and ql), which water EA values using an equation similarto
c c
represent the annual net carbon sequestration equation[3]butreplacingthepriceandquan-
in shrub for the forest regeneration and for- tityvariableswithasingleandconstantwater
estry abandonment scenarios, respectively. environmental price (pw) and constant quan-
After the transition period it is assumed that tities of economic water flows for the forest
qcl remains constant over time. regeneration (qwk) and forestry abandonment
scenarios(ql ), respectively.
w
ForestWater
SilviculturalModelsandTreeSurvival
Water flow figures come from Beguer´ıa et
ProbabilityFunctions
al. (2015) and are based on numerical simu-
lations of the forest water balance on hydro-
The EA model considers a set of 19 sim-
logicalresponseunits(HRUs)in44reservoir
plified silvicultural models applied to seven
catchmentsinAndalusia.Thesimulationuses
differentspeciesthatreproducethemostcom-
dailyhydrologicalandclimaticdataandcov-
monforestrypracticesinAndalusia.18Thein-
erstheperiod2000–2009.Precipitationwater
formation provided by the silvicultural mod-
(and superficial springs in some cases) con-
els allows the estimation of individual tree
stitutestheinputofwatertoeachHRUthatis survival functions. These functions specify
transformedbyforestlandintodifferentwater the survival probability (πij, where 0≤πij≤
t t
output flows. Forest water can be either con- 1) of a tree that belongs to a species i and a
sumed within the HRU by the vegetation silviculturalmodeljateachoneofthetyears
(evapotranspiration flow) or exported out of of the forestrotation(Tij).Thisprobabilityis
t
the HRU (surface discharge and deepaquifer affected by scheduled tree thinning and final
recharge flows). logging (αij),19 natural tree mortality (jij),
t t
Inthewatereconomicsliterature,bluewa- andalsobyforestfirerisk(ρij).Thevariables
t
ter usually defines the fresh surface and αij, ρij, and jij represent the annual proba-
t t t
ground water (i.e., water in rivers, lakes, and bilitiesthattreeswillbefelled,burned,ordie,
aquifers), while the water that is temporarily respectively:
storedinthesoilstobeeventuallyevaporated
ortranspiredbytheplantsistermedgreenwa- (cid:3)1−αitj−ρtij−jtij, if t=1
ter.Afractionofbluewaterflowscanbereg- πitj= πitj−1−αitj−ρitj−jitj, if t>1, [4]
ulatedbythewateragency(collectiblesurplus
where t={1,2,...,Tij}.
offorestwater)andlaterbesoldtotheusers.
The forest water with an economic value is
The individual tree felling probability at
thus made up of the superficial water runoff
eachperiodtisquantifiedastheratiobetween
that reaches a reservoir in Andalusia and is
h, the number of trees that the silvicultural
allocatedbetweenthefinalusers(Beguer´ıaet t
modelsdeterminewillbefelledinthatperiod,
al. 2015).
and N , the initial stand density according to
Estimationsoftheforestwaterbalancede- the si1lviculture model: (αij= h/N ). The
pend on, among other factors, soil and cli- t t 1
mortality ratios are estimated as logarithmic
matic conditions, the distribution of oaks,
conifers,andotherforestspecies,andthefrac-
tionoftreecanopycoverwithintheHRU.We
18See the online supplement for details, available at
assume that the average estimates of eco-
http://le.uwpress.org.
nomic forest water for the period 2000–2009
19Notethattheconditionalprobabilityoftreelogging
would remain constant in the future (station- βdtofequation[1]isestimatedas:βdt=αdt⋅πdt.
93(1) Ovandoetal.:SpatialValuationofEnvironmentalAssets 95
functions of tree age, while the future risk of finedintermsofspeciescomposition,density,
forestfiredependsupontheaveragehistorical age class distribution, slope gradient, the sil-
forest fire ratios by species and province as- vicultural model assigned, and the quality of
sessed for the period 1987–2006 (D´ıaz-Bal- the site for growing timber or cork.
teiroetal.2015).Therotationagebyspecies The EA valuation model is developed in
and silvicultures is exogenously defined by Matlab R2014a.21 Figure 1 shows a scheme
Montero et al.’s(2015) models.20 of the interrelated components of this com-
Asurvivalprobabilitymatrixhasadimen- puting model and the sources of biophysical
sion Tij× Tij and computes the conditional and economic data.
probability(πij)thatatreeofanagedisalive
dt
at each one of the tree ages t that are to be III.RESULTS
reached (d<t): πij =Pr(d/t)=πij/πij.
dτ t d
TheEAmodelincludes152differentprob- EAEstimationsforProvisioningand
ability matrices, one for each of the 19 silvi- RegulatingServices
culturemodelsandeachoneoftheeightprov-
inces of Andalusia. These matrices are used AverageValuesatFarmandVegetationLevels
to simulatetheevolutionofforestsandcould
Table 1 shows the estimated EA value (in
be applied to any initial condition, which is
euros per hectare) by forest species, ESs, sil-
definedbythedistributionoftheexistingtrees
viculturalmodelforthemainscenario,thatis,
byspeciesandageclassinaforestunit.Initial
with a discount rate of 3% and the average
forest inventories and other spatial variables
prices of 2010. This is the main scenario, al-
at the farm level were estimated for the pol-
thoughweanalyzethesensitivityofresultsto
ygons of the Spanish Forest Map using the
discount rates in the range of 2% to 6% and
latestNationalForestInventory(IFN3)inAn-
variations in net benefits, due to changes in
dalusia (MARM 2013) and digital elevation
prices from (cid:3)25% and (cid:3)50%.
maps.
The aggregated EA value of silvopastoral
The IFN3 data were gathered between
provisioning services, water, and carbon
2006and2008inAndalusiaandwereupdated
amounts to 2,813 €/ha (standard devia-
tothebeginningof2010,consideringspecies
tion=2,383 €/ha), on average for the 567
and site-specific growth function (see D´ıaz-
farmsincludedintheanalysis.Corkandgraz-
Balteiro et al. 2015 for details). Those vari-
ing resources represent 9% and 28% of this
ables were assigned to the 567 farms accord-
average EA value, respectively; carbon con-
ing to the weighted average values by forest
tributes39%(58%ofwhichisduetotreenet
speciesandsilviculturalmodelatthemunici-
growth,and42%duetoshrubnetgrowth)and
pality level.
water 23% of this value. Timber, pinenuts,
Thereasonforassigningvaluesatthemu-
and firewood account for the remaining 1%.
nicipalitylevelisthatweignorewhichSpan-
TheEAvaluesoftimberandpinenutsdis-
ish Forest Map polygons correspond to each
playahighervariabilityamongthefarmsthan
farm; rather we observe the municipality in
anyotherES.Variabilityincorkvaluesacross
whichthefarmislocatedanditslandusedis-
farmsisalsolarge,whilethevariabilityofre-
tribution(asstatedbyfarmowners).Thefarm
sults in terms of grazing resources is small.
area is shared out into a set of homogeneous
The relative homogeneity in EA values of
forest units that represent the distribution of
grazingresourcesisduetothefactthatavail-
theforestinventoriesandsilviculturalmodels
abledataongrazingleasingpricesdifferonly
ofprivatelandsinthemunicipality,whilethe
bydominantvegetationandprovince,without
area covered by each forest species, shrub-
connection to other spatial factors. The vari-
land,grassland,andotherlandusesisspecific
ability of the EA value associated with the
tothefarm.Ahomogeneousforestunitisde-
21Matlab is available from MathWorks (www.
20See the online supplement, available at http://le. mathworks.com),andthespecificcodeusedwasdeveloped
uwpress.org. bytheauthors.