Measuring contributions to economic production—use of an Index of Captured Ecosystem Value

Ecological Economics 41 (2002) 479–490

SPECIAL ISSUE: The Dynamics and Value of Ecosystem Services: IntegratingEconomic and Ecological Perspectives

Measuring contributions to economic production—use ofan Index of Captured Ecosystem Value

Kent Gustavson a,*, Stephen C. Lonergan a, Jack Ruitenbeek b

a Department of Geography, Uni�ersity of Victoria, PO Box 3050, Victoria, Canada, BC V8W 3P5b H.J. Ruitenbeek Resource Consulting Limited, RR c2, Site 52, C21, Gabriola, Canada, BC V0R 1X0

Abstract

The production of an economic good derived from a renewable natural resource base involves the extraction ofecosystem function values as represented by the contribution made to production by the originating ecosystem. Theartisanal mixed-species fisheries of Jamaica is used as a case study in the development of a biophysically based indexto account for captured ecosystem values (or embodied ecosystem values) and an examination of the extent to whichthose values are proportionately reflected in monetary exchange values. The Index of Captured Ecosystem Value(ICEV) is developed from a basis in information theory relevant to an analysis of network flows in ecosystems.Technical coefficients, describing the production relationship between ICEV values and market values of catchesassociated with individual fishing efforts in Jamaican fisheries, reveal that captured ecosystem function associated withfisheries using distinct technologies (i.e. China net, trap, hand line, palanca and speargun) are valued differently bythe market. This surplus value is rooted in the observation that certain fisheries target species that are more connectedwithin the coral reef food web than those species typically captured by other fisheries. Consideration of thebiophysically based contributions of coral reef ecosystems to fisheries production reveals distortions between marketand supply-side values, indicating that the role of ecosystems is not being consistently treated. Comment and directionis offered regarding the development of indices of ecosystem function or value that can be applied to policy questionsconcerning the extraction of a renewable natural resource. © 2002 Elsevier Science B.V. All rights reserved.

Keywords: Index; Indicators; Ecosystem function; Food web; Ecosystem value; Supply-side value

This article is also available online at:www.elsevier.com/locate/ecolecon

1. Introduction

Economic valuation of the environment is con-cerned with estimating the monetary value of the

flow of ecosystem services, where ecosystem ser-vices are taken to include all goods, services orother benefits derived from natural capital. deGroot (1994) defined ecosystem functions as theyrelate to economic values as ‘‘the capacity ofnatural processes and components to providegoods and services that satisfy human needs (di-rectly and/or indirectly)’’. Services from suchfunctions include materials, energy and informa-

* Corresponding author. Present address: Gustavson Eco-logical Resource Consulting, PO Box 115, Gabriola, Canada,BC V0R 1X0. Tel.: +1-250-247-9734; fax: +1-250-247-9735

E-mail address: [email protected] (K. Gustavson).

0921-8009/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved.

PII: S0 921 -8009 (02 )00095 -2

mailto:[email protected]

K. Gusta�son et al. / Ecological Economics 41 (2002) 479–490480

tion used to contribute to human welfare (e.g.Costanza et al., 1997). There have been manyattempts to place a value on ecosystem services.The list of potential services is long and varied,including: generating and maintaining soils; main-taining hydrological cycles; regulating gas; nutri-ent cycling and storage; assimilation andelimination of pollution and wastes; regulatingdisturbances; biological control; pollination ofcrops; food production; maintaining species andgenetic resources; maintaining biogeochemical cy-cling; and, regulating weather and climate (Folkeet al., 1994; Bingham et al., 1995; Myers, 1996;Costanza et al., 1997).

The values perceived by humans and the prefer-ences expressed in the market system, or throughother monetary valuation means, may not takeinto account what is necessary or relevant forecosystem integrity for the maintenance of naturalecosystems and their associated services, nor doesit necessarily take into account the co-evolution-ary and complex nature of the environment-econ-omy system (Holling et al., 1995). Yet theproduction of goods and services by the humaneconomy relies on the drawing of services pro-vided by natural capital and, thus, on the produc-tivity of natural biotic systems. The drawing ofservices provided by natural capital may be de-scribed as a process that captures or ‘embodies’function value of the originating ecosystem. It isthrough this involvement of natural biota in eco-nomic production processes that contributions bythe originating ecosystem are made to the value ofthe final economic goods and services (i.e.through a value-added process of which the natu-ral biota is a part or, alternatively, as a form ofcontributory value) (Ulanowicz, 1991).

Isard (1969, 1972) presents a conceptual frame-work for the analysis of the interactions or link-ages between economic and ecological systems(Fig. 1). It is based in a traditional commodity-by-industry input-output modeling framework de-scribing linear interactions between sectors withinthe economic system (e.g. Miller and Blair, 1985),but extends the matrix to include ecological com-modity inputs and outputs to economic activities,economic commodity inputs and outputs to eco-logical processes, and ecological commodity in-

puts and outputs to ecological processes. Each cellentry in the larger matrix represents an input of acommodity into (a positive cell entry) or use of acommodity by (a negative cell entry) a process oractivity. Despite drawbacks, the framework isparticularly attractive in that it does allow for the‘‘…systematic description in relationships andmagnitudes as they exist at a point in time. Thisdescription in itself often provides much insight’’(Isard, 1972, p. 95). Some work has been doneexplicitly applying input–output analysis to thestudy of ecosystem structure and function (e.g.Hannon, 1973; Finn, 1976).

The lower right quadrant (Fig. 1) representsinteractions within the ecological system and isanalogous to what an ecologist envisions throughthe definition of food webs, with food webs de-picting the flow of ecological commodities be-tween trophic species or groupings. Positive cellentries in the upper right quadrant represent out-puts of ecological processes provided as resourcesor products directly as economic commodities.Negative cell entries in the lower left quadrantrepresent the use of ecological commodities withineconomic processes. Whether the product of anecosystem process is recorded as an output of anecological process (economic commodity, upperright quadrant) or as an input to an economicprocess (ecological commodity, lower left quad-

Fig. 1. Framework for the analysis of ecological economicinteractions (adapted from Isard, 1969, 1972).

K. Gusta�son et al. / Ecological Economics 41 (2002) 479–490 481

rant) involves an arbitrary decision on behalf ofthe researcher (Isard, 1969).

The biophysical value of products of naturalecosystems as ecological commodities, expressedas some measure of ecosystem function valueper unit of economic production value within agiven economic sector, defines technical coeffi-cients within the ecological economic model:

aij=Zij

Xj

(1)

where, aij, technical coefficient defining the pro-duction relationship between the ecological com-modity input i (biophysical units) and theeconomic sector j (dollars); Zij, flow of ecologi-cal commodity input i to economic sector j (bio-physical units) and, Xj, gross output value ofeconomic sector j (dollars).

If one is concerned with the direct capture ofbiophysical ecosystem values as represented bytechnical coefficients, what may be used as anindex of ecosystem function value for use in Eq.(1)? Certainly, it would be possible to rely on asimple measure of biomass for all Zij, yet itwould be more advantageous to adopt a mea-sure that more explicitly and consistently acts asa proxy for ecosystem functions that gave riseto the commodity.

There is a growing body of ecological workthat continues to examine the relationship be-tween traditional measures of biodiversity andecosystem function (e.g. Ehrlich and Ehrlich,1981; di Castri and Younes, 1990; Grassle et al.,1991; Schulze and Mooney, 1994; Lamont, 1995;Bengtsson et al., 1997). The evidence is oftencontradictory and, if anything, seems to point inthe direction that the most often relied uponmeasures of biodiversity may not be adequateindices because the relationship is often func-tional process specific, ecological system specific(in space, time and scale), and dependent on thedefinition of biodiversity employed. About allthat can be generally said about the role of bio-diversity in relation to function is that it pro-vides the ‘‘…medium for energy and materialflows, which in turn provide ecosystems withtheir functional properties…’’ (Myers, 1996).

Currently, there are no widely accepted in-dices of ecosystem function per se. Some prelim-inary work in that direction, however, is evidentfrom the literature. Indeed, there appears to bea growing movement by many ecologists to-wards a focus on finding general, measurableattributes that characterize the functioning ofecosystems (e.g. Friend and Rapport, 1991;Holling, 1992; Holling et al., 1995; Martinez,1996). For example, Martinez (1996) advocatesa focus on the conceptualization of biodiversityitself as ‘‘the spatial and temporal variability ofthe structure and function of living systems’’.This definition explicitly recognizes the impor-tance of establishing proxy measures for ecosys-tem function.

There have been efforts to define and measureresilience (e.g. DeAngelis, 1980; Ives, 1995; Lud-wig et al., 1997; Neubert and Caswell, 1997),but ecological theory has yet to advance a gen-eral predictable or supportable theoretical rela-tionship between properties of resilience andfunction. Much has been debated about the pos-sible relationships between stability and struc-ture (e.g. MacArthur, 1955; May 1972, 1973,1975; Pimm, 1982, 1984; Hall and Raffaelli,1993). Indeed, a general and predictable rela-tionship between properties of resiliency andhow ecosystems function is not intuitively obvi-ous and may not exist. Stability statementsalone tend to be specific to the ecological situa-tion under consideration and cannot be assumedto apply to the larger, more general case(Grimm and Wissel, 1997).

Where else can one turn to define an index ofecosystem function? What is ideally required foruse in Eq. (1) is a general proxy measure of thefunctional characteristics of species within acommunity which is supported by theory, calcu-lable using a limited dataset and informationbase, and applicable to a wide array of ecosys-tems. An Index of Captured Ecosystem Value(ICEV) is derived relying on relatively simpledescriptions of the structure of the food web ofinterest and the nature of the trophic relation-ships between its elements. Information theorywill provide the basis for the analysis.


2. Methods

It is reasonable to accept that ecological pro-cesses and ecosystem function are synonymousterms (Lawton and Brown, 1993) or that ecologi-cal functions are more generally interactions withecological processes (Martinez, 1996). Functionaldiversity may be quantified from the type andnumber of functional groups represented in anecosystem and the nature of their interactions.Food webs are simply depictions or models ofhow species in a community interact (e.g. Pimm,1982; Lawton, 1989; Yodzis, 1989; Hall and Raf-faelli, 1993), in which interactions with processesare as either consumers or resources, and thus canserve as the basis for the derivation of an index ofecosystem function.

The development of the index will be ap-proached initially from a basis in informationtheory (e.g. Kullback, 1968). Ulanowicz (1986),Ulanowicz and Norden (1990) outline the theorybehind measures of uncertainty and informationcontent relevant to the analysis of network flowsin ecosystems, building on the previous work ofMacArthur (1955), Rutledge et al. (1976),Ulanowicz (1980), Hirata and Ulanowicz (1984).

Following Ulanowicz (1986), the uncertaintyregarding a specific outcome of an event can bedescribed by

Hi= −K log(pi) (2)

where, Hi, the uncertainty associated with out-come i ; pi, the probability of outcome i ; and K, aconstant (imparts the physical dimensions or ‘size’to the index) (Ulanowicz, 1986).

Where a suite of outcomes are possible, one canestimate the average uncertainty as

H= −K �i

pi log pi (3)

Information concerning events represents a re-duction in the uncertainty associated with predict-ing the outcomes. Thus, one can similarly definethe average gain in information concerning a suiteof possible outcomes as

I=K �i

pi log� pi

p i*�

(4)

where, pi*, the probability of outcome i based onan initial assumed distribution; and pi, the proba-bility of outcome i based on new information.

The linking of events is an organizational char-acteristic of network flows. Applying informationtheory principles to network flows, one can de-scribe the information concerning the outcome ofan event provided by knowledge of a linked orrelated event as

I=K log�p�bj �ai�

p�bj�n

(5)

where, p�bj �ai�, the probability of bj given that ai

has occurred; and p�bj�, the probability that bj

will occur.One can then describe the average mutual infor-

mation as presented by Ulanowicz (1986) as

A=K �i

�j

p(ai, bj)log�p�bj �ai�

p�bj�n

(6)

When applied to a network, Eq. (6) measuresthe information of the flow structure, or specifi-cally, ‘‘…how well, on the average, the networkarticulates a flow event from any one node toaffect any other specific locus’’ (Ulanowicz, 1986,p. 97). To apply this to an open ecosystem andensure that the solution is symmetrical with re-spect to the consideration of either compartmentinputs or outputs, one first defines the matrixdescribing the flows from the ith compartment tothe jth compartment, Tij, where i and j range from0 to n+1 (0 refers to exogenous inputs to thesystem, while n+1 refers to losses or exportsfrom the system) (Ulanowicz and Norden, 1990).The average mutual information (Ulanowicz andNorden, 1990) then becomes

I= �n+1

i, j=0

fij Qi log� fij

�k

fkj Qk

�(7)

where, Qi=�k Tik/�l,m Tlm, the estimated proba-bility that flow passes through compartment i,and; fij=Tij/�k Tik, the fraction of total flowthrough compartment i that also passes throughcompartment j (or the estimated conditional prob-ability that flow will pass through j given that ithas passed through i ).


Ulanowicz (1980, 1986), Ulanowicz and Nor-den (1990) scale the measure of the average mu-tual information (I) by the size of the total flowthrough the system:

A=TI (8)

where, A, ascendancy; and T, total systemthroughput, or �i, j Tij.

System ascendancy, as a measure of the magni-tude of the information flow through the networkstructure of an ecosystem, is effectively an indexof ecosystem function. Moreover, it has manyuseful advantages, such as the ability to be ap-plied to open, far-from-equilibrium systems (Hi-rata and Ulanowicz, 1984; Ulanowicz, 1986).However, its use does demand a relatively com-plete description of the nature and magnitude ofthe interactions among all species. Furthermore,there are the additional complexities associatedwith providing realistic and appropriate al-lowances for temporal and spatial variations insuch relationships (Pahl-Wostl, 1995). The devel-opment of mass-balance models (e.g. the ECO-PATH models) (Christensen and Pauly, 1992)may provide a rich information base, but dorequire specific assumptions regarding biomasses,rates, and food web relationships. As it is theintention of this study to derive an index ofecosystem function which can be applied with arelatively limited knowledge of the ecosystem inquestion, we chose to ‘back away’ from informa-tion demands associated with using the averagemutual information or system ascendancy con-cepts strictly as described above.

Let us assume that one does not have sufficientinformation regarding either the size of the systemthroughput or the associated probabilities of in-terspecific interactions for a given food web, butdoes have knowledge regarding the species ortrophic species involved and whether or not inter-specific interactions may occur between any pairof species. In other words,

fij=1li

(9)

Qi=liL

(10)

�k

fkj Qk=ljL

(11)

where, li, the total number of species i forwardlinks with other species (i.e. the total number ofspecies or trophic species which use species i as aresource or as prey); lj, the total number of speciesj backward links with other species (i.e. the totalnumber of species or trophic species which speciesj uses as a resource or as prey); and L, the totalnumber of links between all species in the set.

Since one is assuming that we do not havesufficient information regarding the probabilitiesassociated with any pair of interactions, all inter-specific interactions are given an equal probabilityof occurring as long as a link is known to exist.Thus, assuming maximum uncertainty regardingthe strength of interspecific interactions, Eq. (7),which defines the average mutual information,becomes

I= �n+1

i, j=0

�1li

�� liL�

log�1/li

lj/L�

(12)

which reduces to

I= − �n+1

i, j=0

�1L�

log�li lj

L�

. (13)

Eq. (13) thus defines the average gain in infor-mation regarding the occurrence of interspecificinteractions or flows through the given food web,given that one has knowledge of the interactionsthat do occur. This can be contrasted to the casein which there is no knowledge of the nature ofthe interactions, and there is simply uncertainty orcomplexity associated with a given set of species(e.g. Ulanowicz, 1986).

Now consider predator compartment j and preyor resource compartment i. The economic extrac-tion of a species from compartment j is ultimatelymade possible through flow from the set of com-partments consisting of R to which compartmentj is backward linked. The mutual informationconcerning the flow to compartment j givenknowledge of the existence of the link with anyone compartment i is

− log�li lj

L�

. (14)


Fig. 2. Conceptual trophic interactions model. The sum of the number of all forward links from prey species i is represented by li.The sum of the number of all backward links from predator species j is represented by lj. The set of all i species utilized by speciesj defines set R. For example, Ra is the set of all species i to which ja is linked. The set of all j species extracted from the environmentdefines the set P.

One can then subsequently consider the mutualinformation for all pairs of compartments i and jlinked, considering only the set of compartmentsconsisting of P from which natural resources areextracted, with each compartment j within the setP having a unique set R of species which it utilizesas a resource. A conceptual example of this modelis shown in Fig. 2.

In allowing the reduction of the average mutualinformation index to Eq. (13), however, we as-sumed no knowledge concerning the size of thenetwork flows. The discussion so far similarlydoes not consider the magnitude of the economicextraction of the natural resource. In order tofactor in the magnitude of this extraction, thetotal biomass (B) extracted in a given period istreated similarly to Ulanowicz (1980, 1986),Ulanowicz and Norden (1990) consideration oftotal system throughput. Thus, using Eq. (13), wearrive at the ICEV:

ICEV= −B �i, j

� 1Lij

�log

�li ljL�

(15)

where, Lij, the total number of links between all i,jpairs considered, in which all i,j pairs are thoselinked within the food web structure for whichcompartment j is utilized or extracted as a naturalresource (i.e. within set P).

To apply the index, it is necessary to develop atypology of functional species groups based onthe variety of interactions with ecological pro-cesses. At the most disaggregated level, individualtaxonomic species or ontogenetic stages withinany one species may be distinguished as func-tional groups. More relevant for food web studies,however, trophic species are identified (Briandand Cohen, 1984; Pimm et al., 1991).

As a case study to illustrate application of theICEV, the existence of surplus value within Ja-maican artisanal mixed-species fisheries is investi-gated. Information regarding the food webstructure and species composition of Caribbeancoral reefs is widely scattered. Information spe-cific to Jamaica is limited—some communitycomposition information exists, but is focused onresolving patterns for sessile benthic species (e.g.Goreau, 1959; Goreau and Goreau, 1973; Liddelland Ohlhorst, 1981, 1987; Liddell et al., 1984).However, Opitz (1996) was able to construct asteady-state model for a generalized Caribbeancoral reef ecosystem, in the process compiling ataxon interaction matrix based on assumed dietcompositions derived from the numerous litera-ture sources with an emphasis on fish species.Opitz (1996) used the comprehensive study of fishdiets by Randall (1967) for the US Virgin Islands,supplemented with fish species listings provided in


Fischer (1978), as the ‘scaffold’ to develop thematrix which identified the fish taxa and keyflora, invertebrate and vertebrate taxa on whichthey directly depend.

Trophic species were identified from the dietcomposition matrix presented in Opitz (1996).Before the analysis was applied, all non-zero in-teraction coefficients in the diet composition ma-trix were recoded to 1, thus recordingpredator–prey relationships as either ‘on’ or‘off’. It was felt necessary to generalize the ma-trix in such a way because: (i) Opitz (1996) hadto employ fairly rough estimation techniques toarrive at specific coefficient values, at times whenonly qualitative data were available regardingdiet; and, (ii) the quantitative data for the matrixwere based using information largely from Ran-dall (1967) for the US Virgin Islands. It was notreasonable to maintain the precision of the inter-action coefficients as presented in Opitz (1996)and apply the data to model present-day Ja-maican coral reefs given the current lack of in-formation regarding the food webs of Jamaicanreefs or other regions of the Caribbean.

For five Jamaican coral reef-based artisanalfisheries (China net [gill net], trap, hand line,palanca [bottom set long-line], and speargunfishing), the capture of ecosystem function valueas measured by the ICEV and the existence ofsurplus value were explored using data from1996 catch and effort surveys conducted by theFisheries Division (Ministry of Agriculture, Gov-ernment of Jamaica, unpublished data). Individ-ual survey samples were eliminated from the setif they failed to record the composition of indi-vidual catches by genus (at a minimum) and byweight. Non-finfish species were ignored for thepresent analysis. Typical market prices are esti-mates of the mean price obtained for the year1996 (Fisheries Division, Ministry of Agriculture,Government of Jamaica, unpublished data).

The technical coefficient associated with eachfishery is represented as

aij=ICEVij

Vj

(16)

where, aij, technical coefficient defining the pro-duction relationship between the ecological com-

modity input i (ecosystem function value asmeasured by the ICEV) and the fishery j (Ja-maican dollars); ICEVij, flow of ecological com-modity input i to fishery j ; and Vj, market valueof fishery j (Jamaican dollars).

For each finfish taxon represented in the catchand effort surveys, the market price for the year1996 and the ICEV value (Eq. (15)) per kg, aswell as the corresponding technical coefficient(Eq. (16)) associated with each individual taxon,was calculated. Subsequently, the technical coeffi-cient (Eq. (16)) was calculated for all individualfishing efforts by fishing technology, consideringthe total ICEV and monetary value of each bun-dle of fishes. In order to explore whether or notfisheries using distinct technologies (i.e. Chinanet, trap, hand line, palanca and speargunfishing) led to ecosystem function value as mea-sured by the ICEV being equally valued by themarket (i.e. is there surplus value?), a non-para-metric Kruskal–Wallis (ANOVA by ranks) testwas conducted to determine if the mean technicalcoefficient of individual fishing efforts within anyone fishery was significantly different from anyother fishery. We initially look simply for differ-ences between the fisheries, lacking a strong apriori basis on which to justify a stronger direc-tional test, and keeping in mind our primarygoal of illustrating an application of the ICEVwithin the context of a production relationshipdefined by technical coefficients. All statisticalprocedures were conducted using SPSS® for Win-dows (Release 7.5.1, 1996) statistical softwarepackage.

3. Results

The identification of trophic species resulted inthe reduction of the 250 original taxonomicgroups as represented in the diet compositionmatrix of Opitz (1996) to 246 groups. Only fourpairs of species of fishes shared both the sameset of predators and prey as recorded in thematrix: (i) Diodon holocanthus (long-spine porcu-pinefish) and D. hystrix (spot-fin porcupinefish)from the family Diodontidae; (ii) Gramma loreto(royal gramma) and G. melacara (blackcap


basslet) from the family Grammitidae; (iii)Kyphosus incisor (yellow sea chub) and K. secta-trix (Bermuda sea chub) from the family Kyphosi-dae; and, (iv) Pomacanthus arcuatus (greyangelfish) and P. paru (French angelfish) from thefamily Pomacanthidae. All other trophic speciesare as listed in Opitz (1996). For the completetrophic interaction matrix, the total number oflinks in the community (L) was 3317.

For each of the five coral reef-based fisheriestechnologies examined in this investigation, sum-mary statistics of the values of the technical coeffi-cients calculated using the results of individualfishing efforts are shown in Table 1. The results ofthe non-parametric Kruskal–Wallis (ANOVA byranks) test were highly significant (�2=106, df=4, P�0.001), indicating that fisheries based indifferent technologies capture different levels ofecosystem function value per dollar of catch. Inother words, surplus value is created. This surplusvalue is rooted in the observation that certainfisheries themselves rely on species representingthe capture of a greater or lesser amount ofecosystem information value— in other words,fishing species which are more or less connectedwithin the food web than is the case with otherfishing technologies, and these differences notbeing reflected in market prices. This interpreta-tion is supported by an analysis of the averagemutual information associated with individualspecies captured by each fishery (Kruskal–Wallistest, �2=80, df=4, P�0.001), indicating thatthe captured ecosystem value is fundamentallydifferent among fisheries utilizing differenttechnologies.

4. Discussion

It is evident that Jamaican artisanal coral reeffisheries based in the use of distinct technologiescapture ecosystem information values that are not,to the same extent, proportionately reflected in themarket values obtained for the catch. Certainfisheries appear to exploit species that are directlyand indirectly more intimately connected within thecoral reef food web as measured by the ICEV, andthus rely on the capture of greater ecosystemfunction value for each unit of monetary value asmeasured by the market. Further detail regardingthese patterns will require subsequent analyses,likely requiring an expanded dataset with improvedrepresentation by certain technologies (Table 1).

This study was primarily interested in derivingand exploring the application of an alternativetheory of value which can be applied to theextraction of a renewable natural resource. Thisanalysis demonstrates that if one considers bio-physical values (i.e. the biophysical contribution ofthe coral reef ecosystem to fisheries production),distortions between market values and supply-sidevalues become evident. What does this creation ofsurplus value mean? In effect, we are not valuingthe ecosystem contributions to the creation of finalproducts to the same extent within a given sector,or perhaps between sectors of the economy. We arenot consistently considering ecosystem functionvalues. But interpretations necessarily becomemore complicated as we examine differences acrossmarkets and over time because of changes inecosystem functions and market supply and de-mand conditions. This is rightfully the subject offurther research.

Table 1Summary statistics for technical coefficients (J$−1) associated with China net, trap, hand line, palanca and speargun fishing inJamaica for the year 1996

Minimum valueS.D.Fishery n Mean value Maximum value

47 0.00698China net 0.00104 0.010880.005290.016590.00407279Trap 0.001900.00887

31 0.00683Hand line 0.00137 0.00507 0.0102716 0.00791Palanca 0.004640.0007960.00632

0.015110.005430.00220Speargun 0.0077120393Total 0.00832 0.00197 0.00407 0.01659


The theoretical development and use of indica-tors and indices is still very much an art ratherthan a hard science, with much room left forfuture developments and improvements. The re-mainder of this discussion will be devoted toproviding comment and direction with that inmind.

4.1. Future directions for index de�elopment

The food web model used for the derivation ofthe ICEV is static. It does not consider or accom-modate dynamic changes to the structure or func-tioning of the ecosystem that may occur througheither human-induced or natural influences. Itmay indeed be possible to modify the structure ofthe ICEV or the method of its calculation toaccommodate food web dynamics, yet to do sowould require a level of scientific knowledge con-cerning the ecosystems in question that may notbe currently sufficient.

In addition to not accounting for ecosystemdynamics, the methodology applied in this analy-sis did not make any explicit distinctions betweenthe ways in which the functioning of ecosystemsof different ‘types’ may systematically vary. Inother words, it would be useful to explore theapplication and behavior of the ICEV as calcu-lated for the extraction of a renewable naturalresource from different ecosystems as representedby distinct food webs.

A pronounced shortcoming of the present anal-ysis is the inability to take into account thresholdsor criticality. Questions concerning the appropri-ate level of exploitation for sustainable use havedominated natural resource economics and bio-logical studies of exploited populations; however,the question before us was not the level of ex-ploitation possible or the fate of the exploitedpopulations, but measuring the contribution tothe value of the economic product. However, itmay indeed be fruitful to explore the incorpora-tion of bioeconomic principles directly into theconstruction and use of an index of ecosystemfunction.

Meredith et al. (1994) discuss thresholds ofenvironmental criticality as both an objective con-struct, as is the case concerning biophysically and

economically defined standards, and a subjectiveconstruct, which may often be the case concerningsocially defined standards. They emphasize thepossibility of ‘multiple anthropocentric assessmentperspectives’ as determined by the particular setof personal values. The ICEV was more objec-tively defined based on ecological and economicprinciples. However, the selection and definitionof the indicator necessarily involved a value judg-ment concerning what was to be measured.Broadening the indicator to consider criticalitieswill increase the subjective content inherent in theindicator construction and use simply because thedimensionality of the indicator will be increased,requiring further value judgments to be made,often where there is a lack of scientific evidence towhich one can appeal. This will not compromisethe validity of the indicator per se, but will requireprudent documentation and justification for thevalues selected.

It would also be potentially fruitful to explorethe junctions between an ecosystem-based appli-cation of information theory, as done in thisstudy, and economic information theory models.Inadequate information and risk as introducedinto economic decision-making models is throughconsideration of either endogenously modeledmarket uncertainties or exogenously modeledevent uncertainties (e.g. Hirshleifer and Riley,1979, 1992). The economics of information, asrelevant here, involves the process of individualeconomic agents attempting to gain informationregarding exogenous event uncertainties in orderto assist in the decision-making process. TheICEV represents a gain in information regardingthe contribution of a natural ecosystem to pro-ductive value and, in essence, reduces uncertaintyregarding the role of ecosystems in the economicproduction process (a role that is normally exoge-nous to neo-classical economic production mod-els). As a more specific example of howinformation economics may potentially be ap-plied, there is a significant body of work examin-ing the effect of asymmetric information betweenprinciples and agents in the formation of contrac-tual relationships (e.g. Feltham et al., 1988; Ma-cho-Stadler and Perez-Castrillo, 1997). As aspecific issue of moral hazard (a gain in informa-


tion by one party after the establishment of thecontract between fishers and management author-ities) or adverse selection (private informationheld by one of the contracting bodies), the ICEVor other aspects of ecosystem information theorymay be able to be applied to economic asymmet-ric information models as represented by conflict-ing objectives through knowledge gainedconcerning the value of ecosystem function. Itwould be interesting to formally pursue this infor-mation-adding angle as it relates to existing eco-nomic information theory and decision-making.

4.2. General issues concerning food web statistics

The ICEV requires information regarding thestructure of the food web of interest, and as suchis effectively a food web statistic. There arecaveats and limitations associated with food webstatistics, primarily stemming from the limitationsof the food webs descriptions themselves. Asstated by Lawton (1989), ‘‘In Pimm’s words(1982), most published webs are, therefore, ‘cari-catures of nature’. My dictionary defines carica-ture as a representation exaggerated for comiceffect, or a ludicrously inadequate or inaccurateimitation! This is a fair assessment of most of thepublished information on food webs.’’

There has been much debate surrounding thecalculation and use of food web statistics, particu-larly concerning connectance (e.g. Paine, 1988;Peters, 1988; Lawton, 1989; Pimm et al., 1991;Hall and Raffaelli, 1993). The calculation of theICEV and food web connectance both rely on adescription of the links between trophic species,and as such are inherently related. The criticaldebate concerning the calculation and use of in-formation theory indices to describe networkflows as it relates to many of the issues raised inthe debate concerning the use of food web statis-tics has yet to develop in the literature.

4.3. General considerations for index de�elopmentfor use in decision-making

Biological indices have been developed with theprimary purpose of being used as tools for scien-tific description and inquiry. However, use has

been extended within the arenas of governmentand management to help fulfill the need forecosystem information relevant for broader mat-ters of ecosystem management, policy and regula-tion. What may often be underappreciated,however, is that such use may not be appropriategiven the theoretical underpinnings and design ofthe specific indicator or index employed.

A useful index must be developed with a spe-cific purpose in mind. Such purpose is reflected inthe range of indicator characteristics and proper-ties, and the derived index implicitly restricts theset of appropriate applications. But generallywhat are the properties of a good indicator orindex? What are the potential uses one shouldconsider and how should these impact the indica-tor design?

The properties of a ‘good’ indicator for use inpolicy development and decision-making regard-ing the environmental management of multi-di-mensional, complex systems include:� the indicator design corresponds to the selected

purpose and application;� the value base behind the indicator is explicit;� the indicator provides a sufficient simplification

or abstraction of the system characteristics;� the theory behind the design of the indicator is

relatively robust;� the sensitivity of the indicator to system

changes has been sufficiently explored anddefined, and the indicator is sufficiently sensi-tive to meet the design purpose; and,

� The information provided by the indicator canbe understood and applied by the user.Most indicators or indices cannot be expected

to meet all of the above criteria; however, itremains the goal of indicator development tosatisfy as many as is possible given shortcomingsof the level of available scientific knowledge, lo-gistical restrictions and the demanding needs ofthe users. It is asserted that the ICEV meets manyof the criteria of a ‘good’ indicator, with the onenotable exception that the sensitivity and behaviorof the index to parameter changes remains to beexplored further.

This present study was confronted with severedata restrictions, given our overall limited knowl-edge of the functioning, structure and dynamics of


most of the world’s ecosystems. This can be ex-pected to be the norm for some time to come, yetmanagement decisions and government policy issignificantly affecting how and in what ways hu-mans are currently interacting with the environ-ment. Science needs to begin to move closer tooperational means to evaluate impacts and ana-lyze the trade-offs involved. The ICEV is pre-sented as one possible tool.

Acknowledgements

Sandra Grant (Data Manager/Analyst) andAndre Kong (Director) of the Fisheries Division(Ministry of Agriculture, Government of Jamaica)kindly provided access to unpublished govern-ment data for Jamaican artisanal fisheries fromtheir Catch and Effort Data Collection Pro-gramme. Comments from anonymous reviewerswere greatly appreciated and improved earlierdrafts. The authors take full responsibility for thecontent of this paper.

References

Bengtsson, J., Jones, H., Setala, H., 1997. The value of biodi-versity. Trends in Ecology and Evolution 12, 334–335.

Bingham, G., Bishop, R., Brody, M., Bromley, D., Clark, E.,Cooper, W., Costanza, R., Hale, T., Hayden, G., Kellert,S., Norgaard, R., Norton, B., Payne, J., Russell, C., Suter,G., 1995. Issues in ecosystem valuation: improving infor-mation for decision making. Ecological Economics 14,73–90.

Briand, F., Cohen, J.E., 1984. Community food webs havescale-invariant structure. Nature 307, 264–267.

Christensen, V., Pauly, D., 1992. ECOPATH II—a software forbalancing steady-state ecosystem models and calculatingnetwork characteristics. Ecological Modelling 61, 169–185.

Costanza, R., d’Arge, R., de Groot, R., Farber, S., Grasso,M., Hannon, B., Limburg, K., Naeem, S., O’Neill, R.V.,Paruelo, J., Raskin, R.G., Sutton, P., van den Belt, M.,1997. The value of the world’s ecosystem services andnatural capital. Nature 387, 253–260.

DeAngelis, D.L., 1980. Energy flow, nutrient cycling, andecosystem resilience. Ecology 61, 764–771.

de Groot, R.S., 1994. Environmental functions and the eco-nomic value of natural ecosystems. In: Jansson, A.M.,Hammer, M., Folke, C., Costanza, R. (Eds.), Investing inNatural Capital: The Ecological Economics Approach toSustainability. Island Press, Washington DC, pp. 151–168.

di Castri, F., Younes, T., 1990. Ecosystem function of biolog-ical diversity. Biology International, Special Issue 22, 20pp.

Ehrlich, P.R., Ehrlich, A.H., 1981. Extinction: The Causes andConsequences of the Disappearance of Species. RandomHouse, New York.

Feltham, G.A., Amershi, A.A., Ziemba, W.T., 1988. Econom-ics Analysis of Information and Contracts. Kluwer Aca-demic Publishers, Boston.

Finn, J.T., 1976. Measures of ecosystem structure and functionderived from analysis of flows. Journal of TheoreticalBiology 56, 363–380.

Fischer, W. (Ed.), 1978. FAO Species Identification Sheets forFishery Purposes: Western Central Atlantic (Fishing Area31), vols. I–VII. Food and Agriculture Organisation of theUnited Nations, Rome.

Folke, C., Hammer, M., Costanza, R., Jansson, A.M., 1994.Investing in natural capital—why, what, and how. In:Jansson, A.M., Hammer, M., Folke, C., Costanza, R.(Eds.), Investing in Natural Capital: The Ecological Eco-nomics Approach to Sustainability. Island Press, Washing-ton DC, pp. 1–12.

Friend, A., Rapport, D., 1991. Evolution of macro-informa-tion systems for sustainable development. Ecological Eco-nomics 3, 59–76.

Goreau, T.F., 1959. The ecology of Jamaican coral reefs I.Species composition and zonation. Ecology 40, 67–90.

Goreau, T.F., Goreau, N.I., 1973. The ecology of Jamaicancoral reefs II. Geomorphology, zonation, and sedimentaryphases. Bulletin of Marine Science 23, 399–464.

Grassle, J.F., Lassere, P., McIntyre, A.D., Ray, G.C., 1991.Marine biodiversity and ecosystem function. Biology Inter-national, Special Issue 23, 19 pp.

Grimm, V., Wissel, C., 1997. Babel, or the ecological stabilitydiscussions: an inventory and analysis of terminology anda guide for avoiding confusion. Oecologia 109, 323–334.

Hall, S.L., Raffaelli, D.G., 1993. Food webs: theory andreality. Advances in Ecological Research 24, 187–239.

Hannon, B., 1973. The structure of ecosystems. Journal ofTheoretical Biology 41, 535–546.

Hirata, H., Ulanowicz, R.E., 1984. Information theoreticalanalysis of ecological networks. International Journal ofSystems Science 15, 261–270.

Hirshleifer, J., Riley, J.G., 1979. The analytics of uncertaintyand information—an expository survey. Journal of Eco-nomic Literature 17, 1375–1421.

Hirshleifer, J., Riley, J.G., 1992. The Analytics of Uncertaintyand Information. Cambridge University Press, Cambridge.

Holling, C.S., 1992. Cross-scale morphology: geometry anddynamics of ecosystems. Ecological Monographs 62, 447–502.

Holling, C.S., Schindler, D.W., Walker, B.W., Roughgarden,J., 1995. Biodiversity in the functioning of ecosystems: anecological synthesis. In: Perrings, C., Maler, K.-G., Folke,C., Holling, C.S., Jansson, B.-O. (Eds.), Biodiversity Loss:Economic and Ecological Issues. Cambridge UniversityPress, Cambridge, pp. 44–83.


Isard, W., 1969. Some notes on the linkage of the ecologic andeconomic systems. Papers. Regional Science Association22, 85–96.

Isard, W., 1972. Ecologic–Economic Analysis for RegionalDevelopment. The Free Press, New York.

Ives, A.R., 1995. Measuring resilience in stochastic systems.Ecological Monographs 65, 217–233.

Kullback, S., 1968. Information Theory and Statistics. DoverPublications, New York.

Lamont, B.B., 1995. Testing the effect of ecosystem composi-tion/structure on its functioning. Oikos 74, 283–295.

Lawton, J.H., 1989. Food webs. In: Cherrett, J.M. (Ed.),Ecological Concepts: The Contribution of Ecology to anUnderstanding of the Natural World. Blackwell ScientificPublications, Oxford, pp. 43–78.

Lawton, J.H., Brown, V.K., 1993. Functional redundancy. In:Schulze, E.D., Mooney, H.A. (Eds.), Biodiversity andEcosystem Function. Springer, Berlin, pp. 255–270.

Liddell, W.D., Ohlhorst, S.L., 1981. Geomorphology andcommunity composition of two adjacent reef areas, Dis-covery Bay, Jamaica. Journal of Marine Research 39,791–804.

Liddell, W.D., Ohlhorst, S.L., 1987. Patterns of reef commu-nity structure, north Jamaica. Bulletin of Marine Science40, 311–329.

Liddell, W.D., Ohlhorst, S.L., Boss, S.K., 1984. Communitypatterns on the Jamaican fore reef (15–56 m). Palaeonto-graphica Americana 54, 385–389.

Ludwig, D., Walker, B., Holling, C.S., 1997. Sustainability,stability, and resilience. Conservation Ecology, 1. http://www.consecol.org/vol1/iss1/art7.

MacArthur, R., 1955. Fluctuations of animal populations anda measure of community stability. Ecology 36, 533–536.

Macho-Stadler, I., Perez-Castrillo, J.D., 1997. An Introduc-tion to the Economics of Information: Incentives andContracts. Oxford University Press, New York.

Martinez, N.D., 1996. Defining and measuring functional as-pects of biodiversity. In: Gaston, K.J. (Ed.), Biodiversity:A Biology of Numbers and Difference. Blackwell Science,Oxford, pp. 114–148.

May, R.M., 1972. Will a large complex system be stable.Nature 238, 413–414.

May, R.M., 1973. Stability and Complexity in Model Ecosys-tems. Princeton University Press, Princeton.

May, R.M., 1975. Stability in ecosystems: some comments. In:van Dobben, V.H., Lowe-McConnell, R.H. (Eds.), Unify-ing Concepts in Ecology. Dr W. Junk B.V. Publishers, TheHague, pp. 161–168.

Miller, R., Blair, P., 1985. Input–Output Analysis: Founda-tions and Extensions. Prentice Hall, Englewood Cliffs.

Meredith, T.C., Moore, C., Gartner, L., Smith, W., 1994.Canadian Critical Environmental Zones: Concepts, Goalsand Resources. A report of the Critical Zones Panel of theCanadian Global Change Program. Canadian GlobalChange Program Technical Report Series Report 94-1. TheCanadian Global Change Program Secretariat, The RoyalSociety of Canada, Ottawa.

Myers, N., 1996. Environmental services of biodiversity. Pro-ceedings of the National Academy of Science (USA) 93,2764–2769.

Neubert, M.G., Caswell, H., 1997. Alternatives to resiliencefor measuring the responses of ecological systems to per-turbations. Ecology 78, 653–665.

Opitz, S., 1996. Trophic Interactions in Caribbean CoralReefs. International Centre for Living Aquatic ResourceManagement, Manila.

Pahl-Wostl, C., 1995. The Dynamic Nature of Ecosystems:Chaos and Order Entwined. Wiley, Toronto.

Paine, R.T., 1988. Food webs: road maps of interactions orgrist for theoretical development. Ecology 69, 1648–1654.

Peters, R.H., 1988. Some general problems for ecology illus-trated by food web theory. Ecology 69, 1673–1676.

Pimm, S.L., 1982. Food Webs. Chapman and Hall, London.Pimm, S.L., 1984. The complexity and stability of ecosystems.

Nature 307, 321–326.Pimm, S.L., Lawton, J.H., Cohen, J.E., 1991. Food web

patterns and their consequences. Nature 350, 669–674.Randall, J.E., 1967. Food habits of reef fishes of the West

Indies. Studies in Tropical Oceanography 5, 665–847.Rutledge, R.W., Basore, B.L., Mulholland, R.J., 1976. Ecolog-

ical stability: an information theory viewpoint. Journal ofTheoretical Biology 57, 355–371.

Schulze, E.D., Mooney, H.A. (Eds.), 1994. Biodiversity andEcosystem Function. Springer, Berlin.

Ulanowicz, R.E., 1980. An hypothesis on the development ofnatural communities. Journal of Theoretical Biology 85,223–245.

Ulanowicz, R.E., 1986. Growth and Development: EcosystemPhenomenology. Springer, New York.

Ulanowicz, R.E., 1991. Contributory values of ecosystem re-sources. In: Costanza, R. (Ed.), Ecological Economics: TheScience and Management of Sustainability. Columbia Uni-versity Press, New York, pp. 253–268.

Ulanowicz, R.E., Norden, J.S., 1990. Symmetrical overhead inflow networks. International Journal of Systems Science21, 429–437.

Yodzis, P., 1989. Introduction to Theoretical Ecology. Harperand Row, New York.

http://www.consecol.org/vol1/iss1/art7

http://www.consecol.org/vol1/iss1/art7

Documents

Measuring contributions to economic production—use of an Index of Captured Ecosystem Value