dynamic cost-effective reduction strategies for acidification in europe: an application to ireland...

Environmental Modeling and Assessment 7: 163–178, 2002. 2002 Kluwer Academic Publishers. Printed in the Netherlands.

Dynamic cost-effective reduction strategies for acidificationin Europe: An application to Ireland and the United Kingdom

Erik Schmieman a, Wim de Vries b, Leen Hordijk c, Carolien Kroeze c, Maximilian Posch d, Gert Jan Reinds b andEkko van Ierland e

a Ministry of Economic Affairs, Economic Policy Directorate (AEP), A/105, Bezuidenhoutseweg 30, 2500 EC The Hague, The NetherlandsE-mail: [email protected]

b Alterra, P.O. Box 47, 6700 AA Wageningen, The Netherlandsc Environmental Systems Analysis Group, Department of Environmental Sciences Wageningen University, P.O. Box 9101, 6700 HB Wageningen,

The Netherlandsd National Institute of Public Health and the Environment (RIVM), P.O. Box 1, 3720 BA Bilthoven, The Netherlands

e Environmental Economics and Natural Resources Group, Wageningen University, P.O. Box 8130, 6700 EW Wageningen, The Netherlands

This paper describes the application of an optimisation model for calculating cost-effective abatement strategies for the reduction ofacidification in Europe while taking into account the dynamic character of soil acidification in a number of countries. Environmentalconstraints are defined in terms of soil quality indicators, e.g., pH, base saturation or the aluminium ion concentration in the soil solutionwithin an optimisation model for transboundary air pollution.

We present a case study for Ireland and the United Kingdom. Our results indicate that reduction of sulphur dioxide emission is more cost-effective than that of nitrogen oxides or ammonia. The reduction percentages for sulphur dioxide are highest, for two reasons: (i) marginalsulphur dioxide reduction costs are relatively low compared to marginal reduction costs of nitrogen oxides and ammonia and (ii) sulphurdioxide reduction is more effective in reducing acidification in physical terms than nitrogen oxides or ammonia abatement. Our dynamicanalysis shows that a (fast) improvement of soil quality requires high emission reduction levels. These reduction levels are often higher thanreduction levels that are typically deduced from the static critical loads approach. Once soil quality targets are reached, in our model, lessstringent emission reductions are required to maintain the soil quality at a constant and “good” target level. Static critical load approachesthat ignore dynamic aspects therefore may underestimate the emission reductions needed to achieve predefined soil quality targets.

Keywords: environmental economics, cost-effectiveness, dynamic soil acidification, Europe, integrated assessment, modelling, transbound-ary

1. Introduction

Acidification is an important transboundary air pollu-tion problem with many economic and environmental as-pects. It was first noticed in Scandinavia in Europe in the1960’s [20]. It was observed that acidifying precipitation(commonly known as “acid rain”) was acidifying the lakesand other ecosystems at such a rate that species died and bio-diversity decreased. It was also discovered that emitted acid-ifying substances are transported through the air over manyhundreds of kilometres, and that acid deposition in a countryis partly (depending on the location and size of the country)caused by emissions in other European countries [21]. Theacidification problem received political attention in the be-ginning of the 1970’s, which resulted in the Convention onLong Range Air Pollution in 1979 [28].

Up to now, European policy targets for acidification aredefined in terms of maximum emissions based on maximumdepositions derived from critical loads used as a thresholdvalue. If acid deposition is below the threshold, it is assumedthat no environmental damage occurs. A critical load is de-fined as a quantitative estimate of an exposure to one or morepollutants below which significant harmful effects on speci-fied elements of the environments do not occur according tothe present knowledge [19].

Integrated assessment models have been used to assistpolicy makers in defining acidification policies in Europe.The Regional Air Pollution INformation and Simulation(RAINS) model [1,2,5] is probably the best known and mostwidely used model for analysing strategies to realise acid-ification policy targets on a European scale. The RAINSmodel was developed at the International Institute for Ap-plied Systems Analysis (IIASA) as an integrated assessmenttool to assist policy advisors in evaluating options for reduc-ing acidification. The RAINS model can be used as an op-timisation model to determine cost-optimal reduction strate-gies based on deposition targets or critical loads for Europe.

In this paper we describe an alternative model foranalysing cost-effective reduction strategies for sulphurdioxide (SO2), nitrogen oxides (NOx) and ammonia (NH3)

in Europe. The paper addresses scientists and policy mak-ers involved in acid rain policies. Contrary to the RAINSmodel we do not include critical loads in our model, butexplicitly account for the dynamic aspects of soil acidifica-tion. Moreover, our model combines this dynamic soil acid-ification model with a dynamic economic cost-effectivenessanalysis that takes into account dynamic aspects related tosoil acidification as an alternative to the steady-state criticalloads approach used so far. Although critical loads are use-ful in integrated assessment modelling and relatively easy to

164 E. Schmieman et al. / Dynamic cost-effective reduction strategies for acidification in Europe

understand for politicians and non-experts, they may not bethe most appropriate basis for an analysis of acidification re-duction policies. It has been stated by several authors thatdynamic aspects of acidification are an important considera-tion in studies on the impact of acidification on ecosystems.For example, Rennings and Wiggering [24] point out that thecritical loads concept is only a first approach for protectingthe quality of complex ecosystems because a critical loadby itself is not an indicator for environmental quality, whilecritical load exceedance can be used as such. In a study oncritical loads and recovery of forest soils, Hettelingh andPosch [16] show the importance of considering the time asoil or ecosystem needs to recover from excess acid deposi-tion. In addition, De Vries [11] states that critical loads giveinformation about the acceptable levels of acid depositionin the long run only. Other natural science studies stressingthe importance of analysing abatement policy options usingdynamic soil acidification models include Holmberg [17],De Vries and Kros [13], Bull [7] and Forsius et al. [15]. In atheoretical paper Schmieman and Van Ierland [26] show howacidification of ecosystems can be studied in economic mod-elling by incorporating dynamic aspects of soil acidification.They show that the temporal development of soil acidifica-tion plays an essential role in identifying optimal reductionpolicies. This study extends that research by an empiricalapplication.

This paper focuses on a transnational application of anoptimisation model that integrates an economic model forabatement strategies with an advanced dynamic soil acidi-fication model. Environmental policy targets are defined interms of soil quality indicators, i.e., the pH, the base satura-tion and the aluminium ion concentration in the soil solution.Dynamic economic aspects of the abatement cost curves aretaken into account and reduction costs are discounted. Weprefer the dynamic approach, because emission reductionstrategies that are based on steady state critical loads do nottake into account that acidified soils need to recover. Such astrategy only focuses on the target load (e.g., the maximumdeposition). The cost-effective strategy would be to reachthat target deposition as late as possible in time and emis-sion reduction will just be sufficient to meet the depositiontargets. At the moment that the critical load is met, soilscan still be acidified because soils only slowly recover if de-positions are equal to the critical load. Since our dynamicapproach taken here focuses on soil quality it incorporatessoil recovery with the corresponding depositions lower thanthe critical load because a lower critical load is needed torealise soil recovery. Schmieman and Van Ierland [26] givea detailed analysis and motivation of this argument. By in-tegrating an economic model and a soil acidification modelthat relates soil quality indicators to acidifying deposition ina dynamic way, it is possible to determine optimal reductionstrategies based on soil quality targets.

The aims of this study are therefore: (i) to develop andapply an economic optimisation model that includes a dy-namic soil acidification model and (ii) to show how cost-effective reduction strategies for SO2, NOx and NH3 can be

investigated while explicitly incorporating dynamics in soilacidification. In principle, the optimisation model covers thewhole of Europe, but as an illustrative case study we limitthe impacts side of the application to Ireland and the UnitedKingdom.

In the next section we describe the optimisation modelwith separate descriptions of the dynamic soil acidificationmodule, the economic submodule and the derivation of thenational abatement cost curves. In section 3 we apply themodel and present our model results. In the final sectionwe draw conclusions and give recommendations for furtherresearch.

2. Model description, application and data

2.1. The dynamic soil acidification module

First we describe the simplified dynamic soil acidifica-tion module used in this study. A full description and jus-tification of the dynamic soil model including assumptions,derivations, data and estimation procedures can be found inReinds et al. [23].

As an indicator for the soil quality one can use the pH orthe aluminium ion concentration in the soil solution. Onecan also use the number of sites at the exchange complexof the soil matrix occupied by base cations, calcium (Ca),magnesium (Mg), potassium (K) and sodium (Na), whichcan be exchanged with protons and/or aluminium ions. Thisnumber of sites is given by B(t) · CECa, where CECa isthe areal cation exchange capacity of the rooting zone, a soilproperty. B(t) is the so-called base saturation, the fractionof the exchange capacity occupied by base cations at time t

(0 � B(t) � 1). The base saturation is a measure for theacid neutralisation capacity of the soil, and also for (part of)the availability of base cations for forest growth. The netacid input (deposition) and the leaching of acidity drive thechange over time of the base saturation. To make the modelsuitable for use in integrated assessment analysis the num-ber of different soil types has been aggregated to five. Thechoices that have been made in the aggregation process arejustified in Reinds et al. [23].

In this study, we focus on the aluminium ion concentra-tion in the soil solution, because there is evidence that ad-verse effects on forest ecosystems can occur with an alu-minium ion concentration of 0.2 molc m−3 or higher. Thislevel is based on results for red spruce, indicating statisti-cally significant biomass reductions at a concentration of2 mg l−1, and therefore considered the most sensitive treespecies [10]. This is a very stringent criterion, since severalcommon tree species in Europe, such as Scots pine, Nor-way spruce, European oak and beech are considered muchless sensitive to aluminium. The base saturation, B(t), isa stock quantity and is expressed as a fraction of the cationexchange capacity. The change over time of the base satura-

E. Schmieman et al. / Dynamic cost-effective reduction strategies for acidification in Europe 165

Table 1Variables and parameters used in the soil model.

Variables (dependent on time) UnitsAcle

s,g(t) acidity leaching molc m−2 yr−1

Acnets,g(t) net acid input (deposition) molc m−2 yr−1

Bs,g(t) base saturation fractionNg(t) total nitrogen deposition molc m−2 yr−1

pHs,g(t) pH is equal to − log10([H](t)) –Sg(t) total sulphur deposition molc m−2 yr−1

[Al]s,g(t) aluminium ion concentration in the soil solution molc m−3

[H]s,g(t) H+ concentration in the soil solution molc m−3

[HCO3]s,g(t) HCO3 concentration in the soil solution molc m−3

Parameters Values Unitsas , bs , cs empirical parameters see table 2 –ds empirical parameters see table 2 m3mol−1

cCECas,g cation exchange capacity multiplied by data baseb molc m−2

bulk density soil thicknessf nets,g the fraction of N that causes acidification after data baseb –

correcting for the occurrence of denitrificationand immobilisation

KAlox Gibbsite equilibrium constant, soil classes 1 to 4 300 mol−2c m6

Gibbsite equilibrium constant, for soil class 5 3 mol−2c m6

KCO2 dissociation constant of CO2 (at 10◦C) 0.0189 (mol l−1)2 atm−1

pCO2 partial pressure of CO2 in soil 0.00948 atmpss,g precipitation surplus data baseb m yr−1

SBCs,g net input of base cations by deposition plus data baseb molc m−2 yr−1

weathering minus uptake and removal of Nby forest growth uptakea

g index for grid cell g = 1 to G

s index for soil class s = 1 to 5

aThese data also depend on forest type, however data are presented as the average over the soil type in a grid cell.bSee Reinds et al. [23].

tion in a soil s in EMEP1 grid cell g, Bs,g(t), is assumed tobe proportional to the difference between the net acid input,Acnet

s,g(t), to the soil and the acid leaching, Acles,g(t), from the

soil, according to (for explanation of the symbols and unitssee table 1):

Bs,g(t) = −1

CECas,g

(Acnet

s,g(t) − Acles,g(t)

), (1)

where CECas,g is the areal cation exchange capacity of therooting zone.

The net acid input, equals the sum of atmospheric de-position of sulphur, S, and nitrogen, N, corrected for basecation, SBCs,g, deposition. In the soil, S removal is gener-ally negligible, but the acidifying impact of N is less due tothe removal of N by denitrification, uptake and immobili-sation (captured by f net

s,g ). Furthermore, there may be a netrelease of base cations by weathering corrected for uptake(captured by SBCs,g). The net input of acidity is thus calcu-lated as:

Acnets,g(t) = Sg(t) + f net

s,g Ng(t) − SBCs,g. (2)

1 The Co-operative Programme for the Monitoring and Evaluation of theLong-Range Transmission of Air Pollutants in Europe (EMEP) has de-veloped atmospheric dispersion models. Most of the input and outputdata of these models are based on the so called EMEP 150 by 150 kmgrid cells (see also figure 2).

The acidity leaching equals the leaching of the sum of H+concentration in the soil solution, [H]s,g(t), the aluminiumion concentration in the soil solution [Al]s,g(t) and theHCO3 concentration in the soil solution, [HCO3]s,g(t), be-ing calculated as

Acles,g(t) = pss,g

([H]s,g(t) + [Al]s,g(t) − [HCO3]s,g(t)).

(3)The concentration [H]s,g(t) is related to the pH, which

was calculated in an empirical way as a function of base sat-uration and acid concentration according to (where as , bs , cs

and ds are estimated parameters, see table 2):

[H]s,g(t) = 103−pHs,g (t), (4)

pHs,g(t) = as + bs

(Bs,g(t) − 0.5

)

+ cs

(Bs,g(t) − 0.5

)9 − ds

Acnets,g(t)

pss,g. (5)

The concentrations [Al]s,g(t) and [HCO3]s,g(t) are cal-culated from chemical equilibria, with KAlox as the Gibbsiteequilibrium constant, according to:

[Al]s,g(t) = KAlox[H]3s,g(t), (6)

[HCO3]s,g(t) = KCO2pCO2

[H]s,g(t) , (7)


Table 2Soil classes and their estimated parameters for the pH-base saturation relationship.

s Soil classes Coefficients

as bs cs ds(m3 mol−1c )

1 Sandy soils 5.100 0.939 465.9 0.05652 Loamy soils 4.861 1.207 550.9 0.11603 Clay soilsa 4.828 0.981 692.6 0.08554 Heavy clay soilsa 4.828 0.981 692.6 0.08555 Peat soils 5.054 1.084 728.1 0.0715

aThe soil classes 3 and 4 have the same estimated coefficients, but differ in inputfrom the database (see for more details Reinds et al. [23]).

where KCO2 is the dissociation constant of CO2 (at 10◦C)and pCO2 is the partial pressure of CO2 in soil.

Following this approach, the base saturation was annu-ally updated (1960 to 1990) as a function of the atmosphericinput of S and N, and related to pH and aluminium ion con-centration. The base saturation depends on soil characteris-tics and the deposition history. The base saturation for thestarting year of 1990 has been derived from a soil data basefor Europe and historical deposition patterns of sulphur andnitrogen [23].

Equation (5) is an empirical relationship between the pHand the base saturation and replaces a set of exchange equa-tions used in deterministic soil acidification models suchas Simulation Model for Acidification’s Regional Trends,SMART [14]. For our application the parameters as , bs , cs

and ds (see table 2) are derived from SMART simulations forthe period 1960–1990 using a European data base of about100,000 forest soil combinations [23].

2.2. The economic optimisation module

The model calculates cost-effective reduction of emis-sions of k pollutants in n countries that lead to acidificationof soils for the period 1990 to 2030, with k = SO2 for sul-phur dioxide, k = NOx for nitrogen oxides and k = NH3 forammonia. Emission of pollutant k after abatement in coun-try n (with n = 1 to L) at time t , Ek

n(t), is the initial emissionof pollutant k in country n at t , Ek

n(t), minus abatement ofpollutant k in country n at t , Ak

n(t),

Ekn(t) = Ek

n(t) − Akn(t). (8)

Both emission and abatement are expressed in kton with1 kton = 106 kg. Emissions of country n are transportedthrough the air to a receptor grid cell g (g = 1 to G).Depositions in a receptor grid cell g can be calculated us-ing source-receptor matrices Mk (with dimensions G × L).These source receptor matrices, that include transfer coeffi-cients and a factor that translates emissions of pollutant k inkton to acid equivalents per hectare, are assumed to be con-stant over time and were taken from the RAINS model [5].Deposition of sulphur S and nitrogen N (in moles of chargeper m2 per year or molc m−2 yr−1) in all grid cells are calcu-lated by multiplication of the emission vectors of SO2, NOx

and NH3, respectively ESO2(t), ENOx (t) and ENH3(t) (di-

mensions L × 1) by the corresponding source-receptor ma-trix

S(t) = MSO2ESO2(t), (9)

N(t) = MNOx ENOx (t) + MNH3ENH3(t), (10)

where S(t) and N(t) are the deposition vectors (with dimen-sions G × 1).

The abatement cost functions, Ckn,t (A

kn(t)), are piece wise

linear functions and increasing in Akn(t) at any time t .

We want to find the abatement path Akn(t) that min-

imises the sum of the discounted total reduction costs, TC,of all pollutants in all countries from 1990 to 2030 plus thediscounted reduction costs beyond 2030 to infinity. Withe−r(t−1990) is the discount factor and in the calculations weused 4% as the discount rate):

MinTC =L∑n

3∑k

(∫ 2030

1990e−r(t−1990)Ck

n,t (Akn(t)) dt

+ 1rC(Ak

n(2030))e−40r

)(11)

subject to

[Al]s,g(t) � [Al]cs,g(t) for t � tc (12)

and equations (1) to (10).The second term in equation (11) is called a “scrap value

function” and it is the value of the discounted reduction coststhat are necessary to keep the reduction level after 2030 con-stant at the level of the year 2030 (which is the approxi-mation of a limit to infinity). The scrap value depends onthe terminal time and the policy target in terms of a criti-cal aluminium ion concentration in the soil that has to bemet after 2030 as stated in equation (12). In equation (12),[Al]s,g(t) is the aluminium ion concentration in the soil so-lution (molc m−3) of soil s in grid cell g and [Al]cs,g(t) is the

critical aluminium ion concentration (molc m−3) in soil s ingrid cell g. Equation (12) states a policy target: the alu-minium ion concentration should be less or equal to a criti-cal aluminium ion concentration, [Al]cs,g(t), after a definedpoint in time, tc. The deposition vectors S and N are in-put to the soil acidification module given in equations (1)to (7). The model was formulated and solved in the GAMSprogramming system [6].


Figure 1. SO2 abatement cost curves in the United Kingdom in 2010, 2000, 1990 as used in our model and the abatement cost curve in 1990 fromRAINS [5].

2.3. National abatement cost curves

The cost curves for unabated emissions in 2010 are takenfrom the RAINS model’s [5]. The methodology to estimatenational abatement cost functions that are used in RAINSfor sulphur emission reduction costs is described by Cofalaand Syri [8], for nitrogen oxides by Cofala and Syri [9] andfor ammonia by Klaassen [18]. Cost curves are constructedby ranking the available emission control options for vari-ous emission sources according to their marginal costs. Thecontrol options are combined with the potential for emissionreduction determined by the properties of the fuel and abate-ment technologies. The cost curves are constructed basedon the calculated unit cost, first for every sector and thenfor the whole country. Technologies characterised by highercosts and lower reduction potentials are considered as notcost-efficient and are excluded from further analysis. Themarginal costs are calculated for each sector. The remainingabatement options are finally ranged according to increasingmarginal costs. They form a piece wise linear cost curve foreach country with each piece representing a particular abate-ment option with constant marginal costs [8]. Figure 1 givesexamples of piece wise linear cost curves.

The national abatement cost curves for each pollutant foreach individual country for the years between 1990 and 2010are derived from the 2010 cost curves from RAINS by a hor-izontal and a vertical shift of each point on the cost curve.Two points are fixed: (i) the zero abatement point (with zerocosts) and (ii) the point for maximum feasible reduction withthe corresponding maximum reduction costs. All intermedi-ate points are scaled based on the shape of the 2010 costcurve. For the years beyond 2010 the cost curves are as-sumed to remain constant at their 2010 level. For more de-tails see [25].

Figure 1 shows the total abatement cost curves for SO2 inthe United Kingdom in the years 1990, 2000 and 2010. Theabatement cost curve in 2010 is identical to the correspond-ing RAINS cost curve. The RAINS abatement cost curve in1990 is shown for comparison with the calculated abatementcost curve in 1990. The shift of abatement cost curves can beup-wards or down-wards depending on the country, the en-

ergy pathway and the pollutant. Abatement cost curves fornitrogen oxides and ammonia show similar characteristics.

3. Model application, data and results: Spatial andintertemporal cost-effective emission reduction

3.1. Model application and data

The aim of our model is to calculate cost-effective strate-gies for the reduction of acidification in Europe taking intoaccount the dynamic processes related to soil acidification.At a European level, this would imply calculations for about600 land based grid cells and about 200 soil types. Solvingthis model for a time horizon of 40 years requires large com-puter capacity and a long calculation time. We reduced thesize of our model because of computer limitations and foranalytical purpose. The number of distinguished soils hasbeen reduced to 5 soil classes as described in section 2.1.The assessment of impacts in our calculations (depositionsand soil acidification) is limited to the land based grid cellsof Ireland and the United Kingdom (the grey shaded gridcells in figure 2).

We selected these two countries because they have rela-tively high depositions of S and N that originate from domes-tic sources (see appendix C). This implies that the two coun-tries themselves can control most of the depositions. Wetake data from the RAINS model whenever possible (emis-sions and emission reduction cost, source–receptor matri-ces). In the optimisation analysis, we only take into accountthe countries from which emissions have a significant im-pact on acid deposition in the United Kingdom and Ireland.The selection criterion is based on the country to grid cellsource–receptor matrices for each pollutant. Countries areexcluded from the analysis when their emissions of eachpollutant contribute less than 0.2% to the total depositionsonto all land based grid cells in the United Kingdom andIreland. In this way at least 99% of the total controllabledeposition (i.e., excluding background depositions) are en-dogenous to our model. Appendix C gives an overview ofthe relative contributions of European countries (or regions).Background depositions and depositions from sources at seaare included in our model but as exogenous values.


Figure 2. The model includes the area of Europe covered by the EMEPgrid cells as shown in this figure. For this application only impacts and soilquality targets for land based grid cells in the United Kingdom and Ireland

are incorporated (the grey shaded grid cells).

Although the impacts considered in our application of themodel are geographically limited to Ireland and the UnitedKingdom, the emission module includes 43 countries (in-cluding sea regions, see appendix A). Emission data aretaken from the RAINS model [5]. The exogenously givenunabated SO2 and NOx emissions largely depend on energyuse based on defined energy pathways which are used to cal-culate SO2 and NOx emissions as follows. For the 15 Euro-pean Union countries (EU-15) the energy pathways are ei-ther “national pathways” or “business as usual” as describedand used in the 6th and 7th interim report to the EU and in thecalculations for the UN/ECE [3,4]. For the non-EU coun-tries the energy pathways are the “official energy pathways”.NH3 emission data are generated using “official agriculturalpathways 1998”. For a detailed description of the energyand agricultural pathways the reader is referred to IIASA’s6th Interim Report to the European Commission DG-XI [3,pp. 49–51] and IIASA’s 7th Interim Report to the EuropeanCommission DG-XI and [4, pp. 9, 35]. Its is important tonote that the unabated emissions are exogenously given tothe model.

We used our model to evaluate two cases:

(i) The reference scenario in which no abatement takesplace in European countries (where abatement is definedas emission reduction by technical means). This is of-ten called “business as usual”. The model calculates soilquality indicators resulting from the reference scenarioemissions.

(ii) An abatement strategy in which our model calculatesoptimal abatement paths for 39 European countries(plus exogenous emissions of four sea areas) givena soil quality target (an aluminium ion concentration of

Figure 3. Total emissions of SO2, NOx and NH3 in Europe in the referencescenario in 1990 and 2010. Changes in emissions result from changes in

energy use and fuel mix.

0.2 molc m−3 or less) for the United Kingdom and Ire-land to be reached in the year 2010 and maintained af-terwards.

3.2. Reference scenario: no abatement

In the reference scenario no technical emission reduc-tion measures are assumed to be implemented in Europe sothat the emissions for each year are the result of the (exoge-neously given) national agricultural and energy pathways asdescribed in section 3.1. The reference scenario emissionsfor 1990 and 2010 are presented in appendix B. The refer-ence scenario emission levels for the years between 1990and 2010 are a linear interpolation of the emissions in 1990and 2010. From 2010, emissions are assumed to remain atthe 2010 level.

The energy pathway describes the fuels used by sectorsand determines the reference scenario emissions of sulphurdioxide and nitrogen oxides. Fuels with high sulphur contentare expected to be replaced by fuels with low sulphur contentin the reference scenario between 1990 and 2010. These fuelswitches largely affect emissions. For example, the switchto fuels containing less sulphur in the United Kingdom leadsto almost 50% reduction of SO2 emissions in 2010 relativeto 1990, without any application of end-of-pipe abatementtechnologies (see appendix B for emission levels).

It is important to note that total European SO2 emissionswill decrease by 26% in 2010 relative to 1990 in the ref-erence scenario (figure 3). Figure 3 also shows that emis-sions of NOx and NH3 remain constant over time. Thus,with the given energy pathway, acid deposition will decreaseand soils will be less acidified in 2010 compared to 1990in the reference scenario. We define abatement as the re-duction of emissions by means of technical measures (e.g.,fuel desulphurisation and end of pipe technologies). The un-abated emission levels are the reference scenario emissionlevels, figure 4 shows the depositions over time.

Figure 5 shows the temporal development of the soil qual-ity as represented by the aluminium ion concentration for thefive soil classes (see table 2). Soils with aluminium ion con-centrations � 0.2 molc m−3 are considered not to be acid-ified (see remarks in section 2.1). The aluminium ion con-


Figure 4. Deposition of S and N (molc m−2 yr−1) in Southeast United Kingdom in the reference scenario.

centration in the soil solution of heavy clay soils (class 4)and peat soils (class 5) remains below 0.2 molc m−3 duringthe whole period. Sandy soils (class 1), loamy soils (class 2)and clay soils (class 3) are most affected by acidification andtherefore we focus on these three classes in the remainder ofthis paper. Soils (except for clay soils, class 3) slightly re-cover from acidification between 1990 and 2030 as a resultof the decrease in SO2 deposition in the reference scenario(see figure 5).

3.3. An optimal abatement strategy

In this section we derive a cost-effective reduction strat-egy (the “Optimal Abatement Strategy”) that minimises thesum of the abatement costs for the European countries in-cluded in our analysis (see appendix C) for all pollutantsfor all years (see equation (11)), necessary to achieve de-fined soil quality targets in the United Kingdom and Ire-land in 2010 and maintained afterwards as formulated inequation (13). In section 2.1 we justified the critical alu-minium ion concentration of 0.2 molc m−3 and we chose2010 because the current acidification reduction strategies[29,30] are formulated based on deposition targets specifiedfor 2010. Therefore, our soil quality target is

[Al]cs,x,g(t) � 0.2 molc m−3 for t � 2010. (13)

The optimal emission reduction time paths are depictedin figure 6 for SO2 and in figure 7 for NOx and as a percent-age of unabated emissions for the most important countries.Table 3 gives the abatement levels for selected years.

The general tendency in our results show that abatementincreases in time with the highest abatement percentages forall pollutants in 2010. Optimal SO2 abatement levels for theUnited Kingdom range up to 92% of the reference scenarioemissions, optimal NOx emissions up to 58%. For othercountries the abatement is lower. NH3 abatement contributesleast to the decrease in deposition both in relative and in ab-solute terms. Optimal emission levels are shown in figure 8.Ammonia emission reduction is not very cost-effective toachieve the soil quality targets in the United Kingdom andIreland. Only ammonia abatement in The United Kingdom,The Netherlands, France and Belgium plays some role.

Given a soil quality target at a certain location, the cost-effectiveness of emission reduction measures in a certaincountry at a certain time is defined by five factors:

(i) the marginal cost,

(ii) the discount rate,

(iii) the impact of emissions of SO2, NOx and NH3 on de-positions of S and N, i.e. the coefficients in the source-receptor matrices,

(iv) the instantaneous impact of S and N deposition on thesoil quality,

(v) the response time or recovery capacity of the soil.

In the cost-effective reduction strategy, emission reduc-tion of SO2 turns out to be most cost-effective, and the re-duction percentage (which is 100% × Ak

n(t)/E) are highest(compare figures 6, 7, 8 and table 3). This can be explainedby the relative low marginal costs of SO2 abatement com-pared to the higher marginal reduction costs of NOx andNH3. The coefficients in the source-receptor matrices dif-fer not too much for sulphur dioxide and nitrogen oxidesso the relative cost-effectiveness of reduction measures indifferent countries is not affected by the physical charac-teristics of atmospheric transport. Ammonia emissions aretransported over smaller distances than sulphur and nitro-gen oxides. Therefore, the cost-effectiveness of abatementmeasures for ammonia is lower outside the United Kingdomand Ireland, meaning that reduction percentages are lowercompared to SO2 and NOx reduction levels. The character-istics of atmospheric transport together with the high mar-ginal reduction costs of NH3 (high relative to SO2 and NOx

marginal reduction costs) ensures that NH3 reduction takesplace only in the United Kingdom, Netherlands, Germany,France, Poland, Denmark and Belgium and for low reduc-tion percentages only (see figure 8).

In explaining differences between cost-effectiveness ofemission reduction in countries the discount rate is not im-portant since we assume the same discount rate for all coun-tries.

The instantaneous impact of a unit (kg) change in SO2emission on the soil quality is higher than the effect of a unit(kg) change of NOx or NH3 emissions because of the net


Figure 5. The temporal development of the aluminium ion concentration [Al3+] (molc m−3) for the five soil classes in the reference scenario (no abatementthrough end of pipe technologies).


Figure 6. Optimal SO2 abatement (% of unabated emissions) in the 10 countries that contribute most to acidification in Ireland and the United Kingdom(see appendix A for country abbreviations and appendix B for unabated emissions).

Figure 7. Optimal NOx abatement (% of unabated emissions) in the 10 countries that contribute most to acidification in Ireland and the United Kingdom(see appendix A for country abbreviations and appendix B for unabated emissions).

Figure 8. Optimal NH3 abatement (% of unabated emissions) in the 10 countries that contribute most to acidification in Ireland and the United Kingdom(see appendix A for country abbreviations and appendix B for unabated emissions).


Table 3Abatement levels for SO2 and NOx (abatement in kton per year) for 2000, 2005, 2010 and

2030 in the optimal abatement strategy (see appendix A for country abbreviations).

SO2 abatement NOx abatement

2000 2005 2010 2030 2000 2005 2010 2030

BELG 389 399 448 293 5 20 32 6CZRE 1212 1188 1171 0 0 0 73 0DENM 190 202 207 0 13 12 24 0FRAN 480 737 848 373 0 0 55 0GERM 3997 3522 3352 674 74 320 422 0IREL 155 175 185 100 4 7 14 4LITH 0 0 60 0 0 0 0 0LUXE 0 0 0 0 0 0 0 0NETH 239 253 245 215 24 40 109 27NORW 0 0 31 0 0 5 5 0POLA 1373 2339 2812 0 0 0 147 0PORT 0 0 0 0 0 0 0 0SPAI 0 0 975 0 0 0 0 0SWED 0 46 156 0 0 0 2 0SWIT 0 0 0 0 0 0 0 0UNKI 2589 2182 1775 1728 302 404 1605 293

effect of the differences between SO2 and NOx in molecu-lar weight and acidifying capacity on a molar base [22] andthe losses of nitrogen from the soil after deposition of NOx

or NH3 because, f net < 0.47 in equation (2). A high re-sponse time or recovery capacity of the soil shifts abatementmeasures to the future.

Remote countries (e.g., Poland) only play a role in theOptimal Abatement Strategy because the environmental tar-gets are stringent (see section 2.1). The peaked pattern ofthe abatement paths in figures 6 and 7 is caused by the waythe national cost curves are implemented in our model. Forthe piecewise linear cost curves every piece represents anabatement technique with constant marginal costs (see fig-ure 1). Because of these constant marginal costs, a tech-nique is cost-effective for its full reduction potential or notcost-effective at all and thus not used. Only if the targets arefulfilled “half way” through the application of a technique,the technique will not be fully applied, and some reductionpotential will not be used. This only applies to the last tech-nique added to meet the targets.

Our model shows that to reach a “good” soil quality highreductions are necessary compared to reductions based ona critical loads approach that does not account for soil re-covery.

At the end of 2010 the targets are reached, and the soilquality has improved in terms of the aluminium ion concen-tration ([Al] � 0.2 molc m−3) for all soils in Ireland and theUnited Kingdom. Up to 2010, reduction levels are consider-ably higher than after 2010 (see figures 6, 7 and table 3)in our dynamic model, because improving a soil requiresa lower acid deposition and higher abatement levels thanmaintaining the soil quality at a given level. The emissionlevels after 2010 can be interpreted as the emission levelsthat correspond to critical load depositions based on a criticalvalue for the aluminium ion concentration of 0.2 molc m−3.

The strong peak reduction in 2010 (illustrated in fig-ures 6–8) and the corresponding temporary decrease in de-

position (illustrated in figure 9) can be explained as follows.The instantaneous response of soil quality to a change in aciddeposition can be relatively large. In finding the OptimalAbatement Strategy our model tries to postpone emissionreduction because of the time preference resulting from thediscount rate.

Basically our model trades off low reduction levels withrelatively low marginal costs early in time against high re-duction levels with relatively high marginal costs just beforethe soil quality target has to be met. The latter option isfound to be most cost-effective. In the present model, weassume that abatement technologies can be implemented inone year and removed in the next year without additionalcosts, which also leads to peak reduction in 2010. In real-ity, most investments in reduction techniques cannot be im-plemented and removed in a short period because they havelong life times. Only some techniques can be used withouthigh investments, for example, the use of low sulphur fuelsis reversible in the short term. Therefore, this assumptionneeds modification in future analysis.

Environmental effects are shown in figure 9 for deposi-tion and in figure 10 for the soil quality (the aluminium ionconcentration in the soil solution). Figure 9 shows the opti-mal deposition for S and N in a grid cell in Southeast UnitedKingdom. Depositions adjust to levels that correspond to thecritical load for soil class 1 in that grid cell after 2010.

Figure 10 shows the temporal development of the soilquality for sandy soils, loamy soils and clay soils. Somesoils (mostly sandy soils, but also loamy soils) in the East ofthe United Kingdom just reach the minimum required soilquality in 2010. Large areas further recover in the period to2030 and no acidified soils remain.

Our model takes into account soil recovery and illustratesthat emissions need to be reduced more than the levels corre-sponding with critical loads. From 2010 the soil quality hasto be maintained at the critical level, and lower abatement issufficient.


Figure 9. Deposition (molc m−2 yr−1) of S and N in the Southeastern part of the United Kingdom under the Optimal Abatement Strategy.

Figure 10. The aluminium ion concentration in the soil solution (molc m−3) of three soil-texture combinations (soil classes 1, 2 and 3) under the OptimalAbatement Strategy.


The policy target formulated in equation (13) turns out tobe a stringent target that requires high abatement levels inseveral European countries. High abatement levels are evennecessary in countries that are far away and have only smallimpacts on depositions in Ireland and the United Kingdom,for example, Poland. Most effects of emission reduction incontinental countries on depositions occur in continental Eu-rope. Therefore, large benefits in terms of improvement ofsoil quality occur outside the models’ geographical area andare not reflected in the results reported here.

4. Discussion and conclusions

This paper shows that it is possible to integrate a dy-namic economic optimisation model with a dynamic soilacidification model to calculate cost-effective emission re-duction strategies. We performed an optimisation analysisand analysed a cost-effective way to realise defined soil qual-ity targets in the United Kingdom and Ireland in 2010 andonwards.

The model as it is presented here has some limitations.We assumed that abatement techniques can be implementedone year, become fully available in the same year, and canbe removed in the next year. It is, however, not realisticto assume that large investments in abatement technologiesare reversible in the short term. Therefore, in this respectthe model deserves modification. The impact of acidifyingdeposition on soils is modelled for Ireland and the UnitedKingdom only. This means that spillovers to other areas interms of decreased acid depositions and increased soil qual-ity are not considered. Also the policy target is rather strin-gent (both in time and by the soil quality target), implyinghigh reduction levels even in countries that have a limitedeffect on depositions in Ireland and the United Kingdom. Inevaluating the results, one has to be aware that our study fo-cuses on acidification only, assuming that an aluminium ionconcentration of 0.2 molc m−3 is a critical limit. First of all,this criterion is stringent since it is based on red spruce, be-

ing one of the most sensitive tree species. As an alternativeit would be better to use for example an aluminium to basecation ratio [27]. Furthermore, other effects of nitrogen dueto eutrophication, such as an increased sensitivity to naturalstress and nutrient unbalances, may be more important thanacidification [12]. Use of additional N targets, such as Ncontents in foliage or nitrate concentrations in water, maylead to different results. Our study only illustrates the effectof a given stringent target related to soil acidification.

Despite the above mentioned limitations of the model wecan draw some important conclusions. Current European re-duction policies are based on deposition targets related tocritical loads and critical levels. An example of this approachis the second sulphur protocol [29] and the most recent pro-tocol to abate acidification eutrophication and ground-levelozone [30]. This approach does not take into account thehigh reduction levels that are required to realise an improve-ment of soils.

Our results indicate that in order to reach the acidificationtargets for the United Kingdom and Ireland sulphur dioxideemissions have to be reduced by almost 90% and nitrogenoxides by about 60% in some European countries. In thiscost-effective solution that only focuses on acidification andnot on eutrophication, ammonia emission reduction playsa minor role. Sulphur dioxide reduction is more effective inreducing acidification in physical terms than nitrogen oxideand ammonia abatement. The analysis indicates that sulphurreduction is most cost-effective, followed by nitrogen oxidereduction and that ammonia reduction is least cost-effectivein acidification abatement. Our results also point out thata fast improvement of the soil quality requires high emissionreduction levels, much higher than reductions that would bebased on a critical loads approach. When the soil quality tar-gets are reached, less stringent emission reductions are suffi-cient to maintain the soil quality on a constant “good”’ level.Including dynamic aspects may have considerable implica-tions for the results of optimisation analyses. Most impor-tantly, static critical loads approaches that ignore dynamicaspects may underestimate the emission reductions neededto achieve predefined soil quality targets.


Appendix A

Table 4Description of country abbreviations.

Country abbreviation Country Region

ALBA Albania whole countryATLO Atlantic Ocean whole sea regionAUST Austria whole countryBALS Baltic Sea whole sea regionBELA Belarus whole countryBELG Belgium whole countryBOHE Bosnia Herzegovina whole countryBULG Bulgaria whole countryCROA Croatia whole countryCZRE Czech Republic whole countryDENM Denmark whole countryESTO Estonia whole countryFINL Finland whole countryFRAN France whole countryGERM Germany whole countryGREE Greece whole countryHUNG Hungary whole countryIREL Ireland whole countryITAL Italy whole countryLATV Latvia whole countryLITH Lithuania whole countryLUXE Luxembourg whole countryMACE FYR of Macedonia whole countryMEDS Mediterranean Sea whole sea regionMOLD Rep. of Moldova whole countryNETH Netherlands whole countryNORS North Sea whole sea regionNORW Norway whole countryPOLA Poland whole countryPORT Portugal whole countryROMA Romania whole countryRUKA Russian Feder. Kaliningrad region Kaliningrad regionRUKO Russian Feder. Kola, Karelia region Kola, Karelia regionRURE Russian Feder. Remaining Russia Remaining RussiaRUSP Russian Feder. St Petersburg region St Petersburg regionSKRE Slovakia Republic whole countrySLOV Slovenia whole countrySPAI Spain whole countrySWED Sweden whole countrySWIT Switzerland whole countryUKRA Ukraine whole countryUNKI United Kingdom whole countryYUGO Yugoslavia whole country

Source: RAINS [5].


Appendix B

Table 5Unabated emission (kton) of SO2, NOx and NH3 for 1990 and 2010.

Country 1990 emissions (kton per year) 2010 emissions (kton per year)

SO2 NOx NH3 SO2 NOx NH3

ALBA 72 24 32 55 36 35AUST 262 262 77 170 288 67BELA 843 402 219 494 316 163BELG 482 375 97 515 429 96BOHE 487 80 31 415 64 23BULG 1841 354 141 1684 314 126CROA 180 83 40 146 93 37CZRE 1922 522 107 1295 437 108DENM 282 290 77 235 249 72ESTO 275 84 29 175 73 29FINL 308 277 40 406 325 31FRAN 1392 1991 807 1037 2059 777GERM 6950 3719 757 3749 3657 572GREE 551 350 80 1139 631 74HUNG 913 214 120 1101 282 137IREL 184 113 127 207 168 130ITAL 3017 2062 462 2875 2288 432LATV 121 117 43 104 118 35LITH 213 152 80 107 138 81LUXE 15 22 7 12 23 9MACE 107 39 17 81 31 16MOLD 197 87 47 117 66 48NETH 327 593 233 297 733 191NORW 52 234 23 62 297 21POLA 2999 1209 505 3271 1404 541PORT 304 208 71 273 328 67ROMA 1331 518 292 1023 458 304RUKA 44 29 11 23 25 11RUKO 739 111 6 571 86 4RURE 3921 3126 1221 2352 2517 845RUSP 308 220 44 173 170 33SKRE 548 207 60 260 209 47SLOV 207 60 23 152 58 21SPAI 2244 1162 352 1405 1592 383SWED 258 378 61 291 438 61SWIT 62 218 72 69 247 66UKRA 3706 1888 729 1926 1433 649UNKI 3805 2839 329 2079 2782 297YUGO 585 211 90 459 166 82

Sea regionATLO 641 911 0 641 911 0BALS 72 80 0 72 80 0MEDS 12 13 0 12 13 0NORS 439 639 0 439 639 0

Sum 43216 26470 7558 31969 26670 6719

Source: RAINS [5].


Appendix C

Table 6Relative contribution by European countries to acidifying deposition in Ireland and United Kingdom.

S N from NOx N from NH3

% cum.% % cum.% % cum.%

Endogenous sourcesIREL 42.04 42.04 37.16 37.16 48.73 48.73UNKI 33.68 75.72 26.93 64.09 41.11 89.83NETH 5.90 81.62 8.13 72.22 2.22 92.06BELG 5.28 86.90 6.76 78.99 2.34 94.39GERM 4.13 94.59 6.21 90.12 1.65 98.19FRAN 3.56 90.46 4.92 83.91 2.14 96.53DENM 1.29 95.88 2.50 92.62 0.72 98.91CZRE 0.94 96.82 1.33 93.95 0.19 99.10SPAI 0.67 97.49 0.98 94.93 0.17 99.26POLA 0.49 97.98 0.89 95.83 0.24 99.51LUXE 0.39 98.37 0.29 96.12 0.02 99.53SWED 0.32 98.69 1.03 97.14 0.08 99.61NORW 0.27 98.96 0.98 98.13 0.07 99.67PORT 0.12 99.09 0.53 98.65 0.04 99.72LITH 0.12 99.23 0.21 99.11 0.05 99.79SWIT 0.02 99.11 0.24 98.90 0.03 99.74

Exogenous sourcesSKRE 0.13 99.49 0.12 99.50 0.03 99.87HUNG 0.13 99.68 0.09 99.70 0.02 99.91AUST 0.07 99.30 0.14 99.25 0.06 99.85FINL 0.06 99.35 0.13 99.39 –ITAL 0.06 99.54 0.11 99.62 0.01 99.89BELA 0.06 99.74 0.08 99.79 0.03 99.94ESTO 0.05 99.90 – –UKRA 0.04 99.78 0.07 99.86 0.03 99.97ROMA 0.04 99.82 0.06 99.92 0.01 99.98SLOV 0.02 99.92 – –RUKO 0.02 99.95 – –YUGO 0.02 99.96 – 0.00 100.00RUSP 0.02 99.98 – –LATV 0.01 99.83 0.05 99.97 0.01 99.99BULG 0.01 99.84 0.02 99.99 0.00 100.00RURE 0.01 99.85 0.01 100.00 0.00 100.00RUKA 0.01 99.99 – –BOHE 0.01 100.00 – –ALBA – – –CROA – – –GREE – – –MACE – – –MOLD – – –

Note: 0.00 means that a source contributes to depositions but its magnitude is very small (rounded 0.00%). A “–” denotes that a source doesnot contribute to depositions (exactly 0%). Country abbreviations can be found in appendix A. Data derived from EMEP source receptormatrices [5].

References

[1] J. Alcamo, R.W. Shaw and L. Hordijk, The RAINS Model of Acidifica-tion: Science and Strategies in Europe (Kluwer Academic Publishers,Dordrecht, The Netherlands, 1990).

[2] M. Amann, RAINS 7.2 Manual and Model Description, inter-net address http://www.iiasa.ac.at/∼rains/index.html, last accessed08.08.2000 (2000).

[3] M. Amann, I. Bertok, J. Cofala, F. Gyarfas, C. Heyes, Z. Klimont,M. Makowski, W. Schöpp and S. Syri, Cost-effective control of acid-ification and ground-level ozone, Sixth Interim Report to the Euro-pean Commission, International Institute for Applied Systems Analy-

sis, Laxenburg, Austria (1998).[4] M. Amann, I. Bertok, J. Cofala, F. Gyarfas, C. Heyes, Z. Klimont,

M. Makowski, W. Schöpp and S. Syri, Cost-effective control of acid-ification and ground-level ozone, Seventh Interim Report to the Euro-pean Commission, International Institute for Applied Systems Analy-sis, Laxenburg, Austria (1999).

[5] M. Amann, I. Bertok, J. Cofala, F. Gyarfas, C. Heyes, Z. Klimont andW. Schöpp, The RAINS 7.2 model of air pollution. Model descrip-tion and general overview, International Institute for Applied SystemsAnalysis, Laxenburg, Austria (1997).

[6] A. Brook, D. Kendrick and A. Meier, GAMS: A User’s Guide(The Scientific Press, San Fransisco, USA, 1992).


[7] K.R. Bull, Critical loads – Possibilities and constraints, Water, Airand Soil Pollution 85 (1995) 201–212.

[8] J. Cofala and S. Syri, Sulfur emissions, abatement technologies andrelated costs for Europe in the RAINS model database, Interim Re-port IR-98-035, International Institute for Applied Systems Analysis,Laxenburg, Austria (1998).

[9] J. Cofala and S. Syri, Nitrogen oxides emissions, abatement tech-nologies and related costs for Europe in the RAINS model database,Interim Report IR-98-88, International Institute for Applied SystemsAnalysis, Laxenburg, Austria (1998).

[10] C.S. Cronan, R. April, R.J. Barlett, P.R. Bloom, C.T. Driscoll,S.A. Gherini, G.S. Henderson, J.D. Joslin, J.M. Kelly, R.M. New-ton, R.A. Parnell, H.H. Patterson, D.J. Raynall, M. Schraedle,C.T. Schofield, E.I. Sucoff, H.B. Tepper and F.C. Thornton, Alu-minium toxity in forest exposed to acidic deposition, Water, Air andSoil Pollution 48 (1989) 181–192.

[11] W. De Vries, Average critical loads for nitrogen and sulfur and its usein acidification abatement policy in the Netherlands, Water, Air andSoil Pollution 68 (1993) 399–434.

[12] W. De Vries, Soil response to acid deposition at different regionalscales. Field and laboratory data, critical loads and model predictions,Ph.D. thesis, Wageningen Agricultural University, Wageningen, TheNetherlands (1994).

[13] W. De Vries, and J. Kros, Assessment of critical loads and the impactof deposition scenarios by steady state and dynamic soil acidifica-tion model, Report 36, The Winand Staring Centre, Wageningen, TheNetherlands (1991).

[14] W. De Vries, M. Posch and J. Kämäri, Simulation of the long-termsoil response to acid deposition in various buffer ranges, Water, Airand Soil Pollution 48 (1989) 349–390.

[15] M. Forsius, M. Alveteg, R. Guardans, M. Holmberg, A. Jenkins,M. Johansson, S. Kleemola, K. Rankinen, M. Renshaw, H. Sverdrupand S. Syri, Assessment of the Effects of the EU Acidification Strat-egy: Dynamic Modelling on the Integrated Monitoring Sites (FinnishEnvironment Institute, Helsinki Finland, 1997).

[16] J.P. Hettelingh and M. Posch, Critical loads and dynamic assessmentof ecosystem recovery, in: Predictability and Nonlinear Modellingin Natural Sciences and Economics, eds. J. Grasman, and G. VanStraten (Kluwer Academic Publishers, Dordrecht, The Netherlands,1994) pp. 439–445.

[17] M. Holmberg, Model of ion dynamics and acidification of soil: Sim-ulating recovery of base saturation, in: Environmental Models: Emis-sions and Consequences, eds. J. Fenhann, G.A. Mackenzie and B.Rasmussen (Elsevier, Amsterdam, The Netherlands, 1989) pp. 359–

368.[18] G. Klaassen, Cost of controlling ammonia emissions in Europe, SR-

91-02, International Institute for Applied Systems Analysis, Laxen-burg, Austria (1991).

[19] J. Nilsson and P. Grennfelt, Critical loads for sulphur and nitrogen,Report from a Workshop Held at Skokloster, Sweden, 19–24 March1988, 1988-15, Copenhagen, Denmark (1988).

[20] S. Oden, The acidification of air and precipitation and its conse-quences in the natural environment, Report number 1, Swedish Na-tional Science Research Council, Stockholm, Sweden (1968).

[21] OECD, A Tentative Analysis of some Data Concerning Long RangeTransport of Air Pollutants (Paris, France, 1978).

[22] M. Posch, J.P. Hettelingh, P.A.M. De Smet and R.J. Downing, Cal-culation and mapping of critical thresholds in Europe: Status Report1997, 259101007, National Institute of Public Health and the Envi-ronment, Bilthoven, The Netherlands (1997).

[23] G.J. Reinds, M. Posch and W. De Vries, A semi-empirical dynamicsoil acidification model for use in spatial explicit integrated assesse-ment models for Europe, Alterra Document 84, Wageningen, TheNetherlands (2000).

[24] K. Rennings and H. Wiggering, Steps towards indicators of sustain-able development: Linking economic and ecological concepts, Eco-logical Economics 20(1) (1997) 25–36.

[25] E.C. Schmieman, Acidification and tropospheric ozone in Europe:

Towards a dynamic economic analysis, Ph.D. thesis, WageningenUniversity, Wageningen, The Netherlands (2001).

[26] E.C. Schmieman and E.C. Van Ierland, Dynamics of soil acidification:An economic analysis, Ecological Economics 31(3) (1999) 449–462.

[27] H. Sverdrup and P. Warfinge, The effect of soil acidification an thegrowth of trees, grass and herbs as expressed by the (Ca+Mg+K)/Alratio, no. 2 1993, Reports in Ecology and Environmental Engineering,Lund University, Lund, Sweden (1993).

[28] United Nations, The 1979 Geneva Convention on Long-Range Trans-boundary Air Pollution, United Nations, Geneva Switzerland (1979).

[29] United Nations, Protocol to the 1979 Convention on Long-rangeTransboundary Air Pollution on Further Reduction of Sulphur Emis-sions and Decision on the Structure and Function of Implementa-tion Committee, as well as Procedures for its Review of Compliance,ECE/EB.AIR/40, United Nations, Geneva Switzerland (1994).

[30] United Nations, Protocol to the 1979 Convention on Long-RangeTransboundary Air Pollution to Abate Acidification, Eutrophica-tion and Ground-level Ozone, ECE/EB.AIR/1999/1, United Nations,Geneva Switzerland (1999).

dynamic cost-effective reduction strategies for acidification in europe: an application to ireland...

Documents