Interaction between Local Air Pollution and Global warming
and its Policy Implications for Belgium

Stef Proost

Denise Van Regemorter

CES KULeuven

Naamsestraat 69, 3000 Leuven, Belgium

e-mail:

Abstract

A policy to reduce greenhouse gas emissions has not only an impact at global level but can also bring benefits locally by reducing other air pollutants linked to energy consumption. Moreover the benefits of the reduction of local air pollutants will accrue to the current generation contrary to the reduction of climate change and also mostly to the population undertaking the mitigation actions, though the transportation of pollutants can be rather extensive. In view of the Kyoto agreement and the target it implies for Belgium, it seems important to take these benefits into account for a more correct evaluation of the cost of GHG reduction policies. Moreover Belgium has signed different agreements to reduce the more local pollutants linked to energy. Therefore, a policy fully exploiting these mutual benefits and reaching the targets for the environmental problems considered is desirable. This paper explores with the Belgian MARKAL model what are the implications of the interactions between pollutants for policy design. MARKAL is a dynamic energy optimisation model, that allows to compute the external cost and benefits of pollutants linked to energy consumption, such as CO2, NOx, SO2, VOC and PM, inclusive their abatement cost and to integrate in the policy evaluation the interactions between pollutants and their abatement. It goes beyond the simple approach to evaluate the secondary benefits/costs of a policy from the decrease/increase of the emissions of other pollutants. The importance of defining policies integrating the interactions between pollutants is evaluated for Belgium by comparing a policy addressing local pollution, a policy addressing global warming and a policy combining both.

Keywords: global warming, ancillary benefits, policy design

This research is financed under the 'Global Change' research program of the Belgian Prime Minister's Office – Federal Office for Scientific, Technical and Cultural Affairs.

Table of Contents

1.Introduction......

2.The modelling framework for MARKAL......

2.1.Basic modelling framework......

2.2.Approach for the implementation in MARKAL......

3.Database Extension......

3.1.Emission coefficients and abatement possibilities......

3.2.Coefficients for the transformation and transport of emissions......

3.3.Damage Parameters and their Monetary Valuation......

3.3.1.Impact on public health......

3.3.2.Impacts on territorial ecosystems and materials......

3.4.Damage from emissions in Belgium......

4.Illustration with Policy Scenarios......

4.1.Definition of the policy scenarios......

4.2.The scenarios comparison......

5.Conclusion......

1

1.Introduction

A policy to reduce greenhouse gasses emissions has not only an impact at global level but can also bring benefits locally by reducing other air pollutants linked to energy consumption. Moreover the benefits of the reduction of local air pollutants will accrue to the current generation contrary to the reduction of climate change and also mostly to the population undertaking the mitigation actions, though the transportation of pollutants can be rather extensive. In view of the Kyoto agreement and the target it implies for Belgium it seems important to take these benefits into account to get a more correct evaluation of the cost of GHG reduction policies. Moreover Belgium has signed different agreements to reduce the more local pollutants linked to energy. Therefore, it might be useful to induce a choice of policy, which fully exploit these mutual benefits, while obtaining the desired impact on the environmental problems considered.

The objective of this paper is to explore with the Belgian Markal model[1] what are the implications of the interactions between pollutants for policy design. MARKAL is a dynamic energy optimisation model, that allows to compute the external cost and benefits of pollutants linked to energy consumption, such as CO2, NOx, SO2, VOC and PM, inclusive their abatement cost. By adding to MARKAL objective function a damage function, it is possible to integrate in the policy evaluation the interactions between pollutant and their abatement. The damage function will depend on the immissions from Belgian origin, i.e. the change in pollutant concentration and in pollutant deposition, the impact of these changes on health and on the ecosystem and their monetary valuation. The development of MARKAL in this direction allows to take the external impact of all air pollutants explicitly into account, when evaluating long term options for air pollution reduction policies. This goes beyond the simple approach to evaluate the secondary benefits/costs of one policy through the decrease/increase of the emissions of other pollutants.

In the first section, we formulate the basic structure of a model where the damage from pollution is internalised and the adaptations to be made to MARKAL. Then, in the second section we give a description of the extension of the database of MARKAL with three types of data: emission coefficients for the main air pollutants and their abatement technologies, data for the modelling of the transport of pollutants and their impact on air quality and finally the damage valuation data. Finally, a first analysis is done with the adapted model to evaluate the importance of defining policies integrating the interactions between pollutants. We look at a policy addressing local pollution, a policy addressing global warming and a policy combining both.

2.The modelling framework for MARKAL

MARKAL is a dynamic optimisation model that represents all energy demand and supply activities and technologies for Belgium with a horizon of up to 40 years, with their associated emissions (CO, CO2, SO2, NOx, VOC and PM). The environmental problems considered in this study, global warming and local air pollution, are both linked to energy consumption and their abatement possibilities are interrelated. This interaction has to be integrated in the modelling framework for a correct policy evaluation. The basic modelling framework is given in the first section to show what the interaction implies for policy evaluation. Then the implementation in MARKAL is described.

2.1.Basic modelling framework

For illustration purposes, we consider the case of two pollutants linked to energy, CO2 and SO2, where CO2 can only be abated through a reduction in energy consumption but SO2 has specific abatement technologies[2]. In a partial equilibrium framework, the problem for the policy maker deciding on pollution reduction policies, can be represented as a maximisation of the consumer and producer surplus, taking into account the production possibilities, the damage from pollution and the abatement possibilities:

(1)

under the production constraint

(2)

wheredemand for an energy service, e.g. heating

the consumer/producer surplus (surface under the demand curve)

the price of capital and energy

the price of the energy service (shadow price of the constraint)

the production inputs, annualised capital and energy

cost of SO2 emission abatement for a reduction of

damage from SO2 and CO2 emissions, assumed constant here[3]

emission coefficients of SO2 and CO2 per energy unit

At the optimum, the first order conditions imply that the marginal utility of the energy service q be equal the price p, the marginal productivity of capital equal to the price of capitaland for energy and emission reduction:

  • the marginal productivity of energy equal to the cost of energy plus the damage from SO2 and CO2 emissions

(3)

or, in terms of ton CO2 abated[4], the marginal abatement cost equal to the damage from SO2 and CO2 taking into account the SO2 abatement

(4)

Because of the interaction between SO2 and CO2 reduction (through energy consumption), the optimum abatement effort for CO2 takes into account both the damage from CO2 and SO2. It will therefore be higher than when there is no interaction. It has also implications for the choice of policy instrument: the policy instrument has to give the incentive to internalise this interaction.

  • the marginal cost of SO2 emission reduction level equal to the damage from the (unabated) SO2 emissions at the optimum

(5)

or, in terms of ton SO2 abated, the marginal abatement cost equal to the damage from SO2 (marginal damage is assumed constant)

(6)

This is the classic result when only one pollutant is considered because neither the SO2 abatement options neither the SO2 emissions have an impact on CO2 emissions or damage.

This basic framework can easily be extended to more interactions between pollution damage and pollution abatement, but it already shows clearly the importance for environmental policy design to take these interactions into account.

2.2.Approach for the implementation in MARKAL

The objective was to adapt MARKAL to be able to take into account in the analysis of policy options the benefits/costs coming from the pollution interactions. The local environmental problems considered are:(i) problems related to the deposition of acidifying emissions and (ii) ambient air quality linked to acidifying emissions and ozone concentration. We consider the energyrelated emissions of NOx, SO2, VOC and particulates, which are the main source of air pollution. NOx is almost exclusively generated by combustion process, whereas VOC’s are only partly generated by energy using activities (refineries, combustion of motor fuels); other important sources of VOC’s are the use of solvents in the metal industry and in different chemical products.

The approach followed for the evaluation of the benefits from the reduction of local pollutants is based on the bottom up damage function approach as developed by the ExternE project. This approach can be illustrated by the following figure (EC, 1995).

EMISSIONS

DISPERSION

IMPACT

DAMAGE VALUATION

It implies an extension of the database towards the different datas needed for the evaluation of the benefits, described in the next section, and the construction of a damage function.

The damage per pollutant or damage function is modelled as follows:

whereEV (env) reflects the relation between the damage cost and the level of emissions; at this stage, to keep the model linear, the damage per unit of emission or the marginal cost per emission is kept constant, i.e. EV (env)= 1,

EVcoef(env) is computed through calibration with the marginal cost, i.e. the derivative of the damage function w.r.t. emission, equal to the value per unit of emission derived from the ExternE results, as explained in the next section.

The sum of the damage-functions per pollutant is added to the objective function and therefore taken into account in the optimisation process. When the damage function is used but not added to the objective function, it allows to compute the environmental damage generated by a policy, without feedback into the optimisation process.

As the computation are based on dose response functions which give the incremental damage from air pollution, the results should also be interpreted in these terms, i.e. in terms of the change in total damage compared to a reference year (the base year).

3.Database Extension

The Markal database has been extended into three directions:

a)emission coefficients for pollutants such as NOx, SO2, VOC and PM, and the emissions abatement technologies,

b)immission coefficients for those pollutants, i.e. coefficient for the translation of emissions into concentration, inclusive the transportation mechanism,

c)impact of emissions and immissions and their monetary valuation.

3.1.Emission coefficients and abatement possibilities

Emission coefficients of NOx, SO2, VOC and PM associated with the energy using technologies have been added to the Belgian Markal database. This is a rather extensive work because, contrary to CO2 emissions coefficients, these coefficients are technology linked. Also the technological options to abate those emissions were added to the database.

3.2.Coefficients for the transformation and transport of emissions

This step establishes the link between a change in emissions and the resulting change in concentration levels of primary and secondary pollutants. The transboundary nature of pollutants leads to the necessity to account for the transport of SO2, NOx, VOC and particulates emissions between countries. In the case of tropospheric ozone (a secondary pollutant), besides the transboundary aspect, the relation between VOC and NOx emissions, the two ozone precursors, and the level of ozone concentration has also to be considered.

Theoretically, the concentration/deposition (IM) at time t of a pollutant ip in a grid g is a function of the total antropogenic emissions before time t, some background concentration[5](BIM) in every country, and other parameters such as meteorological conditions, as derived in models of atmospheric dispersion and of chemical reactions of pollutants:

For the model, the equations are made static and the problem is linearised through transfer coefficients TPC which reflect the effect the emitted pollutants in the different countries have on the deposition/concentration of a pollutant ip in a specific grid, such as to measure the incremental deposition/concentration, compared to a reference situation:

where TPC[g,c] is an element of the transport matrix TPC with dimension GxC. In the model here the grid considered is a country and deposition/concentration levels are national averages.

The transport/deposition coefficients for SO2 and NOx emissions are derived from EMEP budgets for airborne acidifying components which represents the total deposition at a receptor due to a specific source. Basically, the EMEP model is based on a receptor orientated one layer trajectory (Lagrangian) model of acid deposition at 150 km resolution. Characteristics of the various pollutants and their transport across countries, as well as atmospheric conditions are taken into account. For particulates, Mike Holland (ETSU, 1997) has estimated country to country transfers of primary particulates. His computations are based on a simple model which accounts for the dispersion of a chemically stable pollutant around a source, including deposition by wet and dry processes. To convert deposition into air concentration, use was made of linear relations estimated by Mike Holland (1997).

Tropospheric ozone is a secondary pollutant formed in the atmosphere through photochemical reaction of two primary pollutants, NOx and VOC. The source-receptor relationship is not as straightforward as for acid deposition. However, it is recognised (EMEP, 1996) that there is a relatively strong linearity between change in ozone concentration and change in its precursors emissions (both VOC and NOx), allowing an approximation through linear source-receptor relationships.

It would be useful to include the distinction in the source of emission, for instance between emissions from mobile sources and/or low height stationary sources as opposed to high stack sources as it is expected that the deposition of pollutants per unit emitted will be different in each case. However, there is no information available at this moment that allows making such distinction.

3.3.Damage Parameters and their Monetary Valuation

The damage parameters and their monetary valuation are taken from the ExternE project of the European Commission, in which CES and VITO were responsible for its Belgian application. Therefore the approach followed here is entirely based on the framework derived in the project, though at a much more aggregated level. The damage occurs when primary (e.g. SO2) or secondary (e.g. SO4--) pollutants are deposited on a receptor (e.g. in the lungs, on a building) and ideally, one should relate this deposition per receptor to a physical damage per receptor. In practice, dose/exposure-response functions are related to (i) ambient concentration to which a receptor is submitted, (ii) wet or dry deposition on a receptor or (iii) ‘after deposition’ parameters (e.g. the PH of lake due to acid rain). Following the ‘damage or dose-response function approach’, the incremental physical damage DAM per country is given as a function of the change in deposition/concentration (acidifying components or ozone concentration in the model),

The damages categories considered in the model are

  1. damage to public health (acute morbidity and mortality, chronic morbidity, but no occupational health effect)
  1. damage to the territorial ecosystem (agriculture and forests) and to materials, this last category being treated in a very aggregated way at this stage.

The impact on biodiversity, noise or water is not considered, either because there are no data available that could be applied in this study or because air pollution is only a minor source of damage for that category.

For the monetary valuation of the physical damage, a valuation function VAL for the physical damage is used:

The economic valuation of the damage should be based on the willingness-to-pay or willingness to accept concept. For market-goods, the valuation can be performed using the market price. When impacts occur in non-market goods, three broad approaches have been developed to value the damages. The first one, the contingent valuation method, involves asking people open- or closed-ended questions for their willingness-to-pay in response to hypothetical scenarios. The second one, the hedonic price method, is an indirect approach, which seeks to uncover values for the non-marketed goods by examining market or other types of behaviour that are related to the environment as substitutes or complements. The last one, the travel cost method, particularly useful for valuing recreational impacts, determine the WTP through the expenditure on e.g. the recreational impacts.

It is clear that measuring environmental costs at the global level as in this model, raises different problems, which are extensively discussed in ExternE: transferability of the results from specific studies, time and space limits, uncertainty, the choice of the discounting factor, the use of average estimates instead of marginal estimates and aggregation. However, despite all these uncertainties, it is possible, according to ExternE, to give an informative quantified assessment of the environmental costs.

3.3.1.Impact on public health

The ExternE project retains, as principal source of health damages from air pollution, particulates[6] resulting from direct emission of particulates or due to the formation of sulphates (from SO2) and of nitrates (from NOx), and ozone. They retain also a direct effect of SO2 but no direct impact of NOx because it is likely to be small. Direct damages from HC are not yet considered here, because the ExternE figures are still at a preliminary stage. The assessment of health impacts is based on a selection of exposure-response functions from epidemiological studies on the health effects of ambient air pollution (both for Europe and the US). They are reported in the ExternE report (1997) and summarised hereafter. For sulphates, the dose-response functions associated with PM2.5 are taken into account, whereas for nitrates the dose-response functions associated with PM10 are used. ExternE recommends also using the E-R functions related to PM2.5 for the particulates with as primary source transport. When chronic mortality impacts are explicitly accounted for, one must exclude the acute mortality impacts because they are already considered in the former (Hurley et al., 1997).