Co-authors: Berry Meuleman, Wim Turkenburg, Ad van Wijk, Ausilio Bauen, Frank Rosillo-Calle, David Hall.

Accepted for publication in 'Biomass and Bioenergy'.

Abstract - Externalities of electricity production from biomass and coal are investigated and compared for the Dutch context. Effects on economic activity and employment are investigated with help of Input/Output and multiplier tables. Valuations of damage from emissions to air are based on generic data from other studies. In addition external costs are estimated for nitrogen leaching and for the use of agrochemicals for energy crop production.

The average private costs for biomass and coal based power generation are projected to be 68 and 38 mECU/kWh respectively in the year 2005. It is assumed that biomass production takes place on fallow land. Coal mining is excluded from the analysis. If the quantified external damages and benefits are included the cost range for bio-electricity is 53-70 mECU/kWh and 45-72 mECU/kWh for coal.

Indirect economic effects (increment of Gross Domestic Product) and the difference in CO₂ emissions are the most important distinguishing factors between coal and biomass in economic terms. Damage costs of other emissions to air (NO_x, SO₂, dust and CO) are of the same order of magnitude for both coal and biomass (coal mining excluded). In this analysis environmental impacts of energy farming are compared mainly to fallow land focused on the use of fertilizers and agrochemicals. The related damage costs appear to be low but should be considered as a preliminary estimate only.

The quantitative outcomes should not be considered as the external costs of the two fuel cycles studied. Many impacts have not been valued and large uncertainties persist e.g. with respect to the costs of climate change and numerous dose response relations. More detailed analysis is desired with respect to macro-economic impacts. The results serve as a first indication, but the outcomes plead for the support of bio-electricity production and/or taxation of coal based power generation.

Integrated bio-energy systems which rely on specifically produced energy crops are at present not competitive with fossil fuel based energy systems. The situation might change in the somewhat longer term, as a result of the further development and improvement of energy crops, conversion technology and integrated production systems.

Nevertheless, the use of biomass from energy crops to fulfil our energy needs could already now offer several advantages compared to the consumption of fossil fuels. Biomass is a renewable energy source with zero or low net carbon dioxide emissions when the biomass is produced in a sustainable way. Furthermore, the SO₂ emissions could be low because of the inherently low sulphur content of biomass. In addition, biomass could be an alternative crop for farmers in Europe and the USA, where the agricultural sector produces large food surpluses and where arable land is being set-aside. This land can be used to grow energy crops. Consequently, biomass can be an employment generator, especially in rural areas where additional employment is often desired. Biomass would also be an indigenous energy source for most countries; this may improve the reliability of the energy supply and result in additional economic benefits if it replaces the import of fossil fuels. Finally the growth of crops to produce energy can have environmental impacts. These impacts could be negative (compared to fallow land) but also positive if e.g. the growth of SRC (Short Rotation Coppice) or perennial grasses to produce energy would replace the production of food crops.^19,21,44,63 However, the use of biomass to produce energy can also have negative effects, like the emission of pollutants and increase in road transport. Those are all effects of bio-energy production that are not accounted for in the calculation of the private costs of bio-energy use: externalities. A market externality occurs when some costs (or benefits) of a market transaction are borne (or received) by parties not directly involved in the transaction. Externalities, therefore, represent costs or values that are not directly reflected in the market pricing. Accounting for the expected beneficial overall effects of bio-energy could improve its competitiveness with fossil fuels.

One of the first attempts to value the external effects of energy consumption was made by Hohmeyer in 1988.²⁴ To determine the external costs of environmental emissions, he applied a top down approach. Starting from estimates of the total damage done to the environment and human health by air pollutants,⁵⁹ he calculated the damage done by an energy source by determining the contribution of that source to the air pollution. This work was updated in 1991.²⁵ Hohmeyer's study was criticized amongst others because his investigation was based on relatively old estimates of damage caused by environmental emissions. It was also criticized because it did not focus on the marginal effects of those emissions or take into account the current levels of pollution from other sources and the non-linearity of many dose-response curves.

Since 1988 many other studies on the external effects of energy consumption have been done. An extensive follow-up of these studies is currently being undertaken by the Fuel Cycle Studies of the USA and EC.^4,34 In these studies the so-called Damage Function Approach (DFA) is followed in order to investigate the impact of a number of fuel cycles on the environment. DFA can be characterized as a bottom-up approach. It starts by determining the increment in emission levels due to a new activity (e.g. a fuel cycle), then goes on to assess physical impacts (health effects, damage to crops, etc.) followed by an economic valuation of these impacts. This leads to damage estimates of the marginal effects. Drawbacks to this approach are that the results are often strongly site specific and rely on numerous dose-response curves. Therefore, the approach is very data intensive.^4,34 Due to the lack of data often only a part of the actual external effects can be valued. Consequently, the uncertainty in the outcomes could be considerable. As a result the reliability of external cost evaluations and the way they should be used in policy making is heavily debated.^23,54,57

Various methods have been developed for pricing external effects. An overview is given by Pearce et al.⁴⁰ A brief summary has been made Munasinghe.⁶⁴ One approach is to look at the impact of environmental changes on directly observable behaviour valued in conventional markets (mitigation costs of damage; loss of output due to sickness or death; loss of production, etc.). Another approach is to calculate the replacement costs of a shadow project to offset the environmental damage. A third set of methods uses indirect or surrogate market data to determine the value (willingness to pay for bottled water; willingness to pay for travel to nature areas; wage premium to compensate for hazardous circumstances, etc.). A final group of methods simulates market-like behaviour, using marketing experiment or surveys to determine the value. An example is contingent valuation: the willingness to pay for an environmental asset or to accept compensation for its loss, each determined by direct questions.^40,64

Pearce et al. do not advocate any method. The best choice depends on the specific case, desired reliability and availability of data. However, Munasinghe indicates that direct market data are preferable to indirect market data and especially data obtained from simulated market behaviour. Other authors argue that the contingent valuation method (CVM) should be restricted to specific applications²³ and should not be used to value uncertain, non-visible, long-term effects such as climate change or loss of biodiversity.⁵⁷

A specific method for valuing the costs of environmental emissions is the abatement costs approach. In this approach the cost of avoiding certain emissions is calculated. An example is the cost of sulphur removal at coal plants which then serves as a measure of the external costs of SO₂ emissions. The main advantage of this approach is its pragmatic character and relatively exact results. Another benefit is that it allows for least cost planning e.g. for utilities. For a more detailed discussion, see van Wijk et al.⁶⁰ However, a major drawback of taking abatement costs as a basis for external cost evaluations is that the results have no relation to the actual damage done to human health and the environment.

In the literature on externalities, bioenergy fuel cycles have received fairly limited attention, especially with respect to socio-economic effects. Ribera et al.⁴⁵, the National Technical University of Athens³⁶ and Oak Ridge National Laboratory³⁸ (the latter is still ongoing⁴²) have performed studies on the externalities of biomass. The external effects studied cover the economic impact of emissions to air, nitrogen leaching, erosion, employment, damage to roads and occupational risks.

The analysis of the the biomass and the coal fuel cycle will focus on a national scale. The functional unit for comparing the systems will be one kWh. The focus on an evaluation of the external costs on a national scale is based on the notion that the outcomes could be especially meaningful for policy-making by governmental bodies. Decision-making at national level is at present a likely way for internalizing external effects in the costs of energy. Furthermore, a variety of (potential) external effects are location specific. In the case of decision making with regard to energy supply in the future, the indirect consequences for the national economy can be a crucial factor. These consequences can be a justification for applying governmental incentives such as taxes or subsidies. It should be noted that the environmental impact of coal mining will not be included in our analysis since coal mining has ceased in the Netherlands. However, the socio-economic effects of importing coal will be taken into account.

One reason for focusing on coal for comparison is that the bio-energy system studied is expected to operate in baseload power generation, just like coal fired power plants in the Netherlands. Another reason is that an important argument for realising biomass based power generation in the Netherlands is the reduction of CO₂ emissions. Replacement of coal based power generation can substantially contribute to this objective.

Relevant impacts that may lead to external/indirect costs or benefits which will be considered here are: emissions to air (NO_x, SO₂, dust, CO) result in impacts on nature, human health and the built environment. Emissions of CO₂ may contribute to climate change. Energy crop production may result in emissions to water and soil. Also impacts on water-use, landscape and biodiversity are anticipated. In both fuel cycles expenditures on investments, fuel, operation and maintenance lead to indirect economic effects, both on the GDP and with respect to employment generation.

However, at the present moment it is impossible to value all the external effects of fuel cycles. For example, issues like loss of nature and changes in biodiversity, depletion of resources and energy security are hard to value in financial figures. Nevertheless, although there are problems and uncertainties, economic valuation of environmental damages and socio-economic effects is useful since it visualizes, at least partly, what the economic impact of a (new) activity could be. Since the economic attractiveness of a given project is generally a major criterion in decision making, economic valuation of external effects is important, which is why this valuation is carried out in this study.

In this article we begin by briefly discussing the external effects that will be investigated and the methods that will be applied to value these effects in monetary terms. This is done in section 2. The core of the paper deals with the application of these methods in order to investigate and compare the external costs and benefits of power generation from indigenously produced biomass and from imported coal. Section 3 describes the systems considered. Section 4 goes into the actual valuation of the externalities. The meaning and usefulness of the results will be discussed in section 5. Finally we will draw some draw some conclusions from this exercise.

The fuel cycles we want to investigate in this article - electricity production from biomass and from coal in the Netherlands - involve numerous, (potential) external effects. Here we describe which effects will be investigated and which method will be used to value these effects.

Table 1 gives an overview; the first column of the table lists the possible indirect socio-economic effects and environmental external effects of the fuel cycles. The second column briefly summarizes the possible valuation methods of the external effects. Also the main characteristics, advantages and disadvantages are given. The third column mentions which valuation method will be applied in this study and why. Because the methods are at different stages of development, lack of data and because the effects considered are so different, more than one method is needed to value all the effects mentioned.

In this study special attention is given to indirect socio-economic effects of the fuel cycles, in particular effects on the Gross Domestic Product and employment, as these effects may be of major importance for the competitiveness of the fuel cycles. We will try to determine the effects, using Input/Output modelling. This approach allows the quantification of multiplier effects, both on GDP and on the generation of employment.

The Damage Function Approach (DFA) can be followed for the cost evaluation of emissions to air. It is however beyond the scope of this work to perform detailed dispersion modelling or to provide total damage estimates of these emissions. Therefore, we apply generic data determined with a DFA for the Dutch context. For comparison we will do the same with generic data that stem from Hohmeyer. Using two sources with a basically different approach enables a comparison to be made between the two outcomes. The results and comparison will be discussed in detail in chapter 4.

A valuation attempt will be made for two other effects related specifically to energy farming: nitrogen leaching due to the use of fertilizers and environmental impacts due to the use of agrochemicals. However, on these subjects DFA-studies have not yet been performed. For nitrogen leaching it is unclear to what extent leaching in a certain area actually contributes to exceeding standards e.g. for groundwater quality. Effects of fertilizer use for the production of energy crops are dependent on specific conditions and even if these are known determining physical effects can be problematic. We therefore adopt Willingness To Pay (WTP) data with regard to the desire of people to have access to clean groundwater in order to determine the costs of nitrogen leaching. To value the environmental impact of agrochemical use, a shadow-price is calculated based on the reduced agricultural production when fewer or no agrochemicals are used. Both nitrogen leaching and agrochemical use are included in this analysis to obtain a feeling for the order of magnitude of possible damage caused by energy farming. However, on forehand it is recognized that further research on this subject is desired.

For the Dutch context further effects on soil quality, erosion, hydrology, nature and biodiversity due to biomass production for energy are too uncertain to assess, since the actual impacts of energy farming are hardly known yet. From experience in other countries, however, we expect that in the case of SRC Willow (minor) benefits are achieved compared to growing conventional food crops.

The Netherlands is a densily populated country with high pressure on available space. Agriculture is very intensive and has both the highest yields and highest fertilizer and pesticide use per hectare in the world.¹³ Farmers recieve a relatively high income, also per hectare compared to other (European) countries. Since the high cost of land and labour will result in relatively high biomass energy costs, it will be difficult for biomass to compete with other energy sources, especially cheap natural gas. Indeed commercial crop production for energy purposes is currently not competitive with fossil fuels.³¹

The availability of land to produce biomass for energy purposes in the Netherlands is discussed elsewhere in a study by Faaij et al.¹⁶ It is concluded that, despite the demand for land for a number of other functions, a significant amount of land may come available for energy farming provided that energy farming is an economically viable activity for farmers. Land could become available because of a decrease in land use for growing food crops and dairy farming due to expected productivity increments. The total amount of land available for energy crops taking into account other land demanding functions like urban and nature development, could lie between 110,000 and 250,000 ha in the year 2015. In this study we assume that energy farming takes place on fallow land.

Currently there is no commercial energy farming in the Netherlands. Therefore, to investigate its costs and benefits projections have to be used with respect to the performance, yield and costs of biomass production systems in the Dutch context. The potential of energy crop production and the conversion of biomass energy to electricity are discussed below, assuming a time frame of about 5-10 years.

The conversion unit of the selected bio-energy system is assumed to be a 30 MW_e Atmospheric Circulating Fluidized Bed gasifier coupled to a General Electric LM 2500 gas turbine and equipped with conventional low temperature gas cleaning and a flue gas dryer using waste heat. The projected performance of this system is discussed in Faaij et al.¹⁴ and Consonni et al.⁹ The efficiency is expected to be about 42% (LHV; using clean wood with 50% moisture). Investment costs are expected to be in the range 1000-2300 ECU/kW_e; the lower value is expected to represent the situation of a fully proven commercial plant. O&M costs are considered to be about 2.1 MECU/year.¹⁴

Short Rotation Coppice (SRC) Willow is selected as the energy crop. A considerable amount of experience has been gained with this crop in energy production in Sweden. Willow is suited for most Dutch conditions and shows a favourable energy ratio (energy input:energy output = 1:10-15).³¹ Projections for the yield of willow production have been published in the literature. Estimates vary from 8 to 15 tonnes dry matter per hectare per year.^31,35,63 The yield depends on the e.g. soil type, production system and crop variety. An average yield of 12 odt/ha/yr is considered obtainable within 5-10 years. Such a yield is based on crops harvested at intervals of 3 years and a plantation lifetime of about 24 years. Other rotation schemes are however possible (2 to 5 year rotations). At the present stage it is not fully known what production system is best in economic terms.

Biomass is to be transported from the fields to the conversion unit. Optimal logistics are a complex matter and for accurate analysis detailed data of the production area and the road network are required.³ Here a simple approach will be adopted:

It is assumed that biomass will be transported by road. Willow is harvested and transferred to trucks (25 tonnes load) at the side of the production fields and transported directly to the conversion facility. Central, covered storage at the conversion facility is assumed. The conversion unit is assumed to be located in the centre of an area where energy farming takes place. In this study we assume that about 10 to 30% of the land around the plant will be used for energy farming. Also we assume that the chips are used in a power plant with load factor of 80%. On the basis of these assumptions, we can calculate the total number of hectares required (9400 ha), the average transport distance (7.1-12.2 km), and total number of transport kilometres required per year (127,000-220,000 km including return distance). For simplicity, dry matter losses during storage are not taken into account. Such losses can be minimized by covered storage and, partial, storage in the form of sticks.³ Data regarding the cost and energy use of biomass transportation are given in Feenstra et al.¹⁷

On the basis of these assumptions we have calculated the overall production cost per kWh for the entire system. Some details of these calculations and the results are presented in table 2. Major variables influencing the Cost Of Electricity (COE) are the farmers' income, the investment cost of the conversion unit and the applicable subsidies. Other important parameters are the crop yield and the energy conversion efficiency.

In our study three cases for crop production have been determined with subsidy levels between 0 and 700 ECU per ha/yr. Details of the economic aspects of SRC Willow production under Dutch conditions are discussed in Rijk.⁴⁶ The selected cases represent likely situations in the near future.

An interest rate of 3% is applied to all related investments and cash flows; this is the rate which is applicable at the moment in the agricultural sector.⁴⁶ With respect to investments for the conversion unit a depreciation period of 25 years and an interest rate of 5% are assumed. The resulting COE from biomass ranges between 4.0 and 10.5 ECUct/kWh. The lowest value may be obtained in the longer term (e.g. 10-15 years), the highest value represents the current situation assuming that BIG/CC technology has been demonstrated commercially.

For further calculations we use the 'average' case described in table 2: in this case investment costs for the power plant are assumed to be relatively high (i.e. 2000 ECU/kW_e): furthermore, the fallow land subsidy of 500 ECU/ha.yr is assumed to be still available, since we consider the production of biomass on fallow land to be feasible. This case results in a COE of 6.8 ECUct/kWh.

For comparison, the COE of coal based power generation in the Dutch context is (projected to be) 3.8 ECUct/kWh (those costs are obtained with a load factor of 70%).^2,49,55 Despite possible changes in technology such as IG/CC instead of combustion, those COE are not expected to drop much. The world market coal price is expected to remain stable for long periods. The projected direct costs of bio-electricity are therefore expected to be roughly twice the COE produced from coal.

The emissions to air from crop production, harvesting and transport are calculated by using generic data for diesel engines and multiplication by the total energy use for both transport and crop production operations. Total emissions from the biomass conversion unit are calculated from the total fuel gas flow and projected emission levels.¹⁴ The use of trucks will also result in emissions of NOx, SO₂, dust and CO and hydrocarbons. In this

study the emission factors for trucks are derived from Perlack et al.⁴¹ Results and relevant details are given in table 3. The expected emissions for coal based power generation with modern power plants that will be in operation 5-10 years from now are given in the last column. These emissions are based on pulverized coal combustion plants equipped with extensive flue gas cleaning.^18,61 Also, the emission data for coal based power generation in an Integrated Gasification/Combined Cycle, a likely future coal conversion technology, are given.

First the direct expenditures by the coal and biomass fuel cycles are estimated. The expenditures related to the biomass energy system are extracted from studies which describe energy farming of willow and biomass gasification for electricity production.^14,46,56 This breakdown of expenditures and the allocation over various sectors are described in detail in a background report of this analysis.¹⁵ The expenditures of the coal energy system are based on the costs of operating a modern 600 MW_e coal-fired power plant.^2,55,62 These expenditures are attributed to appropriate economic sectors as taken up in the input/output table. The expenditures for biomass fuel production, and for the operation and maintenance and investment costs of both fuel cycles is assumed to take place within the Dutch economy. Only the expenditures for coal are expected to take place outside the Dutch economy, since coal is imported into the Netherlands.

Biomass production is assumed to take place on fallow land for which a fallow land subsidy of 502 ECU/ha.yr is currently available.⁴⁶ The subsidy will be given to farmers in both the biomass and the coal case. The subsidy itself is however not included in the expenditures, also not in the biomass case. In the case of the coal fuel cycle, the agricultural land is assumed to be fallow and no other economic activities take place on this area of land.

We assume that the relation between production and value added will remain constant in the time-frame used in this analysis. As a first order approach this seems reasonable since a linear function is obtained for the period 1985-1994 as well. The indirect effects are calculated for an ideal context in which production capacity and labour are under-utilized. If economic activity is already high, inflatory effects may occur, such as increasing costs of labour and goods. Howeverm we assume that, considering the size of the projects, this will not directly lead to increased inflation and that production capacity is available in relevant economic sectors.

 ^a1 ECU = 2.15 Dfl,- (1995)

Because the expenditures for the biomass fuel cycle are considerably higher per kWh produced than for electricity production from coal it is logical that a higher indirect economic effect is calculated simply because of those higher expenditures. A correction is therefore made for the difference in the Costs Of Electricity (COE) from biomass and from coal (see note table 5). The net indirect effect of biomass based on this correction is shown in table 5. The biomass fuel cycle results in a GDP increment of 6-15 mECU/kWh, but the coal fuel cycle results in a negative increment between 7 and 8.4 mECU/kWh. This mainly due to the fact that coal imports contribute negatively to the GDP.

In the description of the energy crop production (see section 3) subsidies were mentioned that could apply to SRC. Currently subsidies are paid for setting land aside. These set-aside subsidies serve as income support to farmers and help prevent the production of surplus food crops. Those subsidies can be considered as an expenditure that will be made, regardless whether non-food crops are produced or not. Therefore, those subsidies have not been added to the costs of the bio-energy system and are not included when the multiplier effect is calculated.

 ^aCoal price = 1.34 ECU/GJ⁴⁹

bBiomass plant: 30 MW_e; 210 GWh/yr

 Effects on employment can be divided in direct employment, caused by the operation and construction of the power plants and fuel production, and indirect employment, generated in other sectors of the economy because of the expenditures that result from the implementation of the fuel cycle in question.

 Direct labour requirements of the biomass fuel cycle consist of the sum of labour requirements for crop production, construction, operation and maintenance of the conversion plant and for transportation of the biomass. For the coal fuel cycle the direct labour requirement concern the conversion (construction, operation and maintenance) and inland transport of coal.

 ^aSources: Rijk⁴⁶ and Faaij et al.¹⁴

bConstruction: 76.7 man year/TWh; O&M: 30 man year/TWh¹¹

Biomass obviously has a significantly higher labour intensity than coal, the difference being approximately a factor 2.6 per GWh produced. This can be explained partly by the fact that bioenergy production is more expensive. Moreover, coal mining is not included here since coal is imported into the Netherlands.

Economic valuation of employment is a rather arbitrary matter. In this study it is assumed that employment generated by the projects will always reduce the number of long term unemployed persons in the Netherlands (either directly or indirectly). The economic benefits are expressed in terms of unemployment benefits saved. These vary between 156 to 631 ECU/month, depending on various factors. Combining these figures with the employment generated results in a benefit of 0.83 to 4.04 mECU/kWh for the biomass fuel cycle (including the correction for the higher COE of biomass) and of 0.32 to 1.50 mECU/kWh for the coal fuel cycle. In these outcomes the variation through the applied range in interest rates is included.

 ^aResults for the biomass fuel cycle for a 30 MW_e plant producing 210 GWh/year.

The value of employment might however change in the future. In case unemployment is low an increased demand for labour might result in an upward pressure on wages, which may reduce employment in other sectors. In the current situation, however, employment generation is considered desirable and we assume that this will not change dramatically in the time frame considered in this study.

 4.3 Emissions to air

 Generic data from other externality studies are used to estimate the damage costs that result from various emissions to air. The results are summarised for both biomass and coal in table 8. Cost figures per unit of emitted contaminant (like NO_x, SO₂) from both top down (Hohmeyer^24,26) and bottom up (Dorland et al.¹¹) studies are applied. Data from Hohmeyer apply to the German context but are considered reasonably representative for the Netherlands as well. Data from Dorland et al.¹¹ are determined specifically for the Dutch context. They generated those data for a modern pulverized coal fired power plant with extensive gas cleaning in the vicinity of Amsterdam. In order to translate the results of these studies to the two cases in this exercise, the emissions to air per kWh caused by the biomass fuel cycle and the coal fuel cycle (see table 3) are combined with the derived cost data per weight unit of emission. Then the damage per emission per kWh produced for both fuel cycles is calculated. Full details are given in the background report.¹⁵ The main results of this exercise are given in table 8 for SO₂, NO_x and dust emissions.

Although the two approaches of Hohmeyer and Dorland et al. are very different in methodological terms, the numerical results are rather similar. It should be noted however that the two studies take different damage categories into account. The study by Dorland et al. focuses on effects to human health and excludes effects on flora and fauna. Chronic health effects play a minor role in the valuation of Hohmeyer (which can be considered as an omission). Dorland et al. give lower estimates for damage to materials (such as buildings). The focus on SO₂ in Dorland et al. and that the Hohmeyer data are based on damage in the German context, which are assumed to be higher per average building than in the Netherlands, are explanations.¹² The two studies are therefore difficult to compare; both merely give an illustration of quantifiable external costs. They also illustrate that the uncertainties in the overall outcomes of these studies are substantial.

 ^aThe damage costs of the coal fuel cycle are based on emission data for a modern pulverized coal plant with a conversion efficiency of 44%. A main alternative in 2000 could be Coal Integrated Gasfication/Combined Cycle (CIG/CC) technology as demonstrated on a 275 MW_e scale in Buggenum, the Netherlands. Performance data were given in table 2. However, the damage costs caused by emissions of NO_x, SO₂ and dust are of the same order of magnitude (0.99-5.79 mECU/kWh according to data of Hohmeyer and 1.68-4.90 mECU/kWh according to data from Dorland et al.¹¹; which is slightly higher than the case presented here). Although further improvements are possible it is not likely that large differences between pulverized coal and gasification plants will occur. Major advantages of coal gasification are however also related to limited waste water production, which is not taken into account in this study.

The CO₂ emissions per kWh as produced in the fuel cycle of coal and biomass is calculated at 815 and 24 g/kWh respectively. Damage estimates of greenhouse gas emissions that could cause a climate change with severe consequences for nature and mankind, differ widely in literature. Here we will apply recent estimates from a major source: the Intergovernemental Panel on Climate Change.⁸ The IPCC summarizes the results of numerous studies on this topic and presents figures of about 1-25 ECU/ton CO₂ (5-125 U$/ton C). The wideness of the ranges is mainly due to the different assumptions about the interest rates that should be applied to transfer future damage to present costs, differences in assumptions about future greenhouse gas concentrations in the atmosphere as projected in different scenarios and differences in the valuations of damage. Potential damage of climate change occurs on a global level and not only on a national scale (which is the level of investigation in this study). It can however be argued that the burden of potential damage cost by climatic change should be compensated in relation to the amount of greenhouse gases emitted. When those damage estimates are translated into external costs per kWh, this results in 0.02 - 0.59 mECU/kWh for the biomass fuel cycle and 0.83 - 20.73 mECU/kWh for the coal fuel cycle.

It should be kept in mind, however, that some authors suggest that higher damage costs of climate change are possible (see for example Sorensen⁶⁸).

 4.4 External effects related to energy crop production

 Nitrogen leaching

In North Western Europe and in the Netherlands in particular, the abundant use of fertilizers and manure in agriculture has led to considerable environmental problems: nitrification of groundwater, saturation of soils with phosphate, leading to eutrophication and problems is meeting drinking water standards. Also, the application of phosphates has led to increased heavy metal flux to the soil. Energy farming with short rotation forestry and perennial grasses, however, requires less fertilizer than conventional agriculture.^44,63 Data on nitrogen leaching occuring during the production of various crops and projections for short rotation coppice like Willow can be derived from Rijtema et al.⁴⁷ With help of model calculations they consider the annual rainfall, precipitation excess, deposition of nitrogen by rainfall, denitrification rates and uptake efficiency of nutrients by various crops under Dutch conditions. This leads to estimates for the amount of fertilizer that leaches from the soil into ground and surface water. These estimates are summarized in table 9 for willow and three relevant food crops. The leaching of nitrogen relating to Willow cultivation can be about a factor 2-10 less than for food crops.⁴⁷ Biewinga et al.⁶⁵ confirm that leaching levels of Willow production are far lower compared to major food crops. They present values for the total nitrogen gift and expected leaching levels for various crops as well (see also table 9).

 A method of valuing some of these effects has been proposed by Ribera et al.⁴⁵ The costs of the removal of nitrates in connection with production of drinking water is the basis of the economic valuation. However, the identification of causal relations between the production of a certain crop and the need to install denitrification processes is a difficult procedure and will for example depend on the degree of pollution of the groundwater from other sources, which are hard to quantify. Dispersion modelling and allocation of effects to other (agricultural) activities in a given area are too complex to handle. Therefore, to avoid the need for detailed modelling of hydrological aspects and leaching and the estimating of avoided costs, we apply an approach based on consumers Willingness To Pay for improved water quality with respect to nitrogen concentrations. Such data have been presented in a Swedish study.⁵¹ This study describes a WTP of 0.65-6.6 ECU per kg of N avoided in ground or surface water. It seems reasonable to transfer these data to the Dutch context, because of the similarities in environmental awareness and policies between the two countries. No data seem to be available about the willingness to pay to prevent the leaching of phosphate or other components and therefore no attempt has been made to value this aspect here.

 ^a Implies rainfall minus evapotranspiration by the crop.

Table 10 presents the results of applying the WTP cost data to the leaching of nitrogen in the case of Willow. Two different cases are distinguished. The first case assumes that Willow production takes place on fallow or marginal land (which follows our reference system). Here leaching always results in damage (assuming that no leaching takes place from fallow land). For the sake of illustration, in the second case figures are given of avoided damage by reduced leaching when Willow replaces food crops.

The cost ranges found are obviously very wide. In the first place the fertilizer use for Willow might vary. Furthermore, the difference between minimum and maximum WTP data is a factor 10.

 Standards have been set by the EU and the Dutch government in order to prevent the exceeding of nitrate levels in groundwater. To meet health standards for groundwater (i.c. 50 mg nitrate/litre) the nitrogen gift should not exceed 170 kg N/ha.yr (or, for soils with a high adsorption capacity, 210 kg N/ha.yr).⁶⁶ From the values given in table 9 it can be concluded that the nitrogen gifts for SRC-Willow are well below these standards. It is therefore doubtful whether the full cost range of WTP-data is valid for SRC-Willow production.

The Swedish WTP data are based on inquiries using the aforementioned health standards as a reference value. The study shows that the willingness to pay for measures that protect human health increases when (potential) health problems become more serious (and vice versa).⁵¹ This is however a qualitative result from the study and it is not known what the WTP is to reduce leaching to levels that would lower nitrate concentrations in groundwater further. It is therefore difficult to include this notion in a quantitative estimate of the damage costs of leaching caused by Willow production. However, the total nitrogen gift of SRC-Willow production remains well below health standards. We therefore suggest to use the lower value from the WTP cost range to obtain a damage estimate of nitrogen leaching of SRC-Willow production on fallow land. This discussion illustrates, however, that our understanding of the effects and seriousness of nitrogen leaching needs improvement and that the cost estimate of nitrogen leaching of SCR-Willow production should be seen as a first indication only.

 Agrochemical use

The agricultural use of pesticides can affect the health of people and the ground- and surface-water quality and consequently the flora and fauna. Specific effects strongly depend on the type of chemical, the quantities used and the method of application. However, it should be noted that not all effects causing damage to flora, fauna and human health - and the related dose-response curves - are well known or understood. Furthermore the potential number of dose-response curves is very large. Therefore, determining the external costs caused by the application of agrochemicals is very difficult if not impossible if one uses a damage function approach. Consequently, we will adopt another approach to estimate the external costs due to damage caused by agrochemicals based on shadow prices.

Table 11 gives a summary of agrochemical application rates of major food crops (data from LEI²⁹) and a range of projected application rates for Willow.

Table 12 shows the consequences of applying these figures for the COE of the defined bio-energy system. The first case reflects the situation of the selected reference system in which willow production takes place on fallow land which results in damage; the costs of the damage are however highly uncertain. As an illustration the outcomes are shown if SRC Willow were to replace conventional food crops. The very wide range obtained (also with a possible zero effect) is caused by the wide range in avoided chemical use that can be achieved depending on the type of crop that is being replaced by Willow. If relatively large quantities of chemicals are applied to Willow and this crop replaces certain vegetables, there is no net benefit. If potatoes are replaced by Willow the benefit is very large.

Davidson et al.¹⁰ recognize that crop yield losses will increase stronger when agrochemical application rates are reduced further. Also, only limited yield losses will occur when reduction rates remain modest. Although this relation is hard to quantify for each crop under various conditions it is an important notion when the average application rates of chemicals for SRC-Willow and various food crops are considered (see table 11). When the application rates for the mentioned food crops are reduced by the mentioned 50-70% they are still all above the application rate for SRC-Willow as given by Ranney et al.⁴⁴ (see table 11). It should be noted that the chemical application rate for Willow given in Zeijts et al.⁶³ is an assumed level. The value of Ranney et al. is based on field tests with even higher yield levels than are assumed in Zeijts et al. We therefore consider the low value more realistic. When this all is kept in mind it is questionable whether the high shadow price for agrochemical use should be used for calculating the damage costs of SRC-Willow production; or even stronger, if any damage costs should be included at all, since SRC-Willow apparently meets the required environmental standards. Therefore, we suggest to use the lower value of the shadow price range to obtain an estimate for the external costs of agrochemical use of SRC-Willow production.

However, this discussion illustrates that our understanding of the effects and seriousness of agrochemical use for SRC-Willow production needs improvement and that the given cost estimate should be seen as a first indication only.

As a final remark; in this analysis no distinction is made between various categories of agrochemicals (A, B, and C, C being the most harmful to health and environment). In the case of willow production the types of chemicals that are expected to be used are generally A-type.⁶³ So the overall avoided damage estimated by this method is fairly conservative.

Energy farming will also have an effect on soil quality, erosion, water use, nature and biodiversity. With respect to soil quality no attempt was made to valuate any external effect of biomass production, since it is an aspect which at present is hard to quantify. It relates to organic matter content, texture, presence of soil life and absence of diseases. Erosion is hardly a matter of importance in the Netherlands. Since the cover of multi-year rotation willow production is better than of most food crops, any effect that might occur will be very small in the Dutch context. Formation of root systems and less intensive management will most probably lead to improved soil quality when SRC is compared to production of food crops.³³

Effects of SRC-Willow production on groundwater levels are unclear. Willow might give rise to somewhat higher water consumption than other crops³⁵, which might contribute to the problem of dehydration. On the other hand Willow might allow higher groundwater tables than most food crops, which may counteract the effects of dehydration. Also, higher groundwater tables can have benefits in buffer zones of nature areas which could be expressed in avoided yield losses if food crops were cultivated in those areas. In the case of fallow land similar uncertainties occur, also influenced by site specific factors. Due to these uncertainties these effects are not evaluated further in this study.

The effects on biodiversity and nature by the introduction of SRC Willow are to a large extent unknown. Some benefits are expected since the crop provides a habitat for birds, insects, etc.^19,48 The net effects will depend strongly on the previous land-use and the planning of energy crops in the landscape.¹

Fundamental work with respect to the valuation of biodiversity and methodology development is essential to gain insight into this issue.²² Attempts to value biodiversity and nature have been made by several researchers including Groot²⁰ and Markandaya et al.³⁴ but are troubled by large uncertainties. The outcomes are very site specific and of less relevance to energy farming. Furthermore, lack of practical experience with large areas of energy crops is now a bottleneck for making an assessment of the impact on nature and biodiversity in various contexts.

Specifically in the Dutch context SRC might have a value by providing a corridor for certain species. This could save investment in landscape elements, such as described in Sijtsma et al.^52,53 These elements should allow species to migrate from one nature area to another, which is the purpose of policy making with regard to the creation of a National Ecological Network in the Netherlands.³² Again, any benefit will be very site specific and valuation is not possible without further knowledge about whether SRC can actually serve as corridor for the desired species. Actual effects are however uncertain and will depend strongly on how energy farming is implemented in practice. Therefore, no attempt is made in this study to valuate such effects economically.

The results obtained and expressed in either costs or benefits per kWh are summarized in table 13. The figures 1 and 2 illustrate the comparison between coal and biomass including external costs and benefits, added together with the production costs of electricity. To obtain a minimum figure, the maximum benefits and minimum damage have been added up, whereas for the total maximum damage the minimum benefits per externality combined with maximum damage figures were taken together. This sum should however be seen as an indication of the external effects of the two fuel cycles and not as the external costs.

Uncertainties persist and assumptions had to be made in this study that can be disputed. However, the outcomes illustrate possible variations in the amount of external costs and benefits. The outcomes illustrate that the beneficial effects with respect to indirect GDP increments, and to a lesser extent employment, are considerable when biomass is compared with coal. This due mainly to the fact that coal is imported into the Netherlands and biomass is considered to be produced indigenously. However, the indirect effects are calculated for an ideal context in which production capacity and labour are under-utilized. If economic activity is already high, inflatory effects may occur, such as increasing costs of labour and goods. Such effects are not analyzed here and including them would require a completely different approach. However, since this study indicates that the indirect economic effects are important compared to other external costs and benefits, more detailed analysis is desired on this point. For example, the method applied to calculate multiplier effects does not account for a decrease in expenses throughout the economy, because expenditures are limited by the realization and operation of power generation capacity. Expenditures almost always lead to positive GDP effects with an I/O approach. To counteract this effect much more advanced economic models are required.

The potential damage caused by CO₂ emissions is another aspect of this evaluation that causes a big difference between the external costs of coal and biomass. However, the costs of damage for CO₂ are very uncertain and are strongly influenced by the interest rate and included effects, which results in a very wide cost range. Here, a cost range of 5-125 U$/tonne CO₂) from the IPCC was used; the IPCC consulted and integrated the results of numerous sources with respect to potential damage of climage change and included various uncertainties in their cost range. However, it should be kept in mind that some authors suggest that higher damage costs of climate change are possible (see for example Sorensen⁶⁸).