Case Studies

Authored by: Paul H. Brunner , Helmut Rechberger

Handbook of Material Flow Analysis

Print publication date:  December  2016
Online publication date:  December  2016

Print ISBN: 9781498721349
eBook ISBN: 9781315313450

10.1201/9781315313450-4

Abstract

Looking at the graphic result of a material flow analysis (MFA), it seems easy and straightforward to define the system, collect the data, calculate the results, and draw conclusions. In practice, one does not start with the result but quite often with a badly defined problem that is highly complex and that has to be simplified and well structured first. After the goals of an MFA have been clearly defined, the real art consists of skillfully designing a system of boundaries, processes, flows, and stocks that allows solving a given problem at the least cost. Like in any other art, a precondition for mastering the art is to exercise the basic tools as much as possible. The more experienced a user gets, the easier it becomes to set up an appropriate system in a cost-effective way. An expert skilled in MFA will be able to define a metabolic system in any new field quite efficiently, with only a few alterations of the initial draft. Beginners will often find out that they have to revise their systems several times in order to cope with facts such as incomplete information about the important processes, stocks, and flows within the system; inappropriate systems boundaries; missing, bad, or incompatible data; etc.

Case Studies

Looking at the graphic result of a material flow analysis (MFA), it seems easy and straightforward to define the system, collect the data, calculate the results, and draw conclusions. In practice, one does not start with the result but quite often with a badly defined problem that is highly complex and that has to be simplified and well structured first. After the goals of an MFA have been clearly defined, the real art consists of skillfully designing a system of boundaries, processes, flows, and stocks that allows solving a given problem at the least cost. Like in any other art, a precondition for mastering the art is to exercise the basic tools as much as possible. The more experienced a user gets, the easier it becomes to set up an appropriate system in a cost-effective way. An expert skilled in MFA will be able to define a metabolic system in any new field quite efficiently, with only a few alterations of the initial draft. Beginners will often find out that they have to revise their systems several times in order to cope with facts such as incomplete information about the important processes, stocks, and flows within the system; inappropriate systems boundaries; missing, bad, or incompatible data; etc.

MFA is usually a multidisciplinary task. Materials flow through many branches of an economy, and they cross boundaries such as the interfaces anthroposphere–environment or water–air–soil. Hence, it is of prime importance to look for guidance from experts who understand those disciplines that are important for a particular MFA: if regional eutrophication due to poor nutrient management is investigated by MFA, it is necessary to include the knowledge of partners from agriculture, nutrition, sewage treatment, water quality, and hydrology, either by forming a project team or by engaging the experts as consultants when needed. Sometimes, this cooperation leads to new research questions, because the disciplinary research may, so far, not have been directed toward linking their disciplinary knowledge with other fields (cf. Section 3.1.2).

An MFA can be a time-consuming and costly task. This is especially true if an MFA is performed for the first time in a new field, such as a study of regional heavy metal flows (cf. Sections 3.1.1 and 3.4.1). It may well be that the basic data of the region, such as anthropogenic flows and stocks, hydrological data on precipitation, evaporation, and surface and groundwater flows and stocks, have not been assessed before. It should be realized that a minimum amount of information is needed; otherwise, an MFA cannot succeed. Thus, sufficient resources in manpower and funding are required.

It is a distinctly different task to perform an MFA in a particular field for the first time, or to repeat the analysis either for additional materials (e.g., first heavy metals, then nutrients) or for further time periods. The latter two tasks require less effort because the system has been set up and basic data, particularly on the level of goods, have already been collected before. If the costs for an initial MFA seem to be high, it should always be taken into account that the fundamental data can be used for future MFA and similar consecutive studies, such as annual environmental reporting or materials accounting.

The following 18 case studies demonstrate how MFA can be applied for

• Early recognition of beneficial and/or harmful accumulation and depletion of substances in stocks
• Optimization of single processes and of entire metabolic systems
• Policy analysis and policy decisions regarding the three fields of environmental management, resource management, and waste management

In addition, an example of regional materials management (lead) is given in order to show that MFA is especially well suited to address problems related to multiple fields, such as the three described before: the regional lead study by MFA was initially not addressed to any specific problem; it revealed conclusions important for all three fields. The case studies are intended to increase the reader’s experience. It is recommended also to read some of the original literature cited in these case studies. Nevertheless, for those who want to master the fine art of MFA (König, 2002), it will be indispensable to gain additional experience by performing on their own as many MFA studies as possible. Remember that looking at a final graph of an MFA reveals by no means the difficulties even experts encounter when condensing the complex reality of the world into an easily understandable, comprehensive MFA system.

3.1  Environmental Management

Most material flow analysis studies have been undertaken to solve problems related to environmental management. A recent overview of the potential of MFA in this field is given in MAcTEmPo (Brunner et al., 1998). In general, MFA is a tool well suited for

• Linking of emissions to sources and vice versa
• Setting of priorities for management measures
• Designing new processes, goods, and systems in view of environmental constraints

As seen in Chapter 2, an MFA is usually the starting point of any life-cycle assessment (LCA) and environmental impact statement (EIS). It is also useful as a base for an environmental management and audit system (EMAS) at the company level (see Section 3.1.4). If a company’s financial accounting system is linked to a material input–output flow and stock analysis, it can be efficiently used to measure the company’s environmental performance. The following case studies demonstrate that MFA can be used to investigate

• Single-substance issues (e.g., emissions of heavy metals or nutrients)
• Multisubstance problems (e.g., EIS of a coal-fired power plant)

They also show the wide scale of spatial application: a single power plant, a small region of 66 km2, and a large watershed such as the entire River Danube basin with 820,000 km2 can all be investigated using the same MFA approach.

3.1.1  Case Study 1: Regional Lead Pollution

Heavy metals are important substances for both economic as well as environmental reasons. Because of their physical–chemical properties, they can withstand weathering (zinc coatings of steel), improve the properties of other materials (chromium in steel, cadmium as an additive in polyvinyl chloride [PVC]), or serve to improve the efficiency of energy systems (lead in gasoline, mercury in batteries). Some heavy metals are not essential for the biosphere, but many are toxic for humans, animals, plants, and microorganisms. It is thus important to control the flows and stocks of heavy metals to avoid harmful flows and accumulations and to make the best use of heavy metals as resources.

This case study is taken from RESUB, a comprehensive study on the flows and stocks of 12 elements in a Swiss region (Bunz Valley) of 66 km2 and 28,000 inhabitants (Brunner et al., 1990). The purpose was to develop a methodology to assess material flows and stocks within, into, and out of a region in a thorough and integrated way. In addition, the significance of the findings for the management of resources and the environment was to be investigated. There was no given goal in view of environmental management. The case study portrayed in this chapter represents merely a small fraction of the entire RESUB project. Only the flows and stocks of lead relevant to environmental management are discussed. The implications of these flows and stocks for resource management are examined in Section 3.4.1. The detailed procedure described next confirms that an MFA is a multidisciplinary task that requires knowledge, information, and support from many fields.

3.1.1.1  Procedures

In a first step, the region is defined according to Figure 3.1. For the spatial boundary, the administrative boundary of the region “Bunz Valley” is chosen because, by chance, this border coincides well with the hydrological boundary. (This is often the case in mountainous or hilly regions, where the watershed serves well to delineate an administrative boundary.) Because water flow is fundamental for many material flows, it is important that a reliable regional water balance be established. If the spatial boundary does not coincide with the hydrological boundary, it may be difficult to establish a water balance. Hence, it is often crucial to find a good compromise between regional boundaries that match the administrative region, thus allowing the use of data collected by the regional administration, and hydrological boundaries that yield a consistent water balance.

Figure 3.1   Results of the MFA of lead flows (t/year) and stocks (t) through the Bunz Valley. (From Brunner, P. H. et al., Industrial metabolism at the regional and local level: A case study on a Swiss region, in Industrial Metabolism—Restructuring for Sustainable Development, Ayres, R. U. and Simonis, U. E., Eds., United Nations University Press, Tokyo, 1994. With permission. Brunner, P. H. et al., RESUB—Der regionale Stoffhaushalt im Unteren Bünztal, Die Entwicklung einer Methodik zur Erfassung des regionalen Stoffhaushaltes, 1990.)

As a boundary in time, a period of 1 year is selected because existing data about the anthroposphere (e.g., tax revenues, population data, and fuel consumption) and the environment (e.g., data on precipitation, surface water flow, and concentrations in soil and groundwater) show that a sampling period of 1 year is representative for the region during the period of 1985 to 1990.

In this chapter, each process is labeled with a letter and each flow with a number. These letters and numbers help to identify the corresponding processes and flows in Figure 3.1, Tables 3.1 through 3.10, and the calculations at the end of this chapter. The system is defined by the following 10 processes and 20 flows of goods.

3.1.1.1.1  Private Households

The process private households (PHH) summarizes the flows and stocks of materials through 9300 private households of the region. Import goods (1) relevant for lead comprise leaded gasoline (in the process of being phased out) and consumer goods such as lead in stabilizers, caps topping wine bottles, etc. Output goods are exhaust gas from cars (2), sewage (3), and municipal solid waste (MSW) (4).

The lead flows through private households are calculated as follows. Lead input in consumer goods is calculated based on the flows of sewage and MSW. This is a major shortcoming, since neither stocks nor flows of construction materials and appliances in private households are taken into account. To measure flows of lead in such goods is an extremely laborious and costly task. Thus, for this study, it is assumed that all lead that enters households leaves them within 1 year. This hypothesis is incorrect, since it does not account for the lead stock in households. Nevertheless, it is estimated that this error is of little relevance for the conclusions and the overall lead balance of the region.

Figures for lead in sewage are calculated as follows. The number of inhabitants (capita) connected to the sewer system (percentage of the regional population) is multiplied by per capita lead-emission factors (g/capita/year) determined elsewhere in similar regions.

Lead in MSW is similarly calculated. The number of inhabitants (capita) times MSW generation rate (kg/capita/year) times lead concentration in MSW (g/kg) yields the lead flow in MSW. MSW generation rate is available from regional waste-management companies. Lead concentration in MSW is taken from measurements of the residues of waste incineration (see Section 3.3.1).

It is assumed that all lead emitted by car exhausts stems from leaded gasoline. Fuel consumed for room heating contains less than 0.05 t of lead (calculated as the amount of fuel consumed times lead concentration in fuel) and is not taken into further consideration. Lead in car exhaust is calculated as follows. The number of cars licensed in the region is multiplied by the average mileage (in km/year) of a car (taken from national statistics), the average consumption of gasoline per kilometer (l/km, from car manufacturers’ statistics), and the mean lead content of the gasoline (mg/L, from gasoline producers and federal statistics). The results are cross-checked by figures from regional traffic monitoring and a model that takes into account the road network (fractions of highways, urban roads, roads outside of settlement areas) and speed-dependent emission factors (figures for lead concentration in gasoline are kept constant). Lead emissions of trucks are considered to be small and are neglected because, in this region, trucks are operated on diesel only, and diesel does not contain significant amounts of lead.

Table 3.1   Calculation of Lead Flows through Process Private Household

Flow No.

Operator

Description

Units

Value

Stock

Initial value

kg

Not

considered

Rate of change

kg Pb/year

Not

considered

Inputs 1

Consumer goods and leaded gasoline (balanced):

Exhaust gas (2)

kg Pb/year

1596

+

Sewage (3)

kg Pb/year

151

+

MSW (4)

kg Pb/year

5600

=

kg Pb/year

7347

Outputs 2

Exhaust gas:

Number of cars

cars

14,000

×

Mileage

km/car/year

15,000

×

Consumption of gasoline

L/km

0.08

×

mg Pb/L

95

=

mg Pb/year

1.6 × 109

=

kg Pb/year

1596

3

Household sewage:

Number of inhabitants

capita

28,000

×

Connected to sewer system

1

×

g Pb/capita/year

5.4

=

g Pb/year

1.51 × 105

=

kg Pb/year

151

4

Municipal solid waste (MSW):

Number of inhabitants

capita

28,000

×

MSW generation rate

kg/capita/year

400

×

g Pb/kg MSW

0.5

=

g Pb/year

5.6 × 106

kg Pb/year

5600

Note: Process A in Figure 3.1.

Table 3.2   Calculation of Lead Flows through Process Wastewater Treatment Plant

Flow No.

Operator

Description

Units

Value 1

Value 2

Total Value

Stock

Initial value

kg

Not considered

Rate of change

kg Pb/year

–16

Inputs

5

WWTP input:

TP 1

TP 2 + 3

TP 1 + 2 + 3

Wastewater

L/year

6.1 × 109

2.26 × 109

8.36 × 109

flow

×

μg Pb/L

121.7

59.2

concentration

=

μg Pb/year

7.42 × 1011

1.34 × 1011

=

kg Pb/year

742

134

876

Outputs

6

WWTP output:

TP 1

TP 2 + 3

TP 1 + 2 + 3

Purified water

L/year

6.1 × 109

2.26 × 109

8.36 × 109

flow

×

μg Pb/L

20.7

6.3

concentration

=

μg Pb/year

1.26 × 1011

1.42 × 1010

=

kg Pb/year

126

14

140

7a

Sewage sludge

TP 1

TP 2 + 3

TP 1 + 2 + 3

(used):

Sludge flow

kg dry/year

8.06 × 105

2.14 × 105

1.02 × 106

×

mg Pb/kg

875

216

concentration

dry

×

Used inside of

%

92

36

the region

=

mg Pb/year

6.49 × 108

1.66 × 107

=

kg Pb/year

649

17

665a

7b

Sewage sludge

TP 1

TP 2 + 3

TP 1 + 2 + 3

(exported):

Sludge flow

kg dry/year

8.06 × 105

2.14 × 105

1.02 × 106

×

mg Pb/kg

875

216

concentration

dry

×

Exported out

%

8

64

of the region

=

mg Pb/year

5.64 × 107

2.96 × 107

=

kg Pb/year

56

30

86

Note: Process B in Figure 3.1.

Notes:

a  Total value does not equal the sum of values 1 and 2, due to rounded values.

Table 3.3   Calculation of Lead Flows through Process Planetary Boundary Layer

Flow No.

Operator

Description

Units

Value 1

Value 2

Value 3

Stock

Initial value

kg

Not

considered

Rate of change

kg Pb/year

0

Inputs

2

Exhaust gasa:

kg Pb/year

1596

Outputs

8, 9, 10

Deposition:

Forest

Agriculture

Urban

Deposition

3

1

5

ratio

Deposition rate

kg Pb/ha/

0.294

0.098

0.490

year

×

Surface

ha

2000

3700

900

+

kg Pb/year

200

kg Pb/year

588

563

441

Note: Process C in Figure 3.1.

Notes:

a  Data from PHH (see Table 3.1).

Figures about the total input into households are rounded because they do not include lead-containing goods that contribute to the stock and thus have little effect on accuracy. For the overall results and conclusions, this accuracy is sufficient. If it turns out that, for the conclusions, the difference between the calculated and the rounded value is decisive, a more thorough investigation into the lead flows through private households would become necessary.

Table 3.4   Calculation of Lead Flows through Process Forest Soil

Flow No.

Operator

Description

Units

Value

Stock

Initial value

kg

150,000

Rate of change

kg Pb/year

529

Inputs

8

Depositiona:

kg Pb/year

588

Outputs

11

Runoff:

Deposition

kg Pb/year

588

×

Runoff factor

0.1

=

kg Pb/year

59

Note: Process E in Figure 3.1.

Notes:

a  Data from PBL (see Table 3.3).

Table 3.5   Calculation of Lead Flows through Process Agricultural Soil

Flow No.

Operator

Description

Units

Value

Stock

Initial value

kg

240,000

Rate of change

kg Pb/year

982

Inputs

9

Deposition:

+

From PBL

kg Pb/year

563

+

From WWTP

kg Pb/year

665

=

kg Pb/year

1228

Outputs

12

Runoff:

Deposition

kg Pb/year

1228

×

Runoff factor

0.2

=

kg Pb/year

246

Note: Process E in Figure 3.1.

For detailed information about the calculation, see Table 3.1.

Table 3.6   Calculation of Lead Flows through Process Urban Areas

Flow No.

Operator

Description

Units

Value

Stock

Initial value

kg

150,000

Rate of change

kg Pb/year

529

Inputs

8

Depositiona:

kg Pb/year

588

Outputs

11

Runoff:

Deposition

kg Pb/year

588

×

Runoff factor

0.1

=

kg Pb/year

59

Note: Process E in Figure 3.1.

Notes:

a  Data from PBL (see Table 3.3).

Table 3.7   Calculation of Lead Flows through Process Sewer

Flow No.

Operator

Description

Units

Value

Stock

Initial value

kg

n.d.

Rate of change

kg Pb/year

0

Inputs

3

Household sewagea:

kg Pb/year

151

13

Urban area runoffb:

kg Pb/year

265

14

Industry sewage (balanced):

WWTP input (5)

kg Pb/year

876

PHH sewage (3)

kg Pb/year

151

UA runoff (13)

kg Pb/year

265

=

kg Pb/year

460

Outputs

5

WWTP inputc:

kg Pb/year

876

Note: n.d. = not determined. Process G in Figure 3.1.

Notes:

a  Data from PHH (see Table 3.1).

b  Data from UA (see Table 3.6).

c  Data from WWTP (see Table 3.2).

Table 3.8   Calculation of Lead Flows through Process Industry

Flow No.

Operator

Description

Units

Value

Stock

Initial value

kg

Not considered

Rate of change

kg Pb/year

0

Inputs

15a

Used cars:

Number of used cars

cars/year

120,000

×

kg Pb/car

2.5

=

kg Pb/year

300,000

15b

Scrap metal:

Scrap metal

kg/year

6.50 × 107

×

kg Pb/kg

0.0005

=

kg Pb/year

32,500

15

Industry input:

Used cars (15a)

kg Pb/year

300,000

+

Scrap metal (15b)

kg Pb/year

32,500

=

kg Pb/year

332,500

Outputs

16a

Construction iron:

Construction iron

kg/year

1.45 × 108

×

kg Pb/kg

0.0005

=

kg Pb/year

72,500

16b

Filter residues:

Filter residues

kg/year

1.50 × 107

×

kg Pb/kg

0.0133

=

kg Pb/year

200,000

16

Industry output:

Construction iron (16a)

kg Pb/year

72,500

+

Filter residues (16b)

kg Pb/year

200,000

=

kg Pb/year

272,500

14

Industry sewagea:

kg Pb/year

460

17

Automotive shredder

residues (balanced):

Industry input (15)

kg Pb/year

332,500

Industry output (16)

kg Pb/year

272,500

Industry sewage (14)

kg Pb/year

460

=

kg Pb/year

59,540

Note: Process H in Figure 3.1.

Notes:

a  Data from Sewer (see Table 3.7).

Table 3.9   Calculation of Lead Flows through Process Landfill

Flow No.

Description

Units

Value

Stock

Initial value

kg

600,000

Rate of change

kg Pb/year

59,540

Inputs

17

ASRa:

kg Pb/year

59,540

Note: Process J in Figure 3.1.

Notes:

a  Data from Industry (see Table 3.8).

3.1.1.1.2  Wastewater Treatment Plant

In the process wastewater treatment plant (WWTP), wastewater (5) is treated, resulting in cleaned wastewater (6), sewage sludge (7a and b), off-gas, and small amounts of sieving residues and sandy sediments. Due to the chemical species of lead in wastewater, off-gas is of no quantitative relevance for this heavy metal. Preliminary sampling and chemical analysis of the concentrations in sievings and sediments show that the amount of lead in these fractions is small. Hence, most lead leaves the WWTP in sewage sludge and purified wastewater. The flows of wastewater, purified wastewater, and sludge are measured during 1 year (m3/year) and are sampled and analyzed for lead (g/m3). The flow of wastewater is determined by a venturi device at the inflow of the WWTP. Samples are taken continuously from wastewater and purified wastewater by a so-called Q/s sampler that samples proportional to the water flow. The flow of sewage sludge is measured as the total volume of sludge transferred to sludge transport vehicles during 1 year. Samples are taken whenever sludge is transferred to transport vehicles. There are three WWTPs in the region: one large plant and two small plants. Only the large plant and one of the small plants are included in the measuring campaign. For the third plant, the same material flows and balances are anticipated as for the other small treatment plant. The data of the three plants are summarized as a single WWTP process.

For detailed information about the calculation, see Table 3.2.

3.1.1.1.3  Planetary Boundary Layer

The process planetary boundary layer (PBL) denotes the lowest layer of the atmosphere. It is about 500 m high and is well suited as a “distribution” process for the RESUB case study. In regional studies, it is usually not possible to measure a material balance of the PBL, because it is a daunting analytical and modeling task. However, it is possible to make certain assumptions and simplifications that allow using the PBL as a suitable process to account for flows from the atmosphere to the soil and vice versa.

Table 3.10   Calculation of Lead Flows through Process River

Flow No.

Operator

Description

Units

Value 1

Value 2

Total Value

Stock

Initial value

kg

Not

considered

Rate of

kg Pb/year

–948

change

Inputs

18

Surface water

Holzbach

Bunz

Sum

(import):

Water flow

L/year

3.69 × 109

3.19 × 1010

3.56 × 1010

×

μg Pb/L

4.6

18.4

concentration

=

μg/year

1.70 × 1010

5.87 × 1011

flow

kg Pb/year

17

587

604

11

Forest

runoffa:

kg Pb/year

59

flow

12

Agricultural

runoffb:

kg Pb/year

246

flow

6

WWTP

outputc:

kg Pb/year

140

flow

Outputs

19

Surface water

(export):

Water flow

L/year

6.70 × 1010

×

μg Pb/L

29.8

concentration

=

μg/year

2.00 × 1012

flow

=

kg Pb/year

1997

Note: Process I in Figure 3.1.

Notes:

a  Data from Forest Soil (see Table 3.4).

b  Data from Agricultural Soil (see Table 3.5).

c  Data from WWTP (see Table 3.2).

Because the case study region is surrounded by regions that have basically the same metabolic characteristics, it is reasonable to suppose that the emissions into air are similar for all surrounding regions. Thus, it can be assumed that imported lead corresponds to exported lead, i.e., that the amount of lead imported and deposited in the region equals the amount of domestic lead exported and deposited outside of the region. Note that the total flow of lead through the PBL (not given in Figure 3.1) is about three to four times larger than the flow of lead deposited. The flows from the PBL to the soil consist of wet and dry depositions to forest (8), agricultural land (9), and urban (10) soils. The lead flows to the soil are determined by two methods (Beer, 1990). First, based on the assumption of uniform metabolism of all neighboring regions, the total regional emission to PBL is divided among the land areas, taking into account differences in vegetation and surface. Second, preliminary measurements at 11 sampling stations throughout the region show little significant differences for long measuring periods. Therefore, for a period of 1 year, wet and dry deposition of lead is measured only at two sampling stations in the region. Based on the wet and dry deposition results, and on models given by Beer (1990), lead flows are calculated for the corresponding soils. The deposition on urban soils takes into account that most lead is emitted in the proximity of roads; thus, the load per hectare of urban soils is comparatively larger than that of agricultural and forest soils. Results from the two approaches agree fairly well. The method based on actual measurements of dry and wet depositions yields values about 30% higher than the distribution of the total emissions over the region.

For detailed information about the calculation, see Table 3.3.

3.1.1.1.4  Lead Flows and Stocks in Soils

The region consists of 3700 ha soil used for agriculture, 2000 ha forest soil, and 900 ha settlement area. The area actually covered with buildings, roads, and other constructions is much smaller. The hydrological balance reveals the water flow to and from the forest soil and the agricultural soil. Precipitation (measured by continuous automatic rain measurement) minus evaporation, estimated with various models and finally calculated according to Primault (1962), yields the net water flow to the soil. This amount of water is divided among the fractions of agriculture and forest soils, taking into account differences in evapotranspiration of forests and agricultural crops. The water reaching the soil is divided into surface runoff and interflow (both reaching surface waters) and the fraction seeping to groundwater. The concentrations of lead in soil leachate are estimated based on another research project on heavy metal mobility in soils (Udluft, 1981). Erosion is approximated according to von Steiger and Baccini (1990). Both assessments are individually tailored for forest and agricultural soils. The calculations show that 10% of the deposited lead on forest soil (11), 20% of agricultural soil (12), and up to 60% of built-up areas (13) can be found in the runoff. This lead is transported to receiving waters.

For detailed information about the calculation, see Tables 3.4 through 3.6.

3.1.1.1.5  Sewer System

The mixed sewer system receives wastewater from private households (3), urban surface runoff (13), and industry (14), and transports the resulting sewage (5) to the WWTP. Lead in wastewater from industry can be estimated by balancing the process sewer, taking into account lead flows in wastewater from private households, in surface runoff from urban areas, and in the resulting sewage.

For detailed information about the calculation, see Table 3.7.

3.1.1.1.6  Industry

The process industry proves to be a real challenge. Despite the region’s small size and low population, there are 1300 companies with 11,000 employees active in the region. The main task is to find within this large number those companies that play an important role in the regional lead flow. As a first step, all sectors except the production sector are removed from the list. Of the remaining 323 companies, those with less than 20 employees are eliminated. The remaining 102 businesses are included in the investigation, which consists of an interview and a questionnaire about the material flows and stocks of each company. Of these, 61 companies cooperate actively and supply comprehensive data about their material turnover. It appears that only a few are handling lead-containing goods. A car shredder and an adjacent iron smelter using shredded cars to produce construction rods are dominating the process industry. Hence, the inputs into the process industry are used cars (15). The outputs comprise, on the one hand, construction rods (16a) and filter residues (16b) of the smelter that are exported. On the other hand, the car shredder produces organic shredder residues (17) consisting of plastics, textiles, and biomass (wood, paper, leather, and hair) mixed with residual metals of every kind. These so-called automotive shredder residues (ASRs) are landfilled within the region.

The lead flows through industry are assessed as follows. Input in industry is calculated as the number of used cars times the concentration of lead in a car. This yields a minimal figure. It may well be that other goods are shredded and treated in the shredder as well. (Note that this uncertainty is not important for the final conclusion. It would be the same even if double the amount of lead were used in industry.) The number of cars processed by the shredder is supplied by the shredder operator. Figures for lead concentration in used cars are found in the literature or can be received from car manufacturers. The smelter operator supplies figures about the amount of construction steel produced, the amount of filter residue exported, as well as the lead concentration in the steel and the filter residue. The latter figure is confirmed by local authorities, who periodically monitor emissions of the smelter. Note that there is no emission flow given for the smelter in Figure 3.1. This is due to the excellent air-pollution control (APC) device of the smelter, which keeps annual emissions at a level that is orders of magnitude below lead emissions from gasoline. Thus, the smelter is of no relevance for lead emissions in the region anymore. The amount of lead in industry sewage is similar to other regional lead flows.

For detailed information about the calculation, see Tables 3.8 and 3.9.

3.1.1.1.7  Surface Waters

The process river consists of a river flowing through the region and a small tributary originating predominantly within the region. Groundwater entering or leaving the region does not play an important role. The process river receives water from the river inflow (10), the agricultural soil (11), the forest soil (12), and the WWTP (8). The river outflow (13) leaves the region. The direct lead flow from settlement areas to surface waters has not been taken into account. First, the region has a mixed sewer system, and most urban runoff is collected and treated in the WWTP. Second, the settlement area is small (<10%) in comparison with the agricultural and forest soils; hence, neglecting this flow may be justified. The surface water flow has been determined in the course of a complete water balance that is measured for the comprehensive RESUB project. Existing measuring stations continuously record the flow of river water in and out of the region. Since these stations are not located at the exact systems boundaries, the differences are compensated for by taking into account the area contributing water to the river. The river water is sampled continuously at the same stations with Q/s samplers. It turns out that these state-of-the-art samplers are well suited to collect dissolved substances and suspended particles, but they cannot catch aliquots of large particles. In the course of a rainstorm, when the river transports larger particles, debris, and chunks of biomass, the samplers are not working appropriately. In addition to the limits in sampling technology, there are practical problems. On several occasions, large water flows during heavy storms have destroyed or carried off the sampling equipment. Hence, it is advisable to have short sampling periods (e.g., 1 week). In case of invalid sampling, the missing data cover a smaller fraction of the total measuring period and thus are of less weight. Given these shortcomings, values for lead flows in the river must be regarded as minimum flows.

For detailed information about the calculation, see Table 3.10.

3.1.1.2  Results

In this section, conclusions are drawn regarding the use of MFA for environmental management. In particular, the results are used to show how MFA serves to provide early recognition of environmental hazards, how it can be used to establish priorities for environmental measures, and how it can be used for efficient environmental monitoring. In Section 3.4.1, the same case study is further used to point out the potential for regional materials management and resource conservation. The numerical results of the MFA analysis of lead in the region are given in Figure 3.1.

3.1.1.2.1  Early Recognition of Environmental Hazards

The difference between lead import and lead export amounts to approximately 60 t/year. Hence, lead is accumulated in the region. The existing stock of lead totals about 1000 t. A “doubling time” for the lead stock of 17 years can be calculated. In other words, if the regional flows of lead remain the same for the next 100 years, the stock will have increased from 1000 to 7000 t! (Note that according to Chapter 1, Section 1.4.5, there is no indication yet that lead flows will decrease; on the contrary, based on past developments, they are likely to increase further.) Without the present study, this buildup of lead occurs unnoticed. As shown in Chapter 1, Section 1.4.5, such accumulations of substances are a rule for all urban regions. What makes this case study special is the huge extent of the accumulation. About one-sixth of the lead imported stays “forever” within the region. Thus, it is highly important to investigate the fate of potentially toxic lead in the region. Does the accumulation in the soil result in an increase of lead in plants up to a level of concern for animal or human food? Will there be a steady increase in lead flows from the soil to the surface water? When will lead concentrations reach a level that endangers the standards for surface water or drinking water? What about the concentration of lead in dust; will it increase, too?

While MFA is helpful in identifying the problem and formulating relevant questions, the questions cannot be answered by simple MFA alone. It is necessary to engage specific experts, e.g., in the field of metal transfer between soil and plants, between soil and surface water and groundwater, and between soil and air. The merit of MFA is the ability to identify a future environmental problem that has been neither on the agenda nor even known before the study has begun.

From an environmental point of view, the largest flow, stock, and accumulation in stock are caused by the lead imported in used cars, shredded, and landfilled as ASR. The landfill is by far the main regional “accumulator” for lead. Assuming similar practice over the past 10 years, it can be estimated that approximately 600 t of lead is buried in the landfill. This is due to the fact that the separation of lead by the car shredder is incomplete. Some elemental lead as well as lead compounds in plastic additives are transferred to the ASR. The hypothesis that ASR may contribute to the pollution of the regional hydrosphere is discussed next.

Lead is accumulated in the soil, too. The doubling periods are between 170 (urban soil) and 280 (agricultural soil) years. If the use of lead continues in the same way, standards for lead concentrations in soils will be exceeded in the future. It is a matter of soil-protection strategy (and, in a wider perspective, environmental protection) whether such a slow approach to a limit needs to be controlled or not, and if so, when. The “filling up” strategy raises the question of what options future generations will have when they inherit the “full” soil. If soil inputs and outputs can be kept in equilibrium, the concentrations in the soil will remain constant. The case-study region will come close to this condition if leaded gasoline is phased out. (Note that since the residence time of aerosol-borne lead in the atmosphere is several days, which facilitates the transport of lead over long distances, this measure will be effective only if neighboring regions adopt the same strategy.) Lead concentration increases fastest in urban soils. MFA results suggest that persons eating food grown in urban areas, such as home gardeners, consume the highest amount of heavy metals such as lead. Hence, material balances and analysis of urban soils and gardens are needed in order to protect consumers of homegrown products in cities.

3.1.1.2.2  Establish Priorities for Environmental Measures

The landfill is the most important stock of lead, so it must be investigated and controlled first. Based on the following calculation, it can be hypothesized that lead is leaching into the groundwater and surface waters. The balance of the process river reveals a deficit of 0.95 t/year. Compared with the lead flows from soils and WWTP, this is a large figure. The most likely stock that can lead to such a large flow is the landfill stock. At present, the fate of lead in the landfill is not known. Since much of the ASR has been landfilled without bottom and top liners, it may be that lead is leaching into groundwater and surface water. Even if the landfills are constructed according to the state of the art, with impermeable bottom and top liners, it still has to be expected that the liners will become permeable over the long run (>100 years), thus polluting ground and surface waters for long time periods (Baccini and Lichtensteiger, 1989).

The following issues are crucial for possible mobilization and emission of lead from the landfill: the interaction of the landfill body with the surroundings (atmosphere, precipitation), the transformations of ASR within the landfill due to biochemical and chemical reactions, and the chemical speciation of the lead in ASR. Investigations into these aspects are specialized tasks that cannot be performed by MFA. Such investigations require in-depth analysis by experts in the field of transformation and leaching processes in soils and landfills. It is not possible within the case study to follow up this hypothesis (e.g., by collecting leachates and analyzing it for lead). In any case, it is of first priority to ensure that the ASR is landfilled in a manner that ensures long-term immobilization of heavy metals. If disposing of raw ASR leads to leaching of heavy metals and significant water pollution, pretreatment of ASR before landfilling will be mandatory.

The second largest regional flow of lead is due to MSW, which is exported from the region and incinerated. This flow is an order of magnitude larger than the lead flow in sewage. Thus, compost from MSW is less suited for application to land than sewage sludge because it will overload the soil with lead in a comparatively short time. For this region, it is recommended that sewage sludge not be applied to soils as well, since it loads the soil with additional lead. It is clear that a decision to use compost or sludge in agriculture cannot be based on lead alone. The approach taken here is exemplary and also has to be applied to a full range of substances such as other heavy metals, nutrients, and organic substances.

As mentioned previously, MSW is incinerated outside the region. In order to protect the soil from the large flow of lead in MSW, thermal treatment of MSW has to be combined with efficient APC. If incineration transfers less than 0.01% of lead into air, the resulting increase of lead in soil will be below 1% in 8000 years (assuming uniform deposition within the region). State-of-the-art MSW incinerators exhibit such transfer coefficients (TCs) for lead to the atmosphere. Regarding the smelter, the MFA supports the conclusion that the emissions to air are of no priority ever since the furnace has been equipped with a high-efficiency fabric filter system, reducing the lead emissions to less than 50 kg/year.

3.1.1.2.3  Environmental Monitoring

Once an MFA of a region is established, many opportunities for monitoring substance flows and stocks arise.

MFA can replace soil monitoring programs. Such programs are costly and are limited in their forecasting capabilities. If statistically significant changes in soil concentrations are to be detected by traditional soil monitoring, then either (1) very intensive sampling programs with large numbers of samples or (2) sampling over long time periods are required. Because the funds for such intense sampling are not usually available, it takes decades until significant changes in the soil become visible. However, with a single measuring campaign, MFA can predict how the soil concentration will evolve over time. The results indicate whether there is a danger before high concentrations are reached. If inputs to the soil are changed, e.g., through the addition of sewage sludge or a ban on leaded gasoline, the effect of such measures can be evaluated by an MFA before they are implemented. In contrast, traditional soil monitoring would take years to confirm accumulation or depletion of soil pollutants in a statistically significant way.

Combining MFA with the analysis of sewage sludge allows monitoring of the process industry. For example, before the smelter was equipped with high-efficiency fabric filters, a wet scrubbing system removed metals from the off-gas stream. By accident, it happened that some of the lead-loaded scrubber effluent reached the sewer system and severely contaminated the activated sludge in the treatment plant. While wastewater has a short residence time in the treatment plant, sewage sludge in a digester or storage tank represents a “memory” of several weeks. Thus, sludge samples from the digester can show a prolonged increase in lead concentrations. In combination with concentration data for other metals, it may be possible to identify the source of pollution by assigning metal “fingerprints” (concentration ratios) of sewage sludge to those of scrubber liquid.

Likewise, the combination of MFA and monitoring of MSW incineration residue allows one to assess the flow of lead, as well as other substances, through private households (see Section 3.3.1).

Finally, MFA facilitates an additional type of monitoring. If there is no information available about a process, it may be possible to estimate the missing material flows by mass-balancing the adjacent processes. In the aforementioned case study, data about substance flows through the shredder are not available. Rough data about lead in new cars from the last decade, combined with data supplied by the smelter on lead in the two outputs (filter dust and construction rods), allows the flow of lead through the car shredder to be assessed without analyzing the shredder itself.

3.1.1.3  Basic Data for Calculation of Lead Flows and Stocks

In the calculations presented in Tables 3.1 through 3.10, each process is labeled with a letter and each flow with a number. These letters and numbers help to identify the corresponding processes and flows in Figure 3.1. The description of each process is structured as follows:

• Name of the process
• Lead stock inside the process
• Rate of change of the lead stock
• Name of input flows (including a list of quantities that are used to calculate the lead flows)
• Name of output flows (including a list of quantities that are used to calculate the lead flows)

3.1.2  Case Study 2: Regional Phosphorous Management

Nutrients such as nitrogen, phosphorous, potassium, and carbon are essential for the biosphere. They are the key factors controlling growth and enabling species and populations to develop or causing them to vanish. They are especially crucial for the production of food for humans and animals. Because of limitations inherent to the soil–plant system, not all nutrients delivered to the soil can be taken up by plants (Scheffer, 1989). Thus, agricultural losses of nutrients are common, and they cannot be avoided. Yet, they can be reduced by farming practices that are directed toward minimizing losses to the environment. Nutrients in surface waters enhance the growth of algae (eutrophication). As a consequence, the oxygen content in surface water is reduced due to the increased plankton mass, mass death, and decomposition of organisms. As the oxygen concentration decreases, fish and other organisms find it increasingly difficult to survive. Due to transformations in soil and groundwater, nitrogen can also be lost as NOx or NH3 to the atmosphere, contributing to the formation of tropospheric ozone and particulate matter, respectively. Hence, the control of nutrients is of prime importance for the management of resources as well as of the environment.

Case studies 2 and 3 both relate to nutrient pollution. The difference between the two is the scale: a small region of 66 km2 and 28,000 inhabitants in case study 2 versus the entire River Danube with a watershed of 820,000 km2 and 85 million inhabitants in case study 3. It is noteworthy that for both scales, the same MFA approach can be taken. Nevertheless, there are focal differences in the task of balancing nutrients on these two scales. The challenge on the large scale is to put together a team (often international) that uses the same approach along the entire stream, allowing true comparison and combination of the individual results. In addition to the present case study, another case study on P is presented in Section 3.5 about regional materials management. In this case study, the challenge of accounting for varying P flows and stocks over longer time periods is discussed, too.

3.1.2.1  Procedures

Like the lead example in Section 3.1.1, case study 1 is a part of the comprehensive RESUB project; it focuses on flows and stocks of phosphorous (P). The procedure is the same as for lead, with some small changes due to the way phosphorous is used. The systems boundaries in space and time are identical. Only agricultural soil is taken into account, since the flow of P on forest and urban soils is comparatively small. Two additional processes for animal breeding and plant production are introduced. Hence, again, 10 processes and 19 flows of goods are taken into account (Figure 3.2).

As a first step, the water balance is estimated. Water is important for the flow of phosphorus because P can be transported both in a dissolved state (leaching) and as a particle (runoff and erosion). Hence, a comprehensive water balance for the region is needed (Figure 3.3). To minimize the costs for an annual water balance, the relevant hydrological flows and processes must be identified by a systems analysis (Figure 3.4).

By a provisional semiquantitative water balance, the main water flows and stocks are identified in order to set priorities for the following costly assessment and measurement program. The main purpose is to achieve sufficient accuracy with the least number of expensive measurements. A potential problem for water balancing is the mismatch between the regional (administrative) and hydrological boundaries. In this study, the two definitions of the region coincide well. The small deviations are compensated for assuming the same net precipitation for areas within and outside the administrative region. Determining the flows and stocks of groundwater is a necessary but usually difficult and resource-consuming task. It is therefore beyond the possibility of most regional MFAs. If groundwater data are not available, and if there are major groundwater inflows, outflows, or changes in stock, a hydrological balance might not be possible. In such cases, MFA has to be limited to a specific regional problem not related to the hydrosphere, or it fails altogether. Data for evapotranspiration can be calculated using various formulas (according to Penman, 1948 or Primault, 1962) and regional data on climate and vegetation. The path of water from precipitation to groundwater and surface water can only roughly be assessed, too. In the present study, there is sufficient information about groundwater outflow from the region to neighboring regions.

Figure 3.2   Regional phosphorous flows and stocks. (From Brunner, P. H. et al., Industrial metabolism at the regional and local level: A case study on a Swiss region, in Industrial Metabolism—Restructuring for Sustainable Development, Ayres, R. U. and Simonis, U. E., Eds., United Nations University Press, New York, 1994. With permission. From Brunner, P. H. et al., RESUB—Der regionale Stoffhaushalt im Unteren Bünztal, Die Entwicklung einer Methodik zur Erfassung des regionalen Stoffhaushaltes., 1990.)

The following equation is used for the hydrological balance:

$Precipitation + surface water import + groundwater import + drinking water import = evaportranspiration + surface water import + ground water export + drinking water export + change in stock$

The flows and stocks of water in eight goods, listed in Table 3.11, are measured for a period of 1 year. Samples are taken for the same time period for most of these goods. Since drinking water is produced from groundwater, it is assumed that drinking water and groundwater have the same concentrations. Measurement and sampling methods, frequencies, and locations are given in Table 3.11 and Figure 3.4. For more information about establishing regional water balances, refer to Henseler, Scheidegger, and Brunner (1992) After analyzing the water balance, the next step is to measure the flows and stocks of phosphorus. For each good investigated, the flow is multiplied by the concentration of P within that good to determine the phosphorus fluxes. In the following, it will be explained how the data presented in Figure 3.2 are assessed.

Figure 3.3   Results of regional water balance: while the river passes the region, the flow of surface water is doubled by the net precipitation input (precipitation minus evapotranspiration). Bunz and Holzbach are two rivers in the valley. (From Henseler, G. et al., Vom Wasser, 78, 91, 1992. With permission.)

3.1.2.1.1  Private Household

The flow of food-derived P into private households is established using data about household food consumption (BAS, 1987) and the nutrient content of food (Lentner, 1981). Phosphorus in household detergents and cleaners is not taken into account, since federal legislation banned P for these purposes. A rough estimation of other P flows showed that they are so small that they do not have to be taken into account (<1°% of total regional flow, <10% of flow through private households). The P output of households is not measured but instead is calculated according to Figure 3.5 and the conservation of mass. Of the P entering private households, 90% is assumed to leave by means of wastewater and the remaining 10% by MSW. MSW is not considered further, since it is treated in an MSW incinerator outside the region. It is worth mentioning that composting of MSW can hardly be justified on the grounds of nutrient conservation, since its maximum contribution to regional P management is marginal and about 1% of the total nutrient use in agriculture.

Figure 3.4   Determination of regional water balance. (• = flow measurements and sampling points). (From Henseler, G. et al., Vom Wasser, 78, 91, 1992. With permission.)

3.1.2.1.2  River

The good surface water is flowing in and out of the region at rates of 35 × 106 and 67 × 106 m3/year, respectively (Figure 3.3). The P concentrations in the input and output of the region are measured (0.8 mg/L and 1.1 mg/L, respectively) and multiplied by the corresponding water flow, resulting in 28 and 74 t P/year, respectively (Figure 3.2). Some phosphorus flows such as those in precipitation, drinking water import and export, evapotranspiration, and groundwater are less than 1% of the total regional phosphorus flow. Therefore, they are not taken into account for the phosphorus balance.

Table 3.11   Measuring and Sampling Procedure to Establish a Regional Water Balance

Good

Number of Measuring Stations

Method of Flow Measurement

Measuring Period

Method of Substance Sampling

Sampling Period

Precipitation

3

Rain gauge

1 year (365 × 24 h)

Composite sample

27 × 2 weeks

Surface water

3

River gauge

1 year (365 × 24 h)

Composite sample

27 × 2 weeks

Wastewater

2

Venturi

1 year (365 × 24 h)

Composite sample

27 × 2 weeks

Sewage sludge

2

Container

1 year

Composite sample

10 per year

Sievings from WWTa

1

Balance

1 year

Grab sample

3 per year

Sand from WWTb

1

Volume

1 year

Grab sample

3 per year

Drinking water

9

Meter

1 year

Grab sample

9 per year

Groundwater

5

Water table

1 year

No samplingc

Source: Henseler, G. et al., Vom Wasser, 78, 91–116,1992.

Notes:

a  Screenings from wastewater pretreatment.

b  Sediment of wastewater pretreatment.

c  Groundwater identical to drinking water.

Figure 3.5   Flow of food, food dry matter, and phosphorous contained in food through private households. STP: sewage treatment plant. (With kind permission from Springer Science+Business Media: Metabolism of the Anthroposphere, 1991, Baccini, P. and Brunner, P. H.)

3.1.2.1.3  Sewer System

The flows of P through the sewer system are calculated using data from household outputs and measured WWTP inputs. The figure for P in “industry” wastewater is calculated as the difference between WWTP input and household wastewater (38 – 17 = 21 t P/year). In the WWTP, the output to the surface waters is measured by multiplying the volume of treated waste-water by the concentration of P measured in 52 biweekly samples of treated wastewater (19 t P/year). P contained in sludge and applied in agriculture (13 t P/year) is measured in metering the total flow of sludge when transferred to transport vehicles, and samples of P are taken for analysis during this transfer.

3.1.2.1.4  Industry

Two aspects are important for the process industry: food is stored temporarily in a large stock of interregional importance, and polyphosphates are used in some amounts in regional chemical and other companies. The P contained in food leaves the region unchanged as an export good. P in polyphosphates is transferred to the sewer and is a large source for P in the WWTP.

3.1.2.1.5  Landfill

The process landfill is not investigated, although it is possible that phosphorous-containing wastes (biomass, detergents) have been landfilled in the past and that P is leaching to groundwater and surface water.

3.1.2.1.6  Agriculture

Three types of agricultural production systems—animal breeding, crop raising, and miscellaneous—are defined based on their different managerial characteristics. These agricultural practices are investigated under three processes: animal production (production of animals and dairy produce), plant production (production of wheat, corn, vegetables, etc.), and agricultural soil, as shown in Figure 3.2. For each production system, the use of mineral fertilizer, manure, animal products, and harvested goods are measured per unit of agricultural area and monitored for 2 years. Phosphorus contents of all goods are determined analytically to estimate the annual entry of phosphorus to the soil. All data are double-checked against the values taken from agricultural information sources. Input through manure and fertilizer and output through harvest are then extrapolated, taking the values of the three production systems described previously and considering actual farming practice in the region (e.g., number of animals, amount of produce, area for crop production, etc.). Flows of goods in the process animal production, such as animals, fodder, and dairy products, are checked through field accounts. Figures for sewage sludge are collected from WWTPs, and those for deposition, erosion, and runoff are taken from the literature. Figure 3.2 displays the amount of P in the output goods, namely, harvested plants like cereals, vegetables, and fruits (export of 24 t P/year), and animal feedstock cycled within the region (85 t P/year).

The flow of P to plant production including agricultural soil (X) is calculated as follows:

$X = fertilizer + manure + atmospheric deposition + sewage saludge − ( animal feed produced + cereals, vegatetables,fruits ) = 194 − 109 − 85 t P/year$

The amount of P stored in the agricultural soil (S) is calculated as follows:

$S = X − ( erosion + runoff and leaching to surface and groundwater ) = 85 − 17 = 68 t P/year$

The groundwater inflow into the region is close to zero, and the ground-water outflow from the region is small. Therefore, it has been assumed that all P running off and leaching from soils is eventually reaching the regional river.

3.1.2.2  Results

The results of this case study include the regional water balance as well as the P flows and stocks of the region. The water balance is summarized in Figure 3.3. There are two large imports (precipitation and river inflow) and exports (river outflow and evapotranspiration) of water. The main water flow through the region consists of humidity contained in air; this flow has not been considered here because it is not relevant for the P case study. On its way through the region, the surface water flow (river water) is doubled by the input of net precipitation (precipitation minus evapotranspiration). Of the surface water produced within the region, 28% is treated wastewaters from one major and two small sewage treatment plants. Thus, the ratio of wastewater to surface water flow is relatively high, and accordingly, the regional potential for dilution of wastewaters is rather small. Hence, efficient wastewater treatment is highly important for the quality of the river water.

An increase in groundwater stock of ≈10% of net precipitation is observed during the assessment campaign. It reveals that the year of measurement is a rather “wet” year; it distinctly deviates from the 10-year average of the hydrological balance, which shows a tendency toward decreasing ground-water stock.

The results of the analysis of flows and stocks of phosphorus are presented in Figure 3.2. As in the case of lead, the P imports (232 t/year) outweigh the P exports (168 t/year) by far, resulting in an accumulation of 64 t of P/year. The main sink for P is the soil. It contains already 10,000 t, and 68 t/year is added. (Note: the difference between 68 and 64 t/year given for regional accumulation stems from the uncertainty of the processes WWTP and sewer that are not in balance.) In MFA, it is often the case that inputs, outputs, and changes in stocks of processes do not match, and hence, uncertainties remain (see Chapter 2, Section 2.3). The largest amount of P is imported for agricultural activities (45 t/year of fodder for animals and 78 t/year of fertilizer for plant production). By manure (100 t/year), fertilizer (78 t/year), sewage sludge (13 t/year), and atmospheric deposition (3 t/year), 194 t/year of P is applied to the regional soil. Plants take up 109 t/year, and 17 t/year passes to the surface waters by leaching and erosion. A regional silo for food holds large amounts of P that are continuously replenished, accounting for a P flow of 40 t/year. The use of P for industrial water treatment amounts to another 21 t/year. Possibly, more P is used but not accounted for in industry.

3.1.2.2.1  Environmental Protection

Besides the unknown amount of P in landfills, there are two main issues concerning P management in the region. First, P is accumulating in the soil (+68 t/c, corresponding to an increase of the P stock in the soil of +0.68% per annum) and eroding/leaching to the surface waters (17 t/year). Second, P is directly discharged with purified sewage into the receiving waters. The load of P in the river increases from 28 t/year at the inflow to 74 t/year at the outflow. The flow of river water is doubled by the addition of regional net precipitation. Hence, the P concentration in the river increases by ≈30%. If all upstream and downstream regions contribute to the P load in surface waters in the same way, eutrophication is likely to take place in downstream lakes and reservoirs. Thus, the maximum allowable P load to the river must be assessed by also taking into account the potential and limitations for P dilution of the surface waters outside of the region.

3.1.2.2.2  Early Recognition and Monitoring

The MFA of P facilitates early recognition of P accumulation in the soil before it actually happens. This is important for water-pollution control. If the load of P into the river needs to be limited, there are two theoretical options (the uncontrolled landfills are not discussed here as a potential source because they have not been investigated):

1. The removal efficiency for P in the sewage treatment plant WWTP can be increased from about 30–50% to >90% in a relatively short time (months).
2. The flow of P from the soil to the surface waters can be reduced.

The second option does not allow quick reduction of P flows: for a given agricultural practice, the amount of P eroded is mainly a function of the P stock in the soil. Hence, it is necessary to either change agricultural practices or to reduce the P stock in the soil, which takes a long period of time (decades to centuries). MFA makes it possible to forecast accumulation (as well as depletion) of P in the soil long before it actually happens. Taking into account the current amount of P in the soil (≈10,000 t) and the annual accumulation of 68 t/year, it can be assessed that the P concentration in soils will be doubled in about one and a half centuries, if present agricultural practice is maintained. This will lead to a large increase in eroded P, offsetting the reduction of P in the river due to improved elimination of P in sewage treatment.

Direct soil monitoring yields results with large standard deviations. Thus, even an intensive soil sampling and analysis campaign will not identify P accumulation within a few years, because mean values will not be significantly different within one decade. MFA provides timely predictions of the change in soil stocks with one single measuring campaign of soil concentrations and P inputs into the soil. Of course, if agricultural practice is changed, the data and calculations have to be adapted to the new situation, too.

3.1.2.2.3  Priorities

Comparing the various flows of P to the soil, it becomes clear that in this region, sewage sludge is, comparatively, a small source of P, supplying less than 10% of the total soil input. Thus, in terms of resource conservation, the application of sewage sludge on farmland is of little importance and is of low priority. The dominant flows are due to the cycling of a large stock of P within agricultural production by the soil–plant–animal–soil system. The ratio “input to product output” of the different processes is noteworthy: The process animal (production of animals and diary produce) consumes 130 t P/year to produce 30 t P/year; plant production including soil requires 194 t P/year to produce 109 t P/year in plants harvested. This translates to efficiencies of 23% for the use of P in animal production and 56% for plant production including soil. Clearly, if P becomes a limited resource, priorities are either to increase the efficiency of P in animal farming or to shift the dietary intake toward less meat and more vegetarian foods.

It is also noteworthy that composting of household garbage is insignificant and of low priority regarding the P flow within the region. Assuming that less than 20% of food bought by private households is discarded as MSW (80% being eaten, eventually transformed to urine and feces, and collected with sewage), composting of separately collected garbage would supply only about 3 t P/year for agricultural production, equaling ≈2% of total agricultural input.

3.1.3  Case Study 3: Nutrient Pollution in Large Watersheds

Case study 3 is part of an in-depth investigation into water-quality management of the entire Danube Basin comprising 11 different riparian countries. It is described comprehensively in the report “Nutrient Balances for Danube Countries” (Somlyódy et al., 1997).

The case study is included here to demonstrate the following: (1) The application of MFA is independent of scale; hence, the same methodology can be applied to small (farm) and very large (international watershed) systems. Nevertheless, there are clear differences in the focus and procedures according to the scale. In general, multinational MFAs on large scales like an investigation into a transnational watershed require the joint effort of several research groups from each of the participating countries.

(2) The results of a large-scale study directed toward decision making in environmental protection can have different consequences for the partners engaged. While one country may not be a large factor for the pollution of the watershed, another may turn out to be a major contributor. Hence, if a common level of water protection is established and corresponding remediation measures are taken, the financial consequences may be quite severe for the latter and only marginal for the former. It is thus of utmost importance to use the same, appropriate and uniform methodology that is accepted by all partners. It is necessary to acquire an adequate, comprehensive and compatible data set for each country using equal definitions. Otherwise, if terms are not equal, data and results of the different riparian states cannot be compared. This holds true for flows and stocks of nutrients from agriculture, industry and trade, private households, water and wastewater management, and waste management alike. And it is essential to use a uniform methodology in collecting, calculating, and evaluating the data, too. MFA represents such a methodology well suited for multinational teams collaborating in delicate situations. For transnational application of MFA, capacity building and know-how transfer is important to ensure that all groups and participants are applying the same methodology. If MFA is standardized in the future, multinational analysis of metabolic systems will be much facilitated.

High nutrient loads are recognized as one of the most severe problems of the River Danube, the Danube Delta, and the “final sink” Black Sea. The ecosystem of the Danube Delta is severely endangered, and a large part of the Black Sea is critically eutrophic (Mee, 1992) The main objective of this study is to prepare a basis for decisions regarding the protection of the water quality of the Danube, its delta, and the Black Sea. In particular, the goal is to use MFA to establish reliable and uniform information about sources, flows, stocks, and sinks of phosphorous and nitrogen in the Danube Basin (Brunner and Lampert, 1997; Somlyódy, Brunner, and Kroiß, 1999). The main difference between Case Studies 2 and 3 is the scale: Instead of an area of 66 km2 and a population of 28,000 inhabitants (case study 2), the “Danube” case study covers an area more than 1000 times larger (820,000 km2) and includes 12 countries with a population more than 3000 times larger (85 million inhabitants). Despite the large difference in scale, the same MFA methodology is applied.

Key questions of this case study are as follows: what are the main sources of nutrients, and what measures are appropriate to reduce the nutrient flows to environmentally acceptable levels? Traditionally, emission inventories and ambient water-quality measurements are used to answer these questions. As a novel approach, comprehensive material flow analysis is applied to the entire watershed. The main advantage is that all nutrient-related processes in the region are looked at uniformly, and the total inputs, outputs, and stocks are investigated. Nutrient flows are tracked from their very beginning (fertilizer, animal feedstock, agricultural production) to the consumer (private households), to waste management, to surface water and groundwater, and finally to the River Danube. Since the balance principle is applied to all processes, cross-checking of flows and stocks becomes possible at many points within the system investigated.

3.1.3.1  Procedures

As mentioned before, one of the main tasks when exploring such a large system is to set up a broad international group that learns and uses the same MFA methodology. Ten national teams from Austria, Bulgaria, Czech Republic, Germany, Hungary, Moldavia, Romania, Slovakia, Slovenia, and Ukraine, each consisting of several experts, are participating in the study. In a first step, the common MFA methodology as well as water-quality goals and principles are established. The system boundaries are defined in space and time. The least number of processes is selected that allows full description of all necessary nutrient flows and stocks and still does not result in excessive work. To facilitate assembly of the individual data and results, the system is defined uniformly for all teams (see Figure 3.6).

Next, data are collected to balance each of the processes of Figure 3.6. Existing measurements, regional statistics, literature data, expert advice, and sometimes additional measurements are used to assemble a data set as comprehensive as possible. For example, for the process agriculture including soils, this means finding information about all process inputs such as mineral fertilizer, atmospheric deposition, nitrogen fixation, sewage sludge, compost, seedlings, and process outputs such as crops harvested, animal products, eroded soil, gaseous losses, leachate, and percolation. Manure is recycled within the process and thus can be looked at as both an output and an input at the same time if no export or import of manure takes place. Stocks comprise nutrients in the soil, stored manure, animal and vegetable biomass, and stockpiles of fertilizer. The procedure is similar to the one described for case study 2 in Section 3.1.2. All other processes of Figure 3.6 are balanced similarly.

Figure 3.6   System definition for nutrient balancing in the Danube River watershed. The same system is used for all national balances and for the balance of the total catchment area. (From Somlyódy, L. et al., Nutrient Balances for Danube Countries. Final Report Project EU/AR/102A/91, Service Contract 95-0614.00, PHARE Environmental Program for the DanubeRiver Basin ZZ 9111/0102. Vienna, Austria: Consortium TU Vienna Institute for Water Quality and Waste Management, and TU Budapest Department of Water and Waste Water Engineering, 1997.)

Some processes are not easy to balance: erosion from forest and agricultural soils in alpine areas can only be roughly estimated. Denitrification in natural systems (e.g., soil, aquifer) is not well known. The fate of intermediate stocks in the River Danube and in the soil over time is not sufficiently understood yet. Data about the efficiency of wastewater treatment and about corresponding nutrient removal are not available in all of the Eastern European countries. During the time of centrally planned economies, much information on agriculture, water quality management, and waste management was collected and stored on a large scale. However, since the transition of these economies to a free-market economy, much less information is available. In part, this is because it is too costly to gather comprehensive data. On the other hand, the price to access existing information increased dramatically after the economic transition.

It is important that all partners exchange information during collection of data. They also must make sure that they use compatible figures. For instance, it is likely that the balance of a cow (a process in MFA terminology) is similar in most countries of the Danube Basin, and thus that the input and output figures are comparable for most teams. If there are differences, like the significant variation between the nutrient metabolism of a Ukrainian and an Austrian cow, explanations must be available. Often, the balance principle brings such differences to light and allows cross-checking and verifying such differences. Thus, the balance principle can be highly valuable in the negotiation process within a group comprising teams from many countries. It ensures transparency, enables data verification, and results in acceptance of each other’s results.

3.1.3.2  Results

A large-scale multinational MFA proves to be a time- and resource-consuming task. It takes time until all know-how is transferred, incorporated, and well applied in practice. It takes even more time to find all the necessary data. The exchange of information, iterations, and adaptations of the individual work of the different participating groups again takes time. It may be that a partner is not able to perform its task and that a new team has to be engaged in the middle of the project. Given these factors, it is clear that a large-scale MFA cannot be undertaken in a couple of weeks. It is likely to take 1 year or more to complete such a comprehensive and challenging task.

The Danube case study produces a lot of data and many results that can be used to support decisions regarding water quality management. For results regarding wastewater management and water pollution control, see Zessner, Fenz, and Kroiss (1998). The results presented here identify the most important sources and pathways of nutrients in the Danube River basin. The purpose of this presentation is to demonstrate how a very large data set can be compressed to present the relevant key results. In Figure 3.7, the system given in Figure 3.6 is transformed and presented in a format that identifies the major imports and exports of nutrients into the surface waters of the Danube catchment area. This format still shows the same 11 processes, but it centers on the surface waters. Imports and exports are quantified, and conclusions regarding the importance of all flows can be taken according to their mass flow. Note that mass flows alone do not permit one to evaluate the effects of nutrients in surface waters. It is necessary first to transform nutrient flows into nutrient concentrations by dividing nutrient flows through water flows.

Figure 3.7   Nitrogen and phosphorous flows in surface waters of the Danube River basin in 1992, kt/year. The system shown in Figure 3.6 is transformed to Figure 3.7 in order to present the main inputs and outputs of the surface waters. This allows identifying the importance of each process as a source of nutrients for the Danube River. (From Somlyódy, L. et al., Nutrient Balances for Danube Countries. Final Report Project EU/AR/102A/91, Service Contract 95–0614.00, PHARE Environmental Program for the DanubeRiver Basin ZZ 9111/0102. Vienna, Austria: Consortium TU Vienna Institute for Water Quality and Waste Management, and TU Budapest Department of Water and Waste Water Engineering, 1997.)

The same case study data are even more condensed in Figure 3.8. It becomes clear that the more aggregated the data are, the easier it is to get a message across: Figure 3.8 clearly identifies agriculture as the dominant source of nutrients in the Danube Basin. In addition, it suggests the hypothesis that improving existing WWTPs is probably more important for reducing emissions than connecting all households and industries to sewers. This hypothesis of course has to be verified with data about the fraction of people and companies connected to sewer systems, and on nutrient removal in waste-water treatment within the catchment area. In addition, the costs of upgrading WWTP and of connecting households and industries to sewers need to be known. The advantage of an MFA approach is that such hypotheses can be set up, making further investigations more straightforward.

Another way of presenting data is displayed in Tables 3.12 and 3.13, which combine results about sources (agriculture, household, industry, etc.) with results about pathways. Hence, they are well suited to serve as a basis to set priorities for decisions regarding nutrient emission reductions. The following conclusions can be drawn: Agriculture is the main source of nutrient inputs into the River Danube. Erosion and runoff are the main pathway of nutrients from agriculture to surface waters for both P and N. The direct inputs of liquid manure are high and amount to about 12% of total N and 20% of total P loads. Private households are the second largest source of nutrients, contributing around 20% of total N and P. Approximately 10% of both N and P originates from industry. MFA yields the following results regarding pathways: for surface waters, the main nitrogen load (35%) is due to exfiltration of groundwater. Erosion/runoff; direct inputs from agriculture, households, and industry; and effluents from WWTP each contribute about 20–25% of the total load to the Danube River. About 60% of N and 40% of P originate from nonpoint sources. Retention (sedimentation and denitrification) amounts to 15% of N and close to 50% for P.

Figure 3.8   Sources of nutrients in the catchment area of the Danube River in 1992, kt/year. The charts show clearly the importance of agriculture for P and N emissions. Erosion and leaching from agricultural fields dominates all sources (A). Direct discharges and discharges via treatment plants of animal wastes are the second most important path of nutrients to the Danube (B). The direct inflows from private households and industry (others) are smaller than the effluents from WWTPs. Diffusive inputs from forestry are small for P but more significant for N. (From Brunner, P. H. and Lampert, Ch., EAWAG News, 43E, June 1997, pp. 15–17. With permission.)

Table 3.12   Sources and Pathways of Nitrogen in the Danube River (1992)

N, %

Agriculture

Private Household

Industry

Others

Total

Erosion/runoff

17

0

0

4

21

Direct discharges

12

4

6

2

24

Base flow

17

4

0

13

35

Sewage treatment plant

6

10

5

0

20

Total

51

19

10

19

100

Source: Brunner, P. H., and Lampert, C. (1997). “Nährstoffe im Donauraum, Quellen und letzte Senken” (“The Flow of Nutrients in the Danube River Basin”). EAWAG News, 6 (43D+E+F), 15–17.

Note: Total input equals 100%. Base flow represents flows to the Danube via groundwater.

Table 3.13   Sources and Pathways of Phosphorous in the Danube River (1992)

P, %

Agriculture

Private Household

Industry

Others

Total

Erosion/runoff

28

0

0

3

31

Direct discharges

18

6

6

3

33

Base flow

2

2

0

2

6

Sewage treatment plant

9

14

7

0

30

Total

57

22

13

8

100

Source: Brunner, P. H., and Lampert, C. (1997). “Nährstoffe im Donauraum, Quellen und letzte Senken” (“The Flow of Nutrients in the Danube River Basin”). EAWAG News, 6 (43D+E+F), 15–17.

Note: Total input equals 100%. Base flow stands for flows to the Danube via groundwater.

The detailed results of all teams (presented in Somlyódy et al., 1997) make it possible to identify the nutrient contribution of each country. Almost half of the nutrient input into the Danube catchment area comes from Romania, while Austria, Germany, and Hungary together contribute about one-third of the total load. An interesting and not yet resolved question concerns the allocation of the dilution potential of the River Danube to the riparian countries: what is the amount of nutrients a country may release to the Danube? Assuming that the carrying capacity for nutrients of the Danube basin is known, there are several ways to answer this question. A per capita load limit favors countries with large populations; it can be justified on the grounds that every human being has a similar metabolism and thus should have an equal share. This method of allocation neglects the fact that some countries are better suited for agriculture than others and, because of their agricultural activities, will have a larger nutrient turnover. A per-area load limit favors large countries but does not consider population density. A per-net-precipitation limit takes into account the regional dilution of nutrients: if regional nutrient emissions are heavily diluted by a large amount of net precipitation (precipitation minus evapotranspiration), the resulting concentration in the River Danube may still be low and below carrying capacity. However, this argument does not hold for the Black Sea, where the total flow is important.

The River Danube, like many of the large river systems in the world, has become an important path for wastes such as nutrients from countries within its catchment area. The main question is how future loads will develop. At present, eastern European countries are experiencing a low standard of living. It can be assumed that the per capita turnover will rise rapidly in the future and that population will grow again, too. Both factors will increase total nutrient turnover as well as waste generation. If no actions are taken, the capacity of the River Danube, the delta, and the final sink Black Sea will be overloaded, with serious ecological and economic consequences. It is not the “classical” resource problem (lack of nutrients) that will limit the development of the region. Rather, it will be the lack of appropriate sinks that restricts progress. Strategies to limit nutrient loads need to be discussed and developed. Crucial issues will be type of agriculture; population density, lifestyle, and consumption; and standards and enforcement for emissions from industry, households, and WWTP. In any case, transregional and international agreements will be required to solve the allocation problem. As proven by this case study, MFA can play a major role in supporting policy decisions to protect the River Danube, the delta, and the Black Sea.

3.1.4  Case Study 4: Support Tool for EISs

In an environmental impact assessment (EIA), potential impacts of a project such as a new plant (power plant, municipal incinerator) or system (road, harbor) on the environment are identified, quantified, evaluated, predicted, and monitored. In the event that a significant impact is acknowledged, more detailed studies have to be carried out, finally leading to the preparation of an EIS. Meanwhile, more than half of the nations around the world require an EIA for certain projects. The US National Environmental Policy Act of 1969 (NEPA) provided one of the earliest sets of EIA requirements. Many countries followed and modeled their requirements after NEPA (Canter, 1996). In Europe in the 1970s, some federal environmental legislation began to require mandatory EIAs. However, the final breakthrough was not achieved until the European Directive 85/337/EEC on the “assessment of the effects of certain public and private projects on the environment” became effective (European Commission, 1985). MFA that is based on a sound methodological framework is considered a useful tool to support both EIA and EIS (Brunner and Baccini, 1992).

In the case study SYSTOK (Schachermayer, Rechberger, Maderner, and Brunner, 1995), the impact of electricity production from coal on the local and regional environment is investigated. For this purpose, a three-process system comprising coal mining, the coal-fired power plant, and landfilling of the ash is defined (Figure 3.9). The contribution of this system to the anthropogenic and geogenic metabolism of the region is determined (Figure 3.10). The case study focuses on particular technologies of mining, power generation, APC, and landfilling. It is clear that the results cannot be generalized to other coal-fired power plants. If the technology is changed, if the coal composition and heating value is different, or if the landfill leaks to the groundwater, the impact on the environment will also be different. The reason for including this case study is to show that MFA serves well as a base for EIS and EIA. MFA can be applied independently of the technology used or the input composition.

Figure 3.9   Systems definition of electricity production in a coal-fired power plant including the processes coal mining and ash landfilling. The figure includes all exports and imports that are necessary to establish an EIA and EIS.

Figure 3.10   The contribution of a coal-fired power plant to a region’s metabolism. The system consists of the three processes, coal mining, power generation, and landfilling of the ash that can be summarized in a single process, electricity production from coal.

The findings of SYSTOK serve as a basis for the operator to optimize the plant and to prepare an EIS. One of the features of SYSTOK is that a system, in this case electricity production from coal, is to be embedded comprehensively into a region. However, the definition of the region is not clear a priori, so appropriate approaches for the demarcation of spatial and temporal systems boundaries have to be developed. SYSTOK exemplifies how a comparatively simple MFA can lead to valuable new conclusions regarding system boundaries.

3.1.4.1  Description of the Power Plant and Its Periphery

All coal used in the power plant (1 million t/year) is produced in a nearby open-pit coal mine with a surface area of 2 km2. Temporary interruptions in mining or power generation are buffered by interim coal storage of 2 million tons, located at the premises of the power plant. In order to gain 1 ton of coal, an average of 6.7 tons of overburden have to be removed. Of this mining waste, 70% is transported to locations outside of the mine, and the rest is filled back into the mine. For extraction, transportation (trucks and conveyors), and processing of the coal and overburden, 1400 t/year of gasoline and 25 million kWh/year of electricity are needed. The reservoir of coal still in the mine is estimated to be 11 million tons.

Brown coal has a low heating value of 8.4 to 13 MJ/kg and bears much non-combustible ash, making transportation expensive. Thus, the power plant is situated near the coal mine in order to avoid long transport distances. Because of high specific costs of power generation at this plant, it operates only during peak demand. The average time of operation is 4000 h/year, and the maximum electric capacity is 330 MW. The coal (C) is pulverized to coal dust and injected into the combustion chamber at a feed rate of 300 t C/h at full load. The average coal consumption (ca. 265 t C/h) is somewhat lower due to periods of partial loading. Approximately 33 kg/t C of wet bottom ash (40% water content) is removed from the water basin that serves as an air seal toward the combustion chamber. The power plant is equipped with an electrostatic precipitator (ESP) to collect particulates (ESP ash) with an efficiency of 99.85%. The ratio of dry bottom ash to dry ESP residue (or fly ash) is ca. 1:9. The ESP residue is humidified (20% water content) and, together with bottom ash, transported by conveyor belts to the landfill. Sulfur dioxide (SO2) is removed from the flue gas in a wet scrubber using limestone (CaCO3, 20 kg/t C). Absorption of SO2 occurs with an efficiency of more than 90%, producing ca. 37 kg of gypsum per t C (water content is 12%). Finally, nitrogen oxides (NOx) are reduced to molecular nitrogen (N2) in a catalyst by injecting ca. 1.2 kg of ammonia solution (NH3, 33%) per t C. Other chemicals used include hydrochloric acid (HCl, 0.02 kg/t C, 33%) and sodium hydroxide (NaOH, 0.007 kg/t C, 50%) for conditioning of feeder and perspiration water as well as chemicals used to stabilize water hardness, inhibit corrosion, etc. (<0.002 kg/t C). Combustion requires 4000 kg/t C of air, resulting in ca. 5000 kg/t C of off-gas. Water is used to feed the steam cycle (20 kg/t C) and for cooling (2300 kg/t C). Waste heat is dissipated in a cooling tower having an air-exchange rate of 8500 kg/t C. Part of the cooling water (25%) is discharged to surface water; the rest evaporates in the cooling tower.

The landfill for the ashes consists of a natural vale made of quartzite with a surface of ca. 0.35 km2. This basin, which is a former mine area of coal, has a low permeability for water. Hence, precipitation (ca. 1000 l/m2) creates a lake in the landfill. The water is used during dry periods to wet the ashes and is supposed to finally evaporate. No leachate leaves the ash landfill as long as the landfill is maintained by the operator.

3.1.4.2  System Definition

The system displayed in Figure 3.9 is divided further into three processes: coal mining, power plant including APC, and ash landfill. For EIA and EIS, a black-box approach is appropriate. It is not necessary to take into account more detailed subprocesses that would require much additional information. The substances are selected based on knowledge about combustion processes and the main product “coal.” Carbon is the priority substance in any combustion process, since the content of organic carbon in ashes and flue gas is a measure of combustion efficiency and of the formation of organic pollutants. Past experience with coal-fired power plants has shown that they are major sources of emissions of SO2, NOx, HCl, and heavy metals such as arsenic and selenium (Greenberg, Zoller, and Gordon, 1978). Modern legislation (Clean Air Act) sets stringent standards for these emissions. The procedure to select substances relevant for EIAs is given in Table 3.14. The substance concentrations in coal are compared with concentrations in the Earth’s crust. The following six elements are significantly more concentrated in coal than in the crust: As, Se, Hg, S, Cl, and N. The relevant goods are determined by mass-balancing each process and by assuming that goods inducing a substance flow <1% of the total throughput of a substance can be neglected. This requires an iterative approach among the steps establishing substance balances, selection of goods, and establishing total mass balances. For a first step, the premises of mining, power station, and landfilling are selected as a spatial system boundary. Materials balances are determined for an average operational year (temporal system boundary). Figure 3.9 shows the system defined.

Table 3.14   Determination of Relevant Substances for the Analysis of a Coal-Fired Power Plant by Relating Average Concentrations of Selected Substances in Coal and Ash to Concentrations in the Earth’s Crust

Substance

Concentration in Coal (A), mg/kg

Concentration in Ash (B), mg/kg

Concentration in Earth’s Crust (C), mg/kg

Ratio A/C

Ratio B/C

Arsenic

12

64

1

12

64

6

37

13

0.5

2.8

0.1

0.64

0.2

0.5

3.2

Chromium

30

170

100

0.3

1.7

Copper

13

84

55

0.24

1.5

Selenium

0.9

2.2

0.05

18

44

Zinc

27

190

70

0.4

2.7

Nickel

27

130

75

0.36

1.7

Mercury

0.3

0.45

0.08

3.8

5.6

Sulfur

6500

3000

260

25

12

Chlorine

1000

130

7.7

Nitrogen

12,000

20

600

3.1.4.3  Results of Mass Flows and Substance Balances

The mass flows of all goods are listed in Table 3.15. Except for mining, where it is not known whether and how much runoff and leachate is draining to surface water and groundwater, all processes are mass-balanced. The main quantitative features of the system are as follows: (1) The coal mine will be exhausted in about 10 years. (2) Compared with the emissions of the power station, the off-gas from mining is negligible. (3) The power plant is actually a giant fan moving huge amounts of air. The mass of air required for cooling dominates the mass flows of the system, exceeding combustion airflow by more than an order of magnitude. Additionally, the power plant consumes more than 1 t of water per ton of coal. (4) The main flow of solid waste is generated during mining (overburden). The power plant produces a large net “hole,” since coal is extracted and the volume of backfilled ash is much smaller.

Table 3.15   Mass Balance for Goods of the System Electricity Production by a Coal-Fired Power Plant, 1000 t/Year

Input

Output

Process Coal Mining

Combustion air I

29

Off-gas I

31

Storm water I

2000

Vapor

2000–n.d.

Gasoline

1.4

Overburden

4700

Runoff and leachate

n.d.

Coal

1000

Total

2000

7700

Stock (coal)

11,000

1000

Stock (total)

145,000a

Stock change

–5700a

Process Power Plant

Coal

1000

Ashes

280

Cooling air (input)

85,000

Cooling air (output)

87,000

Combustion air II

4000

Off-gas II

5000

Water

2500

Wastewater

760

Limestone

20

Gypsum

390

Total

93,000

93,000

Stock (coal)

2000

Stock change

0

Ash Landfill

Storm water III

350

Vapor III

350

Ashes

280

Total

630

350

Stock (ashes)

4100

Stock change

+280

Note: Values are rounded; n.d. = not determined.

Notes:

a  Estimated, includes coal and overburden.

The substance balances as displayed in Figure 3.11 provide an overview of the qualitative behavior of the system. These balances are calculated based on data for substance concentrations in overburden (“soil” in Table 3.19), gasoline, and coal (Table 3.14), and the TCs for the power plant that have been measured on site (see Table 3.16).

A comparison of sulfur concentration in coal and in the Earth’s crust shows that a large amount of sulfur is extracted from the crust via coal. The power plant transfers sulfur quite efficiently into the product gypsum (86%). Before desulfurization became a part of the APC system of the plant, this sulfur was emitted, too, resulting in a sulfur transfer to the atmosphere of >90%.

The foremost flow of mercury is associated with the good “overburden” or mining waste. During combustion, the atmophilic mercury is evaporated and leaves the plant evenly distributed between ESP residue and off-gas. A small part is precipitated with gypsum. Electricity production in coal-fired power plants extracts significant amounts of mercury from the Earth’s crust and disperses a substantial amount via the stack. So far, 2.2 t of mercury has been deposited in the ash landfill and more than 10 t in the overburden deposit. It is interesting to compare these stocks with other mercury stocks. The consumption of mercury has been assessed to range between 0.66 g/capita/year in Stockholm and 1.5 g/capita/year in the United States. For stock, 10 g/capita has been determined in Stockholm (Jasinski, 1995; Bergbäck, Johansson, and Mohlander, 2001). This means that the landfill contains the same amount of mercury as is stored in buildings, infrastructure, and long-lasting goods associated with a region of 220,000 inhabitants. In the overburden deposit, a mercury stock corresponding to more than 1 million persons is contained. While this mercury is dispersed over a large region (2000 km2), the landfill’s mercury is located in a comparatively small area (0.35 km2) and therefore is easier to control.

Figure 3.11   Substance balances for sulfur, mercury, and selenium for the system electricity production (flows are in t/year and stocks in t). The stock in the process coal mine includes coal and overburden.

Table 3.16   Partitioning of Selected Substances in a Coal-Fired Power Plant, % of Input

Substance

ESP Fly Ash

Bottom Ash

Gypsum

Off-Gas II

Carbon

2

1.3

<0.05

97

Sulfur

5.7

0.93

86

7.6

Mercury

50

0

5

45

Arsenic

99

0.4

0.4

<0.1

Selenium

52

0.6

28

20

About half the amount of selenium that enters the power plant is transferred to the ashes and landfilled, and 20% is emitted into the atmosphere. The relevance of this path for the environment will be discussed in the following sections. Note that gypsum holds about 30% of the selenium contained in coal.

The relation between the power plant and the surrounding region is discussed in the following sections.

3.1.4.4  Definition of Regions of Impact

The mine, power plant, and ash landfill have several impacts. In the first place, they supply power to consumers. Thus, a region can be defined in a product-related way. Second, they create jobs and income and thus serve an economic region. Third, the mine, power plant, and landfill are situated in an administratively defined region such as a community or a province. Fourth, they have an impact on the environment. Depending on the “conveyor belt” that transports an emission, substances released by the coal-fired power plant may affect a small or large area. The ash landfill has only a local impact during the transfer of the ash to the landfill site. Sulfur and mercury emitted by off-gases are distributed over a large (global) area through atmospheric transport. Thus, the size of the region is determined by the distance an emitted substance travels from the plant and by the effect of this substance on the environment. In each of these four regions of impact, there are specific problems, benefits, and stakeholders with regard to the power plant.

In the SYSTOK study, three regions of impact are defined.

3.1.4.4.1  Product-Related Region

The power plant supplies electricity to a certain area. In principle, this area is defined by the amount of electricity the plant supplies, by the demand per customer, and by the population (or customer) density. Due to the liberalization of the electricity market, this area can only be defined as a virtual region, since customers may be served far away from the plant. The product region is changing constantly in response to the market situation. Thus, for SYSTOK, the size of this virtual region is calculated by the average electricity production of the plant, the average consumption per capita, and the national population density.

$A P R = P ⋅ h ⋅ f e ⋅ ρ p = 2100 k m 2$

where

• P = output of the power plant (330 MW)
• h = operating hours per year (4000 h)
• f = factor considering partial-load operation (0.85)
• e = specific demand for electricity in Austria (including private households,industry, service, administration, traffic, agriculture; 5.6 MW•h/ capita/year)
• P = population density in Austria (95 capita/km2)

The region is defined by the borders of the administrative unit, i.e., the district that represents the legislative and administrative authority for the plant operator. The advantage of this definition is twofold:

1. The region as a spatial unit is well accepted and known and is governed by an authority supervising the plant.
2. Data are usually collected on the level of administrative regions, thus facilitating the allocation of data.

3.1.4.4.3  Region Defined by Potential Environmental Impacts

As mentioned before, this area is different for particulate, gaseous, and aqueous emissions, and it is also substance specific. For gaseous emissions, dispersion models help to determine the region. Criteria for selecting the boundaries may be

1. Concentration limit (ambient standard) for a substance (ccrit = clim)
2. A fraction of the concentration limit, since the limit should not be used up by the power plant alone (ccrit = clim/10)

Figure 3.12   Application of a dispersion model to determine the border of a substance-specific, environmentally relevant region. The regional boundary with regard to substance x is defined as the area within cx > cx crit.

Figure 3.13   Regions defined according to three criteria—consumers of electricity (product-related region), administrative designation, and environmental impacts—overlap but are not identical.

Table 3.17   Sizes of Differently Defined Regions for a 330 MW Coal-Fired Power Plant

Region DefinitionSize, km2
Product (electricity)2100
Environmental (example SO2)620

Note: The boundary for the environmentally defined region of SO2 is determined by ccrit = cgeog × 1.1, where cgeog = 10 mg/m3.

3. A proviso that the power plant does not change the current ambient concentrations in a significant way (ccrit = cbackground × 1.1)
4. A proviso that the power plant does not change the geogenic (or “natural,” without present anthropogenic influences) concentrations in a significant way (ccrit = cgeog × 1.1; see Figure 3.12)

Figure 3.13 exemplifies the differently defined regions, and Table 3.17 presents sizes of regions calculated according to different definitions for SYSTOK.

3.1.4.5  Comprehensive Regional Significance

In addition to the procedure given in Figure 3.12, there are other means of determining the relevance of power produced, emissions, and wastes of the coal mine, power plant, and landfill for the region. In Figure 3.14, emissions from the power plant are compared with the total regional emissions of various air pollutants. As a basis for comparison, the product-related region is chosen. The emissions of the region (Ri) are assessed by the following equation:

$R i = X i − P i ⋅ A P R A A U + P i$

with Xi as the average emissions of an appropriate administrative unit (AU) for which data are available (state, district). Pi stands for the emissions of the power plant, and APR and AAU are the respective areas of product-related region and administrative unit.

The power plant is responsible for about half of the region’s CO2 emissions. Removal of SO2 and catalytic reduction of NOx significantly reduced the plant’s contribution to the regional emissions. A further decrease in particulates and NOx will result in modest improvements of air quality only (<10%). For CO, the power plant’s emissions are not relevant at all.

Figure 3.14   Contribution of the power plant to the total emissions of the product-related region (=100%). * Indicates emissions before introduction of advanced air-pollution control.

The throughput of heavy metals by the power plant is set in relation to the product-related region, too. Table 3.18 shows the annual flows of selected metals and nonmetals through the power plant. Comparing these flows to the corresponding total flows through the region is time consuming. Much information, which is usually not available, is required. Therefore, only the materials flows through private households are taken into account. The contributions of industry and the service sector are not considered. Since the consumption of heavy metals in private households is not well known either, the amount of metals in MSW, which is available from measurements (see Section 3.3.1), is taken as a reference. Table 3.18 gives the ratio of substance flows via coal and flows via MSW on a mass-per-capita and year basis. For comparison, the flows of carbon and sulfur in fossil fuels are also shown. The coal-fired power plant is responsible for a high turnover of arsenic, selenium, and sulfur when compared with the generation of MSW in private households.

The relevance of heavy metal emissions can be assessed by the “anthropogenic versus geogenic flows” approach described in Chapter 2, Section 2.5.8. Applied to SYSTOK, the following question has to be answered: does the power plant change substance concentrations in any of the environmental compartments? A simplified model is used to identify the effect of power plant emissions on the soil concentrations in the product-related region of 2100 km2. By deposition, metals such as lead and cadmium may be accumulated in the top 30 cm layer of soil, the soil depth that is turned over by plowing. Assuming an average soil density of 1.5 kg/ m3, the regional compartment soil holds a mass of 2100 × 106 m2 × 0.3 m × 1.5 kg/m3 = 950 × 106 kg.

Table 3.18   Comparison of Substance Flows Induced by a Coal-Fired Power Plant and by MSW in the Product-Related Region

Substance

Flow via Coal, g/Capita/Year

Flow via MSW, g/Capita/Year

Coal/MSW

Arsenic

18

0.8

21

Copper

19

100

0.2

9

170

0.1

0.15

2.3

0.1

Mercury

0.45

0.4

1.2

Selenium

1.3

0.2

8

Chromium

45

53

0.9

Nickel

40

18

2.3

Zinc

40

230

0.2

Carbon

420,000

930,000.0a

0.4

Sulfur

10,500

2000.0a

5

Notes:

a  Figures for carbon and sulfur include the contribution by fossil fuels, which is much larger than MSW.

The cumulative emissions of the power plant can be estimated based on the total coal throughput of 33 million t, the mean substance concentrations in coal, and the TCs for off-gas.

$E i t = 33 × 10 6 t × c i mg / k g × 10 6 × T C$

Table 3.19 shows that only emissions of selenium and mercury are of relevance. Note that the model draws on two major simplifications. First, it is understood that the product-related region is identical to the substance-specific impact-related regions. Second, it is assumed that deposition is evenly distributed over the region. Concerning the first simplification, one has to consider that particulate removal takes place in an efficient ESP. Hence, the emitted particulates are small, most with diameters <2 μm. Such particles (aerosols) have a long residence time in the atmosphere (≈1 week) and do not sediment in the vicinity of the power plant. They are washed out of the atmosphere by precipitation. Thus, the average frequency of rainfall determines the travel distance of these particles. In Central Europe, this is around 1 week and results in a significantly larger region than the product-related region (at least 10 times larger). Hence, the product-related region overestimates substance accumulation in the soil by more than one order of magnitude.

Table 3.19   Accumulation of Metals in Soils due to Emissions of a Coal-Fired Power Plant

Substance

Coal, mg/kg

TC

Emission, t/τa

Soilc, mg/kg

Soil Reservoirb, t

Enrichment, %/τa

As

12

0.001

0.4

8

7600

0

Pb

6

0.005

0.99

25

23,800

0

Cd

0.1

0.041

0.14

0.2

190

0.1

Cr

30

0.001

0.99

40

38,000

0

Cu

13

0.014

6

15

14,300

0

Se

0.9

0.2

5.9

0.1

95

6.3

Zn

27

0.009

8

30

28,500

0

Ni

27

0.003

2.7

20

19,000

0

Hg

0.3

0.45

4.5

0.2

190

2.3

Notes:

a  τ = total time of operation (ca. 33 years).

b  Soil reservoir is calculated based on the product-related region.

c  Scheffer, 1989.

The relevance of the second simplification can be assessed when dispersion models for particulates are analyzed. In most cases, the ratio between maximum concentration and mean concentration is less than 10. This means that the assumption of substances being evenly distributed over the region underestimates the actual accumulation by a maximum factor of 10. Considering both simplifications, it can be concluded that the chosen model rather overestimates the enrichment in soils caused by the power plant, and the emissions can be rated as not relevant with the exception of selenium and mercury, where more detailed investigations are necessary.

3.1.4.6  Conclusions

The case study shows that the power generation based on coal is relevant for enhanced flows of arsenic, selenium, mercury, sulfur, and carbon within the region’s metabolism. The landfill represents a considerable reservoir for certain metals within the region. It can be considered as a point source that is comparatively easy to control. On the other hand, in the event of insufficient immobilization or leaching of the containment, the landfill can substantially contribute to the pollution of the region’s environment. The landfill requires constant water management. Solutions have to be developed for the future, when the power plant is not in operation anymore and when funds are no longer available for landfill aftercare. Since leaching will be a constant threat, it is necessary to investigate whether immobilization of filter ash is more economic than aftercare of the landfill for several thousand years.

The emissions of the power plant are of little relevance for the region, with the exceptions of CO2 (greenhouse gas) and, of minor importance, SO2. The comparatively low retention capacity for mercury and selenium should be a focus for future improvement efforts of the operators after more detailed investigations and measurements. Generally, the study shows that the power plant actually does not pose a severe burden to the region’s environment. The operators use these results for their EIS and for communication with concerned local people.

Problems—Section 3.1

Problem 3.1:

• Assess the effects of the following measures on the regional lead flows and stocks given in Figure 3.1. Show quantitatively and discuss the following effects of reductions in lead concentrations in soil, surface waters, and landfill: (a) ban on leaded gasoline, (b) ban on application of sewage sludge to land, and (c) construction of a new MSW incinerator in the region with an air pollution control efficiency for lead of 99.99% treating the waste of 280,000 persons from the Bunz Valley and neighboring regions.

Problem 3.2:

• Consider a region of 2500 km2 and 1 million inhabitants. Only one river flows through this region. At the inflow, the river has a flow rate of 1 billion m3/year, and the concentration of phosphorous (P) is 0.01 mg/L. The river discharges into a lake that represents a reservoir of 2.8 billion m3; residence time of water in the lake is 1 year. (Precipitation, evaporation, etc., are not considered; assume that the river is unchanged when flowing through the lake.) The anthroposphere of the region comprises the following processes: agriculture, food industry, private households, composting of biomass wastes from private households, and wastewater treatment. Per capita consumption of P for nutrition is 0.4 kg/capita/year; 20% of this demand is supplied by food industry within the region. For cleaning purposes, 1 kg/capita/year of P is used, with 70% contained in detergents for textiles; all detergents are imported. Assume that 90% of total nutritional P and 100% of total detergent-based P are directed to the WWTP. The transfer coefficient (TC) for P into sewage sludge is 0.85. The remaining 10% of nutritional P is contained in biomass waste from households that is composted without loss of P and applied to the soil. The stock of P is assessed at ca. 380,000 t. Agriculture imports 2400 t of P in fertilizers and 1200 t of P in animal feed; 80% of this P flow goes to the soil as manure, dung, and residues from harvesting. The balance is input to the regional food industry. The TC for P into food production wastes is 0.6. Food products that are not consumed within the region are exported. Approximately 1% of P that is annually applied to soil escapes to the surface water (river) as a result of erosion.

1. Draw a qualitative flowchart for the region described (system boundary, flows, processes).
2. Quantify the P flows and stocks (t/year and t) of the system. What is the accumulation of P in the soil?
3. Assume that P in detergents for textiles is phased out. Is this measure sufficient to prevent eutrophication (limit for eutrophication = 0.03 mg/l) of the lake?
4. What further measures do you suggest to prevent eutrophication?

Problem 3.3:

• Discuss and quantify the reaction time of different measures to control phosphorous flows in the Danube River basin. Reaction time is defined as the time span in days, weeks, months, years, etc., between the decision to take an action and a measurable effect in the Danube River. Note that reaction time also includes planning and implementation (construction, startup).

1. Reduction of phosphorous fertilizer input to soils by a resource tax
2. Connecting 95% of all private households to sewer systems
3. Increasing the removal efficiency for P in sewage treatment from 50% to >80%
4. Banning direct discharges from agriculture
5. Assessment of reaction time if P is banned in all detergents (assuming that one-third of the P flow through private households in the Danube basin stems from phosphorus-containing detergents)

Draw a general conclusion regarding the reduction of P flows to the river Danube when you evaluate the effectiveness of the measures discussed.

Problem 3.4:

• Compare the materials turnover of a coal-fired power plant and an MSW incinerator. The feed rate for coal is 300 t/h; for an MSW, it is 30 t/h. Select the substances As, Pb, Cd, Cu, Se, Zn, Hg, S, Cl, and N. Substance concentrations for coal are given in Table 3.19; for MSW, see Table 3.20. Discuss your findings with respect to air pollution control.

The solutions to the problems are given on the website http://www.MFA-handbook.info.

Table 3.20   Example of Average Substance Concentrations in MSW, mg/kg

As

Pb

Cd

Cu

Se

Zn

Hg

S

Cl

N

10

500

10

1000

1

1200

1

3000

7000

5000

3.2  Resource Conservation

The main advantage of the application of MFA for resource conservation is the comprehensive information about sources, flows, and sinks of materials. This makes it possible to set priorities in resource conservation, to recognize early the benefit of material accumulations (e.g., in urban stocks), and to design new processes and systems for better control and management of resources. In this chapter, two groups of substances (nutrients and metals) and two groups of goods (plastic materials and construction materials) are discussed in view of resource conservation.

3.2.1  Case Study 5: Nutrient Management

Nutrients are essential resources for the biosphere. Life without nitrogen and phosphorus is not possible. The atmosphere represents an unlimited reservoir for nitrogen. The industrial transformation of N2 to chemical compounds such as ammonium and nitrate that can be taken up by plants requires energy. Hence, the amount of nitrogen available within the anthroposphere is limited mainly by energy supply. In contrast, phosphorus is taken from concentrated phosphate minerals that are limited in extent. It is assessed that at present consumption rates, concentrated phosphate deposits might be used up in about 100 years (Steen, 1998). Thus, in order to conserve energy and resources, both nitrogen and phosphorus have to be managed with care.

The purpose of the following case study is to show how MFA can be used to set priorities in resource conservation. Measures are analyzed in view of their effectiveness regarding nutrient recycling. The total flows of phosphorus and nitrogen are investigated. The entire activity to nourish is analyzed from agriculture to food processing to private households. Losses and wastes are identified and quantified along the process chain. Because MFA of nutrients is discussed in several chapters of this book (see Sections 3.1.2, 3.1.3, and 3.5.2), the main emphasis is on the interpretation of the results. The procedure for establishing nutrient flow analysis of the activity to nourish is not given in detail. For further information, see Baccini and Brunner (2012).

3.2.1.1  Procedures

The activity to nourish is investigated on a national level. A system comprising five processes for the production (cultivating, harvesting), industrial processing (including distribution), household processing, and consumption (including digestion) of food is defined and investigated (Figure 2.8). For each process, information about inputs and outputs is obtained from available sources, including national statistics about import, export, and production of fertilizer, agricultural produce, and food; agricultural information databases about the use of fertilizer and production of agricultural products; reports from food-processing companies, wholesale companies, and distributors of food; medical literature about human consumption and excretion of nutrients; and databases about concentrations and loadings of nutrients in wastewater, municipal solid wastes, and compost.

It is important to start with reliable data about the structure of the national agricultural sector: what are the main agricultural products; how are they produced; what is the nutrient input required for the production; and how large is the amount of nutrients actually harvested? Internal cycles of agriculture are to be investigated, such as the soil-plant-animal-manure-soil nutrient cycle. Figures for total losses of nutrients in agricultural practice are usually not available. Farmers use different definitions for wastes and losses. Shortfalls have to be calculated as the difference between total input and total output of the agricultural sector. The same method can be applied to the processes industrial processing and distribution to calculate or cross-check figures for losses, wastes, and wastewaters.

Using the sources pointed out previously, the process household can be balanced as shown in Figure 2.8. The average amount of food consumed per capita and per year is taken from national statistics. Note that if such statistics are based on bookkeeping of individual households, they usually do not contain out-of-house consumption; in such cases, it will be necessary to increase the figure for food consumption by 20–30%. Waste analysis data yield the amount of food residues in MSW. If such data are not available, it can be assumed that 5–10% of food purchased is discarded with MSW. Information about kitchen wastewater is taken from studies about sewage production in households. It can also be estimated that about 20–25% of food entering a household is discarded via the kitchen sink. Note that cooking water may contain considerable amounts of dry matter and (dissolved) salts. Of course, the partitioning of food in households between MSW, wastewater, and human consumption is a function of cultural aspects, too: in societies that are traditionally scarce in resources, the amount of kitchen wastes is considerably smaller. If grinders are installed in kitchen sinks, the food fraction in wastewater will be higher. If fast food plays a major nutritional role, food wastes in kitchens will be smaller because most food entering households has already been processed. In such cases, packaging wastes may be larger.

Data about respiration, urine, and feces is found in the medical literature on human metabolism. This information contains figures about N and P in food, urine, and feces, too. It is important to cross-check all data. The output of agricultural production can be compared with the input into the food industry, the output of the food industry to the consumption of the total population, and the output of the total population to the input into waste-water treatment and waste management. If the data for balancing the individual processes have been collected independently for each process, the redundancy of such cross-checking will be high, and the accuracy of the total nutrient balance can be improved significantly.

3.2.1.2  Results

To demonstrate the relevant results, the five processes in Figure 2.8 are combined into the three processes presented in Figures 3.15 and 3.16. Food-related flows of P and N through agriculture, industrial processing and distribution, and consumption are displayed on a per capita basis. In view of resource conservation, agriculture is the most important process, where 80% of P and close to 60% of N are lost during agricultural production. Losses are flows to groundwater, surface water, and air for N, and erosion/surface runoff and accumulation in soils for P. In order to optimize nutrient management, agricultural practice has to be changed as a first priority. Since nutrients are still comparatively cheap, there is no economic incentive yet for such a change. It seems timely to investigate how new or other technologies can make better use of nutrients in agriculture. While the primary objective today is to prevent nutrient losses in order to protect the environment, it is likely that within a century, resource scarcity of phosphorus may become a driving force for changes in agriculture.

Figure 3.15   Phosphorus flow through the activity to nourish, kg capita−1 year−1.

Figure 3.16   Nitrogen flow through the activity to nourish, kg capita−1 year−1

A key factor for nutrient losses in agriculture is consumer lifestyle. During the change from a resource-scarce society to an affluent society, the dietary tradition usually changes from low meat consumption to a diet that is rich in animal protein. The production of meat and poultry requires a much larger nutrient turnover than cereals and vegetables. Hence, the shift in dietary habits causes an increase in nutrient losses, too.

Losses of nutrients in industrial processing and distribution are much smaller than in agriculture, similar to those in households. The main difference between industrial processing/distribution and households is the number of sources: There are about 1000 times more households. Thus, from a reuse point of view, it is much more efficient to collect and recycle wastes from industrial sources than from consumers. The results summarized in Table 3.21 demonstrate clearly the limited contribution of individual housholds to the overall nutrient flows.

Table 3.21   Partitioning of Food-Derived P and N in Private Households

P, g capita−1 year−1

P,a %

P,b %

N, g capita−1 year−1

N,a %

N,b %

Food input

430

100

8.6

3700

100

20.5

Output MSW

40

9

0.8

300

8

1.7

Kitchen wastewater

20

5

0.4

200

6

1.1

Respiration

0

0

0.0

110

3

0.6

Urine

270

63

5.4

2600

70

14.4

Feces

100

23

2.0

490

13

2.7

Total food-related output

430

100

8.6

3700

100

20.5

Source: With kind permission from Springer Science+Business Media: Metabolism of the Anthroposphere (1st Edition), 1991, Baccini, P., and Brunner, P. H.

Notes:

a  Percent of food nutrient input into household.

b  Percent of total nutrient import into activity to nourish from Figures 3.15 and 3.16.

If all food-derived nutrients from households are recycled, less than 10% of P and about 20% of N requirements of agriculture can be satisfied. Table 3.21 also shows the contribution of each household output to nutrient conservation, thus serving as a basis for decisions regarding nutrient conservation and waste management. Composting of MSW is an inefficient measure to recycle nutrients. If all MSW were turned into compost, the contribution to agriculture would only be 1–2%. The fraction of nutrients in wastewater from households is about 10 times larger. Thus, the priority in nutrient recycling should be on wastewater and not on solid waste.

Another interesting fact is revealed by MFA and presented in Table 3.21: the amount of nutrients in urine is three (P) to five (N) times larger than in feces. This opens up new possibilities. Separate collection of urine could allow more than half of all nutrients entering a household to be accumulated in a relatively pure, concentrated, and homogeneous form. Several concepts have been proposed to manage this so-called anthropogenic nutrient solution (ANS) (Larsen et al., 2001). They are all based on a new type of toilet that is designed to collect urine separately from feces. The sewer system would be used after midnight to collect ANS stored in households during the daytime, thus permitting specific treatment and recycling of N, P, and K. Or ANS could be stored in households for longer time periods and collected separately with mobile collection systems. In any case, in order to prepare a fertilizer of high value, hazards such as endocrine substances and pharmaceuticals would have to be removed before ANS could be applied in agriculture.

Note that MFA of the activity to nourish is the basis for identifying the relevant nutrient flows and for developing alternative scenarios. In order to test the feasibility of the scenarios, technological, economic, and social aspects have to be investigated. New ways of managing urine and feces will only be successful if the same or greater convenience for the consumer is guaranteed.

3.2.2  Case Study 6: Copper Management

In Chapter 1, Section 1.4.5.1, it has been documented that modern economies are characterized by unprecedented material growth. Consumption of metals has increased while metal prices have decreased due to more efficient mining and refining technologies (Metallgesellschaft Aktiengesellschaft, 1993). Up to 80–90% of all resources consumed by mankind have been used in the second half of the twentieth century (Figure 3.17).

Within the anthropogenic metabolism, heavy metals are comparatively unimportant from a mass point of view, since they represent less than 10% of all inorganic goods (excluding water) consumed (Baccini and Bader, 1996). However, heavy metals play an important role in the production and manufacture of many goods. They can improve the quality and function of goods and are often crucial in extending the lifetime and range of application of goods. Their importance is based on their specific chemical and physical properties, e.g., corrosion resistance, electrical conductivity, ductility, strength, heat conductivity, brightness, etc.

Figure 3.17   World use of selected resources, t/year. About 80–90% have been used since 1950. (From Kelly, T. D., and Matos, G. R., Historical statistics for mineral and material commodities in the United States (2016 version): U.S. Geological Survey Data Series 140, 2014. Retrieved from http://minerals.usgs.gov/minerals/pubs/historical-statistics/.)

In 1972, the Club of Rome was among the first to point out the scarcity of resources in the book The Limits to Growth (Meadows, Meadows, Randers, and Behrens, 1972). Meadows et al. predicted that resources such as copper will be depleted within a short time of only a few decades. Prognoses about the depletion time (the number of years left until a resource is exhausted) of metals have been constantly revised and extended as a result of newly found reserves and advanced exploitation technologies. For certain metals essential for modern technology—lead, zinc, copper, molybdenum, manganese, etc.—some authors expect shortages within the next several decades (Kesler, 1994). There is controversy about whether this limitation will restrict future growth (for more information, see Becker-Boost and Fiala, 2001). Up to now, some but not all functions of metals can be mimicked by other materials.

Current metal management cannot be considered sustainable. During and after use, large fractions of metals are lost as emissions and wastes. Consequently, in many areas, concentrations of metals in soils as well as in surface water and groundwater are increasing. As discussed in Chapter 1, Section 1.4.5.2, human-induced flows of many metals surpass natural flows. Figure 1.7 displays the example of cadmium (Baccini and Brunner, 2012). While geogenic processes mobilize roughly 5.4 kt/year of cadmium, human activities extract about 17 kt/year from the Earth’s crust. Comparatively large anthropogenic emissions into the atmosphere are causing a significant accumulation of cadmium in the soil. Global emissions of cadmium should be reduced by an order of magnitude to achieve similar deposition rates as those determined for natural cadmium deposition. On a regional basis, the reduction goal should be even higher. Since most of the anthropogenic activities are concentrated in the Northern Hemisphere, the cadmium flows in this region have to be reduced further in order to protect the environment properly. The stock of anthropogenic cadmium grows by 3–4% per year. It needs to be managed, disposed of, and recycled carefully in order to avoid short- and long-term environmental impacts.

Heavy metals are limited valuable resources, but they are also potential environmental pollutants. New strategies and methods are needed for the management of heavy metals. A first prerequisite for efficient resource management is appropriate information about the use, location, and fate of these substances in the anthroposphere (Landner and Lindeström, 1999). Based on such information, measures to control heavy metals in view of resource optimization and environmental protection have to be designed. This case study discusses sustainable management of copper using information about copper flows and stocks in Europe as determined by Spatari and colleagues (2002).

3.2.2.1  Procedures

The copper household is evaluated by statistical entropy analysis (SEA). In Chapter 2, Section 2.5.9, the SEA method was introduced for single-process systems. In this case study, a system consisting of multiple processes is analyzed, requiring additional definitions and procedures. SEA can be directly applied to copper databases with no further data collection and little computational effort. The procedure has been developed and described by Rechberger and Graedel (1999).

3.2.2.1.1  Terms and Definitions

A set of material flows consists of a finite number of material flows. The distribution of a substance represents the partitioning of a substance among a defined set of materials. The distribution (or distribution pattern) is described by any two of the three properties Mi, Xi, ci for all materials of the set (see Figure 3.18).

3.2.2.1.2  Calculations

The following equations are used to calculate the statistical entropy H of a set of solid materials. If gaseous and aqueous flows (emissions) are also to be considered, more complex equations such as given in Chapter 2, Section 2.5.9.3, have to be applied. The system analyzed in this section contains solid materials/copper flows only. The number of materials in the set is k, and the flow rates (mx,...,mk) and substance concentrations (c1,..., ck) are known.

Figure 3.18   (a) Exemplary set of six material flows. (b) Mass flows of the set, mass/time. (c) Concentrations of the substance in the material flows, mass/mass. (d) Distribution of the substance among material flows (fraction).

3.1 $X ˙ i = m ˙ i ⋅ c i$
3.2 $m ˜ i = m ˙ i ∑ i = 1 k X ˙ i$
3.3 $H ( c i , m ˜ i ) = − ∑ i = 1 k m ˜ i ⋅ c i ⋅ ld ( c i ) ≥ 0$

The concentrations in Equations 3.1 and 3.3 are expressed on a mass-per-mass basis in equivalent units (e.g., gsubstance/gproduct or kgsubstance/kgproduct, so that ci ≤ 1. If other units are used (e.g., %, mg/kg), Equation 3.3 must be replaced by a corresponding function (Rechberger, 1999). The variable $m ˜ i$

represents standardized mass fractions of a material set. If the ci and $m ˜ i$ are calculated as described, the extreme values for H are found for the following distributions (see Figure 3.19):

Figure 3.19   (a) A set of material flows representing the distribution of a substance defined by the couple (m˙i,ci). (b) If the substance is only contained in one material flow, the statistical entropy H is 0. If the substance is equally distributed among the material flows, H reaches the maximum. (c) Any other distribution yields an H value between 0 and max. (From Rechberger, H. and Graedel, T. E., Ecol. Econ, 42, 59, 2002. With permission.)

1. The substance is only contained in one of the k material flows (i = b) and appears in pure form Σi = ẋb = ṁb. Such a material set represents the substance in its highest possible concentration. The statistical entropy H of such a distribution is 0, which is also a minimum, since H is a positive definite function for ci ≤ 1 (Figure 3.19b).
2. The other extreme is when all material flows have the same concentration (c1 = c2 = ... = ck). Such a material set represents the substance in its highest possible diluted form. For such a distribution, the statistical entropy is a maximum. Any other possible distribution produces an H value between these extremes (Figure 3.19c).

The maximum of H is expressed as

3.4 $H max ⁡ = ld ( ∑ i = 1 k m ˜ i )$

Finally, the relative statistical entropy (RSE) is defined as

3.5 $R S E ≡ H / H max ⁡$

A material flow system usually comprises several processes that are often organized in process chains. Figure 3.20 displays such a system comprising four processes (P) linked by 10 material flows (F), including one loop (recycling flow F9).

The procedure for evaluating a system by SEA depends on the structure of the system. For the system investigated in this chapter, the statistical entropy development can be calculated as described in the following two sections.

Figure 3.20   (a) Basic structure of a system made up of a process chain including one recycling loop. (b) Allocation of the system’s material flows to five stages. For example, stage 3 is represented and defined by flows F2, F5, F6, and F4. Stages 2 to 5 represent the transformations of the input (stage 1) caused by processes 1 to 4. (c) The partitioning of the investigated substance in each stage corresponds to a relative statistical entropy (RSE) value between maximal concentration (0) and maximal dilution (1). (From Rechberger, H. and Graedel, T. E., Ecol. Econ, 42, 59, 2002. With permission.)

3.2.2.1.2.1  Determination of Number and Formation of Stages

If the number of processes in the system is nP, then the number of stages is nS = nP + 1, where the stage index j = 1, 2, …., nS. The system as a whole can be seen as a process that transfers the input step by step, with each step designated as a stage. Stages are represented by a set of material flows (see Figure 3.20b). The first stage is defined by the input into the first process of the process chain. The following stages are defined by the outputs of processes 1 to nP. So stage j (j > 1) receives (1) the outputs of process j – 1 and (2) all outputs of preceding processes that are not transformed by the system (export flows and flows into a stock). Flows out of a stock are treated as input flows into the process. Flows into a stock are regarded as output flows of the process (see process P3, flow F7 in Figure 3.20a). This means that the stock is actually treated as an independent external process. However, for the sake of clarity, stocks are presented as smaller boxes within process boxes (see Chapter 2, Figure 2.1). Finally, recycling flows are treated as export flows. The allocation of material flows to stages is displayed in Figure 3.20b. The diagram shows how substance flows through the system become increasingly branched from stage to stage, resulting in different distribution patterns of substances.

3.2.2.1.2.2  Modification of Basic Data and Calculation of RSE for Each Stage

The basic data, flow rates of materials, and substance concentrations (i, ci) of the investigated system are determined by MFA. Normalized mass fractions $m ˜ i$

are derived using Equations 3.1 and 3.2. Application of Equation 3.3 to the couple $( c i , m ˜ i )$ or to each stage yields the statistical entropy H for that stage. Hmax is a function of the total normalized mass flow represented by a stage (see Equation 3.4). This normalized mass flow grows with subsequent stages if the concentrations of the materials decrease, since Σci × ṁ1 = 1 (combine Equations 3.1 and 3.2). One can assume maximum entropy when materials of a stage have the same concentration as the Earth’s crust (cEC) for the substance under study. Hmax is then given by

3.6 $H max ⁡ = ld ( 1 C E C )$

The reason for this definition of Hmax is related to resource utilization. If, for example, copper is used to produce a good that has a copper concentration of 0.06 g/kg (the average copper content of the Earth’s crust (Krauskopf, 1967), this product has the same resource potential for copper as the average crustal rock. Thus, a stage with entropy H = Hmax defines a point at which enhanced copper resources no longer exist. Using Equations 3.5 and 3.6, the RSE for each stage can be calculated. Figure 3.20c demonstrates that a system as a whole can be either concentrating, “neutral” (balanced), or diluting, depending on whether the RSE for the final stage is lower than, equal to, or higher than for the first stage.

3.2.2.1.3  Copper Data and Copper System of Study

Figure 3.21 illustrates copper flows and stocks in Europe in 1994, developed as part of a comprehensive project carried out at the Center for Industrial Ecology, Yale University. For a discussion of the quality, accuracy, and reliability of the data, see Graedel et al. (2002) and Spatari et al. (2002). Evaluating copper management practices on the basis of these data poses a challenge. At present, Europe is an open system for copper and depends heavily on imports. The total copper import (2000 kt/year) is more than three times higher than the domestic copper production from ore (≈590 kt/year; ore minus tailings and slag). Large amounts of production residues result from the use of copper, but with the present system boundaries, they are located outside of the system and therefore are not considered in an evaluation of European copper management. For a true evaluation, exports of goods containing copper and imports such as old scrap have to be taken into account, too. Thus, it is necessary to define a virtual autonomous system that (1) is independent of import and export of copper products and wastes and (2) incorporates all external flows into the system. Hence, in this virtual system, the copper necessary to support domestic demand is produced entirely within the system, depleting resources and producing residues. The estimated data for this supply-independent scenario are given in parentheses in Figure 3.21, which represents a closed system that includes all material flows relevant for today’s copper management.

Figure 3.21   Copper flows and stocks for Europe in 1994 (values rounded, kt/year). The values given in parentheses represent a virtual and autonomous copper system with the same consumption level but no copper imports and exports. (From Rechberger, H. and Graedel, T. E., Ecol. Econ., 42, 59, 2002. With permission.)

Table 3.22 gives the data that are used to calculate the entropy trends. The flow-rates for copper (Xi) are from Spatari et al. (2002). The concentrations for copper (ci) and their ranges are either from literature references or best estimates. The ranges provide the basis for assessing the uncertainty of the final entropy trends. The flow rates for materials (i) are calculated using Equation 3.7:

3.7 $m ˙ i = X ˙ i / c i × 100$

Table 3.22   Data on Material Flows of European Copper Management

Material

Material Flow (i), kt/Year

Copper Concentration (ci), g/100 g

Copper Flow ( i),* kt/Year

Ore

69,000

1 (0.3–3)

690

Concentrate

930

25 (20–35)

280

Blister

205

98 (96–99)

200

Cathode I

2200

100

2200

Flow out of stock (production)

290

100

290

Cathode II

1300

100

1300

Tailings

90,000

0.1 (0.1–0.75a)

90

Slag

1700

0.7 (0.3b–0.7)

12

New scrap

260

90 (80–99)c

230

Old scrap I

680

80 (20–99)

540

Old scrap II

250

80 (20–99)

200

Old scrap III

380

80 (20–99)

300

Semialloy and finished products

110

70 (7–80)c

80

Products (pure Cu)

27,000

10 (1–50)c

2700

Products (Cu alloy)

11,000

7 (1–40)c

800

Flow into stock (use)

1,200,000

0.2 (0.1–0.3)c

2600

Wastes

460,000

0.2 (0.1–0.3)c

920

Landfilled wastes

460,000

0.10d

480

Source: Rechberger, H. and Graedel, T. E., Ecological Economics, 42, 59, 2002. With permission. Data from DKI Deutsches Kupferinstitut, Kupfer, Vorkommen, Gewinnung, Eigenschaften, Verarbeitung, Verwendung Informationsdruck. Duesseldorf: Deutsches Kupferinstitut, 1997; Zeltner, C. et al., Regional Environmental Change, 1 (1), 31–46, 1999; Gordon, R. B., Resources, Conservation and Recycling, 36 (2), 87–106, 2002.

Note: Values are rounded.

Notes:

a  Higher value for period around 1900.

b  Lower value for period around 1925.

c  Informed estimate.

d  Calculated by mass balance on waste management process.

*  Spatari, Bertram, Fuse, Graedel, and Rechberger (2002).

3.2.2.2.1  RSE of Copper Management and of Alternative SystemsStatus Quo and Virtual Supply: Independent Europe

The entropy trends are calculated using Equations 3.3 to 3.7, the data given in Table 3.22, and the appropriate flowcharts. Figure 3.22 shows the trend of the RSE along the life cycle of copper for two systems: (1) the status quo of 1994 and (2) the supply-independent Europe (both displayed in Figure 3.21). The assignment of material flows to stages is illustrated in Figure 3.23.

Both systems behave similarly, with the production process reducing the RSE from stage 1 to 2, since ore (copper content = 1 g/100 g) is refined to plain copper (content > 99.9 g/100 g). Note that the RSE for stage 2 is not 0, since mining ores and the smelting concentrates produce residues (tailings and slag). The more efficient a production process is (efficiency being measured by its ability to transform copper-containing material), the more closely the RSE of stage 2 approaches 0, meaning that the total amount of copper appears in increasingly purer form. Note: For the reduction of the RSE from stage 1 to 2, external energy (crushing ores, smelting concentrate, etc.) is required. The impact on the RSE induced by this energy supply is not considered within the system, since the energy supply is outside the system boundary. Whether or not the exclusion of the energy source has an impact on the RSE development depends on the kind of energy source (coal, oil, and hydropower) used. However, in this chapter, the system boundaries are drawn as described by Spatari et al. (2002).

Figure 3.22   Change of the relative statistical entropy along the life cycle of copper for the status quo in Europe in 1994 (open system) and for a virtual, supply-independent Europe (closed or autonomous system). The shapes of the trends are identical, but the overall performances (differences between stages 1 and 5) of the systems are different. (From Rechberger, H. and Graedel, T. E., Ecol. Econ, 42, 59, 2002. With permission.)

Figure 3.23   Assignment of material flows to stages. (a) Status quo, (b) supply-independent system, (c) supply-independent system without recycling, and (d) supply-independent system in steady state. (From Rechberger, H. and Graedel, T. E., Ecol. Econ, 42, 59, 2002. With permission.)

Producing semiproducts and consumer goods from refined copper increases the RSE from stage 2 to 3 because of the dilution of copper that occurs in manufacturing processes. It is obvious that dilution takes place when copper alloys are produced. Similarly, installing copper products into consumer goods (e.g., wiring in an automobile) or incorporating copper goods into the built infrastructure (transition from stage 3 to 4, e.g., copper tubing for heating systems) “dilutes” copper as well. In general, the degree of dilution of copper in this stage is not well known. Information about location, concentration, and specification is a sine qua non condition for future management and optimization of copper. For a first hypothesis, it is sufficient to assume that the mean concentration of copper in the stock is the same as the mean concentration in the residues that leave the stock. This concentration level can be determined from copper concentrations and relevant waste generation rates such as municipal solid waste, construction and demolition debris, scrap metal, electrical and electronic wastes, end-of-life vehicles, etc. (Bertram, Rechberger, Spatari, and Graedel, 2002). During the transition from stage 4 to 5, the entropy decreases, since waste collection and treatment separate copper from the waste stream and concentrate it for recycling purposes. The “V” shape of the entropy trend—the result of entropy reduction in the production (refining) process and entropy increase in the consumption process (see Figure 3.22)—was described qualitatively, e.g., by O’Rourke, Connelly, and Koshland (1996), Ayres and Nair (1984), Stumm and Davis (1974), and Georgescu-Roegen (1971).

The differences in the entropy trends between the status quo and the supply-independent system are noteworthy. First of all, the status quo system starts at a lower entropy level, since concentrated copper is imported in goods. The differences in stage 2 are due to the increased ore production in the supply-independent system, resulting in larger amounts of production residues, which are accounted for in stage 2. In stages 3 and 4, the difference between the status quo and the supply-independent system remains rather constant, since the metabolism for both scenarios does not differ significantly in these stages. The effectiveness of waste management is lower in the supply-independent system, as there is no old scrap imported and the recycling rate is therefore lower. In the following, only the supply-independent system and some variations of it are discussed, since it comprises all processes and flows relevant for European copper management and includes external effects within Europe’s hinterland.

The overall performance of a system can be quantified by the difference between the RSEs for the first and the final stages. In this case,

3.8 $Δ R S E t o t a l = Δ R S E 15 = [ ( Δ R S E 5 − R S E 1 ) / R S E 1 ] × 100$

where ∆RSEtotal > 0 means that the investigated substance is diluted and/ or dissipated during its transit through the system. From a resource conservation and environmental protection point of view, such an increase is a drawback. If maintained indefinitely, such management practice will result in long-term problems. In contrast, scenarios with high recycling rates, advanced waste management, and nondissipative metal use show decreasing RSE trends (RSEtotal ≤ 0%). Low entropy values at the end of the life cycle mean that (1) only small amounts of the resource have been converted to low concentrations of copper in products (e.g., as an additive in paint) or dissipated (in the case where emissions are considered) and (2) large parts of the resource appear in concentrated (e.g., copper in brass) or even pure form (e.g., copper pipes). Wastes that are disposed of in landfills should preferably have Earth-crust characteristics or should be transformed into such quality before landfilling (Baccini, 1989). Earth-crust-like materials are in equilibrium with the environment, and their exergy approaches 0 (Ayres and Martinas, 1994; Ruth, 1995; Ayres, 1998). Thus, waste management systems must produce (1) highly concentrated products with high exergy that are not in equilibrium with the surrounding environment and (2) residues with Earth-crust-like quality. Low- or zero-exergy wastes can easily be produced by dilution, e.g., by emitting large amounts of off-gases with small concentrations in high stacks, or by mixing hazardous wastes with cement, thus impeding future recycling of the resource. A low RSE value for a stage thus means that both highly concentrated (high-exergy) and low-contamination (low-exergy) products are generated.

3.2.2.2.2  Recycling in Supply-Independent Europe

The relevance of recycling on the entropy trend is investigated using Figure 3.24. Numbers in parentheses show the supply-independent system without any recycling of old and new scrap. Compared with the supply-independent scenario displayed in Figure 3.21, this results in a higher demand for ore (+63%) and larger requirements for landfills for production and consumption wastes (+220%).

The entropy trend for the nonrecycling scenario is given in Figure 3.25. All RSEs are higher, showing the effects of not recycling production residues (new scrap) in stages 2 and 3 and the zero contribution of waste management in stage 5. The resulting ΔRSE15 = +28% indicates a bad management strategy. At present, the overall recycling rate for old scrap is about 40%. Some countries within the European Union achieve rates up to 60% (Bertram, Rechberger, Spatari, and Graedel, 2002). Assuming that in the future, all countries will achieve this high rate, ΔRSE15 would be reduced from −1% to -4% (recycling rate of 90%: ΔRSE15 = −11%) for the supply-independent system. This shows that the impact of today’s waste management on the overall performance of the system is limited. The reason is that the copper flow entering waste management is comparatively small.

Figure 3.24   Copper flows and stocks of a supply-independent Europe with no accumulation of copper in the process use, kt/year (steady-state scenario). Values in parentheses stand for a scenario without copper recycling. (From Rechberger, H. and Graedel, T. E., Ecol. Econ, 42, 59, 2002. With permission.)

Figure 3.25   Comparison of the effect of different scenarios on the relative statistical entropy along the life cycle of copper: scenario of supply-independent Europe versus scenario of no recycling and scenario of steady state producing no stocks. The assessment shows that waste management and recycling can play a crucial role in future resource use. (From Rechberger, H. and Graedel, T. E., Ecol. Econ., 42, 59, 2002. With permission.)

3.2.2.2.3  Supply-Independent Europe in Steady State

Figure 3.24 also gives the flows for a steady-state scenario in which the demand for consumer goods is still the same as in the status quo, but the output equals the input of the stock in the process use. This scenario may occur in the future when, due to the limited lifetime, large amounts of materials turn into wastes (Brunner and Rechberger, 2001). Assuming a recycling rate of 60% results in ΔRSE15 = −47%. This shows that in the future, waste management will be decisive for the overall management of copper. A recycling rate of 90% will result in ΔRSE15 = −77%. Such a high recycling rate cannot be achieved with today’s design of goods and systems. Also, better information bases on the whereabouts of copper flows and, especially, stocks are needed. If the design process is improved, if necessary information is provided, and if advanced waste management technology is employed, future management of copper can result in declining RSE rates, contributing to sustainable metal management.

3.2.2.2.4  Uncertainty and Sensitivity

The uncertainty of the data (material flow rates i and substance concentrations ci) and the accuracy of the results are fundamental pieces of information for the evaluation process. In most cases, data availability constrains the application of statistical tools to describe materials management systems. Statistics on material flows do not customarily provide information on reliability and uncertainty, such as a standard deviation or confidence interval. Sometimes, substance concentration ranges can be determined by a literature survey. In Figure 3.26, upper and lower limits of the RSE are presented for the supply-independent scenario.

Figure 3.26   Variance of relative statistical entropy based on estimated ranges of basic data. (From Rechberger, H. and Graedel, T. E., Ecol. Econ., 42, 59, 2002. With permission.)

These limits are calculated using the estimated ranges for copper concentrations as given in Table 3.22. Thus, the limits are not statistically derived but estimated. Since the ranges in Table 3.22 have been chosen deliberately to be broad, the possibility that the actual RSE trend lies within these limits is high. This is despite the fact that the uncertainty in the material flow rates is not considered. The range for ΔRSE15 lies between −23% and 28% (mean, −1%), sufficient for a first assessment. The uncertainties for the different stages vary considerably. The range for stage 1 is due to the range of the copper content in ores (0.5–2%). The range for stage 2 is quite small, meaning that the RSE for this stage is determined with good accuracy. The largest uncertainty is found for stage 3, since the average copper concentrations of many goods are poorly known. The uncertainties for stages 4 and 5 are lower, with a range similar to that for stage 1.

The result emphasizes the hypothesis that the stock in use has the potential to serve as a future resource for copper. Both stages 4 and 5 show the same entropy level for 1 t of copper. When calculating the RSE, the stock is characterized by the estimated average concentration of copper in the stock, meaning that the copper is evenly distributed and maximally diluted in the stock. This can be regarded as a worst-case assumption. Having more information about the actual distribution of copper in the stock would result in lower RSE values for stage 4. Provided that this information can be used for the design and optimization of waste management, the high recycling rates necessary to achieve ΔRSE15 < −70% should be feasible.

3.2.2.3  Conclusions

Contemporary copper management is characterized by changes in the distribution pattern of copper, covering about 50% of the range between complete dilution and complete concentration. Copper flows and stocks through the (extended) European economy are more or less balanced due to recycling of new and old scrap and the small fraction of dissipative use of copper in goods. It is confirmed that the stock of copper currently in use has the potential for a future secondary resource. This can be even further improved by appropriate design for recycling of copper-containing goods. Provided that waste management is adapted to recycle and treat the large amounts of residues resulting from the aging stock, copper can be managed in a nearly sustainable way. Thus, this case study exemplifies how nonrenewable resources can be managed in order to conserve resources and protect the environment.

3.2.3  Case Study 7: Construction Waste Management

Construction materials are important materials for the anthropogenic metabolism. They are the matrix materials for the structure of buildings, roads, and networks and represent the largest anthropogenic turnover of solid materials (see Table 3.23). They have a long residence time in the anthroposphere and thus are a legacy for future generations. On one hand, they are a resource for future use; on the other hand, they can be a source of future emissions and environmental loadings. An example of reuse would be recycling of road surface materials, which is widely practiced in many countries. Examples of emissions are polychlorinated biphenyls (PCBs) in joint fillers and paints, and chlorinated and fluorinated carbohydrates (CFCs) in insulation materials and foams. Hence, construction materials have to be managed with care in view of both resource conservation and environmental protection. A main future task will be to design constructions in a way that allows the separation of construction materials after the lifetime of a building, with the main fraction being reused for new construction, leaving only a small fraction for disposal via incineration in landfills. (Incineration will be necessary to mineralize and concentrate hazardous substances that are required to ensure long residence times of, for example, plastic materials.)

Table 3.23   Per Capita Use of Construction Materials in Vienna from 1880 to 2000

Per Capita Use of Construction Material m3 Capita−1 Year−1

1880–1890

0.8

1890–1900

0.4

1900–1910

0.1

1910–1920

0.1

1920–1930

0.1

1930–1940

1.4

1940–1950

0.1

1950–1960

0.1

1960–1970

0.1

1970–1980

2.4

1980–1990

3.3

1990–2000

4.3

Source: Fischer, T. (1999). Zur Untersuchung verschiedener methodischer Ansätze zur Bestimmung entnommener mineralischer Baurohstoffmengen am Beispiel des Aufbaus von Wien (Diploma Thesis). Technische Universität Wien, 1999.

In this case study, construction materials are discussed in view of resource conservation. Both volume and mass are considered as resources. The purpose is twofold. First, it is shown that MFA can be used to address volume-related resource problems, too. Also, some of the difficulties of bringing construction wastes back into a consumption cycle are explained. Second, two technologies for producing recycling materials from construction wastes are compared by means of MFA.

3.2.3.1  The “Hole” Problem

Excavation of construction materials from a quarry or mine usually results in a hole in the ground. Since construction materials are used to create buildings with residence times of several decades, it takes some 30 to 50 years before these holes can be filled up with construction debris. In a growing economy, the input of construction materials into the anthroposphere at a given time is much larger than the output. Thus, as long as the building stock of a city expands, the volume of holes in the vicinity of the city expands as well.

In Figures 3.27 and 3.28, the total and per capita use of construction materials in Vienna is given for the time span from 1880 to 2000. The extraction of construction materials varies much from decade to decade. The effect of an economic crisis, such as the Great Depression of the 1930s and the postwar periods, on construction activities is evident. If accumulated over the time period of 120 years, a total hole of 207 million m3 results (Figure 3.28). This corresponds to about 140 m3/capita for today’s population (1.5 million inhabitants).

It is interesting to note that the holes created by the needs of a prosperous, growing city of the 1990s are much larger than the volume of all wastes available for landfilling. In Figure 3.29, Lahner (1994) presents a construction material balance established for Austria. The input of construction materials exceeds the output of construction wastes by nearly an order of magnitude. Besides the hole problem discussed here, another important implication arises from input >> output: the amount of construction wastes available for recycling is small when compared with the total need for construction materials. Thus, even if all wastes were recycled, they would replace only a small fraction of primary materials. It may be difficult to create a market for a product with such a small market share, especially if there is uncertainty with respect to the quality of the new and as-yet-unknown material and if there is only a small advantage in price. For successful introduction of recycling materials, it is necessary to establish technical and environmental standards, to develop technologies that produce sufficiently high quality at a competitive price, and to persuade consumers of the usefulness and advantages of the new product.

Figure 3.27   Construction material excavated from the ground and built into Vienna from 1880 to 2000, m3 per decade. (From Fischer, T., Zur Untersuchung verschiedener methodischer Ansätze zur Bestimmung entnommener mineralischer Baurohstoffmengen am Beispiel des Aufbaus von Wien (Diploma Thesis). Technische Universität Wien, 1999.)

Figure 3.28   Cumulative “hole” volume in the vicinity of Vienna due to excavation of construction materials between 1880 and 2000. (From Fischer, T., Zur Untersuchung verschiedener methodischer Ansätze zur Bestimmung entnommener mineralischer Baurohstoffmengen am Beispiel des Aufbaus von Wien (Diploma Thesis). Technische Universität Wien, 1999.)

Figure 3.29   Materials used for construction in Austria (1995), kg capita−1 year−1. The input of construction materials into a growing economy is much larger than the output of construction wastes. (From Lahner, T., Müll Magazin, 7, 9, 1994. With permission.)

In the case of Vienna, the total waste (MSW, construction waste, etc.) generated annually for disposal during the 1990s was about 600,000 tons measuring 800,000 m3 (or 400 kg/capita at 0.53 m3/capita). Wastes that are recycled are not included in this figure. This is approximately eight times less than the annual consumption of construction materials (4.3 m3 capita−1 year−1). Thus, it is not possible to fill the holes of Vienna by landfilling all wastes. Note that the actual volume of wastes to be landfilled in Vienna is considerably smaller due to waste incineration, which reduces the volume of municipal wastes by a factor of 10.

Landfilling is usually not a problem from the point of view of quantity (volume or mass); rather, it is an issue of quality (substance concentrations). The wastes that are to be disposed of in landfills do not have the same composition as the original materials taken from these sites. Thus, the interaction of water, air, and microorganisms with the waste material is likely to differ from the original material, resulting in emissions that can pollute groundwater and the vicinity of the landfill. On the other hand, the native material has been interacting with the environment for geological time periods. Except for mining and ore areas, the substance flows from such native sites are usually small (“background flows and concentrations”) and not polluting.

The conclusion of the “hole balance” problem is as follows: Growing cities create holes; hence, “hole management” is important and necessary. These void spaces can be used for various purposes, such as for recreation or for waste disposal. If they are used as landfill space, qualitative aspects are of prime importance and have to be observed first. Wastes to be filled in such holes need to have stonelike properties. They require mineralization (e.g., incineration with after treatment), and they should be in equilibrium with water and the environment. The new objective of waste treatment thus becomes the production of immobile stones from waste materials.

3.2.3.2  MFA for Comparing Separation Technologies

Construction wastes are the largest fraction of all solid wastes. Thus, for resource conservation, it is important to collect, treat, and recycle these wastes. There are various technologies available to generate construction materials from construction wastes. Their purpose is to separate materials well suited as building materials from hazardous, polluting, or other materials inappropriate for construction purposes. MFA serves as a tool to evaluate the performance of construction waste sorting plants with regard to the composition of the products (e.g., production of clean fractions versus accumulation of pollutants in certain fractions).

In order to design and control construction waste recycling processes, it is necessary to know the composition of the input material that is to be treated in a sorting plant. The composition and quantity of construction wastes depend upon the “deconstruction” process. If a building is broken down by brute force of a bulldozer, the resulting waste is a mixture of all possible substances. If it is selectively dismantled, individual fractions can be collected that represent comparatively uniform materials such as wood, concrete, bricks, plastics, glass, and others. These fractions are better suited for recycling. After crushing, they can be used either for the production of new construction materials or as fuel in industrial boilers, power plants, or cement kilns. Both types of deconstruction yield at least one fraction of mixed construction wastes. While indiscriminate demolition results in mixed contruction wastes only, the mixed fraction obtained in selective dismantling is much smaller and comprises mainly nonrecyclables such as plastics, composite materials, and contaminated constituents.

Construction waste sorting plants are designed to handle mixed fractions. The objectives of sorting are twofold: First, sorting should result in clean, high-quality fractions suited for recycling. Second, sorting should yield non-recyclables that are ready for treatments such as incineration or landfilling. In Figures 3.30 and 3.31, two technologies for construction waste recycling are presented. They differ in the way they separate materials. Plant A (25 t/h) is a dry process, including handpicking of oversize materials, rotating drum for screening, crusher and pulverizer, zigzag air classifier, and dust filters. In plant B (60 t/h), the construction waste is similarly pretreated before it is divided into several fractions by a wet separator. In order to evaluate and compare the performance of the two processes with regard to resource conservation, both plants are investigated by MFA. The results serve as a basis for decisions regarding the choice of technologies for construction waste sorting.

3.2.3.3  Procedures

Since it is not possible to determine the chemical composition of untreated construction wastes by direct analysis, the input material into both plants is weighed only and not analyzed. The composition of the incoming waste is established by sampling and analyzing all products of sorting, and by calculating for each substance the sum of all output flows divided by the mass of construction wastes treated within the measuring period. This procedure is chosen because the sorting plants produce fractions that are more homogeneous in size and composition, and thus, they are easier and less costly to analyze than the original construction waste. The input into both plants is not the same, because the two collection systems that supply construction wastes to plants A and B are also different.

The method of investigation is described in by Schachermayer, Lahner, and Brunner (2000) and Brunner and Stämpfli (1993). Mass balances of inputand output goods are performed for time periods between 2 and 9 h. The wet process is analyzed in five short campaigns, the dry process in a comprehensive investigation of 9 h. Samples of all output goods are taken and analyzed for matrix substances (>1 g/kg) and trace substances (<1 g/kg) at hourly intervals. Off-gases and wastewater are sampled according to standard procedures for such materials. The size of solid samples is between 5 and 500 kg. Aliquots of the samples are crushed and pulverized until particles are smaller than 0.2 mm. Metal fractions such as magnetically separated iron are not crushed; their composition is roughly estimated according to the individual components present. Oversize materials of concrete and stones are not analyzed either. For concrete, literature values are taken; composition of stones is assumed to be the same as in other, smaller stone fractions. For fractions that cannot be analyzed due to the lack of pulverized samples, it is tested to see if the overall material balance is sensitive against these assumptions. The fractions not analyzed amount to less than 5% of the total construction waste treated. Since the matrix (bulk) compositions of these fractions are known (e.g., the magnetically separated fraction contains <80% iron), errors in the assumptions proved not to be decisive for the overall mass balance and the transfer coefficients.

Figure 3.30   Construction waste (CW) sorting plant A, dry process. Fraction A1, large pieces of concrete and stones; A2, metals; A3, oversize combustibles; B, <80 mm; C, concrete and stones; D, metals; E, oversize material; F, light fraction; G, heavy fraction; H1 and H2, dust from cyclones 1 and 2; I, scrap iron; K1, off-gas drum and shredder; K2, off-gas air classifier.● measurement of mass flow, m3/h and t/h; X, measurement of substance concentration, mg/kg.

Figure 3.31   Construction waste (CW) sorting plant B, wet process. Pretreatment 1: crusher and sieve, 0 to 100 cm; pretreatment 2: hand sorting, magnetic separator, pulverizer, and sieve, 0 to 32 cm; CW 0/100 and CW 0/32: construction waste crushed, pulverized, and sieved by a mesh size of 100 and 32 cm, respectively; wastewater includes settled sludge; LF: light fraction; F1 to F3: construction materials for recycling (F1, 16 to 32 mm; F2, 4 to 16 mm; F3, 0 to 4 mm); Fe, scrap iron; W/P, fraction containing wood and plastics. ●, measurement of mass flows, m3/h and t/h; X, measurement of substance concentrations, mg/kg.

3.2.3.4  Results

3.2.3.4.1 Composition of Construction Wastes

As expected, the composition of the construction wastes treated in plant A is not the same as in plant B (see Table 3.24). The material treated in plant A contains more sulfur (gypsum), organic carbon, and iron than the input into plant B, and the concentration of trace metals is about one order of magnitude higher. Construction wastes treated in plant A are more contaminated and contain less inorganic materials than the product for plant B. The reason for this difference has not been investigated. Possible explanations are as follows.

Table 3.24   Composition of Construction Wastes Treated in Dry-Separation Plant A and Wet-Separation Plant B Compared with the Average Composition of the Earth’s Crust

Substance

Construction Waste Plant A (Mixed Construction Wastes)

Construction Waste Plant Ba (Presorted Construction Wastes)

Earth’s Crust

Matrix Substances, g/kg

S

5.8

1.1–2.9

0.3

TC

93

47–79

0.2

TIC

33

35–69

TOC

60

2–21

Si

121

100–150

280

Ca

150

120–200

41

Al

9.5

8–15

81

Fe

40

7–20

54

Trace Elements, mg/kg

Zn

790

24–66

70

Pb

630

3–103

13

Cr

150

13–32

100

Cu

670

8–23

50

Cd

1.0

0.10–0.22

0.1

Hg

0.2

0.05–0.55

0.02

Note: TC, total carbon; TIC, total inorganic carbon; TOC, total organic carbon.

Notes:

a  Data for plant B are the result of four sampling campaigns with different input materials; thus, ranges are given.

1. Plant A is located in Switzerland and was analyzed in 1988, while the mass balance of plant B, operating in Austria, was conducted in 1996. During the time period of 8 years, construction waste management experienced swift development. In the 1980s, mixed construction wastes were treated in separation plants, while the 1990s saw a shift toward selective deconstruction and dismantling, resulting in cleaner and more uniform input fractions for such plants.
2. At the time of analysis, Switzerland and Austria had distinctly different legislation and practices in construction waste management. In Switzerland, no legislative framework had been established at the time of analysis. The MFA of the sorting plant is a first investigation into the power of such plants to produce appropriate secondary construction materials. The results are used to establish a new strategy, giving preference to selective deconstruction (see the results that follow). Eight years later in Austria, it was mandatory to separately collect uniform fractions such as wood, metals, plastic, concrete, etc. when a certain mass flow per construction site is exceeded. The cleaner input into plant B indicates that the decision made in Switzerland (selective deconstruction) is appropriate.
3. Most construction waste stems from demolition and not from new construction sites. Due to different economic cycles (Austria was at a low level of economic development after World War II and was slow in recovering), buildings demolished in Switzerland and Austria are of different time periods. Some of the Swiss construction waste resulted from comparatively new buildings that had been constructed only 20 to 40 years ago, while in Austria, the buildings demolished in the 1990s were older.

Thus, the composition of construction wastes in plant A may resemble the construction materials of the 1950s and 1960s, while for plant B, the input most likely stems from prewar periods (1930–1940) and hence has a different composition in trace substances.

Note that the three reasons stated here have not been investigated in detail; they are merely given as possible explanations for different compositions of construction materials. In order to derive significant results about differences in the composition of construction wastes, the analysis would have to be planned from a statistical point of view, which was not intended when the mass balance was conducted in plant A.

In summary, at the time of investigation, plant A was fed by mixed construction wastes as received when indiscriminately demolishing a building. Plant B received construction debris that resulted from more-or-less controlled dismantling and represented a fraction that looked well suited for recycling, where much of the unsuitable material had already been removed at the construction site.

3.2.3.4.2  Mass Flow of Products of Separation

The balance of plant A is given in Table 3.25. Dry separation generated 14 different products. Four products are wastes and have no further use (dust from cyclones 1 and 2 and off-gases from the drum, shredder, and air classifier). Some of the remaining 10 fractions are, in part, quite similar. Therefore, they have been rearranged into the five fractions I, II, III, metals, and rest seen at the bottom of Table 3.25.

• Major fractionsFraction I, pieces <80 mmFraction II, light materialsFraction III, heavy materials
• Minor fractionsScrap ironRest, consisting of useless residues (filter dust and off-gas)

The rationale for this new grouping will become apparent when the chemical compositions of the individual fractions are discussed.

The balance of goods for plant B is given in Table 3.26. A priori, this plant produces fewer fractions. Only two of the seven fractions generated are of major importance. The light fraction only amounts to 5.1 g/100 g, indicating again that the input into plant B contains less organic waste (plastic, paper, light wood, and the like) than plant A. In contrast to plant A, plant B produces a significant amount of fine-grain material of<4 mm particle size. The operator of plant B finds a good market for this material, while plant A’s customers are asking for coarser materials. Note that due to waste separation on the construction site, the percentage of the scrap-iron fraction is 20 times smaller for plant B than A.

Table 3.25   Mass Flow through Construction Waste (CW) Sorting Plant A

Material

Consisting of

Mass Flow, 103 kg/Day

Fraction, g/100 g CW

Total input

Construction wastes

225.3

100

Fraction

A1

Concrete, stones

8.5

3.8

A2

Metals

3.08

1.3

A3

Oversize combustibles

3.75

1.7

B

<80 mm

102

45.3

C

Concrete, stones

4.14

1.8

D

Metals

2.36

1.0

E

Oversize material

0.43

0.2

F

Light fraction

51.4

22.8

G

Heavy fraction

47.7

21.2

H1

Dust cyclone 1

0.16

0.06

H2

Dust cyclone 2

0.10

0.04

I

Iron metals

1.73

0.8

K1

Off-gas drum/shredder

n.d.

n.d.

K2

Off-gas air classifier

n.d.

n.d.

New fractions

I + II + III + metals + rest

225.3

100

I (= B)

<80 mm

102

45.3

II (= F + E + A3)

Light fraction

55.6

24.7

III (= G + C + A1)

Heavy fraction

60.3

26.8

Metals (= A2 + D + E)

Iron

7.13

3.1

Rest (= H + K)

Dust and off-gases

0.26

0.1

Note: n.d., not determined.

Table 3.26   Mass Flow through Construction Waste (CW) Sorting Plant B (Presorted Construction Waste)

Material

Consisting of

Mass Flow, 103 kg/h

Fraction, g/100 g CW

Total input

370–380

≈500

Presorted CW

75

100

Water

300

400

Total output

200–270

260–360

Wastewater

130–190

170–250

(Wastewater sediment)a

(Wastewater sludge from pond)

(2.5–3.7)

(3.3–4.9)

LF

Light fraction

3.8

5.1

F1

Sorting fraction 16–32 mm

15

20

F2

Sorting fraction 4–16 mm

27

36

F3

Sorting fraction 0–4 mm

25

33

Fe

Scrap iron

0.13

0.17

W/P

Wood and plastic fraction

0.05

0.07

Note: The difference between input and output is due to the loss of water when the drenched fractions leave the wet process and are stored and dewatered on site without measuring water losses. It is not possible to quantify this difference.

Notes:

a  Wastewater sediment is included in wastewater and is generated in a process outside the system’s boundary (sedimentation in a wastewater sludge pond).

3.2.3.4.3  Composition of Products of Separation

The compositions of the products of the two construction waste recycling plants are presented in Table 3.27. In both plants, fractions rich in carbonates and silicates and poor in organic carbon are produced. Also, both plants produce light fractions containing approximately 20% of total organic carbon (TOC) and scrap-iron fractions. The difference in chemical composition of the products obtained in the two plants is mainly due to the difference of input materials.

Because of the given input, all fractions of dry separation in plant A exceed concentrations of heavy metals in the Earth’s crust. Since construction waste treated in plant B is considerably cleaner, the compositions of the wet products come closer to Earth-crust quality. Nevertheless, concentrations of lead and mercury are above that of the Earth’s crust for all fractions analyzed in plant B, too. The fraction most polluted is light fraction II from dry separation. The material is similar to MSW and exhibits a high content of organic carbon (20%). Thus, this fraction is not suited for recycling as a construction material. Instead, it can be utilized to recover energy from waste in an incinerator equipped with sophisticated air-pollution devices to remove acid gases, particulates, and volatile heavy metals like mercury and cadmium.

Table 3.27   Composition of Products from Dry (A) and Wet (B) Construction Waste Separation

Products of Plant A

Products of Plant B

Substance

I

II

III

G

Iron Metals

F1

F2

F3

LF

Wastewater Sludge

Earth’s Crust

Matrix Elements, g/kg

Si

160

n.d.

180

170

n.d.

170 ± 10

170 ± 16

190 ± 13

170 ±8

170

280

Ca

180

91

160

160

n.d.

160 ±9

160 ± 19

140 ± 18

100 ± 17

160 ± 18

41

Fe

12

16

20

22

800

15 ±5

16 ±6

16 ±5

20 ±5

20 ±3

54

TC

62

210

48

47

n.d.

54 ±4

59 ±6

59 ±6

210 ± 90

98 ±23

0.2

TIC

41

17

38

34

n.d.

53 ±5

52 ±10

47 ±8

22 ±8

47 ±6

-

TOC

21

190

9.9

12

n.d.

1.8 ± 1

7±6

11 ±3

190 ± 95

51 ±25

-

Al

8.8

8.3

12

12

8.1

15 ±4

15 ±5

11 ±3

21 ±6

20 ±3

8.1

S

7.3

5.7

3.9

4.3

n.d.

1.6 ± 0.54

1.3 ± 0.2

1.4 ± 0.2

3.8 ± 0.4

2.4 ± 0.5

0.3

Trace Elements, mg/kg

Zn

540

1400

170

200

4900

35 ±8

34 ±8

48 ±5

65 ±9

200 ± 91

70

Cu

47

420

330

410

11,500

16 ±3

21 ±6

22 ±6

30 ± 7

45 ±4

50

Pb

200

940

930

1200

1800

30 ±54

16 ±15

25 ±10

46 ±37

75 ±11

13

Cr

160

90

130

140

760

24 ±3

25 ±9

25 ±10

110 ± 22

41 ±7

100

Cd

0.7

2.3

0.5

0.6

n.d.

0.12 ± 0.01

0.11 ± 0.005

0.13 ± 0.01

0.2 ± 0.07

0.31 ± 0.08

0.1

Hg

0.2

0.3

0.1

0.1

n.d.

0.11 ± 0.07

0.17 ± 0.08

0.47 ± 0.31

0.7 ± 0.03

3.1 ± 1,7

0.02

Note: n.d., not determined.

The light fraction from plant B is similar to the one from plant A. The main differences are that trace-metal concentrations are smaller in B and that the amount of light fraction that the wet plant B produced per unit of construction waste (5.1 g/100 g CW) is about five times smaller than for the dry plant A (24.7 g/100 g CW). Both differences are due to differences in the input materials for the two plants. Plant B produces a large amount of wastewater containing suspended solids. Most of this wastewater is treated in a sedimentation pond, where a sludge (sediment) is formed and deposited. Contaminant concentration of this sludge is higher than in any other product of plant B, confirming the hypothesis that a lot of heavy metals are present on small particles that are removed and transferred to the water phase during wet separation. A significant amount of less contaminated wastewater is not controlled and is “lost” on the site (the plant stands on a river bank).

3.2.3.4.4  Partitioning of Metals and Transfer Coefficients

The main purpose of construction waste sorting is to produce clean secondary construction materials. In chemical terms, sorting must direct hazardous substances contained in construction wastes to those fractions that are not intended for reuse. Preferably, the resulting substance concentration in recycling fractions is close to the concentration of materials used for the primary production of construction materials such as limestone, granite, and gypsum. A second goal is to maximize mass flows of useful and clean fractions. A third goal is to produce separation wastes that are well suited for disposal, by either landfilling or incineration. All of these goals can be achieved if mechanical sorting succeeds in controlling the flow of hazardous substances to certain fractions of sorting. Hence, it is of first importance to know the partitioning of heavy metals among the sorted products.

Table 3.28 lists the transfer coefficients (partitioning coefficients) for the two plants A and B. The results show that neither the dry nor the wet processes achieves the goal of directing the whole array of hazardous substances from recycling fractions to disposal fractions. Transfer coefficients for mass and substances are quite similar for most fractions, showing that true enrichment or depletion does not take place. It becomes clear that the superior qualities of the products of plant B are due to the clean input material and not because of a better separation by the wet process. MFA reveals the potential of the two technologies, and the transfer coefficients allow comparison of the separation efficiencies.

Table 3.28   Transfer Coefficients k of Selected Substances in Construction Waste Sorting Plants A and B, ×10−2

Plant A

Plant Ba

Substance

I

II

III

Metals

F1

F2

F3

LF

Sludge

Mass

45

25

27

3

20

36

33

5.1

≈ 4

Si

60

n.d.

40

n.d.

21

38

33

3.9

4.6

Ca

56

15

29

n.d.

23

41

28

2.7

5.0

Fe

14

10

13

63

18

34

27

4.3

5.5

TOC

16

80

4

n.d.

2.2

15

21

46

15

Al

42

21

34

3

23

40

25

6.0

6.8

S

57

24

18

n.d.

20

29

24

9.1

6.7

Zn

31

44

5

20

16

28

31

5.5

21

Cu

3

15

13

69

16

38

31

5.5

10

Pb

14

37

40

9

25

24

30

7.2

14

Cd

29

57

14

n.d.

19

33

30

5.9

12

Hg

43

36

12

n.d.

5.7

16

35

6.8

37

Note: n.d., not determined.

Notes:

a  Transfer coefficient kFe for scrap metals in plant B is 0.11. Transfer coefficient kS for wastewater in plant B is 0.11. All other kis for wastewater are <0.003.

Transfer coefficients display the partitioning of elements only; they do not yet allow direct comparison of the enrichment or depletion of substances. In Figure 3.32, the quotients substance concentrations in main fractions over concentration in construction waste are presented for plant A on a log scale. These quotients are chosen to measure accumulation and depletion. In plant A, the most enriched elements are iron, copper, zinc, and chromium in the metal fraction. Dry sorting successfully concentrates these metals in the metal fraction. Organic carbon, cadmium, mercury, and lead are enriched in the light (combustible) fraction II. Fractions I and III are similar. In both, the matrix substances Si, Ca, and inorganic carbon are slightly enriched, while organic carbon and some heavy metals are modestly depleted. Except for copper in fraction I, all substances are depleted by less than an order of magnitude in fractions I to III. For mixed construction wastes, this order of magnitude is necessary if the process is to produce materials that are similar to the composition of the Earth’s crust or to primary construction materials (see Table 3.24). There are no mechanical means yet to appropriately control the flow of all hazardous substances in sorting of mixed construction wastes.

3.2.3.5  Conclusions

Dry separation in plant A successfully concentrates combustible materials in the light fraction and construction-like materials in two other fractions. The processing yields about 70% of potentially useful construction products in two fractions and about 3% of metals for recycling. The remaining fraction of 25% is not suited for recycling or landfilling; it has to be incinerated. Plant A is not capable of reducing the contaminant level of any fraction significantly. The main disadvantage of all products is the high trace-metal concentration. When the recycling products from plant A are being used for construction, the buildings will contain heavy metals that are significantly above Earth-crust concentrations. When the light fraction is incinerated, sophisticated and expensive air pollution control is required. Thus, it is most important that contaminants be removed by selective dismantling before entering the construction waste recycling plant.

Figure 3.32   Enrichment, [concentration of X in fraction I]/[concentration of X in CW], of selected substances in main fractions of CW sorting plant A.

Due to a cleaner input, wet separation in plant B results mainly in two comparatively clean fractions well suited for recycling. Although a few of the heavy-metal concentrations are elevated compared with the Earth’s crust, they are (because of the cleaner input) generally of much lower concentration than in plant A. The overall performance of the wet process is similar to that of the dry process. While it is possible to produce a fraction rich in TOC and combustibles, significant accumulation or depletion of hazardous metals in any of the fractions is not observed. As for plant A, the light fraction contains much organic carbon, too, with the content of TOC reaching nearly 20%. Landfilling of a material with such a high TOC requires long aftercare periods. Thus, it seems appropriate to utilize the light fraction as a fuel. However, due to the presence of heavy metals such as Hg (see Table 3.27), boilers designed to utilize the light fraction must be equipped with efficient air pollution control devices for atmophilic metals.

Despite the differences between the inputs into the two separation processes, MFA and transfer coefficients allow a comparison of the performance of the two plants. From a recycling point of view, the main differences are the products, with plant A producing gravel substitutes and plant B producing sand and gravel. The regional market situation determines whether sand or gravel is to be preferred. From an environmental point of view, there are no important differences. Because neither plant can sufficiently enrich or deplete hazardous materials, the substance concentrations of the main product fractions are similar to the concentrations of the incoming construction wastes.

The results of the MFA of the two plants support the strategy of selective deconstruction. Neither of the two processes is able to accumulate or deplete significantly (factor 10) hazardous materials in any of the resulting fractions. Once again, it becomes evident that at today’s stage of development, mechanical processes are of limited use for the chemical separation of waste materials. Thus, wastes from indiscriminate demolishing of buildings are not well suited to produce recycling materials in construction waste sorting plants. For optimum resource conservation, it is important to separately recover materials during the deconstruction process and to recycle uniform fractions such as bricks, concrete, wood, and metals individually. In most cases, the remaining fraction can be mechanically sorted to recover a combustible fraction. Due to the composition of this fraction, containing plastic materials, paints, tubing, and cables, it is mandatory that energy recovery take place in incinerators equipped with state-of-the-art air pollution control devices suited to remove heavy metals such as mercury.

3.2.4  Case Study 8: Plastic Waste Management

Plastic materials were introduced in the 1930s. Ever since, polymers such as polyvinyl chloride (PVC), polyethylene (PE), polypropylene (PP), and polyamide (e.g., nylon) have shown large growth rates. Today they are among the most important man-made materials for many activities. At present, most plastics are made from fossil fuels that represent nonrenewable carbon sources. The production of plastics accounts for about 5% of the total fossil fuel consumption. They are used in cars, construction, furniture, clothes, packaging materials, and many other applications. Often, they contain additives to improve their properties. In particular, long-living plastic materials such as window frames, floor liners, and car fenders have to be protected from degradation and weathering by ultraviolet light, aggressive chemicals, temperature changes, and the like. Hence, plastic materials are usually mixtures of polymers with stabilizers, softeners, pigments, and fillers.

3.2.4.1  Plastic as Significant Fraction of MSW

Plastics make up between 10% and 15% of the total MSW flow. In addition, industrial and construction wastes are important sources of plastic wastes. Some plastic wastes (in particular from plastic manufacturing) are relatively clean and homogeneous and thus suitable for recycling. Others are mixtures of several goods and substances and hence cannot be recycled. Most plastic materials have a high energy content, and turned to waste, they can be used as a fuel. Due to stabilizers that contain heavy metals (lead, tin, zinc, cadmium, and others) and the chlorine content of some polymers (PVC, poly-vinylidene chloride), thus yielding dioxins during incineration, incinerators for plastic wastes generally must be equipped with advanced air pollution equipment.

As shown in Table 3.29, packaging materials are comparatively clean and may be used as a secondary resource. On the other hand, the stock of long-living plastics contains large amounts of hazardous substances that will have to be dealt with in the future. Hence, plastic recycling and waste management needs tailor-made solutions that are appropriate for the individual material and its ingredients.

Figure 3.33 shows the plastic flows and stocks through Austria (Fehringer and Brunner, 1996). The figure was prepared using data from plastic manufacturers, waste management, and other sources. In the following discussion, the focus is on plastic waste management, emphasizing plastic wastes as energy resources and as sources of hazardous materials. In 1992, about 8 million Austrian consumers bought roughly 1.1 million tons of plastic materials. A large portion is used in goods with long residence times (floor liners, window frames, car parts, etc.) and thus is incorporated into the “anthropogenic stock.” In Figure 3.33, this stock is assigned to the process consumption. The rest of the plastic is used for products with short residence times such as packaging materials and other consumer goods. The net flow (input minus output) into the stock of the process consumption amounts to 410 kt/year. Of the 720 kt/year of plastic wastes that leave the process consumption, 590 kt/ year is landfilled, and the rest is either incinerated or recycled. It is interesting to note that the packaging ordinance that was instated in Austria in 1992 does not change much of this situation. Only about 7% (49 kt/year out of 759 kt/year) of all plastic wastes are controlled by the packaging ordinance and are directed toward material recycling. About 71 kt/year is being incinerated together with MSW in MSW incinerators. By far the largest amount of plastic wastes (590 kt/year) was still disposed of in landfills. Hence, much energy was wasted, since 1 t of plastics corresponds roughly to 1 t of fossil fuels. The landfilling of plastic wastes is not only a waste of resources; it also offends the Austrian Waste Management Act (BMUJF, 1990). The goals of this law are directed toward the conservation of resources such as energy and materials, and the law explicitly calls for the minimization of landfill space. Neither of these requirements is observed by plastic waste management practices as presented in Figure 3.33. However, the introduction of a new landfill ordinance prohibiting the disposal of organic material changed the situation significantly.

Table 3.29   Additives in Plastic Materials in Austria

Material

Total Consumption (1992), 1000 t/Year

Packing Material Consumption (1992), 1000 t/Year

Total Stock (1994), 1000 t

Plastics

1000

250

6700

Softener

14

3

180

Ba/Cd stabilizers

0.25

0.0002

4

Pb stabilizers

1.6

0.002

27

Flame retardants

2

0

34

Source: Fehringer, R., and Brunner, P. H. (1996). Kunststoffflüsse und die Möglichkeiten der Kunststoffverwertung in Österreich. Vienna, Austria: Umweltbundesamt Wien GmbH.

Note: Plastics with short residence times such as packaging materials are comparatively clean. The long-lasting stock in construction, cars, and other applications contains large amounts of hazardous materials such as cadmium, lead, and organotin compounds.

Figure 3.33   Plastic flows and stocks in Austria. (From Fehringer, R. and Brunner, P. H., Kunststoffflüsse und die Möglichkeiten der Verwertung von Kunststoffen in Österreich, UBA Monographien Band 80, Umweltbundesamt, Vienna, 1996. With permission.)

3.2.4.2  Plastic Management from a Holistic View Point

In Figure 3.34, the advantage of an integrated MFA approach is visualized: If only MSW is considered (“MSW view”), 200 kt/year of plastic wastes are observed, with 80% being landfilled and 20% being incinerated. When public attention is drawn to packaging wastes, leading to legislation such as the Dual System in Germany or the Packaging Ordinance in Austria, a certain amount of plastic wastes (70 kt/year) is separately collected and thus not landfilled (−60 kt/year) or incinerated (−10 kt/year) anymore (“packaging view”). Due to inferior quality, not all separately collected plastic wastes can be recycled as polymers. Hence, a certain percentage is used as an alternative fuel, e.g., in cement kilns, leaving 50 kt/year for substance recycling.

If all plastic wastes are included in the assessment, a much larger amount of landfilled wastes is observed (590 kt/year) (“total waste management view”). It is important to note that without an investigation into the total national flows and stocks of plastic, it is not likely that the large amount of landfilled plastics can be identified. Only a balance of the process consumption, with estimates of the mean residence time of various plastic materials, allows a reliable assessment of wastes that are leaving consumption. It is a much more difficult, if not impossible, task to directly identify the amount of plastics in the many wastes landfilled. Figure 3.34 shows clearly that rational decisions regarding plastic wastes have to be based on a complete set of flows and stocks of wastes in a national economy (“resource management view”). The sole focus on a single waste category such as packaging wastes results in solutions that are not optimized regarding resource and waste management.

Figure 3.34   MFA as a decision-support tool enables different views of an issue such as management of plastic waste.

The benefit of an MFA approach in resource management as discussed in this case study is as follows: a total plastic balance at a countrywide level shows the important flows and stocks of plastics and helps in setting the right priorities in resource management. First, the large and useful stock of plastics (and thus materials and energy) in consumption and landfills is recognized. Second, potential hazards due to toxic constituents of plastic materials are identified in both stock “consumption” and landfill; the toxics will have to be treated with care in the future. This knowledge is a precondition for controlling the flow of polymers and their hazardous additives to processes well suited for recovery and final disposal such as plastic recycling and waste to energy.

3.2.5  Case Study 9: Aluminum Management

Due to the buildup of substantial anthropogenic aluminum (Al) stocks, the sourcing of secondary raw materials from these Al stocks has become of major interest from an economic as well as from an environmental perspective (EC, 2014). Knowledge about past and present Al use patterns in society is essential to evaluate future anthropogenic resource potentials and provide recommendations on optimized Al management.

In this case study, stocks and flows of Al in Austria are analyzed using static as well as dynamic material flow analysis to create a basis for optimized Al resource management on a national level. By the static MFA, an in-depth understanding of Al use patterns in Austria in the year 2010 (Buchner, Laner, Rechberger, and Fellner, 2014a) is created. Next, the dynamic model is developed based on historical data about Al production and consumption in Austria for the time period between 1964 and 2012 (Buchner, Laner, Rechberger, and Fellner, 2015a). The dynamic MFA allows for determining the in-use stocks of Al in various sectors following a top-down approach (Laner and Rechberger, 2016) and estimating the end-of-life (EOL) Al flows from these sectors to waste management and exports. The dynamic model is calibrated by adjusting model parameters based on independent bottom-up estimates. The model results are cross-checked against independent estimates and data. In a following step, the future development of in-use stocks and old scrap generation is projected by combining the data from the historical dynamic material flow model with forecasts on future Al consumption (Buchner, Laner, Rechberger, and Fellner, 2015b). Future Al consumption is estimated using a stock-driven approach (i.e., stock development is the driver for consumption) for some sectors and an input-driven approach (i.e., based on projections of future consumption, for instance, by using annual growth rates) for others. The model projections are used to evaluate if the domestic Al scrap potentially can satisfy the demand for Al scrap in Austria given current trends in Al consumption and scrap generation. Finally, the quality of Al scrap is introduced in the modeling to account for constraints concerning the recycling of mixed scrap into specific alloy types, e.g., cast alloys cannot be recycled to wrought alloys (Buchner, 2015). The model thereby allows for investigating the potential of cast-alloy production to absorb mixed old scraps under different scenarios considering the application of advanced sorting technologies.

3.2.5.1  Static Al Balance

A static MFA of Al in Austria for the year 2010 is performed to investigate current Al use patterns on the national level. A particular focus is put on the waste management phase and on Al flows on the scrap market to provide a basis for evaluating the resource efficiency of national Al use (Buchner, Laner, Rechberger, and Fellner, 2014a,b). The static material flow model is developed using STAN. It comprises all main stages of the Al life cycle, from production and processing, to utilization and waste management. Foreign trade flows are considered for unwrought Al, semifinished products, and final products (indirect Al flows) to determine the total final Al demand in 2010. The total old scrap (i.e., postconsumer waste) generation is estimated by balancing national secondary production with the amounts of new scrap (i.e., preconsumer waste) and the net-import of foreign scrap and unwrought Al. Al scrap amounts are estimated for individual use sectors by combining top-down and bottom-up estimates (Buchner, 2015). Finally, data quality is assessed based on data quality indicators and then translated into uncertainty ranges for all input data, given by mean values and standard deviations.

The total output of domestic secondary Al production in 2010 is 572 Gg/year, which is either used domestically or exported (Buchner, Laner, Rechberger, and Fellner, 2014a). The Al input to the use phase in final products is around 218 Gg/year or 26 kg capita−1 year−1 in 2010. The major Al-consuming sectors, making up 86% of national Al consumption, are buildings and infrastructure, transport, and packaging (see Figure 3.35). Forty-three percent of the input is accounted for as growth of Austrian in-use Al stock, with particularly strong Al stock growth in buildings. Old scrap generation is 7 kg capita−1 year−1, of which 80% is recovered in waste management processes. The highest losses occur for Al in packaging waste, where roughly 30% of the Al in wastes is either landfilled or oxidized during thermal waste treatment processes. The largest share of EOL Al flows from the transport sector are not directed to waste management but exported in old vehicles for further use outside of Austria. From a production perspective, secondary Al production in Austria is highly dependent on net imports, which constitute around 40% of the production input. Due to this high share of foreign scrap, for which a distinction between new and old scrap is not possible, the qualitative resource demand of national Al is hard to evaluate, with a possible range of old scrap utilization in national production between 0% and 66%.

Figure 3.35   Total Al consumption of each use sector and partitioning into different pathways in the static model based on Buchner, Laner, Rechberger, and Fellner (2014a). Reading example: Al consumption of the sector buildings and infrastructure is around 70,000 metric tons/year, with nearly 70% of the consumed Al adding to the buildup of stock.

3.2.5.2  Dynamic Al Flow Model

The static MFA provides the basis for developing a dynamic material flow model, which enables a detailed investigation of the Al in-use stock developments and the trends in Al scrap generation over time. The focus on the national system provides the opportunity of increasing the confidence in model outcomes based on the comparison with other estimates, which is typically not possible for dynamic material flow studies on a large scale (cf. Buchner, Laner, Rechberger, and Fellner, 2014a). In-use Al stocks are calculated for six sectors following a top-down approach (see Equation 3.9). The growth of in-use stock in a specific year t is determined by subtracting the output O(t) from the input I(t). Summing up the change of stock over all previous time periods (from 1 to T) and accounting for the initial stock in year 0 S(0) results in the total stock at the time T S(T). This is a widely used approach in dynamic material flow modeling to derive estimates of in-use metal stocks based on historic production and consumption data, and to make projections on future secondary resource availability (e.g., Pauliuk, Wang, and Müller, 2013; Müller, Hilty, Widmer, Schluep, and Faulstich, 2014). However, as historic data on the outputs from use sectors are rarely available, the output of obsolete products is typically calculated using sector-specific lifetime functions. Such functions are defined for specific end-use sectors, with outputs being calculated by accumulation of the fraction of all former inputs becoming obsolete in a respective year. This is done by combining the input function I(t) with the lifetime function flt(t) in a convolution operation (Müller, Hilty, Widmer, Schluep, and Faulstich, 2014) as described in Equation 3.10, where T is the year for which the output is determined and d is the duration in years the material has been used. Because it is typically not possible to solve this convolution analytically, the calculations are performed for discrete time steps of single years. With respect to lifetime functions, several types of statistical distribution functions, such as normal, lognormal, beta, or Weibull functions, are available for describing the residence time of material in the in-use stock (Melo, 1999). In the present study, Weibull functions with different sector-specific parameters (i.e., average lifetimes) have been chosen to model the obsolescence behavior of in-use Al products (cf. Buchner, Laner, Rechberger, and Fellner, 2015a).

3.9 $S ( T ) = S ( 0 ) + ∑ t = 1 T I ( t ) − O ( t )$
3.10 $O ( T ) = ( I ⋅ f l t ) = ∑ d = 1 ∞ I ( T − d ) ⋅ f l t ( d )$

where S is stock, I is input, O is output, and T is the time for which the stock and the output is determined.

Apart from the choice of lifetime functions and corresponding parameters, many other model parameters (e.g., sector-split ratios, recycling rates) have to be defined. These model inputs are considered to be uncertain and potentially varying over time. In order to improve the initial parameter estimates, independent bottom-up estimates are used to calibrate the dynamic material flow model. Such estimates can be established for the input flows to the transport and packaging sectors, where it is then possible to adjust the respective sector-split ratios. Furthermore, the model outcomes can be crosschecked with results from other studies or statistical data to evaluate their plausibility. Implementing these calibration and validation steps creates a more reliable basis for assessing historical, current, and future Al resource use and scrap availability.

Projections on the future consumption of Al and the development of in-use stocks enable evaluations of the Al resource availability in Austria until the year 2050 (Buchner, Laner, Rechberger, and Fellner, 2015b). For three of the six in-use sectors (transport, buildings and infrastructure, and electrical equipment), a stock-driven approach is used to determine future Al consumption and calculate Al scrap flows. Although this approach is considered to be more robust in a long-term perspective than inflow projections, it cannot be applied to the other in-use sectors. In the case of the packaging sector, there is no substantial accumulation of Al in stocks, and for the sectors machinery and consumer goods, bottom-up stock estimates are not available. Therefore, the future development of Al consumption in these sectors is calculated by assuming a certain growth rate of annual consumption starting from current levels (in 2012).

The results of the dynamic Al flow model are shown in Figure 3.36 for the total inflow to and the outflow from the use phase and the development of in-use Al stocks over time. It is apparent that Al consumption has been on the rise since the mid-1960s (beginning of the model period) until now (2012 in the model) and is expected to continue increasing also in the future. Al usage is expected to grow at a particularly high rate in the transport sector due to lightweight construction of cars. The generation of old scrap (output in Figure 3.36) will even grow at a slightly higher rate than consumption in the future and increase from a current level of around 130 Gg (14 kg/ capita) per year to 210 Gg (24 kg/capita) per year in 2030 and 290 Gg (31 kg/ capita) per year in 2050 (cf. Buchner, Laner, Rechberger, and Fellner, 2015b). The most substantial increases in old scrap generation are expected for the transport sector and the building and infrastructure sector. The total in-use stock of around 3.0 million metric tons (Tg) (360 kg/cap) in 2012 is projected to increase to 3.9 Tg (440 kg/cap) until 2030 and to 5 Tg (530 kg/cap) until 2050, which corresponds to an average annual growth rate of the in-use Al stock of 1% during the next 40 years.

Figure 3.36   Total final Al demand (input to use phase) and EOL Al flows (output from use phase) as well as in-use stock development for the period from 1964 until 2050 in Austria based on Buchner, Laner, Rechberger, and Fellner (2015b).

3.2.5.3  Potential of Anthropogenic Stock to Satisfy Demand

The ongoing growth of Al resource flows and stocks highlights the significance of efficiently managing anthropogenic Al resources. From a national perspective, the question of whether domestic secondary resources can provide the basis for satisfying domestic Al demand is of particular interest. Therefore, the results of the dynamic model projections for the development of Al scrap generation (cf. Figure 3.36) are used to evaluate the future Austrian “self-supply potential” with respect to secondary Al.

The doubling of scrap generation from 2010 to 2050 is an opportunity for increasing the self-supply rate with respect to final Al consumption in Austria. Assuming no imports of unwrought Al and Al scraps and no exports of Al scrap as well as a ban on exports of Al-rich EOL products (i.e., end-of-life vehicles) and higher Al collection and recovery rates in waste management (sector-specific collection rates of 90-95%, processing losses of 2%), the final self-supply of Al for consumption is not expected to exceed 75% in 2050 [cf. scenario Rhigh in Buchner, Laner, Rechberger, and Fellner, (2015b)]. Hence, given the growth rates assumed for Al consumption, complete self-supply is not achievable in the foreseeable future, even with positive suppositions on recycling efficiency. If per capita consumption remained constant at the level of 2012 (approximately 23 kg capita−1 year−1), still the available anthropogenic Al resources would not suffice to satisfy demand completely in 2050 [self-supply would rise to 83%; see Buchner, Laner, Rechberger, and Fellner, (2015b)]. Thus, satisfying final domestic Al demand based on domestically available secondary raw materials could only be reached through a decrease in consumption, which seems rather unlikely given historic developments.

A major issue with respect to a circular economy, apart from the quantity of recycled materials, is the quality of materials introduced to the recycling loop. In case of Al, the mix of old scraps from different applications contains various alloy elements in different concentrations, which may be a critical constraint to the use of old scrap in secondary production. Thus, increasing future Al self-supply may be jeopardized by qualitative limitations for Al recycling and secondary raw material utilization. In a first screening evaluation, wrought alloys are distinguished from cast alloys in the dynamic Al flow model, and the supply of Al scrap is projected with respect to these two major groups of alloys. Due to product specifications, cast alloys cannot be used to produce wrought alloys, but wrought alloys can be used to produce cast alloys. Consequently, the cast-alloy production represents a sink for Al scrap of different quality. Possible quality constraints regarding Al recycling in a closed national system (self-supply scenario) are investigated by comparing current and future cast and mixed scrap generation to national final cast Al demand (Buchner, 2015). The analysis of various scenarios using the dynamic Al flow model shows that a surplus of mixed scrap is expected to occur in the near future, if no separation between wrought and cast alloys is applied. This points out directions for future technological progress. The introduction of advanced separation technologies applied to old scrap could prevent a substantial surplus of mixed Al scrap compared to national final cast Al demand until 2040. Afterward, new technologies or foreign trade flows may be required to compensate for the misfit between the national final demand pattern and the scrap generation pattern in terms of alloy composition.

The MFA modeling results indicate clearly that the current recycling practice will lead to unsuitable Al scrap qualities on the domestic market, if high recycling rates and closed cycles are aspired to. Though these findings relate to aluminum and Austria, a single metal and a rather small country, they may essentially be transferable to other metals and other highly developed markets such as the European Union, because Al consumption patterns are rather similar. Consequently, moving toward a circular economy in terms of metal recovery requires intensified recycling and sorting of scrap. This implies technology development and appropriate commodity markets for secondary raw materials, because quality and composition of metal scraps are a determining factor for the properties of secondary metal products. Dynamic MFA represents a powerful tool for supporting policy development regarding Al, but also all other metals, in a circular economy.

Problems—Section 3.2

Problem 3.5:

• Assume that the production of food by traditional agriculture can be replaced by “hydrocultures” that do not require soil for plant production. What will be the major change regarding total nutrient (N, P) requirements and losses from the activity to nourish? Use Figures 3.15 and 3.16 for your discussion.

Problem 3.6:

• Use the following information to complete the four exercises listed afterward.
• In 1996, about 8.1 million t/year of zinc (Zn) ores and 2.9 million t/year of Zn scrap are processed in order to produce 9.6 million t/year of Zn. Ore processing resulted in approximately 230 million t/year of tailings from milling with a content of about 0.3% Zn and 14 million t/year of slag from smelting with about 5% of Zn, each material flow representing a Zn flow of ca. 0.7 million t/year. Mining wastes are not considered. Zn is further manufactured into products that can be roughly grouped into five categories: galvanized products (3.3 million t/year), die castings (1.3 million t/year), brass (1.5 million t/year), Zn sheet and other semiproducts (0.6 million t/year), as well as chemicals and other uses (1.4 million t/year).

Galvanizing here stands for all kinds of technologies producing a coating of Zn on iron or steel in order to avoid corrosion. Die casting is a process to produce strong accurate parts in large quantities by forcing molten Zn alloy under pressure into a steel die (mainly used in the automotive industry). Brass is an alloy based on copper (Cu) and Zn. The Zn content ranges up to ca. 40%. Brass is used as sheets, wire, tubes, extrusions, and so on. Zn sheet is produced from Zn or Zn alloy rolled into thin sheets suitable for forming into roofing and cladding and other applications. The last category comprises mainly the dissipative uses, where Zn occurs as a trace metal, for example, in paints, automotive tires, brake linings, pesticides, animal feed and food additives, pharmaceuticals, cosmetics, etc. Manufacturing also results in production waste (Zn content, ca. 1.5 million t/year), mainly in the form of brass and galvanizing residues.

The total amount of Zn in products entering the use phase is 8.1 million t/year. The amount of Zn discarded is estimated at about 2.2 million t/year. Waste management separates 1.4 million t/year of Zn from the waste stream (Zn scrap). The remainder, which has a mean concentration of about 0.1% and comprises waste categories such as municipal solid waste, construction and demolition debris, wastes from electrical and electronic equipment, automotive shredder residues, hazardous wastes, industrial wastes, and sewage sludge, is landfilled (0.8 million t/year). This latter figure can only be regarded as a rough estimate. Mass flows of goods, their Zn content, and the resulting Zn flows are given in Table 3.30.

Table 3.30   Flows of Zn-Containing Materials, Their Zn Content, and Related Zn Flows for the World Economy

Goods

Mass Flow, Million t/Year

Zn Content, %

Zn Flow, Million t/Year

Zn ore

160

5

8.1

Tailings

230

0.3

0.7

Slag

14

5

0.7

Metal

9.6

100

9.6

Production waste

3.0

50

1.5

Zn scrap

17

11

1.4

Products

1500

0.54

8.1

Galvanized products

83

4

3.3

Die castings

1.3

99

1.3

Brass

4.3

35

1.5

Zn sheet and semiproducts

0.6

99

0.6

Chemicals and others

1400

0.1

1.4

Flow into stock

2200

0.27

5.9

(dissipative loss)

(1700)

(0.007)

(1.3)

Wastes

810

0.27

2.2

Landfilled

wastes

800

0.1

0.8

Note: Values rounded to two significant digits.

1. Establish the flow diagram of the described Zn system.
2. Assign the material flows of the flowchart to stages and draw a diagram according to Figure 3.23 in Section 3.2.2.
3. Calculate the statistical entropy trend for the system. Is the trend sustainable?
4. Calculate what happens if 15% of consumed/used Zn neither is transformed to waste nor remains in the stock, but escapes to the environment (assume that the Zn flow is evenly dispersed in the soil.)

Problem 3.7:

• Consider the following quantitative flowchart for the fluxes and stocks of construction materials within a fictitious region (see Figure 3.37).

1. Which stock will be most important for sand and gravel after 100 years (constant materials management assumed)?
2. Which conditions are required in order for recycling of construction materials to make a substantial contribution to the supply of construction materials (both buildings and underground)?
3. Which differences in material quality do you expect in the four stocks (which is the fourth stock)?

The solutions to the problems are given on the website http://www.MFA-handbook.info.

3.3  Waste Management

MFA is an excellent tool to support decisions regarding waste management for the following reasons:

1. In waste management, waste amounts and waste compositions are often not well known. MFA allows calculating the amount and composition of wastes by balancing the process of waste generation or the process of waste treatment. Thus, MFA is a well-suited tool for cost-efficient and comparatively accurate waste analysis.
2. As mentioned in the first paragraph of Chapter 1, inputs and outputs of waste treatment processes can be linked by MFA. Thus, if transfer coefficients are known, one can assess whether a given treatment plant achieves its objectives for a given input. Often, transfer coefficients are not known in waste management, but they can be determined by MFA even if some inputs or outputs are not known.

Figure 3.37   Fluxes and stocks of construction materials.

3. Advanced waste management is a comparatively young branch of the economy. There is rapid development, driven by rising waste amounts, new technologies, and differing interests of stakeholders. There is a need for policy advice regarding future directions: What are the deficits of a given waste management system with regard to the goals set? What is the cost-effectiveness of a waste management system to reach the goals? And how can goal orientation as well as cost-effectiveness be measured and improved?

The following case studies are presented to exemplify how MFA can be used for waste analysis, optimization of waste treatment, waste policy analysis, and upporting policy decisions regarding waste management.

3.3.1  Use of MFA for Waste Analysis

Reliable information on waste composition and waste generation rate is crucial for the following objectives:

1. To identify potentials for recycling (biomass, paper, metals, plastics, etc.)
2. For the design and maintenance of waste treatment plants, including air and water pollution control technologies (recycling, incineration, landfilling)
3. To predict emissions from waste treatment and disposal facilities
4. To examine the effects of legislative, logistic, and technical measures on the waste stream

Because the composition and the generation rate of wastes are changing constantly, it is necessary to analyze them periodically. This is especially true when new consumer goods are being introduced to the market. Thus, routine, cost-effective determination of waste composition and of time trends is essential for waste management. In this chapter, selected methods of characterizing MSW are presented and discussed. These approaches were originally presented in a paper by Brunner and Ernst (1986).

The parameters that are used to characterize waste materials can be divided into three groups:

1. Materials or fractions of MSW (e.g., paper, glass, metals)
2. Physical, chemical, or biochemical parameters (e.g., density, heating value, biodegradability)
3. Substance concentrations (e.g., carbon, mercury, hexachlorobenzene)

To solve a particular problem of waste management, it is usually not necessary to analyze all parameters. For example, for recycling studies, information on the content of certain fractions in MSW such as paper or glass is required. To predict emissions, the elemental composition of MSW needs to be known.

Generally, there are three main methods for solid-waste analysis (see Figure 3.38). The first involves direct analysis of MSW, while the second and third are indirect methods based on MFA and the mass-balance principle.

1. Direct analysis, also known as the “sample and sort” method. A specified, statistically planned amount of MSW is collected. Samples are taken, screened, analyzed for waste goods, dried, pulverized, and finally analyzed for substances. The sample that is analyzed is usually small compared with the total MSW generated. This method has been used in many waste-characterization studies in the United States, Europe, and elsewhere (Barghoorn, Dobberstein, Eder, Fuchs, and Goessele, 1980; BUWAL, 1984; Maystre and Viret, 1995). Several manuals have been published describing how to conduct such analysis (Yu and Maclaren, 1995).
2. Indirect analysis of MSW composition by market-product analysis. This approach requires information about the production of goods and about the fate of these goods during use and consumption. Data collected from industrial sources such as key corporations, professional organizations, or government agencies are used to estimate flows of goods that are produced and consumed. The generation of MSW is calculated by measuring or assuming average life spans for these goods. Various adjustments are made for imports, exports, and stocks in each product category. This method was developed in the early 1970s. Since then, data collection has improved, and databases have evolved. Results are compared with information about wastes that are landfilled, combusted, or recycled and with direct waste-analysis studies. The U.S. Environmental Protection Agency (EPA) applies this approach to estimate MSW generation (US EPA, 2002).

Figure 3.38   Methods for MSW analysis. (From Brunner, P. H. and Ernst, W. R., Waste Manage. Res., 4, 147, 1986. With permission.)

3. Indirect analysis using information about the products of waste treatment to calculate MSW composition. The advantage of this method is that the outputs of waste treatment are usually less heterogeneous than the input waste.

Especially for long-term monitoring, it is more cost-effective and accurate to determine waste composition by indirect methods (Morf and Brunner, 1998).

3.3.1.1  Direct Analysis

Direct waste analysis was the first approach used to determine waste composition. Waste samples are collected from different communities or regions based on statistical evaluations. The sample size usually varies between a minimum of 50 kg up to several tons. Samples are classified by hand into a selected number of fractions (paper, glass, etc.). Mechanical equipment is commonly used to separate magnetic metals and to sieve the remaining unidentified material into several additional fractions of different particle sizes. In order to determine the chemical and physical parameters of each fraction, representative samples are drawn from each material. These samples are further prepared (dried, pulverized, and sieved) for laboratory analysis.

The direct method is useful for

1. Measuring the concentration of most materials in MSW
2. Determining energy and water content of MSW and its fractions
3. Investigating the influence of geographic, demographic, and seasonal factors on the concentration of materials and some parameters in MSW
4. Assessing changes of waste composition with time
5. Evaluating the impact of separate collection measures on waste composition such as content of paper or glass or the impact of different collection systems (e.g., size of waste containers)

However, the direct method of waste analysis also has a number of limitations and disadvantages. First, it is labor-intensive and requires expensive equipment. Provided that adequate technical equipment and sufficient personnel are available, the analysis of one truckload takes at least half a day. A monitoring study on annual changes of MSW is estimated to consume 15 person-months of unpleasant and unhealthy labor. Second, the residue of separation that is not assigned to defined fractions such as glass, paper, etc., is usually quite large, often making up as much as 40–50% of the total MSW analyzed. As long as the composition of this fraction remains unknown, the value of the other results can be questioned, for example, the assessment of recycling potentials. Third, the determination of trace-element concentrations is problematic. If, for example, mercury batteries and their contribution to heavy metals in MSW are analyzed, one may find a few small batteries in 1 ton of MSW. This results in an average sample concentration of one to a few milligrams of mercury per kilogram of MSW. However, if only one or two MSW samples of 2 to 20 kg are collected, there is either a great chance of finding no mercury from batteries at all or of finding a high concentration if a single battery turns up in one of the randomly selected samples (Hg content up to 30% for Zn/Hg batteries).

This challenge is illustrated in Figure 3.39. If the chosen sample size is too small, the result of the analysis will probably be too low. The possible range of results increases with smaller sample sizes. Sufficiently large samples are needed to achieve results that reflect the actual content of unknown substances. Fourth, for technical and economic reasons, the metal fraction is often excluded from the chemical–physical analysis. However, this fraction may contain considerable amounts of heavy metals. Therefore, results of direct waste analysis may represent minimal values. A fifth problem is erosion and contamination from grinding and pulverization equipment.

Figure 3.39   Drift of the most probable result as the sample size becomes very small. (From Pitard, F. F., Pierre Gy’s Sampling Theory and Sampling Practice, Vol. II, Sampling Correctness and Sampling Practice, CRC Press, Boca Raton, FL, 1989, p. 159. With permission.)

These problems highlight some of the distinct limitations of the direct analysis of MSW with regard to the determination of chemical parameters, particularly of trace substances. They indicate that direct chemical analysis of waste materials represents the actual field concentrations only when large sampling campaigns are undertaken, resulting in extremely high costs. The direct approach is well suited for the determination of materials in MSW, but it seems of limited value in analyzing the elemental composition of MSW.

3.3.1.2  Indirect Analysis: Case Studies 11 and 12

The aforementioned problems and limitations of direct waste analysis led to the development of complementary methods, which yield more accurate results with less effort in terms of manpower and costs. Two case studies are presented to illustrate the use of MFA in indirect analysis.

3.3.1.2.1  Case Study 10: Waste Analysis by Market Analysis

Goods are produced and consumed. After use, they are either recycled or discarded as wastes. Since most industrial branches have accurate figures about their production, and since the pathways of many goods are well known, it is often possible to calculate the composition of MSW without field analysis and with high accuracy. This procedure, which can be used to analyze both the contents of materials and the elemental composition, is illustrated by the following examples for paper, glass, and chlorine content in MSW.

Paper: The most abundant single substance in MSW is cellulose, the main constituent of paper.

For paper recycling as well as waste treatment, it is of considerable interest to know the amount of paper in MSW. Figure 3.40 shows the flux of paper through the Austrian economy. Data are collected from pulp and paper manufacturers and checked against other available information. The amount of paper in MSW (48 kg/capita/year) is calculated as the difference between total paper consumption (179 kg/capita/year) and separately collected and recycled wastepaper (131 kg/capita/year). The Austrian population in 1996 was around 8.1 million inhabitants, and MSW generation amounted to 1.3 million t/year, which translates to 160 kg/capita/year of MSW for each resident. Based on these figures, a paper content of 30% (48 kg of wastepaper in 160 kg of MSW) can be calculated for average Austrian MSW. This figure has been confirmed by direct analysis.

Glass: A simple balance for the per capita glass flux in Switzerland in 2000 is given in Figure 3.41 (Kampel, 2002). Only packaging glass (bottles, beverage containers, etc.) is considered. The amount of glass in Swiss MSW (2.8 kg/ capita/year) is calculated as the difference between glass consumed (46.6 kg/ capita/year) and glass recycled (43.8 kg/capita/year). Glass with residence time greater than 1 year is not considered. Accumulation in the stock household is assumed to be less than 1% of consumption. With a Swiss population of 7.2 million inhabitants and 2.54 million t of MSW generated annually, the per capita allocation of MSW is 350 kg/capita/year. Based on these data, an average concentration of packaging glass in MSW of 8 g glass/kg MSW is calculated.

Figure 3.40   Paper flows in Austria (1996), kg/capita/year. (From Austrian Paper Industry, Personal communication, 1996.)

Figure 3.41   Recycling of packaging glass in Switzerland in 2000, kg/capita/year.

Paper and glass products have a short lifetime of less than 1 year. Therefore, it is reasonable to assume that inputs equal outputs over the balancing period. For other products with longer or even unknown residence times (e.g., wood in building materials), attempts to balance are more difficult. Yet, the US EPA studies on MSW generation show that this method is successful. Kampel used this approach to determine differences in waste-glass management among Australia, Austria, and Switzerland (Kampel, 2002).

Chlorine: The main sources of chlorine in MSW are assumed to be PVC and sodium chloride (NaCl). Minor amounts of Cl are contained in plant materials, other plastic materials, and other products. Thus, the content of Cl in MSW can be roughly estimated by the figures on consumption of PVC and table salt and by assumptions on the fate of these products during and after consumption and use. Data about goods such as NaCl and PVC are usually published in annual reports of the specific industrial branch, e.g., salt mine operators and plastics manufacturers. Sodium chloride in private households is utilized for dietary purposes mainly. It is assumed that not more than 10% of the NaCl purchased is discarded with MSW. Most salt is either eaten or discarded with wastewater while preparing food; in both cases, chloride leaves the household via sewage. Residence times of goods containing PVC are difficult to assess. It is estimated that 50 ± 20% of PVC is used in long-life products, and the other part is used for short-residence-time packaging material and consumer goods. Note that there is not yet a steady state for PVC flows. There is a large yearly growth rate on the input side. Because of the long residence time of some products, PVC is accumulated in the anthroposphere. Therefore, the amount of PVC-derived chlorine in MSW is calculated according to varying percentages of PVC. Despite the fact that chlorine estimates are based on several assumptions, the order of magnitude (5 to 10 g Cl/kg MSW) in Table 3.31 compares well with values resulting from product analysis of 7 to 12 g Cl per kg MSW.

Advantages and drawbacks: The main advantage of the analysis of MSW by a material balance of market products is the fact that no measurements are needed. MSW composition can be assessed quickly with little effort. In most cases, such rough estimates can give good results on a nationwide level. However, the method is not well suited to identify regional differences. It is usually more important to have reliable figures on the production/ consumption side of a product than to have exact estimates of the proportion that enters the waste cycle. Another advantage of this method is the potential to predict trends in waste composition. Because today’s products determine tomorrow’s waste composition, this method is the only one that can be used to predict future waste composition.

Table 3.31   Determination of Chlorine in Swiss Municipal Solid Waste by Market Analysis

NaCl

PVC Min.

PVC

PVC Max.

Consumption/use, kg/capita/year

5

8

8

8

10

30

50

70

Mass in MSW, kg/capita/year

0.5

2.4

4

5.6

Cl content, g/kg

610

580

580

580

Mass of Cl, kg Cl/capita/year

0.31

1.4

2.3

3.2

Contribution to Cl in MSW, g Cl/kg MSW

0.9

3.8

6.3

8.8

Total Cl in MSW (market analysis), g Cl/kg MSW

5–10

Direct analysis, g Cl/kg MSW

3.4–4.2

Product analysis, g Cl/kg MSW

7–12

Drawbacks of the method are (1) the dependency on production/consumption data, which are usually known on a national level only, and (2) that data are available only for a limited amount of materials and elements. It is not yet possible to characterize MSW from a physical point of view by this method (e.g., density and particle size).

3.3.1.2.2  Case Study 11: Analysis of Products of Waste Treatment

The analysis of the products of different waste-treatment processes is a powerful tool to characterize MSW (Brunner and Ernst, 1986). The main advantage is the homogenizing effect of treatment processes. This is particularly true if incineration is chosen for analysis. The incinerator acts as a large “thermal digester,” separating substances from each other and releasing products that are of more uniform composition than the initial MSW. If all residues of the incinerator are analyzed and the total input and output mass flows are determined over a given period of time, the composition of the input into the plant can be calculated. This makes it possible to determine the flows of selected elements through an MSW incinerator and calculate the chemical composition of the waste input. The method has been successfully applied to several incinerators (Brunner and Mönch, 1986; Reimann, 1989; Vehlow, 1993; Belevi, 1995; Schachermayer, Bauer, Ritter, and Brunner, 1995; Morf, Ritter, and Brunner, 1997; Belevi and Mönch, 2000; Morf, Brunner, and Spaun, 2000).

Procedure: The procedure employed in a full-scale incinerator is as follows. The total mass flows of all input and output goods are determined during a given measuring period. Typical measurement campaigns last from several hours to several days. A (crane) balance measures the weight of the waste material fed to the incinerator. Consumption of water and chemicals is continuously recorded by incinerator control devices. The volume of air used for combustion is calculated based on a final mass balance and data about the energy consumption of the air blower. Solid incineration products are collected separately and weighed as received. Wastewater and off-gas are measured routinely by online flowmeters and translated into mass flows. To determine the chemical compositions, samples of bottom ash, filter cake, purified wastewater, and fly ash from the electrostatic precipitator (ESP ash) are taken and prepared for analysis. The bottom ash is the most heterogeneous product and requires extensive processing before analysis. First, it is separated from large pieces of iron, crushed, and sieved. The oversize material is weighed but usually not analyzed. It is assumed to consist mainly of iron (an assumption that is not justified for every incinerator). From the pretreated bottom ash, several composite samples are dried (at 105°C for ca. 24 h until a constant weight is achieved) and pulverized in a mill. Again, the oversize material is assumed to be of iron. For balance calculations, all fractions of the bottom ash are taken into account. Composite samples of fly ash are taken as close as possible to the filter device (to avoid time lag) and pulverized to laboratory sample size. Wastewater and filter cake, two rather homogeneous products, are sampled, too. In coordination with the sampling of solid and liquid incineration products, off-gas samples are taken to determine the flows of substances that are not measured continuously (mainly heavy metals). Detailed descriptions of effective sampling plans, procedures, the preparation of samples, and methods of analysis are presented by Morf, Brunner, and Spaun (2000) and Morf, Ritter, and Brunner (1997).

Figure 3.42 gives an example of appropriate locations for sampling and measurements in an MSW incineration plant.

Figure 3.42   Locations of sampling and mass flow metering for indirect waste analysis in an MSW incinerator.

The element balances are calculated by multiplying the mass flow of goods by the respective concentrations of the elements for every period of the campaign. As mentioned before, input composition is not measured but is calculated indirectly by summarizing the mass flows of each element in the incineration products (minus substance inputs in other input goods such as water, air, or chemicals) and dividing by the mass flow of the waste input (see Equation 3.11).

3.11 $C M S W , j = ∑ i = 1 k c i j ⋅ m ˙ i m ˙ M S W$

where

• k = number of incineration products
• j = substance

Concentrations of C, Cl, F, S, and several heavy metals in MSW have been determined by this method. Table 3.32 lists the results from six studies of five incinerators in Austria and Switzerland.

When analyzing wastes, it is highly important to consider uncertainty and to assess aspects of quality control. Bauer (1995) developed a method for quantifying the statistical uncertainty of such indirect waste analysis. Thus, it is possible to determine the effort that is necessary to obtain a given confidence interval for the waste composition. Higher efforts (more samples per time, larger sampling sizes) yield more reliable results (smaller uncertainties). A relationship between cost and accuracy can be established. Results with sufficiently small intervals (below ±20%) with a confidence of 95% can be obtained at reasonable costs. Morf and Brunner (1998) extended this approach. Based on MFA and transfer coefficients, they developed a method that allows routine measurement of MSW composition by analyzing only a single product of incineration per substance. They present procedures and examples of how to select the appropriate incineration residue to be analyzed, how to determine the minimum frequency for analyzing the residue, and how to measure the chemical composition of MSW routinely.

Results: Results of such investigations into MSW concentration are shown in Figures 3.43 and 3.44 (Brunner, Morf, and Rechberger, 2004). The monthly mean values of Cl and Hg vary by up to a factor two. The daily flows of the two selected elements Cl and Hg also vary. For Hg, these variations are quite substantial and up to a factor of four within a period of a few days. This emphasizes that random moment investigations are not a sufficient means of determining MSW composition.

The proposed MFA-based method for routine monitoring of waste composition by analyzing single incineration residues has significant advantages regarding data quality compared with the normally applied direct waste analysis. If waste composition were measured in the same way on several MSW incinerators, this would allow comparison of waste compositions in a more cost-effective and objective way than the present practice of direct waste analysis. Future MSW incinerators should be designed for and supplied with hardware and software to apply routine MFA for waste analysis. The additional costs would be small and the return on investment large when compared with the costs and accuracy of traditional approaches.

Table 3.32   Results of Different Indirect Waste-Analysis Campaigns, g/kg

Biel (CH) 1981

Müllheim (CH) 1984

St. Gallen (CH) 1991

Vienna (A) 1993

Wels (A) 1996

Wels (A) 1996

C

275 ± 55

n.d.

370 ± 40

190 ± 10

252 ± 25

265 ± 28

Cl

6.9 ± 1.7

n.d.

6.9 ± 1.0

6.4

12.2 ± 1.8

10.3 ±1.2

S

2.7 ±0.5

n.d.

1.3 ± 0.2

2.9 ± 0.2

4.2 ± 0.14

4.1 ± 0.17

F

0.14 ± 0.06

n.d.

0.19 ±0.03

1.2 ± 0.1

0.054 ± 0.007

0.060 ± 0.002

Fe

67 + 35

n.d.

29 ±5

42 ±1

37 ± 0.25

43 ±0.2

Pb

0.43 ± 0.13

0.57 ± 0.43

0.70 ± 0.10

0.60 ± 0.10

0.40 ± 0.079

0.49 ± 0.088

Zn

2.01 ± 1.51

1.1 ± 0.5

1.4 ± 0.2

0.83 ± 0.07

1.2 ± 0.069

1.3 ± 0.14

Cu

0.27 ± 0.07

0.46 ± 0.19

0.70 ± 0.20

0.36 ± 0.03

0.59 ± 0.13

0.52 ± 0.076

Cd

0.0087 ± 0.0019

0.012 ± 0.0056

0.011 ± 0.002

0.008 ± 0.001

0.0107 ± 0.0028

0.0084 ± 0.0026

Hg

0.00083 ± 0.00081

0.002

0.003 ± 0.001

0.0013 ± 0.0002

0.0019 ± 0.00039

n.d.

Source: Morf, L. S. et al., Güter- und Stoffbilanz der MVA Wels: Institut für Wassergüte und Abfallwirtschaft, TU Wien, 1997.

Note: n.d. = not determined.

Figure 3.43   Time trends for monthly mean MSW concentrations of chlorine and mercury as determined for an incinerator (Spittelau) in Vienna, Austria, between February 1 and September 30, 2000. The figure shows means as well as the lower and upper limits for an approximately 95% confidence interval. (Reprinted from Solid Waste: Assessment, Monitoring, and Remediation, Twardowsky, I., Allen, H. E., Kettrup, A. A. F., and Lacy, W. J., Eds., Brunner, P. H., Morf, L., and Rechberger, H., Copyright 2003, with permission from Elsevier.)

Figure 3.44   Time trends for daily flows of Cl (kg/day) and Hg (g/day) through the MSW incinerator (Spittelau) in Vienna, Austria, between September 1 and September 30, 2000. (Reprinted from Solid Waste: Assessment, Monitoring, and Remediation, Twardowsky, L., Allen, H. E., Kettrup, A. A. F., and Lacy, W. J., Eds., Brunner, P. H., Morf, L., and Rechberger, H., Copyright 2003, with permission from Elsevier.)

The main disadvantage of the waste product analysis is that waste fractions cannot be determined, e.g., it is not possible to calculate the contents of paper, plastic, or any other single fraction. This means that, in most cases, the product method is limited to the analysis of elemental composition and parameters like energy content, water content, and the content of total inorganic and organic matter.

Conclusion: It is highly important to choose the method of analysis that is most appropriate to solve a particular problem of waste management. In general, direct waste analysis yields good results on some fractions in MSW, but it is expensive and labor-intensive to determine reliably elemental concentrations by this method. Market-product analysis combined with MFA is an inexpensive and quick method to determine with sufficient accuracy the fraction-based and elemental composition of MSW. In many cases, this method of analysis can be applied in favor of direct waste analysis. However, the method is limited to those materials where information from the producing industries is available and where residence times in stocks are more or less known. MFA-based waste product analysis is well suited for determining element concentrations in MSW, but it does not allow analysis of material composition. It is the superior, cost-efficient method for determining time trends in elemental analysis of wastes.

3.3.2.1  Case Study 12: ASTRA

In the case study ASTRA (a German acronym for “evaluation of different scenarios for waste treatment in Austria”), selected scenarios for the treatment of combustible wastes are compared in view of reaching the waste-management goals of “environmental protection,” “resource conservation,” and “aftercare-free landfills” (Fehringer, Rechberger, Pesonen, and Brunner, 1997). The incentive for this Austrian study is a new federal ordinance on landfilling that became effective in 1996 (Austrian Landfill Ordinance, 1996). The ordinance mandates that beginning in 2004, only wastes with a TOC <2–5% may be landfilled. The exact percentage depends on the type of landfill (e.g., monofill, landfill for construction waste, etc.). The reason for banning organic carbon in landfills is that organic carbon is transformed by microorganisms. The metabolic products are organic compounds that may be transferred to landfill leachates, and carbon dioxide and methane in landfill gas that contribute to global warming if not collected and treated properly. In addition, organic acids are produced that may mobilize heavy metals. Therefore, leachates of such reactor-type landfills are contaminated with a broad array of organic and inorganic pollutants. This requires treatment of the leachate over long periods (>100 years) and contradicts one of the Austrian objectives of waste management, which is to avoid shifting waste-related problems to future generations (aftercare-free landfills).

Because of the limit for organic carbon, treatment before landfilling is mandatory for most wastes such as MSW, sewage sludge, and construction debris. Combustion is an efficient means of transforming organic carbon to carbon dioxide. In order to ensure free choice of waste-treatment technologies, the landfill legislation allows an exemption from the TOC limit for the output material of mechanical–biological treatment facilities. These plants produce two fractions: (1) a combustible fraction that is mechanically separated and appropriate for further energy recovery and (2) a product derived from biological digestion. The biological degradation process cannot provide a residue with a TOC <5%, because persistent organic compounds such as plastics and lignin cannot be decomposed within months by microorganisms. Thus, an exception is stipulated for this fraction: it may be landfilled if the heating value is below 6000 kJ/kg. In contrast to the limit for TOC, which minimizes reactions in the landfill body and thus supports the objectives of waste management, the limitation of the heating value does not improve landfilling practice or reduce the need for aftercare. Rather, the exemption is based on political decisions. Both limits prevent direct landfilling of untreated MSW after 2004.

Some industry branches are eager to use combustible wastes as a substitute for fossil fuels. This helps to reduce costs of production, since wastes are usually cheaper than fuel. If the wastes are contaminated (e.g., with PCBs) or otherwise difficult to dispose of, they may even create revenue. Also, wastes made up of biogenic carbon are attractive fuels because they do not contribute to global warming.

The following treatment options are available for combustible wastes: incineration, cocombustion in industrial furnaces (using both conventional fuels and wastes), mechanical sorting, and biological digestion. All of these options have different environmental impacts and different contributions to the goals of waste management as stated in the Austrian Waste Management Act (AWG, 1990). In the case study ASTRA, various scenarios for the management of combustible wastes are developed and compared in view of the requirements of the new Landfill Ordinance and of the goals of the Waste Management Act.

3.3.2.1.1  Procedures

The ASTRA project consists of the following steps:

1. Selection of waste treatment processes and defining waste management systems and scenarios
2. Selection of substances
3. Selection of wastes
4. Establishment of mass balances (see Figure 3.45) and substance balances for the actual system
5. Development and selection of criteria to evaluate the scenarios
6. Development of an optimized scenario for improved management of combustible wastes (optimum assignment of wastes to treatment processes)
7. Establishment of total mass balance as well as substance balances for the optimized scenario
8. Comparison between actual system and optimized scenario

For brevity, not all steps of the comprehensive ASTRA study are presented here in detail. The only steps that are discussed are those relevant to the understanding of the results and implications of the case study.

Figure 3.45   Mass flows of combustible wastes through the system waste management in Austria (1995), 1000 t/year. (From Fehringer, R. et al., Auswirkungen unterschiedlicher Szenarien der thermischen Verwertung von Abfällen in Österreich (Project ASTRA). Vienna, Austria: Institute for Water Quality, Resource and Waste Management, Technische Universität Wien, 1997.)

Table 3.33   Goals of Waste Management and Assignment of Assessment Methods

Goals of Waste Management as Stated in the Austrian Waste Management Act

Assessment Methods and Criteria

1. Protection of human health and the environment
2. Conservation of energy and resources
3. Conservation of landfill space
4. Aftercare-free landfill

• 1. Critical air volume
• 2. Efficiency of utilization of the energy content of wastes
• 3. Volume reduction through treatment
• 4a. Total organic carbon in landfilled wastes
• 4b. Fate of substances on their way to “final sinks”

3.3.2.1.2  Selection and Development of Criteria to Evaluate Balances

The starting point is the goals of waste management as listed in the Waste Management Act:

1. To protect human health and the environment
2. To conserve energy, resources, and landfill space
3. To treat landfilled wastes so that they do not pose a risk to future generations

The latter goal is part of the precautionary principle. Since the long-term behavior of landfills is not known, future emissions have to be prevented by today’s waste treatment and immobilization. In general, these goals are quite abstract and therefore require focus: what are the indicators to decide if human health and the environment are protected?

Table 3.33 lists the criteria that have been applied in ASTRA. The chosen metrics or indicators are not absolute measures, since they cannot quantify the extent to which the goals of waste management have been reached. However, they do allow relative comparison of the actual situation with various scenarios, yielding statements such as “scenario X is Y% better than the actual situation.”

3.3.2.1.3  Assessment Methods and Criteria

1. 1. The critical air volume as used in ASTRA is adapted from the Swiss Eco-points approach. It is defined by the following equations:
3.12 $V i , c r i t = E i L i$
3.13 $V i , c r i t = ∑ i = 1 n V i , c r i t$
where Ei = emission of substance i into the air Li = concentration of substance i in ambient air Vi,crit = hypothetical air volume that is needed to dilute Ei to ambient air concentration for substance i The substance-specific critical volumes are added up to a final assessment indicator (Vcrit), which has its optimum at small volumes.
2. 2. The efficiency of utilization of waste energy content is calculated as follows:
3.14 $Efficiency = substituted fossil fuel [ J ⋅ yr − 1 ] energy content in waste [ J ⋅ yr − 1 ] ⋅ 100$
Fossil fuels can be substituted directly and indirectly by waste combustion. Direct substitution takes place when wastes replace fossil fuels, e.g., when coal is replaced by plastic waste to fire a cement kiln. It is assumed that wastes replace energy-equivalent units of fossil fuels. Strictly speaking, this is only true when the heating values of wastes and fossil fuels are similar (difference smaller than 20%). Indirect substitution is given when wastes are used in an incinerator to produce electricity and/or heat to feed into a network, conserving fossil fuel that would have been required without the MSW incinerator. An efficiency of 100% means that one energy unit of wastes replaces the equivalent energy amount of fossil fuels.
3. 3. Volume reduction by waste treatment is expressed as the difference in landfill space required for the various scenarios.
4. 4a. TOC of final wastes is assessed based on mass and substance balances.
5. 4b. “Fate of substances” means that each substance will finally be transferred to intermediate A and final B sinks. These sinks are
1. (A) Recycling products or other new secondary products (e.g., cement, bricks)
2. (A) The atmosphere
3. (A) The hydrosphere
4. (B) The lithosphere as an underground disposal facility
5. (A + B) The lithosphere as a landfill (see gray boxes in Figure 3.45) Sinks 1, 2, and 3 are intermediate sinks for most substances; sink 4 is designed as a final sink; and sink 5 is a sink that is leaching over very long periods of time. For each substance, a suitable sink must be defined. For example, only very minor amounts of cadmium should reach the atmosphere and the hydrosphere. Also, the transfer into recycled goods or cement is not desirable, since cadmium is not required in these products and has to be disposed of at the end of all life cycles. Cadmium in landfill poses a small long-term risk. The disposal in specially designed underground storage facilities that have not been in contact with the hydrosphere for millions of years (e.g., salt mines) represents a long-term solution with an extremely low risk of environmental pollution. Hence, from the point of view of finding an appropriate final sink for cadmium, the underground storage is the most preferred solution. For nitrogen, recycling as a nutrient and emission into the air as molecular nitrogen (not nitrogen oxide) are positive pathways and sinks. Most other fates such as nitrate in groundwater or NOx in air are considered negative paths. For chloride, transport in river systems to large water bodies such as oceans are acceptable solutions as long as the ratio of anthropogenic to geogenic concentrations and flows is small (e.g., <1%>). The criterion for fate of substances is expressed as the percentage of a substance that is transferred into appropriate compartments. The reference (100%) is the total flow of the substance in combustible wastes.

3.3.2.1.4  Total Mass Balance for the Actual Situation (1995)

The generation and actual flows of combustible wastes in Austria are assessed by analyzing statistics and studies that were commissioned by the authorities responsible for waste-management issues in Austria. As a working hypothesis, combustible wastes are defined as wastes having a heating value >5000 kJ/kg dry substance. This is the range where autarkic combustion is possible. The result is given in Table 3.34. The total amount of wastes is 39 million t/year. It is dominated by construction and demolition debris, including soil excavation. But only a small fraction of this category is combustible (2%). Altogether, about 22% or 8.5 million t/year are combustible wastes. The most relevant fractions are waste wood (41%) and wastes from private households and similar institutions (26%). Waste wood comprises bark, sawdust, chips of wood, and other minor fractions. Wastes from water purification and wastewater treatment mainly consist of municipal and industrial sludge and screenings from the sewer and wastewater treatment plants. Other nonhazardous wastes comprise all sorts of industrial wastes. The composition of this fraction is comparatively unknown. Better statistics are available for hazardous wastes. Generally, one can say that for every average Austrian, about 1 metric ton of combustible wastes is produced per year. Three-quarters of this amount accrues elsewhere (industry, infrastructure) and is not directly visible for the consumer.

Figure 3.45 shows the flows of combustible wastes in Austria in 1995. Approximately 40% (3400 kt/year) is used for feedstock recycling. This can be sawdust for chipboard production or wastepaper recycling. About 30% (2600 kt/year) is directly landfilled. After 2004, the disposal of this latter amount did not comply with the landfill ordinance of 1996. New methods of treatment and disposal were needed. Simple incinerators and boilers without advanced air-pollution standards utilize about 17% (1400 kt/year) of the combustible wastes for energy recovery. These plants are equipped with settling and baffle chambers, multicyclones, electrostatic precipitators (ESPs), or baghouse filters. The standard fuel is oil, coal, or biomass, and emission limits are not as stringent as for MSW incinerators. About 9% (770 kt/year) is incinerated in high-standard facilities. These plants are equipped with advanced air pollution control (APC) systems and easily surpass the most stringent emission regulations. Finally, some 3% (240 kt/year) is treated in mechanical–biological facilities.

Table 3.34   Total Waste Generation in Austriaa and Combustible Fractions

Combustible Portion of Wastes

Waste Category

Total Wastes, t/Year

%

t/Year

% of Total Combustible Fractions

Wastes from private households and similar sources

2,500,000

87

2,170,000

26

Construction and demolition debris including soil excavation

22,000,000

2

500,000

6

Residues from wastewater treatment

2,300,000

41

940,000

11

Waste wood

3,500,000

100

3,500,000

41

Other nonhazardous wastes

7,800,000

14

1,130,000

13

Hazardous wastes

1,000,000

22

220,000

3

Total

39,100,000

22

8,500,000

100

Notes:

a  8.1 million inhabitants.

3.3.2.1.5  Selection of Substances and Characterization of Wastes

The following substances are selected as indicators: carbon, nitrogen, chlorine, sulfur, cadmium, mercury, lead, and zinc. Carbon is selected because of the TOC limit that will apply beginning in 2004. Nitrogen is relevant as a potential nutrient and for the cement industry. Cement kilns are single sources (2.5%) of national NOx emissions, along with traffic (62%), other industries (17%), and home heating (10%) (Hackl and Mauschitz, 1997; Gangl, Gugele, Lichtblau, and Ritter, 2002). Wastes rich in nitrogen may increase this emission load. Compounds of chlorine, sulfur, and heavy metals are major air pollutants. The heavy metals are also of interest for their potential as resources. An overview of the content of the selected substances in combustible wastes is given in Table 3.35. The ranges are broad and show that there are both extremes of wastes: “clean” wastes that have lower contamination than fuel oil and wastes that show a significantly higher level of pollution than MSW.

Table 3.35   Substance Concentrations of Combustible Wastes and Comparison with Other Fuels, mg/kg Dry Matter

C

N

S

Cl

Cd

Hg

Pb

Zn

Mean

450,000

9100

2300

4300

5.7

0.8

230

520

Minimum

100,000

200

60

10

0.01

0.001

<1

1

Maximum

900,000

670,000

17,000

480,000

500

10

4000

16,000

MSW

240,000

7000

4000

8700

11

2

810

1100

Coal

850,000

12,000

10,000

1500

1

0.5

80

85

Fuel oil

850,000

3000

15,000

10

<1

0.01

10

20

3.3.2.1.6  Criteria for Optimized Assignment of Wastes to Treatment Processes

The variance in chemical composition requires tailor-made assignment of combustible wastes to treatment processes. Not every facility is qualified to treat any waste if the goals of waste management are to be reached. The following criteria were developed to decide upon waste treatment.

First, wastes having lower contaminant concentrations than average coal are appropriate for production processes such as cement kilns or brickworks. Concentrations are not determined per mass but per energy content of the fuel (e.g., mg/kJ). The reason is that 1 ton of waste does not necessarily replace 1 ton of coal. Rather, equivalent energy amounts are substituted by waste utilization. The rationale for this criterion is that it prevents products from becoming a sink, for example, for heavy metals. It is not clear whether elevated concentrations in concrete, bricks, asphalt, etc. have an impact on the environment. Thus, the precautionary principle is applied by this criterion, and contamination of products is banned. A second argument is that once substances are transferred into such products, they cannot be recovered again. The second criterion considers the impact on air quality by waste combustion. Emissions of state-of-the-art MSW incinerators are smaller, for some substances orders of magnitude smaller, than modern air pollution regulation demands. Thus, MSW incineration has proved to be environmentally compatible and serves as a reference. The criterion says that emissions from any waste-combustion facility must not exceed typical emissions from state-of-the-art MSW incineration. This can be expressed by the following equation:

3.15 $c max ⁡ = T C I T C C P ⋅ c M S W$

Transfer coefficients of state-of-the-art incineration (TCI) for relevant substances into the air are known from several investigations. Also, average concentrations in MSW (cMSW) are common. Typical values are given in Table 3.35. Transfer coefficients of the specific combustion process (TCCP) have to be determined by an MFA. The maximum allowable concentration of a substance in a waste to be burned in a specific combustion process is then cmax.

3.3.2.1.7  Results of the Optimized Scenario and Comparison with the Actual Situation

Applying these criteria to the actual situation yields a new optimized scenario. In Table 3.36, the optimized assignment of wastes to combustion processes is listed. The capacity for combustion has to be more than doubled from 2.1 to 5.0 million t/year. Some of the required plants already exist (cement, pulp and paper, etc.). Most of them could manage the assigned wastes with little or no process adaptation. These changes can be carried out comparatively quickly. On the other hand, new incineration plants with advanced APC technology and a total capacity of 2.8 million t/year have to be erected. This may take up to 5 years, including the permitting process, financing, planning, and engineering.

The improvement between actual and optimized situations can be seen when the aforementioned assessment criteria are applied to the materials balances (Figure 3.46).

• 1. The critical air volume calculated for NOx, SO2, HCl, Cd, Hg, Pb, and Zn is reduced by 43%. This is surprising because the quantity of combusted wastes is increased by 140%. The reason for this paradox is that in the actual situation, a comparatively small quantity of wastes is combusted in simple furnaces that lack adequate APC. The new scenario assigns all wastes to appropriate plants. Noncontaminated wastes are utilized in furnaces that have a lower (but sufficient) standard in flue-gas cleaning. “Dirty” wastes are treated in well-equipped combustion plants.

Table 3.36   Assignment of Combustible Wastes in an Optimized Scenario and Changes Compared with the Actual Situation

Optimized ScenarioCompared with Actual Situation
MSW incineration1,500,000+1,000,000
High-standard industrial combustion2,000,000+1,800,000
Hazardous waste combustion70,000±0
Wood industry585,000−30,000
Biomass cogeneration power station110,000+10,000
Pulp and paper industry550,000+31,000
Cement industry170,000+77,000
Total5,000,000+2,900,000

Figure 3.46   Comparison between status quo waste management and the optimized scenario based on selected criteria. (a) Critical air volume, (b) energy efficiency, (c) volume reduction, and (d) substance management and “final sink.”

• 2. The efficiency of utilization of the energy content of wastes is improved by 150%. In the optimized scenario, one energy unit of waste replaces the energy-equivalent amount of fossil fuel almost completely. The main reason for this progress is that wastes are no longer landfilled without energy recovery; no waste is processed in a mechanical–biological treatment plant anymore.
• 3. Consumption of landfill space is reduced by 80%. Again, the main reason for this improvement is the ban of direct landfilling of wastes and the abandonment of mechanical–biological waste treatment. Combustion reduces the volume of wastes by 80–98%, depending on the ash content of the specific waste.
• 4a. The TOC of all residues landfilled is below 3%. This is an important step away from reactor-type landfills to “final storage” landfills that require no aftercare.
• 4b. The percentage of substances that are transferred into appropriate final sinks is increased by 180%. This indicates a relevant improvement in substance management.

3.3.2.1.8  Conclusions

This case study shows that MFA facilitates goal-oriented waste management. General goals on a high hierarchic level, as stated in various waste-management acts (e.g., European Waste Framework Directive, Swiss Guidelines for Waste Management, and German “Kreislaufwirtschaftsgesetz”) can be translated into well-defined, concrete assessment procedures with appropriate criteria. ASTRA outlines a way for this to be achieved. Note that a single goal of waste management may require two or more assessment methods for comprehensive evaluation. MFA is used at several levels in the study:

1. To describe the actual situation of the system management of combustible wastes
2. To reveal deficits and develop criteria for waste assignment
3. To compile the optimized scenario
4. To demonstrate the differences between the actual and the optimized situation

The findings of the study reveal the capacities for new plants and may serve as a basis for planning and engineering. A next step is to assess costs (including uncertainties) for the scenarios. In fact, this is also done in ASTRA. The result is that the optimized scenario can be realized without significantly raising total costs for disposal (collection, separation, treatment, and landfill-ing). The main drawback of the optimized scenario, and also the main reason why this scenario will take a long time to be accomplished, is the following: A large proportion of the wastes that are landfilled at present will be incinerated in the future. For landfill owners and operators, this may cause a severe economic situation because the landfills will lose business. Considering that landfills are long-term investments with filling times between 25 and 50 years, it is clear that such a stern change cannot be pushed through in a short time. It is also clear that landfill operators will use every possible legal and economic means to postpone strategic changes endangering landfilling.

3.3.2.2  Case Study 13: PRIZMA

In the case study PRIZMA (German acronym for “positive list for utilizing of residues in the cement industry: methods and approaches”), the utilization of combustible wastes for energy recovery in cement kilns is investigated for Austria (Fehringer, Rechberger, and Brunner, 1999). In this country, production of cement requires about 10 million GJ/year to produce 3 million t/year of clinker that is further processed into cement. This amount of energy corresponds to some 400,000 tons of wastes with an average heating value of 25 MJ/kg. Today, wastes cover about 27% of the energy demand of the cement industry. Energy consumption for clinker production is quite high and represents a significant share of total production costs. For example, the European Cement Association estimates that energy accounts for 30–40% of the production cost of cement (CEMBUREAU, 1999). Hence, the cement industry strives to minimize energy-specific costs. One possibility is to reduce costs for fuel consumption. Wastes are alternative fuels. Compared with standard fuels such as natural gas, oil, or coal, they are less expensive. Other ways to reduce costs are to improve energy efficiency or to use different raw materials and technologies.

The Austrian cement industry has a long history of experience with energy recovery from wastes. Traditional waste fuels are used tires, waste oil, and solvents. Test runs have been carried out with sewage sludge, mixed plastics, and waste wood as well as meat and bone meal [as a result of the disposal crisis caused by bovine spongiform encephalopathy (BSE)], and others. Sorted fractions of MSW are also considered as a fuel alternative. The Austrian cement industry aims to cover 75% of its energy demand by wastes within a few years. Besides the expected cost relief, this goal contributes to the reduction of carbon dioxide emissions by the branch. Since direct landfilling of organic wastes was forbidden after 2004 in Austria, the cement industry supported the provision of the required capacities for treatment. On the other hand, not all wastes are appropriate for the cement process. Wastes with high contamination of heavy metals may lead to environmentally incompatible emissions and a polluted product (cement, finally concrete). Hence, operators as well as authorities want to have clear regulations specifying which kinds of wastes are appropriate for energy recovery. One possibility for establishing such an instrument is to generate a so-called positive list. The positive list specifies and characterizes waste types that are appropriate for recovery in cement kilns. The objective of PRIZMA is to develop criteria to establish such a list.

3.3.2.2.1  The Cement Manufacturing Process

The prevailing technology for cement manufacturing in Austria is the cyclone preheater type. A scheme of such a facility is displayed in Figure 3.47. The heart of each cement factory is a massive steel tube up to 100 m in length and up to 8 m in diameter. It is slightly inclined to the horizontal (3 to 4°) and slowly rotates at about one to four turns per minute. The main raw materials for the feed of the kiln are limestone, chalk, marl, and corrective materials (e.g., ferrous materials). The chemical properties of these materials and the desired properties of the clinker govern the correct mixture. Mixing is an important step in the process to ensure an even distribution of the properly proportioned components of the raw material so that the clinker will be of a uniform quality. The raw material is ground in a mill, from where it is pneumatically boosted into a mechanical predeposition (baffle) and a subsequent ESP. There, the raw material is collected and conveyed into a storage silo. Afterwards, the raw material runs through four stages: evaporation and preheating, calcining, clinkering, and cooling.

Evaporation and preheating remove moisture and raise the temperature of the raw material. This process takes place in the cyclones, where raw material counterflows the hot off-gas from the kiln. The raw material enters the kiln at the back (the upper end of the kiln), and gravity and the rotation of the kiln allow the mix to flow down the kiln at a uniform rate through the burning zone. The tube is lined with refractory bricks to avoid heat damage of the kiln. Calcining takes place at 600 to 900°C and breaks the calcium carbonate down into calcium oxide and carbon dioxide. Approximately 40% by weight of the raw material is lost by this process. Clinkering completes the calcination stage and fuses the calcined raw material into hard nodules resembling small gray pebbles. The clinker leaves the kiln at the front (the lower end of the kiln) and falls onto a reciprocating grate, where it is cooled. The primary fuel is introduced and burnt at the same end of the kiln. The flame is drawn up the kiln to the burning zone, where the heat intensity is highest and fusion of chemicals in the raw material takes place. Hot combustion gases continue to flow up the kiln and exit from the back end. Secondary firing at the upper end of the kiln maintains the energy-intensive calcination process. Product temperatures in the burning zone are around 1450°C. The primary flame has a temperature of 2000°C. The cooled clinker is stored in a silo. Cement is produced by grinding of the clinker and blending with gypsum and other materials (e.g., fly ashes from coal combustion) to produce a fine gray powder. The last stage is bagging of cement and preparing the product for transportation. The cooled flue gas (heat transfer to raw material in the cyclones) is cleaned by the ESP and emitted via a stack.

Figure 3.47   Scheme of a cement kiln (cyclone preheater type).

Generally, for any substance, there are only two ways to enter and to leave the process: enter via raw materials or fuels and exit via off-gas or the product. The partitioning for any substance A between raw materials and fuels can be determined by measurement. The result will be that X% of A enters the process via fuels and Y% via raw materials, with X + Y = 100%. For the off-gas and the product, only total amounts of substance A can be determined. It is not possible to say that $X ˜$

of A stems from fuels and Ỹ stems from raw material in the clinker (again, $X ˜ + ϒ ˜ = 100%$ ). The same applies to the off-gas. Only some qualitative information is available about the behavior of substances in a clinker manufacturing process. However, for the given problem, which is waste combustion in cement kilns, it is mandatory to know how fuel-borne substances (i.e., substances that enter the process via fuel) behave in the process.

A special characteristic of the process is that two kinds of cycles evolve:

1. The so-called inner cycle arises when a substance I vaporizes in the kiln. It is then transferred to cooler parts in the cyclone, where substance i may condense at the surface of raw material particles. So substance i travels back to the hot kiln, where it vaporizes again. The cycle is closed and built up until some kind of equilibrium is established (theoretically). Operators try to break such cycles by bypassing the cyclones.
2. The so-called outer cycle develops because cyclones cannot intercept fine particles (<5 μm). But interception in the cyclones is a prerequisite for raw material particles to enter the kiln and leave the process as clinker. Fine particles are carried by the flue gas to the ESP. There they are removed from the flue gas with high efficiency (>99%) and reenter the cyclones, where they are, again, not intercepted. The loop is closed, and substance built up.

These cycles make it difficult to establish closed substance balances for the process and to predict the behavior of substances in the process. However, the following working hypothesis can be put forward: Fuel-borne substances (e.g., heavy metals) are predominantly embedded in an organic matrix. Organic substances are destroyed in the flame, which is an area with temperatures around 2000°C. This means that inorganic, volatile fuel-borne substances will vaporize to a high extent. Contrary to fuels, heavy metals in the raw material are fixed in a mineral matrix. The raw material is heated up to 1450°C, and it can be assumed that not all metals will vaporize. A portion will remain in the solid phase and contribute in the clinkering process (sintering). In other words, the possibility for a metal to reach the gaseous phase is higher for fuel-borne substances than for substances descending from raw material. When the flue gas is cooled down (in the cyclones), condensation of metals on particle surfaces (raw material, ashes) takes place. This process happens at the same rate for all metals, regardless of their origin (fuel or raw material). Let us assume that the partitioning of a substance χ between fuel and raw material is X:Y. Then the mentioned vaporization and condensation processes imply that there is a different ratio $X ˜ / ϒ ˜$

of χ in particles in the cyclone with $X / ϒ < X ˜ / ϒ ˜$ . Fine particles will pass the cyclones, and a small fraction will also pass the ESP. Evidence of this process is that par-ticulates emitted from cement plants are enriched with heavy metals compared with raw materials (see Table 3.37). The conclusion is that fuel-borne and raw-material-borne substances show different behavior in the process and therefore have different transfer coefficients. However, for the problem of waste combustion, it is crucial to know the transfer coefficients for fuel-borne substances.

Table 3.37   Mean Substance Concentrations in Raw Material and Emitted Particulates

Cl

Cd

Hg

Pb

Zn

Raw material, mg/kg

150

0.15

0.15

15

37

Emission, mg/kg

46,000a

8

2000a

400

150

Enrichment

300

50

13,000

27

4

Notes:

a  Gaseous and solid emissions are related to the total mass of emitted particulates.

Figure 3.48   Assumptions for the fate of fuel-borne heavy metals in the clinker manufacturing process. The left/right assumption yields the lower/upper limit for the transfer coefficient of fuel-borne substances into the atmosphere.

3.3.2.2.2  Assessment of Transfer Coefficients

The uncertainty concerning transfer coefficients for fuels leads to the following approach: A range for transfer coefficients is established by determining hypothetical extreme values (see Figure 3.48). For the lower limit, it is assumed that the partitioning between fuel-borne and raw-material-borne substances is the same in the off-gas and the product. Based on the aforementioned considerations, this will underestimate the transfer of fuel-borne substances into the off-gas. The upper limit is given by the assumption that the emission only contains fuel-borne substances. This assertion certainly overestimates the influence of fuels for emissions. On the other hand, this is a reliable upper limit, since higher transfer coefficients are not possible. The real transfer coefficient has to be somewhere between these extremes. In cases where the range is large (e.g., one order of magnitude), it is safe to apply the upper limit, i.e., the higher value.

3.3.2.2.3  Criteria for Waste Fuels

Which substances should be incorporated into a “positive list” that defines wastes suited for cement kilns? In Section 3.3.2.1, criteria are presented for the selection of substances with regard to waste combustion. The list can be shortened for cement manufacturing because carbon compounds are most efficiently destroyed in the cement kiln. With the exception of carbon dioxide, emissions of carbon compounds are very small. The emission of nitrogen oxide is a problem of cement manufacturing, but this has little to do with waste recovery. The high temperatures and long residence times of gases in the process, which are essential for efficient mineralization of carbon compounds, are responsible for nitrogen oxide formation. The nitrogen content of wastes plays only a minor role in the formation of nitrogen oxide. Sulfur from wastes is efficiently contained into clinker. Recorded emissions of sulfur dioxide stem from certain kinds of raw material. As in the case of nitrogen, these emissions are not a result of waste combustion. Chlorine poses a limit for the quality of the product and may cause blockage in the cyclones. Therefore, it has to be part of the positive list. The heavy metals cadmium and mercury are chosen because of their toxicity and volatility. Lead and zinc are, on one hand, also potentially toxic but, on the other hand, are important resources, too.

Waste as a fuel for clinker manufacturing influences input as well as off-gas and clinker. Therefore, three criteria—A, B, and C—are developed for the positive list:

Criterion A deals with the off-gas and has been described in Section 3.3.2.1. The criterion says that emissions from clinker manufacturing must not exceed typical emissions from state-of-the-art incineration. This can be expressed by Equation 3.16:

3.16 $c max ⁡ = T C I T C C M ⋅ c M S W$

TCI is the transfer coefficient of state-of-the-art incineration (known). Also, the average concentrations in MSW (cMSW) are commonly known (see Table 3.38). Transfer coefficients for clinker manufacturing (TCCM) have to be determined according to aforementioned considerations about a reliable range for transfer coefficients. The result of criterion A, cmax, is the maximal allowable concentration of a substance in a waste.

Criterion B controls the quality of the product clinker. It is based on the principle that anthropogenic material flows must not exceed the natural fluctuations of geogenic flows (see Chapter 2, Section 2.5.8). The criterion now considers raw material chemically as a geogenic flow. As any natural material, raw materials show a certain variance in chemical composition. To apply the criterion, the clinker composition is assessed for (1) average and (2) maximal raw material concentrations. Criterion B says that the changes in clinker concentration caused by waste recovery must not exceed the calculated geogenic variance. For both scenarios 1 and 2, clinker is produced with an average fuel mix consisting of coal (52%), oil (21%), natural gas (3%), used tires (6%), plastics (5%), waste oil (9%), and others (percentages are based on energy equivalents). The calculation of criterion B requires the following assumptions: The mass ratio between raw materials (RM) and fuels (F) is ca. 10:1. Transfer coefficients for substances stemming from raw materials and fuels are identical. The loss of carbon dioxide is 40%, and the ash content of the fuel is considered to be negligible (usually ca. 1% of raw material mass). Then the concentration of clinker (Cl) is

Table 3.38   Data for Criterion A: Transfer Coefficients of the Reference Technology Incineration, Typical Concentrations in the Reference Waste (MSW), and Extreme Transfer Coefficients for Off-Gas of a Cement-Manufacturing Process

Cl

Cd

Hg

Pb

Zn

TCI

0.0005

0.0005

0.02

0.0001

0.0002

cMSW

10,000

10

2

500

1000

TCCM,min

0.01

0.0002

0.4

0.0002

0.0001

TCCM,max

0.02

0.0004

0.8

0.0008

0.0001

3.17 $C C l = ( c F ⋅ 1 + c R M ⋅ 10 ) T C C l 10 ⋅ ( 1 − 0.4 )$

The data needed to calculate criterion B are summarized in Table 3.39. The result of criterion B yields the maximum load of a substance that can possibly be added to the clinker by waste recovery. The load (e.g., mg of a substance per ton of clinker) gives the dependence between the total mass of recovered wastes and the substance concentration in the waste. The result is a curve (see Figure 3.49).

Criterion C considers the input into cement manufacturing. If the total national consumption is assigned a value of 100%, then combustible wastes contain roughly 40% of cadmium and mercury (see Table 3.40). Hence, combustible wastes are important carriers of some heavy metals. This is one reason why waste management plays such an important role for the total turnover of several heavy metals. The consequential question for the cement industry is, Which share of these substances shall enter the processes and finally end up in the product cement? There is no uniform answer to this question. Some cement manufacturers do not want their products associated with hazardous materials and thus are cautious in using contaminated wastes as a fuel. Other manufacturers see a chance for economic advantage over their competitors by using inexpensive waste-derived fuel and use large amounts of wastes. In Table 3.40, results are presented assuming that cement manufacturers take over 15% of the metals contained in combustible wastes. As for criterion B, the result is a curve showing the dependency between total mass of wastes and substance concentration in the wastes (see Figure 3.49).

Table 3.39   Calculation of Maximum Allowable Load on Clinker through Waste Recovery

Cl

Cd

Hg

Pb

Zn

Mean concentration in fuel mix, mg/kg

1100

0.9

0.4

60

65

Mean concentration in raw material, mg/kg

150

0.15

0.15

15

37

Maximum concentration in raw material, mg/kg

400

0.5

0.5

42

110

0.99

0.99

0.6

0.99

0.99

430

0.4

0.19

35

72

840

1.0

0.54

79

190

Maximum load through waste recovery, mg/kg

410

0.6

0.35

45

120

Maximum load through waste recovery, t/year

1200

1.7

1.1

130

360

Note: Based on a clinker production of 3 million t/year.

Figure 3.49   Results of criteria A, B, and C for Hg serve to support decisions regarding the utilization of wastes in cement kilns (Fehringer, Rechberger, and Brunner, 1999). Criterion A, emissions; criterion B, product quality; criterion C, dilution of metals.

Table 3.40   Estimated National Consumption of Selected Substances, Content in Combustible Wastes, and Maximum Flow into Cement Manufacturing

Cl

Cd

Hg

Pb

Zn

National consumption, t/year

450,000

80

10

32,000

43,000

Combustible wastes, t/year

30,000

36

3.9

1600

3900

In combustible wastes, %

6.6

45

39

5

9

Input into cement, %

15

15

15

15

15

Input into cement, t/year

4500

5.4

0.6

240

590

3.3.2.2.4  Results

Criteria A to C are calculated for the selected substances and provide either maximum concentrations for wastes or maximum amounts of substances that can be transferred into clinker. Required parameters for calculation are transfer coefficients of waste-borne substances into the off-gas and into the clinker. They should be determined for each cement plant separately, as results may vary considerably among different technologies. In Figure 3.49, the result is given for mercury. Consider a waste (α) having a mercury concentration of 2 mg/kg. Criterion B allows waste recovery of 500,000 t/ year. This means that, in practice, mercury does not pose a limit for the quality of clinker. Self-restriction of the cement industry (criterion C) still allows 300,000 t/year of waste α. A really severe limit poses criterion A: only wastes having Hg concentrations lower than 0.1 mg/kg are qualified as a waste fuel. This reduces the amount of potentially available wastes considerably. With respect to mercury, it should be evaluated whether the investment costs for improving APC technology are paid back within an acceptable time frame by the savings due to a cheaper waste fuel.

3.3.2.2.5  Conclusion

Case study PRIZMA exemplifies how MFA can be used to establish environmental regulations. The proposed criteria consider system-specific as well as plant-specific constraints. Criteria A and B guarantee that waste utilization in cement kilns does not pollute the atmosphere (and subsequently the soil) or lower the quality of clinker. The extended systems approach assures that only a limited amount of resources is transferred into cement and concrete (criterion C). Note that the selected substances are not required in cement and are lost for recovery and recycling. A limit for this kind of sink is therefore reasonable. Waste recovery in state-of-the-art cement kilns that fulfill criteria A, B, and C can be considered as environmentally compatible. Thus, the utilization of wastes in cement kilns can be a valuable contribution to goal-oriented waste and resource management.

3.3.2.3  Case Study 14: Recycling of Cadmium by WTE

In mineral ores of commercial value, cadmium is usually associated with other metals. Greenockite (CdS), the only cadmium mineral of importance, contains zinc, sometimes lead, and complex copper–lead–zinc mixtures. Hence, zinc and lead producers have no choice, and they usually produce cadmium, too. Most cadmium (>80%) is produced as a by-product of beneficiating and refining zinc metal from sulfide ore concentrates. An estimated 90–98% of the cadmium present in zinc ores is recovered in the zinc extraction process. About 3 kg of cadmium is produced for every ton of refined zinc. Small amounts of cadmium, about 10% of consumption, are produced from secondary sources such as baghouse dust from electric arc furnaces (EAFs) used in the steelmaking industry and the recycling of cadmium products. Total world production of cadmium in 2000 was about 19,300 tons (US Geological Survey, 2001a).

The International Cadmium Association has made the following estimates of cadmium consumption for various end uses in 2001: batteries, 75%; pigments, 12%; coatings and plating, 8%; stabilizers for plastics and similar synthetic products, 4%; and nonferrous alloys and other uses, 1% (U.S. Geological Survey, 2001b). Utilization of cadmium in developed countries is estimated at between 5 and 16 g/capita/year (Llewellyn, 1994; Bergbäck, Johansson, and Mohlander, 2001; US Geological Survey, 2001b). Currently, annual consumption of cadmium amounts to about 20,000 tons. In contrast to most other metals, production does not show an increasing trend, most likely due to increasing regulatory pressure to reduce or even eliminate the use of cadmium. This is a result of the growing awareness of cadmium being potentially toxic to humans and of the risks presented by its accumulation in the environment. Unlike other heavy metals (e.g., zinc, selenium), cadmium is not essential to the biosphere and has no known useful biological functions. It accumulates in the kidneys and liver and affects protein metabolism, causing severe disorder, pain, and even death. In addition, it may cause a variety of severe damages such as lung cancer (Heinrich, 1988; Waalkes, 2000) and bone diseases (osteomalacia and osteoporosis) (Verougstraete, Lison, and Hotz, 2002).

In the input of waste to energy (WTE) plants, the typical concentration of cadmium is between 8 and 12 mg/kg (Brunner and Mönch, 1986; Schachermayer, Bauer, Ritter, and Brunner, 1995; Morf, Brunner, and Spaun, 2000; Verougstraete, Lison, and Hotz, 2002). This is about 50 times the average concentration found in the Earth’s crust (0.2 mg/kg). When MSW is landfilled, cadmium is quite immobile during the anaerobic phase but can be mobilized during aerobic periods by organic acids and reach the groundwater. Cadmium and some of its compounds such as chlorides have low boiling temperatures (Cd, 765°C; CdCl2, 970°C). Therefore, cadmium belongs to the group of atmophilic elements like Hg, Tl, Zn, and Se. In combustion processes, these substances tend to volatilize and escape via the flue gas from the combustion chamber.

The average generation of MSW in Europe is between 150 and 400 kg/capita/ year. In the United States, currently about 700 kg/capita/year is collected. There are several reasons to explain the differences in quantities. First, the term MSW is defined operationally. MSW usually designates all mixed wastes that are collected at the curbside on a daily, weekly, or biweekly basis. For example, in some areas, bulky wastes are collected separately and therefore not included in MSW data. In other areas, waste containers are large and allow collection of MSW together with bulky wastes. MSW includes mixed wastes from private households and may also include wastes from the service sector and small shops and companies. Decisive factors for waste generation include the kind of wastes the collector accepts, the collection frequency, the size of the bins or containers for collection, and how statistical data are compiled. A second reason for differences is the extent of recycling. Separate collection of paper, biowaste, glass, metals, etc. may reduce the weight of waste collected as MSW up to 50%. Third and fourth reasons are differences in lifestyle (e.g., small or large family and household size, packed food versus open food, etc.) and purchasing power of the consumer. At an average MSW generation rate of 250 kg/capita/year, about 2.5 g/capita/year of cadmium is collected via MSW. This is about 25% of the average national per capita consumption of cadmium. The remaining 75% is mostly incorporated in goods with long residence times that will enter waste management in the future. A smaller part of cadmium is expected in other wastes (estimated 20%, mainly contained in industrial wastes and the combustible fraction of construction and demolition waste), and a small quantity is lost to the environment via emissions and fugitive losses.

Modern WTE plants are equipped with advanced APC devices. As mentioned before, cadmium is transferred to the flue gas during incineration and removed by the APC devices. During off-gas cooling in the heat exchanger, volatile cadmium is condensed on very small particles that offer a large surface area. Consequently, more than 99.9% of cadmium can be removed by particle filters such as ESPs or fabric filters if these filters have been designed to capture small particles. Remaining quantities are removed upstream in wet scrubbers with high-pressure drops (venturi scrubbers) or adsorption filters. Very small amounts are emitted via the stack (<0.01°%). Figure 3.50 shows transfer coefficients for cadmium in a state-of-the-art MSW incinerator.

Typical Cd concentrations in incinerator fly ash range from 200 to 600 mg/ kg. In comparison, filter dust from EAFs that is recycled contains 500 to 1000 mg/kg Cd (Donald and Pickles, 1996; Stegemann, Caldwell, and Shi, 1997; Xia and Pickles, 2000; Youcai and Stanforth, 2000; Jarupisitthorn, Pimtong, and Lothongkum, 2002). This shows that incinerator fly ash could be used for cadmium recovery.

Thermal treatment of bottom ash and/or fly ash can improve the potential for recovery. For example, adequate thermal treatment of incinerator bottom ash results in three products:

1. A silicate product, with an average cadmium concentration similar to the Earth’s crust that can be utilized for construction purposes
2. A metal melt containing mainly iron, copper, and other lithophilic metals (metals of high boiling points)
3. A concentrate of atmophilic metals

Applying such technologies to ashes makes cadmium and other metals accessible for efficient recovery. For cadmium, recycling efficiencies from MSW up to 90% are realistic (Figure 3.51). Another possibility is to use thermal processes, not to produce a concentrate for recovery but to immobilize metals in a ceramic or vitreous matrix (vitrification). Such residues may come close to final storage quality (Baccini, 1989).

Figure 3.50   Transfer coefficients for cadmium in a state-of-the-art WTE plant. (From Schachermayer, E. et al., Messung der Güter- und Stoffbilanz einer Müllverbrennungsanlage, Monographien Bd. 56, Bundesministerium für Umwelt, Vienna, Austria, 1995. With permission.)

Figure 3.51   Recycling of cadmium by WTE of MSW. Percentage of flows of Cd in MSW, other wastes, and the rest may vary according to technological and economic situation.

Today, most of the obsolete cadmium enters landfills, where it remains a potential hazard for generations. The advantage of recycling is that the consumption of primary cadmium is reduced. Thus, the quantity of cadmium that enters the anthroposphere is reduced, facilitating the management and control of this resource. Technologies to immobilize cadmium are required to safely dispose of the large amounts already in the anthroposphere. The new task is to continually collect and transform cadmium that is stored in infrastructure and long-living goods into a form where it can be safely stored within the anthroposphere. Note that for both options, a concentration step is indispensable; thermal processes are proven technologies that can achieve such concentrations.

3.3.2.4.1  Introduction

The purpose of case study 15 is twofold: First, the results demonstrate how MFA on the level of substances can be used to support waste management decisions on a city level. And second, it is used to explain a waste management strategy directed toward “clean cycles” and “safe final sinks.” The object of investigation of this case study is PBDEs, a group of chemical compounds called polybrominated diphenyl ethers (Figure 3.52). They pose a challenge in plastic waste management because of their hazardous properties that impede polymer recycling.

PBDEs consist of two benzene rings linked by an oxygen bridge (diphenyl ether):

Figure 3.52   Poly brominated diphenyl ether, PBDE [C12H(10−x)BrxO(x=1,2,...,10=m+n)]

Each benzene ring can be substituted with zero to five bromine atoms in five possible positions, yielding altogether 209 congeners (i.e., structurally related compounds). PBDEs are classified with respect to the number of bromine atoms, such as penta-, hexa-, hepta-, or octa-PBDE. On the one hand, they are useful chemicals and thus are widely applied as flame retardants for plastic materials and present in a wide array of products such as transportation vehicles (cars, airplanes), construction materials, home appliances, furnishings, electronic devices, thermal insulation foams, textiles, and much else. On the other hand, some of the congeners have been proven to pose a serious health hazard for humans and the environment. This is particularly the case for PBDEs averaging one to five bromine atoms. These lower-brominated PBDEs are regarded as more hazardous because they bioaccumulate, affect hormone levels in the thyroid gland, and have been linked to reproductive and neurological risks. Thus, the so-called Stockholm Convention (Stockholm Convention, 2004), which has the objective to protect human health and the environment from hazardous persistent organic pollutants (POPs), has restricted the production of some PBDEs (Sindiku et al., 2014) as well as produced recommendations about the management of wastes containing PBDEs (UNEP, 2015b). PBDEs are commonly used as commercial mixtures of several congeners and not in a pure form. Hence, often, the short form cPentaBDE is used, meaning commercially available pentabrominated diphenyl ether.

At the end of the lifetime of products containing PBDEs, these are either recycled or disposed of in incinerators and landfills. In order to fulfil the goals of waste management, namely (1) protection of human health and the environment and (2) resource conservation, information about the pathways of PBDEs from sources (industrial synthesis) to sinks (thermal destruction, landfilling) is instrumental: it is necessary to know which stocks of PBDEs have been accumulated in the past and are which still present, and which hazardous congeners will reach waste management by which collection system (separate collection of electronic wastes, plastic materials, construction wastes, and end-of-life vehicles). Also, the fate of PBDE-containing wastes during recycling and waste treatment must be known. MFA can supply and link such information about stocks and flows of individual congeners from the anthropogenic metabolism to the environment.

In case study 15, flows and stocks of commercial pentabrominated and octabrominated diphenyl ether (cPentaBDE, cOctaBDE) have been studied on a city level (Vienna, Austria) (Vyzinkarova and Brunner, 2013). The city level was chosen for analysis because of the following:

1. Cities are major hotspots of PBDE emissions. Concentrations on sites close to emission sources, such as the soil adjacent to urban roads (Luo et al., 2009; Gevao et al., 2011) and soils and sediments in the vicinities of landfills and wastewater treatment plants, are higher than at remote sites (Oliaei, 2010).
2. Flows of PBDEs to the hinterland are uncovered, and thus, the dependency of a city on its hinterland for disposing of hazardous substances becomes apparent.
3. Strategies for waste management are determined by the municipality and other urban stakeholders responsible for waste collection, recycling, and disposal. Often, the municipality controls by law the means to manage all hazardous as well as some beneficial waste materials within the city boundaries.
4. Because of the statistical data collected on the level of a city, such as MSW generation, collection rate, treatment capacities and recycling rates, and emissions from treatment processes, municipalities are often able to supply information about flows and stocks of goods and substances within their regime.

3.3.2.4.2  Objectives

The specific objectives of the case study are (1) to identify sources, pathways, stocks, and sinks of cPentaBDE and cOctaBDE in the city of Vienna, Austria, (2) to determine the fractions that either are recycled or reach appropriate final sinks, and (3) to develop recommendations for waste management ensuring a “clean cycles” and “safe final sink” strategy (Kral et al., 2013). Such a strategy demands that products from waste recycling are of high quality containing very low amounts of hazardous substances (clean cycles) and that the hazardous substances removed from the cycle are completely destroyed (e.g., thermal destruction by incineration) or are disposed of in a long-term safe storage without aftercare. The latter two processes, incineration and safe storage without aftercare, are called final sinks because there are no further flows of these materials into either the anthroposphere or the environment.

The motivation for these objectives stems from two key directives of the European regulation on cPentaBDE and cOctaBDE (European Council, 2003a, 2015). These regulations demand removal of pollutants as a fundamental rule for treatment. They prohibit recycling of waste electrical and electronic equipment (WEEE) containing cPentaBDE and cOctaBDE with more than 1 g/kg. On the other hand, the European waste hierarchy defines recycling for waste plastics as the preferred option. Thus, the art of recycling WEEE and other wastes containing PBDE is to remove PBDEs from such plastic wastes down to a concentration of 1 g/kg. Industrial efforts to achieve this goal with commercial mixtures are underway; it remains to be seen how successful they will be.

3.3.2.4.3  Methods and Data

To reach the objectives, substance flow analysis was applied to previously published data about flows and stocks of cPentaBDE and cOctaBDE (Morf et al., 2003; Tasaki et al., 2004). No additional laboratory analyses were performed. For modeling flows and stocks, the software STAN was used taking uncertainties into account. Scenario analysis was applied for addressing the issues of missing data and large uncertainties.

For MFA of cPentaBDE and cOctaBDE, the following system boundaries have been selected: the border in space was the administrative area of the municipality of Vienna, and the boundary in time was the year 2010. The flows and stocks were assessed for both levels of goods and of substances. Goods containing PBDEs are summarized in Table 3.41.

Table 3.41   Former Uses of Commercial Mixtures of PentaBDE and cOctaBDE in Various Sectors

Compound and Sector

Polymer

Application

Mass Flow Estimates

cPentaBDE Vehicles

PUR

Upholstery of seats, ceiling, headrest, textile back-coating

PUR foam in vehicles and construction accounts for 90–95% of total cPentaBDE use

Construction

PUR

Insulation foam

Major use

PVC

Duroplastic sheeting

Minor use

Other

Various

Textiles, printed circuit boards, cable sheets, conveyor belts, etc.

Other applications less than 5% of total use

cOctaBDE Vehicles

HIPS

Dashboard and steering wheel

Estimates range from minor to major use

PBT

PA

Construction

PE

Thermoplastic sheeting

Minor use

EEE

ABS

WEEE categories 3 and 4, with focus on CRT computer monitors and TVs

Major use, estimated up to 95% of total cOctaBDE use in the EU

Source: Based on UNEP. (2015a). Draft guidance for the inventory of polybrominated diphenyl ethers (PBDEs) listed under the Stockholm Convention on Persistent Organic Pollutants. Retrieved from http://chm.pops.int/Implementation/NIPs/Guidance/GuidancefortheinventoryofPBDEs/tabid/3171/Default.aspx; Morf, L. S. et al., Selected polybrominated flame retardants, PBDEs and TBBPA, substance flow analysis. Bern, Switzerland: Swiss Federal Office for the Environment, 2003; flame retardants, PBDEs and TBBPA; Leisewitz, A., and Schwarz, W., Erarbeitung von Bewertungsgrundlagen zur Substitution umweltrelevanter Flammschutzmittel Band II: Flammhemmende Ausrüstung ausgewählter Produkte—anwendungsbezogene Betrachtung: Stand der Technik, Trend, Alternativen. Dessau, Germany: Umweltbundesamt, 2000; Lassen, C., and Løkke, S., Brominated Flame Retardants—Substance Flow Analysis and Assessment of Alternatives. Copenhagen, Denmark: The Danish Environmental Protection Agency, 1999.

Note: ABS, acrylonitrile butadiene-styrene; HIPS, high-impact polystyrene; PA, polyamide; PBT, polybutylene terephthalate; PE, polyethylene; PUR, polyurethane; PVC, polyvinyl chloride. Mass flow estimates vary in different sources, especially in case of cOctaBDE and are therefore only indicative.

The following three main processes have been included in the STAN model: consumption, waste management, and environment (see Figure 3.53).

Each of these main processes acts as a subsystem and is subdivided into further processes. The process consumption serves to quantify anthropogenic stocks and flows of POP-PBDEs in goods that are in use, such as construction materials, vehicles, and electrical and electronic equipment (EEE), and it includes consumer emissions, too. Because of out-phasing of some of the brominated flame retardants, the processes construction and vehicles have no imports. With only exports, stocks are decreasing today, with transfers either to waste management, which is the biggest flow, or to the environment, a much smaller flow. Some of the PBDEs are recycled by waste management. It is not well known into which recycling products these flows are directed. In Figure 3.53b, all flows of recycled PBDEs are inputs to the process use of EEE. In reality, this may be different, with probably significant fractions being directed toward constructions and vehicles. Nevertheless, for an overall picture of the three main processes, the allocation of recycled PBDEs to the subsystem consumption is sufficient and correct. The flow of recycled PBDEs is of particular concern because it prolongs their lifetime. Thus, despite the intention of the authorities to out-phase POP-PBDEs, and regardless if the recycled PBDEs are contained in vehicles, construction materials, or EEEs, the consumer is exposed to these hazardous substances long after they have been abandoned.

PBDEs are contained in a variety of waste materials that are collected and in part treated in Vienna (see also Table 3.41): WEEE [CRT-PCs (computer monitors), CRT-TVs (televisions)]; plastic construction wastes (polyurethane, polyvinylchloride, and polyethylene), and end-of-life (EOL) vehicles. With regard to WEEE, 10 categories have been defined, with category 3 (information technology and telecommunication equipment) and category 4 (consumer equipment and photovoltaic panels) being relevant with respect to POP-PBDEs (Wäger, Schluep, Müller, and Gloor, 2011; UNEP, 2015a). For this case study, categories 3 and 4 have been taken into account. Because EOL vehicles are not treated within Vienna but are exported beyond the systems boundaries, they are not included in the subsystem waste management.

Imports into the “environment” consist mainly of consumer emissions into the air, from where they are subsequently transferred by dry and wet deposition to the soil and hydrosphere, including municipal wastewater. Literature values for consumer emissions are available in the literature and range from 0.054% of the stock for cOctaBDE to 0.39% for cPentaBDE (Morf et al., 2003; UNEP, 2015a). In this study, the existing PBDE stocks were multiplied by these emission factors. Because of the small contribution of the hydrosphere, sedimentation, and other minor flows, only the two major processes atmosphere and soil have been taken into account quantitatively within the subsystem environment. For more information about the calculations of emissions in Vienna, see Vyzinkarova and Brunner, 2013.

Figure 3.53   Model of flows and stocks of PBDEs in the city of Vienna, 2010. (a) Total system, and (b) subsystem consumption. CRT-PCs: cathode ray tube computer monitors; CRT-TVs: cathode ray tube televisions; EEE: electrical and electronic equipment; PE: polyethylene; PUR: polyurethane; PVC: polyvinyl chloride; WEEE: waste electrical and electronic equipment; WEEE-3: WEEE category 3; WEEE-3r: WEEE category 3 excluding screen devices (“rest”); WEEE-4: WEEE category 4; WEEE-4r: WEEE category 4 excluding screen devices (“rest”); WWTP: wastewater treatment plant. (c) Subsystem waste management (including Vienna Hinterland) and (d) subsystem environment. CRT-PCs: cathode ray tube computer monitors; CRT-TVs: cathode ray tube televisions; EEE: electrical and electronic equipment; PE: polyethylene; PUR: polyurethane; PVC: polyvinyl chloride; WEEE: waste electrical and electronic equipment; WEEE-3: WEEE category 3; WEEE-3r: WEEE category 3 excluding screen devices (“rest”); WEEE-4: WEEE category 4; WEEE-4r: WEEE category 4 excluding screen devices (“rest”); WWTP: wastewater treatment plant.

The three key Austrian goals for waste management are protection of human health and the environment, resource conservation, and “aftercare-free” waste management practice. The latter means no landfills requiring aftercare for decades to centuries and no recycling concepts that cycle hazardous substances so that the next generation has to take care of these risks. A large share of PBDEs and other plastic additives have entered waste management in the past; therefore, it is of high significance to focus on waste management as the major subsystem (UNEP, 2015b). In Vienna, the correct path for POP substances fulfilling the aforementioned objectives is incineration in one of the state-of-the-art municipal WTE plants. These plants can act as final sinks for most POPs and also for PBDEs (Vehlow and Mark, 1997) because they are designed for complete mineralization and for very low emissions to air and water. In 2010, the two major imports into the subsystem waste management of Vienna were WEEE and construction wastes. WEEE is collected and divided between reuse, export for treatment abroad, incineration, and recycling. Construction wastes are divided between incineration and landfilling. PBDEs contained in atmospheric deposition are a third, minor import. They are collected by the sewer system and transferred to the municipal wastewater treatment plant, where PBDEs, as a result of their low solubility and lipophilic character, accumulate in sewage sludge.

The data on the level of goods were collected from Statistik Austria (vehicles), Elektroaltgeräte Koordinierungsstelle Austria GmbH (EAK-Austria) (annual collection of WEEE in Austria, distribution of “new” and “historical” devices), and previously published literature in Germany and Switzerland (construction). The mean concentrations of POP-PBDEs in some goods were taken from UNEP (2015a). The emission factors were taken from previously published literature.

3.3.2.4.4  Uncertainty Treatment

Data uncertainties are taken into account by the software STAN. Normal distribution is assumed for data with mean value μ and standard deviation σ. This approximation is often not appropriate, but it offers the possibility to use error propagation and data reconciliation. In reality, however, the higher the uncertainties, the less symmetric the error intervals become (Hedbrant and Sörme, 2000). To overcome the challenge of missing information and highly uncertain data, scenario analysis has been introduced. Three cases were investigated, and for each case, the impact of selected scenarios (starting with most realistic assumptions and continuing to vary one parameter at a time) on the MFA system as a whole was evaluated. The cases are as follows:

1. The split of cOctaBDE contained in WEEE categories 3 and 4 between (a) CRT-PC monitors and TVs, and (b) other products excluding screen devices. For (a), higher concentrations of PBDEs are expected than for (b).
2. cOctaBDE occurrence in vehicles, which is poorly documented in the literature, with large deviations from source to source.
3. Flows of PBDE containing waste plastics in construction wastes, where the uncertainty about the path to incineration or landfilling can vary between 4:1 and 1:4.

Table 3.42 summarizes the outcomes of the scenario analysis for scenarios 1a and 1b. For more details about the sources of these data and the scenario analysis, see Vyzinkarova and Brunner (2013).

Table 3.42   Overview of Measured cOctaBDE Concentrations in Polymers of CRT-PCs, CRT-TVs, and WEEE Categories 3 and 4 without Screens (Wäger et al., 2010); in Housing Shredder Residues from CRT-Glass Recycling (Schlummer et al., 2007); and in CRT-PCs and -TVs Polymers (Single Housing Samples) of European Origin Imported to Nigeria (Sindiku et al., 2012)

Sindiku et al. (2012)

Data Set Good

CRT-TVs

CRT-PCs

WEEE-3r

WEEE-4r

Mixed-3r&4r

HSR (CRT)

CRT-TVs

CRT-PCs

cOctaBDE

P41a

1.03

P31a

0.51

C3a

0.4

C4a

0.15

M3a

0.19

HSR1

0.00

S1–32

0.00

S1–22

0.00

concentration

P41b

0.05

P31b

0.14

C3b

0.05

C4b

0.15

M3b

1.56

HSR2

0.00

sзз

6.60

in sample,

P41c

0.67

P31c

0.66

C3c

0.1

M3c

0.38

HSR3

6.39

S34

59.30

g/Kg

P41d

0.05

P31d

10.6

HSR4

8.10

S35

64.10

P41e

3.54

P31e

0.79

HSR5

2.88

S36

290.00

P41f

0.66

HSR6

13.84

P41g

0.1

HSR7

6.35

Mean μ

0.87

2.54

0.18

0.15

0.71

5.37

11.67

0.00

Median

0.66

0.66

0.1

0.15

0.38

6.35

0.00

0.00

Standard deviation σ

1.14

4.04

0.15

0.00

0.61

4.55

49.12

0.00

Coefficient of variation

131%

159%

84%

0%

85%

85%

421%

0%

3.3.2.4.5  Results

The following three main outcomes were obtained: (1) Waste management plays a crucial role in the life cycle of PBDEs because it controls the path to the consumer and to the final sink. (2) Uncertainty of the data is high, pointing to future research needs. (3) Recycling plants are crucial to reach the objectives of clean cycles, and thus monitoring of plastic recycling products and emissions is required for quality control, in the same manner as, for example, monitoring of WTE residues and emissions.

The MFA shows clearly the key role waste management plays with regard to the objectives protection of human health and the environment as well as resource conservation (Figure 3.54): The largest amount of OctaBDE and PentaBDE flows from subsystem consumption to subsystem waste management. Vehicles are not treated within Vienna and are in part treated in Austrian car shredders outside of Vienna, partly exported. Consumer emissions to the environment are small and no longer play a significant role, neither for cPentaBDE (<10 kg per year [kg/yr]) nor cOctaBDE (<20 kg/yr). They will continue to decline as the consumption stock is depleted. Figure 3.54 shows the amounts of cPentaBDE and cOctaBDE in the consumption stock, estimated at 80 +/− 20 t for cPentaBDE and 20 +/− 40 t for cOctaBDE. Both stocks decrease at approximately the same velocities: dStock is −3 +/− 0.4 t/yr for cPentaBDE and −3 +/− 5 t/yr for cOctaBDE. If this trend continues in a static and linear way into the future, the two stocks of cPentaBDE and cOctaBDE will be depleted within 24 and 7 years, respectively, with high uncertainties for cOctaBDE.

Taking into account the aforementioned objectives, waste management should direct POPs into the final sink thermal treatment, either state-of-the-art WTE plants or cement kilns. Landfills are also sinks for POPs. However, in contrast to the complete thermal destruction during WTE or cement production, landfilling may release small amounts of POPs over a very long time period of several centuries. Thus, a landfill is not a final sink for PBDEs. MFA allows determining which fraction of cOctaBDE and cPentaBDE is directed toward a final sink, thus fulfilling the objectives of waste management. The largest waste flow of cPentaBDE (2 +/− 0.4 t/yr) is contained in construction waste plastics (PUR insulation foam). At the time of the case study, it was unidentified where these construction waste plastics end up. Thus, due to the lack of data, it is basically unknown if the objective of final sink has been reached for cPentaBDE.

The main flows of cOctaBDE are contained in WEEE (1.3 +/− 3 t/yr) and, possibly, EOL vehicles (2 +/− 0.9 t/yr), which leave Vienna for export, pointing to a supranational challenge. According to STAN modeling, 73% of cOctaBDE entering waste management ends up in WTE plants, with a high uncertainty of 1.2 +/− 5 t/yr. Five percent is exported, and 5% is landfilled. By recycling, 17% of cOctaBDE entering waste management returns back to consumption. However, this flow is highly uncertain (+/− 6 t/yr). Scenario analysis of case 1 shows that varying input concentrations of cOctaBDE in polymers of CRT-PCs, TVs, and WEEE-3r and WEEE-4r largely affect the results (Table 3.42). The biggest impacts have variations in the concentration of CRT-PC monitors and WEEE-3r.

Figure 3.54   Stocks and flows of (a) cPentaBDE and (b) cOctaBDE in Vienna, 2010, as modeled by STAN in tonnes per year resp. tonnes, rounded to 1 significant digit. Numbers for both substances are given as commercial mixtures. “?” designates that the stocks of PBDEs in the environment (soil) and in waste management (landfill) are not known. EEE = electrical and electronic equipment; WEEE = waste electrical and electronic equipment.

Both cases of cOctaBDE and cPentaBDE show high uncertainties. MFA clearly shows the need for better data. Part of the uncertainty of the process use of EEE is caused by the fact that there is no information about the stock of old EEE in Viennese households. Thus, statistical per capita data from other regions (Switzerland) had to be used to determine stocks of CRT-PCs and CRT-TVs in Vienna. The high uncertainty in stock changes influences estimates about recycling flows. Substance flows are calculated as (1) flow of goods multiplied by (2) polymer fractions and (3) substance concentrations in the polymer. Thus, the total uncertainty is additive and originates from three parameters.

To support the hypothesis that the current knowledge about the three aforementioned parameters is still small, available information about cOctaBDE concentration was reviewed. References include European flows of WEEE (Wäger, Schluep, Müller, and Gloor, 2011), housing, and mixed WEEE shredder residues (Schlummer, Gruber, Mäurer, Wolz, and Van Eldik, 2007), and CRT-PCs and TVs imported to Nigeria (Sindiku et al., 2012) (see Table 3.43). The data, which are further evaluated and discussed in Vyzinkarova and Brunner (2013), show that the existing data sets are insufficient for advanced decision making and need to be amended by more reliable, large, and profound sampling and analysis. In particular, for goal-oriented waste management and recycling, PBDE concentrations in various wastes such as EOL vehicles, WEEE, and construction wastes must be assessed in a systematic and reproducible way.

Table 3.43   Results of the Scenario Analysis of Case 1, with Different cOctaBDE Input Concentrations in Polymer Fractions of (1a) CRT-PCs and -TVs, and of (1b) WEEE-3r and -4r

Scenario 1a and 1b

Average Flow of Polymer (t/yr), Average cOctaBDE Treated Fraction (%)

cOctaBDE Concentration in the Fraction (g/kg)

Impact on the System: cOctaBDE Recycling Flow Estimate (t/yr)

1a

CRT-PCs

994, 30

Min. c = 0.14

0.19 ± 6.36

Max. c = 10.6

0.68 ± 6.63

Mean c = 2.54

0.29 ± 6.06

Median c = 0.66

0.20 ± 6.29

CRT-TVs

1741, 30

Min. c = 0.05

0.23 ± 6.17

Max. c = 3.54

0.52 ± 6.24

Mean c = 0.87

0.29 ± 6.06

Median c = 0.66

0.28 ± 6.08

1b

WEEE-3r

808, 42

Min. c = 0.05

0.24 ± 6.04

Max. c = 1.56

0.96 ± 6.23

Mean c = 0.18

0.29 ± 6.06

Median c = 0.38

0.39 ± 6.07

WEEE-4r

325, 24

Min. c = 0.15

0.29 ± 6.06

Max. c = 1.56

0.46 ± 6.10

Mean c = 0.15

0.29 ± 6.06

Median c = 0.38

0.32 ± 6.06

Also, there is a need for information about recycling processes: transfer coefficients have to be determined for the various recycling techniques. Despite this discussion about the lack of sufficient data, the MFA displayed in Figure 3.54 indicates clearly that by WEEE management in Vienna, cOctaBDE is partly directed into consumer products. This has been observed by a similar study in Switzerland, too (Morf, Taverna, Daxbeck, and Smutny, 2003; Morf, Tremp, Gloor, Huber, Stengele, and Zennegg, 2005).

Hence, there is need for action. Austrian legislation requires federal states to control plants that treat hazardous waste at a minimum of every 5 years (AWG, 2002). This legislation could be expanded to include the flows of selected POP-PBDEs through recycling plants. If monitoring of products and emissions of PBDEs in recycling plants is introduced, recycling could reach the same high standards that WTE plants fulfill today. This would enable us to follow POP-PBDEs from sources to final sinks and to ensure that the goals of a clean cycles and safe final sink strategy can be reached by waste management.

3.3.2.4.6  Conclusions

The case study allows drawing conclusions with respect to (1) the application of MFA on one hand and (2) waste management decision making on the other hand.

1a. Regarding MFA, the case study focuses on the substance level and shows that defined mixtures of similar substances can be investigated by MFA and STAN, too. The substances investigated and balanced, e.g., cOctaBDE, are a commercial blends of several individual substances that are quite similar but not identical. As long as the commercial mixture contains similar congeners (in the case of cOctaBDE, from hexabromodiphenyl to decabromodiphenyl ether) with similar physical–chemical characteristics, it can be justified to treat the mixture as one substance. In case the mixture comprises also substances of different properties, it will be necessary to analyze and balance each substance individually.

The case study proves that even with little information about flows and stocks of substances and about transfer coefficients, it is still possible to establish an MFA on the substance level. Preconditions are a minimum data set about flows and stocks of goods containing POP-PBDEs and about concentrations of these chemicals in the corresponding goods. In order to reduce uncertainty, scenario analysis is useful. It allows us to identify the crucial parameters, and to focus on these. It is important to realize that even with very little data, an MFA/SFA can be established, although the uncertainty usually will be high. With increasing research, analysis, and expenditure, uncertainty can be reduced. There is a trade-off between costs and uncertainty: the higher the resource input, the lower the uncertainty. The art of designing an MFA system and collecting data in a cost-effective way is to reduce uncertainty only as much as is necessary to draw conclusions regarding the objectives of the project.

1b. For designing the MFA system and for data collection, the choice of system boundaries is crucial. In general, the system boundary in space should be selected with data collection in mind. City is an appropriate system boundary if the data are administered by a municipality or by a body that collects data on a city level. This is sometimes the case for goods but, unfortunately, rarely the case for substances. In hardly any city, data about the flows of PBDE are collected and managed. Thus, it is necessary to link various information sources: urban stocks and flows have to be reconstructed from national data, or from data from other urban regions where the missing information is available.

To choose a system boundary on an urban level makes sense for subjects that can be managed by the city. For instance, municipalities responsible for waste management want to know if their waste management practice fulfills federal regulations, and if not, what the most effective means would be to reach compliance. On a general level, the PBDE case study supports a final sink strategy of a municipality operating a WTE plant that is capable of completely destroying organic substances.

2a. Regarding waste management decision making, it is clear that without the MFA of PBDEs, it is not possible to identify those hotspots of PBDE flows and stocks that are offending legislation. MFA links sources and sinks, in this case PBDE containing consumer goods on one hand and recycling products, WTE plants, and emissions on the other hand. It is interesting to note that the main result of polluted cycles and incomplete flows to final sinks can be reliably estimated based on rather limited data. Thus, based on this SFA, measures to control the flows for compliance with regulations can be designed with confidence.

2b. A dilemma of modern waste management is the need for closing cycles, on one hand, and the fact that sometimes, hazardous substances are enclosed in wastes, rendering them unsuitable for recycling, on the other hand. MFA is instrumental for resolving this dilemma because it can show the level both of goods having a recycling potential and of substances comprising possible hazards for human health or the environment.

2c. In Vienna, the largest flows of POP-PBDEs are contained in three wastes: WEEE, construction wastes, and EOL vehicles. For cOctaBDE, WEEE and, possibly, vehicles are the main flows. Most EOL vehicles are exported from Vienna and pose a continental, rather than a local, challenge. According to the modeling, approximately 73% of cOctaBDE ends up in WTE plants with advanced APC, which represent safe final sinks. In view of the goals of waste management, namely, protection of human health and environment, cOctaBDE in WTE plants fulfills the objectives of complete destruction.

2d. A considerable fraction of POP-PBDEs containing waste is recycled. For cOctaBDE entering waste management, 17% is directed back to consumption, with little information about its fate during preparation and recovery. Secondary plastics, made from WEEE, may thus contain significant amounts of cOctaBDE; however, uncertainties are high. According to uncertainty analysis, the major cause is the lack of reliable values regarding cOctaBDE concentrations in European WEEE categories 3 and 4, including cathode ray tube monitors for computers and television sets. For adequate understanding and decision making in waste management, more information about the recycling processes is required.

In order to protect workers, human health, and the environment, a new, goal-oriented data set and mass balance of flows and stocks of polybromi-nated diphenyl ethers needs to be established. It must contain information about waste constituents, recycling plastic compositions, measured data about transfer coefficients, and emissions of POP-PBDEs in existing treatment plants, particularly recycling plants. Without the same set of information that, for example, WTE plants disclose, effective allocation of PBDE containing wastes to different waste treatment plants will not be possible. Dependable and sufficient information is particularly required because waste management is the key process for a region that has—due to successful regulations—no more inputs of POP-PBDEs but still has a large stock because of the legacies of the past.

2e. The main flows of cPentaBDE are contained in construction materials and are landfilled in construction waste landfills. They represent a long-term stock releasing minor amounts of PBDEs over long time periods. In view of the waste management goals aftercare-free landfills, this practice does not yet comply with legislation. Therefore, EOL construction materials made of plastic and containing POP-PBDEs, especially PUR foam insulation, PVC duroplastic sheeting, and PE roof sheeting, which may account for cOctaBDE flows into landfills, must be separated from construction wastes and properly treated, for example, in a state-of-the-art WTE plant. The example of PBDEs shows well the power of MFA to support a clean cycle and safe final sink strategy in waste management.

Problems—Section 3.3

Problem 3.8:

• Plastic wastes have a high calorific value, which makes them a potential fuel for cement kilns, blast furnaces, and municipal incinerators. Packaging plastics have been successfully incinerated in cement kilns: production costs are reduced; the quality of cement does not change; and emissions are not altered significantly. Wastes from longer-lasting plastic materials (containing about 10% PVC) have a high chlorine content, rendering these wastes unsuitable as a fuel in cement kilns because they exceed the capacity of the process for chlorides. Using Table 3.29, evaluate whether nonpackaging plastics are better suited for blast furnaces or for MSW incinerators. Take into account environmental and resource considerations only, and do not consider economic (additional investments, fuel savings) or technological (pretreatment, adaptation of feed or furnace, etc.) aspects. Discuss final sinks for heavy metals. Use the transfer coefficients in this handbook for MSW incineration, Table 3.38, and look for data about blast furnaces in the library or the World Wide Web.

Problem 3.9:

• Assess the paper content in MSW of a country of your choice. First, determine the appropriate system (processes, flows, system boundaries). Second, carry out an Internet search for the annual report of the pulp and paper industry of the selected country and determine the flows through and within your system. Third, find out the national MSW generation rate (e.g., contact the EPA website) and calculate your result.

Problem 3.10:

• The combustion of biomass is described in Obernberger, Biedermann, Widmann, and Riedl (1997). Calculate the Cd concentration in cereals based on the information given in the paper. Using the approach described in Section 3.3.1.2, calculate the composition of the input (cereals) from the composition of the output (different ash fractions). Compare with Cd values for cereals you find in the literature.

Problem 3.11:

• Figure 3.55 gives the Cd balance for the management of combustible wastes in Austria. Discuss the flowchart together with the total mass balance for combustible waste flows in Austria as given in Section 3.3.2.1, Figure 3.45. Consider resource potentials and potentially dangerous environmental loadings.

Problem 3.12:

• Summarize the reasons why a cement manufacturer association might decide to limit the annual flow of heavy metals into cement kilns with 15% of the national consumption of heavy metals.

Problem 3.13:

• Assume that incineration of 1 ton of MSW [copper (Cu) content, ca. 0.1%] yields the following solid residues: 250 kg of bottom ash, 25 kg of fly ash, 3 kg of iron scrap, and 3 kg of neutralization sludge from the treatment of scrubber water. About 90% of the Cu leaves incineration via bottom ash and 10% via fly ash. The Cu flow via other residues such as off-gas, iron scrap, etc. is <1% and can be neglected. Investigations show that by mechanical processing of bottom ash, approximately 60% of the Cu can be separated in the small fraction of metals concentrate. The Cu content of this fraction (ca. 50%) can be recovered in a metal mill. (a) What is the recovery efficiency for the combined process MSW and mechanical processing of bottom ash? (b) Calculate the substance concentrating efficiency (SCE) for the process chain incineration, mechanical processing, and metal mill. (c) As a decision maker, would you support such a technology, and why?

Figure 3.55   Flows of cadmium induced by the management of combustible wastes in Austria (1995), t/year.

The solutions to the problems are given on the website http://www.MFA-handbook.info.

3.4  Industrial Applications

MFA has a long-standing tradition in chemical engineering. Educts and resulting products and by-products have been balanced by stoichiometric methods for reasons of reaction design, optimization, and quality control of chemical processes. While this has been state of the art for many decades in the production of chemical substances, MFA has just recently been introduced to industrial processes in other fields, such as metal production and automotive or airplane engineering. Particularly in manufacturing, the advantage of applying MFA has been recognized when optimizing processes and process chains (Krolczyk et al., 2015). Krolczyk et al. use MFA as an analysis and optimization tool for reducing costs in an industrial company that manufactures composite elements for automotive, electric, and agricultural industries. They see the main advantage in the comprehensive picture that is produced by analyzing flows and stocks of materials through a production plant in a systematic way. Particularly, they point out how MFA can be used to create an internal transport program and to optimize the working stands arrangement. As a result of the reorganization of the manufacturing plant, the number of transport operations is reduced, material supply and transport are smoothened, and costs are reduced.

Another example of application of MFA in the processing industry has been conducted by Trinkel, Kienberger, Rechberger, and Fellner (2015). These authors attempt to balance different heavy metals in a blast furnace process in order to follow their path from source to products and emissions. However, they face various challenges, particularly because heavy metals are sometimes present at small concentrations in different input and output materials. The composition of these materials is often heterogeneous, making representative sampling and subsequent analysis difficult. In their case, the major challenge for performing an MFA of lead through a blast furnace is the analysis of the content of Pb in the metal produced. Different analysis methods result in different Pb concentrations. In addition, Pb proves to be unequally distributed in the metal product, calling into question current sampling and analysis procedures. This example shows well the power and limitation of MFA in supporting decisions on the production level: if adequate sampling and analysis methods are not available, balancing of substances in complex processing such as a blast furnace becomes a real challenge.

In the following chapter, an MFA of an industry manufacturing interior panels for airplanes is presented (Müller, 2013). The specific feature of this case study is the link between MFA and economic analysis. The flow of values is depicted in parallel to the flows of goods. Also, flows and stocks of materials are associated with working hours, thus enabling an economic optimization of the production lines with less idle time and more productivity. A similar approach has been attempted before by Kytzia (1989). This author linked flows of materials, energy, and financial resources into one model of the residential building stock. In her study, flows of financial resources consist of cost and revenue. The difficulty in such a model is how to allocate revenues to the various cost units.

The same problem is encountered by the MFA work of Eisingerich (2015) on open burning of rice straw. This author links material and substance flows on Thai farms to economic parameters in order to improve the economic situation of small rice farmers, and at the same time to decrease environmental loadings. The allocation of revenue to the individual farming processes could not be accomplished for all flows and processes. Case Study 16 by Müller avoids these difficulties by focusing on production costs only and neglecting the revenue (Müller, 2013). The rationale for this approach is that the goal of process optimization is to minimize cost of production, and not to increase revenue, because revenue is independent of production processes and cannot be increased by process optimization.

3.4.1  Case Study 16: MFA as a Tool to Optimize Manufacturing

To use resources efficiently and without environmental degradation is not only an economic objective of a company; it is also one of the goals of sustainable development. Hence, it is in the interest of both companies and society to minimize resource consumption, emissions, wastes, and cost per unit of good produced. The present case study demonstrates how MFA can be applied on the company level for minimizing resource use and optimizing economic benefits.

The main challenge for an entrepreneur is to identify those production processes that have the highest potential for resource conservation, environmental protection, and economic optimization. Key questions are which methodology to apply, how to get the data, and how to assess the effects of uncertainty on the results. To answer these universal questions, a case study on a state-of-the-art manufacturing system for an advanced product of the aircraft industry was performed (Müller, 2013). Because of the novelty of the linking of MFA with economic parameters, the case study also required new methodological development beyond traditional MFA and STAN.

The company involved in this case study is an internationally leading producer of insulating materials, laminates, and composites. The MFA covers a particular segment of production and focuses only on a small fraction of the entire company. For reasons of confidentiality, the name of the company as well as the names of goods and processes are undisclosed. All numbers of flows, stocks, and economic parameters are changed for reticence. Nevertheless, the results and conclusions serve well to demonstrate the power of MFA to support and optimize manufacturing processes.

3.4.1.1  Objectives

Case Study 16 aims at developing a method for mapping complex manufacturing systems in a transparent and comprehensible way in order to minimize production costs (primary objective) and wastes, and optimize resource use (secondary objective). The goal is to produce one or several models that take into account all relevant stocks and flows of materials, costs, and production time, including uncertainties (Müller, 2013). These models should test the feasibility of MFA for identification, analysis, quantification, and representation of production systems, and allow discerning of the production steps with the highest potential for improvement. They show flows and stocks of educts, products, wastes and emissions of each process, the associated costs, and the working hours required to produce a particular product or waste. Also, the possibility of STAN linking MFA and economic parameters such as costs and time represents an important new and original step.

This facilitates the understanding of the entire manufacturing process and represents the starting point for optimization in terms of resource efficiency, cost, and time.

The option of STAN to include the level of substances is not an objective of this study. However, if issues such as health protection or environmental pollution were to be addressed, hazardous substances could be added for investigation without methodological difficulties.

The following research questions are addressed in the case study:

1. Is it possible to jointly depict flows of material and money of manufacturing processes by STAN?
2. What is the main advantage of using STAN for this combination?
3. How can uncertainties be taken into account for risk assessment?
4. How can STAN diagrams be used by entrepreneurs for optimization of production systems?
5. How can STAN be improved for the specific purpose of mapping production processes physically and economically?

3.4.1.2  Procedures

Basically, three (MFA) models are created (Müller, 2013). In the first model, the whole production system for manufacturing one unit of output is described. In the second model, the effect of uncertainties of input flows is investigated. The third model allows the following of a semiproduct over a defined time period. These three detailed models facilitate comprehensive understanding of the manufacturing process and enable identification of the potential for optimization as well as the impact of uncertainties in the input values. The results of the three models have been compared with those of the enterprise resource planning (EPR) system that is installed in the company. This comparison serves as a plausibility check, too.

The entire manufacturing system is modeled in STAN following these steps:

1. Definition of system boundaries, units, balancing periods, and costs.
2. Structuring the manufacturing system into processes, material flows and stocks, and associated cost flows and working hours.
3. Implementation of steps 1 and 2 in STAN.
4. Collection of production and economic data about flows, stocks, and working hours, and input of this information into the STAN model. For this, all input goods are put in relation to one unit of output product.
5. Validation of the STAN model on the mass flow level by balancing the MFA system for all periods and correcting errors.
6. Validation of the STAN model on the money flow level by balancing the MFA system for all periods and correcting errors.
7. Checking for plausibility by comparing the results of STAN with another planning or quality control system, such as the ERP system implemented in this company.
8. Applying the results for optimizing the manufacturing process.

Simple examples for balancing mass flows, money flows, and working hours of a single process with the same STAN model are presented in Figures 3.56 through 3.58. They show that in principle, money and working hours can be treated the same way as mass flows. This offers the advantage that a single STAN model allows combining of all three aspects of mass, monetary values, and required working time. In this case study, the unit energy, implemented by default in STAN, was replaced by the unit money.

Figure 3.56   Example of flow of materials through production process 1. Expenditure of time (working hours) is expressed in STAN as a virtual material flow and thus is represented as 0 in the mass flow diagram. The numerical values for expenditure of time are given in Figures 3.58 and 3.62.

Figure 3.57   STAN representation of money flows associated with production process 1. The sum of the costs of individual educts plus process costs equals the cost of the product. Process costs are not associated with a material flow, and include all costs emerging from production except for educts such as feedstock and semifinished products. Labor cost is included in the process cost.

Figure 3.58   STAN representation of working hours associated with a production process. The sum of the working hours used to produce the educts plus the time expenditure for producing the process output yields the total working hours to produce a unit of output.

System boundaries in space comprise the area that is required for the production of the product, including machinery, equipment, space for stock, and transport. The boundaries in time change according to the rhythm of the production system, which is dependent of the external economic situation. In this case study, 10 periods of 1-month duration each have been investigated and balanced. Flows and stocks of materials are analyzed and balanced according to MFA standards; for more information, see Müller (2013).

In order to include costs in STAN, the energy level offered by the software was “abused” and exchanged for money flows, with the euro (€) replacing the energy unit joule (J). The advantage of this procedure is that it is easily possible to switch between the two STAN levels of mass and money flows and that complete consistency is given between the two levels (cf. Figures 3.56 and 3.57). By a simple click, a manufacturing system depicted in STAN can be viewed either as a physical material system of flows and stocks, or as a money flow system. However, as of now, the energy level cannot be used for simultaneously mapping energy flows.

The functional unit of this case study is one unit of a product. All processes and flows of goods that are performed within the enterprise contributing to the manufacturing of this unit are taken into account. Wastes, emissions, and by-products are considered as well. If processes are complex, it is recommended to split them into subsystems. This prevents a black-box effect and facilitates understanding of the underlying subsystems.

In order to evaluate the results, operating numbers (key figures) are defined. They allow assessment of the effect of measures on the manufacturing system and are instrumental for comparing performance of different units within and outside of the enterprise. The following two sets of operating numbers are chosen for this case study:

Ecological operating numbers focus on resource efficiency, solvent use, and wastes because these three issues are the main environmental concern of the company: (1) total material efficiency: ratio of total material input per unit of product; (2) efficiency of solid auxiliary material utilization: ratio of solid auxiliary material input per unit of product; (3) solvent utilization: ratio of solvent used per unit of product; and (4) waste generation: ratio of waste produced per unit of product.

As economic operating numbers, the following five key figures are defined: (1) material costs: ratio of costs of raw materials versus total costs to produce a unit of product; (2) processing costs: ratio of process costs versus total costs; (3) solvent costs: ratio of solvent costs versus total costs; (4) solid auxiliary material costs: ratio of solid auxiliary material costs versus total costs; and (5) costs for waste management: ratio of costs for waste management versus total costs to produce a unit of product.

By taking into account uncertainties, production risks can be assessed. This allows us, for example, to set priorities for purchasing, to support the selection of cheaper or more environmentally sound substitutes, or to define tolerances for individual manufacturing processes. For this reason, values for uncertainty are implemented in STAN for all input flows on the level of mass flows as well as money flows. Various scenarios are calculated in order to analyze the effect of uncertainties on the results. For more information and practical application, see Müller (2013).

3.4.1.3  Results

The processes and flows of educts, auxiliary materials, solvents, and wastes for the production of one unit of product are presented in Figures 3.59 through 3.62. Figure 3.59 depicts all mass flows to produce one unit of final product, Figure 3.60 money flows per unit of final product, Figure 3.61 working time required to produce one unit of product, and Figure 3.62 the mass flows for producing one unit of semifinished product.

On the basis of Figures 3.59 and 3.60, the cost driving material flows can easily be detected, and the focus for economic improvement can be put on these flows, respective of the losses of the processes handling these flows. The calculation of the operating numbers yields the ratio of process costs versus material costs, thus allowing us to set priorities in optimization. In this case study, both costs are nearly equal [cf. Figure 3.60: sum of cost for total material import (235 €/P) minus cost of exported product (476 €/P); the difference of 241 €/P is the operating cost]. It is recommended to decrease the costs of auxiliary materials because they have the least impact on the final market product.

The scenario analysis allows checking of the effect of variations in manufacturing. The variations in Table 3.44 are chosen according to actual market and manufacturing conditions. Ten percent uncertainty of the import mass flow in scenario 1 yields only 3.7% uncertainty on the final product flow. In order to maintain high-quality production, new specifications for import materials can be defined that allow a better performance on the output side.

Figure 3.59   STAN mass flow diagram of the entire production from feedstock to product, including all imports, exports, semifinished products and products, and wastes. Exports such as off-gas from thermal post combustion and residues from waste treatment are not considered.

Figure 3.60   STAN money flow diagram of the entire production, including all material and operational costs, per unit of product.

Figure 3.61   STAN cost-per-time-flow diagram (K€/month) to produce one unit of semifinished product during a period of 1 month.

Figure 3.62   STAN mass-per-time-flow diagram for product manufacturing during a time period of 1 month.

Table 3.44   Uncertainty Assessment Based on Scenarios Analysis for Economic Optimization of Production

Uncertainty of Final Product

Scenario

Scenario Specification

Level of Goods, %

Level of Cost, %

1

10% uncertainty in all input mass flows

3.7

1

2

10% uncertainty in the purchase price of all input flows

0

1.8

3

Dependence from suppliers: 45% uncertainty in cost for raw materials

0

6.4

4

Variation in production: 20% uncertainty in processing cost

0

1.9

Ten percent uncertainty in the purchasing price for imports (scenario 2) results in a rather small uncertainty of 1.8% for the cost of the final product. In order to assess the importance of unstable markets, uncertainties can be assumed to be much larger, and the corresponding effects on the total cost can be calculated. This enables a proactive business strategy anticipating future market volatility, e.g., in the resource or energy markets. If the price for the two most important raw materials fluctuates by 45%, the largest effect on the final product results (scenario 3). In scenario 4, the process costs are assumed to differ by 20%. The scenario analysis shows the largest effect on the final product for scenario 3. For the final revenue of the whole manufacturing process, plus or minus 6.4% cost of the final product is a significant number. Thus, to reduce the entrepreneurial risk, strategies must be developed to stabilize the cost of the two crucial raw materials that are, at the moment, purchased from a single supplier. Variations in operating costs are less relevant. To summarize, the scenarios displayed in Table 3.44 serve well for setting priorities for economic stabilization of the production.

The time required for the production of semifinished products is displayed in Figure 3.63. In this figure, the STAN results of 10 consecutive balancing periods of the manufacturing system are summarized for semifinished product 5. This allows, on one hand, identifying the individual workloads of the working places. On the other hand, the total time required to produce a semifinished product or a product can be calculated. Since some of the semifinished products are especially made within the company to supply the manufacturing process, information about available stocks (Figure 3.64) and time required for their production is instrumental for careful planning of the whole operation. Figure 3.64 presents a highly useful overview about mass flows and stocks and their changes over time, and shows that the manufacturing process is not a just-in-time operation yet but shows fluctuations. STAN diagrams and Figure 3.64 serve as means to raise awareness among the personnel for stock and flow issues in order to keep the stocks low and the wastes small. Also, they can be used to optimize labor force employment.

Figure 3.63   Flows and stocks of mass and money to produce a semifinished product over 10 balancing periods.

Tables 3.45 and 3.46 summarize the operating numbers determined in this case study. The ecological operating numbers show considerable promise for improvement of the manufacturing process. Overall material input is about 2.8 times higher than useful product output. Also, solid and liquid (solvents) auxiliary inputs are larger than the product output. Per 1 kg of product, 0.5 kg of wastes is generated. Table 3.45 shows that the production of semifinished product 4 yields the highest operating numbers and thus is of first priority when optimizing production as a whole.

Economic operating numbers in Table 3.46 show that efforts to minimize manufacturing costs should focus on the production of semifinished products 1, 4, and 5. The purchasing department is well advised to negotiate better purchasing conditions for semifinished products 2 and 3 because they have the largest potential for cost saving. The example shows that operating numbers are well suited to effectively support cost reduction in production processes.

3.4.1.4  Conclusions

The conclusions regarding application of MFA in manufacturing on the three levels mass flows, money flows, and time are summarized in Table 3.47. The outcomes show the feasibility of this method for analysis and representation of production systems, and that STAN is well suited for decision support.

Figure 3.64   STAN diagram presenting the time required to produce semifinished product 5. The change in stock of minus 260 h signifies that during this period, more time was consumed for the production of the feed and semifinished products than was supplied and accomplished during that period. This can be due to a decrease in stock when material has been produced in a former period.

Table 3.45   Ecological Operating Numbers of the Semifinished Products (SFPs) and the Final Product (FP)

Ecological Operating Number

SFP1

SFP 2

SFP 3

SFP 4

SFP 5

FP

Total material input per product

140

167

150

200

125

276

Solid auxiliary material per product

40

67

50

100

25

125

Solvent utilization per product

40

67

50

40

0

60

Waste generation per product

0a

0a

0a

0a

0a

50

Note: The numbers stand for material flows per unit of product and are given in %.

Notes:

a  No waste generated.

Table 3.46   Economic Operating Numbers of the Semifinished Products (SFPs) and the Final Product (FP)

Economic Operating Number

SFP 1

SFP 2

SFP 3

SFP 4

SFP 5

FP

Material cost

27

60

55

20

36

33

Processing cost

56

24

24

57

43

34

Auxiliary material cost

17

16

18

19

18

28

Solvent cost

17

16

18

13

15

15

Solid auxiliary material cost

0

0

0

6

3

12

Waste disposal cost

0

0

3

4

4

5

Note: The numbers stand for, e.g., material cost per total costs to produce a unit of product, and are given in %.

Table 3.47   Application and Outcome of MFA in Manufacturing on the Three Levels of Goods (Mass Flows), Cost (Money Flows), and Time

Model

Level

Outcome

Entire   production

Goods

Facilitates understanding of production system Supports the design of optimal material flows

Cost

Points out processes and material flows of high costs

Uncertainty

Goods

Reveals the effect of inaccuracies of manufacturing on the final products

Cost

Exposes the effect of fluctuations of cost of material and labor on cost of end product

Exposes the effect of fluctuations of process cost on cost of end product

Semifinished product

Goods

Delivers actual material flows through the production system Shows the demand for material stock for each manufacturing step and process

Cost

Depicts actual money flows through the manufacturing system Depicts capital required for each working place

Time

Demonstrates workload of each working place Demonstrates minimum processing time of a semifinished product

The answers for the five questions addressed in the beginning (cf. Section 3.4.1.1, “Objectives”) are as follows:

1. Is it possible to jointly depict flows of material and money of manufacturing processes by STAN? Formally, STAN is not yet equipped with a feature that allows us to take economic parameters into account. However, it is possible to substitute energy (in J) for costs (in €). This exchange allows easy consideration of money flows. The drawback is that at the moment, it is not possible to work with both energy and costs. There is a need for a next version of STAN that will provide both possibilities.
2. What is the main advantage of using STAN for this combination? STAN delivers a total view of a production system including both mass flow and economic level. Switching from one to the other level is quick and easy. Full transparency and reproducibility are guaranteed. This facilitates fast comprehension of the entire production system.
3. How can uncertainties be taken into account for risk assessment? STAN is well suited to include and calculate data uncertainty (cf. Chapter 2, Section 2.4). Thus, based on the uncertainty of input data, output uncertainties can be assessed, and entrepreneurial decisions can be based on these uncertainties. Since values and uncertainties of data can be changed easily, STAN serves well for scenario analysis, too. However, at present, only data with standard distribution can be taken into account.
4. How can STAN diagrams be used by entrepreneurs for optimization of production systems? (a) Because both mass flow and economic levels are included, the STAN diagram yields an overview of production cost and allows detection of causes of high cost in a straightforward way. The graphs are very well suited for decision support in planning of future investments, for strategic priority setting, and for optimization of manufacturing systems. (b) The ratios of waste versus final product and resource use versus final product point out potential economic losses and allow the setting of priorities for improvement. (c) The working-hour diagram shows where the production line can be improved by decreasing idle time.
5. How can STAN be improved for the specific purpose of mapping production processes physically and economically? (a) Managing a production line requires appropriate information about mass flows, energy flows, money flows, and expenditure of time flows. For further development of STAN, it is recommended to incorporate these four levels into the software. This means amending the present version with two additional levels for money, and time and labor. (b) The possibilities to present results should be amended by an evaluation showing the composition of the final product in terms of all mass inputs and money flow inputs. This allows exact assignment of the total materials and money flows used in manufacturing to one unit of product. (c) If operating numbers are implemented in STAN, they can be calculated automatically, saving time and effort. A graphical display of the development of operating numbers over time, or for different scenarios, would increase the value of STAN considerably for optimization of manufacturing. (d) At present, for uncertainty calculations, STAN assumes that data are normally distributed. STAN results could benefit by a feature that allows (i) choosing between various distributions when inserting data and (ii) determining a lower and upper limit for uncertainty. (e) If STAN can be linked to an ERP system, the application of STAN would be significantly facilitated, leading to a wider and regular use of STAN.

In summary, Case Study 16 shows that cost analysis and production time can be linked to material flows and stocks by STAN pursuing the following three steps: First is modeling the mass flows, money flows, and working hours of the entire production system for one unit of product. Second is modeling the flow of a semifinished product over a defined period of time from imports (educt, auxiliary materials, and solvents) to exports (semiproduct). In the third model, the effect of data uncertainties in input flows is investigated. These three highly detailed models allow the identification, quantification, and realization of optimization potentials for production systems in terms of resource efficiency, cost, and time.

For such a comprehensive study, the quality of input data is key because it determines the quality of the results and hence of the subsequent business decisions. In order to collect appropriate data of good quality, the expenditures to establish the three models are considerable. However, the graphic display of the results greatly increases the understanding of the entire manufacturing system. In addition, changes in production or input goods can be implemented quickly by STAN, and thus, the models serve as an excellent tool for improving and optimizing manufacturing. The addition of economic parameters such as cost and time in STAN represents an important step in decision support for efficient production systems. Due to the option of STAN to include the level of substances, too, this method can be further advanced to address environmental and health issues of a production line.

3.5  Regional Materials Management

The objective of regional materials management is to protect the environment, to conserve resources, and to minimize wastes in one combined effort. Regional materials management is an integrated approach that links all three issues and strives for an optimum solution. Instead of focusing on one topic alone, all three are taken into account at the same time. This comprehensive procedure requires less effort and results in more information than three separate studies. It also ensures that the results from the different fields are compatible and that conclusions regarding all three fields can be drawn. For regional materials management, it is essential to know the main anthropogenic as well as natural sources, conveyor belts (transport paths), stocks, and sinks of materials in a region. Without this information, regional materials management is not possible. In order to achieve the stated goals, a long-term view must be taken. Material flows and stocks have to be balanced over decades to centuries in order to examine whether harmful or beneficial accumulations and depletions of materials are taking place in the region. All materials used within a region must find a safe final sink. If a safe final sink is not available, use of the material should be phased out or controlled tightly and accumulated over long time periods with a clear purpose and economic plan for future reuse.

3.5.1  Case Study 17: Regional Lead Management

This example of regional lead management is drawn from Case Study 1 (Section 3.1.1), which described system definition and data collection. The following discussion covers only those results and conclusions that are important for regional materials management.

3.5.1.1  Overall Flows and Stocks

A total of 340 t/year of lead is imported into the region, and 280 t/year is exported (see Figure 3.1). The main import consists of used cars that are crushed in a car shredder. The main exports are filter residues from a steel mill that produces steel for construction from the shredded cars, lead contained in construction steel, and lead in MSW. The difference between imports and exports amounts to 60 t/year, which is accumulated mainly in landfills. The geogenic stock soil includes about 400 t of lead (the term geogenic is not actually precise here, since a certain fraction of lead in soils is of anthropogenic origin (Baccini, von Steiger, and Piepke, 1988). The anthropogenic stock landfill is much larger and amounts to >600 t (>10 years of land-filling 60 t/year). Like most materials in urban regions (Chapter 1, Section 1.4.5.4), lead is accumulated in this region. Imports and exports of lead by geogenic conveyor belts (air, water) are marginal and are <1°%. From the point of view of resource management, shredder residues and filter dusts are of prime interest. From an environmental point of view, depositions on the soil as well as potential leaching of lead from landfills to surface water and groundwater are important.

The advantage of a regional material balance is that with one single balance, present and future hot spots for environmental, resource, and waste management can be detected. For example, the potentially large but, at present, unknown flow of lead from local landfills to the hydrosphere cannot be hypothesized without information about local landfills and their constituents. Quantitative information about landfills in general is not available, but it is known that most of the shredder residue is landfilled within the region. By a simple balance of the car shredder, assuming a certain lead input based on car manufacturers’ information and the number of cars treated in the shredder, and the lead in the metal fraction used and analyzed by the smelter, it is possible to roughly assess the amount of landfilled lead.

The existing stock of lead in landfills totals >600 t. A doubling time (t2x) for the lead stock of ≈10 years can be calculated. In other words, if the regional anthroposphere remains the same for the next 100 years, the stock will have increased from 600 to 7000 t. According to Chapter 1, Section 1.4.5.1, there are no indications yet that waste lead flows will decrease. What makes this case study special is the huge extent of the accumulation. Nearly 20% of the lead imported does not leave the region and stays there probably for 10,000 years until erosion slowly removes the landfill. All lead landfilled and deposited on the soil is of no further use. The concentration is comparatively low, and the heterogeneity of the landfilled materials is much larger than that of lead ores. Hence, economic reuse of this stock is, at present, not feasible. Emissions from this stock are likely but not known. A conscientious approach to regional materials management would dictate that, in the future, this lead stock be managed in a different way, turning it from a hazard into a positive asset. Means for upgrading and reuse have to be explored (see item 2 in Section 3.5.1.4).

Lead flows can be divided into flows in products, in wastes, and in emissions. The management goal is to maximize the use of lead in products, to reuse lead in wastes, and to reduce emissions to an acceptable level. MFA shows where the large lead flows are and thus points out key processes and goods for control and management. For each environmental compartment (water, soil, air), potential sources are identified and sometimes quantified. Thus, priorities can be set when measures for the protection of the environment are taken.

Figure 3.1 shows that lead increases by 1.4 t/year in the river between the point of entry in and exit out of the region. While 0.31 t/year is due to leaching from soils, and 0.14 t/year due to treated wastewater, 1 t/year has not been accounted for by MFA. This flow is so large that it is not likely due to an error in measuring soil leaching or effluent from WWTP. The most probable source of this large amount of lead is leachates of shredder residue from landfills. It is not efficient to reduce the comparatively small amount of lead emitted by WWTP effluents. The first step is to investigate the hypothesis that shredder residue landfills are really leaching such a large amount of lead to the surface waters. The second step is to reduce the loadings of the soil, e.g., by banning leaded gasoline (as was done in the late 1980s) or by incinerating sewage sludge and landfilling the immobilized ash.

MFA supports environmental impact assessment and serves as a design tool. Figure 3.1 shows that emissions from landfills (and any other point source) are not relevant if they are in the range between 0.002 and 0.02 t/year (0.1–1% of present aquatic export flow). Considering a total stock of ≈1000 t of lead landfilled, one can calculate that no more than about 2 to 20 ppm (mass) of lead may be mobilized in the landfill if there is to be no significant effect on the river water concentration. This figure can serve as a goal for the design of waste treatment such as immobilization or solidification. Note that this calculation does not consider groundwater pollution. If a local groundwater flow is small and the residence time is high, the lead flows from landfills calculated previously may be large enough to exceed drinking water standards. Hence, it is important to take groundwater into consideration, too.

For the region, it is more efficient to manage lead in a comprehensive way than to segregate the lead issue into different problem areas. This is exemplified by the following three conclusions:

1. Lead not in use should be accumulated actively and purposefully in safe, intermediary stocks with residence times of several decades. The objective is to build up concentrated stocks of lead and other metals and to reuse these stocks once they have reached a size that makes them viable for economic reuse. In order to concentrate lead as much as possible, the shredder residue should be treated in an incinerator with advanced air pollution control. Mineralization will increase lead concentration by at least a factor of 10. Many materials are well suited for such accumulation. The region could offer to take back filter residues from MSW incineration and to accumulate these materials together with car shredder residues. Intermediary lead stocks are distinctly different from fluff or MSW in landfills. They are highly concentrated in metals, and the chemical form is such that economic metallurgical reuse is facilitated. Hence, solidification with cement is not recommended. The intermediary stocks are engineered sites that are designed and constructed to last for a predefined period during which they have to be maintained. The period is calculated according to economic considerations. Due to the economy of scale, and depending on the technology applied, there is a minimum size required for economic reuse of materials. This minimum size divided by the waste generation rate yields the time span needed for accumulation of an amount of material for economic recycling.
2. Accumulation and mineralization for reuse ensure that most lead is controlled within the anthroposphere and no longer poses a threat to the environment. Information about anthropogenic lead flows allows identification of those processes that emit lead. Based on the regional flows and depositions and on dispersion models, the acceptable flows and depositions for lead in water and soil can be calculated. Acceptable can be defined from a toxicology point of view (limiting value for lead content in water or soil) as well as a precautionary principle point of view (lead input into soil or water equals lead output). In any case, concentrations and flows of lead have to be taken into account. Also, potential accumulation of lead in downstream regions needs to be considered. Information about input flows and acceptable output flows of processes is useful in designing transfer coefficients that ensure regional environmental protection over long periods of time.
3. Monitoring based on materials accounting allows one to track accumulation or depletion as well as harmful flows of lead. Efficient monitoring points are as follows:
1. The products construction steel and filter residue of the smelter. These two goods are routinely analyzed for production and quality control purposes. The results allow the determination of lead in shredded cars and indicate whether a change in landfilled lead is to be expected.
2. Concentration of lead in gasoline. This figure is supplied by gasoline producers.
3. Filter residues from MSW incineration. This information combined with known transfer coefficients allows calculation of lead flows in MSW.
4. Sewage sludge. Routinely sampled and analyzed sewage sludge yields information about the sewage network as a potential source. This analysis is instrumental for identifying new emissions or for confirming that loadings to the sewer have been successfully eliminated.
5. Surface waters. For water quality assessment, sampling surface water at the outflow of the region yields adequate information about the total load of the hydrosphere, especially if the same information is available from upstream regions. Monitoring of soil samples may be adequate initially to get an overview of lead in soils. However, as mentioned in Section 3.1.1, routine monitoring by soil sampling is expensive and inefficient, and it does not allow early recognition of harmful accumulations or depletions in soils.

3.5.2  Case Study 18: Accounting of Phosphorus as a Tool for Decision Making

This case study exemplifies how materials accounting can be performed on a routine basis, thereby increasing the power of MFA to understand complex systems and to detect fields of action for the optimization of a region’s metabolism. If the MFA of a region is periodically repeated (e.g., yearly), a resource accounting scheme is obtained. Zoboli, Laner, Zessner, and Rechberger (2016) established a retrospective accounting scheme for the region Austria and the resource phosphorus (P) by compiling yearly P budgets from 1990 to 2011 to demonstrate the feasibility of such a scheme. Their work delivered several important findings:

First, workload and number of budgets (years) are not linearly correlated. Most of the time had to be used to establish the basic system and identify the data sources. Once this is accomplished, the budgets for adjacent years were produced comparably fast.

Second, even in a relatively short and economically stable period of 22 years, the national P budget of Austria, consisting of 122 flows and 8 stock change rates (Figure 3.67), has undergone unexpected significant and partially abrupt changes. This is illustrated in Figure 3.65, where the outcomes of the analysis of the degree of change of the budget with respect to the reference year 1990 are shown (Figure 3.65a). Here, flow changes are divided into three categories, namely, constant, moderate, and extreme. Constant means that a flow did not change since 1990 (rather unrealistic). Extreme means that a flow more than doubled or more than halved compared to 1990. Changes in between are considered as moderate. In order to compare uncertain flows of different years, different tolerance levels were applied (see Figure 3.65). For example, the P flow via import of mineral fertilizer to Austria in 1990 was 44.000 t/year ± 8%. In 2003, the same flow amounted to 32,000 t/year ±8%. Applying tolerance levels of ±0% to 15% and ±σ would classify the change as moderate, while ±20% and ±2σ would give no (significant) change as a result. Consequently, the results are partly sensitive to the applied tolerance levels. If tolerance levels between 0% and ±5% are applied, Figure 3.65 indicates that one-third of the flows and stock change rates changed moderately, and two-thirds were affected by an extreme variation, whereas with ranges from ±10% to ±20%, the fraction of moderately changing flows and stock rates gradually decreases until 15%. The specific standard deviation shows outcomes very similar to the ±20% range, whereas the level of twice the standard deviation decreases both the extreme and moderate fractions to 50% and 5%, respectively. In conclusion, the analysis reveals that half of the flows and stock change rates changed substantially, with certain flows that appeared or disappeared and others that at least doubled or halved their initial value.

Figure 3.65   Degree of temporal change of 122 flows and 8 stock change rates: (a) categorization according to the change with respect to the reference year 1990; (b) categorization according to annual change. Results are shown for different tolerance levels (uncertainty thresholds used to determine whether temporal changes can actually be detected or not). The y-axis indicates the number of flows and stock change rates in each category. (From Zoboli, O. et al., Added values of time series in material flow analysis: The Austrian phosphorus budget from 1990 to 2011. Journal of Industrial Ecology, 2015. doi: 10.1111/jiec.1238.)

The second component of this analysis (Figure 3.65b), instead explores to what extent the flows and stock change rates changed from a given year to the following one, to provide an overview of whether the changes took place gradually or rather abruptly. This analysis reasonably suggests that a large proportion of the flows were affected by gradual and moderate changes, but between 24% and 33% of the flows (depending on the considered tolerance level) recorded at least one extreme variation, indicating the noteworthy presence of substantial and sudden changes. The outcomes also highlight the difficulty of detecting smaller annual changes when uncertainty ranges are applied. However, the main conclusion from this analysis is that national anthropogenic material systems tend not to be stable over time; at least, for P this is the case. This means that classical 1-year MFA studies help to get a common understanding of a system’s metabolism but have to be regularly updated for robust decision making. Zoboli’s work shows that such updating is feasible. Additionally, the multiyear approach also improves the understanding of a system and helps making the model more comprehensive and more suitable to constitute the basis of materials accounting and monitoring.

The analysis of MFA time series directly leads to relevant actions in decision making. This is demonstrated in Figure 3.66 for phosphorus and Austria. In the upper-left diagram, one can see that the total P inputs into the Austrian waste management sector increased considerably since 1990. One of the major tasks of waste management is to collect materials; therefore, such a development can be regarded as positive, in any case showing the rising importance (and responsibility) of the sector. On the other hand, the upper-right time series of Figure 3.66 reveals that large amounts of the waste P are lost in landfills and in concrete. The latter is due to cocombus-tion of sewage sludge and meat and bonemeal (slaughter waste) in cement kilns. Comparing the two time series reveals that the ratio of losses versus input rather increased over the years, a clear negative trend that requires counteraction(s).

Figure 3.66   While the import of P into waste management increased constantly over the past years, (a) the losses of P to landfills and cement increased even more, (b) indicating a clear field for required action. The losses of P from agriculture to the hydrosphere remained rather constant, (c) indicating that efforts for optimized fertilizing and farming practice have been rather inefficient. Contrarily, (d) the emissions of P from wastewater treatment works could be reduced significantly, showing the effectiveness of technical solutions. (From Zoboli, H., Novel approaches to enhance regional nutrients management and monitoring applied to the Austrian phosphorous case study (PhD Thesis). Vienna: Technische Universität Wien, 2016.)

The other two time series of Figure 3.66 provide information on emissions of P to the hydrosphere. While emissions from point sources (here, wastewater treatment plants) could be substantially reduced, areal emissions, which stem from agricultural soils (diffuse, nonpoint sources), stayed rather constant and are now even becoming dominant. A general conclusion is that point sources are easier to control than areal emissions, a finding that has been made several times before, e.g., by Bergbäck (1992), for some heavy metals. The specific conclusion for P is that effective water protection has to put more emphasis on the agricultural sector. This is another hint for decision makers where action (adequate policy) is required.

Zoboli and colleagues (Zoboli, Zessner, and Rechberger, 2016) determined how and to what extent the management of P in Austria could be optimized. They used a detailed national model, obtained for the year 2013, as a reference system (Figure 3.67). Then they selected a range of measures of decision making aimed at reducing consumption, increasing recycling, and lowering emissions of P and discussed them with regard to applicability and limitations. The potential effect of each field of action on the reference system was quantified and compared using three indicators: import dependency, mineral fertilizer consumption, and emissions to water bodies. Table 3.48 presents the potential gain that can be achieved through each field of action, expressed as percentage of the indicators values in the reference year 2013.

Figure 3.67   Austrian phosphorus balance for the reference year 2013 (unit: tP/year). (From Zoboli, O., Zessner, M., and Rechberger, H., Science of the Total Environment. 565, 313–323, 2016.)

Table 3.48   Relative Effect of the Fields of Action on the National P Management, Expressed through Three Indicators

Field of Action

Scope for Reduction of

Import Dependency

Scope for Reduction of Mineral Fertilizer Consumption

Scope for Reduction of Emissions to Water Bodies

Main Data Gaps

Main Challenges

Increase of P recycling from meat and bonemeal

16%

23%

P concentration

Legal framework and market uncertainties for recovered fertilizers

Increase of P recycling from sewage sludge

23%

32%

Performance and product quality for new recovery technologies

Legal framework and market uncertainties for recovered fertilizers

Increase of P recycling from compost

11%

15%

Current use shares; P concentration

Regulation/coordination of sales in large number of composting plants

Increase of P recycling from digestates

Feedstock amounts and composition

Large number and heterogeneity of biogas plants

Increase of P recycling from biomass ashes

2%

3%

-

Current recycling rate; ash quality

Lack of economic incentives that offset logistical costs

Increase of P recycling from manure

Livestock excretion factors; use efficiency of manure as fertilizer

Improvement of municipal and industrial organic waste management

2%

3%

P concentration in MSW; current use of industry, by-products; food waste prevention potential

Resistance of households and similar establishments to further increase separate collection; increase of logistical effort and costs for the municipalities

Achievement of a balanced and healthy diet

20%

5–6%

Complexity of system feedbacks

Resistance to behavioral change; opposition of meat producers

Increase of the use efficiency in crop farming

8%

11%

Livestock excretion factors; P concentration in crops

Optimization of P content in feedstuff

20%

Current state of optimization; complexity of system feedbacks

Reduction of P use in detergents

4%

2%

Reduction of P use in other industrial processes

Materials flows in industrial applications

Substitutability of P

Reduction of surplus accumulation in private and public green areas

11%

15%

Home composting; sales of compost to privates

Resistance to behavioral change; coordination of large number of people

Reduction of point discharges

10%

Loads and perform, of in situ industrial treatment plants

Higher Fe levels in sewage sludge would pose a problem for several P recovery technologies

Reduction of erosion from agricultural soils

12%

17%

13%

Retention processes; long-term behavior of “legacy” P

Implementation at large scale; identification of hot spots

Indicator value in 2013

18,600 tP/y 2.2 kgP cap−1 y−1

13,200 tP/y 1.6 kgP cap−1 y−1

4600 tP/y 0.54 kgP cap−1 y−1

Source: Zoboli, O., Zessner, M., and Rechberger, H., Science of the Total Environment. 565,313–323, 2016.

Note: Percentage values indicate the estimated improvement with respect to the reference year 2013.

Figure 3.68   Optimized Austrian phosphorus balance based on the reference year 2013 (unit: tP/year). Objectives for the optimization are reduction of import dependency, consumption of mineral fertilizers, and emissions to water bodies. (From Zoboli, O., Zessner, M., and Rechberger, H., Science of the Total Environment. 565, 313–323, 2016.)

In a next step, all the gains that could be obtained through the measures (fields of action) were integrated in the reference system to generate an ideal target system (Figure 3.68). The fact that this is characterized by an extremely low import dependency of 0.23 kgP cap−1 y−1 (2.2 kgP cap−1 y−1 in 2013), zero consumption of mineral fertilizer for domestic use, and a 28% decline of emissions to water bodies indicates that governance in Austria offers a large scope for P stewardship.

The systemic approach of MFA in this study allowed quantification of the relative effect of each field of action on the national performance measured with different indicators, and thus performance of a proper comparative assessment. Further, it has made possible the generation and visualization of a target system, obtained through the integration of all potential gains in the reference model. The resulting concise though exhaustive overview can be very useful in supporting decision makers in designing national governance strategies and setting priorities, as well as in assisting domain experts in fitting their work into a broader context. As a next step, such studies need to be complemented with the analysis of the different costs involved in implementing each field of action—therefore another need and chance for interdisciplinary research.

Problems—Section 3.5

Problem 3.14:

• Taking the lead example in Figure 3.1 as a starting point, establish an MFA for cadmium in the same region, assuming no major industrial application of cadmium. Use data given in this handbook, such as Table 3.29 and Figure 3.50, and data from the Internet on cadmium in soils, MSW, etc. Assume that MSW from 280,000 persons is incinerated in the region.

1. What are the major flows and stocks of cadmium in the region with “old” incineration and air pollution control technology (transfer coefficient to air = 0.10)?
2. How do these flows and stocks change if advanced air pollution control equipment is applied and the transfer coefficient is changed to 0.00001?
3. Evaluate environmental and resource implications arising from the two technologies in the region.

The solutions to the problems are given on the website http://www.MFA-handbook.info.

Use of cookies on this website

We are using cookies to provide statistics that help us give you the best experience of our site. You can find out more in our Privacy Policy. By continuing to use the site you are agreeing to our use of cookies.