Lake Superior LaMP

Appendix A
Compound Estimates and Assumptions

This appendix documents the data sources and assumptions used to characterize the compound emission, use, and disposal estimates provided in chapters 1 and 5 of this report. The appendix is organized in three subsections:

• Appendix A.1: Mercury Emission and Disposal Estimates
• Appendix A.2: PCB Use Estimates
• Appendix A.3: Dioxin Emission and Disposal Estimates

The assumptions and data sources underlying the pesticide collection information are documented in chapters 1 and 5.

This section is organized into two subsections: A.1.1, U.S. mercury emission and disposal estimates and A.1.2, Canadian mercury emission and disposal estimates. Following the tabular summaries of the emission and disposal estimates (Tables A.1 and A.2) in each section is a description of the specific data sources and assumptions supporting each estimate.

It is assumed that the final disposition of 10 percent of mercury in total Commercial/Municipal effluent is in sludge (Lohse-Hanson 1999). Therefore, not including the WLSSD, there was 4 kg/yr of mercury in sludge in 1990 and 4 kg/yr of mercury in sludge in 1999.

General and Petroleum refining: The 1990 estimates were used (LSBP 1999).

Copper: White Pines closed (Michigan Mercury Pollution Prevention Task Force 1996)

Iron: Taconite production estimates for Minnesota (Jiang 1999)

Oil: 1990 estimates were used (LSBP 1999).

Natural Gas: 1990 estimates were used (LSBP 1999). The following facilities use natural gas: Hibbing Public Utility, Duluth Steam Plant, GLT-Cloquet, NNG-Carlton, NNG-Wrenshall, USG, Georgia Pacific, and Louisiana Pacific.

Wood

The 1999 estimate is 1 pound/year (LSBP 1999)

MN Power ML Hibbard estimate (3 pounds/year) is based on 1995 emission estimates (Hagley 1999).

Louisiana Pacific and Georgia Pacific emission estimates based on 1998 estimates for the amount of wood burned and emission factor for wood-burning unit with electrostatic precipitators (ESP) control devices. Louisiana Pacific has ESP and catalytic afterburner for 14,289 tons of wood and a centrifugal collector and fabric filter for 5,026 tons of wood. Georgia Pacific has a multiclone and ESP for 6,327 tons of wood and a ESP on 8,789 tons of wood (Kim 1999). An emission factor was only available for ESP control (2.6 * 10-6 pound/ton) (EPA 1997).

Louisiana Pacific: (2.6 * 10-6 pound mercury/ton) * 19,315 tons/year = 0.502 lb mercury/year = 0.023 kg mercury/year

Georgia Pacific: (2.6 * 10-6 pound mercury/ton) * 15,116 tons/year = 0.039 lb mercury/year = 0.018 kg mercury/year

Coal

1990 estimates were based on Minnesota statewide figures, extrapolated to the population of the Lake Superior basin (Tetra Tech Inc. 1996)

1999 estimates are based on facility-specific information for the Lake Superior basin

1997 mercury emissions for LTV Mining (50 lb/yr), MN Power Laskin Units 1 (17 lb/yr) and 2 (16 lb/yr), Northshore Mining Company (26 lb/yr), and Potlach Corporation (<3 lb/yr) (Oliaei 1999)

1998 emissions for NSP Bayfront (2.3 lb/yr) and University of Wisconsin Superior (1.215 lb/yr) (Cabrera-Rivera 1999)

1995 emissions for City of Marquette (16 lb/yr) (City of Marquette 1997) and 1998 emissions for Wisconsin Electric (150 lb/yr) (Michaud 1999)

1998 emissions for Hibbing Public Utility based on amount of subbituminous coal used in cyclone and spreader stoker units (Kim 1999) multiplied by an emission factor for ESP control (EPA 1997)

64,931 tons/year * (0.052 * 10-3 lb mercury/ton coal) = 3.38 lb

mercury/year = 1.53 kg mercury/year

1998 emissions for the Duluth Steam Plant based on amount of pulverized coal used in a dry bottom unit that has a multiclone with a fabric filter (Kim 1999). An emission factor was used for bituminous coal with multiclone control (EPA 1997).

38,198.26 tons of coal/year * (0.78 *10-3 lb mercury/ton coal) = 29.79 lb

mercury/year = 13.51 kg mercury/year

WLSSD: 1999 estimates were provided by the WLSSD (Tuominen 1999).

Small incinerators: 1990 estimated emissions were moved to the use and disposal category for 1999, since most incinerators in this category have ceased operating since 1990.

Other sludge: 1990 estimates were used (LSBP 1999).

Medical waste: Michigan has two medical incinerators remaining with no control devices (Troutman 1999), Minnesota has no medical incinerators remaining in the Basin (Lohse-Hanson 1999), and Wisconsin has no medical incinerators remaining in the Basin (Larson 1999). The 1999 emission estimate was determined by multiplying the amount of medical waste burned by the emission factor for medical waste with combustion control (EPA 1997). This emission factor was the most conservation emission factor available.

Escanaba Hospital: 16.3 tons of medical waste/year * (74 * 10-3 lb mercury/ton waste) = 1.21 lb mercury/year = 0.55 kg/year

Crystal Falls Hospital: 12.3 tons of medical waste/year * (74 * 10-3 lb mercury/ton waste) = 0.412 lb mercury/year = 0.41 kg/year

Cremation: The 1999 estimate was determined by calculating what percentage the Basin population [425,548] (Tetra Tech Inc. 1996) is of the total Michigan, Minnesota, and Wisconsin 1998 population [19,766,161] (U.S. Census 1998). This percentage (2.15%) was multiplied by the number of total projected cremations in Michigan, Minnesota, and Wisconsin for 2000 [46,569] (EPA 1997) to obtain the total number of cremations in the Basin. The number of cremated bodies [1,002.6] was multiplied by the emission factor of 1.50E-03 kg/body for cremation (EPA 1997).

425,548/19,766,161 = 2.15%

.0215 * 46,569 = 1,002.6

1,002.6 bodies/yr * 1.50E-03 kg mercury/body = 1.50375 kg mercury/yr

Batteries: A Hennepin County study showed about a 90-94 percent decrease since the early 90’s (NEMA 1999). In addition, the volume of mercury used in batteries has declined by over 95% (Ross & Associates 1994). Battery sorting studies have shown about a 95% decrease in mercury content since the late 1980’s (Erdheim 1999). Therefore, 1990 estimates were decreased by 90 percent.

Electric lighting :

Air emissions: The 1999 estimates are based on a population extrapolation and Minnesota mercury emission estimates from fluorescent lamp breakage for 2000 [9.07 kg/yr], which are based on the proportion of lamps not recycled and industry figures on mg/lamp (MPCA 1999). A U.S. Basin population of 425,548 was used (Tetra Tech Inc. 1996).

9.07 kg/yr/ 4725419 people in MN = 0.816966 kg/person/yr

0.816966 * 425,548 = 0.82

Disposal/use: The average mercury content of a four foot lamp in 1994 was 22.8 mg; the National Electric Manufacturers Association expects the mercury content of a four foot lamp to be < 12 mg [ 47% decrease] by 2000 (EPA and Environment Canada 1998c). Therefore, 1990 estimates were decreased by 47% to obtain 1999 estimates.

Thermometers, thermostats, light switches, pigments: 1990 estimates were used (LSBP 1999).

Paint and Fungicides: Paint registrations were canceled in 1991 and fungicides were canceled in 1993 (Ross and Associates 1994).Commercial/Municipal

WLSSD: 1999 estimates provided by the WLSSD. Half of sludge being generated is applied to land (Tuominen 1999).

Landfills; dental uses, hospitals, and labs; and residential and other: 1990 estimates were used (LSBP 1999).

Forest Products: The 1999 estimate includes 1995 estimates for Kimberly Clark, Avenor-Thunder Bay, Abitibi Price - Prov. Paper, Abitibi Price Fort William, Northern Wood Preserves, Norampac Packaging-RR, Weldwood of Canada Ltd., and Fort James-Marathon (Brigham 1999

Mining : The Algoma Steel Plant in Wawa, Ontario closed. The 1999 estimate includes the 1995 estimate for Williams Operations gold ore (Brigham 1999).

Metal Finishing and Photoprocessing: The 1990 estimates were used (LSBP 1999).

Oil, Natural Gas, Wood, and Coal: The 1990 estimates were used (LSBP 1999).

Medical waste: The 1999 estimate includes 1995 estimates for St. Joseph’s General and McClausland hospitals (Brigham 1999).

Cremation: The 1999 estimate includes 1995 estimates for Riverside Cemetery and Sunset Crematorium (Brigham 1999).

Batteries - A Hennepin County (in Minnesota) study showed about a 90-94 percent decrease since the early 90’s (NEMA 1999). In addition, the volume of mercury used in batteries has declined by over 95% (Ross & Associates 1994). Battery sorting studies have shown about a 95% decrease in mercury content since the late 1980’s (Erdheim 1999). Therefore, 1990 estimates were decreased by 90 percent.

Electric lighting :

Air emissions: The 1999 estimates are based on a population extrapolation and Minnesota mercury emission estimates from fluorescent lamp breakage for 2000 [9.07 kg/yr], which are based on the proportion of lamps not recycled and industry figures on mg/lamp (MPCA 1999). A U.S. Basin population of 150,000 was used (Thompson 1994).

9.07 kg/yr/ 4725419 people in MN = 0.816966 kg/person/yr

0.816966 * 150,000 = 0.29 kg/year

Disposal/use: The average mercury content of a four foot lamp in 1994 was 22.8 mg; the National Electric Manufacturers Association expects the mercury content of a four foot lamp to be < 12 mg [ 47% decrease] by 2000 (EPA and Environment Canada 1998c). Therefore, 1990 estimates were decreased by 47% to obtain 1999 estimates.

Thermometers, Thermostats, Light switches, Pigments, Instruments (other): 1990 estimates were used (LSBP 1999).

Paint and Fungicides: Paint registrations were canceled in 1991 and fungicides were canceled in 1993 (Ross and Associates 1994).

Wastewater Treatment Plants, Runoff, Landfills, Pharmaceuticals: 1990 estimates were used (LSBP 1999).

Hospital/Medical : The 1990 estimate was used (Brigham 1999).

This section is organized into two sections. Section A.2.1 summarizes PCB use estimates for the U.S. portion of the Lake Superior Basin, and section A.2.2 provides documentation for PCB use in the Canadian portion of the basin.

Methods used to extrapolate MPCA capacitor and transformer data to Lake Superior Basin:

Population of Minnesota in Lake Superior Basin: 232,928

Minnesota MPCA data:

Number of capacitors > 500 ppm (Minnesota Power) 2935

Number of capacitors > 500 ppm (other industry/utilities) 418

Number of transformers and capacitors < 500 ppm 195

Capacitors > 500 ppm PCB per capita (industry/utilities other than MN Power), 1.79x10-3

Transformers and capacitors < 500 ppm PCB per capita, Minnesota 8.73x10-4

Capacitors > 500 ppm PCB in Basin (1.79x10-3 x 232,928 + 2935 MN Power) 3353

Transformers and capacitors < 500 ppm PCB in Basin (8.73x10-4 x 232,928) 195

Method for determining the mass of PCB in U.S. portion of Basin from transformers > 500 ppm PCB and all capacitors:

Capacitors > 500 ppm 3 gallons & 175,000 ppm each

Transformers < 500 ppm**

95.5% 15 gallons & 150 ppm each

0.5% 2500 gallons & 250 ppm each

Transformers > 500 ppm 15 gallons & 550 ppm each

* Equipment volume and concentration estimates based on personal communication with Gene Beadey, Minnesota Power PCB Program Manager (Beadey 1999)

** also applied to capacitors < 500 ppm

Calculations to find mass of PCBs

# caps > 500 ppm 326

Volume of caps > 500ppm (3353 x 3 gal) 10,059 gal

Volume of caps > 500ppm (10,059 gal x 3.785 liters/gal) 38,077 liters

Mass PCB*** (38,077 liters x 175,000 ppm [mg/l] / 1000000 mg/kg) 6664 kg

# caps & tfs < 500 ppm 195

Volume of tfs < 500 ppm, 15 gal (195 caps x .955 x 15 gal) 2793 gal

Volume of tfs < 500 ppm, 15 gal (2793 gal x 3.785 liters/gal) 10,574 liters

Mass PCB***, 15 gal (10,574 liters x 150 ppm [mg/l]/ 1000000 mg/kg) 1 kg

Volume tfs < 500 ppm, 2500 gal (195 caps x .005 x 2500 gal) 2438 gal

Volume tfs < 500 ppm, 2500 gal (2438 gal x 3.785 liters/gal) 9227 liters

Mass PCB***, 2500 gal (9227 liters x 150 ppm [mg/l]/ 1000000 mg/kg) 2 kg

*** assuming ppm = mg/l, thus density of oil = 1

Note regarding U.S. treatment of PCB generating processes

U.S. EPA has concluded that the quantity of PCBs inadvertently generated and released into the environment is inconsequential compared to releases from items with intentional PCBs and, therefore, did not ban these processes. However, U.S. EPA did add certification, recordkeeping and reporting requirements to the facilities that inadvertently produce PCBs. (EPA 1998a)

1997 data for Canada are from Brigham (1999)

In Canada, quantities are reported as PCB contaminated materials and fluids. Liquids are generally reported in litres. Conversion to kilograms was made assuming 1.15 kg/litre.

1990 data for Canada were taken from the Stage 2 LaMP.

Data for total quantity destroyed in Canada are from pgs 30 and 37 of the Zero Discharge report, adding all of the data for the provincially monitored sites and the total for the federally monitored sites. However, pg 36 of the Zero Discharge report provides higher quantities for provincially monitored sites (in the summary table) and would result in a total of 435,949 kg destroyed between 1990-1997, a difference of 91,918 kg. The data presented are for provincially monitored and federally monitored sites and are not presented by sector.

The total amount of PCBs in use in Canada in 1997 is drawn from the Zero Discharge report, pg 31, indicating the total quantity of high level PCB liquids only. It is not known whether there is an additional quantity of low level PCB liquids still in use in 1997.

Though it would appear that Canada has already exceeded the reduction goals for 2005 based upon the quantity destroyed 1990- 1997 (as presented in the Zero Discharge report) and the baseline quantity in use and storage in 1990 (as presented in the Stage 2 LaMP), there is an additional 96,012 kg in use and storage in 1997 (as presented in the Zero Discharge report). The reason for this discrepancy is not known, though it may be the result of the discovery of additional PCB storage and use since completion of the 1990 inventory.

High level liquid and solid PCB materials are defined as containing greater than 10,000 ppm PCBs.

Low level liquid and solid PCB materials are defined as containing 50-10,000 ppm PCBs.

The federally monitored sites do not report whether the stored materials are high or low level waste and, therefore, it is all classified as high level waste.

This appendix is organized in two sections. Appendix A.3.1 summarizes dioxin emission and disposal estimates for the U.S. portion of the Basin, and Appendix A.3.2 provides estimates for the Canadian portion of the Basin.

Table A.3.1 summarizes U.S. estimates for the 1990 baseline and 1999.

a Estimate of dioxin presence in soils at one site in the U.S. portion of the basin. This is not an annual release.

Industrial

Forest products: Dioxins are generated in pulp and paper mills from the paper bleaching process, especially in plants using elemental chlorine as a bleaching agent. In recent years, pulp mills in the Basin have modified their bleaching processes by substituting chlorine dioxide for elemental chlorine, thereby virtually eliminating dioxins from pulp and paper mill effluents (Stromberg et. al. 1996). However, low level monitoring data were not available to assess the degree to which dioxin effluent concentrations have declined since 1990 for the five pulp and paper mills in the U.S. portion of the Basin (two of which discharge directly to the lake). As a result, the 1990 baseline estimate of 0 to 0.6 g TEQ/yr included only the two facilities discharging to Lake Superior, one of which has since closed. The other three mills discharge to Western Lake Superior Sanitary District (WSLSSD). The 1999 estimate has been reduced to 0 to 0.3 TEQ/yr.

Petroleum refining: Dioxins can be formed when catalysts used in petroleum refining are reactivated by burning off coke deposits at 380 degrees C to 525 degrees C in the presence of chlorinated compounds (Bear et. al. 1993). Prior to 1991, 1.5 x 10-5 g TEQ/yr was measured in the effluent of the Murphy Oil facility in Superior, Wisconsin. The dioxin in the effluent was thought to be associated with the regeneration of the catalyst reformer. Waste from this process has since been segregated and is disposed in a hazardous waste facility (LSBP 1996). As a result, dioxin emissions in effluent are assumed to have been reduced to below measurable levels in 1999.

Wood preserving: Past industrial use of pentachlorophenols (PCP) to treat timber, railroad ties, and utility poles are a potential source of dioxins in the Basin (Tetra Tech 1996). The estimate of dioxin contamination in soil is based on an estimate of pentachlorophenol present in soils in the vicinity of the Koppers Inc. facility in Superior, Wisconsin. The facility used PCP to treat railroad ties until 1979. Characterization studies under Resource Conservation and Recovery Act (RCRA) corrective action are ongoing at the site.

Mining: Non-ferrous metal, especially copper, smelting and refining are a known source of dioxin emissions accounting for approximately 1.36 x 10 -2 lb/yr TEQ air emissions in the United States (EPA 1997). In the U.S. portion of the Lake Superior Basin, the Copper Range, White Pine Mine smelter operated in Northern Michigan until 1995. With the closure of the White Pine mine smelter, dioxin emissions from copper smelting were eliminated from the U.S. portion of the Basin.

The combustion of wood and coal as an energy source for industrial and residential use is a known source of dioxins (EPA 1997). Increased attention has been devoted over the past several years to estimate the dioxin emission factors associated with these processes. Table A.3.2 provides estimates of the wood and coal combustion rates in the U.S. portion of the LSB and the current emission factors used to estimate dioxin TEQ emissions from those sources.

a Adapted from Tetra Tech (1996).

b EPA 1998

c Tetra Tech 1996

d 1 ng = 10-9 g

Burn Barrels: In the 1990 baseline estimate, private household waste incineration was not assessed as a source of dioxin air emissions because of an absence of data to characterize the source. In the past several years, additional research has found that household “burn barrels” may be a significant dioxin source. WLSSD (1992) estimated that burn barrels produce 20 times more 2,3,7,8-TCDD per unit of household garbage burned than a controlled incinerator (e.g., a municipal waste combustor (MWC)). Lemieux (1998) estimated that 1.5 to 4 households that burn their waste in the open (e.g., in burn barrels) equal the dioxin generating potential of a fully-operational MWC. Overall, household waste combustion in burn barrels appears to be an overlooked, but potentially significant source of dioxin and other toxic air emissions.

The average person in the U.S. generates between 800 and 1,350 pounds of household waste in a year (MDEQ 1999). The U.S. EPA estimates that 40 percent of people living in non-metropolitan areas burn their waste and that 63 percent of their daily waste is burned in burn barrels. Nationally, this amounts to over 1.8 billions pound of household waste burned in burn barrels every year. Normalized for the U.S. Lake Superior Basin population, this amounts to over 4.5 million pounds of household waste openly burned in the Basin each year.

While such household waste burning is suspected to be a significant source of dioxin and other toxic air emissions, research findings differ as to the rates of dioxin emission per unit of household waste burned (Cohen 1999). Table A.3.3 summarizes dioxin generation emission factors for several recent studies. The table illustrates that emission rate estimates vary over several orders of magnitude. As a result, these emission factor estimates are provided to illustrate the potential significance of the source. Much additional work remains to be completed to properly estimate the dioxin emissions from household waste burning that is occurring in the Basin.

a Recyclers were assumed to reduce the proportion of newspaper, plastic, and some metals in their household waste.

b Expressed as grams TEQ/yr.

To illustrate the potential magnitude of household hazardous waste burning in the U.S. portion of the Basin, Table A.3.4 applies the Cohen (1999) emission factor to potential household hazardous waste burn rates in the U.S. Lake Superior Basin counties to generate an annual TEQ dioxin emission estimate. Extrapolation of national estimates on burning rates to the Lake Superior basin yields an estimate of about 7g TEQ/yr.

Medical and industrial: In the 1990 baseline estimate, medical and industrial incinerators were estimated to contribute 134 g TEQ/yr in dioxin air emissions. As of 1999, all medical incinerators have been closed in the U.S. portion of the Basin. The remaining industrial incinerators are estimated to account for approximately 83 g TEQ/yr (after Jackson 1993) in air emissions. As a result, dioxin air emissions are estimated to have declined to 83 g TEQ/yr for this sector in 1999.

Small incinerators: In the 1990 baseline, small incinerators (e.g., those operated by schools, apartment buildings , and retailers) were estimated to contribute 235 to 2,274 g TEQ/yr in dioxin air emissions. As of 1999, all small incinerators are assumed to be closed in the U.S. portion of the Basin. As a result, no dioxin air emissions are estimated for this sector in 1999.

WLSSD: The Western Lake Superior Sanitary District (WLSSD) operates the only municipal solid waste incinerator in the Basin (Stage 2 LaMP 1999). Estimated dioxin releases of 0.19 g TEQ/yr are based on stack testing. This incinerator is expected to close in 2000.

Wastewater treatment plant sludge: The WLSSD receives indirect discharges from three pulp and paper mills, as well as other industrial and commercial facilities. In addition, new cotton clothing and other household items have also been found to contain dioxins, which come out in the wash and are discharged to the wastewater treatment facility (Horstmann and McLachlan 1994). In 1990, WLSSD treatment plant sludge contained 0.014 g TEQ. Dioxin TEQ concentrations are assumed to remain constant in 1999.

Pentachlorophenol use: Pentachlorophenol has been used to preserve a variety of commercial products, including textiles and leather goods in the United States and abroad. In the past, pentachlorophenol was widely used as a pesticide although most of those uses are now restricted. Dioxin contamination in pentachlorophenol could contribute as much as 10,500 g TEQ dioxins/yr in the United States (Slants and Trends 1995). Based upon the normalized population of the LSB, approximately 18.0 g TEQ/yr of dioxin are assumed to be found in the Basin. The 1990 estimate was based on this national figure. A 1999 estimate should probably show a decrease because of declining use of pentachlorophenol. However, no updated estimates are available.

Table A.3.5 summarizes the estimated dioxin emissions in the Canadian portion of the Lake Superior Basin 1990 to 1999. The assumptions used to generate these estimates are presented in the following section.

a Contaminated soils - not an annual rate of disposal.

b Resulting from spills - not included in annual disposal estimate.

All 1990 estimates are drawn from the Stage 2 LaMP (LSBP 1999) and are expressed in terms of dioxins and furans, rather than TEQs. As a result, the values are not directly analogous to the U.S. estimates reported in Table A.3.1, unless specifically noted . Emissions and dioxin/furan levels in soil and disposal are assumed to remain constant through 1999 except for the following changes:

Forest Products: Yearly average dioxin concentrations in the wastewater effluent form the kraft mills in the Thunder Bay Region have generally declined form 1990 to 1994, although information on total dioxin load has not been reported. As a result, dioxin load in wastewater from this sector is assumed to remain constant from 1990 to 1999 (Brigham 1999).

Mining/Sintering: The Algoma Ore Division iron sintering plant in Wawa, Ontario closed in 1998, thereby eliminating the 21.8 g/yr in dioxin emissions estimated for this sector in 1990.

Medical: The number of medical incinerators in the Canadian Lake Superior Basin has declined from seven in 1990 to three in 1999 (Brigham 1999). As a result, dioxin emissions are assumed to have declined proportionally to 0.07 g dioxin/yr.

EPA Home | Privacy and Security Notice | Contact Us