4.4.2. Off-Site Transport of Air-borne Contaminants

Estimating the dispersion and resulting exposure site concentrations of air-borne contaminants, originating at the site of contamination in a vapor or a particle phase, requires a different solution for the off-site as compared to the on-site situation. A simplified solution, given as a virtual point source model, can be found in Turner (1970). This model approximates the dispersion that occurs from an area source by using an imaginary point source. This point is located upwind of the actual source at a distance calculated to create a lateral dispersion at the site equal to its width:

Equation V3 4-41

The term, FREQ, has been added to this equation to appropriately account for changing wind directions, and hence, obtain a more accurate annual average air concentration. The vertical dispersion, Sz, is estimated as an empirical function of the distance from the source center to receptor:

Equation V3 4-42a and b

The virtual distance, VD, is an empirical function of the width of the contaminated area and the actual distance from source center to receptor:

Equation V3 4-43

Prior guidance on windspeed (Section 4.3.2) indicated windspeeds ranged from 2.8 to 6.3 m/sec, and suggested a mid-range of 4.0 m/sec in the absence of better information. Where the wind blows from all directions equally, then it will blow from one compass sector about 15% of the time. On these bases, a windspeed of 4.0 m/sec and a FREQ of 0.15 were used in the example scenarios in Chapter 5. In most places, however, wind direction is much less variable, and the appropriate value is best determined with site specific information.

4.4.3. Specific Cases of Off-Site Soil Contamination

This section provides background information on specific sites of soil contamination which have been studied for the presence and impact of dioxin-like compounds. These include landfills used for disposal of ash from municipal waste combustion facilities, the disposal of sludge from pulp and paper mills, and sites of soil contamination typified by the sites monitored in the National Dioxin Study (in many cases, Superfund sites or sites that were in some stage of being considered for inclusion on the NPL list at the time of the study; EPA, 1987). Discussion of these particular sites does not imply that they represent the bulk of such sites nationally, or that they are discussed here based on any critical environmental or exposure rational. They are discussed because they present unique issues for emissions and fate and transport of dioxin-like compounds from sites of soil contamination, and because they have been studied. Issues discussed below are pertinent for other types of off-site soil contamination sites. Landfills receiving ash from municipal waste incinerators

Particular issues regarding landfills receiving ash include: the impact of soil cover on releases, the ash concentrations, the size of such landfills, the quantity of ash generated by incinerators, and the fugitive emissions that result from ash management. Key sources providing information for this section include a methodology document describing approaches to estimating environmental releases and exposures to ash (EPA, 1991), and a contractor report applying these types of methodologies using site-specific data from several ash landfills (MRI, 1990).

Each of the identified topics will be discussed in turn.

. Landfill Cover:
Whether or not ash is covered once it is disposed of at the landfill is critical in determining releases and subsequent exposures. Currently, practices at operating landfills vary from no coverage after disposal on active portions of the landfill to daily coverage of disposed ash. MRI (1990) visited six facilities disposing ash, including ash monofills and municipal solid waste landfills. In one of the facilities, an ash monofill located at the site of the combustor, the disposal area encompassed 15 acres and did not use daily cover until final elevation was reached. At that time, a clean cover of 2 feet of soil would be applied. At a second facility, located at the site of the combustor but landfilling municipal solid waste as well as ash, ash was used for different purposes, including a subbase roadbed material, as soil substitute for earth work, and as a daily cover for MSW receipts. An assumption of bare surfaces (i.e., no vegetation) during the period of landfilling activity, with concentrations of dioxin-like compounds equal to concentrations in the ash would appear to be appropriate assumptions for practices at these two landfills.

Where daily cover is employed, however, appropriate assumptions are not straightforward. Of the remaining four sites studied by MRI, two employed daily covers ("clean cover material" in one case and a "HDPE liner" in the other, sic), and daily coverage practices were not discussed for two sites. Approaches described for airborne emissions and erosion losses would have to be modified when daily cover is applied. First, losses of contaminants via overland soil or wind erosion could not be expected to occur when cover (soil cover or otherwise) is in place, although the active part of the landfill would be subject to erosion during an operating day. Even in that case, however, site-specific practices might include little or no ash disposal during periods of soil-erosion-producing storms. Depending on site-specific practices, one might estimate annual erosion losses using methodologies described in this assessment, and then empirically reduce erosions losses based on these practices and scientific judgement.

Air emissions from active portions of the landfill, as in wind erosion and volatilization, also are obviously impacted by cover practices. These emissions would occur during the actual disposal. Wind erosion and volatilization fluxes could be estimated as given in earlier sections, and then reduced by two-thirds, which might correspond to an assumption of disposal during 1/3 of a day or a year, etc. When covered by soil or a synthetic cover, wind erosion losses would not occur. However, buried residues may diffuse through layers of clean soil and be released via volatilization.

Estimates of volatilization release via diffusion through clean cover have been made. A rigorous approach for such estimates is detailed in Hwang, et al. (1986). Use of this approach requires a computer to iteratively solve a partial differential equation, expressed in terms of a Fourier series. It can be shown, with these equations, that the vapor emission rate through such a cover will not reach steady state for hundreds of years. Hwang's approach was applied to an earlier assessment for 2,3,7,8-TCDD (EPA, 1988b). Calculations were performed for 2,3,7,8-TCDD contamination with a thickness of contamination of 8 ft, and clean caps ranging from 10 to 25 cm. The results of this exercise suggest that the average emission rate of a 70-year period are 1/4 to 1/5 of what they would be without the cap. Based on this exercise, a simple assumption might be made that a clean cap will reduce the average emission rate calculated without a clean cap by 80%. However, these results are not consistent with those described in Jury, et al. (1990).

The analytical solution developed by Jury was demonstrated on 35 organic compounds. One exercise conducted by Jury was to estimate the cover thickness required to restrict volatilization to less than 0.7% of the mass incorporated in soil. For 2,3,7,8-TCDD, the thickness was estimated at 0.7 cm for a sandy soil and 0.2 cm for a clayey soil. This appears to contradict the work of Hwang since it shows an essentially insignificant loss for a cap much less thick than the 10-25 cm cap in the exercises using Hwang's approach. However, Jury's approach allows for assumptions on degradation of the buried compound. For that exercise, Jury assumed that the half-life for 2,3,7,8-TCDD was 1 year.

This is a very rapid degradation rate, given information that the dioxin-like compounds resist degradation, particularly when not exposed to sunlight. On the other hand, the Hwang model assumes no degradation loss, and as such, the generalization from his exercise might be an overestimate. Hwang's exercise might also have overestimated since it assumed a rather thick 8-ft layer of subsoil contamination. From these arguments, it would appear that neither exercise appropriately evaluated the difference in volatilization in a no cover versus a cover situation.

The above discussions concerned flux calculations when cover practices are used. One set of adjustments discussed reduced a total potential flux of volatilized or wind eroded losses based on a portion of the time that the ash would be uncovered. A second discussion indicated that some loss via volatilization might be modeled with a clean cap. In any case, it is clear that cover practices will reduce losses. Cover practices must be considered when evaluating the exposure to ash disposed of in landfills.

. Ash Concentrations:
A key consideration, of course, in modeling transport of dioxin-like compounds from an ash landfill is the concentration on the ash. Ash concentrations of dioxin-like compounds have been found to vary widely, from non-detect (generally less than 0.1 ppb) to the hundreds and thousands of part per billion. Table 4-5 appears in EPA (1991) and summarizes concentrations of dioxin-like compounds and PCBs found in fly, bottom, and combined ash. These data are a summary of 19 references, ranging in publication date from 1974 to 1990. It should be noted that, except for 2,3,7,8-TCDD and 2,3,7,8-TCDF, results listed are for congener groupings defined by degree chlorination.

. Size of Landfill and Amount of Ash Landfilled:
The size of the landfill and the amount of ash applied daily or over time are both required for estimating exposures nearby. These can both be obtained from site-specific observations. Amounts of daily disposed ash are required to estimate fugitive particulate emissions, as will be discussed shortly. Amounts of daily or ultimate disposal are also tied to landfill size, or the portion of a landfill that is active on a daily basis. One common practice is to fill cells of a landfill one at a time, and once filled, to cover with a 2-ft (or so) layer of clean soil.

The appropriate size in this case is the average size of a landfill cell. If daily coverage is applied, than the size for modeling purposes corresponds to the area over which daily coverage occurs. This can also vary depending on the depth of disposal during a day. A six-inch daily coverage, for example, would take twice as much space as a 1-ft depth of daily disposal. If the intent of a day's disposal is to cover over the entire area of an active cell, then depth of coverage need not be considered in determining landfill size.

Determination of landfill size (or the size of the active portion of the landfill) may be required in the absence of site-specific information, such as in the planning stages for a new incinerator. This is where details on landfill management need to be determined. One important detail, as already noted, is the amount of ash generated for daily disposal. Cook (1991) assumes that bottom and fly ash combined comprise about 11% of total receipts on a volume basis. However, a relationship between ash generated and solid waste received by an incinerator on a mass basis is more useful for estimating daily disposal amounts. In a recent EPA (1990f) report on ash characterization, ash mass was estimated as an average of 29.5% of municipal solid waste received in five facilities studies, with a narrow range of 25-35%. ...

table Table 4-5. Ranges of concentrations of PCDDs, PCDFs, and PCBs in municipal waste combustor ash (results in ng/g or ppb).
... This mass was estimated on a wet weight basis. Ash is wetted when exiting the incinerator, and water comprises 20-30% of the total weight at that point. If the ash is immediately trucked for landfill disposal, its total weight includes the weight of this quench water. Often ash is stored at the incinerator site in piles prior to disposal, that storage ranging from hours to days. In this circumstance, much of the quench water would have drained off or evaporated, and then the total weight hauled would be closer to a dry weight estimate. In summary, the amount of ash generated to be disposed of a daily basis can be estimated as: the daily receipt of municipal solid waste (tons) * a wet weight ash fraction (0.25-0.35) * a wet to dry weight conversion if appropriate (wet weight * 0.80, e.g.).
expand table Table V3 4-5

. Fugitive Particulate Emissions:
Fugitive emissions can occur from the time ash exits the incinerator for temporary storage at the facility site (or immediate loading onto trucks for disposal) until ultimate disposal. Approaches to estimate fugitive releases from incinerator ash management are described in EPA (1991), and will be summarized here.

As noted, ash can be wet when exiting the quench tank. If stored at the facility site prior to disposal in a landfill, leaching from piles can occur. Because dioxin-like compounds are strongly hydrophobic, however, the impact of leaching is unlikely to occur much beyond the soil beneath and near the storage piles. If loaded onto trucks when very wet, leaking onto roadways may also occur. If these storage piles are left uncovered, they would of course be subject to erosion losses, which might move residues further from the piles than just leaching of water from the piles.

Of more concern than water-borne losses due to ash management are fugitive emissions of dry ash. Wind erosion, which can occur from open storage piles or uncovered portions of the landfill, is a fugitive emission that has been discussed for soil contamination. Specific practices in the management of ash can also result in fugitive emissions. Such practices include:

1) loading onto and dumping out of trucks,

2) truck transport from the incinerator facility to the landfill site,

3) truck or other traffic over paved or unpaved roadways at the incinerator site, at the landfill site, or other roadways containing contaminated dust, and

4) spreading and compacting of ash at the landfill site.

A set of empirical emission factor equations for estimating fugitive particulate emissions, called "AP-42" equations, have been developed by EPA's Office of Air Quality Planning and Standards (EPA, 1985a; EPA, 1988a). Specifics on applying these equations for ash management are described in EPA (1991). An example of their application using site-specific information for ash management is detailed in MRI (1991). An abbreviated listing of emission factor equations that have been used in these two publications are:

. Vehicular traffic over unpaved roadways.
Dust on the surfaces of roads, both unpaved and paved, can become suspended due to vehicular traffic. When these roadways are near ash storage piles or within the landfill, that dust can become contaminated. The emission factor equation for emissions from unpaved roadways is:

Equation V3 4-44

. Emissions off trucks in transit.
Although no emission factor equations have specifically been developed for trucks while in transit from the incinerator facility to the landfill, such emissions can occur if the ash is dry, and partially or completely uncovered. The following equation for estimating emissions from open storage piles has been suggested for use in estimating fugitive emissions from trucks in transit (EPA, 1991; the emission factor equation from EPA, 1985a). Note that use of this equation will require specific management assumptions in order to estimate the number of uncovered hectares per day: the number of trucks in use per day, the surface area of trucks, the percent of uncovered area if a tarpaulin is used, the moisture content of ash, and so on.

Equation V3 4-45

. Loading and unloading.
The unloading operations at the disposal site may result in the release of fugitive dust. The following emission factor equation provides emission factors for kilograms of particulate emitted per megagram (metric ton, or 1000 kg) of soil loaded and unloaded:

Equation V3 4-46

. Spreading and compacting of ash at the landfill.
An emission factor specifically for ash spreading and compacting has not been developed. However, emission factor equations for similar applications have been applied for estimating fugitive emissions due to spreading and compacting. MRI (1990) used an AP-42 emission factor developed for dozer moving of overburden in western surface coal mines. Kellermeyer and Ziemer (1989) assumed that the spreading and compaction of ash was analogous to vehicular transport on unpaved surfaces, and used the emission factor for that process. A third possible assumption is that the processes of spreading and compacting are analogous to agricultural tillage. That emission factor equation for agricultural tillage is:

Equation V3 4-47

When applying such equations, there are further key issues to consider.

These include:

. Concentrations on fugitive ash emissions:
When such an emission occurs from ash surfaces, such as from storage piles, off trucks in transit, in spreading and compacting, and so on, than there is a good argument to assume that such concentrations on such emissions are "enriched" in comparison to an ash average. The argument here is similar to the argument for enrichment assumed for eroded soils: processes resulting in fugitive air emissions favor lighter particles with more surface area and hence more sites for binding.

No data could be found to assign a value to an ash enrichment ratio. MRI (1990) did, however, take data on municipal waste combustor facility roadway dust, and based on that data and statistical evaluations, speculated that fly ash constituted the principal source of lead and cadmium found on paved surfaces. Since fly ash is finer than bottom or combined ash, one hypothesis for this finding is that fugitive emissions from ash management at the combustor site transported these finer particles to roadway surfaces.

This is not to imply, however, that concentrations in dust suspended from roadways due to traffic should be higher in concentration than concentrations in ash - this enrichment concept only applies to ash surfaces themselves. Rather, the concentration on roadway suspended dust should be lower than on the ash.

This is because contaminated dust on roadways mixes with clean dust from other sources. As noted, MRI (1990) did take roadway dust samples, and their data appears to place such a dilution factor (concentration on roadway dust divided by concentration on ash) in the range of 0.1 to 0.3. Specifically, they took particulate samples from landfill haul routes while at the same time taking samples of incinerator ash being delivered for disposal the same day.

Each paired sample (roadway particulate and ash), were measured for four metals: As, Cd, Cr, and Pb. Several paired samples were taken on both paved and unpaved haul routes. Ratios were then generated for roadway particulate metal concentrations over ash metal concentrations.

Results were:

- paved and unpaved ratios were similar and consistently near 0.1 (roadside particulate concentrations of As were 10% of ash concentrations of As), Cd
- paved and unpaved ratios were similar and ranged between 0.0 and 0.4, Cr
- paved ratios ranged from 0.3 to 0.6, while unpaved had a wide range of 0.3 to 2.0, Pb
- paved and unpaved ratios were similar between 0.0 and 0.2. For analogous situations
- daily deliveries of contaminated ash
- one might assume a dilution factor in the 0.1-0.2 range.

. Selection of values for emission factor equations:
As noted, all these equations are empirical equations. They were developed from data on sites where such emissions occur, such as strip mining sites. EPA (1988a) describes the range of conditions over which such equations were developed. What is meant by "conditions" are such factors as the range of vehicle weights in the data set, the range in number of wheels on such vehicles, and so on. Application of these equations for situations not included within these ranges should be done cautiously.

Very critical also is the selection of the particle size multiplier variable, k. These values range from about 0.10 to no higher than 1.0. Lower k values are used to estimate emissions of the smallest sized particles; generally particles less than 5 m m in diameter. Higher k values are used to estimate emissions of all sized particles less than a higher diameter, usually either 15 or 30 m m. If these equations are used to only estimate particulate inhalation exposures, than the k value corresponding to 10 m m sized particles, or inhalable sized particles, should be used. When used to estimate total emissions, than the highest k value listed should be used. Such estimations are appropriate when also evaluating impacts to off-site soils or vegetation.

. Controls for fugitive emissions:
All these equations were developed when no fugitive emission controls were in place. Common controls for roadway dust suppression include wetting or use of a chemical dust suppressant. Ash transported in trucks is commonly wetted and/or a tarpaulin is used to control emissions off trucks. There is no guidance or data on the effectiveness of such controls, but they must be considered. In demonstrating these procedures, EPA (1991) assumed that controls on emissions resulted in 90% reductions in potential emissions. If a control is known to be in place and used on a regular basis, than this percent reduction is probably a reasonable starting assumption. Land application of sludge from pulp and paper mills

This discussion focuses on an assessment on the land application of sludge from bleached kraft and sulfite pulp and paper mills (EPA, 1990e). Focusing on this source of sludge does not imply that pulp and paper mills produce more sludge than other industries, or that sludge from pulp and paper mills contains more dioxin-like compounds than other sludges.

However, it is known that dioxin-like compounds are found in pulp and paper mill sludges. Also, because of the 104-mill study in 1988, much information is available on the content and disposal of this sludge (further information on the 104-mill study can be found in EPA (1990c) and EPA (1990d)). Some of the issues briefly discussed below for pulp and paper mill sludges would also pertain to sludges containing dioxin-like compounds from other sources.

EPA (1990e) described frequency distributions of concentrations of 2,3,7,8-TCDD and 2,3,7,8-TCDF for 79 mills reporting this information and also broke out the data based on disposal option. Although EPA (1990e) used the disposal option breakout of concentrations in their assessment of the impacts of the various options, it is not felt that the disposal option of choice is based on concentration. Over all options, the median (50% percentile as given in EPA (1990e)) and maximum 2,3,7,8-TCDD concentrations found in sludges were 51 and 3800 ng/kg (ppt), respectively. The median and maximum 2,3,7,8-TCDF concentrations found were 158 and 17100 ppt.

Fate and transport for contaminants is sludge is dependent on disposal means. Of the approximate 2.5 million metric tons of pulp and paper mill sludge generated annually (as estimated in the 1988 104-mill study), five principal options for disposal were noted: landfilling (44% of all sludge disposed), surface impoundments (24%), land application (12%), incineration (12%), and distribution and marketing (8%). Impacts by incineration were not discussed in EPA (1990e) and are not discussed in this section.

Key issues pertaining to each disposal issue are now discussed.

. Landfilling:
The issue of coverage as discussed above for ash landfills is relevant for any landfill. However, fugitive particulate emissions during sludge handling and disposal is not an issue as it was for disposal of ash from incinerators due to the differences in moisture content. Sludge is much higher in moisture at the time it is disposed of in comparison to ash - with moisture contents as high as 90%.

. Surface Impoundments:
It was assumed in EPA (1990e) that sludge disposed of in surface impoundments have a higher moisture content as compared to sludge disposed of in landfills. Surface impoundments were located at the mill site, explaining the assumption for a higher moisture content. A surface impoundment in the EPA (1990e) assessment was defined as a facility in which the sludges are stored or disposed on land without a cover layer of soil. For this type of management, soil cover would not be an issue. Concentrations would be those measured in the sludge. Also, vegetative cover would be expected to be minimal, which would influence parameters associated with soil erosion.

. Land Application:
Twelve percent of all sludge produced annually was land applied. Four of the 104 mills applied the sludge to forest land, two mills land applied the sludge to agricultural land, and two mills used the sludge for abandoned mine reclamation. The high organic matter content (EPA (1990e) assumed a 25% organic carbon fraction in sludge) and high fraction of clay-sized particles make sludge an attractive soil amender. Sludge is either applied to the land surface with or without incorporation. When not incorporated, sludge can be assumed to replace surface soils and concentrations would be those in the sludge. When incorporated, soil concentrations can be estimated simply as (in mg/kg): (mass of contaminant added, mg)/(mass of sludge added, kg + mass of soil in mixing zone, kg).

One key issue when incorporated is the number of years of such treatments. Most of the land application uses of paper and pulp mill sludges reported in EPA (1990e) made applications in only one year. As easily seen in the above suggested equation, higher concentrations result with more years of incorporation. The other key issue with incorporation, of course, is the depth of incorporation. For agricultural applications, the depth of incorporation assumed in EPA (1990e) was 15 cm, similar to the 20 cm incorporation assumed for home vegetable gardening in this assessment. For silvicultural uses, the assumption in EPA (1990e) was 2.5 cm, which corresponds to some but minimal mixing. For abandoned mine reclamation, the assumption was 0 cm incorporation.

Routes of exposure might also vary from focuses in this document depending on land application choice. When applied to agricultural land, impacts to food crops would demand particular attention (the procedures in this assessment were demonstrated with home grown vegetables, although of course impacts to food crops are critical when agricultural field soils are impacted by dioxin-like compounds). When applied to forest land, ecological impacts might warrant particular attention, as was discussed and demonstrated in EPA (1990e). A final issue to consider when land applying sludge to land is a rate of dissipation/degradation of dioxin-like compounds.

Landfills and surface impoundments have ongoing surface applications and over time, the total depth of applications in the range of meters, so an assumption of a constant source strength over a period of exposure, as was assumed in this assessment for soil contamination sources, is reasonable. However, if only a few centimeters of surface soil are impacted, which might be the case for single applications to land and/or surface applications with no incorporation, an assumption of dissipation may be warranted. EPA (1990e) assumed no degradation of 2,3,7,8-TCDD in their assessment of impacts from land applications.

. Distribution and Marketing Uses:
The volume of sludge distributed and marketed was approximately 8% of the total amount of sludge generated for the 104-mill study. For this use, sludge was composted and then sold as a soil amendment in residential, agricultural, and commercial settings. More attention to the dermal contact pathway appears appropriate for this usage. Site-specific factors, and the values for these factors used in EPA (1990e), include:

1) depth of incorporation - 0, 15 and 25 cm in assumptions characterized as high, best, and low estimates,

2) garden size - 0.016 and 0.022 hectares characterized as low/best estimate and high, and referencing a national gardening survey,

3) rate of application - between 5 and 20 dry metric tons per hectare references a USDA publication on use of sewage sludge compost for soil improvement and plant growth, and

4) years of using such compost - 20 without specific reference.

The years of application is needed for estimating soil concentrations during and after the period of exposure, using a simple ratio as discussed above in land application.