PAGE 4 OF 9 Vegetation concentrations

Vegetation concentrations are required for the estimation of exposure to homegrown fruits and vegetables, and also for the beef and dairy food chain algorithms. Three principal assumptions are made to estimate vegetative concentrations:

. Outer surfaces of bulky below ground vegetation are impacted by soils which contain dioxin-like compounds. Inner portions are largely unimpacted.

. Translocation of dioxin-like compounds from roots to above ground portions of plants are negligible compared to other mechanisms which impact above ground portions of plants. As such, translocation into above ground portions will be assumed to be zero.

. Similar to the assumption concerning transport of contaminants from outer to inner portions of below ground vegetation, it will be assumed that outer and not inner portions of above ground bulky vegetation are impacted.

Concentration of contaminants in below ground vegetation is only required for vegetables (carrots, potatoes, e.g.) grown underground. The basis for the below ground algorithm is the experiments of Briggs, et al. (1982) on uptake of contaminants into barley roots from growth solution, and their elaboration of a Root Concentration Factor.

The below ground concentration is given by:

Equation V3 4-23

Two processes, air-borne vapor phase absorption and air-borne particle deposition, are assumed to contribute to above ground vegetation concentrations:

Equation V3 4-24

The basis for a vapor-phase bioconcentration factor for various airborne contaminants, including 1,2,3,4-TCDD, from the atmosphere to vegetation was developed by Bacci, et al. (1990, 1992), with amendments suggested by McCrady and Maggard (1993), and McCrady (1994). Bacci and coworkers conducted laboratory growth chamber experiments on the vapor-phase transfer of 14 organic compounds from air to azalea leaves, and developed a generalized model to predict the vapor-phase bioconcentration factor based on a contaminant Henry's Constant, H, and octanol water partition coefficient, Kow.

A similar experiment by McCrady and Maggard (1993) conducted for 2,3,7,8-TCDD vapor transfer to grass leaves suggested that the Bacci empirical algorithm to estimate the transfer factor would greatly overestimate it. Further details on these experiments are in the section below on this critical bioconcentration parameter, termed Bvpa in this assessment. The algorithm estimating plant concentrations as a function of vapor-phase air concentrations is:

Equation V3 4-25

Several exposure efforts for 2,3,7,8-TCDD (Fries and Paustenbach, 1990; Stevens and Gerbec, 1988; Connett and Webster, 1987; Travis and Hattemer-Frey, 1991), have modeled the accumulation of residues in vegetative matter (grass, feed, vegetables) resulting from deposition of contaminated particulates. Key components of their approach, as well as the one for this assessment, include:

. Vegetative concentrations result from particulate deposition onto plant surfaces.

. Vegetative dry matter yield is the reservoir for depositing contaminants; this reservoir varies according to crop.

. Not all particulate deposition reaches the plant, some goes directly to the ground surface; the "interception fraction", less than 1.0, reduces the total deposition rate. This fraction can be related to the percent ground that is covered by the vegetation.

. Weathering processes, such as wind or rainfall, remove residues that have deposited onto plant surfaces via particle deposition, and this process is reasonably modeled as a first-order exponential loss with an associated weathering dissipation rate.

All the above references have justified a dissipation rate derived from a half-life of 14 days (based principally on field measurements described in Baes, et al. (1984)); this is the value used for all dioxin-like compounds in this assessment as well. As well, a portion of particles depositing as wet deposition are not retained on the vegetation after the rainfall. A retention factor reduces total wet deposition considering this.

. Vegetative concentrations may not reach steady state because of harvesting or grazing, but a steady state algorithm is used.

The steady state solution for plant concentrations attributed to wet plus dry particle deposition is:

Equation V3 4-26

The unit contaminant wet plus dry deposition rate, F, is given as:

Equation V3 4-27

Following is brief guidance on assignment of values to the terms in Equations (4-23) to (4-27).

. Cs and Kds:

This is the soil concentration and soil/water partition coefficient, respectively. The soil concentration is specified for the on-site source category. For the two source categories where soil contamination is distinct from the site of exposure, the soil concentration at the site of exposure is estimated. As discussed in Section 4.4.1 below, two soil concentrations including one for a no-till and one for a tilled situation, are estimated. For estimating below ground vegetable concentration, the tilled concentration is required. The soil partition coefficient is a function of the contaminant organic carbon partition coefficient, Koc, and the soil organic carbon fraction, OCsl, as discussed above in Section 4.3.1. Division of Cs by Kds results in the equilibrium soluble phase concentration of the contaminant, in mg/L.

. RCF:

Briggs, et al. (1982) conducted experiments measuring the uptake of several compounds into barley roots from growth solution. He developed the following relationship for lipophilic compounds tested (lipophilic compounds were identified as those tested that had log Kow 2.0 and higher (n=7, r=0.981):

Equation V3 4-28

Since his experiments were conducted in growth solution, the RCF is most appropriately applied to soil water in field settings. This is why the Cs was divided by Kds in Equation (4-23).

. VGbg:

This correction factor and the one used to correct for air-to-leaf transfer of contaminants, VGag, are based on a similar hypothesis. That hypothesis for VGbg is that the uptake of lipophilic compounds into the roots of this experiments is due to sorption onto root solids. High root concentrations were not due to translocation to within portions of the root hairs. Direct use of the RCF for estimating concentrations in bulky below ground vegetation would greatly overestimate concentrations since an assumption (stated above) is that there is insignificant translocation to inner parts of below ground bulky vegetation for the dioxin-like compounds. Concentrations in outer portions of edible below ground vegetation would mirror concentrations found in barley roots, by this hypothesis.

VGbg can be estimated by assuming that the outer portion, or skin, of below ground vegetables would contain concentrations that can be predicted directly using the RCF, but that the inner portions would effectively be free of residue. The correction factor can be estimated as the ratio of the mass of the outer portion to mass of the entire vegetable:

Equation V3 4-29

Simplifying assumptions are now made to demonstrate this ratio for a carrot and a potato. First, it will be assumed that the density of the skin and of the vegetable as a whole are the same, so the above can become a skin to whole vegetable volume ratio. The thickness of the skin will be assumed to be same as the thickness of the barley root for which the RCF was developed. Without the barley root thickness in Briggs, et al. (1982), what will instead be assumed is that the skin thickness is equal to 0.03 cm. This was the thickness of a leaf from broad-leaved trees assumed by Riederer (1990) in modeling the atmospheric transfer of contaminants to trees.

The shape of a carrot can be assumed to be a cone. The volume of a cone is given as (p /3)r2l, where r is a radius of the base and l is length. Assuming a carrot base radius of 1 cm and a length of 15 cm, the volume is 16 cm3. The curved surface area of a cone is given as: p r(r2 + l2)1/2, which equals 47 cm2, given the r and l assumptions. The volume of the cone surface area is 47 cm2 * 0.03 cm, or 1.41 cm3. The skin to whole plant ratio for this carrot is 0.09 (1.41/16). A similar exercise is done for a potato, assuming a spherical shape with a radius of 3 cm. The volume is given as 4/3p r3, or 113 cm3. The surface area of a sphere is 4p r2, or 113 cm2, and the volume of this surface area is 3.39 cm3. The skin to whole plant ratio for the potato is 0.03.

This exercise indicates upper bounds for such an empirical parameter. For exposure assessments, other factors which reduce vegetative concentrations should also be considered and will be considered in this empirical correction factor in this assessment. Additional reductions in concentration result from peeling, cooking, or cleaning, for example. Wipf, et al. (1982) found that 67% of unwashed carrot residues of 2,3,7,8-TCDD came out in wash water, and 29% was in the peels.

A peeled, washed carrot correction factor might instead be, 0.09*0.04, or 0.004 (0.09 from above; 0.04 = 100% - 67% - 29%). A 96% reduction in the estimated VGbg for the potato (the potato is cleaned and the skin is not eaten; additional reductions possibly when cooking the potato) would equal 0.001. In a site-specific application, the type of vegetation, preparation, and so on, should be considered. The VGbg for underground vegetables for this assessment is assumed to be 0.01. This is less than the estimates of 0.09 and 0.03 for the carrot and potato above, but greater than it might be if based on this discussion on cleaning, washing, peeling, and so on.

. Cva, Cpa:

The vapor-phase concentration of contaminant in air, Cva, used in this algorithm is estimated using procedures described in Section 4.3.2 above. The particle-phase concentration of contaminant in air, Cpa, is estimated using procedures described in Section 4.3.3.

. Bvpa:

Bacci, et al. (1990, 1992) conducted laboratory experiments on the air-to-leaf transfer of vapor-phase concentrations of 14 organic contaminants to azalea leaves. With their results, they developed an empirical relationship for a vapor-phase bioconcentration factor from air to azalea leaves, termed in this assessment the Bvpa, but which was termed BCF by Bacci and coworkers. They related the Bvpa to the chemical octanol-water and air-water partition coefficients, Kow and Kaw. The air-water partition coefficient, Kaw, is a dimensionless form of Henry's Law constant, H, derived by dividing H by the product of the ideal gas constant, R, and the temperature, T. The most general form of the air-to-leaf transfer factor is on a unitless volumetric basis: [ng contaminant/L or leaf]/[ng contaminant/L of air], and is given as:

Equation V3 4-30

Bacci, et al. (1990) showed that the volumetric transfer factor can be transformed to a mass-based transfer factor by assuming that 70% of the wet leaf is water, the leaf density is 890 g/L, and the air density is 1.19 g/L:

Equation V3 4-31

Bacci's experiments were conducted under conditions which would not account for photodegradation of his test chemicals from the leaf surfaces. A recent study by McCrady and Maggard (1993) which investigated the uptake and photodegradation of 2,3,7,8-TCDD sorbed to grass foliage suggests a significant difference in experimental Bvol for grass plants. The authors note that the log Bvol for 2,3,7,8-TCDD and azalea plants, using Bacci's empirical relationship, was estimated as 8.5. The experimental log Bvol for 2,3,7,8-TCDD and grass plants reported by McCrady was 6.9 when photodegradation was accounted for, and 7.5 in the absence of photodegradation. Since the photodegradation experiments by McCrady best represent outdoor conditions, their work suggests that the air-to-leaf transfer factor estimated by Bacci's algorithm may be 40 times too high for vapor-phase transfer of 2,3,7,8-TCDD onto grass leaves.

While McCrady's experiments included consideration of photodegradation of 2,3,7,8-TCDD, it is uncertain as to how their results can be generalized to other dioxin-like compounds and vegetations other than grass. There is very little information in the literature on the photodegradation of dioxins and furans on plant surfaces. McCrady and Maggard (1993) cite Crosby and Wong (1977) as the only other work measuring photodegradation of 2,3,7,8-TCDD from leaf surfaces. In that work, 2,3,7,8-TCDD was applied as a 15 ppm concentration in Agent Orange, and McCrady speculated that the rapid photodegradation measured in those experiments occurred because the herbicide formulation contained carriers and organic solvents that may have promoted photodegradation.

Some experiments conducted in organic solvents (Crosby, et al., 1971; Buser, 1976) and in water (Friessen, et al., 1990) noted reductive dechlorination resulting in dioxin compounds of lower chlorination. Other experiments did not find such reductive dechlorination (Dulin, et al., 1986; Friessen, et al., 1990 who found reductive dechlorination in one experiment, but not in another). An important issue to consider, at least, for the process of photodegradation of dioxins and furans on leaf surfaces is the possible formation of lower chlorinated congeners of non-zero toxic equivalency.

Another issue discussed by McCrady is that the theoretical time for the grass tissue to reach a steady state in his experiments is much shorter than that indicated in the Bacci experiments. Using Bacci's results, McCrady noted that the azalea leaves theoretically take greater than 400 days to reach equilibrium, in comparison to less than 20 days to reach equilibrium for the grass plants in his experiments. This difference is not entirely due to photodegradation. McCrady (personal communication, J. McCrady, Corvallis Environmental Research Laboratory, EPA) suggests that the 50-day exposure time used in Bacci's experiments may allow for considerable diffusion into the newly formed plant surface wax.

The sorbed TCDD residues may be trapped and unable to volatilize. Thus, for estimating contaminant concentrations in animal feeds such as relatively short-lived grass plants, the equilibrium Bvol from the Bacci azalea model may overestimate the contaminant concentration in grass. On the other hand, McCrady's experiments may have been conducted in too short a time frame, with the sum of uptake and elimination phases being less than 10 days in the various experimental designs. The volatilization and photodegradation rates reported by McCrady may be higher than what might occur for the longer exposure times expected in real world situations, where growth and residue trapping may occur.

These arguments are being presented to demonstrate the uncertainty in choosing either of the two reported Bvol values for estimating plant contaminant concentrations. McCrady's results pertaining to 2,3,7,8-TCDD cannot be generalized to other dioxin-like compounds or other contaminants in terms of commonly available contaminant parameters such as H or Kow. Therefore, a McCrady framework similar to Bacci's for estimating congener-specific Bvpa cannot be offered at this time. On the other hand, their work strongly suggests that the Bacci model may be inappropriate for terrestrial vegetations of this assessment, including vegetables/fruits and vegetations of the beef/milk food chain model, and Bacci's experiments, because of their length of time, the use of an azalea leaf of high wax content, and lack of an artificial light source simulating photodegradation, are likely to have overestimated the air to leaf transfers.

What will be done for this assessment is to first estimate a congener-specific Bvol using the Bacci algorithm of Equation (4-30) above. Then, it will be transformed into a mass-based Bvpa as in Equation (4-31), except that the assumptions McCrady and Maggard used for fraction of grass plant that is wet weight, 85%, and the grass leaf density, 770 g/L, will instead be used as more representative of vegetations of this assessment. Most importantly, the Bvpa calculated this way will be empirically reduced by a factor of 40 for all dioxin-like congeners as suggested by the difference in McCrady's experiments as compared to Bacci's.

It should be noted that all bioconcentration or biotransfer parameters, such as the Bvpa, are qualified as second order defaults for purposes of general use. Section 6.2. of Chapter 6 discusses the use of parameter values selected for the demonstration scenarios, including a categorization of parameters. Second order defaults are defined there as parameters which are theoretical and not site specific, but whose values are uncertain in the published literature. The parameter values in this category should be considered carefully by users of the methodology.

. VGag:

The same discussion for this correction factor for below ground vegetation applies here. Fruits such as apples, pears, plums, figs, peaches, and so on, can be approximated by spheres, and upper bound estimates of correction factors would be less than 0.05. Peeling, cooking, and cleaning further reduces residues. The VGag for unspecified above ground fruits and vegetables in this assessment is assumed to be 0.01. Like VGbg, this value is assigned considering that it should be less than estimated just based on surface volume to whole fruit volume ratios.

Two other VGag values are required for this assessment. One is for pasture grass and the other for other vegetations consumed by cattle. Both are required to estimate concentrations in these vegetations consumed by cattle in order to estimate beef and milk concentrations. A VGag value of 1.0 was used to estimate pasture grass concentrations since there appears to be a direct analogy to exposed azalea and grass leaves. However, VG should be less than the other general category of cattle vegetations defined in this assessment, "hay/silage/grain". Recognizing that pasture grass is important in terms of amount consumed in the lifetime of a beef cow, and the fact that it is a leafy vegetation, it is considered seperately, whereas other cattle vegetations are lumped together in this second category.

As described below in Section, this second general category of non-grass cattle vegetations include some thin leafy (hay) as well as bulky (corn silage and other grains) vegetations to consider. A volume ratio of outer surface to whole surface area to volume vegetation could be used to assign a value to VG, if specific assumptions concerning proportions of each type of vegetative cattle intake were made. An appropriate assumption for a fully protected vegetation such as grain would be zero. Silage can be considered part protected and part leafy. Since specific assumptions concerning hay/silage/grain intake are not being made for this exercise, a simple assumption that VG equals 0.50 for hay/silage/grain is instead made, without rigorous justification.

The only experimental evidence that a VGag for vapor transfers of dioxin-like compounds is justified came in a recent study by McCrady (1994). McCrady experimentally determined uptake rate constants, termed k1, for vapor phase 2,3,7,8-TCDD uptake into several vegetations including kale, grass, pepper, spruce needles, apple, tomato, and azalea leaves. Recall that the similar experimental design of both McCrady and Maggard (1993), and Bacci, et al. (1990; 1992), included an initial phase where vegetations in experimental chambers were exposed to the vapor-phase organic chemicals. The uptake which occurs during this initial phase is described with the rate constant, k1.

A second "elimination" phase then occurs where organic vapors are removed from the chambers and the chemicals allowed to volatilize or otherwise dissipate from the vegetation. The rate constant for this phase is termed k2. A steady state bioconcentration factor, or Bvpa in this assessment, is then estimated as k1/k2. The uptake rate constants from air to the whole vegetations estimated in the recent experiments by McCrady (1994) demonstrate the concept behind the VG parameter. The uptake rate for an apple divided by the uptake rate for the grass leaf was 0.02 (where uptake rates were from air to whole vegetation on a dry weight basis).

For the tomato and pepper, the same ratios were 0.03 and 0.08. The VGag was 0.01 for fruits and vegetables in this assessment, but note above that the simple exercise with a conical carrot and spherical potato estimated a surface volume to whole fruit volume ratio of 0.09 (carrot) to 0.03 (potato); a value of 0.01 for fruits and vegetable empirically considers factors such as washing or peeling which would reduce exposures. McCrady (1994) then went on to normalize his uptake rates on a surface area basis instead of a mass basis; i.e., air to vegetative surface area instead of air to vegetative mass. Then, the uptake rates were substantially more similar, with the ratio of the apple uptake rate to the grass being 1.6 instead of 0.02; i.e., the apple uptake rate was 1.6 times higher than that of grass, instead of 1/50 as much when estimated on an air to dry weight mass basis. The ratios for tomato and pepper were 1.2 and 2.2, respectively.

In his article, McCrady (1994) concludes, "The results of our experiments have demonstrated that the exposed surface area of plant tissue is an important consideration when estimating the uptake of 2,3,7,8-TCDD from airborne sources of vapor-phase 2,3,7,8-TCDD. The surface area to volume ratio (or surface area to fresh weight ratio) of different plant species can be used to normalize uptake rate constants for different plant species." McCrady does caution, however, that uptake rates are only part of the bioconcentration factor estimation, and is unsure of the impact of surface area and volume differences on the elimination phase constant, k2 (personnal communication, J. McCrady, US EPA, ERL-Corvallis, Corvallis, OR 97333). Still, his recent experiments do appear to justify the use of a VG parameter since the Bvpa were developed on an air to whole plant mass basis, and his results are consistent with the assignment of 0.01 for fruits and vegetables.

. kw:

Fries and Paustenbach (1990) note that this approach may overestimate concentrations because crops can be harvested or pastures grazed before the plant concentrations reach steady state, and that a kw based on a weathering half-life of 14 days may be too long given experimental results of Baes, et al. (1984) which showed a range of 2-34 days, and a median value of 10 days. Stevens and Gerbec (1987) considered harvest intervals by including the exponential term, (1-e-kt), and assigning values of t based on harvest intervals of different crops. This assessment uses a kw of 18.02 yr-1, which is equivalent to a half-life of 14 days.