Modelling of the uptake of organic pollutants in vegetables and fruit
Summary and conclusions
The objective of the project was to review a method or a basis for describing/assessing the relation between the level of contamination with organic pollutants in the soil (and pore air) and the level in vegetables and fruits. Based on the above method/basis, a proposal should be made for the preparation of guidelines to be applied in risk assessment of the uptake of pollutants in vegetables and fruits from soil contaminations.
Previous studies had illustrated the consequences of direct contact between edible parts of vegetables and soil contaminiated with PAHs and metals at levels close to the Danish EPA cut-off criteria. This project was thus focussed on whether concentrations above the cut-off criteria in the underlying soil - where the soil is not in direct contact with the edible parts of the plant but only with the roots of the plant - pose a problem because of the transport from the roots to other parts of the plant or evaporation from the soil with supsequent uptake in the parts of the plant above the soil. Furthermore, all common organic pollutants should be included, among these especially oil products and chlorinated solvents.
This statement of aims resulted in the development of a decision-support tool based on mathematical models, in which great importance was attached to the inclusion of water-soluble and volatile substances in the calculations.
The report introduction describes the mechanisms governing the concentrations of organic substances in the crops including the mechanisms of uptake and transport of organic substances in plants, which part of the plant is used as well as dilution because of the growth of the plant and metabolization of the substances. Furthermore, the importance of the residence time of the substances in the soil ("aging") to their availability to the plants is discussed together with the assumptions regarding availability included in existing models.
Exising models for the description of the distribution and transport of the substances in the soil and their uptake in plants are examined and discussed, and a brief summary of existing model-based risk assessment tools is given.
With regard to a selection of models for the decision-support tool, previous validation studies of the different models were examined and estimated values compared with measured data from two investigations carried out in soils contaminated with PAH and PCB. No experimental data on more mobile substances are available. In general, comparison with measured values indicates that the models tend to overestimate the uptake in the plants with a factor of up to 10 though exceptions were seen. This is probably due to the fact that several circumstances reducing the concentration in the plants are not included in the particular models, e.g. aging in the soil, degradation/metabolization of the substances in the plant. It should, however, be emphasized that the availabe data from experimental studies are scanty and that the variation is large. The available data for a validation of the description of the uptake of substances in fruit are not adquate but as the evaporation of the substances from fruits is not included, we estimated that the uptake in fruit is probably overestimated.
Criteria were established for the selection of models for the project and models were chosen for the description of the uptake and transport in plants of the substances and their movements in the soil. Furthermore, the scenarios and crops used in the subsequent estimations were described.
The model tool composed on the basis of the above criteria and assumptions comprises the estimation of
 | Concentration in the three soil phases (water, air and solid phase) by simple partition equations | ||||||||
| Transport in the soil by use of the Jury model | ||||||||
| Evaporation from the soil by use of the Jury model and concentration in the air by use of the principles in the Jagg model in which the mixing altitude is, however, fixed at the height of the individual crop | ||||||||
| Uptake in the crops by use of crop-specific models with respect to
|
Estimations were made
| For eight substances: MTBE, toluene, n-dodecane, trichloroethene, naphthalene, benzo(a)pyrene, benzene og tetrachloroethene |
| For two soils: one soil with a normal and one soil with a low content of organic carbon |
| For a cover layer of clean soil: No layer and a cover layer of 0.5 m. In the estimations with cover layer, the average concentration in the root depth of the plant was used to estimate the amount taken up together with the water whereas the average concentration in the layer, in which the root crop is situated, was used to estimate the diffusive transport into the root crop. The average concentration in the surface layer (fixed at 2 cm in the estimations) was used with regard to deposition of soil on the plant (soil splashes). The estimations were made one year and ten years after the establishment of the cover layer. |
| With and without regard to aging |
| For the following crops: Potato, carrot, lettuce, kale, apple, berries from shrubs, strawberry, nut |
The intention with the selection of the parameters included in the models - and the chosen values - was that they should represent a realistic worst-case sitution. In some cases, the conditions are, however, unrealistic and the plant uptake is overestimated, e.g. the metabolization of the substances in the plant is fixed at zero in all the calculations and evaporation of substance from the fruit is not included in the fruit tree model.
With regard to a selection of the crops that contribute most to the ingestion of organic substances, the estimated plant uptake was linked with the expected daily intake of selected types of crops. In this way, the expected daily intake of each individual substance together with each of the eight crops was estimated. The estimates showed that
| When a layer of clean soil is not established on top of the contaminated soil, potato is the crop contributing the most to the potential daily human intake of organic contaminants through crops |
| When a layer of clean soil is established on top of the contaminated soil, other root crops than potato are the most critical crops with respect to PAHs while potato is still the most critical with respect to n-dodecane and fruit is the most critical type of crops with respect to the other more mobile model substances. |
To some extent, this reflects that potaoes and apples/pears are the most commonly eaten crops in Denmark. The results for the less mobile substances are in agreement with the results of previous investigations performed in contaminated soil showing that direct contact with the contaminated soil could result in high concentrations of the substances in root crops (including potato) whereas crops that are not in direct contract with the soil contained considerably lower concentrations. It may seem surprising that the uptake in fruit of contaminants from the soil should be of greater importance to the daily intake than the uptake in potatoes or other root crops. This is, however, only the case if a cover layer of clean soil has been established on top of the contaminated soil so that the root crops are not in direct contact with the contamination and with respect to mobile substances, which can be taken up through the roots of the plants, as tree roots are assumed to reach deeper than the roots of vegetables. Furthermore, it should be emphasized that, in the fruit tree model, evaporation of the substances from the surface of the plant (fruit) is not included. No experimental data are available for a verification or modification of the estimations of the uptake in fruit of mobile substances.
A sensitivity analysis of the potato, carrot (other root crops) and fruit tree models was performed. It showed that if all included parameters are varied at the same time, the standard deviation on the estimated BCF values for both potato and fruit was less than one order of magnitude. I.e., it is of minor importance if one or more of the parameters applied in the model are not precisely determined. When using the models, it is of relevance to know for which parameters it is most import to apply a precisely determined value. Therefore, the importance of the variation in the individual parameters was analysed for the three most essential models.
The potato model is the most sensitive with respect to variation in the parameters when the not very mobile substances, for which equilibrium between the root crop and the soil cannot be assumed, are estimated. For these substances, the parameters resulting in the largest variation in the results of the potato model, are the size of the potato and, to some extent, the air volume together with the growing period.
For the carrot model (used for several root crops), the sensitivity analysis showed that the lipid content in the root crops and the log KOW of the substances are of importance to the uptake in root crops while the air-water partition coefficient is only of minor importance to the results.
The fruit tree model is most sensitive to changes in the log KOW of all investigated substances with the exception of the least sorbing substance, MTBE. For MTBE, the model is very robust towards variation in all parameters and the characterization of the soil is the most important parameter. The characteristics used for the plant are of minor importance to the estimated BCF value.
The developed model tool is built into an Excel file in which the models recommended for the individual types of crops and for estimation of the transport in the soil of the substances, have been programmed. In the spreadsheet, small databases have been incorporated with data on the model substances of the project, on different types of Danish soils and on the individual model plants. These databases are easily extended with more data.
A model tool has thus been developed that
| Can estimate the concentrations, which may occur in typical Danish crops grown in contaminated soil with or without a cover layer of clean soil. |
| By comparing with ADI (in which other exposure routes are taken into consideration) can clarify whether the soil concentrations corresponding to the soil quality standards or to compliance with the evaporation criteria above surface will result in critical concentrations in the relevant crops. |
| Can identify which type(s) of crops that must be expected to be the critical one(s) rather than others in the specific situation. |
| Can prioritize the areas for which preventive measures should be taken. |
The models are only valid for non-ionizable and non-polar substances.
It is an important precondition for the use of modelling for risk assessments that the obtained estimates are fairly realistic (but still conservative).
The main limitation of the model tool is that it is based on a number of conservative assumptions and parameters (e.g. no evaporation from fruits, no metabolization in the plant), which will not be realistic in all situations; In other words, the tool is expected to overestimate the uptake in the crops. When basing risk assessments on these estimations, assessments showing low risk - and thus no need for (further) preventive measures will be well-founded whereas the preventive measures based on assessment of high risk may actually be unnecessary.
The estimations performed and comparisons with the limited amount of available data from experimental investigations show, however, that the model tool is at least on the level with corresponding tools used in other countries with respect to the "precision" of the estimates as well as the robustness towards variations in the individual parameters. The submodels for single crops (especially the potato and the fruit tree models), which were developed and included in the tool, further increase its applicability considerably.
Based on the results of the project, the model tool is considered to be applicable for elaboration of risk assessments. Several preconditions must, however, be fulfilled:
| Human toxicological assessments – e.g. as values for acceptable daily intake (ADI) - must be available for each substance |
| Investigations are conducted, which can confirm or invalidate the results of the model calculations; especially the estimate that mobile substances will be taken up from the soil and accumulated in fruits |
| The applicability of the model tool is evaluated through calculations based on different scenarios in which the parameters in the databases are based on "worst case" assumptions (like now) as well as on more realistic assumptions. |
Depending on the outcome of such scenario calculations, the results can be used for preparation of guidelines describing how to assess the risk in typical contamination situations – including the extent of the necessary preventive measures.