SCOPE 49 - Methods to Assess Adverse Effects of Pesticides on Non-target Organisms

8.1 INTRODUCTION

For many years the information on the environmental distribution and fate of chemical substances was acquired largely in retrospect from empirical observations after large-scale monitoring efforts with an enormous number of chemical analyses. This retrospective approach left wide margins for error in the environmental management of chemical substances, and the possibility of large-scale undesirable effects remained largely uncontrolled. It is sufficient to note the widespread distribution of organochlorine pesticides and their detection in areas very far from the application zones, and the presence of herbicides in groundwater used for human consumption.

One way to face the problem might have been to model the environmental distribution of individual substances. In the sixties, the modelling era started when technical facilities made the solution of complex numerical calculations easier; but after an initial period of euphoria, complex models were considered much less applicable and useful than expected. For example, large-scale models such as that of Randers (1973) worked for DDT, but only with large quantities of input data; their predictive capability was therefore very limited.

In the early seventies, several groups of scientists began thinking in predictive terms. Moreoever, legislative mandates, such as the Toxic Substance Control Act (USEPA,1978) in the US and the Directive on Dangerous Substances (EEC, 1979) in Europe, gave a strong impetus to the predictive approach. Initially, several investigators advanced the hypothesis that one could predict the behaviour of a new chemical by comparing properties measured in a laboratory with those of compounds for which more environmental data were available. The concept of environmental chemodynamics was then proposed as a holistic approach for the comprehensive understanding of the behaviour of chemical substances in the environment.

Contemporary management of chemical substances assesses the hazard as one step needed for a comprehensive risk assessment of potentially dangerous chemicals. In essence, a hazard assessment is a comparison of the no-observed-effect level and the expected environmental concentration. For making such an evaluation, the prediction of exposure has become a key issue in research on specific substances.

 8.2 DEVELOPMENT OF EVALUATIVE MODELS

To predict environmental distribution and fate of chemicals, Baughman and Lassiter (1978) introduced the concept of an evaluative model with the aim of developing a quantitative approach for exposure estimation. According to these authors, evaluative models `incorporate the dynamics of no specific environment but are based on the properties of stylized environment or hypothetical pollutants for which we specify (rather than measure) inputs'.

 In the ensuing years, many publications appeared on the same subject (see, for example, Haque,1980; Neely,1980; Hutzinger,1980, 1982; Gunther,1983; Neely and Blau, 1985; Sheehan et al., 1985; GSF, 1985). The Organization of Economic Cooperation and Development, within the framework of its Chemical Group and Management Committee in the Hazard Assessment Project, prepared a report on the practical approaches for the assessment of environmental exposure (OECD, 1986).

Several authors produced simple evaluative models (Mackay, 1979; Neely, 1979; Roberts and Marshall, 1980; Frische et al., 1982). A `fugacity' model was proposed to calculate the relative amount of a substance that would ultimately partition into each environmental compartment (Mackay and Paterson, 1982). In these models, no attempt was made to simulate the precise or actual environmental conditions; but they were intended to allow the prediction of the potential distribution of the chemical compounds.

 The most relevant possibilities for application of the results of such types of models were, according to Calamari and Vighi (1988), the following:

1. Preliminary screening of the environmental behaviour of a number of chemicals,
2. Indications of the partition and affinity for the main environmental compartments,
3. Identification of the compartments where high levels are more likely to be present, and of the matrices where the transformation processes are more likely to be relevant, 
4. Information to assist with the planning of laboratory or field experiments and/or monitoring campaigns,
5. Data for preliminary hazard assessment.

Fugacity models of increasing complexity were produced by Mackay and others (Mackay and Paterson, 1982; Mackay et al., 1983) and laboratory and field studies were undertaken to clarify fully the concept of fugacity and to identify the value and limitations of such an approach, the areas of applicability, and the predictive capability of evaluative models.

In the following sections the fugacity approach will be described, because the experience of the authors deals mainly with this kind of model; various evaluative models of comparable applicability have been developed recently, but are not discussed here.

8.3 THE FUGACITY MODEL

8.3.1 CONCEPT AND APPLICATION

The concept of fugacity was introduced by Lewis at the beginning of this century as a criterion for equilibrium between phases. Fugacity can be regarded as the tendency of a chemical substance to escape from a phase. It has units of pressure, and can be linearly related to concentration (Mackay and Paterson, 1982).

The relation between fugacity (f) and concentration (C) for a given phase (or environmental compartment) can be written as:

where Z is a `fugacity capacity constant' depending, at a given temperature and pressure, on the nature of the substance and of the environmental compartment under examination. Z quantifies the capacity of a given phase or compartment for fugacity.

Each substance tends to accumulate in compartments where Z is high and where high concentrations can be present with relatively low fugacities. Where a chemical is introduced in a multicompartment system, the fugacity at equilibrium is equal in all media, but concentrations are different functions of the different capacities. To evaluate equilibrium concentrations on a comparative basis and express them in commonly accepted units, the concept of `unit of world' was then introduced (Mackay, 1979). This is a hypothetical model environment that includes the main environmental compartments at defined volumes.

 If one can find Z values for a substance for each environmental phase, the distribution of the substance at equilibrium can be easily calculated. The highest concentrations will correspond to the highest values of Z. To obtain Z values for the different environmental compartments, it must be noted that in the atmosphere, fugacity can be considered in general as equivalent to the partial pressure (P) for the vapour phase of a chemical. Concentration (C) is related to partial pressure through the gas law:

	PV = nRT		8.2
			8.3

From equation (8.1) it follows that

		8.4
Thus
		8.5

For all other compartments, one must remember that in a multicompartment system at equilibrium,

wher f_i is the fugacity of a substance for the compartment i. Considering two compartments i and j, from equation (1),

		8.7
		8.8

where K_ij is the partition coefficient between the two phases. From equation (8.8)

Therefore, one can easily calculate Z for each compartment, if one knows the partition coefficient between two phases and the Z values for one of the two compartments. For example, for water,

 where K_wa is the partition coefficient between water and air. It follows that

8.11

From equation (8.3)

8.12

Through this procedure, one can easily include new environmental compartments in the original `unit of world'.

 8.3.2 APPLICATIONS

In a series of papers, the concept of fugacity and its potential to describe environmental distribution and fate phenomena were explored and explained. Fugacity, which at the beginning was no more than a simple definition of `escaping tendency', has been and could be used for different purposes.

The first application could be in the prediction of environmental distribution with a fixed scenario (e.g., the standard `unit of world') to compare different substances and rank them on the basis of their affinity for a certain compartment. Fugacity can be used, for example, to select which of a number of herbicides with similar agronomical performances has the lowest affinity for the water compartment, providing a basis for minimization of the probability of water contamination.

Some of the essential aspects of the environmental distribution and fate of cypermethrin (Bacci et al., 1987) have been elucidated by means of a combination of evaluative models and a few ad hoc experiments in a simulation chamber. The experiments produced results that had been predicted by the fugacity model. In fact, according to its physical and chemical properties, cypermethrin should have strong soil affinity, be practically non-volatile in air, unable to reach water, and unable to accumulate in plants. Degradation occurred in soil over a period of time comparable with that previously reported.

Although evaluative models are not intended for prediction of detailed fate in a specific environment, one can vary the standard scenario to improve the simulation and obtain an enhanced understanding of the behaviour of a single molecule (e.g., a unit of world in interstitial water to mimic fate in a cultivated field, or the removal of the soil compartment to simulate a lake, etc.). 

The mass balance concept and biogeochemical approach should be considered for a complete understanding of the behaviour of a chemical in an environment; therefore, these types of simulations can be of great help for the preparation of monitoring campaigns or large-scale field experiments. Clark et al. (1988) discussed the utility of such an approach in relation to a monitoring program to identify persistent and biocumulative contaminants such as chlorinated hydrocarbons. From a theoretical example, they stated that, if reliable bioconcentration relationships exist between water, fish, birds, and bird eggs, it could be more convenient routinely to analyse a few eggs instead of high numbers of samples from any media or organisms with lower and more variable concentration levels.

The fugacity approach was applied to evaluative models, laboratory experiments, and field work to understand mechanisms regulating environmental distribution and fate of a pesticide under specific conditions and to define acceptable loads in a given territory. In the watershed of Lake Chiusi (Sienna, Italy), an area of about 100 ha of clay soil was studied after treatment with a single dose of atrazine (Bacci et al., 1990). The fugacity model was adapted to the specific environmental system (watershed plus lake) considering the differences at various times (rainfall, growing of crops, etc.). Laboratory experiments were performed to evaluate degradation in conditions comparable to those existing in the examined system. Concentrations of atrazine were measured at various times in different media, and compared with those calculated by the model. A good agreement was shown between the prediction of the model and the measured values (Table 8.1). Moreover, advection by runoff was demonstrated as the only route of transport of atrazine from the terrestrial to the aquatic environments.

Table 8.1. Comparison of predicted and measured levels of atrazine (from Vighi and Calamari, 1990) in environmental compartments of the Lake Chiusi area at different times after treatment

Time	Soil		Lake water		Lake sediments
(days)	Predicted	Observed	Predicted	Observed	Predicted	Observed
0	401		62	66*	5
90	104			46
130	54	42		32
180	26	20		25		1.8
230	11	9		21
*Calculated from the loss rate constant

This integrated approach allows the estimation of the acceptable load of atrazine in fields during `medium' rainfalls to obtain acceptable levels of the herbicide in lakes whose water is used for drinking purposes (Vighi and Calamari, 1989).

Besides the practical and concrete applications of the fugacity approach already cited, other potential implications have been suggested by Clark et al. (1988). These authors proposed that by converting disparate concentration units related to various media into one common unit (e.g., fugacity in Pa), it is reasonably easy to evaluate the significance of various concentrations in different environmental compartments (i.e., a trend for bioaccumulation). In case of equifugacity in a medium, an equilibrium status or a rapid exchange among compartments could be recognized. The same type of calculation (concentration expressed as fugacity) could identify hot spots and clean areas.

8.3.3 PROBLEMS AND CONSTRAINTS IN THE USE OF THE FUGACITY MODEL 

A limitation inherent in the original form of the fugacity model is the absence in the unit world for biota other than aquatic animals. Terrestrial plants are a very important component due to their biomass and to the possibility of transfer of contaminants to higher levels, including man, of the trophic chain.

On the other hand, as has been previously emphasized, the inclusion of a new compartment into the model is conceptually an easy task, requiring only the calculation of the Z value once partition coefficients are known.

A wide range of literature is available on bioconcentration factors for aquatic animals, whereas very scanty information exists for terrestrial plants. This situation can explain why these species have not been included in the standard form of the model. The inclusion of plant biomass in the fugacity model was attempted by Calamari et al. (1987). Three equations have been used relating accumulation in roots and in the stem with the log of the octanolwater partition coefficient, log K_oW (Briggs et al., 1982, 1983) and accumulation in foliage with the Henry's law constant, utilizing data reported by Bacci and Gaggi (1987). By means of the resulting bioaccumulation factors, terrestrial plants have been included as a sum of three separate compartments (namely roots, stem, and foliage).

To produce a theoretical validation of the consistency of the improved model, the hypothesis was proposed that bioaccumulation in foliage depends on log K_ow and H. This hypothesis has been successively confirmed by Travis and Hattemer-Fret' (1988), who used the same data from Bacci and Gaggi (1987) and found that the logarithm of the bioconcentration factor is directly related to log K_ow and inversely to log H. Thus, an improved equation for bioaccumulation in foliage, based on new experimental data, has been proposed by Bacci et al. (1990).

Major problems in the use of the fugacity model can derive from the lack of availability and the poor quality of basic data on the physico-chemical properties of the molecules. Theoretically, the only data essential to apply the fugacity model are water solubility and vapour pressure. Even for these elemental properties, much uncertainty exists. In the most frequently consulted reference books, such as The Pesticide Manual (Worthing and Walker, 1987), data are often expressed in vague terms or are not available at all (e.g., solubility in water: `negligible'; vapour pressure: `less than . . .'). A recent review (Suntio et al., 1988) of Henry's law constants for pesticides demonstrates that data in the literature are often conflicting and that the selection of the most reliable data is a very vexing task.

All other parameters needed for the fugacity model (octanolwater partition coefficient, soil sorption coefficient, bioconcentration factors, etc.) could be obtained, in theory, by means of property-to-property equations (e.g., from solubility). Nevertheless, the applicability of property-to-property equations has been shown to be relatively limited and to have great potential for error, particularly in the calculation of K_ow, a key property in each type of evaluative model.

A more reliable approach is the calculation of K_ow by means of the fragment constants method (Hansch and Leo, 1979) or other comparable techniques. Computer programs exist for the calculation of log K_ow, by means of this method. Recently the US Environmental Protection Agency has modified these programs for the personal computer (USEPA, 1987).

Many authors consider the fragment constants method the best way to obtain reliable values for log K_ow; in some cases, this method is even preferred to experimental measurements. Nevertheless, for very complex chemicals (such as many pesticides and drugs) and in particular for values of log K_ow, values and other basic physico-chemical data must be carefully examined before use.

8.3.4 WIDENING THE FUGACITY CONCEPT

 In addition to the previously discussed limitations of these models, another relevant constraint is the absence of any indication of the time scale for the attainment of equilibrium. A second observations is that, for the same system of reference (i.e., the standard `unit of world'), fugacities varied over a range of several orders of magnitude even if one considers pesticides alone (Figure 8.1). 

Figure 8.1. Range of variability of fugacity values for some pesticides at equilibrium after an emission of 100 moles into the system; physico-chemical data taken from Worthing and Walker (1987) and from Suntio et al. (1988)

Accordingly, based on these two observations, an attempt was made to explore whether a difference in the fugacity could also be an indication of a potential difference in mobility among molecules. Fugacity by definition is not an intrinsic property of the molecule, since it depends on the compartments of the environmental system and the relative concentrations in the various phases of the system besides the characteristics of the molecule.

In a dynamic system, fugacity is not constant, but varies with time and concentration. However, if the parameters of the system (i.e., standard unit of world) and the time (i.e., initial condition or equilibrium condition) are fixed, fugacity can be considered independent of the system and, therefore, an `intrinsic' property of the molecule.

If one considers the fugacity at equilibrium in the standard unit of world, it can be observed that sometimes different molecules may show similar distribution in the environmental compartments, notwithstanding different fugacity values.

As an example, Table 8.2 shows the physico-chemical properties needed for the application of the fugacity model to four pesticides (part A) and the percentage distribution and fugacity in the standard unit of world (part B). The percentage distribution of the four molecules is very similar in all compartments, with the exception of air, while fugacity ranges over about three orders of magnitude.

Considering the relation between concentration and fugacity (C= f Z and that  Z for the air compartment is independent of the characteristics of the molecule (Z_a=1/RT), one concludes that it is impossible to obtain similar concentrations in air with different fugacities.

In compartments other than air, partition is regulated by log K_ow or other partition coefficients directly related to it (soil sorption coefficients, bioconcentration factors, etc.). These coefficients are practically identical for the molecules considered here. In Figure 8.2, data from Table 8.2 for sulfotep and chlorfenvinphos are graphically presented; in each box, representing an environmental compartment, a similar distribution can be observed. When the system is at equilibrium, by definition the same quantities of the substance move in and out from each compartment. The total mass transport in a fixed unit of time is different for the two chemicals and depends on their escaping tendencies. This is shown in Figure 8.2, where escaping tendency (fugacity) is represented as a vector. If we consider two molecules having different fugacities but similar equilibrium distributions in the standard unit of world, one could expect that, by placing the same quantity in the same compartment, the mass transfer in a fixed time should be different for the two chemicals due to differences in fugacity.

 A numerical example for the two molecules examined is shown in Figure 8.2. Placing the same quantity (e.g, 100 moles) into the same compartment (e.g., water), one can calculate the initial fugacity from equations (8.1) and (8.12):

 Table 8.2. Physico-chemical properties needed for the application of the fugacity model to four pesticides (part A); percentage-distribution and fugacity in the standard unit of world (part B)

8.13

It follows that values of 4.1 x 10⁶ and 3.9 x 10⁴ can be calculated for sulfotep and chlorfenvinphos, respectively. This means that the former exerts an escaping tendency from water to all other compartments about a thousand times higher than the latter. In other words, the higher the fugacity, the shorter the time needed for the attainment of equilibrium. This intuitive statement is also supported by the fact that fugacity is expressed in units of pressure (i.e, lower pressure, lower mobility).

 Unfortunately, if this is true, it can be proposed only on a qualitative basis, and the role of fugacity as a mobility parameter cannot at present be quantified. There is a need for experimental work to confirm this hypothesis and eventually to define the relationship between fugacity and rate of transport.  

 8.4 CONCLUSIONS

Ideally, evaluative models should allow calculation of the potential exposure concentration in each ecosystem; however, this is still a distant prospect. This paper has presented some positive achievements in the use of evaluative models. Recently, other evaluative models, based on simple physico-chemical properties, have been developed, such as those proposed by Rao et al. (1985) to predict the potential for groundwater contamination from pesticides. Jury et al. (1987) indicate that further development of integrated models will provide an interesting methodology to understand and quantify transport and transformation pathways of organic chemicals in the environment.

Figure 8.2 Illustration of equilibrium distribution for sulfotep (S) and chlorfenvinphos (C) in 'unit of world'; fugacities shown as vectors

Figure 8.3. Overall research strategy on valuative models for the environmental management of pesticides

The scientific community has a general interest in this subject, as evidenced by the number of references in the joint report, this paper, and in specialized publications, such as the proceedings of a workshop on environmental modelling for priority setting among existing chemicals (GSF, 1985). Figure 8.3, from Calamari and Vighi (1988), illustrates a strategy of the use of an evaluative model to minimize the impact of pesticides on the environment. The overall strategy is aimed predominantly at developing predictive systems to confidently calculate the potential exposure in various compartments of the ecosystem. Recently, this field of investigation has increased exponentially, and offers great promise for future research in environmental sciences.

8.5 REFERENCES 

 Bacci, E., Calamari, D., Gaggi, C. and Vighi, M. (1987) An approach for the prediction of environmental distribution and fate of cypermethrin. Chemosphere 16, 1373-1380.

Bacci, E. and Gaggi, C. (1987) Chlorinated hydrocarbons vapours and plant foliage: kinetics and applications. Chemosphere 16, 2515-2522.

Bacci, E., Renzoni, A., Gaggi, C., Calamari, D., Franchi, A., Vighi, M. and Severi, A. (1990) Models, field studies, laboratory experiments: an integrated approach to evaluate the environmental fate of atrazine (s-triazine herbicide). In: Paoletti, M. G., Stinner, B. R., and Lorenzoni, G. G. (Eds) Agricultural Ecology and Environment: Proceedings of an International Symposium, Padova, Italy, 5-7April, 1988, Elsevier, Amsterdam.

Baughman, G. L. and Lassiter, R. R. (1978) Prediction of environmental pollution concentration. In: Cairns, J., Dickson, K. L., and Maki, A. W. (Eds) Estimating the Hazard of Chemical Substances to Aquatic Life, ASTM Special Technical Publication 657, American Society for Testing and Materials, Philadelphia, pp. 35-54.

Briggs, G. G., Bromilow, R. H. and Evans, A. A. (1982) Relationship between lipophilicity and root uptake and translocation of non-ionized chemicals by barley. Pest. Sci. 13, 492-500.

Briggs, G. G. Bromilow, R. H., Evans, A. A. and Williams, M. (1983) Relationships between lipophilicity and the distribution of non-ionized chemicals in barley shoots following uptake by roots. Pest. Sci. 14, 492-500.

Calamari, D. and Vighi, M. (1988) Experiences on QSARs and evaluative models in ecotoxicology. Chemosphere 17, 1539-1549.

Clark, T., Clark, K., Paterson, S., Mackay, D. and Norstrom, R. J. (1988) Wildlife monitoring, modelling, and fugacity. Environ. Sci. Technol. 22, 120-127.

EEC (1979) Council directive 79/831. Off. J. Europ. Commun. L259, 10pp. 

Frische, R., Esser, G., Schoenborn, W. and Kloepffer, W. (1982) Criteria for assessing the environmental behaviour of chemicals: selection and preliminary quantification. Ecotoxicol. Environ. Saf. 6, 283-293.

GSF (1985) Environmental Modelling for Priority Setting among Existing Chemicals, Workshop, 11-13 November 1985, Munchen-Neuherberg Gesellschaft fur Strahlen und Umweltforschung mbH, Munchen.

Gunther, F. A. (Ed.) (1983) Residues of pesticides and other contaminants in the Total Environment. Residue Rev. 85, 1-307.

Hansch, C. and Leo, A. J. (1979) Substituent Constants for Correlation Analysis in Chemistry and Biology, John Wiley, New York.

Haque, R. (Ed.) (1980) Dynamics, Exposure and Hazard Assessment of Toxic Chemicals, Ann Arbor Science, Ann Arbor, Michigan.

Hutzinger, O. (Ed.) (1980) The Handbook of Environmental Chemistry, Reaction and Processes, Vol. 2, Part B, Springer-Verlag, Berlin, Heidelberg, New York.

Jury, W. A., Winer, A. M., Spencer, W. F. and Focht, D. D. (1987) Transport and transformations of organic chemicals in the organic chemicals in the soil-air-water ecosystem. Rev. Environ. Contam. Toxicol. 99, 119-164.

Lyman, W. J., Reehl, W. F. and Rosenblatt, D. H. (1982) Handbook of Chemical Property Estimation Methods, McGraw-Hill, New York.

Mackay, D. (1975) Finding fugacity feasible. Environ Sci. Technol. 13, 1218-1223.

Mackay, D., Joy, M. and Paterson, S. (1983) A quantitative water, air, sediment interaction (QWASI) fugacity model for describing the fate of chemicals in lakes. Chemosphere 12, 981-997.

Mackay, D. and Paterson, S. (1982) Fugacity revisited. Environ. Sci. Technol. 16, 654-660.

Mackay, D., Paterson, S. and Joy, M. (1983) A quantitative water, air, sediment interaction (QWASI) fugacity model for describing the fate of chemicals in rivers. Chemosphere 12, 1193-1208.

Neely, W. B. (1979) A preliminary assessment of the environmental exposure to be expected from the addition of a chemical to a simulated aquatic ecosystem. Int. J. Environ. Studies 13, 101-108.

Neely, W. B. (1980) Chemicals in the environment. Distribution, Transport, Fate Analysis, Marcel Dekker, New York, Basel.

Neely, W. B. and Blau, G. E. (Eds) (1985) Environmental Exposure from Chemicals, Vol. 1, CRC Press, Boca Raton, Florida.

OECD (1986) Hazard Assessment Project, report of workshop on practical approaches for the assessment of environmental exposure, ENV/CHEM/CM/86.8.

Randers, J. (1973) DDT movement in the global environment. In: Meadows, D. L. and Meadows, D. H. (Eds) Toward Global Equilibrium, Wright-Allen Press, Cambridge, Mass., pp. 49-83.

Rao, P. S. C., Hornsby, A. G. and Jessup, R. E. (1985) Indices for ranking the potential for pesticide contamination of groundwater. Proc. Soil Crop Sci. Soc. 44, 1-24. 

Roberts, J. R. and Marshall, W. K. (1980) Retentive capacity: an index of chemical persistence expressed in terms of chemical-specific and ecosystem-specific parameters. Ecotoxicol. Environ. Saf. 4, 158-171.

Sheehan, P., Korte, F., Klein, W. and Bourdeau, P. (Eds) (1985) Appraisal of Tests to Predict the Environmental Behaviour of Chemicals, SCOPE 25, John Wiley & Sons, Chichester, New York, Brisbane, Toronto, Singapore.

Suntio, L. R., Shiu, W. Y., Mackay, D., Seiber, J. N. and Glotfelty, D. (1988) Critical review of Henry's law constant for pesticides. Rev. Environ. Contam. Toxicol. 103, 1-59.

Travis, C. C. and Hattemer-Frey, H. A. (1988) Uptake of organics by aerial plant parts: a call for research. Chemosphere 17, 227-283.

USEPA (US Environmental Protection Agency) (1978) Toxic Substances Control Act; chemical information rules. Fed. Reg. 43, 11318.

USEPA (1987) Graphical Exposure Modelling System (GEMS), Personal Computer Version, PC CHEM, US Environmental Protection Agency, Environmental Criteria and Assessment Office, Research Triangle Park, North Carolina 27711.

Vighi, M. and Calamari, D. (1989). Evaluative models and field work in estimating the exposure. In: Sommerville, L., and Walker, R. (Eds) Pesticide Effect on Terrestrial Wildlife, Taylor & Francis, New York.

Worthing, C. R. and Walker, S. B. (1987) The Pesticide Manual. A World Compendium (8th edn), The British Crop Protection Council, Thornton Heath, UK.