Review Articles
Chemicals Regulation Assessment
Environmental Risk Assessment of Existing Chemicals Jan Ahlers, Robert Diderich, Ursula Klaschka, Annette Marschner, Beatrice Schwarz-Schulz
German Federal Environmental Agency (Umweltbundesamt), Bismarckplatz 1, D-14193 Berlin, Germany
Abstract Most of the existing chemicals of high priority have been released into the environment for many years. Risk assessments for existing chemicals are now conducted within the framework of the German Existing Chemicals Program and by the EC Regulation on Existing Substances. The environmental assessment of a chemical involves: a) exposure assessment leading to the derivation of a predicted environmental concentration (PEC) of a chemical from releases due to its production, processing, use, and disposal. The calculation of a PEC takes into account the dispersion of a chemical into different environmental compartments, elimination and dilution processes, as well as degradation. Monitoring data are also considered. b) effects assessment. Data obtained from acute or long-term toxicity tests are used for extrapolation on environmental conditions. In order to calculate the concentration with expectedly no adverse effect on organisms (Predicted No Effect Concentration, PNEC) the effect values are divided by an assessment factor. This assessment factor depends on the quantity and quality of toxicity data available. In the last step of the initial risk assessment, the measured or estimated PEC is compared with the PNEC. This "risk characterization" is conducted for each compartment separately (water, sediment, soil, and atmosphere). In case PEC > PNEC an attempt should be made to revise data of exposure and/or effects to conduct a refined risk characterization. In case PEC is again larger than PNEC risk reduction measures have to be considered.
1
Introduction
Contrary to the so-called new substances (i.e. substances notiffed after 1981), the existing chemicals have not a priori been subject to registration procedures or, consequently, risk assessment analysis. In 1982, the Chemicals Act came into force in Germany. Thereupon, increased activities started in the field of existing chemicals, parallel to the notification w o r k on new chemicals. The Chemicals Act was the legal framework but, for existing chemicals, it did not provide instructions for risk evaluation and control. Thus, on a voluntary basis, the German government initiated a cooperation between industry, government, and the scientific community. This Existing Chemicals Concept implies the following steps: - registration of all relevant substances - compilation of basic data sets for high production volume chemicals - priority setting - comprehensive reports - risk assessment - risk management. Up to the present time - basic data sets have been published for almost 600 high production volume chemicals
ESPR- Environ. Sci. & Pollut. Res. 1 (2) 117-123 (1994) 9 ecomedpublishers, D-86899 Landsberg, Germany
- about 150 comprehensive reports have been finalized - priority setting has been performed for 400 substances - proposals for the classification "dangerous for the environment" have been prepared for 200 substances by the German Federal Environmental Agency - environmental risk assessments have been performed for 60 substances. These results of the German Chemicals Program are available to the EU. 2
T h e E u r o p e a n C o u n c i l (EC) Regulation f o r Existing Chemicals
In June 1993, the EC regulation No. 7 9 3 / 9 3 on the Evaluation and Control of the Risk of Existing Substances came into force, which includes several elements of the German Existing Chemicals Conception. The producer has to provide all available data for substances with a production volume of more than 1,000 t/a. The comparison of exposure and effects allows a priority setting for health, occupational, and environmental aspects conducted by the EC. A list of priority substances will be annually published by the Commission. For such substances a risk assessment on environmental, health, and occupational aspects has to be performed by the member countries. This assessment is performed according to an European Commission Regulation on Risk Assessment for Existing Chemicals to be adopted by June 4th, 1994. The producer has to submit additional data for priority substances to allow a refined assessment on the basis of which it can be decided whether or not measures will be neccessary. 3
Risk Assessment
The German Federal Environmental Agency (UBA) is responsible for the environmental risk assessment of chemicals. Within the national Existing Chemicals Program it cooperates with the German Advisory committee on Existing Chemicals (BUA). In 1992, a comprehensive concept for the environmental assessment of existing chemicals was published (AHLERS et al. 1992; AHLERSet al. 1993). This concept was applied to about 60 chemicals and is n o w being amended as to the requirements of the EC Regulation on Risk Assessment for Existing Chemicals. 3.1
Evaluation of Data on New and Existing Chemicals
Despite the similarities in the assessment of new and existing chemicals, there are also differences; e.g., for existing substances data may be available from environmental monitoring or from simulation tests regarding environmental fate etc. In these cases there is the potential for greater insight
117
Chemicals Regulation Assessment
Review Articles
into the risks presented by a particular substance and hence a better base from which to propose control measures, if any. The evaluation of new substances depends on the availability of test data almost always obtained from tests conducted according to internationally accepted test guidelines and carried out in a laboratory subject to GLP. Although many data are available for existing chemicals, their quality is variable. Furthermore, when assessing the exposure of the environment to existing chemicals, it has to be borne in mind that, unlike for new chemicals, large quantities of the chemicals have possibly been released over decades into the environment, and that accumulative processes may have resulted in a so-called "background concentration" in the environment.
3.2
General Procedure
Basically, measured or estimated environmental concentrations (PEC) are compared with concentrations which are expected not to have negative effects on organisms (Predicted No Effect Concentration, PNEC). PNECs result from acute or long-term laboratory tests or, in a few cases, from model ecosystem tests ( - Fig. 1). priority substances
_~
f
available data
effect
exposure assessment (see section 3.3)
assessment
~_
(see section 3.4)
V--
Predicted No Effect Concentration (PNEC)
Predicted Environmental Concentration (PEC)
risk characterization (see section
3.5)
V
PEC/PNEC risk reduction t strategy
no immediate I concern
i furthertesting
improvementof data
Fig. 1: Environmental risk assessment for existing chemicals
118
3.3
Environmental Exposure Assessment
3.3.1 Procedure The objective of the exposure assessment is to predict the environmental concentration of a chemical. The general approach consists of the following steps: - analysis of all potential emission sources - identification of relevant environmental compartment(s) - estimation of quantities released to particular environmental compartments and of elimination and dilution processes to compute a Predicted Environmental Concentration (PEC) and to reflect the resulting concentration immediately after leaving the technosphere - estimation of distribution and degradation of substances as a function of time and space to calculate a refined PEC, which reflects a longer-term exposure - consideration of monitoring data. The basic required information for a PEC-calculation, the production volume, should be provided by the producer/importer. If the quantities released into the environment cannot be deduced from the submitted data, worst-case emission factors have to be applied to the production volume. 1. Elimination Processes Before Entering the Environment In sewage treatment plants a substance can be eliminated by aerobic/anaerobic biodegradation, volatilisation, adsorption, or precipitation. The elimination factor in a waste water treatment plant should be based on analytically determined influent and effluent concentrations. Elimination in treatment plants is quite diverse, depending on operational conditions, such as retention time in the aeration tank, aeration intensity, influent concentration, age and adaptation of sludge, extent of utilization etc. Mathematical models are available to estimate the elimination of a chemical in a waste water treatment plant due to its physico-chemical properties and based on the results from laboratory biodegradation tests. These models can be used for a first approximation, if measured data are not available. 2.
Dilution Processes Upon Entering the Environment
If exaCt hydrological data are available, the calculation should be performed with the 10-percentile value of the flow rate Qnwr (this relatively low value is not reached by 10 % of the measurements in the given period). A comparison between the 10-percentile values and the average flows for German rivers shows that the average flows are by a factor 3 - 4 higher than the 10-percentile values. Therefore, if the average flow is known, the calculation can be based upon one third of the long-standing average. If the above data are not available, a dilution factor of 10 is assumed. The dilution of a chemical entering the atmosphere through stacks or vents is estimated with a Gaussian plume model. 3. Degradation and Distribution Processes in the Environment After its release into the environment, a chemical can undergo several processes which infuence the PEC: the distribution processes between hydrosphere, atmosphere and geosphere (leaving the substance unchanged) as well as chemi-
ESPR-Environ. Sci. & Pollut. Res. 1 (2) 1994
Review Articles
cal conversion processes resulting in a veritable elimination of the substance. Also, due to degradation processes, stable metabolites occur, which have to be assessed separately.
(1) Volatilization from Water The vaporization is the mass-transfer pathway from water to the atmosphere. The Henry's law constant H is an indicator of a substance's volatility from water. Reversely, a substance released into the atmosphere can be washed out with rain back into the hydrosphere and geosphere (wet deposition). (2) Adsorption/Desorption The concentration of a substance in water can be diminished by sorption to suspended matter or sediment. By way of an adsorption/desorption experiment, the organic carbon adsorption coefficient Koc can be determined. On the basis of the PEC for the hydrosphere, the concentration of the substances in sediment can be calculated. For the exposure of the geosphere, the Koc value informs on the potential of a chemical to accumulate or to contaminate the ground water. (3) Biodegradation Via biodegradation, substances are decomposed by microorganisms of soil or water. This degradation can proceed to complete mineralization or can stop at the stage of stable metabolites. For the calculation of a PEC, information on the degradation percentage after the release from the technosphere is necessary. However, it is difficult to adapt the results from laboratory experiments to field conditions. (4) Hydrolysis In exposure assessments, hydrolytical degradation is of special significance for chemicals which are not biodegradable. Hydrolysis leads to primary degradation. The potential environmental hazards of the hydrolysis reaction products have to be considered. (5) Photodegradation in the Hydrosphere Direct photolysis can only occur in the upper water layers. The photolysis process leads to primary degradation. Indirect phototransformation processes are also possible. They are especially important for compounds which do not absorb wavelengths > 290 nm and include reactions with photoreactants (e.g. 02, R O - , O H - , H O O " radicals). (6) Photodegradation in the Atmosphere In addition to direct photolysis, a chemical can be degraded in the atmosphere by reactions with a number of reactive intermediate species, including O H - , HO2", N O 3" radicals, and 03. For the majority of organic chemicals, the reaction with O H radicals during daylight hours is expected to be the dominant atmospheric removal process.
3.3.2
Determination of Long-Term PECs
In many cases, the calculation of an initial environmental concentration is sufficient for assessing a chemical. In some cases it is inevitable to calculate a long-term PEC: - if an environmental compartment is indirectly exposed by distribution processes ESPR-Environ. Sci.& pollut. Res. 1 (2) 1994
Chemicals Regulation Assessment
- if a chemical is released through different point sources into the same water body - if a chemical is released diffusely or simultaneously through diffuse and point sources - for the evaluation of a concentration in soil, especially if the input rate exceeds the output rate by elimination processes. For the determination of long-term PECs, the following exposure scenarios have to be considered: - release through point sources implying a local (sitespecific) evaluation - diffuse release, implying a regional evaluation - simultaneous diffuse and point source release. 3.3.3
Determination of Regional PECs
For a chemical released diffusely into one or several environmental compartments, a regional distribution model (e.g. the Fugacity Models by MACKAY) is appropriate to determine long-term PECs in the compartments air, water, soil, and sediment. With regional distribution models, a steady-state concentration in every compartment can be calculated.
1. Relation Between Regional and Local PECs A more complex situation arises when chemicals are released both diffusely and through point sources. This is probably the case for chemicals with a wide dispersive use. During production and processing high concentrations can occur in the vicinity of the production site, whereas the widespread use of the chemical leads to a lower but ubiquitous (regional) concentration in the environment. This "background" concentration in the environment due to diffuse release has to be added to concentrations arising from point sources.
3.3.4 Use of Monitoring Data 1. Selection of Representative Data It has to be ascertained whether the data are the result of sporadic studies or whether they are detected at the same site over a certain period of time. Data from a prolonged program, including seasonal fluctuations, are of special interest for the exposure assessment. The 90-percentile values are of highest preference. 2. Evaluation of the Geographical Relation Between Emis-
sion Sources and Sampling Site If a substance is released into the environment through point sources, this emission might be already included in the measured concentrations (e.g. data from a monitoring program in an industrial area). If there is no regional proximity between the sampling site and the emission source, the data (e.g. from rural regions) might represent a background concentration that has to be added to the calculated release through point sources. 3. Verification of the Quality of the Applied Measuring
Techniques The applied techniques of sampling, processing and detection have to consider the physico-chemical properties of the
] 19
Chemicals Regulation Assessment compound. The detection limits of the analytical method must be beneath the PNEC established in the effect assessment; otherwise even with the result "not detectable" from measurement, a concentration of concern has to be expected. The environmental concentration of a "not-detectable" chemical should be assumed to equal the detection limit. After the determination of a calculated PEC and selection of qualified monitoring data, calculated and measured concentrations have to be compared. Analysis and critical discussion of divergences are important steps for developing an environmental risk assessment of existing chemicals. 3.4 Assessment of Environmental Effects The principle of the effects assessment is 1. to use toxicity data from acute and long-term experiments for the selection of an appropriate assessment factor 2. to calculate a predicted no-effect concentration (PNEC). The data from laboratory tests used for the extrapolation to the environment include uncertainties due to the following aspects: 1. differing sensitivities of each individual of one species (biological variance) 2. differing sensitivities of species 3. variances of the results between different laboratories 4. extrapolation of acute toxicity (LC/ECs0) to chronic toxicity (NOEC) 5. extrapolation of laboratory data to ecosystems 6. additive or synergistic effects. These aspects have to be considered for the assessment factor. Almost all ecotoxicity data on existing chemicals refer to the aquatic compartment. However, the same general conclusions for aquatic tests can be applied to tests with species from other environmental compartments. 3.4.1
Assessment of Aquatic Effects
The quantity and quality of data available for existing chemicals vary considerably, ranging from less than base-set information for new chemicals to very well studied chemicals. Within the framework of the EC Regulation for Existing Chemicals only priority chemicals with a production volume of more than 1,000 t / a will have to be assessed. For these chemicals base-set information has to be provided by the manufacturer. For the aquatic environment the base set includes acute toxicity studies with the standard organisms fish, Daphnia and algae. In accordance with the EC Environmental Risks Assessment Regulation for Existing ChemiCaLlSand the Technical Guidance Documents for New Chemicals (1993) an assessment factor of 1,000 is considered to be appropriate to calculate a PNEC from the lowest valid LC/ECs0 values obtained ( ~ Table 1). If, in addition to the base set, long-term toxicity tests have been performed, considerably lower assessment factors can be applied on the NOEC derived from these studies. They range from 100 for one long-term study to 10 for three of such studies ( ~ Table 1). From the scientific point of view factor 10 is too low as the uncertainties listed in section 3.4 cannot be covered completely. However, since the factor has been established and used in the past, it is employed for pragmatic reasons only. Assessment factors resulting from field studies or from investigations using model ecosystems are derived case by
120
Review Articles case. These assessment factors are used in accordance with the technical guidance documents for new chemicals if, in long-term tests, the most sensitive organism in the acute test has been investigated. If this test is not available, it should be performed. However, to reduce the number of tests, all valid data available should be taken into consideration. Therefore, higher assessment factors may be applied in the absence of a long-term study with the most sensitive species in the base set. Table 1: Assessment factors for the aquatic c o m p a r t m e n t Assessment factors to be applied to the lowest EC/LC50 resp. NOEC
Available valid data
AF At least one acute test of the three trophic levels of the base set (fish, Daphnia, and algae)
1,000
Long-term toxicity tests in addition to the base set: Long-term toxicity test on one species, either fish, or Daphnia, or algae
100
Long-term toxicity tests on two species, either fish, and/or Daphnia and/or algae
50
Long-term toxicity tests on three species, fish, Daphnia, algae
10
Field data/data of model ecosystems
case by case
A large quantity of data on ecotoxicological effects for existing substances were obtained several years ago and do not correspond to standard assays. Nevertheless if the validity of the test results can be confirmed, these data should also be considered. The assessment factors in Table 1 are selected for a clearly defined data basis. To adapt the assessment factors to the actual data situation, it is suggested that the assessment factors ( ~ Table 1) should be modified in a case by case decision. The factor of 1,000 should not be increased because less information than the base set is not acceptable. A factor smaller than 10 cannot be applied for mono species laboratory tests, due to the remaining uncertainties. A system with fixed assessment factors for common "data situations" and the possibility for the expert to modify them to a limited extent seems to be the best compromise between the desire for strict rules and the necessity to reflect the degree of uncertainty as exactly as possible. Substances that bioaccumulate cause a higher risk for a longer time period, which is not considered in an acute test. Therefore, the base set is not sufficient for the assessment (measured BCF > 50). If no measured BCF is available, the bioaccumulation potential (e.g. log Pow > 3) can be used. We consider a chronic test necessary for the assessment of those substances. The chronic test should be appropriate to indicate effects due to bioaccumulation (e.g. modified repro-
ESPR-Environ. Sci. & Pollut. Res. 1 (2) 1994
Review Articles
duction test for Daphnia including clearing phase, early life stage test, full life cycle test). Alternatively, it should be discussed whether for substances with bioaccumulation potential the assessment factor could be increased. 3.4.2
Assessment of Effects on Sediments
There are only few tests for sediment organisms available. The selection of both representative organisms and standardized sediments is being discussed. Various approaches are being developed to investigate the effects of chemicals on sediment and sediment organisms (OECD, 1992). As soon as standardized test methods are available and the assessment factors agreed upon, the calculated PNEC~ed can be compared with the estimated concentration in sediment (PECk) or with the measured concentration in sediment. In the absence of ecotoxicological effect data on sediment-dwelling organisms, the PNEC~ed may provisionally be calculated with the PNEC for aquatic organisms and the sediment/water partitioning coefficient (OECD, 1992). 3.4.3
Assessment of Effects on Terrestrial Compartments
If the following two criteria are fulfilled, an assessment of effects on the terrestrial compartment is to be conducted: 1. The substance is discharged into the soil in considerable quantities and remains in this compartment. 2. The substance has a toxic potential for aquatic organisms. As the information on terrestrial toxicity data is limited for most of the existing substances, tests on aquatic organisms could be used for an initial approach. Aquatic organisms might show a similar sensitivity to chemicals as soil organisms which are exposed mainly to the pore water of the soil. Further effects which chemicals, adsorbed to soil particles, execute on soil organisms by ingestion cannot be considered by this initial approach. This is one of the reasons why this approach can give only a first idea regarding terrestrial toxicity. If in this first step the PEC~oa-vo. . . . . t e r is larger than the PNECwater, tests with terrestrial organisms are indispensable for the assessment of terrestrial effects. Especially for highly lipophilic chemicals, other indications for a toxic potential for the terrestrial compartment beside the aquatic toxicity should be considered, e.g. high toxicity for mammalians (LDs0 oral --> 200 mg/kg BW or LDs0 dermal 400 mg/kg BW), or for birds (LDs0~c~ >-- 50 mg/kg BW). The effects of a substance on the compartment soil can be divided into two fields: 1. Effects on biotic parameters: The application of a chemical can cause a considerable change of the species pattern of soil organisms and a reduction of the variety of the species. Terrestrial tests with higher plants or earthworms have been conducted for only a few existing chemicals. The minimal assessment basis should be one test for fauna and one test for flora. 2. Effects on abiotic parameters: Chemicals can affect processes in the soil like filtration, buffering, metabolism and degradation (UBA 1993). For the characterization of abiotic effects the investigation of degradation has first prioriry. This information allows to characterize the change of the soil properties and to revise the estimation of PECso~.
ESPR-Environ. Sci. & Pollut. Res. 1 (2) 1994
Chemicals Regulation Assessment
In this stage of the initial risk assessment, adsorption, desorption, and mobility can be estimated by model calculations. Corresponding tests of percolation and adsorption may be necessary for the refined risk assessment. The same assessment factors as for the aquatic system (-* Table 1) should be used for the terrestrial system according to the type of investigations (acute or long-term toxicity test) and the number of trophic levels tested. If only acute toxicity data are available, a factor of 1,000 is applied to the lowest E(L)C50. If in addition to the acute tests one, two, or three long-term test(s) of different trophic levels are conducted, an assessment factor of 100, 50, or 10 has to be applied to the lowest NOEC to calculate the PNECso ~. The abiotic effects should be assessed qualitatively. A substance should be considered as hazardous to the soil if its degradation is smaller than its discharge into the soil or if it changes the soil properties in such a way that noxious effects result (see German Federal Government, 1985 and ISO, 1991). If according to the initial risk characterization PEC~o,j > PNEC~oil, a further testing program is to be initiated to improve the PNEC~o~. Depending on the effect of a chemical on vascular plants, earthworms, or microorganisms, the information about the ecotoxicological effect on the most sensitive group of organisms (primary producers, consumers, or destruents) has to be refined (further testing). If a third trophic level has not been tested yet, an organism of this trophic level has to be tested (acute investigations). For further testing, usually long-term tests (with e.g. earthworms, springtails, enchytraeids, and staphylinids) are preferred. 3.4.4
Assessment of Atmospheric Effects
An assessment of atmospheric effects should only be performed under the following conditions: half-life tl/2 > 10days or diffuse emission > 10 t / a or one point-source emission > 1 t/a. No internationally accepted test guidelines have been developed for the assessment of air effects. For most of the chemicals only data on mammalian toxicity are available. For some chemicals investigations on the toxicity of invertebrates, mostly insects (e.g. Apis mellifera, Syrphus corollae) might be available which have been conducted according to guidelines for the testing of plant protection agents. Only in some of these tests it is possible to determine an effect concentration. Therefore, a PNEC can usually not be derived. Special test guidelines for the effects in the atmosphere have to be developed. Consecutively, the development of an appropriate assessment factor is needed. For the evaluation of an atmospheric risk the following physico-chemical effects of a chemical on the atmosphere have to be considered as well: - degrading air quality (DAQ) - tropospheric ozone building (TOB) - acidification - depletion of the ozone layer - warming up of the atmosphere - greenhouse warming potential (GWP). The potential to cause adverse effects to the atmosphere by DAQ, TOB or GWP should be determined by an expert
121
Chemicals Regulation Assessment judgement for organic chemicals (gaseous, volatile liquids or solids), which show a vapour pressure >__ 100 Pa or a Henry's Constant > 0.01 P a x m3/mol. Organic chemicals which contain F a n d / o r CI or Br atoms and which are gaseous or volatile liquids with a vapour pressure _> 100 Pa should immediately be subject to a first qualitative assessment of the ozone depletion potential (ODP).
3.4.5
Assessment of Secondary Poisoning
Secondary poisoning is the result of the transfer of chemicals via food chain to different trophic levels (biomagnification) forming residues which may be detrimental to the last organism(s) in the chain. The food chain (water ~ fish -~ fish-eating bird or fisheating mammal) is one example of a secondary poisoning pathway. Safe levels for fish-eating animals do not exclude risks for birds or mammals feeding on other aquatic organisms (e.g. mussels, worms). An assessment of secondary poisoning has to be performed if the following criteria are fulfilled: First, it has to be decided whether or not indirect exposure of ecosystems (or man) is likely to occur. Second, it has to be considered whether there are indications for a bioaccumulation potential. Subsequently, it is necessary to consider whether the substance is classified on the basis of its mammalian toxicity. It is assumed that the available mammalian toxicity data can indicate the possible risks of the chemical to higher organisms in the environment. Effects on avian and mammalian populations should be representative for fish-eating birds or mammals (PNECor~). For example, results of the 28-day repeated dose test have to be used preferably for the assessment of secondary poisoning effects. In this case, PNECoral is calculated on the basis of the factor 100. Only toxicity studies on dietary and oral exposure are relevant, as the pathway for secondary poisoning refers exclusively to the uptake through the food chain. The results of these tests can be expressed as a concentration in the food (mg/kg) or a dose (mg/kg bodyweight/day) causing no effect. For the assessment of secondary poisoning, the results are converted to the concentration in food (mg/kg food). To extrapolate to predicted no-effect concentration from the test results in food that should be protective for other mammalian and avian species, an assessment factor is necessary. For the assessment of secondary poisoning the predicted concentration in food (provisional PECoral = PEC ..... x Pow x 0.05) has to be compared with the predicted no effect concentration for fish-eating birds or mammals (PNECor~I). A measured BCF can be used instead of the log Pow for the calculation of a final PECo~a (PECo~~ = PECwaterx BCF). If the provisional PECor ~ exceeds the PNECot~, secondary poisoning could be a critical pathway for fish-eaters and a bioaccumulation test (BCF-test OECD guideline 305) has to be conducted.
3.5
Risk Characterization and Further Testing
As soon as all available information are evaluated and a PEC as well as a PNEC have been derived, an initial risk charac-
122
Review Articles terization can be performed by comparing the PEC with the PNEC (-~ Fig. 2) for the relevant compartment. An initial risk characterization is also performed if only few investigations on a chemical are available. The uncertainty in the initial risk characterization due to the low quality and quantity of the data is to be compensated by sufficiently high assessment factors. If in this initial assessment PEC _< PNEC no further tests are required unless the production volume or the use pattern change considerably. If PEC > PNEC an environmental hazard is apparent. In this case more information on exposure and/or effects should be obtained (e.g. by repeating certain tests under valid conditions or performing long-term tests or tests with organisms of trophic levels not yet considered). In addition, detailed information on exposure helps to avoid worst-case assumptions for calculating a refined PEC. In many cases measurement of environmental concentrations should be initiated. With this additional information the new refined PEC and/or PNEC can differ from the initial values. Therefore, whenever new and relevant data are obtained, PEC and PNEC have to be compared again in a refined risk characterization. If PEC _< PNEC the assessment procedure is completed, i.e. no further information is necessary. This iterative process has precautionary aspects especially in the initial phase as data gaps are filled in by worst-case assumptions or high assessment factors. Morever, this approach saves resources as only a minimum of tests have to be performed. By the preliminary assessment possible hazards are identified at an early stage and a specific testing plan can be elaborated for a maximum of additional information at low expenses.
3.6
Risk Management
If after a refined risk characterization PEC exceeds PNEC and if an improvement of the data basis does not seem to be feasible or does not change the result of the assessment, measures for risk reduction have to be discussed. The EC Existing Chemicals Regulation demands cost/benefit analyses and investigations on possibilities for a substitution of the chemical. However, the most specific measure which efficiently reduces the risk for the concerned environmental compartment should always be proposed. About 60 existing chemicals have been assessed since 1992 in Germany. For 8 % of these chemicals the assessment revealed the need for measures to reduce emissions. According to the initial assessment for 26 % more data on exposure a n d / o r effects have to be collected or the results of measures already taken have to be observed. The majority of substances assessed did not show current concern.
4
Discussion
The task to protect the environment against the hazards of the more than 100,000 substances listed in the European Inventory of Existing Chemicals (EINECS) seems to be immense. However, the conception of the German Existing Chemicals Program and, hopefully, the EC Existing Chemi-
ESPR-Environ. Sci. & PoUut. Res. 1 (2) 1994
Review Articles
Chemicals Regulation Assessment
effect assessment determination of PNEC*
r
I
J
v initial risk
identification of representative monitoring data
exposureassessment / determination of PEC*
~
1,000 t/a) a systematic priority setting has been performed. These substances of highest priority are the first candidates for a comprehensive risk assessment including risk reduction measures. For this process all available information should be used. The German practice during the recent years showed that the data situation is very different for each chemical and therefore the assessment of these chemicals cannot be stereotyped. Expert judgement is important in each case. An indication that relevant chemicals are covered by this procedure comes from recent accidents in German chemical plants. In most cases, critical data had already been collected by the German Existing Chemicals Program, so that an assessment could be performed or had already been available. These incidents confirm that the ongoing process will contribute to decrease the hazards for the environment, especially as soon as all EU coun-
ESPR-Environ. Sci. & Pollut. Res. 1 (2) 1994
tries participate and as soon as a close cooperation with other industrial countries within the O E C D can be achieved. 5
Literature
AHLERS,J.; KOCH, W.; LANCE, A.; MARSCHNER,A.; WELTER, G. (1992): Bewertung der Umweltgef~ihrlichkeityon Mten Stoffen nach dem Chemikaliengesetz (ChemG). ChemikaliengesetzHeft 10, Texte 19/92, Umweltbundesamt, Berlin, M~irz 1992 AHLERS,J.; KOCH,W.; MARSCHNER,A.; WELTER,G. (1993): The Sci. Total Environm. Supplement 1993,1587-1596 Deutscher Bundestag (1985): Bodenschutzkonzeption der Bundesregierung BT-Drs. 10/2977 vom 7. Miirz 1985 ISO (International Organisation for Standardization) (1991): Soil Quality, Vocabulary, Part 1: Terms and Definitions Relating to Soil Protection and Soil Pollution, ISO/CD 11071 OECD (1992): Report of the Workshop'on Effects Assessment in Sediment. Copenhagen, Denmark, 13-15 May 1991 Technical Guidance Documents in Support of the Risk Assessment Directive (93/67/EEC)for Substances Notified in Accordance with the Requirements of Council Directive 67/548/EEC.Brussels, 1993 UBA (Umweltbundesamt) (1993): Entwurf zur Bewertung yon Bodenbelastungen, Fachgebiet I 3.7
123
Chemicals Regulation Assessment
Environmental Risk Assessment of Existing Chemicals Jan Ahlers, Robert Diderich, Ursula Klaschka, Annette Marschner, Beatrice Schwarz-Schulz
German Federal Environmental Agency (Umweltbundesamt), Bismarckplatz 1, D-14193 Berlin, Germany
Abstract Most of the existing chemicals of high priority have been released into the environment for many years. Risk assessments for existing chemicals are now conducted within the framework of the German Existing Chemicals Program and by the EC Regulation on Existing Substances. The environmental assessment of a chemical involves: a) exposure assessment leading to the derivation of a predicted environmental concentration (PEC) of a chemical from releases due to its production, processing, use, and disposal. The calculation of a PEC takes into account the dispersion of a chemical into different environmental compartments, elimination and dilution processes, as well as degradation. Monitoring data are also considered. b) effects assessment. Data obtained from acute or long-term toxicity tests are used for extrapolation on environmental conditions. In order to calculate the concentration with expectedly no adverse effect on organisms (Predicted No Effect Concentration, PNEC) the effect values are divided by an assessment factor. This assessment factor depends on the quantity and quality of toxicity data available. In the last step of the initial risk assessment, the measured or estimated PEC is compared with the PNEC. This "risk characterization" is conducted for each compartment separately (water, sediment, soil, and atmosphere). In case PEC > PNEC an attempt should be made to revise data of exposure and/or effects to conduct a refined risk characterization. In case PEC is again larger than PNEC risk reduction measures have to be considered.
1
Introduction
Contrary to the so-called new substances (i.e. substances notiffed after 1981), the existing chemicals have not a priori been subject to registration procedures or, consequently, risk assessment analysis. In 1982, the Chemicals Act came into force in Germany. Thereupon, increased activities started in the field of existing chemicals, parallel to the notification w o r k on new chemicals. The Chemicals Act was the legal framework but, for existing chemicals, it did not provide instructions for risk evaluation and control. Thus, on a voluntary basis, the German government initiated a cooperation between industry, government, and the scientific community. This Existing Chemicals Concept implies the following steps: - registration of all relevant substances - compilation of basic data sets for high production volume chemicals - priority setting - comprehensive reports - risk assessment - risk management. Up to the present time - basic data sets have been published for almost 600 high production volume chemicals
ESPR- Environ. Sci. & Pollut. Res. 1 (2) 117-123 (1994) 9 ecomedpublishers, D-86899 Landsberg, Germany
- about 150 comprehensive reports have been finalized - priority setting has been performed for 400 substances - proposals for the classification "dangerous for the environment" have been prepared for 200 substances by the German Federal Environmental Agency - environmental risk assessments have been performed for 60 substances. These results of the German Chemicals Program are available to the EU. 2
T h e E u r o p e a n C o u n c i l (EC) Regulation f o r Existing Chemicals
In June 1993, the EC regulation No. 7 9 3 / 9 3 on the Evaluation and Control of the Risk of Existing Substances came into force, which includes several elements of the German Existing Chemicals Conception. The producer has to provide all available data for substances with a production volume of more than 1,000 t/a. The comparison of exposure and effects allows a priority setting for health, occupational, and environmental aspects conducted by the EC. A list of priority substances will be annually published by the Commission. For such substances a risk assessment on environmental, health, and occupational aspects has to be performed by the member countries. This assessment is performed according to an European Commission Regulation on Risk Assessment for Existing Chemicals to be adopted by June 4th, 1994. The producer has to submit additional data for priority substances to allow a refined assessment on the basis of which it can be decided whether or not measures will be neccessary. 3
Risk Assessment
The German Federal Environmental Agency (UBA) is responsible for the environmental risk assessment of chemicals. Within the national Existing Chemicals Program it cooperates with the German Advisory committee on Existing Chemicals (BUA). In 1992, a comprehensive concept for the environmental assessment of existing chemicals was published (AHLERS et al. 1992; AHLERSet al. 1993). This concept was applied to about 60 chemicals and is n o w being amended as to the requirements of the EC Regulation on Risk Assessment for Existing Chemicals. 3.1
Evaluation of Data on New and Existing Chemicals
Despite the similarities in the assessment of new and existing chemicals, there are also differences; e.g., for existing substances data may be available from environmental monitoring or from simulation tests regarding environmental fate etc. In these cases there is the potential for greater insight
117
Chemicals Regulation Assessment
Review Articles
into the risks presented by a particular substance and hence a better base from which to propose control measures, if any. The evaluation of new substances depends on the availability of test data almost always obtained from tests conducted according to internationally accepted test guidelines and carried out in a laboratory subject to GLP. Although many data are available for existing chemicals, their quality is variable. Furthermore, when assessing the exposure of the environment to existing chemicals, it has to be borne in mind that, unlike for new chemicals, large quantities of the chemicals have possibly been released over decades into the environment, and that accumulative processes may have resulted in a so-called "background concentration" in the environment.
3.2
General Procedure
Basically, measured or estimated environmental concentrations (PEC) are compared with concentrations which are expected not to have negative effects on organisms (Predicted No Effect Concentration, PNEC). PNECs result from acute or long-term laboratory tests or, in a few cases, from model ecosystem tests ( - Fig. 1). priority substances
_~
f
available data
effect
exposure assessment (see section 3.3)
assessment
~_
(see section 3.4)
V--
Predicted No Effect Concentration (PNEC)
Predicted Environmental Concentration (PEC)
risk characterization (see section
3.5)
V
PEC/PNEC risk reduction t strategy
no immediate I concern
i furthertesting
improvementof data
Fig. 1: Environmental risk assessment for existing chemicals
118
3.3
Environmental Exposure Assessment
3.3.1 Procedure The objective of the exposure assessment is to predict the environmental concentration of a chemical. The general approach consists of the following steps: - analysis of all potential emission sources - identification of relevant environmental compartment(s) - estimation of quantities released to particular environmental compartments and of elimination and dilution processes to compute a Predicted Environmental Concentration (PEC) and to reflect the resulting concentration immediately after leaving the technosphere - estimation of distribution and degradation of substances as a function of time and space to calculate a refined PEC, which reflects a longer-term exposure - consideration of monitoring data. The basic required information for a PEC-calculation, the production volume, should be provided by the producer/importer. If the quantities released into the environment cannot be deduced from the submitted data, worst-case emission factors have to be applied to the production volume. 1. Elimination Processes Before Entering the Environment In sewage treatment plants a substance can be eliminated by aerobic/anaerobic biodegradation, volatilisation, adsorption, or precipitation. The elimination factor in a waste water treatment plant should be based on analytically determined influent and effluent concentrations. Elimination in treatment plants is quite diverse, depending on operational conditions, such as retention time in the aeration tank, aeration intensity, influent concentration, age and adaptation of sludge, extent of utilization etc. Mathematical models are available to estimate the elimination of a chemical in a waste water treatment plant due to its physico-chemical properties and based on the results from laboratory biodegradation tests. These models can be used for a first approximation, if measured data are not available. 2.
Dilution Processes Upon Entering the Environment
If exaCt hydrological data are available, the calculation should be performed with the 10-percentile value of the flow rate Qnwr (this relatively low value is not reached by 10 % of the measurements in the given period). A comparison between the 10-percentile values and the average flows for German rivers shows that the average flows are by a factor 3 - 4 higher than the 10-percentile values. Therefore, if the average flow is known, the calculation can be based upon one third of the long-standing average. If the above data are not available, a dilution factor of 10 is assumed. The dilution of a chemical entering the atmosphere through stacks or vents is estimated with a Gaussian plume model. 3. Degradation and Distribution Processes in the Environment After its release into the environment, a chemical can undergo several processes which infuence the PEC: the distribution processes between hydrosphere, atmosphere and geosphere (leaving the substance unchanged) as well as chemi-
ESPR-Environ. Sci. & Pollut. Res. 1 (2) 1994
Review Articles
cal conversion processes resulting in a veritable elimination of the substance. Also, due to degradation processes, stable metabolites occur, which have to be assessed separately.
(1) Volatilization from Water The vaporization is the mass-transfer pathway from water to the atmosphere. The Henry's law constant H is an indicator of a substance's volatility from water. Reversely, a substance released into the atmosphere can be washed out with rain back into the hydrosphere and geosphere (wet deposition). (2) Adsorption/Desorption The concentration of a substance in water can be diminished by sorption to suspended matter or sediment. By way of an adsorption/desorption experiment, the organic carbon adsorption coefficient Koc can be determined. On the basis of the PEC for the hydrosphere, the concentration of the substances in sediment can be calculated. For the exposure of the geosphere, the Koc value informs on the potential of a chemical to accumulate or to contaminate the ground water. (3) Biodegradation Via biodegradation, substances are decomposed by microorganisms of soil or water. This degradation can proceed to complete mineralization or can stop at the stage of stable metabolites. For the calculation of a PEC, information on the degradation percentage after the release from the technosphere is necessary. However, it is difficult to adapt the results from laboratory experiments to field conditions. (4) Hydrolysis In exposure assessments, hydrolytical degradation is of special significance for chemicals which are not biodegradable. Hydrolysis leads to primary degradation. The potential environmental hazards of the hydrolysis reaction products have to be considered. (5) Photodegradation in the Hydrosphere Direct photolysis can only occur in the upper water layers. The photolysis process leads to primary degradation. Indirect phototransformation processes are also possible. They are especially important for compounds which do not absorb wavelengths > 290 nm and include reactions with photoreactants (e.g. 02, R O - , O H - , H O O " radicals). (6) Photodegradation in the Atmosphere In addition to direct photolysis, a chemical can be degraded in the atmosphere by reactions with a number of reactive intermediate species, including O H - , HO2", N O 3" radicals, and 03. For the majority of organic chemicals, the reaction with O H radicals during daylight hours is expected to be the dominant atmospheric removal process.
3.3.2
Determination of Long-Term PECs
In many cases, the calculation of an initial environmental concentration is sufficient for assessing a chemical. In some cases it is inevitable to calculate a long-term PEC: - if an environmental compartment is indirectly exposed by distribution processes ESPR-Environ. Sci.& pollut. Res. 1 (2) 1994
Chemicals Regulation Assessment
- if a chemical is released through different point sources into the same water body - if a chemical is released diffusely or simultaneously through diffuse and point sources - for the evaluation of a concentration in soil, especially if the input rate exceeds the output rate by elimination processes. For the determination of long-term PECs, the following exposure scenarios have to be considered: - release through point sources implying a local (sitespecific) evaluation - diffuse release, implying a regional evaluation - simultaneous diffuse and point source release. 3.3.3
Determination of Regional PECs
For a chemical released diffusely into one or several environmental compartments, a regional distribution model (e.g. the Fugacity Models by MACKAY) is appropriate to determine long-term PECs in the compartments air, water, soil, and sediment. With regional distribution models, a steady-state concentration in every compartment can be calculated.
1. Relation Between Regional and Local PECs A more complex situation arises when chemicals are released both diffusely and through point sources. This is probably the case for chemicals with a wide dispersive use. During production and processing high concentrations can occur in the vicinity of the production site, whereas the widespread use of the chemical leads to a lower but ubiquitous (regional) concentration in the environment. This "background" concentration in the environment due to diffuse release has to be added to concentrations arising from point sources.
3.3.4 Use of Monitoring Data 1. Selection of Representative Data It has to be ascertained whether the data are the result of sporadic studies or whether they are detected at the same site over a certain period of time. Data from a prolonged program, including seasonal fluctuations, are of special interest for the exposure assessment. The 90-percentile values are of highest preference. 2. Evaluation of the Geographical Relation Between Emis-
sion Sources and Sampling Site If a substance is released into the environment through point sources, this emission might be already included in the measured concentrations (e.g. data from a monitoring program in an industrial area). If there is no regional proximity between the sampling site and the emission source, the data (e.g. from rural regions) might represent a background concentration that has to be added to the calculated release through point sources. 3. Verification of the Quality of the Applied Measuring
Techniques The applied techniques of sampling, processing and detection have to consider the physico-chemical properties of the
] 19
Chemicals Regulation Assessment compound. The detection limits of the analytical method must be beneath the PNEC established in the effect assessment; otherwise even with the result "not detectable" from measurement, a concentration of concern has to be expected. The environmental concentration of a "not-detectable" chemical should be assumed to equal the detection limit. After the determination of a calculated PEC and selection of qualified monitoring data, calculated and measured concentrations have to be compared. Analysis and critical discussion of divergences are important steps for developing an environmental risk assessment of existing chemicals. 3.4 Assessment of Environmental Effects The principle of the effects assessment is 1. to use toxicity data from acute and long-term experiments for the selection of an appropriate assessment factor 2. to calculate a predicted no-effect concentration (PNEC). The data from laboratory tests used for the extrapolation to the environment include uncertainties due to the following aspects: 1. differing sensitivities of each individual of one species (biological variance) 2. differing sensitivities of species 3. variances of the results between different laboratories 4. extrapolation of acute toxicity (LC/ECs0) to chronic toxicity (NOEC) 5. extrapolation of laboratory data to ecosystems 6. additive or synergistic effects. These aspects have to be considered for the assessment factor. Almost all ecotoxicity data on existing chemicals refer to the aquatic compartment. However, the same general conclusions for aquatic tests can be applied to tests with species from other environmental compartments. 3.4.1
Assessment of Aquatic Effects
The quantity and quality of data available for existing chemicals vary considerably, ranging from less than base-set information for new chemicals to very well studied chemicals. Within the framework of the EC Regulation for Existing Chemicals only priority chemicals with a production volume of more than 1,000 t / a will have to be assessed. For these chemicals base-set information has to be provided by the manufacturer. For the aquatic environment the base set includes acute toxicity studies with the standard organisms fish, Daphnia and algae. In accordance with the EC Environmental Risks Assessment Regulation for Existing ChemiCaLlSand the Technical Guidance Documents for New Chemicals (1993) an assessment factor of 1,000 is considered to be appropriate to calculate a PNEC from the lowest valid LC/ECs0 values obtained ( ~ Table 1). If, in addition to the base set, long-term toxicity tests have been performed, considerably lower assessment factors can be applied on the NOEC derived from these studies. They range from 100 for one long-term study to 10 for three of such studies ( ~ Table 1). From the scientific point of view factor 10 is too low as the uncertainties listed in section 3.4 cannot be covered completely. However, since the factor has been established and used in the past, it is employed for pragmatic reasons only. Assessment factors resulting from field studies or from investigations using model ecosystems are derived case by
120
Review Articles case. These assessment factors are used in accordance with the technical guidance documents for new chemicals if, in long-term tests, the most sensitive organism in the acute test has been investigated. If this test is not available, it should be performed. However, to reduce the number of tests, all valid data available should be taken into consideration. Therefore, higher assessment factors may be applied in the absence of a long-term study with the most sensitive species in the base set. Table 1: Assessment factors for the aquatic c o m p a r t m e n t Assessment factors to be applied to the lowest EC/LC50 resp. NOEC
Available valid data
AF At least one acute test of the three trophic levels of the base set (fish, Daphnia, and algae)
1,000
Long-term toxicity tests in addition to the base set: Long-term toxicity test on one species, either fish, or Daphnia, or algae
100
Long-term toxicity tests on two species, either fish, and/or Daphnia and/or algae
50
Long-term toxicity tests on three species, fish, Daphnia, algae
10
Field data/data of model ecosystems
case by case
A large quantity of data on ecotoxicological effects for existing substances were obtained several years ago and do not correspond to standard assays. Nevertheless if the validity of the test results can be confirmed, these data should also be considered. The assessment factors in Table 1 are selected for a clearly defined data basis. To adapt the assessment factors to the actual data situation, it is suggested that the assessment factors ( ~ Table 1) should be modified in a case by case decision. The factor of 1,000 should not be increased because less information than the base set is not acceptable. A factor smaller than 10 cannot be applied for mono species laboratory tests, due to the remaining uncertainties. A system with fixed assessment factors for common "data situations" and the possibility for the expert to modify them to a limited extent seems to be the best compromise between the desire for strict rules and the necessity to reflect the degree of uncertainty as exactly as possible. Substances that bioaccumulate cause a higher risk for a longer time period, which is not considered in an acute test. Therefore, the base set is not sufficient for the assessment (measured BCF > 50). If no measured BCF is available, the bioaccumulation potential (e.g. log Pow > 3) can be used. We consider a chronic test necessary for the assessment of those substances. The chronic test should be appropriate to indicate effects due to bioaccumulation (e.g. modified repro-
ESPR-Environ. Sci. & Pollut. Res. 1 (2) 1994
Review Articles
duction test for Daphnia including clearing phase, early life stage test, full life cycle test). Alternatively, it should be discussed whether for substances with bioaccumulation potential the assessment factor could be increased. 3.4.2
Assessment of Effects on Sediments
There are only few tests for sediment organisms available. The selection of both representative organisms and standardized sediments is being discussed. Various approaches are being developed to investigate the effects of chemicals on sediment and sediment organisms (OECD, 1992). As soon as standardized test methods are available and the assessment factors agreed upon, the calculated PNEC~ed can be compared with the estimated concentration in sediment (PECk) or with the measured concentration in sediment. In the absence of ecotoxicological effect data on sediment-dwelling organisms, the PNEC~ed may provisionally be calculated with the PNEC for aquatic organisms and the sediment/water partitioning coefficient (OECD, 1992). 3.4.3
Assessment of Effects on Terrestrial Compartments
If the following two criteria are fulfilled, an assessment of effects on the terrestrial compartment is to be conducted: 1. The substance is discharged into the soil in considerable quantities and remains in this compartment. 2. The substance has a toxic potential for aquatic organisms. As the information on terrestrial toxicity data is limited for most of the existing substances, tests on aquatic organisms could be used for an initial approach. Aquatic organisms might show a similar sensitivity to chemicals as soil organisms which are exposed mainly to the pore water of the soil. Further effects which chemicals, adsorbed to soil particles, execute on soil organisms by ingestion cannot be considered by this initial approach. This is one of the reasons why this approach can give only a first idea regarding terrestrial toxicity. If in this first step the PEC~oa-vo. . . . . t e r is larger than the PNECwater, tests with terrestrial organisms are indispensable for the assessment of terrestrial effects. Especially for highly lipophilic chemicals, other indications for a toxic potential for the terrestrial compartment beside the aquatic toxicity should be considered, e.g. high toxicity for mammalians (LDs0 oral --> 200 mg/kg BW or LDs0 dermal 400 mg/kg BW), or for birds (LDs0~c~ >-- 50 mg/kg BW). The effects of a substance on the compartment soil can be divided into two fields: 1. Effects on biotic parameters: The application of a chemical can cause a considerable change of the species pattern of soil organisms and a reduction of the variety of the species. Terrestrial tests with higher plants or earthworms have been conducted for only a few existing chemicals. The minimal assessment basis should be one test for fauna and one test for flora. 2. Effects on abiotic parameters: Chemicals can affect processes in the soil like filtration, buffering, metabolism and degradation (UBA 1993). For the characterization of abiotic effects the investigation of degradation has first prioriry. This information allows to characterize the change of the soil properties and to revise the estimation of PECso~.
ESPR-Environ. Sci. & Pollut. Res. 1 (2) 1994
Chemicals Regulation Assessment
In this stage of the initial risk assessment, adsorption, desorption, and mobility can be estimated by model calculations. Corresponding tests of percolation and adsorption may be necessary for the refined risk assessment. The same assessment factors as for the aquatic system (-* Table 1) should be used for the terrestrial system according to the type of investigations (acute or long-term toxicity test) and the number of trophic levels tested. If only acute toxicity data are available, a factor of 1,000 is applied to the lowest E(L)C50. If in addition to the acute tests one, two, or three long-term test(s) of different trophic levels are conducted, an assessment factor of 100, 50, or 10 has to be applied to the lowest NOEC to calculate the PNECso ~. The abiotic effects should be assessed qualitatively. A substance should be considered as hazardous to the soil if its degradation is smaller than its discharge into the soil or if it changes the soil properties in such a way that noxious effects result (see German Federal Government, 1985 and ISO, 1991). If according to the initial risk characterization PEC~o,j > PNEC~oil, a further testing program is to be initiated to improve the PNEC~o~. Depending on the effect of a chemical on vascular plants, earthworms, or microorganisms, the information about the ecotoxicological effect on the most sensitive group of organisms (primary producers, consumers, or destruents) has to be refined (further testing). If a third trophic level has not been tested yet, an organism of this trophic level has to be tested (acute investigations). For further testing, usually long-term tests (with e.g. earthworms, springtails, enchytraeids, and staphylinids) are preferred. 3.4.4
Assessment of Atmospheric Effects
An assessment of atmospheric effects should only be performed under the following conditions: half-life tl/2 > 10days or diffuse emission > 10 t / a or one point-source emission > 1 t/a. No internationally accepted test guidelines have been developed for the assessment of air effects. For most of the chemicals only data on mammalian toxicity are available. For some chemicals investigations on the toxicity of invertebrates, mostly insects (e.g. Apis mellifera, Syrphus corollae) might be available which have been conducted according to guidelines for the testing of plant protection agents. Only in some of these tests it is possible to determine an effect concentration. Therefore, a PNEC can usually not be derived. Special test guidelines for the effects in the atmosphere have to be developed. Consecutively, the development of an appropriate assessment factor is needed. For the evaluation of an atmospheric risk the following physico-chemical effects of a chemical on the atmosphere have to be considered as well: - degrading air quality (DAQ) - tropospheric ozone building (TOB) - acidification - depletion of the ozone layer - warming up of the atmosphere - greenhouse warming potential (GWP). The potential to cause adverse effects to the atmosphere by DAQ, TOB or GWP should be determined by an expert
121
Chemicals Regulation Assessment judgement for organic chemicals (gaseous, volatile liquids or solids), which show a vapour pressure >__ 100 Pa or a Henry's Constant > 0.01 P a x m3/mol. Organic chemicals which contain F a n d / o r CI or Br atoms and which are gaseous or volatile liquids with a vapour pressure _> 100 Pa should immediately be subject to a first qualitative assessment of the ozone depletion potential (ODP).
3.4.5
Assessment of Secondary Poisoning
Secondary poisoning is the result of the transfer of chemicals via food chain to different trophic levels (biomagnification) forming residues which may be detrimental to the last organism(s) in the chain. The food chain (water ~ fish -~ fish-eating bird or fisheating mammal) is one example of a secondary poisoning pathway. Safe levels for fish-eating animals do not exclude risks for birds or mammals feeding on other aquatic organisms (e.g. mussels, worms). An assessment of secondary poisoning has to be performed if the following criteria are fulfilled: First, it has to be decided whether or not indirect exposure of ecosystems (or man) is likely to occur. Second, it has to be considered whether there are indications for a bioaccumulation potential. Subsequently, it is necessary to consider whether the substance is classified on the basis of its mammalian toxicity. It is assumed that the available mammalian toxicity data can indicate the possible risks of the chemical to higher organisms in the environment. Effects on avian and mammalian populations should be representative for fish-eating birds or mammals (PNECor~). For example, results of the 28-day repeated dose test have to be used preferably for the assessment of secondary poisoning effects. In this case, PNECoral is calculated on the basis of the factor 100. Only toxicity studies on dietary and oral exposure are relevant, as the pathway for secondary poisoning refers exclusively to the uptake through the food chain. The results of these tests can be expressed as a concentration in the food (mg/kg) or a dose (mg/kg bodyweight/day) causing no effect. For the assessment of secondary poisoning, the results are converted to the concentration in food (mg/kg food). To extrapolate to predicted no-effect concentration from the test results in food that should be protective for other mammalian and avian species, an assessment factor is necessary. For the assessment of secondary poisoning the predicted concentration in food (provisional PECoral = PEC ..... x Pow x 0.05) has to be compared with the predicted no effect concentration for fish-eating birds or mammals (PNECor~I). A measured BCF can be used instead of the log Pow for the calculation of a final PECo~a (PECo~~ = PECwaterx BCF). If the provisional PECor ~ exceeds the PNECot~, secondary poisoning could be a critical pathway for fish-eaters and a bioaccumulation test (BCF-test OECD guideline 305) has to be conducted.
3.5
Risk Characterization and Further Testing
As soon as all available information are evaluated and a PEC as well as a PNEC have been derived, an initial risk charac-
122
Review Articles terization can be performed by comparing the PEC with the PNEC (-~ Fig. 2) for the relevant compartment. An initial risk characterization is also performed if only few investigations on a chemical are available. The uncertainty in the initial risk characterization due to the low quality and quantity of the data is to be compensated by sufficiently high assessment factors. If in this initial assessment PEC _< PNEC no further tests are required unless the production volume or the use pattern change considerably. If PEC > PNEC an environmental hazard is apparent. In this case more information on exposure and/or effects should be obtained (e.g. by repeating certain tests under valid conditions or performing long-term tests or tests with organisms of trophic levels not yet considered). In addition, detailed information on exposure helps to avoid worst-case assumptions for calculating a refined PEC. In many cases measurement of environmental concentrations should be initiated. With this additional information the new refined PEC and/or PNEC can differ from the initial values. Therefore, whenever new and relevant data are obtained, PEC and PNEC have to be compared again in a refined risk characterization. If PEC _< PNEC the assessment procedure is completed, i.e. no further information is necessary. This iterative process has precautionary aspects especially in the initial phase as data gaps are filled in by worst-case assumptions or high assessment factors. Morever, this approach saves resources as only a minimum of tests have to be performed. By the preliminary assessment possible hazards are identified at an early stage and a specific testing plan can be elaborated for a maximum of additional information at low expenses.
3.6
Risk Management
If after a refined risk characterization PEC exceeds PNEC and if an improvement of the data basis does not seem to be feasible or does not change the result of the assessment, measures for risk reduction have to be discussed. The EC Existing Chemicals Regulation demands cost/benefit analyses and investigations on possibilities for a substitution of the chemical. However, the most specific measure which efficiently reduces the risk for the concerned environmental compartment should always be proposed. About 60 existing chemicals have been assessed since 1992 in Germany. For 8 % of these chemicals the assessment revealed the need for measures to reduce emissions. According to the initial assessment for 26 % more data on exposure a n d / o r effects have to be collected or the results of measures already taken have to be observed. The majority of substances assessed did not show current concern.
4
Discussion
The task to protect the environment against the hazards of the more than 100,000 substances listed in the European Inventory of Existing Chemicals (EINECS) seems to be immense. However, the conception of the German Existing Chemicals Program and, hopefully, the EC Existing Chemi-
ESPR-Environ. Sci. & PoUut. Res. 1 (2) 1994
Review Articles
Chemicals Regulation Assessment
effect assessment determination of PNEC*
r
I
J
v initial risk
identification of representative monitoring data
exposureassessment / determination of PEC*
~
1,000 t/a) a systematic priority setting has been performed. These substances of highest priority are the first candidates for a comprehensive risk assessment including risk reduction measures. For this process all available information should be used. The German practice during the recent years showed that the data situation is very different for each chemical and therefore the assessment of these chemicals cannot be stereotyped. Expert judgement is important in each case. An indication that relevant chemicals are covered by this procedure comes from recent accidents in German chemical plants. In most cases, critical data had already been collected by the German Existing Chemicals Program, so that an assessment could be performed or had already been available. These incidents confirm that the ongoing process will contribute to decrease the hazards for the environment, especially as soon as all EU coun-
ESPR-Environ. Sci. & Pollut. Res. 1 (2) 1994
tries participate and as soon as a close cooperation with other industrial countries within the O E C D can be achieved. 5
Literature
AHLERS,J.; KOCH, W.; LANCE, A.; MARSCHNER,A.; WELTER, G. (1992): Bewertung der Umweltgef~ihrlichkeityon Mten Stoffen nach dem Chemikaliengesetz (ChemG). ChemikaliengesetzHeft 10, Texte 19/92, Umweltbundesamt, Berlin, M~irz 1992 AHLERS,J.; KOCH,W.; MARSCHNER,A.; WELTER,G. (1993): The Sci. Total Environm. Supplement 1993,1587-1596 Deutscher Bundestag (1985): Bodenschutzkonzeption der Bundesregierung BT-Drs. 10/2977 vom 7. Miirz 1985 ISO (International Organisation for Standardization) (1991): Soil Quality, Vocabulary, Part 1: Terms and Definitions Relating to Soil Protection and Soil Pollution, ISO/CD 11071 OECD (1992): Report of the Workshop'on Effects Assessment in Sediment. Copenhagen, Denmark, 13-15 May 1991 Technical Guidance Documents in Support of the Risk Assessment Directive (93/67/EEC)for Substances Notified in Accordance with the Requirements of Council Directive 67/548/EEC.Brussels, 1993 UBA (Umweltbundesamt) (1993): Entwurf zur Bewertung yon Bodenbelastungen, Fachgebiet I 3.7
123
