the role of laboratory terrestrial model ecosystems in the testing of potentially harmful substances

Ecotoxicology 3, 213-233 (1994)

MINI REVIEW

The role of laboratory terrestrial model ecosystems in the testing of potentially harmful substances

E D W A R D M O R G A N *a and T H O M A S K N A C K E R z

1Environmental Research Unit, Dept of Science and Chemical Engineering, The University of Glamorgan, Pontypridd, Mid-Glamorgan, CF371DL, South Wales, UK eBattelle Institute e. V., Am ROmerhof 35, D-60486 Frankfurt a. M., Germany

Received 2 August 1993; accepted 30 September 1993; revised 21 October 1993

A classification of terrestrial model ecosystems (TMEs) was introduced which is based upon the physical properties of intactness of the physical medium and openness to the atmosphere. This gave rise to four types of system, namely open and closed intact systems and open and closed homogeneous' systems. These systems have different capabilities with respect to fate and effect end-points with various substances. The large closed TMEs are generally very complex, require a high degree of operator skill, expensive and therefore not replicable. Whilst these can provide estimates of losses due to volatility, they are not useful for determining effect end-points because of low replicability; high replicability being necessary because of natural variation in organism response. Open systems, especially those having intact soil-cores, are usually smaller, less complex and therefore more replicable. These have provided useful information on integrative functional effect end-points, but can only produce mass balances with non-volatile substances. Homogeniza- tion of the medium has also helped elucidate ecotoxicological effects by increasing replicability, but may introduce artifacts because of the disruption to soil organisms.

A major limitation of TME studies would seem to be that few effect end-points can be non- destructively sampled. Further investigations into these may provide information on recovery of terrestrial ecosystems over time after substance application, perhaps using multivariate statistical techniques. Other problems concerning TMEs are related to complexity and scale. In this respect ecosystem functions in which microorganisms play a major role, such as nutrient cycling, provide the greatest similarity when compared to field evaluations of the same substances, especially where the TME is intact. However, effects upon structural aspects of biological communities have in general not been well researched in TMEs. Once these have been added to the more complete set of functional end-points, TMEs will provide a very useful tool in hazard assessments of potentially harmful substances.

Keywords: terrestrial model ecosystems; harmful substances; fate; ecosystem-level effects

Introduction

When a new substance is produced or its manufacturing levels increased, hazard assessments using physico-chemical and environmental data, and estimates of toxicity

*To whom correspondence should be addressed.

0963-9292 © 1994 Chapman & Hall

214 Morgan and Knacker

from single species tests are performed to screen out substances for which no environmental danger is indicated. However. it has been gradually realised that potentially harmful substances could also exert effects at levels of ecosystem structure and function that could not be accounted for in single species tests (Cairns et al., 1972). These concerns prompted the development of laboratory-scale physical terrestrial model ecosystems (TMEs). Odum defined these aspects of ecosystem structure as including the whole of the biological community, the quantity and distribution of abiotic materials, and the range and gradient of conditions of existence (Odum, 1962). Odum also stated that functions of ecosystems are the rate of energy flow through the system, biological or ecological regulation, and the material and nutrient cycling rates (Odum, 1962). Each of these properties involves a substantial degree of interaction between many ecosystem components that should be assessed when considering the potential environmental impact of a substance.

Whilst ecosystem-level effects may be studied in the field, terrestrial model ecosystem studies have an advantage of safety since the substance may be screened without unnecessarily risking the environment. They can also indicate a correspondence between the location of the substance or its break-down products and particular ecotoxicological effects. They have in fact provided the first quantitative view of the fate and effects of certain pesticides and industrial chemicals (Nash et al., 1977; Gile et al., 1982), and waste products (Malanchuk et al., 1980; Van Voris et al., 1982) in the terrestrial environment.

When TMEs were first developed there was a great deal of interest in developing a generic system that could be adopted as an alternative to field studies. Several of the systems contained both terrestrial and aquatic phases. These multicompartmental systems provided a valuable initial insight into the bioaccumulation and biodegradability of many radio-labelled agrochemicals (Metcalf et al., 1971; Metcalf et al., 1973; Lichten- stein et al., 1978; Lu et al., 1978). Purely terrestrial systems followed later (Lichtenstein et al., 1974; Beall et al., 1976; Cole et al., 1976; Draggan, 1976; Gillett and Gile, 1976; Ausmus et al., 1979). However, limits on the extrapolability of the data from many of these systems resulted in rejection of the generic system idea but did enhance the design and value of ad hoc or site-specific systems (Gillett, 1989). Thus, a number of studies have been published using a wide diversity of TMEs. It is the intention to review these and classify them according to physical characteristics. The substances used in these studies will be cross-referenced to the various fate and effect end-points determined. Other topics to be reviewed will include aspects of the experimental design and analysis of TME studies and the limitations of the various systems. This should help to highlight the capabilities and inadequacies of the different systems, thereby pointing the way forward for TMEs in the future.

Definitions

Before attempting a review of TME use a working definition is required. Many authors have used the term 'microcosm' to describe their TME. The term TME is preferred here because of the frequent misuse of the term 'microcosm' to describe almost any type of model system that includes terrestrial components, without reference to location, physical characteristics or capabilities. A TME is therefore a controlled, reproducible system that attempts to simulate the processes and interactions of components in a portion of the terrestrial environment (Gillett and Witt, 1980). It has a boundary, although it may be an open or semi-enclosed system, and is subject to investigator

Terrestrial mode l ecosystems 215

control of environmental factors such as temperature, light, humidity, and fluxes of air and water. Obviously, for this degree of control, the system could be operated in a greenhouse or in an environmental chamber.

To fulfil the role of a model ecosystem the system should have at least some ecosystem- level characteristics of structure and function, e.g. more than one trophic level (Giesy and Odum, 1980). Thus a simple pot system for studying plant uptake of a chemical from a synthetic medium or sterile soil is not a TME, whereas a similar system with a homogeneous soil may be classified as such if a natural community of microorganisms is allowed to develop.

Scope of the review

The systems and studies considered in the state-of-the-art review are those which have been published since 1979. Some literature was already available in-house. To supple- ment this, a search of chemical and pollution abstracts was made for studies in which the terms 'terrestrial', 'soil' or 'soil-litter', 'model ecosystem' or 'microcosm' were found in the title, abstracts or key-words. Studies involving other named systems were also examined and included if the systems fitted the above definition of a TME. However, systems designated by the study authors as microcosms but which were field-based have been excluded (e.g. Seastedt et al., 1983; Kelly et al., 1984; Donnelly et al., 1990; Greville and Morgan, 1991; Mieth et al., 1993) since the degree of control over environmental conditions is lower than for laboratory-based systems. Also, in moving to the field, system size often comes more into line with that normally associated with lysimeters, extensively used for fate studies (Ftihr and Hance, 1992) or mesocosms (Gillett, 1989). In addition, systems having a separate aquatic phase have also been excluded since ecosystem structure and function will be completely different.

A classification of terrestrial model ecosystems

Combinations of physical properties are used here as the main basis of the classification scheme as these will have a great influence upon model ecosystem capabilities. The first property considered is the integrity of the terrestrial medium. The soil or litter may either be extracted intact, e.g. as a core, so disrupting the soil micro-organisms and meso-fauna as little as possible, or extracted and homogenized to reduce natural variability. A second property that has a great influence upon TME capabilities is whether it is a sealed (closed) or an open system. Closed systems have air pumped through at a controlled rate and can often be used to trap radio-labelled volatile parent compounds and their break-down products, providing the investigators with the potential for complete mass balance of volatile substances. In contrast open systems, whilst being technically less complex and therefore less expensive, cannot be used with volatile substances because of possible cross-contamination. These two properties give rise to four types of system. For each type the data obtained from the studies are presented according to system characteristics (by increasing system size) in Tables 1-4.

INTACT TERRESTRIAL MODEL ECOSYSTEMS

Intact terrestrial model ecosystems (Tables 1 and 2) came to prominence in the 1980s and 1990s whereas they were hardly known previously. The few examples that did exist had no typical set of characteristics. They were either small and replicable (Gile et al.,

216 o

e ~

O

.=.

.=_ 0

Q

8 0

o

w~

,;.2 ~. .

: ~ o . ~ ~ "~

= ,3 .=- .~

• - ~ ~

= o ~ R o

{ ~ & ~ o- z-r, & ~ ~.~

i i ~ ~ ~ ~ o o ~

o o ~

~5 ,4 ,-4 ,..; ~ ,5 ~

0 0

"a .P-

E- ;>

o . ' . a

g_ ~ _~.=.

ID

Z

~ ' ~

=~ .= .~ .~.

Z

i;

..,,

*g,

'~,

O

Z a. ~

..=

"= ==

.=. .._. "a ~ ' a

* a , g ,

o

. " g "

,*=, ~ ~ ,..-= , o

o ~ID ¢xl

"4

O ¢/)

:O

"d

o

,.d

o

ila

"8

o a.

E o

2 1 7

218

o

©

8

" 0

C/)

,.4

.o

r~

o

o e~

0

0

" 0

0

" 7

0 o

• . ~ ~ . ~ " ~

z ~

~ ~ ' ~

Z Z = . ~

I

Z ~+

c~

0 0 o 0

.o ,-

o

U

,.-2~ , - ~ b ~

~ o o o S

0

-d

'2 0

~ ' ~ " ~ ; ~

e~

o o

. , ~ : ~ , . ~ ~ . .

< ~

. o ~:

-~ ~ . -

b ~

, " ~ o

~ ~ . ~

~.~.~ ~ ~

m

L~

:..: .."

< ~ < <

~ m . . . . . . ; . : o

~.~. .~ .~.= ~ .~.~.

*~ o ~ .o_-~

r,.)

- . ~ ~-~ 6 ' ~

o0

z

:0 t- .

¢..)

z

-6

,.g:

:1

219


1979) or relatively large with few replicates (Ausmus et al., 1978; Jackson et al., 1978a, b). In contrast, modern intact TMEs are often intermediate in size and capable of a reasonable degree of replication under controlled conditions.

Open intact TMEs

Many of these systems are roughly based upon the 'soil-core microcosm' proposed by Draggan and others (Draggan, 1976; Ausmus et al., 1979; Gile et al., 1979), and developed by Van Voris and co-workers (Van Voris et al., 1982; Van Voris et al., 1984). Some of these systems included natural vegetation (e.g. Van Voris et al., 1980), whereas another has been used in several studies with a variety of crop plants (Van Voris et al., 1984; Tolle et al., 1990). In reality this system used with crops does not have a perfectly intact soil core as the first 15-20 cm are removed, homogenized, replaced and planted. This is justified by the necessity to mimic normal agricultural practice. Also, model ecosystems of the same genre have been used to examine effects on a natural grass ecosystem (Arrhenatheretum elatoris association, Knacker et al., 1989a, b; R6mbke et al., 1993).

Using this system as its basis a protocol was published in the US (Environmental Protection Agency, 1987; American Society for Testing and Materials, 1988). In contrast, neither the European Community nor OECD has such a protocol. This protocol, whilst setting out methods for examining chemical fate (applicable only to non-volatile substances), is limited for effect end-points to describing methods for assessing effect upon plant productivity and nutrient cycling (Environmental Protection Agency, 1987; American Society for Testing and Materials, 1988). Whilst there are some publications more or less based upon this system (Knacker et al., 1989a, b; Tolle et al., 1990; R6mbke et al., 1993), it cannot be said to have profited from its adoption as a regulatory tool.

A wide variety of compounds have been examined in open intact TMEs and include inorganic elements and compounds, synthetic agrochemical compounds, inorganic and organic wastes (Table 1). The studies with the smaller open intact TMEs (Table 1) have tended to concentrate on integrated functional effect end-points in which soil micro- organisms play major roles, such as nutrient loss in the leachate (e.g. Cronan, 1980; Hinchman and Zellmer, 1986; Van Voris et al., 1980) or respiration as measured by the Anderson-Domsch substrate-induction method (Knacker et al., 1989a, b). These respiration measurements have also been converted to estimates of microbial biomass which might be indicative of effects upon ecosystem structure. Surprisingly few other attempts have been made to examine effects upon microbial and meso-fauna community structure (Van Voris et al., 1980; R6mbke et al., 1993) even though effects upon individual species have sometimes been assessed (Van Voris et al., 1982; Wright and Coleman, 1988). Whilst it is unlikely that large numbers of samples for faunal counts may be taken from TMEs, thereby limiting the indices that can be applied, some indices such as Renkonen coefficients or plots of dominance-rank curves are applicable to low sample numbers and could be used.

Open intact TMEs are limited in their usefulness for studying the fate of volatile compounds since mass balance determinations will not be possible. Nevertheless other fate end-points for non-volatiles have been successfully assessed, e.g. bioaccumulation by plants (Van Voris et al., 1984; Knacker et al., 1989a) and movement of the substance through the soil column by monitoring leachate losses (Hinchman and ZeUmer, 1986).

Terrestrial model ecosystems 221

Closed intact TMEs

Substances examined in closed intact TME studies (Table 2) have included some radio- labelled agrochemical compounds (e.g. Nash and Beall, 1980a, b; Winkelmann and Klaine, 1991), but were mainly the 'heavy' metal elements (e.g. Shirazi et al., 1984) and inorganic compounds (e.g. Billings et al., 1984).

The closed intact TMEs (Table 2) whilst being slightly fewer in number than the open intact systems are nevertheless diverse in character, varying in size from 0.125-1125 dm 3. Some are small to medium size soil-cores with limited air space above the soil (Piwoni et al., 1986; Winkelmann and Klaine, 1991). Others are larger and more box-like (Branham et al., 1985; Nash and Beall, 1980a, b), giving greater air-space above the soil and plants, and allowing larger number of air changes per hour, so that realistic measurements of volatility might be made. As would be expected most studies with these large closed systems have concentrated on fate determinations, usually with radio- labelled agrochemicals, although one study did examine a large range of non-labelled organic chemicals in an intact soil-core type system (Table 2, Piwoni et al., 1986). However, four of the systems were instead used for monitoring COz evolution (e.g. Shirazi et al., 1984; Larkin and Kelly, 1988), providing the investigators with an integrative non-destructive functional end-point.

HOMOGENEOUS TERRESTRIAL MODEL ECOSYSTEMS

Homogeneous terrestrial model ecosystems are a large group of systems, varying in size from 0.05 dm 3 to 663 dm 3, in the type of media used, and in the range of end-points assessed (Tables 3 and 4). The smallest open and closed systems, i.e. those having volumes of less than 0.5 dm 3, have been used mainly for effect end-point determinations involving micro-organisms and meso-fauna (Tables 3 and 4). Many of these studies are highly replicated experiments, often using over 100 individual TME units to examine the effects of one compound (e.g. Parker et al., 1985; Van Wensem and Adema, 1991; Van Wensem et al., 1991). This level of replication has been used to overcome the high natural variability in organism response that is often found with effect end-points, even with completely homogenized systems. At the same time, fate end-points are not usually assessed with these small systems. One reason for this might be the difficulty in detecting amounts of added substances at ecologically relevant concentrations without completely destroying the system. This lower size limit is therefore likely to remain unless there are substantial improvements in analytical detection limits.

Open homogeneous terrestrial model ecosystems

Many of these systems, whilst fitting the criteria for a model ecosystem are often little more. Some systems, e.g. that of Nagpal (1986), have low degrees of control over environmental factors and are only included on the justification that the soil column will have its own natural microflora. The substances examined in open homogeneous TMEs (Table 3) have mainly been synthetic organic compounds, waste products or inorganic nutrients.

Ecosystem-level end-points typically assessed in these systems (Table 3) are nutrient loss in leachate (O'Connor et al., 1980), microbial population (e.g. Parker et al., 1985), microbial and meso-fauna biomass (Parker et al., 1985; Knacker et al., 1993) and litter

0

0

0

O

O

°~,,~

o

e~

©

o

0

e~ 0

N

0

b ~

o

o o =

2 ~J

6 o O

O ~5

~oo =

2

c~

O

~U

o

, o

r~

O

o

0J

o3

"O

~J

O

Z

O b0

~J

r~

O

"O

"O

O

o~

o ~J

v

o i=.

~J

__=

O

©

O

"O

_= O


loss (Parker et al., 1985; Knacker et al., 1993). Litter loss is a particularly interesting end-point not assessed in intact systems. It may be considered to be an integrative functional end-point, at least for forest and grassland ecosystems, since it could reflect an overall impact upon the micro-organism community. It could also be used in a non- destructive manner as the litter bags may be replaced periodically. Several studies (Table 3) did assess acute effects upon particular species such as earthworms and wood lice (e.g. Parker et al., 1985; Knacker et al., 1993), but here again these were not converted to indices which might reflect effects upon community structure.

Closed homogeneous T M E s

These formed the largest single group of systems and were used for a diverse group of compounds (Table 4). In addition to the small systems already discussed in general earlier, many were very large systems (Table 4). These appear to be fairly similar to their box-like intact counterparts (Table 2) as far as the actual mechanism of trapping the volatile components are concerned. Some are also gnotobiotic, including assembla- ges of organisms representing various parts of trophic chains, e.g. voles and crickets (Gillett and Gile, 1983), and centipedes and field slugs (Schuphan, 1986; Schuphan et al., ~1987). These systems have mainly been used to examine the fate of radio-labelled agrochemicals with few or no replicates (e.g. Gile and Gillett, 1981; Gile et al., 1982; Schuphan, 1986), undoubtedly a function of the cost and time. This will limit their potential usefulness for assessing effect end-points because of high natural variability in organism response.

Typical effect end-points assessed in the small closed homogeneous TMEs include respiration and population numbers (e.g. Fairbanks et al., 1984; Blair et al., 1989). Several studies also used rate of litter loss as an integrative effect end-point (Van Wensem and Adema, 1991; Van Wensem et al., 1991).

The design and analysis of terrestrial model ecosystem studies

In TME studies there are severe limitations imposed on the investigators by the lack of ecosystem-level effect end-points that may be easily monitored non-destructively. At present, the only such parameters that are readily amenable to this are the measurements of nutrient concentrations in the leachate and respiration as determined by CO2 evolution. Therefore, replicates have to be sacrificed to measure other end-points. This problem is exacerbated in larger systems that are difficult to replicate. Thus, studies with these large systems are often not rigorously designed and analyzed statistically. In contrast, studies with the smaller systems are often highly replicated and the results thoroughly analyzed; analysis of variance (ANOVA) being the usual statistical tool for establishing significant effects (e.g. DeCantanzaro and Hutchinson, 1985; Dougherty and Lanza, 1989), often followed by multiple comparison tests (e.g. Van Wensem and Adema, 1991). When it has proved difficult to maintain the same conditions for all the replicates, either blocked designs (e.g. Malanchuk et al., 1980; Fairbanks et al., 1984; Larkin and Kelly, 1988) or latin square designs (e.g. Van Voris et al., 1984) have been used.

Often ecotoxicologists only vary the concentration of the added substance in their studies. However, other ecological parameters are also likely to be important and may interact with observed concentration effects. Surprisingly, therefore, only one study used

2 2 4

0

c

0

o

o o ~2

c

c

o

.= r ~

A~

o

.o

C

o=

o=

¢.

©

0J

0

.o ~.,,q o E , o "~ c

t.q

+ . "0 ~

Z

z ~

.o

..= .~ ..= .~

o

.o .o .~

~ <

o '~ o =

o

. . . . . .

._.9. E

Y:

< <

225

Z

0

. o .o

.o

- i

E

2 <

t"q

<

e ~

r.)

0

= ~

2 r~ .o

(.)

¢.)

o

O 0 0

o

o r.~

.=.

O " '

226

0

0

U,.I

0

"0

©

o

~"-, o " 0

<

"7

z

g ~

~ ' : . . ~

0 - ~

. o

o

"o "

.~" ~ .~ -~ . . .~ .,- .~

o

0 ~ ~-

.o

o

.o

..° o ~ =- G ~ .~-~

c .o ~" o .u ;.: o :,.= , . ._. ~ ~ ~ ~ ~ .~ ~ ~ f ~

0 = ,-g-,

r.~ r.~ ~/0~ 0 c~

0,,)

¢) "E

E

E

0.,) E

0 ;>

:n ~r'

0


a multivariate design, a factorial with 159 homogeneous TME units to examine effects of two substances upon the rate of litter loss (Parker et al., 1985). Other multivariate designs, requiring fewer units, such as fractional factorials and central composite designs (Box et al., 1978), could also prove useful in TME studies.

In general, ANOVA seems to be the accepted statistical technique in TME studies, although some authors have used regression techniques (e.g. Hickman and Novak, 1987). ANOVA can be a reasonable technique for analyzing results at the end of a TME experiment if potential problems with normality of distribution and homogeneity of variances are dealt with. There are also limitations of graphically representing the results. However, TME experiments often run for weeks or months. If some parameters are then assessed over time, this may introduce additional problems, including the increasing likelihood of introducing a Type II error (accepting a false null-hypothesis) and temporal dependence of the variables.

In the last 10 years, the advent of the personal computer has led to the increased development and dissemination of multivariate statistical techniques such as principal component analysis, correspondence analysis and cluster analysis (Massart et al., 1988). These techniques are capable of graphically displaying information from many dimen- sions in simple projections, indicating the important variables at the same time. Some work on aquatic ecosystems, for example, has used non-metric clustering with a large number of measured ecosystem parameters to demonstrate ecosystem recovery over time, determining the significance of the different variables at the various points in time (Matthews et al., 1991). Whilst no comparable work with TMEs has been published, such multivariate techniques could have potential applications. This would, however, be dependent on being able to determine end-points non-destructively.

The limits to terrestrial model ecosystem technology

TME studies have been criticized for being too short-lived to demonstratc effects upon ecologically significant processes such as natural succession or other multi-species multi- generation phenomena (Gillett and Witt, 1980). It must be debatable, howcver, whether a long-term TME over several seasons would represent an efficient use of time and resources. Also, whilst this criticism may be relevant for larger soil animals and plants, TMEs can still be used in studies of processes for which questions of duration are insignificant, e.g. when micro-organisms are involved. These species will reproduce many times during the course of most TME studies which normally last between 60 and 120 days (Gillett, 1989).

Critics of model ecosystems in hazard assessments have also argued that they should fully predict or mimic the field situation in all respects. In reality this is not possible for any system. Therefore, it has been said, they are of little value (Goodman, 1982). Nevertheless, if thcy can provide answers to specific questions of fate or biological effects at the ecosystem level their use may be justified (Frcdcrickson et al., 1991). Clearly, ecosystem-level effects require an ecosystem or at least a model ecosystem for their determination. A more appropriate consideration is that the information obtained should realistically represent what would happen in the natural ecosystem. Of course, natural components when brought into the laboratory environment may not behave in an entirely natural manner (Pritchard, 1982). Therefore, to ensure that the data are not simply anecdotal, structural and functional aspects of particular TMEs should be qualitatively validated and quantitatively calibrated against equivalent field systems.


Of course no model ecosystem can include all facets of structure and function of the ecosystem it represents. All are limited in size to an extent where only certain components will be selected. Therefore, biological complexity is reduced in model ecosystems (Gillett and Witt, 1980), limiting what may be achieved in a particular study. For example, effects upon a plant's full life-cycle cannot be studied if a pollinator is not included, e.g. honey bee, but studies of the rhizospheric interactions of chemicals, soil, soil micro-biota and plants are still possible (Gillett, 1989).

Organism distribution and the properties of physical media are inevitably hetero- geneous in the natural environment. A reduction in scale from the field to a TME will therefore tend to increase variability amongst the model ecosystem components (Freder- ickson et al., 1991). This will in turn cause a reduction in biological diversity. Natural ecosystems tend to have a lot of functional redundancy and may not be drastically upset by displacement or loss of an individual species, e.g. when an intact soil-core is extracted (Frederickson et al., 1991). In contrast, agricultural systems are usually less diverse and may lose components that play key roles in ecosystem function. Hence a relatively greater emphasis should be placed on ensuring relative conquerability between agricultural type model ecosystems and the field.

Potential problems could result from the uncoupling of ecosystem processes linked to a reduction in complexity (Frederickson et al., 1991). Whilst networks of organisms involving only micro-organisms that act upon processes such as nutrient cycling might be relatively complete in TMEs, networks involving larger meso-fauna might either be absent or incomplete. This could result in differences in the rates at which some processes proceed, such as chemical transformation of the added substance. The half- life of the herbicide atrazine in the field, for example, was 14 days as compared to 21 days in an intact TME (Winkelmann and Klaine, 1991). Apparent half-lives are also shorter for chemicals in TMEs than in simple laboratory tests (Gillett and Gile, 1976). This trend might be due to the presence of additional pathways not accounted for in laboratory studies, and/or similar possibilities typical of the 'attributive properties' of ecosystems (Odum, 1971). Therefore, the relative completeness of each process should be considered before including it as an end-point.

Whether or not the data obtained in a TME study can be extrapolated to the field situation will also depend upon the scale of the process that is being observed (Wolfe et al., 1982; Shirazi et al., 1984). The heterogeneity of the soil, its structure, the plant-soil relationship and relationships of soil invertebrates in the plant rhizosphere in many TMEs are thought to be on an explicit 1:1 scale with the real world (Gillett, 1989). Many of the important processes in the soil are contained within these relationships, e.g. those involving soil-water, micro-organisms and organic matter. Therefore, TME estimates of effects upon processes such as nutrient cycling and rates of chemical transformation may not require scaling. Nevertheless, out-of-scale phenomena have also been noted in several TME studies. For example, in a large gnotobiotic system, a vole totally denuded a lush crop of ryegrass so that none of the data obtained were comparable to the field (Gillett and Gile, 1976). Crop plant species which on maturity have insufficient room for root growth could have similar effects. Inclusion in model ecosystems of trophic levels above small invertebrate animals and small rooted plants is therefore not feasible because containerization forces a restrictive condition that has no correlation with the real world.

Some ecosystem functions have been suggested as means of practically scaling TMEs


with the field e.g. primary productivity and nutrient cycling (Pritchard and Bourquin, 1984). Most importantly their rates and magnitudes will be integrative of the terrestrial community. Shirazi et al. (1984), for example, developed a method of scaling the biological response of soils amended with heavy metals using CO2 evolution to classify the impacts of these metals upon diverse soil types. Primary productivity of field plots has also been reasonably well predicted in the fly-ash studies with crops in intact TMEs after three years (Tolle et al., 1983; Van Voris et al., 1984). Structural attributes have seldom been used for scaling, probably because they are more difficult and time consuming to measure. Only one study was found in which comparisons of effects of two pesticides upon field and intact TME populations and biomasses of earthworms and enchytraeids were made (Rrmbke et al., 1993). Until more studies validating appropriate structural parameters are carried out, researchers may not be sure that the data are not artifacts of the study conditions.

In conclusion it would seem that intact TMEs with natural communities provide the most comparative information as far as functional end-points are concerned but there remains a definite requirement for more research on appropriate structural end-points.

Acknowledgements

This work was supported by the Umweltbundesamt (UBA), Berlin, Germany.

References

American Society for Testing and Materials (1987) Standard guide for conducting a terrestrial soil- core microcosm test, standard El197-87. ASTM 1104, 743-55.

Arthur, M.F., Zwick, T.C., Tolle, D.A. and Van Voris, P. (1984) Effects of fly-ash on microbial CO2 evolution from an agricultural soil. Water, Air and Soil Poll. 22, 209-16.

Ausmus, B.S., Dodson, G.J. and Jackson, D.R. (1978) Behaviour of heavy metals in forest microcosms; III. Effects on litter-soil carbon metabolism. Water, Air and Soil Poll. 10, 19-26.

Ausmus, B.S., Jackson, D.R. and Van Voris, P. (1979). The accuracy of screening techniques. In J.M. Witt and J.W. Gillett (eds) Terrestrial Microcosms and Environmental Chemistry, Proceedings of a Colloquia at Oregon State University, National Science Foundation, Corvallis, pp. 123-30.

Beall, M.L. Jr., Nash, R.G. and Kearney, P.C. (1976) Agroecosystem - a laboratory model ecosystem to simulate agricultural field conditions for monitoring pesticides. Proceedings of EPA Conference on Environmental Modelling and Simulation, Cincinnati, OH, pp. 790-3.

Bengtsson, G. and Annadoter, H. (1989) Nitrate reduction in a groundwater microcosm determined by gas chromatography-mass spectrometry. Appl. Env. Micro. 55, 2861-70.

Biggar, J.W., Nielson, D.R. and Tillotson, W.R. (1984) Movement of DBCP in laboratory soil columns and field soils to groundwater. Environ. Geol. 5, 127-31.

Billings, W.D., Peterson, K.M., Luken, J.O. and Mortensen, D.A. (1984) Interaction of increasing atmospheric carbon dioxide and soil nitrogen on the carbon balance of tundra microcosms. Oecologia 65, 26-9.

Blair, J.M., Crossley, D.A. Jr. and Rider, S. (1989) Effects of naphthalene on microbial activities and nitrogen pools in soil-litter microcosms. Soil Biol. Biochem. 21,507-10.

Box, G.E.P., Hunter, W.G. and Hunter, J.S. (1978) Statistics For Experimenters. An Introduction to Design, Data Analysis and Model Building. New York: Wileys.

Branham, B.E., Wehner, D.J., Torello, W.A. and Turgeon, J. (1985) A microecosystem for fertilizer and pesticide fate research. Agronomy J. 77, 176-9.


Buldgen, P. and Remacle, J. (1981) Influence on environmental factors upon the leaching of cations from undisturbed microcosms of beech and spruce litters. Soil Biol. Biochem. 13, 143-7.

Cairns, J., Lanza, G.R. and Parker, B.C. (1972) Pollution related structural and functional changes in aquatic communities with emphasis on freshwater algae and protozoa. Proceedings Academy National Science, Philadelphia, 124, 79-127.

Cole, L.K., Sanborn, J.R. and Metcalf, R.L. (1976) Inhibition of corn growth by aldrin and the insecticide's fate in the soil, air, crop, and wildlife system of a terrestrial model ecosystem. Environ. Entomol. 5, 583-9.

Cronan, C.S. (1980) Control of leaching from coniferous forest floor microcosms. Plant and Soil 56, 301-22.

DeCatanzaro, J.B. and Hutchinson, T.C. (1985) Leaching and distribution of nitrogen and nickel in nickel-perturbed jack pine microcosms. Water, Air and Soil Poll. 26, 281-92.

Donnelly, P.K., Entry, J.A., Crawford, D.L. and Cromack, K. Jr. (1990) Cellulose and lignin degradation in forest soils: response to moisture, temperature and acidity. Microbial Ecol. 20, 289-95.

Dougherty, J.M. and Lanza, G.R. (1989) Anaerobic subsurface soil microcosms: methods to monitor effects of organic pollutants on indigenous microbial activity. Toxicity Assessment 4, 85-104.

Draggan, S. (1976) The microcosm as a tool for estimation of environmental transport of toxic materials. Int. J. Environ. Studies 10, 65-70.

Environmental Protection Agency (1987). Soil core microcosm test 797.3995. Federal Register (USA) 52(187), 36363-71.

Fairbanks, B.C., Woods, L.E., Bryant, R.J., Elliott, E.T., Cole, C.V. and Coleman, D.C. (1984) Limitations of ATP estimates of microbial biomass. Soil Biol. Biochem. 16, 549-58.

Fermanich, K.J., Daniel, T.C. and Lowery, B. (1991) Microlysimeter soil columns for evaluating pesticide movement through the root zone. J. Environ. Qual. 20, 189-95.

Figge, K., Klahn, J. and Koch, J. (1983) Testing of chemicals by evaluation of their distribution and degradation patterns in an environmental standard system. Regulatory Toxicol. and Pharmacol. 3, 199-215.

Figge, K. and Sch6berl, P. (1989) LAS and the application of sewage sludge in agriculture. Tenside Surfactants Detergents 26, 122-8.

Fredrickson, J.K., Van Voris, P., Bentjen, S.A. and Bolton, H. Jr. (1991) Terrestrial microcosms for evaluating the environmental fate and risks associated with the release of chemicals or genetically engineered microorganisms to the environment. Toxic Subst. J. 11, 65-110.

Fiihr, F. and Hance, R.J. (eds) (1992) Lysimeter Studies of the Fate of Pesticides in the Soil. British Crop Protection Council.

Giesy, J.P. Jr. and Odum, E.P. (1980) Introductory comments. In J.P. Giesy, Jr (ed.) Microcosms in Ecological Research, pp. 1-13. US Dept. of Energy.

Gile, J.D. (1983) 2,4-D - its distribution and effects in a ryegrass ecosystem. J. Environ. Qual. 12, 406-12.

Gile, J.D., Collins, J.C. and Gillett, J.W. (1979) The Soil Core Microcosm - a Potential Screening Tool. EPA-600/3-79-089. Environmental Protection Agency.

Gile, J.D., Collins, J.C. and Gillett, J.W. (1982) Fate and impact of wood preservatives in a terrestrial microcosm. J. Agric. Food Chem. 30, 295-301.

Gile, J.D. and Gillett, J.W. (1981) Transport and fate of organophosphate insecticides in a terrestrial laboratory ecosystem. J. Agric. Food Chem. 29, 616-21.

Gillett, J.W. (1989) Terrestrial microcosms and mesocosms in ecotoxicologic research. In S.A. Levin (ed.) Ecotoxicology: Problems and Approaches, pp. 280-313, Berlin: Springer-Verlag.

Gillett, J.W. and Gile, J.D. (1976) Pesticide fate in terrestrial ecosystems. Int. J. Environ. Studies. 10, 15-22.


Gillett, J.W. and Gile, J.D. (1983) Predator-prey (vole-cricket) interactions: the effects of wood preservatives. Environ. Toxicol. Chem. 2, 83-93.

Gillett, J.W. and Witt, J.M. (1980) Chemical evaluation: projected application of terrestrial microcosm technology. In J.P. Giesy (ed.) Microcosms in Ecological Research, pp. 1008-33. US Dept. of Energy.

Goodman, E.D. (1982) The Limits of Microcosms: Problems in the Interpretation of Toxicity Results from Laboratory Multispecies Systems, ERC-13. Ecosystems Research Center, Cornell Univ. Ithaca, NY.

Greville, R.W. and Morgan, A.J. (1991) A comparison of (lead, cadmium and zinc) accumulation in terrestrial slugs maintained in microcosms: evidence for metal tolerance. Environ. Poll. 74, 115-27.

Hamilton, W.E. and Dindal, D.L. (1989) Influence of earthworms and leaf litter on edaphic variables in sewage-sludge-treated soil microcosms. Biol. Fert. Soils 7, 129-33.

Hamilton, W.E., Dindal, D.L., Parkinson, C.M. and Mitchell, M.J. (1988) Interaction of earthworm species in sewage-sludge-amended soil microcosms: Lumbricus terrestris and Eisenia foetida. J. Applied Ecol. 25, 847-52.

Hickman, G.T. and Novak, J,T. (1987) Microcosm assessment of biodegradation rates of organic compounds in soils. Toxicity of Hazardous Wastes Proceedings 19th Conf., 153-62.

Hinchman, R.R. and Zellmer, S.D. (1986) Determination of the Environmental Fate of Decontami- nation Agent C-8 Using Soil Microcosms, EES-TM-329. Argonne National Laboratory.

Jackson, D.R., Selvidge, W.J. and Ausmus, B.S. (1978a) Behaviour of heavy metals in forest microcosms: I..Transport and distribution among components. Water, Air and Soil Poll. 10, 3-11.

Jackson, D.R., Selvidge, W.J. and Ausmus, ,B.S. (1978b) Behaviour of heavy metals in forest microcosms: II. Effects of nutrient cycling processes. Water, Air and Soil Poll. 10, 13-18.

Kelly, J.M., Strickland, R.C., Weatherford, F.P. and Noggle, J.C. (1984) Evaluation of Simulated Acid Precipitation Effects on Forest Microcosms, Final Report EA-3500. Palo Alto: Electric Power Research Institute.

Kloskowski, R., Scheunert, I., Klein, W. and Korte, F. (1981) Laboratory screening of distribution, conversion and mineralisation of chemicals in the soil-plant system and comparison to outdoor experimental data. Chemosphere 10, 1089-100.

Knacker, T., Marcinkowski, A. and Schallnass, H.-J. (1989a) Ecotoxicological effects of artificial smokes on a terrestrial microcosm. Arch. Toxicol. Suppl. 13, 398-401.

Knacker, T., Marcinkowski, A., FOrster, B., Arthur, M.F. and Tolle, D.A. (1989b) Refinement of terrestrial microcosms for evaluating fate and effects of chemicals. Proceedings Brighton Crop Protection Conference- Weeds, 259-65.

Knacker, T., R6mbke, J., Beck, L., Eder, M., Federschmidt, A., F6rster, B., Diehlmann, H., Jacobi, K. and Coenen-Stass, D. (1993) Ubertragbarkeit und Priizisierung der Wirkung- smechanismen chemischer Belastung in verschiedenen Okosystemen. Special Report of the Kernforschungsanlage Jiilich GmbH Projectleitung, Biologie, Okologie, Energie (in prepara- tion).

Larkin, R.P. and Kelly, J.M. (t988) A short-term microcosm evaluation of CO2 evolution from litter and soil as influenced by SO2 and SO4 additions. Water, Air and Soil Poll. 37, 273-80.

Lichtenstein, E.P., Fuhremann, T.W. and Shulz, K.R. (1974) Translocation and metabolism of 14C-phorate as affected by percolating water in a model soil-plant ecosystem. J. Agric. Food Chem. 22, 991--6.

Lichtenstein, E.P., Liang, T.T. and Fuhremann, T.W. (1978) A compartmentalized microcosm for studying the fate of chemicals in the environment. J. Agric. Food Chem. 26, 948-53.

Lichtenstein, E.P.,:Liang, T.T. and Koeppe, M.K. (1982) Effects of fertilizers, captafol, and atrazine on the fate and translocation of [14C]fonofos and [14C]parathion in a soil-plant microcosm. J. Agric. Food Chem. 30, 871-8.


Lichtenstein, E.P., Liang, T.T. and Koeppe, M.K. (1983) Effects of soil mixing and flooding on the fate and metabolism of [14C]fonofos and [14C]parathion in open and closed agricultural microcosms. J. Econ. Entomol. 76, 233-238.

Lu, P.-Y., Metcalf, R.L. and Carson, E.M. (1978) Environmental fate of five radiolabelled coal conversion by-products evaluated in a laboratory model ecosystem. Environ. Health Perspect. 24, 201-8.

Malanchuk, J.L. and Joyce, K. (1983) Effects of 2,4-D on nitrogen fixation and carbon dioxide evolution in a soil microcosm. Water, Air and Soil Poll. 20, 181-90.

Malanchuk, J.L., Mueller, C.A. and Pomerantz, S.M. (1980) Microcosm evaluation of the agricultural potential of fly ash amended soil. In G.P. Giesy, Jr. (ed.) Microcosms in Ecological Research, pp. 1034-49. US Dept. of Energy.

Massart, D.L., Vandeginste, B.G.M., Deming, S.N., Michotte, Y. and Kaufinan, L. (1988) Chemometrics: A Textbook. Elsevier.

Matthews, G.B., Matthews, R.A. and Ehinger, W.J. (1991) Mathematical analysis of temporal and spatial trends in the benthic macroinvertebrate communities of a small stream. Can. J. Fish. Aquatic Sci. 48, 2184-90.

Metcalf, R.L., Sangha, G.K. and Kapoor, I.P. (1971) Model ecosystem for the evaluation of pesticide biodegradability and ecological magnification. Environ. Sci. and Tech. 5, 709-13.

Metcalf, R.L., Booth, G.M., Schuth, C.K., Hansen, D.L. and Lu, P.-Y. (1973) Uptake and fate of di-2-ethyl hexyl phthalate in aquatic organisms and in a model ecosystem. Environ. Health Perspect. 4, 27-34.

Mieth, A., Erode, M., Janzen, M. and Nissanga, J. (1993) Testung yon Pflanzenschutzmitteln und Umweltchemikalien in Semi-Freiland-Moddl6kosystemen, Final Report Nr. 126 05 083 to Umweltbundesamt, Christian-Albrechts-Universitiit, Kiel.

Mitchell, M.J., Parkinson, C.M., Hamilton, W.E. and Dindall, D.L. (1982) Role of the earthworm, Eisenia foetida, in affecting organic matter decomposition in microcosms of sludge- amended soil. J. Applied Ecol. 19, 805-12.

Nagpal, N.K. (1986) Effect of soil and effluent characteristics on phosphorus sorption in dosed columns. J. Environ. Qual. 15, 73-8.

Nash, R.G., Beall, M.L. Jr. and Harris, W.G. (1977) Toxaphene and 1,1,1-trichloro-2,2-bis(p- chlorophenyl) ethane (DDT) losses from cotton in an agroecosystem. J. Agric. Food Chem. 25, 336-41.

Nash, R.G. and Beall, M.L. Jr. (1980a) Fate of maneb and zineb fungicides in microagroecosy- stem chambers. J. Agric. Food Chem. 28, 322-30.

Nash, R.G. and Beall, M.L. Jr. (1980b) Distribution of silvex, 2,4-D and TCDD applied to tuff in chambers and field plots. J. Agric. Food Chem. 28, 614-23.

O'Connor, G.A., Wieranga, P.J., Cheng, H.H. and Doxtader, K.G. (1980) Movement of 2,4,5- T through large soil columns. Soil Sci. 130, 157-62.

Odum, E.P. (1962) Relationships between structure and function in ecosystems. Jap. J. Ecol. 12, 108-18.

Odum, E.P. (1971) Fundamentals of Ecology. Saunders, Philadelphia. Overdieck, D. (1986) Long-term effects of an increased CO2 conc. on terrestrial plants in model

ecosystems; morphology and reproduction of Trifolium and Lolium. Int. J. Biometeorol. 30, 323-32.

Overdieck, D. and Reining, E. (1986a) Effect of atmospheric CO2 enrichment on perennial ryegrass and white dover competing in managed model ecosystems. I. phytomass production. Acta Oecologica/Oecologia Plant. 7, 357-66.

Overdieck, D. and Reining, E. (1986b) Effect of atmospheric CO2 enrichment on perennial ryegrass and white dover competing in managed model ecosystems. II. nutrient uptake. Acta Oecologica/Oecologia Plant. 7, 367-78.

Parker, L.W., Ryder-White, J., Thomas, S. and Whitford, W.G. (1985) Effects of oxamyl and chlordane on the activities of non-target soil organisms. Biol. Fert. Soils 1, 141-8.


Piwoni, M.D., Wilson, J.T., Waiters, D.M., Wilson, B.H. and Enfield, C.G. (1986) Behaviour of organic pollutants during rapid infiltration of waste water into soil 1. Processes, definition and characterization using a microcosm. Hazard. Waste Hazard. Materials 3, 43-55.

Pritchard, P.H. (1982) Model Ecosystems. Environmental Risk Analysis for Chemicals, Conway, R.A. (ed.), pp. 257-353. Van Nostrand Reinhold.

Pritchard, P.H. and Bourquin, A.W. (1984) The use of microcosms for evaluation of interactions between pollutants and microorganisms. Adv. Microbial Ecol. 7, 133-217.

R6mbke, J., Knacker, T., F6rster, B. and Marcinkowski, A. (1993) Comparison of effects of two pesticides on soil organisms in laboratory tests, microcosms and in the field. In H. Eijsackers, F. Heimbach and M.H. Donker (eds) Ecotoxicology of Soil Pollution (in press). Lewis.

Schuphan, I. (1986) Determination of the quantitative ecochemical and ecotoxicological behaviour of pesticides using vegetation chambers with controlled ventilation conditions. Plant Research and Development; a Biannual Collection of Recent German Contributions 23, 92-108.

Schuphan, I., Sch/irer, E., Heise, M. and Ebing, W. (1987) Use of laboratory model ecosystems to evaluate quantitatively the behaviour of chemicals. In R. Greenhalgh and T.R. Roberts (eds) Pesticide Science and Biotechnology, pp. 437 44. Oxford: Blackwell.

Seastedt, T.R. and Crossley, D.A. Jr. (1983) Nutrients in forest litter treated with naphthalene and simulated throughfall: a field microcosm study. Soil Biol. Biochem. 15, 159-65.

Shirazi, M.A., Lighthart, B. and Gillett, J. (1984) A method for scaling biological response of soil microcosms. Ecol. Modelling 23, 203-26.

Smith, S. and Willis, G.H. (1985) Movement of pesticides in soil columns as affected by ammonia. Environ. Toxicol. Chem. 4, 425-34.

Tolle, D.A., Arthur, M.F. and Van Voris, P. (1983) Microcosm/field comparison of trace element uptake in crops grown on fly ash-amended soil. Sci. Total Environ. 31,243-61.

Tolle, D.A., Arthur, M.F., Duke, K.M. and Chesson, J. (1990) Ecological effects evaluation of two phosphorus-smoke producing compounds using terrestrial microcosms. ASTM STP 1091, 127-42. American Society for Testing and Materials.

Van Voris, P., Arthur, M.F. and Tolle, D.A. (1982) Evaluation of Terrestrial Microcosms for Assessing Ecological Effects of Utility Wastes. Electric Power Research Institute.

Van Voris, P., O'Neill, R.V., Emanuel, W.R. and Shugart, H.H. Jr. (1980) Functional complexity and ecosystem stability. Ecology 61, 1352-60.

Van Voris, P., Tolle, D.A., Arthur, M.F., Chesson, J. and Zwick, T.C. (1984) Development and Validation of a Terrestrial Microcosm Test System for Assessing Ecological Effects of Utility Wastes. Electric Power Research Institute. Research Report Centre, Palo Alto,

Van Wensem, J. and Adema, T. (1991) Effects of fluoride on soil fauna mediated litter decomposition. Environ. Poll. 72, 239-51.

Van Wensem, J., Jagers op Akkerhuis, G.A.J.M. and Van Straalen, N.M. (1991) Effects of the fungicide triphenyl tin hydroxide on soil fauna mediated litter decomposition. Pesticide Sci. 32, 307-16.

Vishwanathan, R. (1992) Study of pesticide impact on earthworms using a closed laboratory model ecosystem. In P.W. Greig-Smith, H. Becker, P.J. Edwards and F. Heimbach. (eds) Ecotox- icology of Earthworms, pp. 217-19. Andover UK: Intercept.

Winkelmann, D.A. and Klaine, S.J. (1991) Atrazine metabolite behaviour in soil-core microcosms. Formation, disappearance, and bound residues. ACS Symp. Ser., 459 (Pesticide Transformation Products: Fate and Significance in the Environment), 75-92. American Chemical Society.

Wolfe, N.L., Burns, L.A., and Stern, W.C. (1982) Use of linear free energy relationships and an evaluative model to assess the fate and transport of phthalate esters in the aquatic environment. Chemosphere 9, 393-402.

Wright, D.H. and Coleman, D.C. (1988) Soil faunal vs. fertilization effects on plant nutrition: results of a biocide experiment. Biol. Fert. Soils 7, 46-52.

the role of laboratory terrestrial model ecosystems in the testing of potentially harmful substances

Documents