ORNL/EPA-4
EPA-600/3-80-042
Microcosms as Potential Screening
Tools for Evaluating Transport
and Effects of Toxic Substances
W. F. Harris
ENVIRONMENTAL SCIENCES DIVISION
Publication No. 1506
Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Athens, Georgia 30605
-------
Printed in the United States of America. Available from
National Technical Information Service
U.S. Department of Commerce
5285 Port Royal Road, Springfield, Virginia 22161
NTIS price codes—Printed Copy: A18 Microfiche A01
This report was prepared as an account of work sponsored by an agency of the
U nited States G overnment. Neither the U nited States Government nor any agency
thereof, nor any of their employees, makes any warranty, express or implied, or
assumes any legal liability or responsibility for the accuracy, completeness, or
usefulness of any information, apparatus, product, or process disclosed, or
represents that its use would not infringe privately owned rights. Reference herein
to any specific commercial product, process, or service by trade name, trademark,
manufacturer, or otherwise, does not necessarily constitute or imply its
endorsement, recommendation, or favoring by the United States Government or
any agency thereof. The views and opinions of authors expressed herein do not
necessarily state or reflect those of the United States Government or any agency
thereof.
-------
ORNL/EPA-4
EPA-600/3-80-042
MICROCOSMS AS POTENTIAL SCREENING TOOLS
FOR EVALUATING TRANSPORT AND EFFECTS
OF TOXIC SUBSTANCES
by
B. S. Ausmus, G. K. Eddlemon, S. J. Draggan, J. M. Giddings,
D. R. Jackson, R. J. Luxmoore, E. G. O'Neill,
R. V. O'Neill, Monte Ross-Todd, P. Van Voris
Edited by
W. F. Harris
Environmental Sciences Division
Oak Ridge National Laboratory
Oak Ridge, Tennessee 37830
Operated by
Union Carbide Corporation
under Contract W-7405-eng-26
with the
U. S. Department of Energy
Interagency Agreement No. IAG-D6-0713
Environmental Sciences Division Publication No. 1506
Date Published: June 1980
Project Officer
Donald L. Brockway
Environmental Systems Branch
Environmental Research Laboratory
Athens, Georgia 30605
Project Managers
W. F. Harris and R. I. Van Hook
Oak Ridge National Laboratory
Oak Ridge, Tennessee 37830
OAK RIDGE NATIONAL LABORATORY
OAK RIDGE, TENNESSEE 37830
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ATHENS, GEORGIA 30605
U.S. Environm-rP^' Election ARency
Region 5, Li->;' '<)
77 West Jack. ^rd, 12th Floor
Chicago, IL 6u--wt-3L'^0
-------
DISCLAIMER
This report has been reviewed by the Environmental Research Laboratory,
U.S. Environmental Protection Agency, Athens, Georgia, and approved for pub-
lication. Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency, nor
does mention of trade names or commercial products constitute endorsement
or recommendation for use.
ii
-------
FOREWORD
Environmental protection efforts are increasinglydirected towards
preventing adverse health and ecological effects associated with specific
compounds of natural or human origin. As part of this Laboratory's research
on the occurrence, movement, transformation, impact, and control of environ-
mental contaminants, the Environmental Systems Branch studies complexes of
environmental processes that control the transport, transformation, degrada-
tion, and impact of pollutants or other materials in soil and water and
assesses environmental factors that affect water quality.
Concern about environmental exposure to toxic substances has increased
the need for accurate information on the transport, fate, and effects of
trace contaminants in natural waters. One technique that shows promise as
a useful and inexpensive tool for providing some of this information is the
use of laboratory microcosms as ecosystems for the assessment of pollutant
exposure to natural ecosystems. This report evaluates microcosms as research
tools for providing accurate and reliable data concerning the potential trans-
port, distribution, transformations, and persistence of inorganic and organic
contaminants released to the environment.
David W. Duttweiler
Director
Environmental Research Laboratory
Athens, Georgia
-------
ABSTRACT
Terrestrial and aquatic microcosms were evaluated for use in research
on environmental contaminants. Research completed in this project attempted
to (1) evaluate relationships among size, complexity, stability, and replica-
bility; (2) assess the simularity between microcosm results and actual envi-
ronmental transport and effects; (3) identify and quantify system-level para-
meters that might be sensitive indicators of effects of chemical contaminants;
(4) determine the relationship between system-level parameters measured in
microcosms and in natural ecosystems; and (5) suggest protocols for establish-
ing, maintaining, and interpreting results from microcosms. A number of ex-
perimental approaches using several microcosm designs and contaminants are
reported. The use of terrestrial microcosm results was evaluated in conjunc-
tion with mathematical simulation models as a means of extending short-term
experimental results to interpret conditions as they might occur in the natu-
ral ecosystem. Although microcosms offer an excellent experimental system,
their application to toxic substance testing is not a straightforward matter.
They are characterized by complex dynamics and counterintuitive responses
just as is the ecosystem to which they are an analog. These factors notwith-
standing, microcosms do offer an excellent means of studying specific aspects
of contaminant behavior and ecosystem processes. With appropriate attention
to the design of specific questions, answers to which are relevant to inter-
preting ecological transport and effects of contaminants, microcosms can be
useful tools.
This report was submitted in fulfillment of Interagency Agreement No.
EPA-IAG-D6-0713 by the Oak Ridge National Laboratory under the sponsorship
of the U.S. Environmental Protection Agency. This report covers the period
30 April 1975 to 30 June 1978, and work was completed as of 30 September 1979.
-------
ACKNOWLEDGMENTS
A number of people contributed to the preparation of this report.
Natalie Millemann edited all drafts of the document; Cheryl Phillips
typed initial drafts of material; and the staff of the ESD Editorial
Production Office, especially Karen Hoi brook and Patsy Thornton, typed
review drafts and the final copy. The contributions of these people
were invaluable and are gratefully acknowledged.
vii
-------
TABLE OF CONTENTS
Page
ABSTRACT vii
LIST OF TABLES xiii
LIST OF FIGURES xix
1.0 INTRODUCTION 1
2.0 SUMMARIES, CONCLUSIONS, AND RECOMMENDATIONS 5
2.1 AQUATIC STUDIES 5
2.1.1 Summary 5
2.1.2 Conclusions 6,
2.1.3 Recommendations 8
2.2 TERRESTRIAL STUDIES 9
2.2.1 Summary (9;
2.2.2 Conclusions (10
2.2.3 Recommendations (12/
3.0 AQUATIC STUDIES 15
3.1 INTRODUCTION 15
3.1.1 State of the Art, Criteria, and Rationale .... 15
3.1.2 Summary of Research Strategy 17
3.2 SUMMARY OF RESULTS 24
3.2.1 Properties of Microcosms 24
3.2.1.1 Basic Ecological Characteristics of
Pond Microcosms 24
3.2.1.2 Comparison of Microcosms with an
Actual Pond 39
3.2.1.3 Replicability 40
3.2.1.4 Repeatability 49
3.2.1.5 Effects of Variations in Experimental
Condition and Microcosm Design 53
3.2.2 Transport Studies 71
3.2.2.1 Arsenic 71
3.2.2.2 Chromium 95
3.2.2.3 Selenium 103
3.2.2.4 Mercury 115
3.2.2.5 Anthracene 129
-------
3.2.3 Ecological Effects 142
3.2.3.1 Production/Respiration Ratios 143
3.2.3.2 Water Chemistry 148
3.3 DISCUSSION 155
3.3.1 Protocol for the Construction of Pond
Microcosms 155
3.3.1.1 Collection of Components 157
3.3.1.2 Assembly of Microcosms 157
3.3.2 The Role of Pond Microcosms in Contaminant
Research 160
3.3.2.1 Transport 164
3.3.2.2 Fate 168
4.0 TERRESTRIAL STUDIES 171
4.1 INTRODUCTION 171
4.1.1 State of the Art, Criteria, and Rationale .... 171
4.1.2 Summary of Research Strategy 172
4.2 SUMMARY OF RESULTS 175
4.2.1 Behavior of Heavy Metals in Forest Microcosms -
Excised Tree Microcosms 176
4.2.1.1 Materials and Methods 176
4.2.1.2 Results and Discussion 180
4.2.2 The Behavior of Arsenic in Terrestrial
Microcosms 196
4.2.2.1 Materials and Methods 198
4.2.2.2 Results and Discussion 199
4.2.2.3 Summary and Conclusions 204
4.2.3 Extraction of Nutrients from Intact Soils
Cores 204
4.2.3.1 Materials and Methods 205
4.2.3.2 Results and Discussion 209
4.2.4 Hollow Fiber Membranes for Detecting Microbial
and Nutrient Cycling Effects of Contaminants . . 214
4.2.4.1 Materials and Methods 214
4.2.4.2 Results and Discussion 217
4.2.4.3 Summary 219
-------
Page
4.2.5 The Effect of Substrate on Elemental
Transport Through Terrestrial Microcosms .... 220
4.2.5.1 Results and Discussion 222
4.2.6 Functional Complexity and Microcosm Stability . . 226
4.2.6.1 Materials and Methods 229
4.2.6.2 Results 236
4.2.6.3 Discussion 239
4.3 SYNTHESIS OF TERRESTRIAL MICROCOSM RESULTS 239
4.3.1 Introduction 239
4.3.2 Minimal Monitoring Parameters 240
4.3.3 Size Effects 250
4.3.4 Substrate Effects 257
4.4 SIMULATION MODELING AND APPLICATION TO MICROCOSM AND
FIELD DATA 261
4.4.1 Parameter Sensitivity Studies 264
4.4.1.1 Materials and Methods 264
4.4.1.2 Results 268
4.4.1.3 Discussion 291
4.4.2 Simulation of Chemical Transport in a Deciduous
Forest Using Microcosm Calibration 295
4.4.2.1 Tree Microcosm Study and Model
Parameterization 296
4.4.2.2 Field Study and Test Case Simulation . . 300
4.4.2.3 Discussion 304
4.5 DISCUSSION 306
4.5.1 The Role of Terrestrial Microcosms in
Contaminant Research 306
4.5.1.1 Introduction 306
4.5.1.2 Measured Parameters 310
4.5.1.3 Replicability, Reproducibility 315
4.5.1.4 Accuracy 317
4.5.1.5 Problems 322
4.5.1.6 Recommendations 323
XI
-------
4.5.2 Protocol for Construction and Operation of
Terrestrial Microcosms 324
4.5.2.1 Objectives 324
4.5.2.2 Rationale 324
4.5.2.3 Protocol 1 327
4.5.2.4 Protocol 2 329
5.0 LITERATURE CITED 335
APPENDIX A - Synopsis of Aquatic Experiments 351
APPENDIX B - Microcosm Biota 371
APPENDIX C - Project Publications 375
-------
LIST OF TABLES
Table Page
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3.13
3.14
3.15
3.16
3.17
3.18
3.19
Comparison of water chemistry in microcosms, pond
enclosures, and a pond
Coefficients of variation: Microcosm water chemistry . .
Coefficients of variation: Ecosystem metabolism ....
Coefficients of variation: Contaminant concentrations
Arsenic concentrations in microcosm components -
Comparison of Experiment 1 and Experiment 2
Mean nutrient concentrations: Experiment 1
Summary of water chemistry data: Experiment 1
Experiment 1 final arsenic concentrations (ppm)
Mean coefficients of variation (for 3 replicates) of
chemical parameters in 7- and 70- liter microcosms ....
Coefficients of variation for arsenic concentrations
(%) in Experiment 1
Temperatures of microcosm experiments
Gross primary production in fish microcosms
Water chemistry in fish microcosms after 12 weeks
(all values are in ppm)
Summary of kinetic loss rate constants for microcosm
Adsorption and desorption parameters for arsenic in
microcosms (calculated from equation 1)
Arsenic concentrations and bioaccumulation ratios
from the literature
Final mean arsenic concentrations (ppm): Experiment 2 .
Final arsenic mass balance - Experiment 2
Chromium kinetic loss rate constants, r, for microcosm
water
41
43
47
48
S?
Sfi
*>7
61
65
66
68
7?
7?
77
78
8?
89
P?
99
XTM
-------
Table Page
3.20 Final Cr distribution .................. 100
3.21 Final Se concentrations (^g/g dry wt) .......... 106
3.22 Final Se mass balance ........ . ......... 109
3.23 Final Hg mass balance .................. 119
3.24 Final Hg concentrations in microcosm components due
to initial dose of 0.42 ng/ml (ppb) in water ...... 120
3.25 Final C activity in microcosm components twelve
weeks after addition of 10.4 x 106 cpm as anthracene . . 134
3.26 14C budget ....................... 136
3.27 Percentage distribution of radioactivity in extracts . . 137
3.28 Effects of arsenic on microcosm water chemistry:
Experiment 2 ...................... 152
3.29 Inorganic nitrogen concentrations following arsenic
treatment: Experiment 2 ................ 154
3.30 Surnnary of bioaccumulation ratios ............ 163
4.1 Chemical and physical characteristics of experimental
soils .......................... 174
4.2 Source of heavy metals (Pb, Zn, Cd, Cu) deposited as
contaminated litter and as baghouse dust on forest
microcosms ....................... 178
4.3 Heavy metal concentrations in litter, soil, and vegetation
(Acer rubrum) components of forest microcosms (N=3) . . . 181
4.4 Total leachate export of heavy metals from forest
microcosms over the 20-month experiment ......... 183
4.5 Heavy metal enrichment ratios for forest microcosms . . . 185
4.6 Initial litter and soil nutrient pools in contaminated
and control microcosms ................. 186
4.7 Total dissolved nutrient export from contaminated and
control forest microcosms over the 20-month experiment . 188
4.8 Final litter nutrient pools and total changes in
nutrient pools over the 20-month experiment ....... 189
xiv
-------
Table Page
4.9 Mean nutrient concentrations of forest microcosm
components at experiment termination 191
4.10 Comparison of average C02 efflux rates (mg C02/microcosm/
day) of contaminated and control microcosms during spring
and summer of the two years of observations 194
4.11 Simple linear regression of cumulative C02 efflux (mg)
upon time (days) for contaminated and control microcosms
for the two years of experimentation 195
4.12 Microbial parameters in soil of contaminated and control
microcosms at the end of the 20-month experiment .... 197
4.13 Final As concentrations of As-amended forest and
grassland soil cores (± S.E., n=3) 200
4.14 Aqueous efflux of Ca and N03-N from forest and
grassland soil cores amended with As (± S.E., n=18) . . . 202
4.15 Extraction of nutrients from intact ^pil cores using
various extractant solutions [ng/g (X ± S.E.)] 207
4.16 Mean nutrient concentrations (fig/ml) of soil solutions
extracted from 10- and 20-cm depths with hollow fibers
and lysimeters (± 1 S.E., n=10) 218
4.17 Cumulative elemental transport (mg) through
terrestrial microcosms of three substrates over
a 118-day experiment 223
4.18 Rates of essential element transport (mg/microcosm/week)
through terrestrial microcosms constructed of three
substrates (N = 18 weekly observations) 225
4.19 Probability (Z) that successive observations of elemental
transport through any microcosm unit differ by > 100%
(N = 17/repli cates) 227
4.20 Rankings of functional complexity and ecosystem stability
based on monitoring of C02 and total Ca export in soil
leachate 234
4.21 Summary of microcosm experiments and corresponding
parameters 241
4.22 Comparison of variability among elements leached from
terrestrial microcosms 242
4.23 Comparison of variability among elements leached from
terrestrial microcosms 244
xv
-------
Table Page
4.24 Alpha values for the significance of treatment (toxic
substance) effect in eight experiments 246
4.25 Average "t" values compared among elements 247
4.26 Correlation coefficients for Ca and Mg export 249
4.27 Microcosm size comparison on intact cores in Grassland I
where size was a variable 251
4.28 Microcosm size comparison on intact cores
in Grassland II 252
4.29 Comparison of variability of Ca and N03 between small
and large microcosms 254
4.30 Comparison of C02 efflux of small (10x10 cm) and large
(30x15 cm) intact core microcosms 255
4.31 Calculated F values for treatment effect in two microcosm
experiments 256
4.32 Homogenized grassland soil cores and intact grassland
cores were compared in Grassland II 259
4.33 Comparison of variability of Ca and N03 between intact
and homogenized microcosms 260
4.34 Some attributes of coupled models describing carbon,
water, and chemical dynamics in the soil pi ant-litter
system 263
4.35 Standard parameter values and tested values for
sensitivity analyses 265
4.36 Model output variables considered in this study 267
4.37 Subjective evaluation of parameter effects on total
chemical level and distribution in soil, litter,
and plant components 293
4.38 Comparison of water, carbon, and chemical data for
the microcosm and modeling studies 298
4.39 A comparison of initial simulation, experimental data,
and simulation results from this study for Crooked Creek
Watershed 302
4.40 Uses that various microcosm units have for specific
screening applications 308
xvi
-------
Table Page
4.41 Microcosm test systems, substances tested, and
citations 311
4.42 Population and system level parameters measured in soil
microcosms, forest microcosms, and forest under control
and chemical stress conditions 314
4.43 Indices of microbial activity measured at termination
of a microcosm experiment using complex heavy metal
emissions from a lead smelter 316
4.44 Comparison of three parameters measured in control
replicates of three soil core microcosm experiments . . . 318
4.45 Comparison of extrapolated annual export among
substrates 321
xv ii
-------
LIST OF FIGURES
Figure Page
3.1 Production and respiration rates in microcosms
from Experiment 9 28
3.2 Production/respiration ratios from three microcosm
experiments 30
3.3 Dissolved organic carbon in microcosm water from
Experiment 1 (mean of three 70-liter sediment
microcosms), Experiment 2 (mean of three control
microcosms), and Experiment 9 (mean of two microcosms) . 31
3.4 Dissolved oxygen and pH in microcosm water,
Experiment 2 33
3.5 Nitrate, ammonium, and total phosphorus concentrations
in microcosm water, Experiment 9 34
3.6 Nitrate, ammonium, and total phosphorus concentrations
in microcosm interstitial water, Experiment 9 36
3.7 Coefficients of variation of conductivity, dissolved
oxygen, and pH, Experiment 2 44
3.8 Coefficients of variation of nitrate, ammonium,
phosphate, total phosphorus, and dissolved organic
carbon, Experiment 2 45
3.9 Dissolved arsenic concentrations, Experiments 1
and 2 51
3.10 Dissolved oxygen concentrations, Experiment 1 54
3.11 pH, Experiment 1 55
3.12 Final algal biomass, Experiment 1 58
3.13 Final chlorophyll-a concentrations, Experiment 1 .... 59
3.14 Arsenic concentrations in water, Experiment 1 60
3.15 Equilibration of dissolved oxygen in 7- and 70-liter
sediment microcosms, Experiment 1 62
3.16 Equilibration of pH in 7- and 70-liter sediment
microcosms, Experiment 1 63
xix
-------
Figure Page
3.17 Final distributions of arsenic among microcosm
components. Experiment 1 ................ 79
3.18 Regression of snail arsenic concentrations (^g As/g
dry wt) on substrate arsenic concentrations,
Experiment 1 ...................... 81
3.19 Dissolved arsenic concentrations, Experiment 2 ..... 87
3.20 Dissolved arsenic concentrations, Experiments 2
and 3 ......................... 88
3.21 Sediment arsenic profiles, Experiment 2 ........ 91
3.22 Dissolved chromium concentrations, Experiment 4 .... 97
3.23 Sediment chromium profiles, Experiment 4 . . ...... 101
3.24 Dissolved 7^Se concentrations, Experiment 6 ...... 107
3.25 Sediment 75$e concentrations, Experiment 6 ....... 108
3.26 75se concentrations in algae and Elodea tips,
Experiment 6 ...................... 110
3.27 75$e concentrations in Elodea stems, Experiment 6 ... Ill
3.28 75Se concentrations in Elodea roots, Experiment 6 ... 112
3.29 75$e concentrations in fish and invertebrates,
Experiment 6 ...................... 114
3.30 Dissolved 203ng concentrations, Experiment 7 ...... 117
3.31 Sediment 203ng concentrations, Experiment 7, top cm
only .......................... 118
3.32 Elodea 203ng concentrations, Experiment 7,
replicate Cl ...................... 122
3.33 Elodea ng concentrations, Experiment 7,
replicate C2 ...................... 123
3.34 Bioaccumulation ratios for 203ng in Elodea tips
and stems, Experiment 7 ................ 124
3.35 Snail tissue Hg concentrations, Experiment 7 .... 125
3.36 Bioaccumulation ratios for 203^g n-n snaii tissue,
Experiment 7 ...................... 126
xx
-------
Figure
3.37
3.38
3.39
3.40
3.41
3.42
3.43
3.44
3.45
3.46
3.47
4.1.
4.2.
4.3.
4.4.
4.5.
4.6.
4.7.
Zooplankton 203ng concentrations, Experiment 7,
replicate Cl
l^C activity in microcosm water, Experiment 8
Distribution of l^C activity in extracts from
microcosm water, Experiment 8
Comparison of ^C distribution in extracts from
microcosm water and distilled water, Experiment 8 ...
Net production and nighttime community respiration in
microcosms at three arsenic levels, Experiment 2 ....
Production/respiration ratios in microcosms at three
arsenic levels, Experiment 2
pH, Experiment 2
Dissolved oxygen concentrations, Experiment 2
Alkalinity, Experiment 2
Dissolved organic carbon, Experiment 2
Total phosphorus concentrations, Experiment 2
Effect of heavy-metal contamination on extractable
soil nutrients by the intact and slurry methods
Mean weekly Ca efflux from untreated forest and
grassland soil cores
Nutrient pools extracted from Emory and
Captina soils by leaching intact and homogenized
soil cores, and by the slurry method
Schematic diagram of an intact soil core containing
devices for extracting in-situ soil solution
Schematic of glass manifolds for field use of
polycarbonate hollow fibers
Cross-section of old-field microcosm showing two
system-level monitoring points of C02 efflux and
nutrient leachate
Example of time series of C02 efflux from pasture
microcosms
Pag
128
133
138
140
144
146
149
150
151
153
156
192
203
211
216
221
230
231
xx i
-------
Figure Page
4.8. Mean DOC ( ) and N03-N ( ) concentration
in leachate (± 1 S.E., n = 13) 233
4.9. Relative lead distribution in plant, litter, and
soil components with change in lead distribution
coefficient (Kd) 270
4.10. Influence of zinc Kj on zinc transport in the
vegetation: (a) leaf-to-stem phloem, and (b) stem-
to-leaf xylem 271
4.11. Relative zinc distribution in plant, litter, and soil
components with change in zinc solubility (SP) 272
4.12. Relative lead distribution in plant, litter, and
soil components with change in diffusion coefficient
of lead (DL) 274
4.13. Relative zinc distribution in plant, litter, and soil
components with change in litter decomposition rate
constant (DMAX) 275
4.14. Relative chemical distribution in plant, litter,
and soil components with change in leaf-to-stem
phloem resistance (LSPHLO) for (a) lead, and (b)
zinc 277
4.15. Relative chemical distribution in plant, litter,
and soil components with change in standard root
respiration rate (RRESTD) for (a) lead, and
(b) zinc 279
4.16. Influence of standard stem respiration rate (SRESTD)
on sugar translocation in stem-to-root phloem 281
4.17. Relative chemical distribution in plant, litter,
and soil components with change in leaf area weight
ratio (ARI) for (a) lead, and (b) zinc 282
4.18. Influence of root radius (R) on root uptake of lead
during the sunlight period 284
4.19. Relative chemical distribution in plant, litter,
and soil components with change in maximum leaf
storage (L MAX) for (a) lead, and (b) zinc 286
4.20. Relative chemical distribution in plant, litter,
and soil components with change in leaf
permeability to chemicals (PERM) for (a) lead, and
(b) zinc 287
xxn
-------
Figure Page
4.21. Relative chemical distribution in plant, litter,
and soil components with change in leaf cuticle
thickness (FILM) for (a) lead, and (b) zinc ...... 289
4.22. Relative zinc distribution in plant, litter,
and soil components with change in root chemical
conductivity (CONDUC) ................. 290
4.23. Relative zinc distribution in plant, litter,
and soil components with change in atmospheric
C02 (C02X) ....................... 292
4.24. Replicability of two parameters monitored in
microcosms. (A) - COe efflux; (B) - DOC efflux .... 319
4.25. Protocol 1 soil core excised intact from a forest
ecosystem ....................... 328
4.26. Protocol 2 grassland microcosm excised from a
Festuca- dominated pasture ............... 331
4.27. Protocol 2 forest microcosm containing an Acer
rubrum sapling excised from a mesic hardwood forest . . 332
xxm
-------
1.0 INTRODUCTION
There is continuing concern about the environmental transport,
fate, and effects of toxic substances in the biosphere. Legislation
now requires, in part, a determination of the environmental fate of
chemicals before large-scale production, marketing, and dissipation can
occur. Implementation of this legislation requires reliable, adaptable,
and inexpensive research tools capable of evaluating environmental
transport, fate, and effects of trace contaminants. Sensitivity and
precision are requirements of the research techniques used to determine
the effects and distribution of chemicals released to the environment.
A recent survey of testing procedures that have proven useful in the
identification and quantification of the environmental transport of
chemicals revealed that laboratory microcosm experiments can provide
substantial information on the transport and accumulation of chemicals
in the environment (Witherspoon et al. 1976). We define microcosm as
any confined ecosystem or part thereof that may be subjected to labora-
tory controls due to its reduction in size or complexity. Such systems
may be created artificially or they may represent portions of natural
ecosystems which can be brought into the laboratory for study. The
report also indicated that there currently exist no standardized micro-
cosms for the determination of chemical fate, and that the reliability
and validity of results obtained from microcosm studies have not been
fully determined.
Testing procedures (possibly including microcosms) need to be
designed expressly for the determination of chemical fate in order to
-------
give results which are relevant to actual environmental occurrences.
In this approach to microcosm research we initially recognized that the
development of a single, complex microcosm for the assessment of chemi-
cal fate in numerous ecosystems is not realistic. Complex microcosms,
which contain all the major functional groups of natural ecosystems,
can be useful in identifying chemical accumulation, but are insensitive
for measuring rates of synergistic and/or competitive transport pro-
cesses. Microcosms of reduced complexity containing the essential
functional components of natural ecosystems can be used to estimate the
chemical fate in isolated ecosystem components, and are suitable for
measuring a defined number of rate processes over short time periods.
These synthetic microcosm systems do provide for estimation of rate
constants for identifiable transport processes in specific environments.
Several problems exist in the use of microcosm results for both
the description and prediction of environmental behavior of pollutants.
First is the problem of applying results from processes measured in
isolation to the more complex natural ecosystems. Secondly, the rela-
tionships among microcosms of different size and structure have not
been evaluated. Finally, there are no estimates of confidence limits
on parameter measurements and of microcosm reproducibility.
The work undertaken in this project was designed to provide some
insight to the above questions through the following objectives:
(1) To evaluate the relationships among microcosm size, complex-
ity, and replicability to application of microcosm results in both
-------
characterizing and mathematically predicting environmental transport,
accumulation, and persistence of contaminants.
(2) To determine the realism and validity of microcosm results in
relation to the ability of these results to reflect actual environmental
occurrences.
(3) To utilize results obtained in objectives 1 and 2 in evaluat-
ing microcosms as potential "screening tools" for trace toxic contami-
nants released from industrial sources.
The products of this research initially were conceived to include
(1) a greater understanding of the ecological processes governing micro-
cosm behavior; (2) an evaluation of microcosms as research tools for
providing accurate and reliable data concerning the potential transport,
distribution, transformations, and persistence of both inorganic and
organic contaminants released to the environment; and (3) an evaluation
of microcosms as "screening tools" for use by industry and federal
agencies in assessing potential environmental transport and distribution
of trace contaminants.
-------
2.0 SUMMARIES, CONCLUSIONS, AND RECOMMENDATIONS
2.1 AQUATIC STUDIES
2.1.1 Summary
A protocol has been developed for the construction and operation
of pond microcosms for use in research on environmental contaminants.
The microcosm components (water, sediment, and a complete biotic com-
munity) are taken directly from a natural pond, assembled in 70-liter
glass aquaria, and held in an environmental chamber at a constant
temperature on a 12-hr-light:12-hr-dark cycle. Pond microcosms equili-
brate (achieve internal constancy) within 6 to 8 weeks, and remain
stable for several months or even longer. The ecosystems which develop
in the microcosms are ecologically similar to the parent ecosystem.
Variations between replicate microcosms are small (coefficients of
variation generally less than 20%) for most measured parameters. The
protocol has been tested and refined through a series of experiments on
the environmental transport and fate of arsenic, chromium, selenium,
mercury, and anthracene (a polycyclic aromatic hydrocarbon). Pond
microcosms are too complex for analysis and measurement of individual
transport processes and degradation pathways, but they do provide
information on overall net rates and pathways under realistic environ-
mental conditions. These microcosms could also be used to study effects
of contaminants on community interactions and ecosystem processes.
-------
2.1.2 Conclusions
(1) Pond microcosms constructed according to the protocol described
herein are realistic simulations of shallow pond ecosystems.
(2) Pond microcosms achieve a state of internal constancy within 6 to
8 weeks, and persist in this condition for several months or
longer.
(3) Pond microcosms are about as replicable as most other biological
entities. Coefficients of variation among replicate microcosms
are generally below 20% for most measured parameters.
(4) Experiments on contaminant transport in microcosms are repeatable.
Results are not highly dependent on the time of year the micro-
cosms are constructed.
(5) Smaller (7-liter) microcosms behave similarly to the 70-liter
microcosms described in the protocol, but are less stable, more
easily disturbed, and more difficult to sample than the larger
microcosms.
(6) Substrate type greatly influences contaminant transport in pond
microcosms.
(7) Small fish are compatible with a stable and persistent microcosm
community, although they may affect community composition and
«T
metabolism. Snails and crayfish, on the other hand, have a
significant impact on microcosm ecology.
(8) Results of two experiments on arsenic transport were consistent
with each other and with published field data.
-------
(9) The loss of a contaminant from microcosm water can be described
by a first-order kinetic constant "r." Of the five contaminants
tested, values of r were highest for mercury (implying rapid
sorption by sediments) and lowest for selenium (implying greater
potential for hydrologic transport of this metal).
(10) Most of the contaminants tested were concentrated in organisms to
2 3
about 10 to 10 times the equilibrium water concentration.
Mercury was an exception, with bioaccumulation ratios (concentra-
tion in organism/concentration in water) of 10 to 10 ; this
is consistent with the known bioaccumulation tendencies of this
metal. Bioaccumulation ratios are most useful for comparisons
between contaminants; extreme values appear to be reliable indi-
cators of highly bioaccumulative substances.
(11) Microcosm experiments can provide information on the net degrada-
tion rate of a compound in a whole ecosystem and the identity,
quantity, and persistence of derivatives. Major degradation pro-
cesses may sometimes be inferred from the data, but confirmation
must come from experiments with simpler systems in which the
hypothesized pathways are isolated from competing processes.
(12) Although the primary goal of the project was to develop microcosms
for testing contaminant transport and fate, pond microcosms can
also be used to test for effects on whole ecosystems. Possible
"pulse points" for indication of ecosystem dysfunction include
the concentrations of nutrients in interstitial water and the
ratio of primary production to total ecosystem respiration.
-------
2.1.3 Recommendations
(1) The major advantage of pond microcosms over other contaminant
testing tools is their realistic simulation of an entire aquatic
ecosystem. Therefore, microcosms are most useful in the advanced
stages of contaminant testing, when accurate predictions of the
overall fate and effects of a substance are required, rather than
in the earlier screening stages.
(2) For accurate simulation of a particular aquatic ecosystem, micro-
cosm components (water, sediment, and biota) should be collected
from that ecosystem.
(3) Whenever possible, contaminants to be tested should be labeled
with an appropriate radioisotope. This reduces sample sizes
necessary for analysis, improves detection limits, and permits
identification of derivatives of the parent compound.
(4) Microcosm experiments integrate, and thereby obscure, individual
transport processes and degradation pathways. Single-species
experiments facilitate the analysis of individual processes, but
provide no direct means of assessing overall contaminant behavior.
These two complementary experimental approaches are most fruitful
when used in combination.
(5) Microcosm results can be used to identify weaknesses in analytical
models and can provide an overall picture of contaminant behavior
when models fail. Models, on the other hand, can be used in con-
junction with data from simplified experimental systems to assist
in the interpretation of microcosm results.
-------
(6) Further comparisons are needed between microcosms and the natural
ecosystems they represent. Comparisons of contaminant transport
and fate in microcosms and natural ecosystems are also needed.
(7) The feasibility of improving the accuracy of simulation through
the use of larger systems, natural light and temperature regimes,
and continuous nutrient input should be investigated.
2.2 TERRESTRIAL STUDIES
2.2.1 Summary
Terrestrial microcosms were evaluated for use in research on envi-
ronmental contaminants. Research completed in this project attempted
to: (1) evaluate relationships among size, complexity, stability, and
replicability; (2) assess the similarity between microcosm results and
actual environmental transport and effects; (3) identify and quantify
system-level parameters which might be sensitive indicators of effects
of chemical contaminants; (4) determine the relationship between
system-level parameters measured in microcosms and in natural ecosys-
tems; and (5) suggest protocols for establishing, maintaining, and
interpreting results from terrestrial microcosms. In addition to a
number of experimental approaches using several microcosm designs and
contaminants (principally trace metals), the use of microcosm results
was evaluated in conjunction with mathematical simulation models as a
means of extending short-term experimental results to interpret condi-
tions as they might occur in the natural ecosystem. While microcosms
offer an excellent experimental system, their application to toxic sub-
stance testing is not a straightforward matter. They are characterized
-------
10
by complex.dynamics and counter-intuitive responses just as is the eco-
system to which they are an analog. These factors notwithstanding,
terrestrial microcosms offer an excellent means of studying specific
aspects of contaminant behavior and ecosystem processes. With appro-
priate attention to the design of specific questions, answers to which
are relevant to interpreting ecological transport and effects of con-
taminants, terrestrial microcosms can be useful tools. As a panacea to
toxic substance testing, they leave many questions unanswered.
2.2.2. Conclusions
(1) The suite of seven chemical species (Mg, Ca, DOC, K, NO--N,
P, NH^-N) is too large a number to be practicable in a routine screen-
ing program. On the basis of variability and sensitivity tests, it
would appear that NO^-N plus either Ca or Mg would serve as a practi-
cal pair of chemical species for a screening program.
(2) Based on nutrient variability and sensitivity, the effect of
size (10 x 10 cm vs 30 x 25 cm) was significant (P < 0.01). With few
exceptions, the large (30- x 25-cm) microcosms showed greater concentra-
tions of leachates.
(3) With respect to effects of size on variability, no unambiguous
pattern was apparent; indeed, in many cases, the larger microcosms
showed greater variability both through time and between replicates.
(4) Size effects also appeared to play a role in C02 measure-
ments. Large microcosms showed smaller coefficients of variation with
great consistency.
-------
11
(5) Large microcosms (30 x 25 cm) showed a clearer treatment
effect (larger F ratio) than did small microcosms. Smaller size micro-
cosms, however, have sufficient sensitivity to detect an effect. Choice
of size will require greater operating experience to decide the relative
advantages of using a few large microcosms versus a larger number of
smaller microcosms.
(6) Because the major argument for using a homogenized substrate
is the potential for decreased variability, and because our data indi-
cate that there is not a consistent trend for variability with this
substrate, there seems little argument for use of homogenized substrate.
(7) Forest soils used in these experiments were more sensitive
(particularly Ca loss) than grassland soils.
(8) Others have identified the need to relate quantitatively the
data from microcosms to chemical behavior in natural ecosystems. Our
test case application in this study showed that such a simulation is
limited by the field relevance of the data set. Microcosm data provide
one means of model parameter estimation; however, the parameter values
may not be representative of the natural ecosystem.
(9) Forest microcosms treated with heavy metals in smelter bag-
house dust yielded final results qualitatively consistent with observed
conditions in the immediate environs of the smelter.
(10) Techniques were developed for measuring soil nutrients which
appear to be more sensitive to detecting subtle effects on soil chem-
istry than conventional slurry techniques.
-------
12
(11) Long-term monitoring of nutrient export dynamics in grass
microcosms (soil intact) showed that seasonal changes in sensitive
parameters might obscure contaminant effects on nutrient dynamics.
Therefore, a control series should be included in each experimental
series.
(12) Based on detailed analysis of microcosm metabolism (CCL
efflux) and responses to a chemical perturbation (Cd applied to intact
grassland microcosms), it was experimentally shown that increasing
stability appeared to accompany increased complexity. This finding is
important not only to further use of microcosms in experimental ecology
but also to use of microcosms in toxicant screening since stability and
complexity are fundamental attributes of ecosystems. Therefore, how
these attributes compare between microcosm and natural systems as well
as how these attributes influence toxicant behavior is important.
2.2.3 Recommendations
(1) Microcosms are isolated portions of natural ecological assem-
blages. Research to date shows that on at least a qualitative basis
there is a correspondence between transport and effects of toxicants in
microcosm versus natural ecosystems. Furthermore, terrestrial micro-
cosms from different temperate ecosystems (mesic forest vs old-field
grassland) are qualitatively, if not quantitatively, similar in their
behavior with respect to baseline conditions and in response to toxicant
input. There should be no reason to expect a single terrestrial micro-
cosm design to show exact correspondence with nature. The variety of
nature and particular properties of microcosms both point to this
-------
13
Inescapable conclusion. Therefore, the utility of microcosms in evalu-
ating potential toxicants lies in how will we ask questions about eco-
logical effects and transport. As with any rigorous scientific study,
the value of the information derived is dependent upon how carefully
the initial hypothesis is structured. Therefore, there remains the
need to decide upon that set of important questions about the ecological
transport and effects of a potential toxicant. The experimental design
will then follow naturally. Experimental ecologists, indeed biologists
generally, have used microcosm-like laboratory systems long before the
recognition of their applicability to toxicant screening.
(2) The exercise undertaken in this project to employ mathematical
models based on physicochemical and biological data in microcosm studies
as a means to develop models for environmental behavior of chemicals
suggests additional areas for analysis. There is a need to determine
how best to use microcosm data to evaluate model algorithms, to validate
models, to estimate model parameters in a manner representative of the
real world, and to combine to simplified representations (microcosms
and simulation models) in order to gain insights into longer time-frame
behavior.
(3) One feature consistently exhibited by terrestrial microcosms
subjected to contaminants was the accelerated loss of nutrients (Ca,
NO.,-N, and others on a less consistent basis). Nutrient loss gener-
ally accompanies disturbance, especially physical disturbances, of
terrestrial ecosystems. The exact mechanisms underlying the greater
loss of nutrients are not completely known. Furthermore, because of
the cyclic nature of nutrient metabolism in terrestrial ecosystems,
-------
14
there is no reason to expect a single mechanism or set of mechanisms to
be responsible. Therefore, a major uncertainty to the use of terres-
trial microcosms involves studies of nutrient metabolism processes to
determine exactly the causal factors underlying the accelerated nutrient
loss. This is essential for two reasons: (1) while the phenomenon is
general, it is logical to expect it to be accepted as an index of eco-
logical effects without being in a position to explain the origin of
the phenomenon; and (2) if the mechanisms leading to excess nutrient
leakage can be identified and the number of physicochemical factors
involved are not extensive, this knowledge may suggest very simple
experimental systems applicable to screening purposes.
(4) Finally, while it is possible to design a biological system
in the laboratory which has many properties in common with the real
world and which demonstrates an effect when perturbed with a foreign
substance, there remain questions about recovery, chronic vs acute
exposure, and ecological significance of any effect demonstrated in the
laboratory. Therefore, there is a need to calibrate real-world
responses with microcosm responses. Such a calibration assumes that an
acceptable level of ecological behavior can be agreed upon. This is
not a trivial assumption.
-------
3.0 AQUATIC STUDIES
3.1 INTRODUCTION (J. M. Giddings)
3.1.1 State of the Art, Criteria, and Rationale
Shortly before this project was initiated, a review of the state
of the art of assessing the environmental fate and transport of chemi-
cals was completed (Witherspoon et al. 1976). The review indicated
that simplified microcosms, such as that developed by Metcalf and his
associates (Metcalf et al. 1973), were suitable for measurement of rate
processes over short time periods. Three critical problems were
identified:
(1) REALISM: the difficulty of applying results obtained in
simplified systems to complex natural environments;
(2) REPRODUCIBILITY: the lack of information on the comparability
of data from similar but nonidentical microcosms; and
(3) REPLICABILITY: the inability to set confidence limits on
parameters measured in microcosms.
The present study was undertaken to resolve these three problems.
The type of microcosm which formed the basis for discussion in the
state-of-the-art review, and which was described in our first proposal,
was the model food chain, of which the Metcalf microcosm is the best-
known example. Measurements of the bioaccumulation and biodegradation
of chemicals in model food chains can be used to rank a group of related
chemicals according to their food chain mobility, potential for biomag-
nification, and degradability. Good replication (i.e., narrow confi-
dence limits on measured parameters) is necessary for consistent
-------
16
rankings. In principle, Metcalf microcosms are reproducible and stand-
ardizable. Isensee (1976) has measured the replicability of Metcalf-
type microcosms and found them to be comparable to other types of bio-
logical experiments. If Isensee's results are typical, then Metcalf
microcosms satisfy most of the criteria for a screening test: replic-
ability, rapidity, and simplicity. They are not representative of real
ecosystems, but an accurate simulation of nature is not essential in a
screening test. The objective of screening is to identify the most
hazardous substances out of a large group; this is quite different from
understanding the behavior of a chemical in an actual environment.
When a compound is determined (through the screening program) to
be hazardous, it is necessary to arrive at a more detailed prediction
of its fate and effects under natural conditions. To make predictions
concerning a given environment, it is important to simulate that envi-
ronment as closely as possible. Model food chain microcosms differ
from real ecosystems both structurally and functionally. They lack
important ecosystem components, especially detritus and the detrital
food web. They cannot be made complex enough to include the diversity
of transport pathways and opportunities for degradation which would be
found in a natural ecosystem. Recycling of essential elements does not
occur, nor do many other interactions responsible for homeostasis in a
natural ecosystem. Most model food chains are unstable and short-lived,
unlike ecosystems, which are self-regulating and persistent through
time. Because of these differences, model food chains are incapable of
accurately simulating chemical transport and fate in real ecosystems.
-------
17
The three problems identified by Witherspoon et al. (1976), listed
above, do not all apply to one testing system. Reproducibility and
replicability are critical in a screening test, while realism is of
greater concern when detailed predictions are required. The research
reported here centered on the problem of realistic simulation of a
selected aquatic environment, although reproducibility and replicability
were also investigated. We did not re-invent the Metcalf microcosm,
but concentrated instead on maximizing the predictive capabilities of a
more complex, more naturalistic, microcosm design.
3.1.2 Summary of Research Strategy
The immediate goal at the outset of the project was to design a
microcosm which would provide a relatively constant biotic and abiotic
environment in which to examine the behavior of contaminants. Stability,
in the sense of internal constancy, was a prime objective. It was also
deemed essential that the microcosm produce realistic data, i.e., data
consistent with observations in the field. Beyond this, replicability
and reproducibility of results and convenience of operation were desired.
Our first task was to select appropriate microcosm components.
Aquatic organisms were collected from local ponds and streams and tested
for their ability to survive under laboratory conditions. We observed
that natural communities of filamentous algae and their associated biota
were often little affected by transfer from their native habitats to
static aquaria. Spring water was found to be more acceptable to some
organisms than standard reference water. Several potential substrate
types were compared, and two (lake sediment and silica sand) were
selected for further study.
-------
18
We assembled prototype 7-liter (1.85-gal) microcosms containing
spring water and filamentous algal communities, with and without lake
sediment. A third microcosm, with an algal culture (Scenedesmus
abundans) instead of the natural community, was also constructed. The
three systems were placed in a growth chamber and observed for six weeks.
The system containing spring water, sediment, and a natural community
was found to be the most stable, and this design was selected for the
first major experiment.
Experiment 1 (February-April 1976) was intended to examine the eco-
logical properties of the microcosms and to establish basic procedures
for set-up, monitoring, and sampling. Two aquarium sizes [7- and
70-liter (1.85- and 18.5-gal)] and two substrate types (lake sediment
and sand) were compared. After a seven-week equilibration period, a
radiolabeled test compound (sodium arsenate as Na?H AsCL) was added
to the water in subtoxic concentrations, and its distribution was fol-
lowed for five weeks. The microcosms were found to be stable, replic-
able, and persistent, and gave results on arsenic transport which were
consistent with published field data. Size had no significant effect
on the mean values of most parameters measured, but 70-liter microcosms
were more replicable and equilibrated sooner than 7-liter microcosms.
Substrate type, as expected, had a marked influence on water chemistry,
biological productivity, and arsenic transport. Seventy-liter micro-
cosms with lake sediment were the most satisfactory in terms of stabil-
ity, replicability, and longevity, and this design was adopted for
further work. The results of this experiment were summarized in an
open-literature publication (Giddings and Eddlemon 1977).
-------
19
In July 1976, we constructed 12 70-liter microcosms for
Experiment 2. These were similar to the 70-liter sediment microcosms
in the first experiment, except that the sediment was not dried and
homogenized before use, and a macrophyte community was used instead of
an algal community. During the first two months of the experiment we
refined and extended our operating techniques, and gathered data by
which to evaluate replicability of the microcosms. After an equilibrium
had been reached, nine of the microcosms were treated with radiolabeled
sodium arsenate at 0.066 ppm (approximately the same as the first
experiment), at 11.5 ppm, and at 143 ppm. Arsenic distribution was
studied as in the first experiment, and effects of arsenic on produc-
tion, respiration, nutrient concentrations, and community structure
were observed.
Concurrently with the microcosm experiment, a field experiment
(Experiment 3) on the transport and effects of arsenic was conducted in
a shallow pond. Six 60- x 120-cm Plexiglas enclosures were placed in
the pond, each isolating about 700 liters of water and a 5-cm layer of
sediment. Three of the enclosures were treated with 10 ppm sodium
arsenate. The enclosures and the pond were monitored similarly to the
microcosms.
The baseline stability, replicability, and persistence of the
microcosms confirmed the results of the first experiment. Where com-
parisons could be made between the two experiments, the agreement was
excellent, even though quite different communities were involved.
Arsenic transport was found to be concentration-dependent. Effects of
higher arsenic concentrations were seen on the primary production to
-------
20
ecosystem respiration (P/R) ratio (Giddings and Eddlemon 1978), commu-
nity structure, and the levels of some nutrients. Water chemistry in
the control microcosms was very similar to that in the untreated pond
enclosures and in the open pond. Effects of 10 ppm arsenic were similar
in the microcosms and the enclosures, although arsenic transport dif-
fered somewhat between the two systems.
On the basis of our experience to this point, we drafted a tenta-
tive protocol for the use of microcosms to study contaminant fate and
effects. The procedure for constructing a pond microcosm is simple.
All microcosm components are collected from the pond on the day they
are to be used. A 5-cm layer of sediment is placed in an aquarium which
is then filled to the top with pond water. The macrophyte community is
quickly drained and a weighed aliquot is placed in the microcosm.
Except for distilled water to replace evaporation, and new pond water
to replace the samples removed, no further additions to the microcosm
are necessary.
When the protocol was drafted, our analysis of the first experiment
had not yet revealed the effects of size on replicability and equilibra-
tion, and we felt that smaller microcosms would be better suited for
screening purposes. Therefore we recommended 7-liter microcosms in the
protocol. The protocol described a straightforward sampling program
for assessing contaminant transport, and the P/R response was proposed
as an index of ecosystem-level effects.
We tested the protocol using chromium as the test contaminant
(Experiment 4). Sodium chromate was added to 7-liter microcosms in
concentrations of 0.1, 0.25, and 1.0 ppm; this treatment was carried
-------
21
out four times at two-week intervals. The transport of chromium was
studied as in previous experiments, except that no radiotracer was used;
instead, the results were based on stable chromium analysis. Primary
production and ecosystem respiration were measured once or twice weekly.
The experiment proved successful in measuring the sediment-water
exchange and bioaccumulation potential of chromium. However, P/R
results were inconclusive because the microcosms were unstable and
replication was poor. This was thought to be due to disturbances
caused by sampling and P/R measurements, suggesting that 7-liter micro-
cosms are too small for experiments of this type. The small size also
made it difficult to collect sufficiently large samples for chromium
analysis. Seventy-liter microcosms have been used in all subsequent
experiments.
In connection with another project, two additional microcosm
experiments on trace element transport were conducted. In one
(Experiment 6), radiolabeled selenium (sodium selenate) was added to
four microcosms containing fish (to be described below). Experiment 7
examined the transport of mercury in two microcosms without fish. Both
of these experiments preceded very smoothly (one was conducted by an
undergraduate student).
We had originally planned to follow the trace metal experiments
with a test of hexachlorobenzene (HCB), as a representative organic
contaminant. It was later decided to use anthracene (a polycyclic aro-
matic hydrocarbon) instead of HCB, because the environmental transport
and fate of this compound were already under intensive study by another
research group within the Environmental Sciences Division. These
-------
22
researchers (Dr. Stephen Herbes, Mr. George Southworth, and Ms. Linda
Schwall) had measured sorption coefficients, octanol:water partitioning,
volatilization, photolysis, hydrolysis, biodegradation, and bioaccumula-
tion of anthracene under controlled laboratory conditions in one- or
two-component systems (water, water plus sediment, or water plus organ-
ism). These background data would be useful for comparison with the
interpretation of the microcosm results.
The microcosm experiment (Experiment 8) was begun in July 1977,
f 14 1
with the addition of 0.5 ^Ci [9- CJ anthracene to each of two
6-month-old microcosms. The same general procedure was used as in the
trace metal experiments. In this experiment, however, the chemical
fate of the contaminant was studied in detail; extracts of water, sedi-
ment, and organisms were analyzed by thin-layer chromatography and
14
radioautography to determine the amounts of anthracene and C-labeled
derivatives present. We found that anthracene sorbed rapidly to the
sediment, and that anthracene remaining in the water was degraded by
photolysis within 20 days. Anthracene persisted for the length of the
experiment (84 days) in the sediment, where 17% of the original addition
was recovered unchanged. These results were not inconsistent with the
laboratory experiments, but direct comparisons were impossible without
a unified model of transport and fate. Deriving a net degradation rate
for anthracene from the microcosm data was relatively simple. Predict-
ing a net degradation rate by combining single-process data in a mecha-
nistic model is currently beyond our capability.
Explorations of the properties of pond microcosms continued through
the final year of the project. A set of eight replicates was set up in
-------
23
July 1977, following the routine procedure except that small fish
(bluegill, fathead minnow, or mosquitofish) were added to six of the
microcosms (Experiment 5). Water chemistry, biota, and ecosystem
metabolism were monitored for two months, at which time the bluegill
and mosquitofish microcosms were used for Experiment 6. The other
microcosms were studied for an additional 6.5 months. One of the
fathead minnow microcosms was very different from the other fish micro-
cosms and the controls, with turbid water, low pH and dissolved oxygen,
low productivity, and El odea die-offs. It was discovered in early
November that a large crayfish (about 10 cm) had been inadvertently
included in this microcosm. None of the other microcosms exhibited
similar destructive effects. The presence of fish did affect the
benthic and periphyton communities, probably indirectly by eliminating
snails; zooplankton were also sparse in the fish microcosms. Effects
on nutrient concentrations were small, but gross primary production was
over 50% higher in the bluegill and fathead minnow microcosms than in
the others. Most of the fish survived without added food. Snails and
crayfish have a significant impact on microcosm ecology. Our general
conclusion was that fish are compatible with a stable and persistent
microcosm community, although they may affect the community composition
and metabolism.
Concurrently with the third year of the project, aquatic microcosm
research was incorporated into an ongoing ORNL program dealing with the
environmental effects of synthetic fuel production (sponsored by the
Department of Energy). Our standard pond microcosms are now being used
to study (a) the environmental transport and degradation of organic
-------
24
compounds present in wastewater from coal conversion facilities, and
(b) chronic effects of coal conversion effluents on pond ecosystems,
with particular emphasis on the microbial community. In conjunction
with this research, we are continuing to evaluate the microcosms in
terms of ecological similarity to actual ponds, replicability, and
repeatability.
3.2 SUMMARY OF RESULTS
3.2.1. Properties of Microcosms (J. M. Giddings)
3.2.1.1 Basic Ecological Characteristics of Pond Microcosms
Since the ecological properties of pond microcosms change with
time, they are best described in terms of ecosystem development or suc-
cession. Odum (1969) defined ecosystem development by three parameters:
(1) orderly changes in species structure and community processes;
(2) modification of the physical environment by the community; and
(3) culmination in a stabilized ecosystem. These events occur annually
in pond ecosystems, and we have observed them in microcosms as well.
Much of the following summary is drawn from Experiment 9 (see Appendix
A), which was undertaken specifically to observe details of microcosm
succession.
Species Structure
The pond community which is the source of our microcosm inocula is
dominated by El odea (pondweed) and Physa (snail) throughout the year,
and consequently these organisms dominate the microcosms in terms of
biomass and activity. Elodea masses are initially anchored by pushing
-------
25
a few stems into the sediment, leaving most of the plants floating a
few centimeters from the bottom. The floating stems produce dense roots
within a few days, and after a week, most of the roots have begun to
penetrate the sediment. The Elodea mass gradually spreads to cover 75
to 95% of the sediment surface, increasing in height until, after 3 to
6 months, nearly the entire microcosm volume is occupied. Physa,
introduced as part of the Elodea community, begins depositing egg masses
almost immediately, and newly hatched snails soon become abundant. Egg
masses and snails of all sizes are always present in mature microcosms
(except when fish are present - see 3.2.1.5).
Because the microcosms are spatially heterogeneous it is a practi-
cal impossibility to quantify the subdominant and minor components of
the biota. To obtain replicable sampling units, we place artificial
substrates (glass slides and cubes of laboratory sponge) in the micro-
cosms, and remove them for examination after 2 to 14 days. The organ-
isms colonizing these substrates are usually the same as those found in
samples of surface sediment, detrital material, and periphyton. Fila-
mentous blue-green algae (Oscillatoria, Anabaena), which often dominate
the slide communities, form patches on the sediment surface within one
or two weeks after a microcosm is constructed, and remain there unless
disturbed. Other important components of the benthic flora are pennate
diatoms and motile unicellular algae (Cryptomonas, Gymnodinium,
Chlamydomonas, and Euglena). The algal community loses some rarer com-
ponents during the first one or two weeks, but changes little after
this time. From 20 to 30 algal species are normally found on a slide
after three days in a microcosm, regardless of the age of the microcosm;
25 to 35 species are found on slides in the pond.
-------
26
Animal populations in the microcosms are more changeable than the
algae. Oligochaetes and dipteran larvae are usually abundant in the
beginning, accompanied by Hydra, planaria, damselfly larvae, and water
mites. Of these groups, only the oligochaetes persist in large numbers
beyond the first few weeks. Protozoa, especially ciliates, appear early
and become more numerous and diverse through time. Nematodes, rotifers,
and gastrotrichs increase somewhat more slowly, and cladocera and cope-
pods do not become abundant until the microcosms stabilize (about eight
weeks). The animal communities which develop on glass slides in mature
microcosms are very diverse; a typical example from a 10-week-old micro-
cosm included more than 14 genera of protozoans (amoeboid, flagellate,
and ciliate), plus several genera of rotifers, nematodes, gastrotrichs,
and oligochaetes. A list of organisms that have been encountered in
the microcosms is presented in Appendix B.
Community Metabolism
During Experiment 2, we began measuring net community primary pro-
duction (P) and total ecosystem respiration (R) by the three-point
dissolved oxygen method of McConnell (1962). In this procedure, the
dissolved oxygen (D.O.) concentration is measured at the beginning of
one dark period, the next light period, and the following dark period.
The decline in D.O. during the dark period, corrected for diffusion
between water and atmosphere, is a measure of total ecosystem respira-
tion, and the rise in D.O. in the light represents net community primary
production. In measurements made during the first year, we floated a
sheet of Saran Wrap (PVC) on the surface of the water, eliminating dif-
fusion and the need for a diffusion correction. Later, in a series of
-------
27
measurements with the Saran Wrap alternately on and off, we found that
the plastic caused a reduction in primary production, but did not affect
respiration. From differences in nighttime oxygen loss rates with and
without Saran Wrap, we were able to calculate diffusion rates as a
function of D.O. saturation. The empirical relationship derived was:
D = 0.03 (S - 110) ,
where D = oxygen diffusion rate (ppm/12 hr) and S = % saturation. This
relationship was used to correct for diffusion in subsequent experi-
ments, and Saran Wrap was omitted. The diffusion rate was usually less
than 10%, and rarely more than 15%, of the measured rate of change of
oxygen concentration.
The most complete time-series of P and R measurements obtained so
far came from Experiment 9 (Fig. 3.1). Primary production increased
rapidly during the first 40 days, then leveled off abruptly, remaining
within + 10% of a mean value from Day 50 until the termination of the
experiment. Mean equilibrium values for the two replicates in this
experiment were 8.9 and 7.5 ppm/12 hr. In other experiments, microcosms
have shown equilibrium production rates between 4 and 8 ppm/12 hr.
Assuming a photosynthetic quotient of 1.0, this range corresponds to
2
0.5 to 0.9 g C/m /day, which is typical of ponds or shallow lakes
(Odum 1969; Ruttner 1963). We have not collected enough data to deter-
mine the reasons for differences in P between experiments; variation in
temperature or the nutrient content of the sediment may be responsible.
Respiration was initially higher than primary production. The
respiration rate increased gradually, reaching a plateau after about
-------
28
ORNL-DWG 78-16996
CM
O
6
Q.
O.
cr
4
20
40
60 80
TIME (doys)
100
120
in
Figure 3.1. Production and respiration rates
microcosms from Experiment 9.
P = dissolved oxygen increase during
the light period; R = dissolved oxygen
decrease during the dark period. Both
rates are corrected for diffusion, as
explained in the text. SI and S2 are
the two replicates.
-------
29
50 days (Fig. 3.1). The mean equilibrium values for the two replicates
shown were 7.3 and 6.1 ppm/12 hr.
Figure 3.2 shows production/respiration ratios for the two micro-
cosms depicted in Fig. 3.1, plus data from two other experiments. The
ratio was well below 1 at the beginning of the experiment but rose
rapidly. Values of P/R in equilibrated microcosms fell in the range
1.0 to 1.4. It is worth noting that the P/R ratio, an integrative
measure of ecosystem function, is more consistent between experiments,
and between replicates within an experiment, than either P or R alone.
A P/R ratio consistently above 1 implies that more organic matter
is being formed than is being remineralized in the system. The excess
production accumulates in the microcosm as living biomass, detrital
particles, and dissolved organic matter. A steady increase in the
latter has been observed in every experiment where analyses were made
(Fig. 3.3). Accumulation of particulate and dissolved organic matter
during the growing season occurs in ponds and shallow lakes (Allen
1969). The concentration of dissolved organic carbon is the only
chemical parameter we have measured which continues to change after the
first 60 days in a microcosm.
Modification of the Physical Environment
During the course of ecosystem development in pond microcosms,
water chemistry and sediment structure are progressively altered by the
activities of the biota. As a result of rapid plant growth during the
first two months, dissolved oxygen and pH steadily increase, while
concentrations of inorganic nitrogen, phosphorus, and carbon decline.
-------
30
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
ORNL-DWG 78-16997
S1
S2
EXPERIMENT 2
EXPERIMENT 5
20 40 60 80 100 120
TIME (days)
140
160 180
Figure 3.2.
Production/respiration ratios from three microcosm experi-
ments. SI and S2 are the two replicates in Experiment 9.
Values from Experiment 2 are averages from the three control
microcosms. The Experiment 5 line represents control
microcosm Cl. Data for Experiment 2 are results of weekly
measurements; all others are 3-day moving averages.
-------
31
8
ORNL-DWG 78-16995
6 -
o>
5 4
o
o
Q
EXPERIMENT 9/
EXPERIMENT 1
_ A
V
EXPERIMENT 2
1
0
20 40 60 80
TIME (days)
100
120
Figure 3.3. Dissolved organic carbon in microcosm water
from Experiment 1 (mean of three 70-liter
sediment microcosms), Experiment 2 (mean of
three control microcosms), and Experiment 9
(mean of two microcosms).
-------
32
Changes in mid-afternoon D.O. and pH in Experiment 2 control microcosms
are shown in Fig. 3.4. At the beginning of the experiment, respiration
exceeded photosynthesis, which would have resulted in a drop in D.O.;
this was counteracted by diffusion of oxygen from the atmosphere into
the water, because D.O. levels were below saturation for most of the
day. Later, when P/R was greater than 1, the D.O. concentrations were
above saturation, and the excess oxygen produced diffused out of the
water into the air. The relatively constant D.O. concentrations
(measured at the same time each day) in mature microcosms reflect a
balance between photosynthesis, respiration, and diffusion.
Changes in pH during Experiment 2 closely paralleled changes in
D.O. (Fig. 3.4). The increase over the first seven weeks was
accompanied by a decline in dissolved inorganic carbon from 2.4 to
0.4 meq/liter. Most of the inorganic carbon was probably incorporated
into biomass, but some was seen to precipitate (presumably as CaC03)
on the surfaces of macrophyte leaves as the pH rose.
Figure 3.5 shows the concentrations of total P, NH.-N, and NO^-N
in the water of microcosms from Experiment 9. Phosphate-P concentra-
tions are not included in this figure because they were nearly always
below 0.001 mg/liter. Total P varied between 0.004 and 0.02 mg/liter,
with no obvious trends over the first 64 days. Nitrate concentrations
declined during the first few days, rose to a secondary maximum at about
15 days, and then declined to low levels for the rest of the experiment.
Ammonium behaved similarly, except that the secondary peak was spread
over a period of about two weeks. After about 30 days, inorganic N was
scarce in the microcosm water; nevertheless, plant biomass and primary
production continued to increase for several more weeks (Fig. 3.1).
-------
33
O
00
cu
-G
O
in
O
CM
O
Q
I
CC
O
LU
OJ
cu
5
OJ
O
OJ
oo
O
CO CU
cu
CU 3
c
'^- cu
Q.
.Ł=
o
(O 4->
LU C
CU
. cu
CM S~
C «3
CU CU
$- LO
CO
Q. O
X -!->
CD O
-t->
(O TD
2 CU
10 jD
O
O «
O I"
J- OJ
0-P
•i- "3
E O
•i- Q.
CO
x: s-
Q.
CU
TD CU
C S-
CU $-
D^ O
-a cu
cu E
o •—
10 ^
(ludd) N39AXO
-------
34
ORNL-DWG 78-16994
0.04
0 10 20 30 40
TIME (days)
50 60
Figure 3.5. Nitrate, amnonium, and total phosphorus concentrations
in microcosm water, Experiment 9. SI and S2 are the
two replicates.
-------
35
A better picture of plant-nutrient relationships in the microcosms
was obtained by analyzing interstitial water extracted from the sedi-
ment. These samples were taken by means of Ami con hollow fibers buried
2 cm below the sediment surface during construction of the microcosms.
The data are shown in Fig. 3.6. Total P was even lower in the inter-
stitial water than in the overlying water. Interstitial nitrate concen-
trations were initially higher than in the overlying water, but were
reduced to undetectable levels by Day 9. The interstitial water con-
tained high concentrations of NH.-N on Day 2, and the concentrations
increased dramatically to a maximum of 1.4 to 1.9 mg/liter on Day 11.
The source of this ammonium may have been bacterial decomposition of
organic matter or excretion by benthic animals. After Day 11, inter-
stitial ammonium concentrationsdeclined gradually, becoming undetectable
on Day 50, at which time the net primary production rate stopped
increasing (Fig. 3.1).
The observation that inorganic nutrient pools are very low in
mature microcosms implies that nutrient uptake by primary producers
(and possibly bacteria) is equal to nutrient regeneration by animals
and decomposers. This situation seems inconsistent with P/R ratios
greater than 1. If oxygen production in the microcosms is greater than
oxygen consumption, it might be expected that nutrient uptake would be
greater than nutrient regeneration. There are several possible explana-
tions:
(1) The P/R estimates may be in error due to retention of gas in
the internal spaces of the macrophytes (Wetzel 1964). This factor would
-------
36
0.04
ORNL-DWG 78-16993
o>
E
o. 0.02 —
O
0
0.06
0.04
-T 0.02
0
2.0
CP
i
1.0 —
0
0
Figure 3.6.
10 20 30 40
TIME (days)
50 60
Nitrate, anroonium, and total phosphorus concentrations
in microcosm interstitial water, Experiment 9. SI and
S2 are the two replicates.
-------
37
lead to underestimates of both P and R, although its effect on P/R cal-
culations would probably be small once equilibrium was reached. Over-
estimation of diffusion could also introduce a systematic error into
P/R calculations, but long-term constancy of D.O. maxima (Fig. 3.4) and
minima would be compatible with P/R = 1 only if diffusion were zero
(which is unlikely since D.O. concentrations in the microcosms are
usually well above saturation).
(2) The nutrients incorporated into excess organic production may
be drawn from reservoirs in the sediment other than interstitial water.
For example, Elodea roots may be capable of removing phosphate ions
adsorbed to sediment particles.
(3) Nutrients may be regenerated from organic matter faster than
carbon. Thus, more nutrients would become available per molecule of
oxygen consumed than would be assimilated per molecule of oxygen pro-
duced. This has been observed in studies of detrital decomposition in
streams (Kaushik and Hynes 1971, Triska et al. 1975).
(4) Nutrients may be regenerated by anaerobic processes in the
deeper sediment.
Which of these explanations is the correct one cannot be determined
from our data. Perhaps each of these factors contributes to the situa-
tion we observed. In any case, it is certain that a considerable amount
of nutrient recycling occurs in mature microcosms. This subject
deserves greater study.
Progressive stratification of the sediment was mentioned at the
beginning of this section as another indication of biological modifica-
tion of the physical environment. The sediment is nearly homogeneous
-------
38
when a microcosm is first assembled, and vertical differentiation
becomes apparent within a few days. This phenomenon has not been
closely examined in our microcosms, but the development of three dis-
tinct zones can be seen even by casual visual inspection. The upper
zone, 2 to 3 cm thick in a mature microcosm, consists of loose, light-
brown material and is heavily worked by oligochaetes, snails, and other
benthic animals. Beneath this zone is a 1- to 2-cm rust-colored layer,
below which the sediment is dark brown, compact, and probably anaerobic
(the presence of hLS is noticeable). The structuring of the sediment
results from the combined activities of bacteria, benthic fauna, and
primary producers. The existence of microzones within the sediment is
undoubtedly a significant factor in the ecology of the microcosms,
especially nutrient cycling, and also influences the fate of added
contaminants.
Equilibration
In the first stage of microcosm development, nutrient flux is
mainly from inorganic reservoirs into biomass; this is a one-way flux,
and the nutrient cycle is said to be open. The second, or equilibrium
stage, begins when nutrient reservoirs are exhausted (Fig. 3.6) and the
nutrient cycle becomes closed. This progressive closing of the nutrient
cycle is a general feature of ecosystem development (Odum 1969). In a
mature system, nutrient concentrations and flux rates are under biologi-
cal control. Stability (in the sense of internal constancy) is main-
tained by a complex web of interactions between organisms (Reichle et
al. 1975, O'Neill 1976). Relative constancy of conditions in mature
-------
39
microcosms has been described above for primary production and ecosystem
respiration (Fig. 3.1), the P/R ratio (Fig. 3.2), pH and dissolved
oxygen (Fig. 3.4), and nutrient concentrations (Figs. 3.5 and 3.6). Our
microcosms are stable because they are assembled in such a way that many
of the biotic interactions which confer homeostasis are preserved. The
accumulation of dissolved organic matter (Fig. 3.3), and probably also
of particulate organic detritus, indicates that organic production and
remineralization are not completely balanced, even in mature microcosms,
and this may eventually lead to a collapse of stability. However, we
have maintained microcosms for as long as nine months with no sign of
weakened stability, and we expect that experiments lasting more than a
year are possible. This property makes the microcosms suitable for
long-term studies of contaminant degradation, chronic effects, and
patterns of recovery from pollution. Model food chain microcosms
(Metcalf et al. 1971) do not reach an equilibrium, and experiments must
be terminated after about 30 days.
3.2.1.2 Comparison of Microcosms with an Actual Pond
Experiments 2 and 3 were conducted during the summer of 1976 to
compare the ecological characteristics of pond microcosms with condi-
tions in the pond itself and with conditions within enclosures in the
pond. A shallow pond (mean depth approximately 1 m), formerly used for
holding fish, was selected. This pond has a rock bottom overlain by
5 to 10 cm of fine sediment and covered with a dense bed of Elodea.
Twelve 70-liter microcosms were established using sediment and Elodea
-------
40
from the pond (Experiment 2). Six Plexiglas enclosures, each 0.6 x
1.2 m and extending from the sediment to above the water level, were
then placed in the pond (Experiment 3). After 10 weeks, sodium arsenate
was added to nine of the microcosms (three replicates each at 0.066,
11.5, and 143 ppm) and to three of the enclosures (9.2 ppm). Six weeks
later, concentrations of nutrients and major ions were measured in the
microcosms, the enclosures, and the pond.
The final concentrations of the major ions and nutrients in the
pond, the control enclosures, and the treated enclosures are compared
with the control and the 11.5-ppm microcosms in Table 3.1. Differences
between control microcosms and the pond were significant (p < 0.05)
only for Na, NH4, and total P. There were no significant differences
between control microcosms and control enclosures. In both the micro-
cosms and the enclosures, arsenic concentrations in the 10-ppm range
caused rapid disintegration of the Elodea community and replacement by
Spirogyra, accompanied by temporary declines in pH and D.O. The effects
of arsenic treatment on water chemistry (elevated conductivity, alka-
linity, and Na, K, and Ca concentrations) were similar in the enclosures
and in the microcosms, with significant differences found only for Na,
NH4, and P04.
3.2.1.3 Replicability
The replicability of microcosms is one of the principal advantages
of microcosms over field studies. The subject of microcosm replicabil-
ity (similarity between replicates in a single experiment) has been
specifically addressed by Abbott (1966) and by Isensee (1976). Both
-------
41
Table 3.1. Comparison of water chemistry in microcosms, pond
enclosures, and a ponda
Microcosms
Na
K
Ca
Mg
Cl
S04
N03
NH4
P04
Total P
DOC
Alkalinity
Controls
0.03
(0.02)
0.10
(0.0)
11.23
(0.47)
5.13
(0.67)
0.30
(0.04)
3.18
(2.50)
0.006
(0.005)
0.007
(0.002)
0.002
(0.001)
0.006
(0.003)
7.18
(1.39)
49
(4.4)
11.5 ppm
8.25
(1.00)
1.87
(1.86)
19.13
(8.20)
6.17
(1.17)
1.36
(0.63)
2.94
(0.47)
0.007
(0.002)
0.010
(0.004)
0.003
(0.001)
0.029
(0.017)
6.19
(1.26)
117
(24)
Pond enclosures
Controls
1.01
(0.71)
0.17
(0.28)
13.9
(3.2)
4.4
(0.3)
1.25
(0.86)
1.59
(0.62)
0.012
(0.007)
0.011
(0.002)
0.004
(0.002)
0.015
(0.010)
6.97
(2.15)
52
(10)
9.2 ppm
5.89
(0.28)
0.60
(0.11)
19.9
(4.1)
4.7
(0.5)
2.05
(0.44)
2.34
(0.87)
0.008
(0.002)
0.030
(0.011)
0.011
(0.003)
0.013
(0.004)
7.19
(0.62)
75
(13)
Pond
0.87
0.23
11.9
3.4
0.40
3.44
0.006
0.019
0.002
0.014
3.75
43
aAll concentrations in ppm; standard deviations in parentheses.
-------
42
authors concluded that coefficients of variation (C.V.) of most micro-
cosm measurements are about 10 to 30%, comparable to those of typical
laboratory or greenhouse experiments. Our results support this conclu-
sion.
Table 3.2 summarizes the C.V. in thirteen chemical parameters as
determined in Experiments 1 and 2. The values in this table were
derived by calculating the C.V. among 3, 6, or 12 replicates on each
sampling date, and averaging across 1 to 12 dates within a single
experiment. Coefficients of variation averaged less than 20% for total
dissolved solids, alkalinity, Ca, Mg, and Cl. Sodium probably also
belonged in this group—the absolute concentrations in the final samples
from Experiment 2 were very low (0.01-0.05 mg/liter), so the 66% C.V.
may be misleading. The same was true of K concentrations, which were
below 0.01 mg/liter in mature microcosms. Concentrations of SO.,
dissolved organic carbon (DOC), and dissolved nutrients were more vari-
able. The variability of nutrients is not surprising, since these
small, rapidly turning-over reservoirs are highly variable even in a
single microcosm (or other aquatic ecosystem) over short time periods.
Conductivity, pH, and dissolved oxygen, the most frequently mea-
sured parameters in our experiments, have low coefficients of variation.
Data from the 10-week pretreatment period of Experiment 2, when 12
replicates existed, are presented in Fig. 3.7. It can be seen that the
individual microcosms diverged somewhat over the first four weeks, but
then converged again. Convergence was also evident for P0», total P,
and DOC, but not for NO., or NH. (Fig. 3.8). The period of diverg-
ence reflects differences in rates of microcosm succession, while the
-------
43
Table 3.2. Coefficients of variation: Microcosm water chemistry
Variable
NH4
N03
P04
Total P
Na
Ca
Mg
Cl
so4
K
HC03
TDS
DOC
Experiment la
51 (ll)c
73 (11)
63 (11)
41 (11)
7 (12)
9 (12)
5 (12)
10 (12)
63 (12)
86 (12)
6 (7)
6 (12)
-
Experiment 2^
49 (9)
51 (9)
50 (9)
39 (9)
66 (1)
9 (1)
18 (1)
12 (1)
47 (1)
0 (1)
9 (9)
-
35 (9)
aC.V. calculated for three replicates.
bC.V. calculated for six replicates for Na, Ca, Mg, Cl, $04, K;
other variables, 12 replicates.
cValues expressed as percentages; the number of weekly C.V. averaged
is given in parentheses.
-------
44
ORNL-DWG 78-16992
14
O
1>
O
12 -
10
CONDUCTIVITY
8 8
o"
O
OC.
e e
I
Q.
(T
2
^
>
6
10 20 30 40
TIME (days)
50
60
70
Figure 3.7.
Coefficients of variation of conductivity, dissolved
oxygen, and pH, Experiment 2. C.V. expressed as
percentages for twelve replicate microcosms.
-------
45
ORNL-DWG 78-16991
100
90
80
70
60
o
50
40
30
20
0
Figure 3.8.
10 20 30 40
TIME (days)
50
60
Coefficients of variation of
nitrate, ammonium, phosphate, total
phosphorus, and dissolved organic
carbon, Experiment 2. C.V.
expressed as percentages for twelve
replicate microcosms.
-------
46
subsequent convergence implies that the limits to succession are similar
among replicate microcosms. It would be interesting to see whether
disturbances (chemical or otherwise) imposed at the beginning of succes-
sion might increase the divergence, or prevent later convergence, by
interfering with the self-regulating capacity of the microcosm.
It has been suggested (Dudzik et al., 1979) that trace contaminant
experiments in pelagic microcosms should focus on the initial develop-
mental phase. In pond microcosms, this strategy would be expected to
increase the variability of data.
Data from Experiments 2 and 4 permit an assessment of the vari-
ability of metabolic measurements in microcosms. In both studies, rates
of primary production and total community respiration were measured
repeatedly by the diurnal oxygen technique (McConnell 1962) in a set of
replicate microcosms. In Experiment 2, primary production and respira-
tion were measured in 12 70-liter microcosms on three occasions during
the pretreatment period. The average coefficients of variation among
microcosms were 15% for production, 11% for respiration, and 12% for
the P/R ratio. In Experiment 4, eight 7-liter microcosms were monitored
19 times in 11 weeks. Coefficients of variation were 24% (P), 21% (R),
and 18% (P/R). These results are in the same range as those of previous
investigators (Table 3.3).
Coefficients of variation for concentrations of several test con-
taminants in microcosm components are shown in Table 3.4. Except in
the case of snails, all of these values are below 30% and the average
is 19%. (It should be noted that these data are based on analyses made
at the termination of each experiment.) For comparison, Table 3.4 also
-------
47
Table 3.3. Coefficients of variation: Ecosystem metabolism
Experiment 2
Experiment 4
Beyers 1963
Mclntire et al. 1964
Abbott 1966
Production
15
24
15
14
13
Coefficient of variation (%)
Respiration
11
21
14
33
7
P/R
12
18
7
21
8
-------
48
Table 3.4. Coefficients of variation: Contaminant concentrations
Arsenic
Experiment 1
Experiment 2C
Chromium
Low dose
High dose
Selenium
Bluegill
Mosquitofish
Mercury
Anthracene
AVERAGE
(Range)
Isensee 1976
AVERAGEd
(Range)
Water
28b
26
6
6
8
15
26
2
15
(2-28)
5
(1-11)
Sediment
27
6
5
12
1
9
23
12
12
(5-27)
-
-
Algae Elodea
29
21
22
9
11
39
10
19 6
20 17
(9-29) (6-39)
20
(1-59)
Physa i
33
51
10
35
-
-
8
43
30
(8-51)
32
(12-59)
i*
3
3
2
2
2
2
2
2
aNumber of replicate microcosms.
^Values expressed as percentages,
C0.066-ppm treatment.
10 contaminants tested.
-------
49
summarizes Isensee's data from 10 experiments with insecticides in
modified Metcalf microcosms (Isensee 1976). The agreement between the
two types of microcosms is strong, suggesting that system complexity
per se does not affect the variability of microcosm results.
3.2.1.4 Repeatability
One important rationale for defined (or gnotobiotic) microcosms,
such as the Metcalf system, is that an experiment can be repeated at
any time by reassembling microcosms from the same stock cultures. Since
our microcosms are constructed from natural communities, seasonal and
spatial variability within the pond ecosystem might lead to irreproduc-
ibility of experiments. This situation would be a drawback in screening
work, although the ability to reflect natural variability could be
considered an asset when realistic simulation is desired.
A high degree of consistency between experiments has already been
implied in the discussion of ecosystem development (Section 3.2.1.1).
Differences in the species composition of inocula are usually minor,
but even when irregular inocula are used (such as in Experiment 8, where
Elodea were not added directly but allowed to grow from propagules in
the sediment), typical Elodea-Physa communities develop in mature micro-
cosms. In other words, convergence occurs not only within an experiment
(Section 3.2.1.3) but also among experiments. Some examples of inter-
experiment comparisons may be seen in Figs. 3.2 and 3.3.
From the standpoint of contaminant research, reproducibility of
transport data is perhaps more critical than reproducibility of the
test system. We have not normally repeated transport experiments, but
-------
50
a comparison is possible between the 70-liter sediment microcosms of
Experiment 1 and the 0.066 mg/liter arsenic treatment of Experiment 2.
The time course of loss of arsenic from the water was nearly identical
in the two experiments (Fig. 3.9). Final arsenic concentrations in
water, sediment, and snails were not significantly different among
experiments, whereas algae accumulated significantly more arsenic in
the first experiment (Table 3.5). In view of the similarity between
results of these two experiments, it is worth enumerating the differ-
ences in procedure followed in each case. Experiment 1 was begun in
February, Experiment 2 in July. The sediment used in Experiment 1 was
collected from two local reservoirs and consisted mainly of terrigenous
silt (4.9% organic matter), while that in Experiment 2 came from a small
fish pond and was autochthonous in origin (6.8% organic matter).
Experiment 1 sediment was oven-dried, ground with a mortar and pestle,
and homogenized before use, while Experiment 2 sediment was transferred
directly from the pond to the microcosms. The inoculum for Experiment 1
was a filamentous algal community from an outdoor concrete tank, whereas
an Elodea community from the fish pond was used in Experiment 2. In
Experiment 1, adult snails were counted and added to the microcosms
directly; in Experiment 2, they were included as a natural component of
the inoculum. Finally, water temperature was 17 to 19°C in Experiment
1, and 19 to 21°C in Experiment 2. These two sets of microcosms, con-
taining completely different communities collected at different seasons
from different places, gave very similar results on arsenic transport.
We are, therefore, encouraged to believe that results from microcosms
constructed according to our standard procedure will not be greatly
-------
51
ORNL-DWG 77-8288R
0.07
o EXPERIMENT 1
• EXPERIMENT 2
Figure 3.9.
123456
TIME AFTER ARSENIC ADDITION (weeks)
Dissolved arsenic concentrations, Experiments 1 and 2.
Points are weekly averages for three replicate 70-liter
microcosms.
-------
52
Table 3.5. Arsenic concentrations in microcosm components - Comparison
of Experiment 1 and Experiment 2a
Water
Sediment
Algae
Snails
Experiment 1
(ppm)
0.010 ± 0.003
0.529 ± 0.144
15.24 ± 4.38
2.41 ± 0.78
Experiment 2
(ppm)
0.007 ± 0.002
0.522 ± 0.032
2.14 ± 0.45
3.22 ± 1.63
aMean ± standard deviation, three replicate microcosms.
-------
53
dependent on season or source of inoculum. In spite of this convergence
between experiments, however, the most reliable simulations of a given
aquatic environment would be expected from microcosms whose components
come from that environment.
3.2.1.5. Effects of Variations in Experimental
Condition and Microcosm Design
Microcosm Size
In Experiment 1, we compared 7-liter and 70-liter microcosms in
terms of ecological properties and arsenic transport. We found no
significant effect of size on equilibrium dissolved oxygen (Fig. 3.10),
pH (Fig. 3.11), or nutrient levels (except Fe, Table 3.6), nor on the
concentrations of most ions (Table 3.7). Final algal biomass and
2
chlorophyll-a per cm were the same for both sizes (Figs. 3.12 and
3.13). Patterns of arsenic transport were similar in the two sizes,
but the rate of movement of arsenic from water to sediment was higher
in the small systems (Fig. 3.14). This in turn resulted in higher con-
centrations in snails and lower concentrations in algae in the smaller
microcosms (Table 3.8). Effects of size in this experiment were minor,
however, compared to the effects of substrate type (see below).
Baseline stability and equilibration times were related to micro-
cosm size. Coefficients of variation for D.O. and pH in the sediment
microcosms were calculated for overlapping two-week intervals, with the
results shown in Figs. 3.15 and 3.16. Coefficients of variation
declined over time for both variables, indicating progressive stabiliza-
tion of the microcosms. The declines in coefficients of variation
occurred sooner in the 70-liter microcosms than those in the 7-liter
microcosms; that is, equilibration was more rapid in the larger systems.
-------
54
14
12
10
E
Q.
3 8
X
o
O
UJ
O
CO
CO
ORNL-DWG 76-41588
(<7) 70-liter MICROCOSMS
7-liter MICROCOSMS
100
YEARDAY, 1976
110
120
130
140
Figure 3.10. Dissolved oxygen concentrations, Experiment 1. Each line
represents the mean of three replicate microcosms.
-------
55
O)
«o
o
Q.
O)
i-
CJ
s-
4->
tf-
o
(O
(U
0)
a)
CO
O)
Q.
a;
S-
(U
c
-M
a)
Q. O
X O
LU O
S-
•> o
CO
en
X
a.
X
a.
-------
56
Table 3.6. Mean nutrient concentrations: Experiment la
Feb,c
N03-Nb
NH4-N
P04-P
Total P
70- liter
sediment
0.027
0.058
0.012
0.004
0.021
70-liter
sand
0.000
0.014
0.013
0.008
0.017
7-liter
sediment
0.145
0.085
0.079
0.013
0.029
7- liter
sand
0.000
0.014
0.016
0.005
0.018
aAll values are in ppm. Each is a mean for three microcosms, 13
weekly determinations.
^Significant difference between sand and sediment microcosms
(p < 0.05).
cSignificant difference between 70-liter and 7-liter microcosms
(p < 0.05).
Source: Giddings and Eddlemon 1977.
-------
57
Table 3.7. Summary of water chemistry data: Experiment la
Nac
Kc
Cad
Mgd
Clc
so4
HC03
TDSd
Spri ng
water"
0.42
0.53
18.8
9.4
1.10
2.25
84.0
117.0
70- liter
sediment
1.06
0.51
22.52
10.76
1.72
3.78
97.8
138.0
70-liter
sand
0.90
0.46
17.34
8.69
1.68
4.38
75.9
109.0
7-liter
sediment
1.39
0.95
22.39
10.46
3.10
3.85
95.8
138.0
7-liter
sand
1.47
1.55
19.70
10.30
3.26
7.80
85.2
129.0
aAll values are ppm. HC03 expressed as equivalent CaC03. Means
of three replicates, 13 weekly determinations.
^Spring water values are the result of a single analysis at the
beginning of the experiment.
cSignificant difference between 70- and 7-liter microcosms (p <, 0.05).
^Significant difference between sand and sediment microcosms
(p < 0.05).
Source: Giddings and Eddlemon 1977.
-------
58
OJ
E
o
.? 6
0>
o>
E
c/)
en
3
CD
70-liter
SEDIMENT
ORNL-DWG 76-11586
70-liter
SAND
7-liter
SEDIMENT
K
L H
7-liter
SAND
Figure 3.12. Final algal biomass, Experiment 1. Bars represent
replicates identified by letter.
-------
59
ORNL-DWG 76-41587
8
o
CP
X
Q_
O
01
O
_l
X
o
0
70-liter
SEDIMENT
D
70-liter
SAND
7-liter
SEDIMENT
J K L
7-liter
SAND
Figure 3.13. Final chlorophyll-a concentrations, Experiment 1. Bars
represent replicates identified by letter.
-------
60
ORNL-DWG 76-11590
0.06
0.05
70-liter
SEDIMENT
7-liter
SEDIMENT
Figure 3.14.
1234
WEEKS AFTER ARSENIC ADDITION
Arsenic concentrations in water, Experiment 1.
Each point represents the mean of three replicate
microcosms.
-------
61
Table 3.8. Experiment 1 final arsenic concentrations (pptn)
Microcosm
70-
liter
A
B
C
70- liter
7-1
7-1
D
E
F
iter
G
H
I
iter
J
K
L
sediment
0
0
0
sand
0
0
0
sediment
0
0
0
sand
0
0
0
Water
.0108
.0125
.0070
.0495
.0456
.0481
.0033
.0014
.0035
.0310
.0367
.0293
Sub-
strate
0
0
0
0
0
0
0
0
0
0
0
0
.398
.507
.683
.174
.133
.180
.511
.326
.276
.120
.085
.106
Algae
10.47
19.09
16.17
33.81
39.42
65.74
6.54
12.87
6.00
11.58
9.28
14.26
B.R.a
969
1527
2310
683
846
1367
1982
9193
1714
374
253
487
Snails
1
2
3
0
0
0
1
1
1
0
0
0
.77
.17
.28
.435
.559
.576
.53
.44
.24
.366
.550
.788
Zoo-
B.R.a plankton
164 6.76
174
469
8.8 8.33
12.3
12.0 10.4
464
1029
354
11.8
15.0
26.9
B.R.a
626
—
—
168
__
216
—
—
—
—
--
—
aB.R. = bioaccumulation ratio = concentration in organism/concentration in
water. All concentrations are yg As/g dry wt, except water = mg As/liter.
-------
62
ORNL-DWG 77-8285R
20
z
CC Lul
§ i
O LJ
UJ O
Lu
Li_ Lu
LU O
O ^_
10
0
0
Figure 3.15.
10 20 30 40 50
TIME (days)
60
70
80
Equilibration of dissolved oxygen in 7- and 70-liter
sediment microcosms, Experiment 1. Coefficients of
variation among measurements made over a two-week period
were calculated for each individual microcosm for 11
overlapping two-week periods. The points shown here are
average C.V. from three replicate microcosms.
-------
63
ORNL-DWG 77-8286R
O
LJ_ CO
o < 2
O CL
U_ Li_
U. O
UJ _
o
70-liter
0
Figure 3.16.
10 20
30 40 50
TIME (days)
60 70
80
Equilibration of pH in 7- and 70-liter sediment microcosms,
Experiment 1. C.V. calculated as in Fig. 3.15.
-------
64
Variation between replicates was lower in the 70-liter microcosms
than that in the 7-liter microcosms on the whole. Coefficients of
variation among replicate 70-liter and 7-liter sediment microcosms were
calculated for each of the 13 weekly water chemistry samples; the aver-
ages of these coefficients of variation are shown in Table 3.9. For
total dissolved solids, alkalinity, Ca, Mg, Cl, P04, and total P,
70-liter microcosms showed better replicability than 7-liter microcosms,
whereas the reverse was true for K, S0», NO.,, and NH.. Variabil-
ity of arsenic concentrations was generally greater in the small micro-
cosms (Table 3.10).
Smaller microcosms have some obvious advantages over larger ones:
more replicates can be maintained in a given space; less sediment,
water, and biota are required; and the systems are movable. Small size
also imposes some limitations, however, such as smaller sample sizes,
and greater tendency to be disturbed by sampling and monitoring. For
long-term transport studies, 70-liter microcosms are preferable, but
7-liter microcosms might be useful in experiments where repeated sampl-
ing is not necessary.
Substrate
We have used four different substrates in our microcosms: Ottawa
sand (Experiment 1), sediment from local reservoirs (Experiments 1
and 4), sediment from a small fish-holding pond (Experiment 2), and
sediment from spring-fed Target Range Pond (Experiments 5-9). The eco-
logical characteristics of microcosms with sediment from the reservoirs
-------
65
Table 3.9. Mean coefficients of variation (for 3 replicates) of
chemical parameters in 7- and 70-liter microcosms
Variable
TDS
Alkalinity
Na+
Ca^
Mg^
cr
so;
K+
ml
N03
po4E
Total P
7-liter
14.73
13.61
5.61
16.59
11.73
21.59
32.29
57.68
43.08
58.01
83.69
50.04
70-liter
6.18b
5.59
7.07
8.60
4.90
10.47
63.19
85.60
51.34
72.62
62.94
40.65
na
12
7
12
12
12
12
12
12
11
11
11
8
aNumber of weekly C.V. averaged.
^Values expressed as percentages.
-------
56
Table 3.10. Coefficients of variation for arsenic concentrations (%)
in Experiment 1
70- liter
70-liter
7- liter
7- liter
Sediment
Sand
Water
sediment 28
sand 4
sediment 42
sand 12
Average coefficients
Abiotic components
32
12
Substrate
27
16
33
17
Algae
29
37
45
21
Snails
33
15
11
37
of variation (%)
Biotic
components
29
27
Overall
31
20
-------
67
and the ponds are similar (see Sections 3.2.1.1, 3.2.1.4), but micro-
cosms with sand behaved much differently in the one experiment in which
they were used. Differences in water chemistry between sand microcosms
and reservoir sediment microcosms are shown in Tables 3.6 and 3.7.
Microcosms with sand had far less algal growth than those containing
sediment (Figs. 3.12 and 3.13), a difference reflected in dissolved
oxygen (Fig. 3.10) and pH (Fig. 3.11). Apparently, sand does not supply
enough nutrients to maintain closed microcosms in a stable condition;
it is doubtful that the sand microcosms could have persisted for more
than a few weeks beyond the termination of this experiment.
Arsenic transport in sand microcosms was affected by the
low sorptive capacity of the sand (Fig. 3.14 and Table 3.8,
Section 3.2.2.1). One presumed advantage of using a standard substrate
material such as sand is improved replicability; Table 3.10 shows that
this improvement is limited to the abiotic components of the microcosms.
Light, Temperature, and Season
We did not examine the effects of light and temperature on the
microcosms. In all experiments, a bank of cool-white fluorescent tubes
provided 16,000 lux on a 12-hr-light:12-hr-dark cycle. Most experiments
were carried out at 16 to 18°C. Experiment 2 was at a slightly higher
temperature, and in Experiment 4 (and part of Experiment 8) the environ-
mental chamber was set at 8°C, the ambient water temperature in Target
Range Pond at the time the experiment began (Table 3.11). Since tem-
perature was only one of many variables which differed between experi-
ments, we cannot determine the effects of temperature alone. The season
-------
68
Table 3.11. Temperatures of microcosm experiments
Temperature
Experiment 1 17-19°C
Experiment 2 19-21°C
Experiment 4 8-9°C
Experiments 5-7 16-18°C
Experiment 8 (1/25-6/9/77) 8-9°C
(6/9-6/30) 12-14°C
(6/30-10/6) 16-18°C
Experiment 9 16-18°C
-------
69
in which a microcosm is assembled probably influences the course of
ecosystem development, but the properties of mature pond microcosms
converge under our usual experimental conditions (Sections 3.2.1.3,
3.2.1.4).
Macrophytes vs Algae
Although most of our microcosms were assembled with macrophyte
communities, several experiments (1, 4, and 8) began with filamentous
algal communities. In all microcosms containing untreated pond sedi-
ment, the mature communities were dominated by macrophytes with filamen-
tous algae as subdominants; this occurred even in Experiment 8, where
macrophytes were introduced only as propagules in the sediment. It is
apparent from Figs. 3.1 and 3.6 that macrophytes utilize nutrients
contained in the sediment, which is probably the reason they normally
outcompete the algae. We have not performed any experiments directly
comparing microcosms with macrophytes and microcosms with algae,
although the data from Experiments 1 and 2 indicated that water chem-
istry and arsenic transport were similar regardless of the autotrophs
present (Section 3.2.1.4.).
Fish
Eight 70-liter microcosms were established in July 1977 to examine
the effects of including fish in the microcosm community (Experiment 5).
The microcosms were constructed following our regular procedure, using
water, sediment, and an Elodea community from a shallow pond. Fish
were then added to the microcosms as follows:
-------
70
2 microcosms: 10 mosquitofish (Gambusia affinls)
2 microcosms: 5 fathead minnows (Pimephales promelas)
2 microcosms: 2 bluegill (Lepomis macrochirus)
2 microcosms: no fish added (controls)
One of the fathead minnow microcosms became very different from
the other fish microcosms and the controls, with turbid water, low pH
and dissolved oxygen, low productivity, and Elodea die-offs. It was
discovered in early November that a large crayfish (about 10 cm in
length) had been inadvertently included in this microcosm. None of the
other microcosms exhibited similar destructive effects. The other
fathead minnow microcosm and the bluegill microcosms were devoid of
snails, however, presumably due to predation by the fish. These three
microcosms differed from the mosquitofish and control microcosms in
that they had extensive blue-green algal growth on the sediment surface,
dense periphyton on the aquarium walls, a thin algal mat floating on
the surface, and greater growth of Elodea. It seems likely that the
absence of snails (which are mainly herbivorous) rather than the
presence of fish was responsible for these community effects. Zooplank-
ton populations were dense in the control microcosms (dominated by the
cladoceran Simocephalus and a cyclopoid copepod), but very sparse in
the fish microcosms.
Gross primary production, as measured by the diurnal oxygen tech-
nique, averaged 7.3 ppm O^/day in the control microcosms, 7.9 with
mosquitofish, and 12.0 with bluegill. The fathead minnow microcosm
without crayfish averaged 12.1; with the crayfish, the gross primary
-------
71
production was only 5.8 (Table 3.12). There were only small differences
in nutrient levels among treatments, the most noticeable being elevated
ammonium concentrations in all fish microcosms (except with crayfish),
and somewhat more total phosphorus in the bluegill microcosms than the
others (Table 3.13). The greater primary production in bluegill and
fathead minnow microcosms was probably attributable to the reduced
grazing pressure (due to absence of snails), but the elevated nutrient
levels may also have been important.
At the end of Experiment 6, both bluegills were recovered from
microcosm Bl, but only one remained in B2; mean individual weights had
declined by 31 to 35% over the 18-week period. Microcosms Ml and M2
contained 5 and 9 Gambusia, respectively. The mean individual weight
had increased by 65% in Ml and 25% in M2. We do not have data on the
fathead minnows, but live fish were still present after eight months.
We conclude that fish are compatible with a stable and persistent
microcosm commmunity, although they may affect the community composition
and metabolism. Snails and crayfish, on the other hand, have a signifi-
cant impact on microcosm ecology.
3.2.2 Transport Studies (G. K. Eddlemon and J. M. Biddings)
3.2.2.1 Arsenic
Experiment 1
Arsenic (As) in the form of sodium arsenate, was selected as a
test contaminant for our initial experiments on microcosm design. The
availability of a radioisotope, arsenic-74 (half-life 17.5 days), makes
-------
72
Table 3.12. Gross primary production in fish microcosms
Microcosm
Cl
C2
Bl
B2
Ml
M2
Fl
F2
n
28
25
10
11
12
12
27
25
G.P.P.a
5.8 (0.1)
8.7 (0.2)
12.4 (1.4)
11.6 (1.8)
6.8 (0.7)
9.0 (0.9)
12.1 (1.2)
5.8 (0.6)
aGross primary production in ppm 02/12 hr; mean of n
measurements on equilibrated microcosms; standard
deviation in parentheses.
-------
73
Table 3.13. Water chemistry in fish microcosms after 12 weeks (all values are
in ppm)
Microcosm
Na
K
Ca
Mg
so4
Cl
Alkalinity
DOC
N03
NH4
P04
EP
Fe
0
0
21
7
32
1
35
6
0
0
0
0
< 0
Cl
.49
.03
.2
.5
.7
.2
.78
.003
.015
.001
.016
.10
C2
0.60
0.04
21.4
7.1
47.2
1.5
35
6.99
0.005
0.020
0.001
0.006
< 0.10
Bl
0.46
0.04
20.2
5.6
27.5
1.3
37
7.61
0.005
0.104
0.003
0.018
< 0.10
0
0
21
6
37
1
36
8
0
0
0
0
< 0
B2
.47
.04
.4
.0
.2
.1
.10
.010
.040
.003
.032
.10
0
0
20
6
33
1
38
6
0
0
0
0
< 0
Ml
.88
.04
.2
.9
.8
.5
.08
.005
.074
.002
.010
.10
M2
0.76
0.04
20.6
5.9
32.2
1.1
37
8.32
0.005
0.075
0.003
0.016
< 0.10
Fl
1.24
0.04
23.4
7.8
34.5
1.6
42
6.79
0.011
0.071
0.009
0.012
< 0.10
F2
0.66
0.05
23.0
7.0
28.5
1.6
52
9.20
0.004
0.012
0.005
0.014
< 0.10
-------
74
arsenic a convenient subject for studies of environmental transport.
Arsenic compounds occur widely in herbicides, pesticides, industrial
waste, and fossil fuels. Reduced arsenic, arsenite, has been applied
to many lakes to control the growth of aquatic weeds. A recent review
of the aquatic arsenic cycle indicates that "the fate of arsenic in
natural waters has received little attention" (Ferguson and Gavis 1972).
Methods. Six 7-liter and six 70-liter freshwater microcosms were
assembled in si ate-bottomed, glass aquaria, three of each size receiving
a sand substrate, and the others a reservoir-sediment substrate. The
12 microcosms were kept in an environmental chamber at 18°C. A bank of
cool-white fluorescent lights provided 16,000 lux on a 12-hr-light:
12-hr-dark cycle. Appendix A provides details of microcosm construction
for all experiments.
Seven weeks after the microcosms were assembled, the long-term
constancy and reduced daily variation in pH, dissolved oxygen, and con-
ductivity indicated that a condition of relative stability or equilib-
rium had been achieved in each microcosm, at least in terms of these
parameters. At this time, a test contaminant, sodium arsenate (N^HAsO.)
was introduced. In this first experiment, we were primarily interested
in the behavior of undisturbed microcosms and the contaminant within
the undisturbed systems. Therefore, a relatively low concentration of
0.05 ppm of As was selected. The arsenate was labeled with H_'^AsO^
(20.4 ftCi/mg As) obtained from Amersham/Searle Corporation. A gamma
spectrometer was used to count the radioarsenic activity in 2-, 5-, or
10-ml water samples removed weekly from each microcosm.
-------
75
At the conclusion of the experiment, five weeks after arsenic
addition, the nonbenthic filamentous algae were removed by hand from
each microcosm, drained in a sieve, and weighed.
Algal subsamples were dried (55°C, 6 days) and the dry weight
obtained. The dried algae were then placed into sample tubes for
radioarsenic determination.
Following the removal of algae from the microcosms, two 100-ml
samples from each microcosm were vacuum-filtered (Whatman GF/C) and the
radioactivity of the seston on the filters was counted. Zooplankton
were collected from the three microcosms containing them in sufficient
numbers by straining the water through a 0.150-mm sieve. The zooplank-
ton were washed into beakers with distilled water, filtered onto pre-
weighed paper, dried (5 days at 55°C), weighed, and their radioactivity
determined. All snails over 2 mm in length were removed from the
aquaria, placed in spring water for about 1.5 hr, rinsed with distilled
water, weighed, dried, reweighed, and their radioactivity counted.
Samples of the substrate from the completely drained microcosms
2
were removed with cork borers. Three 2.21-cm cores were placed in
preweighed aluminum dishes, dried (6 days at 55°C) and reweighed. The
cores were then ashed (24 hr at 550°C) to determine the organic content.
2
Two 0.68-cm cores from each microcosm were assayed for radioactivity.
Results. The concentrations of arsenic in the water are shown in
Fig. 3.14. The decline in dissolved arsenic concentrations over the
first 7 to 21 days of the experiment may be described by a simple model
of logarithmic decay,
Nt = N0 e-rt , (1)
-------
76
in which N. = concentration at the time t, N = initial concentra-
tion, and r is the first-order kinetic constant. Values of r from this
and subsequent experiments are presented in Table 3.14. The rate con-
stant for arsenic loss from water in the 70-liter sediment microcosms,
0.10/day, compared favorably with the 0.12/day calculated for similar
arsenic microcosms in Experiment 2. Of the five contaminants tested
under similar conditions, only mercury exibited a higher loss rate con-
stant (0.23/day over the first three days). Loss rate constants in
7-liter sediment microcosms averaged 0.25/day, reflecting the greater
sediment surface to water volume ratio of these microcosms. In the
7-liter sand microcosms, the loss rate constants averaged only
0.018/day, due to the low affinity of sand for arsenic.
Dissolved arsenic concentrations over the entire experiment in the
sediment and the 7-liter sand microcosms followed the equation
"t X+M " X+M '
where \= fraction lost from the water per unit time, and n= fraction
lost from the substrate per unit time (Hayes et al. 1952). The esti-
mated values of X and \L are shown in Table 3.15. It was not possible
to fit this model to the 70-liter sand microcosms, which did not
approach equilibrium before the experiment was terminated. Equation (2)
was also inadequate for describing data from later experiments, leading
us to adopt Equation (1) for comparative purposes.
Virtually all of the arsenic lost from the water was found in the
substrate at the conclusion of the experiment (Fig. 3.17). Only 2 to 4%
-------
77
Table 3.14. Sumnary of kinetic loss rate constants for microcosm water
Contaminant
Arsenic
Selenium
Chromi urn
Mercury
Anthracene
First-order
constant, r
0
0
0
0
0
0
0
0
0
0
.0
0
0
.10
.12
.018
.25
.034
.026
.053
.044
.030
.227
.091
.055
.007
day"1
day1
day1
day1
day1
Initial
Interval concentration, N
Days 1-7
1-24
1-21
1-7
1-25
Days 1-40
Days 1-15
1-15
1-15
Days 0-3
10-41
Days 0-15
20-65
0
0
0
0
9
0
0
0
3
0
0
0
0
.05
ppm
.066 ppm
.05
.05
.2
.05
ppm
ppm
ppm
ppm
.025 ppm
.68
.20
.4
.1
.5
.2
ppm
ppm
ppb
ppb
ppb
ppb
Expt.
Expt.
Expt.
Expt.
Expt.
Expt.
Expt.
Expt.
Expt.
Expt.
Expt.
Expt.
Expt.
1,
2,
1,
1,
3,
6,
4,
4,
4,
7,
7,
8,
8,
Remarks
70- liter
70- liter
7-liter
7- liter
sediment microcosms
sediment microcosms
sand microcosms
sediment microcosms
500-liter pond enclosure
70-liter
7-liter
7-liter
7-liter
70-liter
70-liter
70- liter
70- liter
sediment microcosms
sediment microcosms
sediment microcosms
sediment microcosms
sediment microcosms
sediment microcosms
sediment microcosms
sediment microcosms
-------
78
Table 3.15. Adsorption and desorption parameters for
arsenic in microcosms (calculated from
equation 1)
70- liter
sediment
70- liter
sand
7- liter
sediment
7- liter
sand
A 0.83 ~ 1.57 0.19
y 0.18 -- 0.09 0.23
-------
79
cu
O)
Q.
X
10
N-
O
$
Q
I
o:
o
1
I
ID Q
Ł U
*- ^
LJ
CO
Q
2
CO
I-
I Q
O UJ
N- CO
O)
o
Q.
I
O
tn
o
o
o
s-
o
o
o o
•i- O
c s-
O) O
10 T-
t. E
fO
O)
«f- C
o o
1/1 (/)
C -(->
o c
•r- OJ
+J to
3 S-
10
c o
O
O
O
CO
O
CD
O
OJ
DINBSdV 1V101 JO !N3Dd3d
-------
80
was found in the algae, and amounts in snails, zooplankton, and seston
were negligible fractions of the total.
The final concentrations of arsenic in all microcosm components
are presented in Table 3.8 Included in this table are bioaccumulation
ratios (dry weight concentration in organism/concentration in water)
for the biota. A lower substrate surface to water volume ratio and the
relative inactivity of sand with regard to arsenic resulted in higher
concentrations of arsenic in the water of the large microcosms and sand
microcosms, respectively, than in the others. As a result, more arsenic
was available for uptake by algae. Algal arsenic concentrations were
significantly greater in the 70-liter microcosms and in the sand
microcosms (P < 0.05) than in the 7-liter and sediment microcosms.
Bioaccumulation ratios ranged from 253 to 9193 (mean 1810). Arsenic
concentrations in snails were directly proportional to substrate arsenic
concentration (r = 0.947; Fig. 3.18). If the major route of entry of
arsenic into snails is via the substrate, as implied by these data, the
calculated bioaccumulation ratios are meaningless.
Discussion. Table 3.16 presents published values for arsenic
concentrations and bioaccumulation ratios of organisms from marine,
freshwater, and laboratory habitats. These concentrations and bioaccu-
mulation ratios are in substantial agreement with the results of the
present experiment, despite the fact that the values cover a wide range
of ambient arsenic concentrations and include a variety of chemical
forms. Our data support the conclusion that arsenic, unlike mercury,
does not accumulate in the food chain (Ferguson and Gavis 1972, Isensee
et al. 1973). Within each group of organisms, the range of reported
-------
81
ORNL-DWG 76-16838
E
Q.
Q.
LJ
CO
o:
CO
0
0
0.2 0.4 0.6
SEDIMENT ARSENIC (ppm)
0.8
Figure 3.18. Regression of snail arsenic concentrations (yg As/g
dry wt) on substrate arsenic concentrations,
Experiment 1. Each point represents one microcosm.
The line fitted to the data by least-squares analysis
has the equation y = 4.251x - 0.0141 (r = 0.9474).
-------
82
Table 3.16. Arsenic concentrations and bioaccumulation ratios from the
literature
Organism
Bioaccumulation
As(ppm) ratio
Source
Plants
Oedogonium
Lake Michigan phy top lank ton
Lake Superior phytoplankton
Pond aufwuchs
Pond seston
Chara globularis
Myriophyllum
Potamogeton praelongus
Fucus
Ascophyllum nodosum
Macrocystis pyrifera
Filamentous algae
Molluscs
Physa
Littorina littorea
Mytilus edulis
Physa
Crustaceans
Daphnia
Lake Michigan zooplankton
Lake Superior zooplankton
Homarus americanus
Zooplankton
4.5-71.4
5-10
3.2-4.3
11-250
16-250
5-122
15-63
0-15
12-17
10-17
72-144
6-66
1.4-2.6
4-15
1.6-5.4
0.3-3
5-16
4-8
1.2-4.4
4-8
7-10
1248-1635
1750-12,800
—
6000-122,000
7000-80,000
200-250
—
—
3900
3900
Isensee et al.
Seydel 1972
Seydel 1972
Sohacki 1968
Sohacki 1968
Sohacki 1968
Sohacki 1968
Sohacki 1968
Penrose et al.
Penrose et al.
1973
1975
1975
— Boothe and Knauer
250-9200
299-446
3011
1205
8-1000
736-2175
1833-4500
—
818
170-630
1972
This study
Isensee et al.
Penrose et al.
Penrose et al.
This study
Isensee et al.
Seydel 1972
Seydel 1972
Penrose et al.
This study
1973
1975
1975
1973
1975
-------
83
concentrations is smaller than the range of bioaccumulation ratios; our
microcosms, therefore, may be able to predict the former more accurately
than the latter.
Particularly in the lake sediment microcosms, the substrate acted
as the major arsenic sink. This appears to occur in nature as well.
Nearly all of the arsenic entering Lake Michigan accumulates in the
sediments (Seydel 1972). Of the arsenic brought to the surface from
geothermal sources at Wairekei, New Zealand, only 5% remains in dis-
solved form; the remainder is rapidly precipitated on the mud (Ritchie
1961). Penrose et al. (1975) examined the arsenic distribution in
Moreton's Harbor, Newfoundland, where arsenic-containing leachate is
released from a stibnite mine. Within 1 km of the mine drainage out-
flow, the sediments contained 9 to 30 ppm arsenic, while the concentra-
tions in the water were 0.2 to 2.6 ppb. Sohacki (1968) administered
sodium arsenite to a small pond, and found that 60 to 70% went into the
sediments in 10 days with over 90% lost after 30 days. Seydel (1972)
found that the highest sediment arsenic concentrations in Lake Michigan
occurred in clayey sediments (26 to 29 ppm), while the lowest occurred
in sandy sediments (7.2 to 7.6 ppm). Ball and Hooper (1966) observed
that sodium arsenite was taken up much faster (in aquaria) by mud sub-
strates than by sand or gravel. In the pond studied by Sohacki (1968),
areas of the bottom covered with silt adsorbed significantly more
arsenic than sandy areas.
These studies all corroborate the major result of our experiment,
namely, that substrate type is critical in determining the movement of
arsenic into the sediments. According to Seydel (1972), the movement
-------
84
of arsenic into the sediments is due to (1) coprecipitation with iron,
(2) adsorption by sediment particles, and (3) adsorption by organic
matter. The presence of dissolved iron in the sediment microcosms and
its absence from the sand microcosms have already been mentioned.
Furthermore, the lake sediment contained 5% organic matter at the end
of the experiment, while the sand contained only 0.1% organic matter.
These facts, together with the smaller particle size (and hence greater
adsorptive surface area) of the sediment, explain the differences in
substrate arsenic uptake observed (Fig. 3.17).
Experiments 2 and 3
The objectives of Experiments 2 and 3 were, in part: (1) to
examine the effects of dosage level on transport and distribution of
arsenic; (2) to repeat part of Experiment 1 for comparison of results
(reproducibility); and (3) to compare some aspects of arsenic transport
and distribution under semi-natural conditions (Experiment 3 pond
enclosures) with microcosm results.
Materials and methods. Twelve microcosms were established in
70-liter glass aquaria in July 1976. Each microcosm contained 6 cm
(about 10 kg) of untreated sediment from a shallow pond, 40 g (wet wt)
of a mixed Elodea - Potamogeton community from the same pond, and
56 liters of spring water. A diverse assemblage of animals was present
in the sediment and the macrophyte inoculum, including protozoans,
rotifers, copepods, cladocerans, oligochaetes, nematodes, snails, and
insects. The microcosms were maintained in an environmental chamber at
18°C on a 12-hr-light:12-hr-dark cycle. Details of the microcosm
technique may be found in Appendix A.
-------
85
Ten weeks after the microcosms were established, sodium arsenate
was added to nine of them. Appropriate volumes of a sodium arsenate
74
stock solution and a solution of carrier-free hL AsO, were mixed
with 100 ml of water from each microcosm. The mixture was slowly intro-
duced into each microcosm through a glass tube so that the mixture
entered in a horizontal stream approximately 10 cm below the water sur-
face. Arsenic concentrations in the sodium arsenate stock solution and
in untreated pond sediment, water, and Elodea were measured by an arc-
emission technique (Feldman 1977) and used to calculate specific activ-
ities of As added at each treatment level. Arsenic concentrations in
various components of the microcosms were then determined by gamma
spectroscopy. The initial As concentrations in water were 0.066, 11.5,
and 143 ppm, with three replicates per treatment. These concentrations
bracket the range of reported toxic concentrations for most freshwater
organisms (Becker and Thatcher 1973, Cushman et al. 1977). Three micro-
cosms were left untreated as controls.
At the same time that the microcosms were inoculated with As, three
of six 500-liter pond enclosures described in Appendix A, Experiment 3,
also received As at an initial dose of 9.2 ppm As in water.
Water, sediment, and macrophyte samples were regularly removed
from all microcosms for radioassay: water by pipette, sediments by
glass tube corers, and macrophytes by forceps. Sediment and macrophyte
samples were freeze-dried and weighed prior to radioassay. Final sedi-
2
ment samples were extracted using a cork borer of 1.7-cm area after
all standing water had been removed to facilitate removal of whole,
intact, replicable cores. Cores were sectioned at 1.0-cm intervals to
-------
86
determine arsenic distribution by depth. Less frequently we applied
similar sampling techniques to the outdoor pond enclosures.
Results. As determined by an arc-emission technique (Feldman
1977), background As levels in pond sediments, water, and Elodea used
in Experiment 2 averaged 4.3 ppm, 0.75 ppb, and 2 ppm, respectively.
Snails (Physa), Potamogeton, and Spirogyra had less than 3 ppm. The
sediment and water background levels compare favorably with the lower
end of concentration ranges reported for U.S. rivers and lakes. Elodea
canadenis from a New Zealand geothermal area had 3 ppm As; Potamogeton
less than 6 ppm (NAS 1977). Experiment 3 pond enclosures were placed
in the source pond for Experiment 2 microcosm components and therefore
the same concentrations apply, except that pond water was slightly lower
in As at 0.57 ppm.
Figures 3.19 and 3.20 show that dissolved As concentrations
declined over time in all treatments. The 0.066-ppm (A, B, C) and
11.5-ppm (D, E, F) microcosms and the 9.2-ppm (V, W, Y) pond enclosures
reached mean equilibrium concentrations of 0.007, 3.9, and 4.1 ppm,
respectively, within four weeks. Arsenic concentrations in water of
the 143-ppm microcosms (6, H, I) failed to equilibrate by the end of
the experiment; final concentrations averaged 87 ppm.
As in Experiment 1 and subsequent experiments with selenium,
mercury, and anthracene, and probably chromium, the sediments were the
major As sink. Table 3.17 presents the final arsenic distribution due
to added As for all Experiment 2 microcosms and Experiment 3 pond
enclosures. Arsenic concentrations in the top centimeter of sediment
-------
87
ORNL-DWG 78-<87<4
0.10 -
z
LJ
O
Z
O
O
0.08 -
0.06
0.04 -
0.02 -
15
20 25
TIME (days)
30
35
40
45
Figure 3.19. Dissolved arsenic concentrations, Experiment 2. Lower
graph: 0.066-ppm treatment. Upper graph: 11.5-ppm
treatment. Each point represents one sampling date for
one microcosm.
-------
88
ORNL-DWG 78-18713
20 25
TIME (doys)
30
35
40
45
Figure 3.20.
Dissolved arsenic concentrations, Experiments 2 and 3.
Lower graph: 143-ppm treatment, Experiment 2. Upper
graph: pond enclosures, Experiment 3. Each point
represents one sampling date for one microcosm or
enclosure.
-------
89
i B
p- OJ
OOOO OOOO !Ł><Ł> -—i
OOOO
O O O
OOOO
kO •* Tj- CO
c_ g
-------
90
averaged 1.1 ppm in the 0.066-ppm treated microcosms (A, B, and C),
246 ppm in the 11.5-ppm microcosms (D, E, and F), and 1410 ppm in the
143-ppm microcosms (G, H, I). Sediments of the 9.2-ppm treated pond
enclosures yielded only 25 ppm As, about eight times lower than might
have been predicted on the basis of the 11.5-ppm microcosm results.
Although this large difference can be explained in part by the approxi-
mately 2.5 times greater sediment surface to water volume ratio of the
microcosms, only tentative explanations can be offered for the
remainder:
(1) Uptake by periphyton. Periphytic growth on the Plexiglas walls
of the pond enclosures was much greater than that on the glass walls of
the microcosms. Further, periphyton yielded the highest concentrations
of As measured in any of the 11.5-ppm microcosm components other than
surface scum.
(2) Stratification of pond enclosure water. Temperature measure-
ments of enclosure water indicated some stratification which conceivably
could have reduced contact of As with sediments.
(3) Sediment sampling errors due to difficulties in extracting
replicable and representative cores.
(4) Possible leaks at enclosure seams and/or bottom.
Sediment As depth profiles are shown in Fig. 3.21. Based on core
sections, the top centimeter of sediment contained 63.3% of all sediment
As introduced into 0.066-ppm microcosms, 71.9% of all sediment As
introduced into 11.5-ppm microcosms, and 69.2% of that introduced into
143-ppm microcosms. The top 2 cm of sediment in the first two treat-
ments contained about 90% of all sediment As. Approximately 85% resided
in the top 2 cm of sediment in the 143-ppm microcosms.
-------
91
ORNL-DWG 78-18716
100
200
500 1000
ARSENIC (ppm)
u
1
1
4
e c
o n
<— U
Q_
UJ o
o 2
u „
S 4
Q *
\ \
0.066 ppm TRTMT
1 1
) 0.5 1.0 1.
1 1
11 C •> «. _ TC3»*T
1.5 ppm TRMT
r
5
300
u
4
R
1
1 1
4/1^ nnm TPMT
icrO ppm 1 lAlYI 1
1 1
1500
Figure 3.21. Sediment arsenic profiles, Experiment 2. Each
line represents the mean of three replicate
microcosms. Cores were sectioned into 1-cm
intervals; points are plotted at the midpoint
of each interval.
-------
92
Final As mass balances are presented in Table 3.18. The
distribution of As between sediment and water of 0.066-ppm microcosms
is about the same as in the 0.05-ppm treated 70-liter sediment micro-
cosms of Experiment 1 (Fig. 3.17; 80% in sediment in both cases).
Although percentage values for sediment As listed for the 11.5-ppm
microcosms in Table 3.18 are nearly the same as for the 0.066-ppm
microcosms, the greater amounts in water as listed in the adjacent
column in concert with our greater confidence in the accuracy of the
water determinations, indicate that, in fact, more of the introduced As
is remaining in the water at the higher treatment levels. This observa-
tion is further corroborated by the results from the 143-ppm microcosms
(Table 3.18).
El odea and Potamogeton accumulated impressive quantities of As as
shown by Table 3.17. Elodea As concentrations averaged 49 ppm (bio-
accumulation ratio, B.R. of 7060) in the 0.066-ppm microcosms while
Potamogeton averaged 4.6 ppm (B.R. of 582). In the 11.5-ppm microcosms
most macrophytes were killed. Even so, As burdens had climbed to an
average of at least 2440 ppm at experiment's end, although the B.R. had
dropped dramatically to 652, less than one tenth of the 7060 reported
above for 0.066-ppm microcosms. The same trend continued in proceeding
from 11.5-ppm to 143-ppm microcosms, except that Elodea (dead) As
burdens merely doubled. Interestingly, As concentrations in Elodea
from pond enclosures were only about one-third of concentrations in
these plants from 11.5-ppm microcosms, even though initial enclosure
water concentrations (9.2 ppm) were almost as high as in the 11.5-ppm
microcosms. Algae (Spirogyra) averaged 2.1 ppm As in the 0.066-ppm
-------
93
Table 3.18. Final arsenic mass balance - Experiment 2
Treatment
(mg/liter)
0.066
0.066
0.066
11.5
11.5
11.5
143
143
143
9.2
9.2
9.2
Microcosm
A
B
C
Mean
D
E
F
Mean
G
H
I
Mean
V
W
Y
Mean
% of initial
As
in
sediment
85.3
75.8
80.3
80.5
77.9
85.6
85.2
82.9
39.9
44.2
45.8
43.3
4.0
4.1
4.8
4.3
% of initial
As
in
water
13.5
9.7
8.0
10.4
40.8
30.2
30.0
33.7
71.0
59.6
50.7
60.4
67.2
43.4
54.1
54.9
Total
98.8%
85.5
88.3
90.9
119
116
115
117
111
104
96.5
104
71.2
47.4
58.9
59.2
-------
94
microcosms, considerably less than the average 15.2 ppm exhibited by
algae in the 0.05-ppm microcosms of Experiment 1. The exceedingly
rapid growth of algae offers one possible explanation for this
difference; i.e., algae in 0.066-ppm microcosms may have bloomed late
in the course of the experiment and thus had less time in which to
accumulate As. This may also explain the large difference between
Spirogyra As burdens in the 11.5-ppm microcosms (55-285 ppm) and those
in the pond enclosures (1970 ppm).
Periphyton As burdens averaged 11 ppm in 0.066-ppm microcosms,
3500 ppm in 11.5-ppm microcosms, and 5200 ppm in 143-ppm microcosms,
while the blue-green algal mats on the sediment surface averaged
4310 ppm in the 143-ppm microcosms.
Snails from 0.066-ppm microcosms averaged slightly less As
(1.6 ppm) than snails (2.4 ppm) from 0.05-ppm microcosms of
Experiment 1. Snails exposed to the higher concentrations of the
11.5-ppm microcosms averaged 309 ppm As. These concentrations are
based on whole, dry snails including shells, and would undoubtedly be
much higher if based on tissue alone.
An estimated 23 and 36 ppm As were found in mosquitofish of pond
enclosures V and W, respectively, while tadpoles from enclosure Y con-
tained a remarkable 402 ppm As. A single sample of chironomid midge
larvae yielded an As concentration of 374 ppm.
The unusual shape of the early part of the 11.5-ppm microcosm con-
centration curves (Fig. 3.19) suggests two possible explanations:
(1) The initial decline in As concentration due to uptake by macro-
and microbiota was followed by death of much of the biota, with the
-------
95
resultant release of As back to the water. Succession to a more As-
tolerant community and subsequent As uptake then produced the more
typical As decline observed after Day 17.
(2) The sudden decline in microcosm water pH remobilized As from
sediments to which it had first adsorbed (see Section 3.2.3.2,
Fig. 3.43).
The very low concentrations measured in the 143-ppm microcosms on
the third day are probably a result of the heavy As solution initially
sinking to the bottom of the microcosms in spite of our efforts to
ensure mixing with the lighter water.
3.2.2.2 Chromium
Chromium's high potential toxicity, capacity for bioaccumulation
in aquatic systems, and its presence as an important contaminant in
coal conversion effluents (Hildebrand et al. 1976) justified its inclu-
sion in our microcosm transport studies. Chromium (Cr) was the only
test contaminant used in this series of studies which was not labeled
with a radiotracer and which was introduced into the microcosms in a
series of recurrent doses rather than in a single slug.
Methods. Eight pond microcosms of 7-liter capacity each were con-
structed in the same manner as the arsenic sediment microcosms.
Appendix A provides further details (Experiment 4). Each microcosm
received locally obtained whole river sediment (clayey; 3% organic
matter) to a depth of 5 cm, approximately 4.6 liters of spring water,
and a 6.0-g inoculum of wet algae, primarily Spirogyra, with the usual
assemblage of snails and microorganisms. All microcosms were placed in
-------
96
a growth chamber at 7 to 8°C on a 12-hr-1ight:12-hr-dark cycle. After
a 13-week equilibration period, during which several chemical and
physical parameters were monitored (Section 3.2.1), the microcosms were
dosed with sufficient hexavalent chromium as Na^CrO. to produce 0.1 ppm
Cr in water of microcosms Bl and B2, 0.25 ppm Cr in water of microcosms
Cl and C2, and 1.0 ppm Cr in water of Dl and D2. Microcosms Al and A2
served as controls. Appropriate amounts of NaHCO, were added to Al,
A2, Bl, B2, Cl, and C2 to ensure equality of sodium content among all
microcosms. These treatments (same doses as above) were repeated three
times at two-week intervals (Fig. 3.22). In contrast to our other
microcosm experiments in which test contaminants were labeled with
radioisotopes, only stable chromium was used; chromium concentrations
were determined by atomic absorption spectroscopy.
Water of each microcosm was sampled regularly by pipetting three
5-ml subsamples from different locations within the microcosm, pooling,
filtering through a Whatman No. 42 filter into sample vials, and
acidifying with Ultrex sulfuric acid.
Sediment samples were first obtained using a glass tube. Later a
No. 5 cork borer permitted greater precision in removing replicable
cores after most water had been drained from the microcosms. Sediment
cores were then centrifuged and the supernatant pipetted into sample
vials. After oven-drying at 55°C, weighing, and grinding with mortar
and pestle, sediment-chromium extractions were performed by soaking in
hot analytical reagent grade HC1 and HNO., for several hours. Extract-
ants were filtered and stored for Cr determination via atomic absorp-
tion. Sediment chromium profiles were performed by applying the above
methods to cores carefully extruded onto aluminum foil and sectioned.
-------
97
ORML-DWG 78-18052
0.5
0.4
0.3
0.2
0.1
0.8
I 0.6
o
cc
g 0.4
0.2
0
5
4
3
2
1
0
.1 ppm TRMT
0.25 ppm TRMT
C1
^
1.0 ppm TRMT
0 6 13 20 27 34 41 48 55 62 TIME (days)
t t t t
4/12
4/27
5/11
5/25 CHROMIUM ADDED
Figure 3.22. Dissolved chromium concentrations,
Experiment 4. Each point represents
one sampling date for one microcosm.
Chromium was added at the points
indicated.
-------
98
Algae and snails were picked out of the microcosms with forceps,
blotted dry, and ashed at 550°C. Chromium extraction proceeded as for
sediments.
Results. Figure 3.22 shows, for each treated microcosm, the
observed Cr concentrations in water and the points of Cr addition during
the 58 days between first treatment and final sampling. After semi log
plots of the concentrations (observed following the final treatment
closely) approached the straight lines of logarithmic decay, Cr loss
rate constants r, for water, were determined:
r = -In(nt/n0)/t,
where n = initial concentration, and n. = concentration at time t.
U U
Table 3.19 presents these loss rate constants for each microcosm. Loss
rates ranged from 0.03 to 0.05/day and varied inversely with Cr concen-
tration among the three treatments, possibly due to increased saturation
of available adsorptive surfaces in the higher concentration treatments
following more than 40 days of exposure. Chromium loss rates were
intermediate among the five test contaminants used in this series of
experiments. Final Cr concentrations in water ranged from below the
limits of detection in the controls (< 0.01 ppm) to 2.1 ppm in the 1-ppm
treatment microcosms, Dl and D2 (Table 3.20). Estimated chromium con-
centrations in rivers of the Tennessee drainage basin ranged from 2 to
20 ppb with a mean of 6 ppb (Braunstein et al. 1977). As was the case
with every other contaminant tested in these studies, the sediments
were probably the major sink for the introduced Cr. However, as shown
by the Cr-sediment profiles in Fig. 3.23, natural Cr levels in the
-------
99
Table 3.19. Chromium kinetic loss rate
constants, r, for
microcosm water3
Mi crocosm
Bl
I2
X
Cl
C2
X
Dl
D2
~x
0.0547
0.0515
0.0532
0.0349
0.0532
0.0441
0.0297
0.0295
0.0296
following final Cr inoculation.
-------
100
Table 3.20. Final Cr distribution
Treatment
Control
0.1 ppm x
0.25 ppm
1.0 ppm x
Microcosm
Al <
A2 <
x~ <
Bl
4 B2
x~
Cl
x 4 C2
x~
Dl
4 D2
x
Water
(ppm)
0.01
0.01
0.01
0.11
0.12
0.12
0.45
0.27
0.36
2.14
1.96
2.05
Sediment
(ppm)a
18.7(19.0)
21.0(22.5)
19.9(20.8)
20.7(23.0)
22.3(19.5)
21.5(21.3)
21.0(19.5)
22.1(21.0)
21.6(20.3)
24.5(24.0)
20.7(23.5)
22.6(23.8)
Algae
(ppm)
6.2
38.5
22.4
63.5
46.4
55.0
179
290
235
1020
1286
1150
, Snails
B.R. (ppm)
N.D.
0.36
2240
25.4
22.1
458 23.8
36.1
42.2
653 39.2
84.7
50.8
560 67.8
B.R.b
> 36
198
109
33
aBased on sum of sections; whole core values in parentheses.
^Nonequilibrium bioconcentration ratio.
-------
101
ORNL-DWG 78-48054
0-1 cm
1-2cm
icm
A1
A2
A1
A2
A1
A2
CONTROLS
I
I
4
t
t
t
0-1 cm
1-2 cm
>2 cm
c
B1
B2
81
B2
Ql
bl
B2
>.1 ppm TRMT
1
•J-
-|-
0.25
C1
C2
C1
C2
C1
C2
ppm TRMT
H-
-f-
+
|
+
1
02
01
02
01
02
1.0 ppm TRMT
J
10
Figure 3.23.
15 20 25 30 0 5 10 15 20 25 30 35 40
CHROMIUM (ppm)
Sediment chromium profiles, Experiment 4. Each bar
represents one depth interval for one microcosm.
-------
102
sediments were far greater (~ 20 ppm) than the increase due to intro-
duced Cr and, in concert with the considerable variability among sedi-
ment samples, thus tended to mask this increase. Use of a radiotracer
such as Cr would have avoided this problem. In any event, a crude
mass balance based on the mean Cr concentrations in sediments of con-
trols and the 1-ppm treatment microcosms (Dl and D2) was attempted. At
the end of the experiment, an estimated 51% of Cr introduced into the
water remained there, while 29% found its way into sediments, leaving
nearly 20% unaccounted for.
Algae accumulated considerable quantities of Cr naturally present
in the control microcosms (Table 3.20) even though dissolved Cr was
less than 0.01 ppm. Probable errors associated with chemical analysis
of the very small quantity of algae obtained from control microcosm A2
may explain the wide difference between Cr estimates in Al algae (6 ppm)
and A2 algae (38 ppm). All algae compare favorably with a reported
value of 3.5 ppm for phytoplankton (NAS 1974). Algal Cr burdens in the
treated microcosms were roughly proportional to the dose administered
(Table 3.20). Algae in the two microcosms receiving the most Cr, Dl
and D2, attained concentrations of 1020 and 1286 ppm, respectively.
Such high concentrations without observable injury are not unknown among
some of the lower terrestrial plants (NAS 1974).
Chromium concentrations in whole, dry snails (Physa) varied with
level of Cr treatment but not at all linearly, as revealed by both
absolute concentrations and concentration factors presented in
Table 3.20. As snail Cr concentrations increased with increasing
treatment levels, bioconcentration factors, [snails]/[water], decreased.
-------
103
Although the microcosms yielded some useful information regarding
the gross behavior and distribution of Cr in semi-natural "pond" condi-
tions, the extent and replicability of data suffered from the adverse
effects of small size (7 liter) on sampling and system stability.
3.2.2.3 Selenium
We chose selenium as a "test" trace element for a transport study
because of its toxicity, apparently high bioaccumulation potential, and
occurrence in coal, coal ash, and coal conversion effluents (Hildebrand
et al. 1976, Bolton et al. 1975). Gissel-Nielsen and Bisbjerg (1970)
reported that selenate in particular is the chemical species most avail-
able to agricultural plants. Its stability, solubility, and high bio-
logical activity (NAS 1976) further underscore the need to study its
transport and fate in aquatic systems. Specifically, we chose to follow
for seven weeks the gross transport of selenium in lentic freshwater
microcosms after having been introduced in a single "slug" as Ma^SeO,.
Materials and Methods
The microcosms used in this study were constructed in July 1977.
Each of four 70-liter aquaria received a 5-cm layer of sediment,
60 liters of pond water, and 100 g (wet wt) of Elodea from a spring-fed
pond on the Oak Ridge Reservation. The sediment and Elodea provided an
inoculum of the fauna and flora in the pond, including protozoa,
rotifers, copepods, cladocerans, nematodes, oligochaetes, algae, insect
larvae, and, in one case, a salamander. In addition, two of the macro-
cosms, Ml and M2, received ten mosquitofish (Gambusia affinis) each.
Microcosms Bl and B2 received two bluegill (Lepomis microchirus) each.
The fish fed only on food produced by each microcosm.
-------
104
The microcosms were maintained in an environmental chamber at 17°C
on a 12-hr-light:12-hr-dark cycle and monitored regularly for dissolved
oxygen, pH, specific conductance, and other chemical parameters. The
microcosms were allowed to stabilize for six weeks before introduction
of selenium. Into each microcosm 3.09 mg Se as Na^SeO, radiolabeled
with 30 /zCi of 75Se (specific activity = 2.16 x 107 dpm/mg Se) were
introduced via a funnel apparatus which facilitated even and gentle
distribution throughout the water column. These concentrations provided
initial conditions of 1100 dpm/ml and 0.0515 ppm selenium in water.
The sampling regime was a compromise between one which would allow
frequent selenium assay of all major components and one which would
keep disturbances in the microcosms to a minimum. Thus, water and
Elodea tips could be removed regularly, sediments and El odea roots less
frequently. Snails were captured alive, weighed, counted, and returned
to their respective aquaria. Fish were not removed until the end of
the experiment. Plant samples were obtained with long-handled tongs
and forceps. These were divided into tips (defined as last 2.5 cm),
stems, and roots. All samples were washed with distilled water, dried
for 24 hr at 60°C, weighed, and transferred to counting tubes. Sediment
cores were collected from the top 1.0 cm of the sediment surface, and
were also dried, weighed, and placed in counting tubes. Water was
filtered through 0.45-ji Millipore filters, and 1.0 ml was placed into
counting tubes.
All samples were radioassayed via gamma spectroscopy using a
well -type Nal crystal. The counting data were corrected for isotope
decay and graphed as dpm/g over time.
-------
105
At the termination of the experiment, the microcosms were taken
apart, and additional components (fish, dragonfly nymph, snail eggs)
were counted, freeze-dried, and weighed. Snails were dissected into
flesh and shell components. Final selenium concentrations in all com-
ponents were computed and mass balance and distribution tables prepared.
Results
Comparison of the kinetic rate constants of the different test
contaminants in Table 3.14 reveals that selenium (0.026/day) moved out
of the water the slowest. Seven weeks after introduction of selenium
into microcosm waters, the concentration dropped from the initial
0.0515 ppm to 0.009 to 0.0165 ppm (Table 3.21), an 82 to 68% loss.
Further, the concentration curves for both water (Fig. 3.24) and sedi-
ment (Fig. 3.25) indicate that dissolved selenium would have continued
to move from water to sediment had the experiment been allowed to con-
tinue. Based on filtration of some water samples, suspended particu-
lates generally accounted for less than 5% of selenium in whole water.
Not surprisingly, the sediments proved to be the major sink for
selenium. After seven weeks, the estimated fractions of the total,
introduced selenium present in the sediments ranged from 57 to 89% with
a mean of 76% (Table 3.22). Final selenium concentration estimates in
sediments ranged from 3.9 to 4.9 ppm dry wt. Variability among sediment
samples within and between microcosms was high due to sediment hetero-
geneity and difficulties in sampling these wet, rather oozy, sediments.
The macrophyte community in each microcosm was composed almost
exclusively of Elodea. Figures 3.26, 3.27, and 3.28 present Se
-------
106
Table 3.21. Final Se concentrations (yg/g dry wt)
Microcosm
Component
Sediment
Water
Elodea
Tips
Stems
Roots
Algae
Snails
Whole
Flesh
Fish
Fish gut
Dragon fly
Bl
4.92
0.0090
3.13
3.39
3.87
11.5
::
3.23
13.0
—
B2
4.87
0.0101
3.79
3.94
4.42
9.13
::
2.81
11.5
5.74
Ml
4.41
0.0133
2.67
2.30
3.61
4.34
9*. 66 (wet)
3.52
—
__
M2
3.86
0.0165
4.78
4.44
5.53
—
6.37
—
-------
107
cn
O)
c
•r~
(O
+->
c
o
o
CM
to
"O
ro
.—I
CO
c
OJ
Q)
o.
X
to
C T-
O M-
•r- O
4-> 4->
(O T-
S_ 3
•4-> cr
C VI
0) O
O E
c
o c
O •!-
(O
OJ4J
oo c
m o
r-. o
T3 CM
oj s:
i— -o
o =
CO
-------
108
(A|uo 35 pappo) oidd
)
.Q
O
O
CM
CO
CQ
C
eu
CU
Q.
X
UJ •
_E
•> to
(/) T-
C <4-
O O
•i- 4->
J-> -r-
(O 3
i- CT
+-) (/)
c o
0) E
o
c: c
o •<-
o «
CU C
oo o
in o
r-
^
CU T3
•a
CU .
oo :
in
CM
CO
cu
1
§o o
GO 1C
-------
109
Table 3.22. Final Se mass balance
Microcosm
Component
Sediment
Water
Elodea
Algae
Snails
Fish
Total
Bl
(*)
89.4
17.6
2.97
0.17
—
0.14
110.3
B2
(%)
85.7
19.8
2.62
0.17
—
0.07
108.4
Ml
(*)
70.1
26.1
2.07
0.30
0.008
98.6
M2
(X)
57.0
32.4
3.89
0.30*
0.027
93.6
Mean for
all microcosms
(X)
75.6
24.0
2.89
0.08
0.30
0.06
102.9
aAsstimed same value as Ml.
-------
110
( in * ro CJ
GO CO *^
_ O' O 6
'Mill
c
cu
0)
Q.
X
CO
Q.
(U
"O
O
LU
O)
n3
C71
"(0
to
c
o
O)
o
o
o
cu
t/o
IT)
CJ
•
CO
0)
en
o o
§o oo o o o
oo
-------
Ill
(X|uo »s P*ppo) uidd
4->
c
0)
E
•r~
§s_
Q)
CL
X
LlJ
m
l/l
o E
to O)
•4-»
IO
«*rt
»w
O)
T3
o o
to i —
LLl
C
*«
0 (/)
«3 §
M bl °
1 +J
1- »O
S-
4->
C
<1)
8 o
o
o
-------
112
(<|UO 8$ P»PPO)
CO (X) lO
_ O bo
»
in
00
h-
O
I
K
O
III I
- \ v
\ \
- \ '
: \
1
I
I
/
!\
: /
- i
- ti <
-\
- \
\
\
i
i
\
— \
\
\
-
-
-
-
-
—
Iiii
j||
iiii
~
COODZZ
1 1
1 Illl 1 1
9
i
i
i
*
\
* \
1
-
i
i
1 i —
1 i
li
1 i
A J>
. \ ,
i ' /
u :
\\
A\ --
\ ;
|\ "
/ \ -
/ \
i \ '
K \ n1
^ \ 1
/ \ \ -
\ -
\ V
m \
\ /
\ 1 ~
\ i
\ i -
\ /
>-"
/\-
/ ~
/ Z"
2 /
1 1 1 1*
Q
(O
If)
*
O
V
m
10
O
10
m
t\j
o
CVJ
m
0
in
o
c.
cu
I
O-
X
O
O
to
O
T3
O
C
O
•r-
O
c
O
O
O)
CO
LO
oo
CM
CO
3
cn
U-
§o o oo o o o o
OOD<04O^lO ^ *"
w —
(6 *jp/iudp(oO WOIN313S
-------
113
concentrations over time for Elodea tips, stems, and roots. These
plants provided the only significant sink for selenium besides sediment,
accounting for about 3% of the total selenium added to all microcosms
(Table 3.22). Final selenium concentrations in Elodea tips (terminal
25 mm of shoot), stems, and roots ranged from 2.3 to 5.5 ppm dry wt and
are presented in Table 3.21 for each microcosm. Roots had slightly
higher final concentrations of selenium than the remainder of the plants
in all microcosms, possibly due to sediment contamination rather than a
greater uptake of selenium. With the possible exception of roots,
uptake in these plants appeared to have equilibrated with elimination
and water loss rates of selenium (Figs. 3.26, 3.27, 3.28).
Bluegill predation on snails (Physa) in microcosms Bl and B2
apparently removed all grazing pressure on algae (primarily Spirogyra)
since algae occurred in measurable quantities in these two microcosms
alone. Selenium uptake by algae had apparently equilibrated with
selenium loss rates from water by the tenth day following introduction
of the contaminant (Fig. 3.26). The concentrations were considerably
higher in algae than in any other component (11.5 ppm dry wt in Bl, and
9.13 ppm dry wt in B2).
Snails were limited to microcosms Ml and M2, apparently due to
predation by bluegills in Bl and B2. Uptake of selenium (Fig. 3.29)
appeared to have equilibrated with elimination and loss from water after
27 days exposure. Final concentrations (dry wt) were estimated to be
4.3 ppm for snails of Ml and 3.4 ppm for snails of M2. The flesh of
several Ml snails was removed, pooled, and radioassayed, with the result
that nearly 10 fig/g (wet wt) was selenium. On the assumption that snail
-------
114
8
(
CO
h-
i
o
_J
z
(X|uo 93 pappojuidd
oo toin » to PJ
oo (Oin * ro pj — d OO OO O
1 1 1 1 f 1 1 1 1 1 1 1 1 1 1 1 1
— D« j DO O —
II i ltn
- «a s/H-j
-5? 0^5
HZ IH§U
" 5> '533
JŁ !^S
— OZ | QJJQQ
S8 i
-si is I
° i
- S !
i
_
-
-
h_
/ i ~~
/ 1
/ !
/ 1
/ i
- / I
— / ' —
- 4 \
\ I
1
\ i
- \ \
- \ \
\ '
\
b i
A |
\ i _
\ 1
- \ ' ~
x
\
\
\
1*
l\
1 \ -
1 \
. «l
\ | |
\ 1
\ 1 -
x I I
1 \ O 4
\ II
r i \ 1 i :
\ \ M. -
\ V in v>
\ \ _i _i
\ \ | §
\ \ *5 -
\ v> y y
tn * =! Q Q
_i <— I I
— AJ ^ 5 ^ ^
l^^s**
1 1 1 1 i i i i wi mill!
8S SS? S 8 2« «"°*
(6 Ajp/iudp CQO WniN313S
.9
IT)
•
10
\i*J
j.^
S c
* 5
E
•r—
s_
O)
Q.
X
O UJ
*
o
ff~-
J ^
o re
PJ s-
4->
E
(U
o
o
o
1O
*" O)
oo
LO
r~~
•
CT1
9 S
•
CO
O)
S-
^
en
iZ
Q
-------
115
flesh dry wt/wet wt ratio is comparable to that of fish (0.21), this
would yield a remarkable 46 ppm Se in dry flesh.
Due to difficulties in sampling fish without seriously disturbing
the microcosms, only final selenium concentrations were determined for
these animals. Selenium burdens in fish were comparable to those in
El odea and snails, ranging from 2.8 to 6.4 pg/g dry wt. The final dis-
tribution of selenium between water and sediment in the four microcosms
as shown in Table 3.22 suggests that bluegills and/or mosquitofish may
have influenced this distribution. Possibly the sediment disturbing
activities of the much larger bluegill accelerated adsorption of dis-
solved selenium onto temporarily suspended sediment particles.
3.2.2.4 Mercury
We selected mercury (Hg) as our fourth test contaminant in this
series of transport studies because of its well-documented mobility,
toxicity, propensity to be concentrated by many organisms, and the
manifold routes of entry into aquatic environments existing today, such
as through agricultural and industrial applications and the burning or
conversion of fossil fuels.
Materials and Methods. The microcosms were constructed in July
1977 from water, sediment, and organisms collected from a spring-fed
pond on the ORNL reservation. The same set-up procedures and conditions
discussed in detail in the previous section on selenium transport were
followed with the exception that fish were excluded from the 70-liter
microcosms. Further details are given in Appendix A. Five months after
the microcosms were constructed (December 12, 1977), a solution of
-------
116
Hg(N03)2 radiolabeled with 203Hg (specific activity = 2.84 x 106
decays/min/^g Hg) was introduced into each microcosm sufficient to
fi ?n?
establish an initial concentration of 1.2 x 10 dpm/liter ( Hg)
and 0.42 /zg/liter (ppb) as stable Hg.
Water, sediment, macrophytes (Elodea), snails (Physa), and zoo-
plankton were regularly sampled and assayed for Hg, using the same
methods described in previous experiments. Upon termination of the
experiment, algae and water mites were also collected and assayed.
Results. Mercury moved out of the water faster than any other
203
contaminant tested. Plots of Hg concentration in water vs time
(Fig. 3.30) for the two microcosms were fitted into two components each:
one for the first three days of Hg loss and one for Days 10 to 41.
First-order rate constants of 0.260 and 0.194/day were determined for
the first components of microcosms Cl and C2 respectively. Loss rates
declined to 0.094 and 0.089/day for Cl and C2, respectively, for the
remainder of the experiments.
203
As expected, Hg concentrations in the sediment rose concur-
rently with decreasing concentrations in water (Figs. 3.30 and 3.31),
although apparently not as smoothly, as a result of the difficulty in
removing whole cores of sediment intact and due to the heterogeneous
nature of the sediment itself. Up to 31% of the Hg introduced into the
system could not be accounted for (Table 3.23). Figure 3.31 indicates
that sediment Hg concentrations had attained a rough equilibrium after
seven days in microcosm Cl and after 16 days in C2. At 8.6 ppb in the
top centimeter of Cl and 8.1 ppb in C2 (Table 3.24), final sediment Hg
-------
117
ORNL-DWG 78-18050
1000
5
~ 100
E
1 5
x
10
10
1 1
1 1 1
-(CALCULATED)
0.5
0.1
0.05
0>
x
•o
0)
0.01
0.005
1
-10 -5 0 5 10 15 20 25 30 35 40 45 50
TIME (days)
Figure 3.30. Dissolved 203Hg concentrations, Experiment 7,
Cl and C2 are the two replicates.
-------
118
(A|uo BH pappo oj anpjqdd
o
in
(O
CM
o
•o
«
o
o
Q.
O
c
OJ
<1J
o •
c 1/1
o
n3
O) O
31 T-
co •—
O Q.
CM OJ
S-
4->
C P
O) S
- a>
0) J=
co
•
CO
§>
(6 Ajp/uudp
-------
119
Table 3.23. Final Hg mass balance
Cl
Water
Sediment
Elodeaa
Snails
Total
Missing
% of total Hg
added
1.1
68.3
5.3
0.40
75.1
24.9
% of total Hg
remaining
1.5
90.9
7.1
0.5
100.0
% of total Hg
added
1.7
60.6
5.7
0.53
68.5
31.5
C2
% of total Hg
remaining
2.5
88.5
8.3
0.8
100.1
aAssuming 23 dry g Elodea/microcosm as observed in Experiment 6 (Se).
-------
120
Table 3.24. Final Hg concentrations in microcosm
components due to initial dose of
0.42 ng/ml (ppb) in water
Microcosm
component
Cl
ng/g
(ppb)
C2
ng/g
(ppb)
Water
Sediment
Top
Mid
Bottom
El odea
Tips
Stems
Roots
Algae
Snails
Whole
Meats
Shell
Eggs
Zooplankton
Water mites
0.0046
8.6
3.2
2.6
64
42
40
86
48
540
6.3
15
240 (wet)
130
0.0074
8.1
3.3
2.5
73
41
51
43
370
6.3
-------
121
concentrations (due to introduced Hg only) were slightly lower than
peak concentrations measured 25 to 30 days earlier. Final average
concentrations for mid-depth and bottom sediments are also presented in
Table 3.24.
As shown in Figs. 3.32 and 3.33, El odea rapidly accumulated Hg,
with peak concentrations in the actively growing tips (224 ppb in Cl
and 229 ppb in C2) occurring after three days in both microcosms. The
subsequent decline in Hg concentrations in both tips and stems after
three or four days reflected Hg losses from water (Fig. 3.30). After
an initial period of wide fluctuations in our estimates of mercury con-
centrations in El odea roots (due in part to the difficulty in sampling
roots), Hg concentrations as expected did not decline, but rather
remained fairly level (Cl) or slowly increased (C2) as sediment concen-
trations increased. Bioconcentration ratios over time are presented in
Fig. 3.34.
Snail tissue accumulated Hg nearly as rapidly as macrophytes
(Fig. 3.35). Maximum mean Hg concentrations in snails were attained in
microcosm Cl after nine days exposure (933 ppb dry wt). Highest
observed concentrations in C2 snail tissues were observed after four
days (484 ppb). In contrast to what was observed in macrophyte tips,
concentrations remained fairly level after 16 days until the end of the
experiment. This suggests that, whereas the Elodea tips obtained most
of their Hg burden from the continually declining Hg in water, the
snails were accumulating Hg primarily from food in and on the sediment
or from the sediment itself. Final mean Hg concentrations in dry snail
tissue after 30 days exposure were 545 ppb in Cl and 365 ppb in C2.
Snail Hg bioconcentration curves are plotted in Fig. 3.36.
-------
122
ORNL-DWG 76-18048
700 R
600 —
500 -
400
Q.
•o
10
O
~ 300
0>
10
o
M
200 -
100
250
- 200
- 150
0)
T3
•o
O
- 100
Q.
Q.
50
14 21 28
TIME (days)
35
42
49
Figure 3.32. El odea 203Hq concentrations, Experiment 7, replicate Cl.
Bars around each point indicate 1 S.D.
-------
123
ORNL-DWG 78-18047
900
800 -
— 300
14
21 28
TIME (days)
42
49
Figure 3.33. El odea
concentrations, Experiment 7,
replicate C2. Bars around each point indicate 1
S.D.
-------
124
ORNL-DWG 78-18046
E
Q.
TJ
X
E
M
'c
o
o>
O>
E
Q.
O
or
LU
o
z
o
o
o
DO
9000 —
8000 —
7000 —
6000 —
5000 —
^ 4000 —
C2 ELODEA TIPS
X C2 ELODEA STEMS
3000 —
2000 —
1000 —
Figure 3.34.
21 28
TIME (days)
Bioaccumulation ratios for 203ng in El odea
tips and stems, Experiment 7. Cl and C2 are
the two replicates.
-------
125
ORNL-DWG 78-18045
4 —
E 3
Q.
T>
(0
O
UJ
CO
CO
z
(O
0»
I
IO
O
CM
— 1.4
— 1.2
1.0 t
O
o»
I
•o
0.8 o
— 0.6
a.
a.
— 0.4
— 0.2
21 28
TIME (days)
35
42
Figure 3.35.
Snail tissue 203Hg concentrations, Experi-
ment 7. Cl and C2 are the two replicates.
around each point indicate 1 S.D.
Bars
-------
126
ORNL-DWG 78-18053
(XIO3)
E
fe 70
Q.
E
Q.
TJ
I
\
O>
60 h-
I 50
•o
»
3
M
tf)
.- 40 I—
o
c
(A
O
l 30
5 20
a:
UJ
o
I 10
o
CD
H 21 28
TIME (days)
35
42
Figure 3.36. Bioaccumulation ratios for 203ng in snail
tissue, Experiment 7.
-------
127
Table 3.23 shows the final distribution of the Hg addendum among
the several components of each microcosm. Nearly 25% of Hg introduced
into Cl could not be accounted for with certainty at the end of the
experiment; 32% was missing in C2. Sediment sampling error, methylation
and subsequent volatilization of Hg, and adsorption onto glass walls of
the microcosms probably account for most of the missing Hg. In any
event, about 90% of Hg remaining in both systems was in the sediment, 7
to 8% in macrophytes, 1.5% in water, and about 0.5% in snails.
Zooplankton, primarily cladocerans, occurred in numbers large
203
enough to sample in microcosm Cl only. Zooplankton concentrated Hg
with remarkable rapidity, having concentrations on a wet basis as high as
4.5 x 10 dpm Hg/g or 1.6 ppm at the first sampling after only
21 hr of exposure (Fig. 3.37). This represents a bioconcentration ratio,
[zooplankton]/[water], of nearly 4500 times. Subsequent samplings
revealed a gradual decline in Zooplankton Hg burdens, reflecting the
trend but not the rate observed in the water itself. Mercury concentra-
tions in zooplankton had dropped to about 0.24 ppm (680,000 dpm/wet g)
at the last sampling, 37 days following introduction of Hg into the
water. It should again be noted that the above concentrations are on a
wet-weight basis and that a concentration of 1.6 ppm of Hg would indicate
a figure 5 to 10 times higher on a dry-weight basis (i.e., 8-16 ppm).
It would not be surprising should such high concentrations in zooplankton
prove hazardous to some of their natural predators, particularly plankti-
vorous fish, if not to the zooplankton populations themselves.
-------
128
(X|uo SH pappo oj anp) ujdd
in o «
10
ro cj
(SXjp/uudp) NOIlVdlN3DNOO
CJ
O)
(O
o
Q.
S-
(U
cu
Q.
X
c
o
c
cu
u
c
o
o
CJ>
CO
o
CM
(O
"o.
8
co
CO
O)
3
CD
Lu
-------
129
3.2.2.5 Anthracene
Aqueous effluents from coal conversion facilities are expected to
include hundreds or thousands of organic compounds, including phenols,
aromatic amines, monoaromatic hydrocarbons, thiophenes, and polycyclic
aromatic hydrocarbons (PAH) (Herbes et al. 1976). PAH are of particular
concern for several reasons: (a) many are carcinogenic; (b) PAH may
not be removed efficiently by biological waste treatment; (c) some
appear to be resistant to degradation; and (d) their environmental fate
is practically unknown (Herbes et al. 1976).
We selected anthracene, a 3-ring PAH, for intense study as a
representative compound of this class. Anthracene is intermediate in
size (M. W. = 188), its solubility in water is 73 //g/liter (Mackay and
Shin 1977), and its octanol:water partitioning coefficient is 2 x 10
(Leo 1975). The characteristics of anthracene which determine its fate
in aquatic environments have been summarized by Southworth (1977).
Anthracene is moderately volatile, with a predicted volatilization half-
life of 300 hr in a 1-m deep water column under quiescent conditions.
It sorbs strongly to suspended particulate organic matter (K . = con-
centration on particulates/concentration in water = 25,000), an order
of magnitude less to clay particles (K, = 1600), and only slightly to
silt particles (K, = 100) (Southworth 1977). Anthracene is rapidly
degraded in sunlight. The phytolytic half-life in a 2-cm column of
distilled water is approximately 35 min under midday, midsummer sunlight
at 35°N. Because photolysis is a function of light intensity, adsorp-
tion of light by the water would probably limit photolysis to the upper
100 cm in most natural waters (Southworth 1977). Hydrolysis of anthra-
cene has not been found to occur.
-------
130
Microbial degradation of anthracene has been measured in short-term
radiotracer experiments. A mixed culture of microbial strains derived
from soil around an oil-drilling site rapidly converted anthracene to
polar compounds, with over 90% conversion within 90 min (Herbes et al.
1977). In samples of sediment from an oil-contaminated stream, anthra-
cene was readily broken down (half-life = 12 days), but only 10% of the
radiotracer appeared in the polar fraction. In sediment from an uncon-
taminated stream, the degradation rate was an order of magnitude slower
(Herbes and Schwall 1978).
As explained above (Section 3.1.2), this experiment went beyond
the previous experiments by examining the chemical fate of the test
contaminant in addition to its distribution among microcosm components.
This was possible because analytical procedures had been developed as
part of other programs in the Environmental Sciences Division. This
experiment has been described by Giddings et al. (in press), from which
the following sections are taken.
Materials and Methods
f 14 1
Two 6-month-old microcosms were treated with 19- CJanthracene
(specific activity = 32 fjCi/M, Amersham/Searle) in July 1977. Approxi-
mately 1 fiCi of anthracene was dissolved in 20 ml of acetone and diluted
to 500 ml with distilled water. The solution was divided into two
250-ml aliquots. One milliliter of each aliquot was withdrawn for
radioassay, and the remainder was mixed with about one liter of micro-
cosm water and slowly poured into the microcosm. The initial concentra-
tion of anthracene in the water was calculated to be 0.5
-------
131
Two 1-ml water samples were removed from each microcosm, two to
five times weekly, placed in a scintillation vial with 10 ml of Bray's
solution, and counted on a Packard Tri-Carb Liquid Scintillation
Spectrometer. Samples of sediment, Elodea, and filamentous algae were
taken approximately once a week. Only the upper 1 cm of sediment was
collected, except for the final samples in which entire cores were
removed and sectioned. Samples of sediment and organisms were dried
overnight at 55°C, weighed, oxidized with a Packard Tri-Carb Sample
Oxidizer, and counted.
The final samples were taken 70 to 84 days after anthracene addi-
tion. In addition to the samples described above, periphyton (from the
aquarium walls), snails, snail eggs, zooplankton, and water mites were
collected. The snails were removed from their shells, and all samples
were then dried and oxidized as described above.
Water samples for extraction and thin-layer chromatography (TLC)
were taken from the microcosms 11 times during the course of the experi-
ment. Sediment samples for extraction were taken at the end of the
experiment; only the upper 2 cm were used. Algae samples were removed
and extracted on Day 42. Daphnia magna were abundant at the time of
anthracene addition, and 10 to 12 individuals were removed daily until
the population was depleted.
Samples were extracted as follows: (1) Water samples were acidi-
fied with a mineral acid and extracted twice with ethyl acetate in a
1:5 ratio of solvent to water. For each extraction, solvent-water
mixtures were agitated for 45 min on a wrist-action shaker. (2) Wet
sediment samples were extracted overnight with acetone in a Soxhlet
-------
132
apparatus. Acetone was evaporated from the extracts, and the remaining
liquid was extracted three times with equal volumes of ethyl acetate.
(3) Excess water was blotted from Daphnia and algae. Samples were
weighed, homogenized in acetone, and centrifuged.
The extracts from each sample were pooled, dried over anhydrous
calcium sulfate, centrifuged, and reduced to 2 ml under nitrogen. One
milliliter of the extract was counted by liquid scintillation spectrom-
14
etry to determine total extractable C, and the remainder of the
extract was subjected to thin-layer chromatography followed by radio-
autography. Toluene was used as a solvent for thin-layer chromatography
on silica gel 60 TLC plates. Radioautograms were prepared with x-ray
no-screen film.
Residues remaining after extraction of algae and sediment were
oxidized to determine unextractable radioactivity.
Results
14
The C activity in the water declined rapidly for the first
15 days, then declined more slowly for 50 days thereafter (Fig. 3.38).
The mean first-order kinetic rate constant was 0.055/day for Days 0-15,
and 0.007/day for Days 20-65. There was a suggestion of accelerated
14
C loss over the last 20 days of the experiment.
14
Table 3.25 shows the C activity in all microcosm components
12 weeks after anthracene was added. Approximately 30 cpm/ml (15% of
the initial C) remained in the water. Most of the C accumulated
in the upper 2 cm of sediment, with much lower activity below this level.
The total activity in the sediment was estimated to be 10.2 x 10 and
-------
133
o
OJ
to
in
I
oo
i^
o
Jj
Z
cr
o
o
oo
o
-o
O
rO
O
00
O
O
O
c\J
O
o
o
IT)
o
00
o
OJ
CO
+->
c
QJ
Q.
X
S-
-r-
•r— 1 —
•P Q.
O
-------
134
Table 3.25. Final C activity in microcosm components twelve weeks
after addition of 10.4 x 10° cpm as anthracene
Component
Water3
Sediment
Upper 2 cm
Lower 5 cm
Plants
Elodea
Algae (benthic)
Periphyton
Animals
Zooplankton
Water mites
Snails (tissue)
Snails (eggs)
Ml
31.6
7,220
875
25,800
22,500
17,800
12,400
31,900
30,100
21,300
M2
(cpm/g dry wt)
30.5
6,080
374
28,000
29,800
20,100
26,700
45,700
56,800
21,600
aWater activity in cpm/ml.
-------
135
8.0 x 10 cpm for the two replicate microcosms. The activity in the
14
water and the sediment accounted for 92 to 114% of the C initially
added (Table 3.26). It is therefore unlikely that volatilization of
14
C compounds was significant in this experiment.
All organisms accumulated activity to about 10 times the activ-
ity in the water. In each microcosm, the highest activities were found
in water mites and snail tissue. These animals probably represent the
ends of the grazing and detrital food chains, respectively, implying
that anthracene or its degradation products may undergo biomagnification
(accumulation up the food chain).
Chromatographic analysis of water, sediment, Daphnia, and algal
extracts showed the presence of anthracene, four radiolabelled deriva-
tives, and an unresolved polar fraction (Table 3.27). Co-chromatography
of sample extracts with internal nonradioactive standards permitted
identification of anthracene (R.r = 0.84); however, Rf's of
9-anthraldehyde (Rf = 0.40), anthraquinone (Rf = 0.31), and
anthracene-9-carboxylic acid (Rf = 0.01) did not agree with Rf
values of radioactive unknowns. The contribution of each compound to
total extractable activity in the water during the experiment can be
seen in Fig. 3.39. Only trace amounts of anthracene were recovered
from the water after Day 20. The disappearance of Compound 5 closely
followed the disappearance of anthracene, suggesting that anthracene
was converted directly to Compound 5, which was not persistent.
Compound 3 was present in large quantity throughout the experiment.
This compound constituted as much as 49% of the extractable radioactiv-
ity in the water (Day 28), and although its relative abundance decreased
-------
Table 3.26. 14C budget
136
Anthracene microcosms
I (cpm) II (cpm)
Added 10.4 x 106 10.4 x 106
Recovered
Water 1.7 x 106 1.6 x 106
Sediment 10.2 x 106 8.0 x 106
Total 11.9 x 106 9.6 x 106
% recovered 114 92
-------
137
Table 3.27. Percentage distribution of radioactivity in extracts
Compound
Anthracene
5
4
3
2
1
Polar
R
0.84
0.76
0.34
0.27
0.08
0.04
0.00
Day
Daphnia
22.3
19.3
-
22.7
-
-
35.7
6
Water
9.4
5.6
-
39.5
4.9
-
40.6
Day
Algae
1.5
1.6
-
40.5
-
1.9
54.5
42
Water
0.9
0.8
-
24.7
-
2.0
71.7
Day
Sediment
32.1
15.7
-
29.5
-
6.7
16.0
84
Water
-
-
-
13.1
-
2.4
84.5
-------
138
CO
00
sr
CD
i
GO
o
I
o:
o
CO
O)
(O
O
S- CL
co OJ
O O)
o 10
O i-
$- <1J
O)
E i-
o *o
<4- (O
CO »O
4-> O
O
03
i- •
X CM
CD •
co
c
•r- C
O
«o co
0}
o
«d- T3
,-1
CO _ _
55-
&b
o
o
o
GO
o
CD
o
C\J
CO
CO
O)
3naviovdix3 do %
-------
139
thereafter, it was still present when the experiment was terminated on
14
Day 84. Forty-five percent of the total C in the water was not
extractable on Day 84.
r 141
Distilled water to which |_9- CJ anthracene was added yielded
anthracene, a polar fraction, the four derivatives found in microcosm
extracts, and an additional compound with Rf = 0.34, after 7 days in
the environmental chamber. The latter compound, which constituted 14%
of the radioactivity extracted from the distilled water, was never
detected in microcosm extracts. Figure 3.40 is a comparison of extract-
able fractions from microcosm water and from distilled water. After
seven days, the C profile from distilled water was very similar to
that of microcosm water on Day 12. Anthracene was nearly absent from
both systems. The large polar fraction recovered from distilled water
(45%) as well as from microcosm water (48%) is particularly striking
since it indicates that anthracene is readily degraded to polar com-
pounds in the absence of microcosm biota. In studies of abiotic degra-
dation of anthracene, Southworth (1977) found that photolysis was rapid
but hydrolytic breakdown was slow; therefore, the anthracene transfor-
mation products detected in the microcosm water and in the distilled
water were probably due to photolysis.
14
After 84 days, unextractable C accounted for 40% of the activ-
ity in the upper 2 cm of sediment and 30% of the activity in the lower
14
sediment. At least a part of this C was probably incorporated into
bacterial cells. One-third of the extractable activity, or 17% of the
initial addition, was anthracene, and another large fraction was
Compound 5 (Table 3.27). The polar fraction was only 16% of the
-------
140
ORNL-DWG 78-6483
50
40
o
tf
w
LU
or
u.
o
30
20
10
0
DISTILLED WATER
MICROCOSMS
ANTH
43
COMPOUND
1 POLAR
Figure 3.40.
Comparison of C distribution in extracts from
microcosm water and distilled water, Experiment 8.
Distilled water was extracted 7 days after anthracene
addition. Microcosm data for Day 12 were inter-
polated from Fig. 3.39. Compounds are numbered as
in Table 3.27.
-------
141
extractable activity. These findings are consistent with observations
of Herbes and Schwall (1978). The data imply that microorganisms in
the sediment do not readily degrade anthracene, and that anthracene in
the sediment could provide a source for continued exposure of benthic
organisms to the parent compound.
Daphnia was found to contain higher levels of extractable anthra-
cene than those found in water (Table 3.27). The distribution of
activity in extracts of algae was not markedly different from the dis-
tribution in water extracts taken at the same time, except that more of
Compound 3 was present (Table 3.27). Sixty-five percent of the C
in the algae was unextractable.
Conclusions
(1) Anthracene disappeared rapidly from the water. After 84 days,
17% of the anthracene initially added was recovered unchanged in the
sediment. Approximately 40% was degraded to more polar compounds,
apparently by photolysis. An additional 35% was unextractable from the
sediment, probably reflecting incorporation into bacterial cells, and
7% was transformed into an unextractable form in the water.
(2) Several of the degradation products persisted in the water or
the sediment for the duration of the experiment. The most abundant
product (Compound 3) was found in all microcosm components examined.
Compound 4, one of the main degradation products formed in distilled
water exposed to light, was never found in the microcosms, implying
that it was broken down further by the microcosm biota.
-------
142
(3) Anthracene was much less susceptible to degradation when sorbed
to sediment than when dissolved in water. Since sorption occurs rapidly
in shallow water, a significant fraction of the anthracene entering a
pond or stream would accumulate in the sediment. This would probably
be available for uptake by organisms interacting with sediments (Walter
and Johnson 1977).
3.2.3 Ecological Effects (J. M. Giddings)
Although the primary goal of this project was to develop microcosms
for testing contaminant transport and fate, we recognized that our pond
microcosms could also be used to test contaminants for their effects on
whole ecosystems. One obstacle to the inclusion of ecosystem-level
effects in a testing program is the lack of readily measurable indica-
tors of contaminant-induced stress. This methodological deficiency
should not obscure the basic importance of studies at the ecosystem
level, since disruptions of ecosystem functions may have far-reaching
effects that are not predictable from single-species toxicity tests.
We sought to identify ecosystem-level effects of contaminants by moni-
toring ecosystem metabolism and nutrient pools in treated and untreated
microcosms. Our objective was to locate "pulse points," parameters
which could be easily monitored in microcosms to indicate ecosystem
dysfunction. The following discussion is based on results of
Experiment 2, in which replicate microcosms were treated with sodium
aresenate at 0, 0.066, 11.5, and 143 ppm.
-------
143
3.2.3.1 Production/Respiration Ratios
The ratio of primary production to total community respiration
(P/R) is an integrative index of ecosystem metabolism which can be
easily measured in many aquatic ecosystems. Odum (1956) proposed the
use of P/R for classifying ecosystems as autotrophic (P/R > 1) or
heterotrophic (P/R < 1), and noted that either type of system tends to
approach P/R = 1 over time. Odum (1969) listed P/R = 1 as an attribute
of mature ecosystems, concluding that the ratio could be used as an
index of relative maturity. P/R ratios of approximately 1 have been
found in many aquatic ecosystems (Riley 1956, Odum 1957, Odum and
Hoskins 1958, Jordan and Likens 1975).
In studies with aquatic microcosms, P/R was found to depart from 1
when a system was disturbed. This effect was observed whether the
disturbance was in the form of temperature stress (Beyers 1962), light
reduction (Copeland 1965), increased grazing (McConnell 1962, Beyers
1963), or toxicants (Whitworth and Lane 1969). Thus it appears that
P/R is a sensitive indicator of stress-induced changes in ecosystem
metabolism. We tested the effects of arsenic on the P/R ratio as part
of Experiment 2.
Net primary production and total ecosystem respiration are shown in
Fig. 3.41. Before treatment with As, the mean net primary production
for all 12 microcosms increased gradually from 3.7 ppm D.O./12 hr at
week 6 (S.D. = 0.5) to 4.6 ppm/12 hr at week 10 (S.D. = 0.7). This
trend continued in the control and the 0.066-ppm microcosms, reaching
5.4 (S.D. = 0.5) at week 16. The two higher As concentrations, however,
resulted in immediate reductions in net primary production. In the
-------
144
ORNL-DWG 77-8290
(
6
5
4 (
c
3
2
1
n
1 1 1
~ CONTROL 9 ~
- ^tJ v
-^'"^ ^ ^
— —
>
? ,
1
_ • NET PRIMAF
0 NIGHTTIME C
iiii!
i i i
~ 0.066 ppm A —
i^r
/ ^^x V-^A^
!L"°'' _
— —
Y PRODUCTION
OMMUNITY RESPIRATION
IIII
8 10 12
14 16 6 8 10
TIME (weeks)
12 14 16
Figure 3.41.
Net production and nighttime coirmunity
respiration in microcosms at three
arsenic levels, Experiment 2. Open
circles: nighttime community
respiration; filled circles: net
production. Each point represents the
mean of three microcosms. Arsenic was
added, in the concentrations indicated,
immediately following P and R
measurement at week 10.
-------
145
143-ppm microcosms, dissolved 0? concentrations declined during the
light period one week after treatment (i.e., gross photosynthesis was
less than daytime community respiration in these microcosms). Net pro-
duction remained near zero for two weeks in the 11.5-ppm microcosms,
and for three weeks in the 143-ppm microcosms. Toward the end of the
experiment, productivity recovered in both treatments; in fact, the net
production at week 16 in the 143-ppm microcosms was considerably higher
than that in the control. Recovery was accompanied by shifts in
community structure. Macrophytes were replaced by Spirogyra in the
11.5-ppm microcosms, and by blue-green algae in the 143-ppm microcosms.
The effects of 11.5 and 143 ppm As on nighttime community respira-
tion were similar to, but less drastic than, effects on production.
The decline in respiration was about one week behind the decline in
productivity, and recovery was also slower. Respiration declined toward
the end of the experiment in the 11.5-ppm microcosms; this effect is
unexplained. Respiration increased very slightly during the experiment
in the control and 0.066-ppm microcosms, averaging 3.4 and 3.9 ppm
D.O./12 hr at week 6 and week 16, respectively.
The ratio of net primary production to nighttime community respira-
tion averaged 1.2 for all microcosms before As treatment. The ratio
remained fairly constant (1.0 to 1.5, mean 1.4) in the control and
0.066-ppm microcosms throughout the experiment (Fig. 3.42). In the
11.5- and 143-ppm microcosms, P/R dropped to zero or below for two
weeks following As addition. The decline was greater at 143 ppm than
at 11.5 ppm. By the third week after As treatment P/R increased, even
in the 143-ppm microcosms where net production remained very low for an
-------
146
ORNL-DWG 77-8289
8
10 11 12
TIME (weeks)
13
15 16
Figure 3.42.
Production/respiration ratios in microcosms at three
arsenic levels, Experiment 2. Each point represents
the ratio of mean net production to mean nighttime
community respiration for three microcosms. Arsenic
was added during week 10.
-------
147
additional week. The highest P/R ratios recorded in the entire experi-
ment were in the 11.5-ppm microcosms at week 16 (P/R = 2.5).
The P/R recovery in both the 11.5- and 143-ppm treatments was
similar to that observed by Copeland (1965) in marine microcosms exposed
to a reduction in light intensity. In both cases, recovery was accom-
panied by changes in community structure. The ability to compensate
for stress by species shifts is an important property of ecosystems.
Until the theoretical relations between population dynamics and ecosys-
tem dynamics are more fully understood, it will be difficult to predict
the effects of a particular stress on a particular ecosystem. Certainly
the P/R response of these microcosms to As treatment could not have
been predicted by measuring the responses of the separate components.
While further refinement of this technique is necessary, the P/R
ratio appears to be a useful indicator of ecosystem response to stress.
The lowest treatment in this experiment had no effect on the P/R ratio
of the microcosms, while higher concentrations produced progressively
greater reductions in P/R. The As concentrations used here were appar-
ently at the extremes of the dose-response curve, but the method appears
usable for determining the response of an aquatic ecosystem to different
concentrations of a contaminant.
We attempted to test this hypothesis in Experiment 4. However,
P/R was highly variable between replicates in this experiment, which we
believe was due to the disturbing effect of monitoring dissolved oxygen
in the 7-liter microcosms. Because of the inconsistency of the results,
no clear trends in P/R were discernable. Further research is needed in
this area.
-------
148
3.2.3.2 Water Chemistry
Effects of contaminant stress were evident in nearly every
chemical parameter monitored in Experiment 2. Declines in the P/R
ratio at 11.5 ppm and 143 ppm arsenic (Fig. 3.42) resulted in dramatic
changes in pH (Fig. 3.43) and dissolved oxygen (Fig. 3.44). These two
parameters are clearly under biological control in the microcosms.
Since pH and D.O. are more easily measured than P/R, they might be used
as indirect indicators of contaminant effects on ecosystem metabolism.
Concurrently with the drop in pH, alkalinity increased in the
11.5- and 143-ppm microcosms (Fig. 3.45), probably due to dissolution
of carbonates which had precipitated during the first two months of the
experiment. This conclusion is supported by the significant increases
in Ca, K, and Mg in these microcosms at the end of the experiment
(Table 3.28). Sodium concentrations also rose, but the increases
corresponded to the amount added as sodium arsenate. Concentrations of
Cl and SO, were also significantly higher in the microcosms receiving
high arsenic doses.
Significant increases in dissolved organic carbon were noted in
the 11.5-ppm microcosms one week after treatment, and in the 143-ppm
microcosms three weeks after treatment (Fig. 3.46). These increases
might represent excretion and/or autolytic release of DOC from plants
and animals, or temporary inhibition of DOC utilization by bacteria.
DOC concentrations returned to the same level as the controls within
one to two weeks.
Concentrations of NO., and NH^ showed a tendency to increase,
following treatment with 11.5 or 143 ppm arsenic (Table 3.29), but the
-------
149
ORNL-DWG 77-12370R
11
10
pH 9
8
7
CONTROL
0.066 ppm
11.5 ppm
143 ppm
0
6 8 10
TIME (weeks)
12 14
16
Figure 3.43. pH, Experiment 2. Each line represents the mean of three
replicate microcosms. Arsenic was added during week 10.
-------
150
20
18
E
Q.
,3 14
X
o
LJ
10
8
o 6
CO
Ł 4
2
0
0
ORNL-DWG 77-12371R
CONTROL
0.066 ppm
11.5 ppm
143 ppm
I
l!
V
8 10
TIME (weeks)
12
14
16
Figure 3.44.
Dissolved oxygen concentrations, Experiment 2. Each line
represents the mean of three replicate microcosms. Arsenic
was added during week 10.
-------
151
ORNL-DWG 78-16990
200
180 —
160
140
ro
O
O
"
cn
120
E
SlOO
>
H
Z
_l
<
_l
<
80
60
40
20
I I I I I I I I I I
I
I
143 ppm
PRETREATMENT
MEANS - 12
MICROCOSMS
CONTROL
^^^«^B>
0.066 ppm
ARSENIC
ADDED
I I I I I I I I
I
I I I I I
7 9
TIME (weeks)
11
13
15
Figure 3.45.
Alkalinity, Experiment 2. The line from
week 1 to week 10 represents the mean of 12
replicate microcosms; lines after week 10
represent the means of three replicate
microcosms. Arsenic was added at the point
indicated. Alkalinity is expressed in ppm
as
-------
152
Table 3.28. Effects of arsenic on microcosm water chemistry:
Experiment 2a
Final (ppm)
Na
K
Ca
Mg
Cl
so4
Initial
(ppm)
0.74
(0.06)
5.56
(0.55)
30.51
(2.49)
10.13
(0.84)
1.44
(0.17)
3.21
(0.24)
Control
0.03
(0.02)
0.10
(0.0)
11.23
(0.47)
5.13
(0.67)
0.30
(0.04)
3.18
(2.50)
0.066
ppm As
0.03
(0.02)
0.10
(0.0)
10.63
(1.44)
3.97
(0.57)
0.28
(0.04)
5.36
(1.79)
11.5
ppm As
8.25
(1.00)
1.87
(1.86)
19.13
(8.20)
6.17
(1.17)
1.36
(0.62)
2.94
(0.47)
143
ppm As
61.17
(1.60)
7.27
(0.23)
32.13
(1.33)
5.97
(0.21)
2.50
(0.42)
8.39
(0.58)
aValues are means for three replicates; standard deviations in
parentheses.
-------
153
ORNL-DWG 78-16989
12
8 —
E
Q.
- 6
o
o
I I I I
i I i FIJI
H
/Li 1.5
A
«»•" i i • ^^ g *
/ ipprry \
/ i i
' \ l<
•143
\ ~
ppm \
PRETREATMENT
MEANS-12 MICROCOSMS
ARSENIC ADDED
I I I I I I I I I I I I I I
7 9
TIME (weeks)
11
Figure 3.46. Dissolved organic carbon, Experiment 2,
3.45.
13 15
Lines as in Fig.
-------
154
Table 3.29. Inorganic nitrogen concentrations following arsenic
treatment: Experiment 2a
Treatment (ppm As)
Date
N03
9/20/76
9/27
10/4
10/11
10/18
10/25
11/1
NH4
9/20/76
9/27
10/4
10/11
10/18
10/25
11/1
Days after
As addition
3
10
17
24
31
38
45
3
10
17
24
31
38
45
0
0.007
0.008
0.018
0.007
0.017
0.006
0.006
0.013
0.015
0.008
0.013
0.015
0.007
0.007
0.066
0.015
0.012
0.009
0.017
0.014
0.012
0.007
0.018
0.022
0.004
0.032
0.009
0.007
0.006
11.5
0.010
0.011
0.012
0.009
0.049
0.006
0.007
0.023
0.034
0.010
0.019
0.024
0.019
0.010
143
0.009
0.011
0.012
0.005
0.011
0.002
0.004
0.015
0.054
0.019
0.052
0.026
0.040
0.015
aAll concentrations in ppm, means for three replicates.
-------
155
data were too variable for statistical confirmation of this trend (see
Section 3.2.1.3., and Table 3.1). Effects on phosphorus concentrations
were large enough to be significant (P < 0.01) in spite of the vari-
ability between replicates (Fig. 3.47). The dramatic increase in total
P implies a temporary disturbance of normal phosphorus cycling; a dis-
ruption in ecosystem function is evident even though we do not know
which component(s) were affected. The usefulness of such "black box"
nutrient parameters as microcosm "pulse points" might be limited, how-
ever, due to the high variability of nutrient measurements. Nutrients
in interstitial water (see Section 3.2.1.1.) may prove to be more sensi-
tive to nutrient cycling disruptions than nutrients in the overlying
water.
Many of the effects of arsenic on microcosm water chemistry were
also observed in pond enclosures (Section 3.2.1.2., Table 3.1).
Nutrient concentrations in the enclosures were quite variable, and
consistent trends such as those shown in Figs. 3.21 and 3.22 were not
found.
3.3 DISCUSSION
3.3.1 Protocol for the Construction of Pond Microcosms (J. M. Giddings)
The following protocol describes the basic procedure for construct-
ing pond microcosms. Since microcosms designed in the manner described
are suitable for a wide range of applications in contaminant research,
standard experimental techniques are not defined in the protocol.
Specific uses of pond microcosms in contaminant testing are discussed
in Section 3.3.2.
-------
156
ORNL-DWG 78-16988
E
Q.
Q.
0.35
0.30
0.25
0.20
0.15
0.10
0.05
/ \
/ \
/ \ -I
/ \
/ \
/— 143ppm \
/
\
ARSENIC ADDED ;
PRETREATMENT
MEANS-12 MICROCOSMS
CONTROL
x'll.Sppm _
7911
TIME (weeks)
13
15
Figure 3.47. Total phosphorus concentrations, Experiment 2. Lines as
in Fig. 3.45.
-------
157
3.3.1.1 Collection of Components
All microcosm components are collected from a single ecosystem,
preferably on the same day they are to be used. A shallow pond is the
best source of material for pond microcosms, but littoral zones of
lakes, or slow-moving rivers, may be acceptable alternatives.
Water is collected first, before the pond has been disturbed by
sediment collection. It is sufficient to simply dip the water from the
pond with a bucket. At least 60 liters are required for each microcosm.
Sediment is scooped, shoveled, or dredged from the bottom of the
pond. Stones, twigs, and other large debris are removed if possible.
Ten liters or more of sediment are needed per microcosm.
If a macrophyte community is present, a portion is shoveled from
the bottom and placed in a bucket or tub. All of the organisms
naturally associated with the macrophyte community can be included,
except crayfish, which have a devastating effect on macrophytes in a
closed system.
If no macrophyte communities occur in the pond, filamentous algal
communities may be collected instead.
The water, sediment, and biota are protected from bright sunlight
and extreme temperatures, and placed in an environmental chamber as
soon as possible. The temperature of the chamber is set to equal that
of the pond water.
3.3.1.2 Assembly of Microcosms
We have found 70-liter si ate-bottomed glass aquaria to be conven-
ient microcosm containers. Each aquarium is washed with tap water,
rinsed with dilute acid, and rinsed several times with distilled water.
-------
158
The aquaria are positioned in the environmental chamber before filling,
since they are virtually unmoveable when full.
A 5-cm layer of sediment is placed in each aquarium. If replicate
microcosms are to be constructed using more than one tub of sediment,
each tub should be equally apportioned among all aquaria. The sediment
within each aquarium is thoroughly mixed, and remaining stones and
debris are removed. If interstitial water sampling is planned, one or
n
two Ami con hollow fiber interstitial water sampling units are buried
2 cm below the sediment surface. The surface of the sediment is leveled
and smoothed.
Each aquarium is then filled with pond water to within 2 cm of the
top. Since many types of sediment are easily resuspended, the water
must be added slowly with a minimum of turbulence. A simple diffusion
device, consisting of the bottom of a 4-liter polyethylene jug with
holes punched around the perimeter, is useful for dissipating the force
of the inflowing water. Once the aquarium is about one-half full, it
is relatively easy to add more water without disturbing the sediment.
If the pond water has been collected in more than one container, each
container is evenly distributed to all aquaria.
If the water is murky after the aquaria are filled (cloudy enough
so the bottom cannot be seen), they are allowed to settle until the
water is clear. This precaution is usually unnecessary, if the water
has been added carefully.
Representative samples of the macrophyte community are briefly
drained (not blotted or seived, since this would damage some organisms
in the community) and weighed in tared 1-liter beakers. Each microcosm
-------
159
receives 100 g (drained wet wt) of the community. A few stems are
pushed into the sediment to anchor the plants. After all the microcosms
have been inoculated, 1 to 2 liters of the remaining water from the
macrophyte collection tub are added to each microcosm as an additional
source of pond organisms.
The completed microcosms are covered with clear 0.25-in. (0.64-cm)
Plexiglas to retard evaporation and to protect the microcosms from dust
and debris. The Plexiglas covers are also useful as supports for dis-
solved oxygen probes, pH probes, and other monitoring equipment. The
covers must be kept clean so that light transmission is not affected.
The temperature of the environmental chamber is set to equal that
of the pond. Light is supplied by an overhead bank of fluorescent
tubes; cool-white tubes are satisfactory, but lights whose spectral
output simulates natural sunlight are preferable. The lights are con-
trolled by a timer on a 12-hr-on:12-hr-off cycle. Light intensity is
probably not a critical variable, as long as it is constant and repro-
ducible; in experiments performed during development of this protocol,
an illumination of 16,000 lux was used. For special applications, it
may be useful to simulate diurnal or seasonal fluctuations in tempera-
ture, photoperiod, or light intensity.
It is necessary to add distilled water to the microcosms periodi-
cally to replace evaporative losses. If a significant volume of micro-
cosm water is removed in sampling, it is replaced with an equal volume
of fresh pond water. No other additions to the microcosm are necessary.
Pond microcosms undergo predictable successional changes during
the first two months after construction, eventually reaching a fairly
-------
160
constant condition. By monitoring pH or dissolved oxygen, it is pos-
sible to detect the end of the successional period and the onset of
steady-state conditions. Studies of contaminant transport, degradation,
and chronic effects should be performed on equilibrated microcosms.
3.3.2 The Role of Pond Microcosms in Contaminant Research
(J. M. Giddings)
Pond microcosms constructed according to the protocol (Section 3.3.1)
are semi-realistic representations of shallow pond ecosystems. They
differ from the natural prototypes in lacking inputs from adjacent ter-
restrial ecosystems, seasonal variations in light and temperature, large
predators, and disturbances caused by storms. The first two items could
be incorporated into the protocol without difficulty; their absence
probably does not significantly affect the model ecosystem over periods
of three to four months or less, since similar conditions prevail in
natural ponds during the sunnier. Populations of some animals (zooplank-
ton, snails, and oligochaetes) are strongly influenced by the presence
or absence of fish. Ramifications of these population effects undoubt-
edly extend to lower trophic levels but may be insignificant; ecosystem
metabolism and nutrient pools are not seriously affected (Section 3.2.1.5)
The effects of storms on pond ecosystems include increased turbidity,
turbulence, nutrient input, and flushing rate; sediment is resuspended,
macrophytes are uprooted and washed away, and water temperature and
chemistry are altered. Pond ecosystems, and organisms that inhabit
them, have evolved in the context of such disturbances, and we do not
know the ecological consequences of imposing constant conditions. We
believe that for most experimental purposes the artificialities of our
-------
161
pond microcosms do not significantly detract from their ability to mimic
natural ponds. The data in Table 3.1 support this conclusion, but much
more comparative research is needed.
One question which concerned us at the beginning of the project was
the extent to which microcosm data would be representative of a range
of aquatic environments. While our microcosms are good models of 20-cm
deep ponds, they probably do not simulate deeper ponds or littoral zones
of lakes as accurately, much less pelagic zones, rivers, etc. Since
lakes and rivers are of greater ecological and economic importance than
20-cm deep ponds, why model the latter? The significance of pond micro-
cosms lies in their degree of realism, which is greater than for most
microcosms designed after other ecosystem types. Of all natural aquatic
ecosystems, shallow ponds are least distorted by encapsulation under
laboratory conditions. Moreover, it has been recognized since Forbes
(1887) that processes occurring in ponds are to some extent typical of
any ecosystem. Accurate simulation of a pond ecosystem is more valuable
for many purposes than poor simulation of a more "important" aquatic
system.
It is often supposed that an advantage of microcosms over field
experiments is their simplicity, which aids in analyzing ecosystem pro-
cesses. This supposition is unjustified. Whittaker (1961) noted that
"aquarium communities are small but not simple; they are scarcely
simpler to study and interpret than full-scale natural communities."
Model ecosystems, including our pond microcosms, are too complex for
analysis of individual processes, pathways, or rates. For example, it
is impossible to distinguish between bioaccumulation of contaminants
-------
162
through the food chain and direct absorption from water. We measure
concentrations over time and calculate net rates, but we cannot isolate
the various processes involved. Net rates are useful for comparisons
between contaminants. The five contaminants we tested, for example,
differed in the kinetic constants for disappearance from water
(Table 3.14) and in bioaccumulation ratios (Table 3.30). It must be
recognized that the absolute values of these parameters are not proper-
ties of the contaminants themselves but are "performances" of the system
(Warren and Liss 1977). The absolute values are system-specific and
cannot be applied with confidence to different aquatic environments.
Because of the difficulty of untangling components and processes
in a pond microcosm, additional information about the autecology of the
dominant organisms or rate constants of major processes is valuable in
interpreting the results. This information may be obtained in simpler
systems (the anthracene photolysis experiment conducted in conjunction
with Experiment 9 is an example). In some situations, the necessary
supplementary information may be available from earlier stages of con-
taminant testing.
A complementary approach to the use of microcosms is to measure
individual processes in simple experimental systems and to integrate
the results by means of a mechanistic ecosystem model. Some modelers
have proposed that the only appropriate use of microcosms is the valida-
tion of models. Since microcosms can only provide estimates of net
processes, validation of a model must be based on comparisons of
model-derived and microcosm-derived net rates. If discrepancies are
encountered, they may be due to one or more of the following factors:
-------
163
Table 3.30. Surmiary of bioaccumulation ratios
Contaminant
As
Cr
Se
Hg
Anthracene
Algae
969-2310
458-653
904-1278
18,700
712-977
Elodea Physa
164-469
33-198
215-401 326
726a
7400-10,600 5900-10,400
28, 000-75, 000a
816-918 953-1862a
Zooplankton
626
-
-
5200
392-875
aSnail tissue only.
-------
164
(a) inadequacies and omissions in the model;
(b) inaccurate measurement of single processes in the simple
systems;
(c) inaccurate measurement of net rates in the microcosms;
(d) inadequate adjustment of single process results to correct
for differences between the standard test conditions and the
set of environmental conditions present in the microcosms; and
(e) interactions between individual processes or components,
modifying behavior of the contaminant in unforeseen ways.
When mechanistic models do not agree with microcosm results because
of factors (d) or (e), microcosm experiments provide the most direct
means of predicting contaminant behavior in specific ecosystems. Micro-
cosms simulate a natural set of environmental conditions, including
their temporal and micro-spatial variations. Likewise, microcosms auto-
matically integrate individual mechanisms, including interactions of
which the modeler may not be aware. Microcosm results can thus be used
to identify weaknesses in analytical models, and can provide an overall
picture of contaminant behavior (under one set of quasi-natural condi-
tions) when models fail. Models, on the other hand, can be used in
conjunction with data from simplified experimental systems to assist
the interpretation of microcosm results.
3.3.2.1 Transport
The contaminants tested in our transport experiments ranged from
elements with relatively high natural background levels (Cr, As) to
compounds not normally encountered except in polluted ecosystems
(anthracene and its derivatives). In the experiments in which radio-
-------
165
isotopes of arsenic, selenium, and mercury were tracked through the
model ecosystem, some of the transport we measured may have represented
one-half of a dynamic equilibrium between labeled and, as yet, unlabeled
pools. For example, when adsorption and desorption from the sediment
are equal, so that there is no net transport in either direction, we
will still see an apparent movement into the sediment if we add a radio-
isotope to the water. In the chromium experiment (the only one in which
radioisotopes were not used), we established microcosms with four levels
of total chromium and measured the concentrations in various components
at each level; however, the processes and pathways by which each distri-
butional pattern was achieved can only be guessed at. Inferring cycles
of naturally occurring substances is somewhat more complicated than
following the dispersal of a foreign substance after it is added to the
water. Most of the contaminants to undergo testing with microcosms
will presumably be substances not naturally present in aquatic ecosys-
tems. The experiment with anthracene is therefore the best example of
how our microcosms might be used in contaminant transport studies.
The use of radiolabeled contaminants confers several advantages in
microcosm transport experiments. Smaller samples are normally required
for radioassay than for chemical analysis. Chemically undetectable
contaminant levels can be monitored with radioisotopes. In theory, and
often in practice, the use of labeled contaminants allows for the con-
struction of mass budgets for the added material. Perhaps most impor-
tantly, isotopic labels are almost essential for determining the nature
and fate of derivatives of the test contaminant.
Changes in the distribution of an added contaminant are rapid at
first, becoming more gradual as equilibrium is approached. Accordingly,
-------
166
sampling frequency must be highest at the beginning of an experiment—at
least daily. We have found that water samples taken immediately after
a contaminant is added do not always contain the calculated concentra-
tions, due to incomplete mixing. Sampling the entire water column with
a glass tube, rather than taking samples from a single depth, reduces
this discrepancy. In most of our experiments, the contaminants became
uniformly distributed in the water within one or two days.
Sediment sampling has been a major problem in our transport experi-
ments. The oozy clay substrates we have used do not form compact cores,
and a variable amount of sediment is always lost from the bottom of a
core. Elodea roots become dense as a microcosm ages, further adding to
the difficulty of coring. Because of the heterogeneity of the sediment,
several cores are needed for a reliable estimate of contaminant concen-
trations at any point in time. Ten or fifteen coring attempts are some-
times necessary to get enough good cores for each sample. The effects
of repeated sampling on one microcosm could be significant. We have
compromised by limiting sediment samples to the upper 1 cm, which norm-
ally contains most of the contaminant (Figs. 3.21, 3.23, Tables 3.24,
3.25). Samples of the upper centimeter are relatively easy to take (we
use a glass tube connected to a syringe), and do not disturb the rest
of the system appreciably. At the conclusion of an experiment, entire
cores may be taken after the water has been siphoned off, and the
vertical distribution of the contaminant may be accurately determined
at that time.
By sampling the water, sediment, and biota at intervals throughout
the experiment, we can measure (a) the distribution of the contaminant
-------
167
among microcosm components, and (b) uptake of the contaminant by par-
ticular organisms. In the first case, we are interested in accounting
for 100% of the contaminant over time, whereas in the second we wish to
obtain precise measurements of contaminant concentrations in tissues.
With all of the substances we tested, the water and sediment components
together contained 90% or more of the contaminant most of the time.
Water concentrations are measured more easily and accurately than sedi-
ment concentrations, so it is possible to obtain an approximate indica-
tion of the overall distribution of a contaminant by monitoring the
water (Figs. 3.14, 3.19, 3.20, 3.22, 3.30, 3.38). Changes in water
concentrations over time approximated exponential decay curves in nearly
all of our experiments, enabling us to compare test contaminants by
their first-order kinetic constants (Table 3.14). The kinetic constant
reflects losses from water to organisms as well as to sediment, but
generally the biota have a small effect on water concentrations.
Another point of comparison for contaminant distributions is the ratio
of water concentration to sediment concentration (or total contaminant
mass in water to total mass in sediment). The dynamic parameter r is
more convenient to estimate than the static ratios because only water
samples are needed. Also, the ratios are meaningful only when the
contaminant distribution has stabilized, which it sometimes fails to
do. We recommend the use of r to characterize the water-to-sediment
transport of test contaminants in pond microcosms.
Concentrations of the test contaminant in Elodea, algae, and Physa
can be monitored throughout an experiment, but most other organisms are
not abundant enough for more than a single sample. Net uptake rates
-------
168
can be calculated if samples are taken frequently enough. Algae take
up many contaminants very rapidly (Fig. 3.26), and require sampling
every few hours for estimation of uptake rates. We have found it
necessary to divide Elodea into tips (2.5 cm), stems, and roots, and to
separate snails into tissue and shell, since contaminant distributions
are frequently nonhomogeneous in these organisms. Elodea uptake is
difficult to interpret, because uptake varies in different parts of the
plant and translocation occurs within the plant as well.
Bioaccumulation ratios (concentration in organism/concentration in
water) may be useful for comparisons between contaminants. Data from
our experiments (Table 3.30) indicate that mercury was concentrated by
all organisms to a greater extent than the other substances tested,
which is in accordance with the known behavior of this element. Most
bioaccumulation ratios for the other contaminants fell between 100 and
1000. There were no consistent differences between organisms. It
should be pointed out that bioaccumulation ratios tend to increase over
time in microcosm experiments; concentrations in organisms reach con-
stant levels in spite of gradually declining water concentrations. The
meaning of nonequilibrium bioaccumulation ratios is unclear. No sig-
nificance can be attached to small differences between bioaccumulation
ratios, but extreme values appear to be reliable indicators of highly
bioaccumulative substances such as mercury (Table 3.30) and DDT (Metcalf
1977).
3.3.2.2 Fate
Our remarks in Section 3.3.2 concerning the difficulty of analyz-
ing single processes in microcosms apply to fate studies as well as
-------
169
transport experiments. No matter how thoroughly the microcosm compon-
ents are sampled, it is difficult or impossible to reconstruct pathways
of contaminant degradation from microcosm data alone. If a degradation
product is detected in the tissues of an organism, it may be a product
of that organism's metabolism, or it may merely have been absorbed from
the water. The case of anthracene is a good example. Lu et al. (1978)
tested anthracene in a Metcalf microcosm and found several derivatives
in water and organisms after 33 days. They inferred biodegradation
pathways within each organism from the derivatives detected in that
organism. Our anthracene experiment revealed that all of the nonpolar
derivatives in the microcosm biota were probably photolysis products
absorbed from the water. The notion of determining degradation pathways
in several species simultaneously in one experimental unit is intriguing
but impractical. Microcosm data tell us about the overall fate of a
compound in a whole system, integrating (and thereby obscuring) indi-
vidual degradation processes. Single-species experiments, on the other
hand, facilitate the analysis of individual processes, but provide no
direct means of assessing overall contaminant fate. These two comple-
mentary experimental approaches are most fruitful when used in combina-
tion.
Within the context of the measurement of net degradation rates,
the major limitation of a microcosm experiment is the chemical analysis
of samples. Separation and quantification of organic contaminants is
often difficult in the presence of complex mixtures of naturally occur-
3 14
ring compounds. Using H- or C-labeled forms of a contaminant to
be tested greatly enhances the range of information that can be derived
-------
170
from a pond microcosm. The methods used in Experiment 8 (extraction of
the parent compound and its derivatives in organic solvents, separation
by thin-layer chromatography, location of radiolabeled spots by radio-
autography, and final quantification by liquid scintillation counting)
could be applied to any compound available in radiolabeled form. With
this procedure, information can be obtained on the net degradation rate
of the parent compound in the system as a whole, the identity and quan-
tity of derivatives of the parent compound, and the persistence of
derivatives. Major degradation processes may sometimes be inferred
from the data, but confirmation must come from experiments with simpler
systems in which the hypothesized pathways are isolated from competing
processes.
We have mentioned earlier that the sediment microbial community is
probably the most realistic component of pond microcosms, although this
assumption needs to be verified by a competent aquatic microbial ecolo-
gist. The rest of the detrital food web is also well represented; pro-
tozoa, nematodes, rotifers, gastrotrichs, and oligochaetes are abundant
and diverse. The decomposer community appears to be little affected by
enclosure in an aquarium or by the exclusion of large predators. Since
most of the biodegradation of organic contaminants in aquatic ecosystems
occurs in the sediment microbial community, the fate of a test compound
in microcosm sediments is a reliable indication of its fate in nature.
Such data are quite valuable because even qualitative information on
the persistence of a compound and its derivatives is difficult to obtain
experimentally (this was the subject of a recent workshop sponsored by
the Gulf Breeze Environmental Research Laboratory of the EPA).
-------
4.0 TERRESTRIAL STUDIES
4.1 INTRODUCTION
4.1.1 State of the Art, Criteria, and Rationale
The purpose of this portion of the project was to determine the
value of microcosms as test systems to examine the rates of transport,
transformation, accumulation, and effects of chemical contaminants in
terrestrial environments. A recent review of microcosm construction
and manipulation suggested six critical study areas (Witherspoon et al.
1976). First, the effects of biotic community complexity on rates and
pathways of chemical transport required determination. Effects of the
decomposer community on elemental transport and accumulation are espe-
cially great. The presence of primary producers, whether grasses or
trees, significantly alters transport rates and chemical states of con-
taminants. Second, information about the replicability and reproduci-
bility of microcosms was needed. A major shortcoming of microcosm
studies conducted to date has been the uniqueness of experiments; very
few have been repeated and the results compared. Third, the determina-
tion of the optimal size of a microcosm for realistic estimation of
contaminant fate was necessary for managed (pastures) and natural
(forests) ecosystems potentially receiving contaminant inputs. Two
major ecosystem processes affecting chemical transport and fate, primary
production, and decomposition, have successfully been studied, using
small experimental units in separate microcosm and pot studies. Micro-
cosms constructed to document the transport and fate of contamnants
should incorporate these experimental units to ensure the realistic
171
-------
172
functioning of their environmental components. Therefore, relative
size becomes an important determinant, in both forest and pasture
analogs, for reliable estimation of conditions in the field. Fourth,
determinations of equilibration time necessary for newly constructed
microcosms and identifiable and measurable parameters to monitor within
microcosms during equilibration and experimentation were required.
Equilibrium, for the purposes of these microcosm studies, is defined as
that state from which reproducible metabolic parameters (such as CCL
efflux) can be expected from field-derived soil cores or larger soil
sections collected at different times. Fifth, the response time of
each microcosm component to contaminant additions (dose, incubation
time, chemical form) until equilibration with the chemical is reached
was needed. Chemical redistribution and transformation within micro-
cosms was assessed after disturbance to or removal of critically impor-
tant microcosm components. Finally, comparisons were needed between
rate parameters and transformation products determined in microcosms
and those occurring in the field. Field values and ecosystem models
were available from the IBP and other ecosystem analysis studies. These
comparisons allowed preliminary extrapolation of microcosm results to
the field; field validation was not within the scope of this study.
4.1.2 Summary of Research Strategy
The approach taken in these studies was to assess the importance
of size, complexity (abiotic and biotic), replicability, and stability
of microcosms in each of two ecosystem types, a pasture and a forest.
The construction of microcosms and the analysis of these as test systems
began in the first year. Size, replicability, and biotic complexity of
-------
173
microcosms from pasture and forest were assessed using routine measure-
ments of certain ecosystem processes (CC^ efflux, plant growth, con-
sumers, population growth, etc.)- These routine measures provide a
basis from which to establish equilibration times for the microcosms
and stability after equilibrium. In addition, comparisons were made
between values obtained from these metabolic measurements and those
from other microcosm and field studies to assess microcosm reality.
Emory silt loam soil was chosen as the forest soil because of know-
ledge gained in previous experiments. This soil is a bottom-land soil,
alluvial in origin, and found in many study areas on the ORNL
reservation (Table 4.1). The soil is slightly acid (pH 5.5) and
moderately high in organic matter (4.7%). The soil has been studied in
detail and its physicochemical characteristics documented.
Pasture soil used in microcosms was a Captina silt loam from a
grassland study area on the ORNL reservation. This is a well-drained,
fairly fertile (3.2% organic matter), slightly acid (5.6 pH) soil occur-
ring near the Clinch River (Table 4.1).
The two terrestrial ecosystems were chosen as those for which we
have complementary field and microcosm data. The suite of plants and
microbial and invertebrate populations used in microcosms will reflect
the ecosystem types chosen. The three microcosm sizes were chosen for
(1) simplicity, ease of handling, and replicability (2.5 x 5 cm);
(2) use in other studies (5 x 15 cm); and (3) suitability for tree
seedling growth (46 x 46 x 15 cm) and grass growth (46 x 46 x 15 and
5 x 15 cm). Within the grass and forest microcosms, biotic complexity
was defined to include various combinations of biota known to modify
-------
174
CO
^
o
CO
, —
to
c
CU
O)
Q.
X
CD
4-
O
CO
O
•r—
CO
•r-
i-
O)
+->
ro
S-
u
r-—
ro
CJ
CO
>^
-C
QL
"g
"3
•—
CJ
'i
OJ
t— 1
•
^J-
0}
r—
JO
to
h-
"d
ro ^J
CO — '
+3
•r- >?
oo —
^
CO ^-N
' ^^
0
o
C O) s—
ro -(-> ^«
S- rt)
o E
CD
o
0 O
LU i— 1
O -v.
CT
CD
E
in
Q.
,
•r-
O
E
QJ
CO
>5
CO
o
CJ
LLJ
CO
OO
I— i
CO
CO
un
CO
^>
! 1
(^
•
"*
E
ro
0
40
r—
• p—
CO
S-
o
LJU
•M
CO
O)
^_
0
Ll_
CM
I^J
T— 1
cr>
CM
•
oo
CO
CM
OO
LO
,
CO
CO
LD
ro
0
r—
4->
•r-
CO
ro
Ł
CL
CJ>
-o
c
ro
, —
CO
CO
to
i.
O
-------
175
some chemical transfer rates and transformations (i.e., presence or
absence of microbes, invertebrates, and plants).
4.2 SUMMARY OF RESULTS (W. F. Harris, ed., from contributions by
B. S. Ausmus, D. R. Jackson, E. G. O'Neill and P. Van Voris)
Evaluation of terrestrial microcosms was initiated in recognition
of their potential application as screening tools for assessment of
contaminant transport. To this end, first-year activities included
evaluation of microcosm construction, complexity, stability, and repro-
ducibility (Van Hook et al. 1976). Emphasis solely on contaminant
transport was broadened to include investigation of contaminant effects
that might occur, based on early identification of sensitive system-
level monitoring points in microcosms. Thus, second-year activities
were undertaken to: (1) further examine and document the validity of
nutrient export and C0? efflux as sensitive measures of ecosystem
effects, (2) continue examination of chemical transport in microcosms,
and (3) refine and document protocols for standardizing microcosms for
use as screening tools for chemical contaminants. The third-year effort
on terrestrial microcosms involved no new experimental work. The exten-
sive data sets developed in the first and second years were subjected
to further analysis across experiments to elucidate patterns of terres-
trial microcosm behavior. A second major analytical task in the third
year dealt with application of simulation modeling to microcosm and
field data as a means of (1) extending the interpretation of microcosm
results in time, and (2) through the use of sensitivity analysis, estab-
lishing the range of physical and biotic conditions in the terrestrial
environment that need to be included in measurements on terrestrial
microcosms.
-------
176
4.2.1 Behavior of Heavy Metals in Forest
Microcosms - Excised Tree Microcosms
This experiment (TREECOSM I) was conducted to assess the transport,
fate, and effects of heavy metals on forest microcosms. Specifically,
objectives were to: (1) document the transport and assimilation of Pb,
Cd, Zn, and Cu in forest components, and compare microcosm results to
similar findings reported on Crooked Creek Watershed (CCW) (Jackson and
Watson 1977); (2) document the effects of heavy metals on nutrient cyc-
ling within forest microcosms, and verify that the observed changes in
nutrient dynamics of CCW were indeed the result of heavy metal contami-
nation; and (3) discern the effect of heavy metal contamination on
litter-soil biota in forest microcosms. Efflux of C0? was used as an
index of biotic activity because it can be monitored nondestructively
and has been historically viewed as a measure of decomposer community
respiration (Gray and Williams 1971, Parkinson et al. 1971). In this
study C02 efflux was used as an indicator of the integrated litter-
soil community dynamics, including roots.
4.2.1.1 Materials and Methods
Six intact microcosms (45 x 45 x 25 cm) were excised from a mesic
hardwood forest near Oak Ridge, Tennessee (see Ausmus et al. 1976 for
site description) in December 1974. Soil was alluvial, Emory silt loam
(% clay = 28, % silt = 59, % sand = 13), slightly acidic (pH = 5.5),
with approximately 5% organic matter. Each soil section contained an
Acer rubrum sapling approximately 2 m in height and the associated
ground flora. Sections were placed in wooden boxes, set against two
-------
177
walls and sealed with quick-setting epoxy resin. Microcosms were placed
in a greenhouse and monitored for a period of 20 months.
Microcosms were equilibrated for two months, December and January,
then litter layers were removed and weighed. Litter from control micro-
cosms was replaced after weighing. Original litter from three micro-
cosms designated for heavy metal treatment was replaced with equivalent
weights (35 g 01 litter, 100 g 02 litter) of litter contaminated with
heavy metals collected 0.4 km from the smelter on Crooked Creek
Watershed (CCW) (Watson et al. 1976). After three weeks, additional
heavy metal input to treated microcosms was added in the form of bag-
house dust obtained from the smelter adjacent to CCW. Dust was applied
weekly as an aqueous slurry for 12 weeks, approximating one annual
deposition which would occur 0.4 km from the smelter stack on CCW
(Jackson and Watson 1977). The total combined dose to the treated
222
microcosms was: 1 mg/cm Pb, 0.13 mg/cm Cd, 0.75 mg/cm Zn, and
2
0.16 mg/cm Cu (Table 4.2). Lead and Cu were primarily associated
with contaminated litter, while the primary source of Cd and Zn was
baghouse dust (Table 4.2). Soil leachate was collected biweekly follow-
ing addition of distilled water in excess of field capacity and analyzed
for heavy metals (Pb, Zn, Cu, and Cd) by flame atomic absorption and
for nutrients (Ca, Mg, and K, and NO^-N and PO. by flame atomic
absorption and Autotechnicon techniques, respectively).
Blackened plastic boxes approximately one-fourth the surface area
of a microcosm were inverted over alkali traps (25 ml, 0.18 M KOH) for
24-hr periods. Carbon dioxide efflux was calculated from titration
data using 0.10 M HC1 as acid titrant. Open and closed blanks were
-------
178
CO
03
-Q
c~
rO
S-
CD
+J
'^
O)
-P
c
1
o
o
CO
(O
T3
O)
CO
0
Q.
Ol
*^3
'S'
t\
T3 CO
0 E
co
•> O
C (J
rvi o
i-
•> o
*•— ^
co co
r- QJ
03 S-
4-> O
O> 4-
E
c
>> 0
fO •(->
O) CO
f~ ^
-o
M-
O
c
hg
O)
CO
O
T3
^
^ — '
^-^
CM
E
O
cn
cu
CO
O
-a
,-^
OJ
E
u
•^^
cn
E
o
^^
01
CO
o
*^j
2-
^~^
OJ
E
O
^^
cn
E
^•"^
r^
Q.
CD
CO
O
T3
^5
— *
^^
OJ
E
u
cn
^
^ f
o;
o
S-
^
o
CO
^~
(O
r^
S-
a>
4_J
1 ^
•r-
OJ
o
r^^.
•
oo
<~o
o
o
•
o
OJ
•
CO
VD
O
1 — 1
LO
•
o
cn
•
LO
CO
o
1 — 1
1— 1
•
o
^.
•
OJ
OJ
LO
•
OJ
rO
4_>
CO
"^
•a
a>
CO
"^
o
cn
to
ca
o
o
.— i
i— i
10
1— (
•
o
o
o
r— 1
00
r*-^
•
o
o
o
r-H
CO
OJ
I— 1
•
o
o
o
I— <
o
I— I
t— 1
01
CO
o
•a
r—"
fO
4^
0
1—
•
to
J^
o
-f-*
fO
0
•r-
r—
cx
CL
fO
>^
^~
XX
OJ
OJ
OJ
<— 1
'I
o
1 —
rO
O
I—
n3
-------
179
used to subtract ambient COp concentrations from CCL efflux esti-
mates. Between yeardays 300 and 350, C0~ traps were not maintained,
resulting in a gap in the data base. Other parameters monitored were:
ambient air temperature, soil moisture, nutrient concentrations in soil
leachate, leaf nutrient concentrations, and aboveground plant growth.
Six Acer leaves were collected monthly during the first growing
season and analyzed for heavy metal content. Foliage from each micro-
cosm was pooled and ashed at 450°C for 24 hr. Ash was dissolved in
25 ml 0.1 H HC1 and analyzed for Pb, Cd, Zn, and Cu.
Microcosms were harvested during mid-September 1976. Microcosm
components were separated, dried, ashed, and dissolved in 0.1 N, HC1
prior to heavy metal analysis. Root tissue (< 0.5 cm diam., > 0.5 cm
diam.) was thoroughly washed to remove adhered soil particles. Soil
was partitioned into three depths, 0-5, 5-10, and 10-25 cm, and heavy
metals were extracted by shaking 5-g aliquots of dried and sieved soil
with 25 ml 0.1 N. HC1.
A new procedure for the determination of extractable nutrient pools
(Jackson and Hall 1978) was used at harvest to extract soil (0-5 and
5-10 cm depths) from treated and control microcosms (see Section 4.2.6
for technique description). The technique involved excision of cores
(5 cm diam. x 10 cm depth) from microcosms, division of cores into 5-cm
depths, and sealing in heat shrinkable polyvinyl chloride. Aliquots of
200 ml each 1 M NaHCOg were leached through each core. The resulting
KC1 and NaHC03 extracts were analyzed for Ca and NH--N; and PO.-P,
N03-N, and dissolved organic carbon (DOC), respectively.
-------
180
Soil microbial communities were characterized using three param-
eters. Total soil microbial protoplasm was estimated using the ATP
assay by extraction of 1 g soil with 3 ml chloroform mixed with 6 ml of
0.02 M TRIS buffer (pH = 7.4). After centrifugation, 2 ml CC14 were
added to the decanted buffer to remove traces of chloroform. The aque-
ous phase of each solution was then assayed for ATP by measuring absor-
bance of NADH (340 nm) produced by enzymatic (hexokinase) reaction of
ATP with glucose. Bacterial density and fungal mass were determined by
standard dilution plating (Parkinson et al. 1971) and modified Jones-
Moll ison slide technique (Jones and Mollison 1948, Parkinson et al.
1971), respectively.
4.2.1.2 Results and Discussion
Transport and Distribution Among Components
Characterization of baghouse dust. Chemical analysis of baghouse
dust revealed Pb, Cd, Zn, and Cu concentrations of 590, 26, 120, and
1.5 mg/g, respectively. Analysis of the dust by x-ray diffraction
indicated the presence of metal sulfates, oxides, and sulfides. In
addition, metals in the dust were readily soluble in 0.01 M^ EDTA
(Fulkerson et al. 1974). Heavy metal particulates removed from leaf
surfaces on CCW were similar in chemical composition and solubility to
the baghouse dust (Jackson and Watson 1977).
Transport and distributon between soil and vegetation. All metals
(Pb, Cd, Zn, and Cu) in treated microcosms were significantly greater
in concentration down to a soil depth of 25 cm relative to the controls
(Table 4.3). Export of dissolved metals in soil leachate of treated
-------
181
Table 4.3. Heavy metal concentrations in litter, soil, and vegetation (Acer
rubrum) components of forest microcosms (N=3)
Litter
01
02
Soil
0-5 cm
5-10 cm
10-25 cm
Leaves
Branches
Bole
Roots
< 0.5 cm diam
> 0.5 cm diam
Pb
10
80
24
18
19
7
2
3
10
11
Control
Cd
0.4
0.5
0.1
0.02
0.02
0.6
0.2
0.2
1.0
0.2
(mg/g)
Zn
84
48
10
4
6
29
30
15
32
39
Treated (mg/g)
Cu
18
26
8
4
6
7
4
4
11
5
Pb
7,800
11,000
1,100
320
460
320
22
62
29
23
Cd
68
100
9.0
3.8
3.2
6.4
1.1
0.6
1.8
0.3
Zn
410
760
40
40
18
56
180
27
115
88
Cu
190
520
20
10
10
18
4
4
11
4
-------
182
microcosms, although greater than controls, was less than 0.2% of the
total applied dose (Table 4.4). These results reflect the mobility of
metals from baghouse dust and contaminated litter, as well as the great
capacity of soil for retention of heavy metals.
The sites of greatest concentration of metals in the microcosms
was 02 litter and 0-5 cm soil in both treated and control microcosms
(Table 4.2). These results reflect the presence of humic substances,
predominant in 02 litter and surface soil, which have high affinity for
heavy metals. The 02 litter was also the site of greatest heavy metal
accumulation on CCW (Watson et al. 1976).
Analysis of leaf tissue during the first growing season (February-
November 1975) indicated no significant uptake of heavy metals. How-
ever, uptake of all heavy metals was significant during the second
growing season, as indicated by metal concentrations in leaves at the
time of harvest (Table 4.3). These results show a lag time of one
growing season before significant translocation of metals occurred into
Acer leaf tissues.
In most cases the distribution of metals in Acer rubrum of treated
microcosms was different from the control microcosms. For example, Pb
in the bole component of treated microcosms was comparable to Pb in
leaves and roots; whereas, in controls the bole contained significantly
lower Pb concentrations than either leaves or roots. Thus, mechanisms of
heavy metal translocation within Acer rubrum were apparently different
in the presence of highly concentrated metals from smelter emissions.
Enrichment in soil and vegetation. Enrichment ratios (ER) were
computed by dividing the heavy metal concentration of a treated
-------
183
.c
o
E
|
0
CM
CD
-C
4->
CU
>
o
E
00
O
(J
o
S-
t_)
•1—
E
4_>
00
CU
O
4-
E
O
i.
00
r~*~
re
4->
cu
>5
>
re
cu
.c
o
4-5
s_
o
Q.
X
cu
cu
4_)
re
^-
0 4->
re c
cu cu
•— E
*^~
re cu
4-> Q.
o x
(— cu
•
**
^
cu
^~-
f">
re
I—
0
c
NJ
oo
o
T3
J«
c^T
g
o
"^^
en
— '
'o?
00
o
"U
^
^_^
. — .
CM
E
O
en
^~J
—
•o
0
CU
oo
o
>s
^ '
CM
E
O
^s.
en
^
-Q
CL.
**~~ N
OJ
oo
O
TD
"^^
*._ ^
^ ^
CM
E
o
*^^
S
•^
CU
E
4-5
re
cu
s-
CM 1
• 1
O
re
*3- o
0 C)
r-H 1
. 1
0
re
r-^ oo
CM «— 1
«-H O
, — |
O 1
o
re
CM tn
t-i O
o o
t— i
CD 1
1
O
re
CM CM
»Ł> WD
r-H O
00
C
o
S- 00 r—
cu oo o
4-> -i- J_
r- E 4->
cu cu c
E 0
h- 00 0
.
^_^
LO
O
CD
V
Q-
,_
O
S-
4-J
O
O
C
re
4->
j_
cu
re
cu
S-
cn
^
i—
"^
re
o
14-
c
co
oo
re
-------
184
microcosm component by the same metal concentration of a control micro-
cosm component, i.e., for Pb in leaves, ER = 320/7 = 45 (Table 4.5).
These ratios were computed to evaluate the mobility of heavy metals in
treated microcosms relative to controls.
With the exception of Cu in branches, bole, and roots, all metals in
every component of treated microcosms were enriched (ER > 1) (Table 4.5).
Within Acer rubrum, the greatest enrichment occurred in leaf tissue for
all metals except Cu. All metals were least enriched in root tissue.
Enrichment ratios show that heavy metals are greatly dispersed within
Acer tissue after assimilation by roots.
Greatest enrichment of metals in soil occurred in 01 and 02 litter
(Table 4.5). This was primarily the result of placing highly contami-
nated litter on the surface of treated microcosms. Cadmium was more
enriched in soil than the other metals. This result reflects the
extremely low indigenous concentration of Cd in control soil (Table 4.3)
as much as the high dosage to the treated microcosms. Enrichment of Zn
and Cu was more uniform through the soil profile than Pb or Cd, suggest-
ing that the mobility of Zn and Cu in contaminated soil was more similar
to that in controls than that in Pb and Cu.
Effects on Nutrient Cycling Processes
To discern differences in initial litter substrate quality between
treated and control microcosms which might bias experimental results,
several parameters were measured. These were: total carbon, labile
carbon (water- and ethanol-soluble), nitrogen, phosphorus, and calcium
(Table 4.6). There were no significant differences in total carbon,
labile carbon, or nitrogen between contaminated and control litter
-------
185
Table 4.5. Heavy metal enrichment ratios for forest
microcosms
Component
Vegetation
Leaves
Branches
Bole
Roots (< 0.5 cm)
Roots (0.5-2.0 cm)
Soil
01 litter
02 litter
0-5 cm
5-10 cm
10-25 cm
Pb
45.0
10.4
26.8
2.9
2.1
783.0
138.0
44.2
18.2
24.5
Cd
10.7
5.5
3.0
1.3
1.5
168.5
214.0
89.0
190.0
160.0
Zn
1.9
5.6
1.8
3.6
2.3
4.9
16.0
4.1
10.2
3.0
Cu
2.6
0.8
1.0
1.0
0.9
10.7
20.0
2.5
2.8
2.4
-------
186
Table 4.6. Initial litter and soil nutrient pools in contaminated and
control microcosms
Stratum
01 litter
02 litter
Microcosm
Treated
Control
Treated
Control
Total
C
8.1
34.7
Labile Ca
(mg/cm2)
3.7
4.0
6.9
7.3
N
0.14
0.18
0.94
1.26
Ca
0.12
0.43b
0.46
4.35b
P
(yg/cm2)
6.5
17. 3b
48.3
167. 9b
aLabile C = water- and ethanol-soluble carbon
bSignificantly different at P Ł 0.05.
-------
187
(P Ł0.05). Initial calcium and phosphorus concentrations in 01 and
02 litter of control microcosms were greater than those in treated
microcosms (Table 4.6). Therefore, slightly greater nutrient turnover
and decay rates may be expected in control microcosm litter, but were
not measured in this study.
Nutrient leachate. Total export of Ca, Mg, K, and N03-N was
significantly increased from treated microcosms during the 20-month
experiment (Table 4.7). Export of PO«-P was not affected by heavy
metal contamination. The variance of nutrient leachate among replicate
microcosms was also increased in treated microcosms.
Greatest losses of nutrients from treated microcosms were Ca and
NOg-N. Similar results have been reported in response to other
stresses, e.g., arsenic (Jackson et al., 1978) and clear-cutting
(Likens et al. 1970). However, significant PO.-P did not occur in
response to heavy metal contamination. Similarly, P was not greatly
depressed near the smelter at CCW (Jackson and Watson 1977).
Nutrients in litter and Acer rubrum. Over the 20-month experiment,
litter mass did not significantly change in either treated or control
microcosms (Table 4.8). While litter nutrients were lost from control
microcosms during the period, contaminated litter retained or gained
nutrients (Table 4.8). This is consistent with observations of
increased nutrient retention in response to heavy metal contamination
at CCW (Jackson and Watson 1977). However, these results cannot defini-
tively validate field results because of the initial differences between
contaminated and control litter substrate quality.
-------
188
Table 4.7. Total dissolved nutrient export from con-
taminated and control forest microcosms
over the 20-month experiment3
Microcosm
Treated
Control
Ca
216
(47)
169
(1.5)
mg total
Mg
59.5
(11.6)
46.5
(1.3)
export/microcosm
K
39.0
(5.1)
26.7
(3.4)
P04-P
133
(30)
158
(12)
N03-N
42
(36)
6.5
(5.9)
aNumbers in parentheses are standard errors.
-------
189
JC
•I-3
c
o
i
O
CM
CU
+j
S-
CD
>
O
IT)
"o
o
Q.
4-3
C
CU
•f—
S-
-t->
E
•i —
in
CD
to
c~
O
to
o
T3
C
to
o
0
a.
+j
E
cu
S-
13
5-
cu
4-> -f_>
4-> E
•r- O)
-— E
•i —
r— S-
tO CU
c o.
•r- X
U_ O)
CO
*
^~
CU
1 —
to
1—
^-^,
CM
E
u
cr
O-
to
0
Q.
z
to
s:
E
(/)
0
u
0
o
• r—
2:
E
3
to
s-
oo
CM
o
4^
CM
oo
o
LO
t— i
•
I— 1
1
LO
•
1^
,_,
o
o
+
LO
t— 1
•
o
^
•
t— 1
10
LO
rH
TO
CD
fO
CU
S-
1—
i.
CU
+->
•r—
1 —
1— 1
o
IO
I — 1
o
1
1^
CM
C5
T— 1
«^J-
•
CO
1
en
CO
CO*
CM
o
10
>— 1
•
o
CO
o
o
(^
r-H
r—
0
S-
4->
E
o
o
CO
o
1 — 1
-1-
00
CO
1— 1
CO
•
^,
1 — 1
+
CT)
LO
10
o
I— 1
o
^J-
o
I— 1
CO
l^
(O
10
o
t-H
T3
CU
IO
O)
h-
J_
CU
-M
•r—
1 —
CM
0
r— 1
«— 1
CO
1
^t-
CM
, — 1
LO
•
oo
1 — 1
'
o
oo
oo
LO
LO
o
I— 1
r^*.
•
o
r-H
•
<Ł>
r^
CM
CTl
1 —
o
t-
-l->
c
0
(^
-------
190
Distribution of nutrients in Acer rubrum tissue did not signifi-
cantly differ between contaminated and control microcosms (Table 4.9).
Leaf and root tissue sampled on CCW near the smelter had depressed
nutrient concentrations relative to those sampled in control areas
(Jackson and Watson 1977). Presumably, primary producer exposure to
heavy metal concentrations requires more than two growing seasons to
cause depressed nutrient uptake due to the great reservoir of nutrients
solely within the tree. Depressed nutrient assimilation by trees could
result from depleted nutrient pools in soil, damaged mycorrhizal asso-
ciations, or both. As shown in Fig. 4.1, extractable soil nutrients
were depleted by heavy metal contamination due to increased leaching.
While mycorrhizal structures were not assessed in this experiment,
chemical toxins have been reported to affect mycobionts (Trappe et al.
1973).
Effects on Litter-Soil Carbon Metabolism
Carbon dioxide efflux from control and contaminated microcosms was
weakly correlated with temperature (r ~ 0.6). Neither soil moisture
nor watering schedule was significantly correlated with daily C02
efflux.
Rates of CO efflux. Contamination of litter-soil strata with
heavy metals significantly increased daily C02 efflux rates following
initial treatment (P Ł0.05). To closely examine COp loss, two
periods of each year were partitioned: spring (mid-February, year-day
50, to early May year-day 124) and summer (early May, year-day 125, to
mid-July, year-day 200). This partition was made to compare C02
-------
191
Table 4.9. Mean nutrient concentrations of forest microcosm com-
ponents at experiment termination
(mg/g dry wt)
Component
Leaves
Branches
Bole
Roots (< 0.5 cm)
Roots (> 0.5 cm)
01 litter
02 litter
Ground flora
Microcosms
Treated
Control
Treated
Control
Treated
Control
Treated
Control
Treated
Control
Treated
Control
Treated
Control
Treated
Control
Ca
6.60
6.10
5.40
7.70
6.31
2.82
5.70
6.31
5.42
5.63
20.9
15.6
13.0
13.4
11.5
9.9
Mg
1.45
1.55
0.48
0.64
0.29
0.55
1.14
1.33
0.72
0.89
2.67
1.74
1.38
1.34
3.01
3.08
K
4.25
5.70
1.43
1.18
0.96
0.79
1.63
2.64
1.15
1.52
1.49
1.89
7.9
8.0
7.56
28.32
P
0.78
1.26
0.38
0.42
0.32
0.36
0.67
0.86
0.55
0.75
0.49
0.52
0.62
0.36
0.74
0.82
N
14.8
13.1
5.9
3.4
5.5
3.3
12.3
9.1
9.4
8.1
7.5
1.0
9.9
7.5
-------
192
9
cc
cc
K> OJ
>
en
CE
R^^^^
en
tr
(scro)QS"^
8
CO
O
0
o
E '
O ."
(6/67/j DO
a. E
URRY
AC
(6/67*)
S
o
<
O .o
j:
Q.
.a o
4J -r-
c c
OJ CD
rtJ
i— CU
•r— i — •
O
t/1 t/>
•— o
J3 C
(C OJ
•f-> T3
O
(O s_
S- tl
4-> X)
X
CD i—
E -O
(O O
o a)
o
LO
fO S-
4-> S- O
O) 3
E •— V
I I/)
>i a.
> -o
(O C 4->
a> (0 a>
M- O O
•M C S_
O ••- O)
0) M-
LU 4-> T3
2!
-------
193
efflux between periods of two consecutive years having similar
temperature-moisture regimes and equal numbers of observations (n = 75).
During the first year, C02 efflux rates averaged 3.20 and
2.60 mg/day from contaminated and control microcosms, respectively
(Table 4.10). Carbon dioxide efflux rates from control microcosms
remained constant through spring of the second year. During summer of
the second year, C0? efflux rates from both contaminated and control
microcosms decreased to a level of 2.0 mg C02/microcosm/day
(Table 4.10).
Total gaseous carbon loss. Cumulative C02 efflux was computed
by sunming daily C02 efflux rates and expressed as a regression
against time. Due to the gap in the data base, total gaseous C02
loss was computed separately for the first (year-days 0-300) and second
years (year-days 0-205).
Total gaseous carbon loss between treated and control microcosms
was identical until year-day 130, after which time greater loss was
incurred from the treated microcosms. These results indicate that the
application of highly contaminated litter did not significantly affect
total gaseous carbon loss until three months post-treatment. After
this time, however, C02 loss from control microcosms remained
constant, while C02 loss from treated microcosms was increased as
indicated by a greater regression coefficient between years 1 and 2
(Table 4.11).
Regression analysis (Table 4.11) indicates that cumulative C02
efflux from control microcosms did not differ between years. Therefore,
the control data were pooled into a single regression equation for both
-------
194
Table 4.10. Comparison of average COg efflux rates (mg CC^/microcosm/day)
of contaminated and control microcosms during spring and summer
of the two years of observations
Year
1
2
Spring
(year-days 50-124)
Contaminated Control
3.209 2.60
3.46a»c 2.67
Summer
(year-days 125-200)
Contaminated
3.20a>b
2.00
Control
2.60b
2.00
Significantly greater than controls at P Ł0.05.
bYear 1 > year 2 at P ^0.05.
cYear 2 < year 1 at P < 0.05.
-------
195
Table 4.11. Simple linear regression of cumulative C02 efflux (mg) upon time
(days) for contaminated and control microcosms for the two years of
experimentation
Year
1
2
1,2
Treatment
Contaminated
Control
Contaminated
Control
Control
N
203
203
186
186
389
Intercept
-14.8
-11.8
-8.9
-5.1
-4.6
Slope
0.4953
0.450
0.635a'b
0.476
0.433
F
10.000C
9,74ic
11,596C
12.915C
4,238C
R2
0.98d
0.98d
0.98d
0.99d
0.92^
aSignificantly different from controls at P _< 0.05.
bSignificantly different from year 1 at P Ł 0.05.
cSignificantly linear regression at P Ł 0.01.
dSignificantly fit with a + bx model at P Ł 0.01.
Significantly fit with a + bx model at P Ł 0.05.
-------
196
years of the experiment. Efflux rates of CCL were independent of
soil moisture and weakly correlated with soil temperature (r = 0.60).
Thus, variable environmental conditions in the greenhouse did not
significiantly affect (XL efflux during the experiment.
Microbial community characterization. The three parameters mea-
sured to characterize the soil microbial community were significantly
different between contaminated and control microcosms at soil depths of
0 to 5 and 5 to 10 cm. Results indicate that metal contamination
increased ATP concentration and bacterial density and decreased fungal
biomass (Table 4.2). Elevated ATP concentrations resulted from the
elevated bacterial biomass, which more than compensated for decreased
fungal biomass in soil of contaminated microcosms.
4.2.2 The Behavior of Arsenic in Terrestrial Microcosms
The purpose of this research was to evaluate the use of soil
microcosms for screening chemicals potentially toxic to terrestrial
ecosystems, and to determine the sensitivity of monitoring Ca and
NOo-N efflux from forest and grassland soil treated with subacute
levels of arsenic.
Mesic hardwood forest and managed pasture were chosen for study
because they are dominant ecosystems of the southeastern United States
and represent possible targets for chemical contaminants in this
region. In addition, these ecosystems are represented on the Oak Ridge
Reservation and have been monitored extensively. Their selection allows
comparison of parameters measured on microcosms to field-established
values. Soil characteristics of these sites have been summarized in
Table 4.12.
-------
197
Table 4.12. Microbial parameters in soil of contaminated and control
microcosms at the end of the 20-month experiment
Microbial parameter
ATP concentration (ppm)
Bacterial density (108/g)
Fungal lengths (m/g soil)
Treatment
Contaminated
Control
Contaminated
Control
Contaminated
Control
Soil
0-5 cm 5-10 cm
13.?a 4.2b
5.6 2.1
21. 8a 10. 2b
8.0 5.1
75a 135a
1300 750
aSignificantly different from controls at P Ł0.01.
^Significantly different from controls at P < 0.05.
-------
198
The mesic hardwood study area is within a 30-ha (70-acre), second-
growth mesophytic assemblage dominated by Liriodendron tulipifera L.
Mean canopy height is 30 m; mean stand age is 48 years. The climatic
regime is humid-mesothermal with a mean annual soil temperature of 13°C.
This system is characteristic of southern Appalachian Highland cove
hardwoods in karst topography and has been the intensive study site in
the Eastern Deciduous Forest Biome of the International Biological
Program.
The managed pasture study area is comprised of 16 ha (40 acres)
and is characterized by codominance of Festuca arundinacea Schub. and
Andropogon virginicus L. The pasture has a perched water table during
winter which causes saturated flow in the silt loam soils during this
high rainfall period. The pasture was in agricultural use prior to
1942 and planted in Pinus taeda L. in 1956 after lying fallow for 14
years. In 1964, the pines were cut and the area planted in Festuca.
Mean annual temperatures resemble the forest but extremes are greater.
The pasture has received occasional mowing to maintain Festuca
dominance.
4.2.2.1 Materials and Methods
Nine soil cores (5 cm diam. x 10 cm depth) each were excised from
a managed grassland and a mesic forest. Growth of plants was eliminated
from each grassland microcosm by cutting stems at the soil surface.
All soil cores were encased in shrinkable PVC plastic sheeting and
placed on a filter holder to collect leachate. The cores were incubated
in a temperature-regulated growth chamber at 25°C. Leachate was
-------
199
collected weekly by adding enough water (20-30 ml) to exceed field
moisture capacity and supply approximately 10 ml in the leachate
collector.
74
Microcosms were treated with As solutions containing As tracer.
10 9
Solutions contained specific activities of 1.09 x 10 and 1.08 x 10 dpm
"7/1 0
As/g As for application rates of 10 and 100 ^ig As/cm , respectively.
Three microcosms were assigned to each treatment level. Each treatment
was applied by leaching As solutions through individual microcosms.
Retention of As by soil was 40 to 60%; therefore, final As concentra-
tions in soil cores were more indicative of actual As treatments than
the initial application rate (Table 4.13).
Leachate solutions were filtered through Whatman 40 paper and
analyzed for pH. Subsamples of 5 ml each were placed in counting tubes
74
for radiochemical determination of As. Samples originating from
untreated soil cores were not routinely assayed for As. However,
analysis of two leachate samples from an untreated core indicated an As
concentration at least 1000-fold lower than leachates from As-treated
soil cores. Therefore, background As in these soils was assumed not to
be a factor in the experiment. Analytical determinations for Ca were
made by flame atomic absorption, and Autotechnicon procedures were used
for determination of N03-N.
4.2.2.2 Results and Discussion
Transport and Effects of Arsenic
The depth-profile of As concentrations in forest and grassland
microcosms indicates greatest retention in the top 2.5 cm of soil
-------
200
Table 4.13. Final As concentrations of As-amended forest and
grassland soil cores (± S.E., n=3)
As treatment
Soil type
Soil depth
(cm)
10
(yg/cm )
100
Forest 0-2.5 5.2(1.1) 34.9(11.2)
2.5-5.0 1.0(0.5) 7.3(1.6)
5.0-10.0 1.0(0.3) 8.0(1.8)
Grassland 0-2.5 4.7(1.4) 24.6(2.6)
2.5-5.0 1.1(0.2) 4.3(1.5)
5.0-10.0 0.5(0.1) 3.8(0.2)
-------
201
(Table 4.13). Presumably, As was retained in this area by humic sub-
stances. Forest soil retained a greater concentration of As than
grassland, which may reflect a slightly higher organic matter content
of the forest soil (Table 4.1). The highest level of As retention
(34.9 ppm) in this experiment is well below a value of 90 ppm reported
to cause reductions in plant growth (Carrow et al. 1975). Therefore,
the levels of As used in this study were considered subacute for
terrestrial ecosystems.
Efflux of Ca and NOo-N was significantly increased (P Ł0.05)
from forest microcosms amended with the highest concentration of As
(Table 4.14). Grassland microcosms were unaffected at this level of
amendment. Forest microcosms appear more sensitive to contamination by
As. This effect may be due to the nreater retention of As by forest
soil. Various indicators of microbial activity (amylase, cellulase,
invertase, and urease enzyme activity and ATP concentration) failed to
detect a treatment effect for both soils and levels of As amendment.
Nutrient Comparisons
Calcium efflux diminished for approximately four weeks from
untreated microcosms of both ecosystems (Fig. 4.2). Similar results
were obtained from treated microcosms which were unaffected by As
amendment. Physical disruption of the soil system created by the
excision process may account for the high initial concentrations of
Ca. A pre-equilibration period of approximately four weeks is recom-
mended for future experiments on the basis of this study.
-------
202
Table 4.14. Aqueous efflux of Ca and N03-N from forest and
grassland soil cores amended with As (± S.E., n=18)
Soil type
Forest
Grassland
As
treatment
(yg/cm2)
0
10
100
0
10
100
Ca
efflux
(yg/ml )
20.3(1.7)
24.7(3.8)
29.4(2.9)
10.6(1.3)
12.5(2.2)
11.3(0.9)
NOa-N
efflux
(ug/liter)
32.4(4.8)
48.9(7.5)
132.2(26.0
20.8(3.5)
23.3(5.0)
39.9(24.0)
-------
40 -
203
30 -i
fO
o •
CD
20
10 -
FOREST
GRASSLAND
1 23456
Time (Weeks)
Figure 4.2. Mean weekly Ca efflux from untreated forest and grassland
soil cores.
-------
204
4.2.2.3 Summary and Conclusions
This study supports the use of intact soil cores for screening
chemicals which may be toxic to terrestrial ecosystems. Nutrient efflux
from the cores was stabilized within four weeks, and the cores were
amenable to routine handling. Treatment of the soil cores with a sub-
acute dose of As resulted in increased nutrient efflux from forest soil,
while population parameters were unaffected. These results suggest
that efflux of nutrients in soil water is a more sensitive monitoring
point than population parameters for assessing chemical stress in soils.
Nutrient losses appear to provide an early warning indicator of more
drastic effects from the exposure to toxic contaminants.
4.2.3 Extraction of Nutrients from Intact Soil Cores
A sensitive soil extraction technique could be invaluable for
assessing nutrient pools in soil cores at the end of a prescribed incu-
bation period. This technique might be more quantitative (in terms of
assessing changes in nutrient pools) and more economical than monitoring
nutrient concentrations in soil leachate.
The technique developed to extract nutrient pools from intact soil
must be: (1) highly sensitive to slight changes in nutrient availabil-
ity; (2) applicable to the analyses of a wide spectrum of nutrients (Ca,
NH3-N, N03-N, P04-P, and DOC), and (3) capable of utilizing one
extractant for cations and another for anions which can be successively
leached through a single soil core. These critieria were established
in a preliminary experiment involving various extractants and volumes.
This experiment was followed by a comparison of three extraction
techniques, using soil sections which were treated with heavy metals.
-------
205
4.2.3.1 Materials and Methods
Soil Preparation
Soils from a mesic hardwood forest (Emory silt loam) and managed
pasture (Captina silt loam) were collected from the Oak Ridge
Reservation (Table 4.1). All soil cores (5 cm diam. x 5 cm depth) were
excised with a steel coring device. Aboveground plant parts were
removed from each grassland soil core by cutting stems at the soil
surface.
Intact soil cores were constructed by encasing undisturbed cores
in heat-shrinkable polyvinyl chloride plastic sheeting. A reservoir
for containing 50 ml of extracting solution was made by extending the
PVC sheeting 5 cm above the soil surface. Free drainage was allowed
from the base of the cores.
Homogenized soil columns were prepared from oven-dried 5- x 5-cm
soil cores which were pulverized and thoroughly mixed. The homogenized
soil was then loosely packed into polyethylene cylinders (5 cm diam. x
10 cm depth). Each cylinder was fitted with fine-mesh nylon screening
on the base to prevent particulate loss.
Intact and homogenized soil columns were placed in a Plexiglas
assembly designed to hold the columns in place and to direct leachate
from the columns into collection beakers. This assembly allowed con-
venient leaching of 20 to 30 soil columns simultaneously. After col-
lection, all leachate samples were centrifuged at 1000 g's for 20 min
prior to analysis for nutrients.
Soil slurries were prepared by the conventional method of mixing
2.5 and 5.0 g soil with 50 and 25 ml of 0.5 M NaHC03 and 0.5 M KC1,
-------
206
respectively (Soil Science Society of America 1967). The mixtures were
agitated for 30 min to ensure thorough mixing and subsequently centri-
fuged to remove the extractant solution.
Chemical Analysis
Standard Autotechnicon techniques were used for analysis of NhL-N,
NOg-N, and PO^-P. Calcium analyses were performed by flame atomic
absorption. Dissolved organic carbon was determined by digestion of the
leachate solutions and subsequent infrared analysis of evolved C02-
Analysis of extractants. Three extractant solutions for intact
cores were evaluated in terms of the operational criteria previously
described. Experimental treatments were comprised of duplicate intact
soil cores of Captina and Emory silt loam soil from grassland and forest
ecosystems, respectively.
All intact soil cores were constructed by the previously described
procedure. The cores were initially leached with 400 ml of distilled
water to remove nutrients and soil colloids solubilized by the excision
process. Volumes of 200 ml of extractant (0.5 M KC1, NH4OAc, or NaHCO,)
were then leached through each core (Table 4.15). This volume was
selected to approximate a soil:solution ratio of one. Soil core weights
varied from 180 to 200 g.
The feasibility of successive extractions from a single soil core
was evaluated using KC1 and NaHCOo for extraction of cations and
anions, respectively (Table 4.15). For this test, soil cores were
successively leached with 400 ml distilled water, 200 ml of 0.5 M KC1,
200 ml distilled water, and 200 ml 0.5 M NaHC03-
-------
207
\
CD
j^L
1 — 1
t/1
C
0
[ S
3
+->
c
4_>
U
re
i-
X
CO 4->
•r-
-!-> E
o .a
re 3
c
•i — •*
i —
E i—
o re
s- n:
Cl-
nQ
CO C
+j rd
OJ C
•r- O
S- I/)
3 O
c re
(—3
<4— ^— •"
O
i — r
C «•— s,
0 .
•r- LU
-t-> .
<-) oo
re
S- +1
xlx
LU » ,
•
LO
i— 1
*
«^*
CU
^~
f~l
re
t-
C_)
g
Q_
O-
z.
00
•z.
i
00
"T~
z.
re
o
a>
a.
>j
i *
1—
•r—
O
00
4J
C
re
4J
O
re
s-
-l->
X
LU
i-H
CM
i-H
*
CO
I— 1
I— 1
o
o
o
oo
o
CM
O
O
O
OO
o
o
x-^x
c^
•
t-H
•*^s
CO
CM
*-~^
CO
cn
*-^^*
t-H
r-4
fO
c:
•r-
^ ^
a.
re
to
,—
C_3
NyX
SI I
LT)
«
0
oo
i-H
CM
^*
i-H
i— 1
O
O
CM
O
O
CM
O
O
O
§
0
^-~+.
LT>
•
o
v_x
CO
•
CM
^— ^.
LO
«=}•
*^-^
f"*^
O
1^
>^
S-
o
E
LU
LO
O
o
CO
o
o
LO
o
o
o
OJ
o
o
.. — .
o
o
oo
«^prf-
LO
SJ-
re
c
•r-*
^ '
DL
*o
O
O
^^
O
^T"
Z
SI
LO
O
i— 1
0
o
LO
0
o
^— ^
o
0
CM
O
O
*—^
O
oo
•v^^
1— 1
^^
LO
>^
S—
o
E
LU
O
O
OO
CM
CT>
O
O
CM
OO
0
^-^
>-H
0
0
o
d
^-^,
LO
•
o
**^*
lŁ>
^H
^ ^
I— 1
•
r— (
LO
•
IO
i-H
re
c
•r-
^_>
Q.
re
o
oo
o
o
ni
re
^^
sr* 1
LO
O
UD
OO
LO
Ol
i-H
CM
O
O
^
o
,*—*
I— 1
o
LO
i-H
O
^-*^
i-H
•
o
s_^
o
CM
^ ^
CO
f__l
^h^x
^o
•
tf^
1r_l
>fc
s-
o
E
LU
co"
f">».
CO
CM
O
O
ID
CM
0
,_,
r-H
O
O
cr>
o
o
^^-^
t— i
o
o
s^^-
^"
•
f-H
^_^
f*x.
•
C\J
CM
•
r-.
r-l
fO
c:
•r—
-M
Q.
fl3
O
re
oo
o
o
re
z.
^"1
LO
0
, — ,
1— I
CM
+*~*
CT>
CO
O
0
CO
I— 1
0
LO
o
0
0
o
o
s~*.
CO
•
o
N_^
p^
•
oo
^ ^
oo
•
t-_ 1
s^^x
LD
•
i-H
i-H
>^
S-
o
E
LU
,
^~
o
sx
5C|
LO
•
o
.E
•5
-a
CU
4-9
o
re
S-
4-^
X
O)
>,
r—
in
o
•r-
>
ai
s-
o.
-------
208
Comparative extraction techniques. Nutrients extracted from intact
and homogenized soil cores were compared to those extracted from slur-
ried soil. Intact and homogenized cores were successively leached with
200 ml of 0.5 M KC1 (for extraction of Ca and NHg-N), 200 ml distilled
water, and 200 ml of 0.5 M NaHCO-, (for extraction of N03~N, PO.-P,
and DOC). Slurried soil samples were prepared by the above procedure.
Soil samples extracted with KC1 and NaHC03 were analyzed for Ca and
NH3-N and for N03-N, P04-P, and DOC, respectively. Triplicate
extractions were made of Captina and Emory silt loam soils for all three
techniques.
Sensitivity analysis. Soil cores for this experiment were excised
from six enclosed blocks (microcosms) of intact Emory silt loam soil.
Each microcosm contained one red maple seedling (Acer rubrum). approxi-
mately 2 m high, with associated herbaceous ground cover. Three of the
microcosms were treated with primary lead smelter stack emissions
2
(equivalent to 12 mg Pb/cm ) while three served as controls. The
microcosms were maintained in a greenhouse for 20 months prior to the
excision of soil cores.
Duplicate intact soil cores (5 cm diam. x 10 cm depth) were excised
from each microcosm. Each core was divided at a depth of 5 cm to pro-
vide a depth-profile analysis of the soil. Cores from each microcosm
were encased in shrinkable PVC for leaching with extractants. The
duplicate core was dried and ground for extraction using the soil
slurry technique. Extractions and nutrient analyses were done as in
the previous experiment.
-------
209
4.2.3.2 Results and Discussion
Analysis of Extractants
Nutrient cations (Ca and NH--N) were extracted with 0.5 M^ KC1
and NH.OAc (Table 4.15). The extractants were chosen on the follow-
ing basis: 0.5 M KC1 is commonly used for extraction of NH--N and
NOo-N, and 0.5 M NH»OAc is commonly used for extraction of exchange-
able nutrient cations. Potassium chloride was chosen as the best
extractant for Ca and NHo-N because: (1) the amount of Ca extracted
from the tested soils was comparable to that extracted by NH.OAc, and
(2) extractable NHo-N could not be determined using NH.OAc as an
extractant. In most cases, concentrations of nutrient anions (NO-.-N,
P04-P, and DOC) extracted with KC1 and NH4OAc were 10 fold lower
than the same nutrients extracted with NaHCO~ (Table 4.15). Thus,
KC1 and NH^OAc were not used in subsequent experiments for extraction
of nutrient anions.
Nutrient anions were extracted with 0.5 M_ NaHCO., (Table 4.15).
This extractant removed high concentrations of NO.,-N, PO.-P, and
DOC (expected due to its high alkalinity, pH = 8.3), but was clearly
ineffective in removing Ca and NhU-N.
A successive leaching technique using KC1 followed by NaHC03 was
necessary to extract all nutrients from a single soil core. This
criterion was based on the fact that in some experiments only one soil
core is available for the extraction of all nutrients. Therefore, a
procedure was adopted whereby a soil core was leached with 200 ml of
KC1 extractant, rinsed (to remove excess KC1) with 200 ml of distilled
water, and subsequently leached with 200 ml of NaHC03 extractant.
-------
210
Extraction of PO^-P and DOC was lowered when NaHC03 was leached
through the cores after 0.5 M KC1 and distilled water (Table 4.15).
This indicates that a significant loss of these nutrients occurred from
the extraction with KC1. However, the advantages offered by the succes-
sive leaching technique outweighed the loss of PO.-P and DOC. Thus,
the successive leaching of soil cores with 0.5 M KC1, distilled water,
and 0.5 M NaHCOo was adopted as the standard method for the following
experiments. Extracts of KC1 were analyzed for Ca and NhL-N and
extracts of NaHC03 were analyzed for N03-N, P04~P, and DOC.
Comparative Extraction Techniques
All nutrients with the exception of DOC were extracted in much
greater quantities using the slurry technique (Fig. 4.3). The high
soil-to-solution ratio (1:5 for KC1 and 1:50 for NaHC03) and the
dispersion of soil colloids may account for these results. Values
obtained by this technqiue are generally regarded as the total exchange-
able or available amount of nutrients in the soil system (Soil Science
Society of America 1967).
Lower quantities of nutrients were extracted using the leaching
technique for intact and homogenized soil cores. The soil-to-solution
ratio was approximately 1:1. However, the difference in soil-to-
solution ratios did not account for the greater amounts of nutrients
extracted by the slurry technique.
The volume of extractant solutions (KC1 and NaHCOO was 200 ml
for this technique. Further analysis showed that up to 500 ml of solu-
tion could be leached through the soil cores before nutrient concentra-
tions were diminished in the effluent solutions. Therefore, the values
-------
211
(O
m
-------
212
reported for intact and homogenized cores are interpreted as an index
of available or exchangeable soil nutrients and not quantitative esti-
mates of their gross size.
Sensitivity Analysis
The sensitivity of the intact soil core and soil slurry techniques
was compared by extracting nutrients from microcosms which were treated
with Pb smelter emissions (Fig. 4.1). Results clearly show that levels
of Ca, NH3-N, N03-N, and DOC to a 5-cm depth extracted by the
intact method were decreased from the application of smelter emissions.
Identical results were obtained for NHL-N, NCU-N, and DOC to a 10-cm
soil depth. These data corroborate results showing increased gaseous
COp-C losses from the soil surface in addition to increased nutrient
losses via soil leachate from the Pb-treated microcosms.
Analysis of slurried soil samples did not distinguish between con-
trols and Pb-treated microcosms for all nutrients. In addition, all
nutrient pools extracted by this method except for Ca and NH,-N were
a factor of 10 greater than those extracted by the intact technique.
These results are similar to the previous experiment (Fig. 4.1).
The significance of this experiment is that extraction of gross
nutrient pools by the slurry technique was not sensitive to changes in
nutrient availability resulting from application of Pb smelter emis-
sions. However, the method of extracting intact soil cores was sensi-
tive to this effect.
In addition to indicating nutrient losses, the intact leaching
method distinguished differences between nutrient pools at 5 and 10 cm
soil depth. Extraction with the intact soil core technique indicated
-------
213
greater nutrient pool size at a 5-cm depth for all nutrients except
PCL-P and NHU-N. No differences were detected using the slurry
technique.
The detection of soil nutrient loss by the intact-core leaching
technique provides additional evidence that the application of heavy
metals affects nutrient cycling in forested ecosystems. Remaining
questions are: How much and at what rate are the soil nutrient pools
depleted? Will a loss of nutrients result in decreased autotrophic and
microbial growth? The answers to these questions lie in the degree of
heavy metal toxicity to these organisms. If growth is inhibited by
heavy metal toxicity, then the loss of nutrients will be further
enhanced, resulting in gross nutrient depletion, as in the case of
Crooked Creek Watershed (Jackson and Watson 1977).
Summary
A procedure was developed for extracting nutrient pools from intact
soil cores. Intact soil cores were leached with solutions of 0.5 N[ KC1,
distilled water, and 0.5 M NaHC03. In this manner, analyses were
completed for Ca, NH3-N, N03-N, P04-P, and DOC from a single soil
core. This technique was compared to the conventional slurry extraction
methods. Results indicate that a much greater quantity of nutrients
were extracted by the slurry method.
Experiments in this study were conducted on silt loam soil which
was easily excised intact. No significant differences were found
between nutrients extracted from intact and homogenized soil. There-
fore, leaching homogenized soil cores may be feasible on soils which
are not easily excised from the field.
-------
214
The leaching technique was found to be more effective than the
slurry method in detecting nutrient losses from Pb-treated microcosms.
Results from this experiment further substantiated the effects of heavy
metals on nutrient cycling in forested ecosystems.
Ultimately the leaching technique may be incorporated into a
protocol for assessing the potential toxicity of new chemicals to
terrestrial ecosystems.
4.2.4 Hollow Fiber Membranes for Detecting Microbial
and Nutrient Cycling Effects of Contaminants
This experiment tested the applicability of polycarbonate hollow
fibers within microcosms for collection of soil solution at various
depths to monitor contaminant and essential element concentrations in
solution. We also examined the feasibility of glass manifolds (fitted
with fibers for collection of litter and soil water in field plots) to
validate microcosm transport data.
The specific objectives of this study were to: (1) compare the
nutrient composition of soil water extracted from intact soil cores by
tn\
Diaflov ' hollow fibers to that collected by porous ceramic cup
lysimeter and soil leachate; and (2) evaluate the sensitivity of hollow
fiber and lysimeter extraction techniques to discriminate soil depth in
terms of dissolved nutrients in soil water from intact grassland soil
cores.
4.2.4.1 Materials and Methods
Three intact soil cores (30 cm diam. x 25 cm depth) were excised
from a managed pasture on the Oak Ridge Reservation. All intact soil
cores were fitted with a lysimeter and a bundle of hollow fibers
-------
215
mounted horizontally at 10- and 20-cm depths (Fig. 4.4). Each bundle
of four hollow fibers was joined at one end with glass tubing (9.3 cm
diam., 2 cm length) and plugged with RTV 3140 clear silicone rubber
adhesive. The opposite end was secured in a similar fashion, but was
left open, allowing a tube connection to pull suction from a vacuum
pump. The fiber bundles were installed by tunneling a passage through
the side of the core with a metal tube (0.3 cm diam., 40 cm length) and
threading the fibers through. The metal tube was then removed. The
entire core was wrapped in polyethylene sheeting and encased inside a
plastic tub with concrete for support and insulation. A leachate port
was constructed at the bottom of each soil core. Polyethylene tubing
(0.63 cm diam.) was imbedded in the concrete at the 10-, 20-, and 25-cm
depths for separate connections from the fiber bundles and the leachate
port to the outside of the tub. The soil cores were also fitted with
(D\ fD)
Irrometersv ' at 10- and 20-cm depths (Fig. 4.4). The Irrometersv '
were previously calibrated to indicate percent soil moisture.
Using estimates of soil core weight and measurements from the
(R}
installed Irrometersv ', the soil cores were watered to 130% field
capacity (F.C. = 25% soil moisture) once per week. The soil core was
allowed to equilibrate to field capacity. Water volumes greater than
field capacity (estimated to be 795 ml HpO) drained through the
leachate port and were collected for chemical analysis. Subsequently,
a small vacuum pump was attached with a sample collection vial to the
fiber bundle connections outside the tub (Fig. 4.4). A vacuum of 25 cm
Hg was applied for 5 to 6 hr to obtain 15 to 30 ml of solution from
soil surrounding each fiber bundle. Concurrently, porous cup lysimeters
-------
216
CO
o
o
cc
o
o
in
O)
o
a)
T3
en
c
(O
-l->
c
o
u
-r-
O -M
"3 3
IO ^
<4- CO
O
E 4J
(O -r-
i- in
fO o
E «
CD i-
.c ••->
O X
ai
Ł
3
D)
-------
217
were evacuated to a tension of 0.8 atm. Approximately 10 to 30 ml water
were pulled into the lysimeters through the porous cups after 5 to 6 hr
of extraction. Solutions were pumped from the lysimeters into sample
containers at the end of the collection period. Five successive weekly
collections of soil water were made in this manner. All solutions were
subsequently passed through 0.4-/I Nuclepore membrane filters. Concen-
trations of Ca and Mg were determined by flame atomic absorption.
lo\
Automated Techniconv ; procedures were used for NOo-N and PO.-P
determinations. Electrical conductivity (EC) measurements were made on
all samples.
4.2.4.2 Results and Discussion
Concentrations of Ca and Mg and mmhos/cm EC in soil solutions
extracted by lysimeter and hollow fiber techniques were significantly
greater (P Ł0.05) at 10-cm than at 20-cm soil depths (Table 4.16).
However, concentrations of NO^-N and PO^-P were the same between
these depths. Close agreement was found in EC and all nutrient concen-
trations between solutions extracted with lysimeters and hollow fibers.
In addition, no significant difference of nutrient concentrations except
NOg-N was noted between soil solutions extracted at 20 cm by lysim-
eters, hollow fiber techniques, and soil leachate collected at 25 cm
(Table 4.16). Screening of N03-N by lysimeters and hollow fibers was
indicated from these results (Table 4.16). Closer scrutiny is necessary
to determine if this effect is real or a unique characteristic of this
experiment.
Samples of sufficient quantity for chemical analyses were extracted
by hollow fiber and lysimeter techniques in reasonably short periods of
-------
218
O
O
CM
•o
c
(0
1
0
t— 1
E
o
&-
M-
T3
CU
-M
(J
*O
S-
4->
X ^"^
CO O
f"H
to II
c: c
o
• (— A
4-> •
3 LU
O OO
to
[
•r- +1
O — *
to
to
4- S-
0 CU
4_J
1 -^
— to
CT> >>
^^
•o
to c
c *o
o
•i- tO
4-> S-
(O CU
S- X5
•«-> •!-
C 4-
cu
0 3
c o
O i—
O i—
o
+J -C
c
CU -C
•1- +->
4J 3
3
c to
jr
c: 4J
n3 Q.
CU CU
2! T3
VO
,— 1
«^-
cu
S
to
\ —
:>>
r- -l->
tO -r~ ^~.
0 > E
$1 -Ł>^
•M O tO
0 3 O
cu -a ^:
UJ o •> •
0
Q.
'«*
O
Q.
cn
s:
(O
0
z
1
CO
S
0
V|- -i-
O 4->
CJ
cu
o
o
d
r-.
i— i
0
d
cu
-a
c
3
_a
cu
J3
•t—
LU.
0
r— 1
CO
O
O
O
CM
LO
0
d
CM
O
O
O
LO
o
0
d
r^
0
d
en
o
•
r— 1
en
co
o
«d-
CM
LO
(_^
0
0
d
ccT
o
o
o
.ysimeter
i
0
CM
LO
O
O
O
*±
LO
O
O
*3-
0
0
O
LO
0
o
o
o
t— 1
o
VO
0
•
r-H
r^
LO
•
o
CO
CM
LO
co
O
o
CD
CO
i— 1
O
CD
CU
-o
C
3
J3
CU
_a
LU
S
VO 3
VO <—
• to
f ;=»
Q
I/O
_l
.
*T
O
0
•
o
*d-
CM
0
CD
-I-"
eachate po
i
LO
CM
>1
i— -t->
tO -I-
O >
S- 4->
4-> O
O 3 O
CU TD t-H
r— C O
LU O
u o
CL CO
I O
«*• o
o
CL. O
o
r--.
CM
CD
s: o
co
LO
fO
0 t-H
Z LO
1 0
CO O
§ CD
•*— •"*
e oJ
O 4->
C CU
O O E
to CM M-
•i- to
S- tO >>
03 > r—
Q.
E Ł «
O OS-
O CU
O -O
1— 1 •!-
VI-
VO
O
O
CD
CM
i-H
o
d
•3-
Cn
r— 1
CD
Cr>
I--.
o
f— 1
CO
0
O
*
o— -
1 1
*T~
S-
o
Q.
E -
0 S_
CU
LO 4->
CM CU
to •!-
> to
>>
u
**
O i-
CM CU
J3
vT
-------
219
time. The solutions were clear and no visible residues were left on
the Nuclepore membrane filters after solutions were passed through.
Hollow fibers were flexible, easy to manipulate, and resistant to
breakage. The fiber bundles continued to extract water at comparable
rates throughout the experiment despite cyclic wetting and drying con-
ditions. Clogging was not a problem with either hollow fibers or porous
ceramic-cup lysimeters.
4.2.4.3 Summary
This experiment demonstrates that porous ceramic-cup lysimeters and
hollow fibers are viable techniques for in situ extraction of soil solu-
tion. Hollow fiber and porous ceramic-cup lysimeter techniques did not
affect EC and concentrations of Ca, Mg, and PCL-P, and were sensitive
to differences in dissolved nutrients with respect to soil depth. How-
ever, apparent screening of NOg-N was found to occur with both tech-
niques.
Hollow fibers are considered more useful for extraction of water
from greenhouse pots and intact soil cores because this technique
displaces a smaller volume of soil compared to the cup lysimeter. Cup
lysimeters are more useful for field sampling.
Laboratory Soil Water Extraction
(Rl
Ami conv ' hollow fibers can be used (in pairs) in intact soil
microcosms. They can be inserted as described above and monitored on a
regular schedule. Preliminary studies have shown that fibers can be
reused after long intervals of inactivity and can be successfully trans-
ferred to new experiments.
-------
220
Field Use of Fibers
/D\
Manifolds of glass might be useful in adapting Ami conv ' fibers
to field use, especially to extract litter solution entering soil
(Fig. 4.5). The manifolds were used by placing one unit at the inter-
face of 01 and 02 litter and a second at the litter-soil interface. A
moisture-activated switch engages a vacuum pump which applies tension
to fibers. Water extracted from litter percolate is collected in a
bottle attached to the vacuum pump.
4.2.5 The Effect of Substrate on Elemental Transport
Through Terrestrial Microcosms
The objective of this experiment was to determine the effect of
substrate on transport of essential elements through terrestrial micro-
cosms. Substrates used for microcosms are not now standarized. Micro-
cosm results have been reported using soil, sand, synthetic, and mixed
substrates. Substrate may significantly influence parameters estimated
by microcosm screening techniques. Essential element transport through
substrates used in microcosm units should furnish a good test of sub-
strate effect because (1) essential element transport is well documented
in terrestrial ecosystems, and (2) rates of essential element transport
have been shown to increase under perturbation.
The experimental design was randomized complete block design with
three treatments (intact soil, homogenized soil, and sand/soil sub-
strates), with two blocks with three replicates each. After encasement,
microcosms were equilibrated for three weeks and then monitored for 118
days without treatment. For 18 successive weeks leachate was collected,
following rainfall additions in excess of soil water-holding capacity.
-------
221
CO
00
CO
Q
cu
03
E T-
o -o
Qj -I—
_Q to
to
to
co 03
s- •—
0) CD
.a
•r- O
M- +->
'.
O
re
03 ^
> en
«4-
O
O
«o to
E S-
O) CU
-c .a
o -r-
M-
Lf>
CO
O
00
-------
222
Concentrations of essential elements in rainfall and leachate were then
determined. Elemental transport was calculated from concentration data
multiplied by the leachate volume from each microcosm, respectively.
4.2.5.1 Results and Discussion
Substrate had a significant effect on the amounts, rates, and pat-
terns of essential elemental transport through terrestrial microcosms.
Cumulative Transport
Substrate significantly affected total transport of essential
elements from microcosms over the 118-day experiment. Table 4.17
summarizes the total loss from each substrate of Ca, K, DOC, PO.-P,
NH3-N, N03-N, Na, and Mg. For each element examined there was a
significant difference in transport between intact soil and either or
both of the other substrates. There was a significant block effect
within homogenized soil microcosms for all elements measured. Transport
of Ca, Mg, Na, and DOC was greater in one of two blocks of homogenized
substrate and the sand/soil mixture. Potassium transport was greatest
from sand/soil, least from homogenized soil. Ammonia-N transport was
greater from intact soil than that from other substrates. Ammonia-N
loss averaged 59% nitrate-N loss weekly from intact soil microcosms
over the interval. Phosphate-P and N03-N transport were most variable
of any element examined. Block A homogenized soil microcosms trans-
ported more PO.-P but less NHU-N than intact soil. Maximum NO^-N
transport was observed in homogenized soil, Block B. Sand/soil micro-
cosms transported significantly less P04~P and N03~N than intact
soil.
-------
223
V)
0)
^>
rO
S-
10
to
CU
OJ
^
-M
4-
O
to
E
to
O
u
o
i-
o
'i
,
TO
•r-
to
CU
^.
i-
Q.
c x
CU CU
F
CU >>
.— ro
CU "O
CU CO
> i-H
•r— r-H
•4—*
*O (O
r—
3 S-
E CU
3 >
C_> 0
.
[^
i-H
•
^f
CU
^«-
JO
ro
1—
^
to
>•
ro
~C3
CO
i— I
i-H
•
'rt
O
1—
S-
o
Q
I/)
C
ro
S-
4_>
f»
rO
•P
C
CU
F
CU
^«-
UJ
o
S
Q.
1
^j-
0
Q_
•z.
1
CO
0
z
z:
i
CO
z
rO
"^
CD
s:
sx
ro
0
^— ^
to
CU
rO
•o
CU
I \ _
T^ •*
ro to
S- CU
•M -(->
to ro
JO 0
3 T-
00 •—
Q.
d)
i~
• — •
If)
Ol
CO
if)
t— t
0
O
00
(Ł1
00
•
O
o
CM
O
0
•
O
p^
ir>
•
o
CM
^~
•
CO
CO
to
co
X — ,
00
r^H
9\
to
4-^
0
ro
.fj
C
I—H
_Q
^S"
^
*^
T-— (
JCt
CO
un
o
CD
-Q
CM
^j-
CM
•
0
JO
CT>
0
<— t
If)
O
fO
^*
[N^
•
O
»O
10
<0
CM
<0
CO
i— 1
VO
^Ł
sx
O
o
CO
•o
CU
IM ^"^
•i- 00
C i-H
CU
CD «
O CO
Q
Z
00
0
co
i-H
I— 1
o
0
*
JO
If)
o
Lf)
•
rH
_Q
CT>
CO
i— 1
0
ro
O
CO
f— (
f-x^
LO
•
o
ro
CM
If)
•
CM
VO
•
co
CO
NX
U
o
CO
^— ^
CO
i-H
ft
CO
rO
CT>
Lf)
t
O
I— 1
ro
CO
0
O
•
O
JO
CM
"sj"
f— I
•
o
JO
LO
r- 1
O
ro
O
CO
O
fO
cr>
•
O
CT>
•
ir>
ro
cr>
•
^>
^— ^
CO
i-H
"
^,
^~
•r—
O
(/I
•o
C
rO
oo
•
LO
O
o
v|
o.
4^
(j
rO
^J
C
•r—
E
O
S-
^_
C
CU
S-
O)
M-
•f~
^^
>j
r—
C
ro
U
•r—
M-
C
CD
•r—
to
ro
•
T- 1
O
o
v|
D_
+J
U
rO
«(_>
C
•^
E
O
l-
4-
i %
•H*
C
CU
S-
cu
4-
4-
no
>j
r—
C
rO
O
•^
4-
C
CD
OO
-------
224
Transport Rates
Substrate significantly affected rates of elemental transport
through terrestrial microcosms (Table 4.18). For example, rates of DOC,
Ca, Mg, and Na were significantly less from intact soil than those from
other substrates.
Rates of elemental transport were quite constant in intact soil
microcosms through the 18-week experiment. However, rates of elemental
transport were erratic from homogenized soil and sand/soil substrates
through time. Too, variation in transport rates among replicates within
substrates was often 2 to 5 fold. For example, NO.,-N and NhL-N loss
rates were radically different between blocks of homogenized soil micro-
cosms. Over the 18-week experiment, rates of NhL-N loss decreased
within one block but increased within the other. The converse was
observed in NOo-N transport. Rates of DOC and Ca transport increased
with time in four of six sand/soil microcosms but remained relatively
constant from the remaining two replications. At the end of the experi-
ment, cumulative DOC and Ca transport was twice as great from this group
of four replicates of sand/soil microcosms as that from the others.
There was no apparent cause for these differences among replicates
within substrate groups. These differences were not detected in CO,,
efflux measured concomitantly and were minimal for transport of some
elements such as K and Mg.
Replicability
The replicability of observations made over the 18 weeks on ele-
mental transport through intact soil microcosms was equal to and/or
greater than that of other substrates. Mean coefficients of variation
-------
225
1
o
S-
u
• r—
E
'to
S-
4_)
CO
CU
S-
i-
1 \ ^^^^
CO
CD O
Z5 T-
0 4->
S- <0
-C >
+-> S-
O)
<— - CO
OJ 0
CU
s >>
E vx
CO CU
O CU
0 3Ł
o
S- CO
O t-i
• 1—
E ll
cf^
^
to
4-> CU
S- 4->
O ftf
Q. S-
CO 4->
C CO
tO *"**
S- :3
+-> CO
4-> CU
C 0)
CU S-
E -C
r—
CU 4-
tO T3
•r- CU
[ ^ 4_>
c o
0) 3
CO S-
CU CO
c
O (_>
CO CO
CU E
4J CO
CO O
o: o
^
CO
I— 1
^!
CU
r_
-Q
1—
- — <
LU
•
oo
Ix
4_}
S-
0
Q.
to
c
to
X-
1 \
r—
fO
4_3
c
O)
E
CU
1 —
LU
C_}
o
Q
Q.
1
^^J-
o
D_
^y
1
oo
0
•z.
00
~r~
•^
Z.
CD
s:
^
C_3
CD
•*->
i-
4_)
CO
f~\
•3
oo
^- cy>
• *
(T> ^D
rH r-l
CM
x— ^
«^~ f*-x
CO r- i
• •
o o
.. — ^
r^> ^o
• •
^^ ^"d~
^" *• — '
UD ^t-
r^. oo
CM •— -
f — x
LT> OO
CM i— 1
CM
r *
LT> O~l
i-H CM
oo —
. — x
^H OO
O O~>
Q^ ^ ^
!— 1
^-^
LO UO
«d- co
O i-l
CM
4_3
O
^_>
sz
1— 1
CO
•
ID
CTl
^d~
T3
oo
O
*
CM
LO
•
CO
^J-
CM
CO*
to
*3-
OO
i^Q
^J-
•
^o
oo
to
vr>
^.
oo
i— i
oo
^-
CO
CM
^
CU
N
c
CU
co
o
^
LO
•
LO
«xj*
•* — *-
^ — ^.
r-l
CO
•
o
f — -x
^~
•
^O
Si •-
CO
vo
^ — »
LO
CO
". •
S—**
OO
r-H
N ^
^ N»
CO
«^-
^ ^
, .
CO
oo
1— 1
*. ^
T— 1
•
LO
^J-
CTl
tO
CO
•
O
o
p-x.
•
r*x.
en
r^.
fO
oo
^j-
^j-
CO
CTi
OO
to
CM
to
CO
CM
O1
00
Oi
CM
;—
o
CO
i
T3
C
fO
oo
to
•
LO
OO
rH
'
^ —
00
o
•
o
.
^^
•
.— I
V '
r-H
t — 1
*.
^ — ,
to
CM
x
^—^
LO
•
OO
N '
^- *
CM
LO
^~
^ r
, s
CM
P^.
CM
s. ^
• •
LO i— 1
O O
0 0
V | V |
Q- D_
0 0
fO n3
C. d
E E
O O
S- S-
4- 4-
C C
CU CU
CU O)
4- 4-
4- 4-
•r- 'r^
•O T3
^ ^
c. c
fO fO
0 0
4- 4-
•i — -i —
Ł Ł
CD CD
•r— -r—
oo oo
fO ("">
-------
226
(C.V.) over the 18-week experiment were calculated for essential ele-
ments studied. Mean C.V. was not significantly different among sub-
strates for Ca, Mg, PO.-P, and total P. Mean C.V. was significantly
less in intact soil than that in homogenized soil for DOC and Na but
was significantly greater for K. Mean C.V. of sand/soil microcosms was
significantly less than that of intact soil for NCL-N and NH.,-N
transport but was significantly greater than that for transport of K
and S04-S.
Greater insight into the replicability of substrate elemental
transport was provided by calculating the probability (Z) that succes-
sive measurements of transport of any element from an individual micro-
cosm unit differ by greater than 100% (Table 4.19). In no case was the
probability greater for intact soil microcosms than that for other sub-
strates (P <. 0.05). Three elements showed a significant increase in Z
in homogenized soil microcosms: NHo-N, NO.,-N, and Mg. Sand/soil
microcosms showed the greatest probability of change (> 100%) between
successive observations. The Z values were significantly greater in
sand/soil microcosms than those in intact soil for six of the eight
elements measured: DOC, NH3-N, NOo-N, Mg, K, and Na. Neither Ca
nor POa-P showed significant differences among substrates. We, there-
fore, conclude that replicability was greatest in microcosms constructed
of intact soil.
4.2.6 Functional Complexity and Microcosm Stability
To rely on microcosms as useful screening tools for environmental
impact assessment, it is necessary to understand their long-term
response to extraction and maintenance under controlled environmental
-------
227
e
o
o
o
t-
o
• r—
E
c
ro
CD
O
S-
r~
4->
i_
0
Q.
C.
ro
4_>
1 —
1 1
CD
VU
'o;
Q
*^—
o
CO
f~
o
•r— * — .
+-> CO
(O -4-^
S- ro
OJ (j
CO -c—
_0 r—
O D-
O)
O) S-
> *^^
• r- (^
CO •— 4
CO
II
O
O Z
CO
"Ǥ
_C «-H
"^ A
—.Ł
>) S-
+-> Ol
•r— Vf_
•— M-
•i — 'i~~
t~*i "Q
ro
-Q 4->
O -i-
S- C
o_ rs
.
o^
.— 1
•
^J-
O)
r^
_o
ro
1—
• — •
UJ
oo
>,
-t-"
^
1^
ro
0
S-
Q.
S-
o
M
(O
^j.
O
D_
rj)
CO
O
^^
CO
•z.
ro
O
O
o
CD
4_>
ro
S-
1 ^
CO
J3
13
00
^ — ,
CvJ VO
^H O
O O
, .
VO VO
0 0
0 0
LO VO
VO O
• •
0 O
. — ,
CvJ VO
r-H 0
d o
„
LO CO
CO O
• •
0 0
^ — ^
vn vo
CvJ O
0 0
„ — ^
*3- VO
CvJ O
O O
^^
^" vo
CvJ O
O O
»-
4_3
O
rO
r—
t— <
00
T— 1
O
0
o
I— 1
fv^
•
0
ro
^j-
CvJ
Q
ro
CO
LO
•
0
ro
^
•
0
oo
o
LO
CO
d
-a
O)
IM
•t—
C
VII
en
O
E
0
JC
. — ^
VO
o
0
^
VO
0
0
Vo"
o
•
o
N •
.
VO
o
o
— •
^-^
8
•
o
.*
^-^
8
d
J — ^
8
o
. _
CVJ
1 — 1
•
o
^^^n
ro
^j-
CM
o
ro
CvJ
O
LO
VO
•
o
ro
.— I
O
ro
LO
VO
*
O
rO
VO
LO
C)
.—1
O
ro
^>
•
O
r—
•r—
O
CO
1
•^
c
(O
00
. — ^
VO
o
o
* — ^
§
o
VO
O
•
o
, *
VO
0
d
, — ,
VO
o
•
o
„ — ^
I—)
•
o
, ^
•—I
T— 1
o
. ,
I — 1
I— 1
•
0
,
LO
o
o
V
D_
^_>
ro
O)
+->
ro
S-
co
r*i
13
CO
0
ro
•-
ro
O)
S-
CT)
3?
C
ro
O
• r—
*|
• r—
C.
CD
•r—
OO
rO
-------
228
conditions. It is also important to ascertain the stability of each
microcosm and to determine if differences in stability are related to
the functional complexity that is intrinsic to each microcosm system.
The objectives of this experiment were: (1) to develop valid
procedures for encasing terrestrial microcosms and to determine if
materials used have any effect on parameters being measured, (2) to
develop techniques for measuring total terrestrial microcosm metabolism,
and (3) to develop a method of determining the relative functional
complexity of each microcosm and to investigate its relationship to
stability.
A central hypothesis in ecology is that system stability is in
some way associated with complexity (e.g., MacArthur 1955, May 1973).
The issue has been examined in some detail through mathematical models.
In addition, limited data on interacting populations have been collected
which in some cases support a stability-complexity relationship. In an
effort to extend the experimental evidence on this theme to the ecosys-
tem level, a stability experiment using old field microcosms was per-
formed.
Correlated questions also were investigated: (1) how do terres-
trial microcosms behave under regulated environmental conditions over
long periods of time without being perturbed; (2) what is a reasonable
equilibrium time for microcosms of this type; and (3) how will season-
ality affect microcosm results. These objectives and correlated ques-
tions relate directly to project goals of determining the reliability
and utility of microcosms as reasonable screening tools for transport,
fate, and effects of toxic substances.
-------
229
4.2.6.1 Materials and Methods
Thirteen intact cores were extracted from the approximate center of
an old field located on the U. S. Department of Energy reservation at
Oak Ridge, Tennessee.
Each microcosm, with litter and autotrophs intact, was trimmed to
predetermined size (15 x 10 cm) and sleeved with 0.10-cm thick polyvinyl
chloride (PVC). A Plexiglas disk, with leachate port attached, was
pressed to the bottom of the soil profile, and the PVC heat shrunk until
taut. The encased microcosm was placed in a small polyethylene pot,
medium grain quartz sand packed between the PVC and pot, and the micro-
cosm placed in the environmental chamber (Fig. 4.6).
Complexity of each microcosm was characterized by analyzing system
dynamics evident in COp efflux measurements. Carbon dioxide efflux
is an ecosystem-level variable which to some degree integrates the
dynamics of all system components. Hourly measurements of C0? efflux
were obtained for 2050 hr under controlled environmental conditions
using infrared gas analysis techniques. Sample data are displayed in
Fig. 4.7. The C02 time series is dominated by cyclic dynamics. The
power spectral density was calculated for the C0~ time series for
each microcosm using standard methods (Blackman and Tukey 1958). This
spectral analysis elucidates the relative importance of cyclic compon-
ents of different periods required to decompose the time series into
sinusoids. It is assumed that each significant cyclic component (i.e.,
peaks in the power spectral density) was the result of functional inter-
actions (e.g., predator-prey interaction) in the microcosm, so that the
number of peaks in the power spectral density is hypothesized to be a
measure of the dynamic or functional complexity of the system.
-------
230
ORNL DWG 76-18158
IRGA
(1 liter/mm)
RAIN INPUT PORT
DEFOGGER
3.2mm THICK, 18 4cm
DIAM. PLEXIGLAS
6.4mm THICK,16.5cm INSIDE
DIAM X 430cm HIGH PLEXI-
GLAS
AMBIENT AIR INTAKE
THERMOCOUPLES
BLACK PLASTIC CONTAINER
, ] /'"• i ' ' / / /
HEAT SHRINKABLE POLYVINYL
I CHLORIDE
DOW CORNING 3140 RTV
SEALANT
PLEXIGLAS DISK
PRESSURE BYPASS
FIBERGLASS
INSULATION
MICROCOSM
15cm DIAM X 10 cm DEEP
LEACHATE COLLECTION BOTTLE
(250ml)
IGNITED, DISTILLED WATER
WASHED SAND
Figure 4.6. Cross-section of old-field microcosm showing two
system-level monitoring points of C02 efflux and
nutrient leachate.
-------
231
15
~ 10
k_
.C
1 5
o>
- 0
X
ORNL-DWG 77-13383
h
UJ
o
o
-5
i
-10
-15
0
-15 -
100
200 300
TIME (hr)
400
500
-20 4
0 1000 2000 3000 4000 5000
TIME (hr)
Figure 4.7. Example of time series of C02 efflux from
pasture microcosms.
-------
232
The stability of each microcosm was established by performing a
perturbation experiment using cadmium in toxic concentrations. A dose
2
per unit area of 0.4 mg/cm was chosen. The response of the micro-
cosms to this perturbation was monitored by weekly measurement of the
export of Ca in leachate. Typical response of this variable to the
perturbation is illustrated in Fig. 4.8. Stability, in the sense of
resistance and ability to recover from perturbation, was summarized by
calculating the area under the Ca response curve. Stability ranking by
this measure and complexity are summarized for each microcosm in
Table 4.20.
Encasement Procedures
Encasement of microcosms in epoxy proved to be an unreliable
method due to both the chemical nature and the lack of predictability
of hardening of the epoxy. The epoxy was found to infiltrate the entire
soil column and harden incompletely. In addition, tests of Teachability
of epoxy were made. These tests indicated that the epoxy leached dis-
solved organic carbon over a long period of time. The amount of leach-
ing was directly related to the relative hardness of epoxy. Thus, a
new encasement technique was developed and tested for this type of
intact microcosm.
The new encasement method follows the procedure described below:
(1) Place a 15.5-cm diam. sleeve of polyvinyl chloride (PVC)
around the 15-cm diam. microcosm.
(2) Place a 0.64-cm thick x 15 cm-diam. Plexiglas disk with a
1.23-cm diam. hole in the center, inside the PVC sleeve.
-------
233
40
30 -
Ł
0
ORNL-DWG 77-9486A
0.028
0.021
MAY JUN JUL AUG SEP
TIME
Q)
0.014
ro
o
0.007
Figure 4.8. Mean DOC ( ) and N03-N (-
leachate (± 1 S.E., N = 13).
0
OCT
—) concentration in
-------
234
Table 4.20. Rankings of functional complexity and ecosystem stability
based on monitoring of C02 and total Ca export in soil
leachate
Microcosm number Peaks
1 10
2 12
3 13
4 11
5 16
6 16
7 9
8 10
9 13
10 11
Functional
complexity
Rank
8.5
5
3.5
6.5
1.5
1.5
10
8.5
3.5
6.5
Ca loss
289.5
203.6
315.5
379.3
133.3
314.0
434.4
543.5
353.0
420.6
Ecosystem
stability
a Rank
3
2
5
7
1
4
9
10
6
8
aCa loss expressed as total peg.
-------
235
(3) Place a small bead of Dow Corning RTV sealant along the edge
of the Plexiglas disk.
(4) Heat-shrink the PVC until skin-tight around the microcosm and
the edge of the Plexiglas dish.
(5) Remove excess PVC from both top and bottom of the microcosm
leaving approximately a 2.5-cm lip on top.
(6) Tighten a hose clamp around the outside edge of the Plexiglas
disk which has been covered by the PVC. This ensures bonding
between the Plexiglas and PVC. Additional RTV may be applied
around the edges from the bottom as needed.
(7) Allow 24 hr for the RTV to harden, insert plastic port into
hole in Plexiglas disk, and place encased microcosm into a
(18-cm diam. x 14-cm high) plastic bucket painted black to
reduce light transmission.
(8) Back fill between the PVC-wrapped microcosm and container
with pre-ignited quartz sand that has been washed with acid
and distilled water.
Figure 4.6 shows a cross-sectional view of an encased microcosm
and the cuvette designed for CO^ monitoring.
Materials Tests
Due to the problems of leaching encountered with epoxy, all
materials used for encasement have been tested for their potential
effect on nutrient concentrations in the leachate. An experiment was
designed to test each of the materials that potentially could come in
contact with the leachate under similar pH conditions (rainwater
-------
236
pH = 6.40) and constant contact over a reasonable length of time. The
materials tested include vinyl tubing, leachate collection bottles,
leachate ports, sand, PVC, RTV sealant, nuclepore filters, Plexiglas,
and teflon. All materials were placed in 125-ml flasks (N=3) with
100 ml of distilled water (pH 6.4) and shaken at 250. Aliquots of the
100-ml solution were removed from each sample after 2, 4, and 8 days.
After each sampling period, the distilled water was changed. Each
sample at each time period was subsequently analyzed for Ca, DOC, and
N03-N.
Analytical results, after adjusting element concentrations of each
sample at each time period for distilled water concentrations as back-
ground, showed that the only encasement material likely to leach any of
the nutrients analyzed is RTV coating. It was found that the RTV coat-
ing, under these rigorous test conditions, might have a small effect on
the DOC concentration. However, this possible effect diminished with
time. Due to the limited contact that might occur between the leachate
and the RTV, this effect is judged to be insignificant. All other
materials (vinyl tubing, leachate collection bottles, leachate ports,
sand, PVC, nuclepore filters, Plexiglas, and teflon) did not signifi-
cantly affect Ca, DOC, or N03-N concentrations during any sampling
period.
4.2.6.2 Results
Nutrient Losses
Nutrient leachate data from 13 intact old-field microcosms (15 cm
diam. x 10 cm deep) containing naturally occurring autotrophs have been
-------
237
compiled during an eight-month equilibration period. During this
period, weekly amendments of collected rainwater (of a known volume and
nutrient concentration) were added to each microcosm. Leachate volumes
for each microcosm were recorded, an aliquot was filtered through a
0.45-/zm Nuclepore filter, and analytical determinations for Ca, DOC,
N03-N, and NhU-N were made.
Results show that there was a significant increase in the DOC con-
centration in early September and a corresponding decrease in NO^-N
at the same time (Fig. 4.8). These results were compiled during unper-
turbed conditions and are postulated to be associated with the natural
senescent behavior of the autotrophic component and microbial immobili-
zation of essential nutrients during simulations of fall conditions of
light and temperature. This demonstrates that seasonality, if not
accounted for, might obscure results of contaminant effect on nutrient
dynamics.
Microcosm Metabolism
The objective of monitoring C02 was to test whether or not
individual temperate grassland microcosms, with the autotrophic
component intact, exhibit significant differences in their functional
complexity. This was ascertained through the use of power spectral
density analysis of C0~ efflux measurements. The Fast Fourier
Transform algorithm was used to compute the power spectral density of
the equally spaced time series of COp only after the data have been
detrended and smoothed. Once this was accomplished, the next step was
to determine if there is a correlation between these dynamic patterns
and the relative stability of each microcosm. In this case, stability
-------
238
was measured as the percent deflection of nutrient losses from normal
operational patterns and the length of time necessary to return to some
stable state after perturbation.
Due to the complexity of the question, it was necessary to design
and fabricate an automatic sampling system which would serially sample
an air stream for C0? concentration being drawn from the enclosure of
each microcosm (Fig. 4.6). The sampling system is designed to draw air
from each microcosm at the rate of 1.0 liter/min continuously, with the
air stream from each microcosm being sampled hourly for CCL concentra-
tion using a Beckman 215A Infrared Gas Analyzer (IRGA). This allows
for collection of an uninterrupted, equally spaced, time series of mg
C02/nr for all eleven microcosms. To date, the system has been
operational for 130 days with > 1% downtime.
A representative example of the (XL time series is shown in
Fig. 4.7. These were examined using power spectral density analysis
for each microcosm. There was little or no difference in rhythmic
behavior among microcosms as ascertained through power spectral density
analysis. This interpretation was highly influenced by the magnitude
of the peak associated with the diel rhythm. After fitting and removing
from the data a sinusoid of this dominant frequency, we were able to
detect high period patterns which occurred at low frequency. Thus, the
C0? data indicate that all microcosms seem to be operating with the
same general trends and oscillatory patterns; however, there are some
subtle differences in rhythmic patterns in the low frequency regions.
We conclude that power spectral density analysis is a useful analytical
tool to reveal differences in functional complexity and subsequently
show its relationship to the relative stability of each microcosm.
-------
239
4.2.6.3 Discussion
The microcosms were rank-ordered, based on stability and complexity
measures as described above. Spearman's rank correlation test was used
to test the hypothesis that stability does not increase as complexity
increases. The hypothesis was rejected at the 95% confidence level.
Further details of this experiment to study the relationship
between complexity and stability in laboratory ecosystems may be found
in Van Voris et al. (1978). The experimental verification of an
increasing stability with increasing complexity hypothesis described
here is of course limited. Rather than providing results which con-
clusively establish the relationship between complexity and stability,
this experiment demonstrates clearly that the manner in which stability
and complexity are characterized and the level at which experiments are
performed are critical to verifying theoretical conjectures associated
with stability and complexity.
4.3 SYNTHESIS OF TERRESTRIAL MICROCOSM RESULTS (Monte Ross-Todd,
E. G. O'Neill, and R. V. O'Neill)
4.3.1 Introduction
The microcosm project involved more than a dozen individual experi-
ments designed to examine the scientific basis of microcosms as screen-
ing tools. The individual experiments focused on specific problems of
size, substrate, parameters to be monitored, etc. These experimental
results form the main output from the project and have been documented
in a series of open literature papers.
While the papers focus on results of single experiments, the number
of experiments performed over the course of the project requires that
-------
240
some effort be expended in looking across the suite of experiments to
identify patterns that might not be identified in individual papers.
The objective of the synthesis project was to compare the results of
all terrestrial experiments with a view toward answering the following
questions: (1) What is the minimum set of nutrient parameters which
should be measured in a large-scale monitoring system?; (2) Is there a
significant effect of size of the microcosms on their ability to detect
ecological changes resulting from toxic substances?; (3) Is there a
significant effect of soil type or soil preparation (i.e., intact versus
homogenized core) on ability to detect change? These three questions
provided the focus for this project and will be considered in individ-
ual subsections of this report.
Table 4.21 outlines the characteristics of the eight terrestrial
experiments which formed the material for this synthesis effort. Each
experiment has been designated by a code name given in the first column
of the table. These names will be used for convenience throughout this
section. It is clear from the table that a great variety of experimen-
tal designs were utilized and this seriously constrains the types of
cross-comparisons that are possible. Thus, it was difficult to find a
single factor which differed between two experiments while all other
factors remained constant. In spite of the difficulties inherent in
the effort, a number of interesting results were discovered.
4.3.2 Minimal Monitoring Parameters
The suite of seven chemical species which were measured through
most of the terrestrial microcosm experiments (see Table 4.22) is too
large a number to be practicable in a routine screening program. The
-------
241
Table 4.21. Summary of microcosm experiments and corresponding parameters
Grassland I
Grassland II
Van Voris
Treecosm I
Treecosm II
Washburn
O'Neill
Cd dosage
Substrate(s)
Grassland
Grassland-
1 . Homog .
2. Intact
Grassland
Forest
Forest
Grassland
and forest
Forest-
1. Intact
2. Homog.
3. Sand-soil
Forest
Pine stand
Size(s)
(cm)
10x10
30x25
15x25
30x25
15x10
63x63x45
63x63x45
5x5
5x5
5x5
Location(s)
Greenhouse
and field
Slathouse
and field
Environ.
chambers
Greenhouse
Slathouse
and field
Environ.
chambers
Environ.
chambers
Environ.
chambers
Nutrients
measured Vegetation
Ca, Mg, DOC,
N03, P, K
Ca, Mg, DOC
N03, P, K
Ca, Mg, K
P, DOC, N03,
NH3
Ca, Mg, K
Cd, Cu, Pb, Zn
Ca, P, N03
NH3, DOC, K
Ca, Mg, K, P
P, N03, NH3
Ca, K, DOC
P, N03, NH3
Ca, DOC, N03
Yes
Yes
Yes
Yes
Yes
No
No
No
-------
242
Table 4.22. Comparison of variability among elements leached from terrestrial
microcosms. Values are coefficients of variation
(S.D. x 100/mean) averaged across all treatments in the
experiment. "R" indicates variability among replicates
(C.V. calculated for each leaching date and then averaged).
"T" indicates variability through time (C.V. calculated for each
core and then averaged).3
Grassland I
Grassland II
Van Voris
Treecosm I
Treecosm II
Washburn
O'Neill
Dosage Cd
R
T
R
T
R
T
R
T
R
T
R
T
R
T
R
T
Mg
57
5ff
57
55
23
5T
36
5ff
30
5T
32
IX
Ca
63
39
73
60
27
FT
41
59
51
55
34
53
35
5?
49
55
DOC
65
78
62
5?
45
79
72
Zi
33
92"
52
84
41
64
K
84
Ł8
88
80
52
55
50
161*
70
Z5
53*
71
42
48
N03
126*
144*
122*
117*
72
85
174*
67*
136*
131
52
100*
69
163*
68*
66*
P
116*
147*
116*
147*
113*
183*
90*
115*
103*
149*
72*
116*
81*
169*
NH4
78*
96*
93
203*
89*
156
aThe two lowest coefficients are underlined in each comparison and the two
largest are indicated by an asterisk.
-------
243
first objective of the synthesis subproject was to examine all the
experimental data and determine objective, analytical approaches to
reducing the number of species to be considered in later applications
of the soil core microcosms.
The first step was to analyze variability in the leachate data in
an effort to locate chemical species which showed little variability
during the experiments and would, therefore, be likely candidates for
reliable, consistent indicators. Chemical species which showed them-
selves to be highly variable either between replicates (R in Tables 4.22
and 4.23) or through time (T in Tables 4.22 and 4.23) would be expected
to be unreliable indicators of disturbance to the system. They would
be unreliable because it would be difficult to differentiate random
fluctuations from treatment effects. All other factors being equal, it
would require fewer replicates to detect a disturbance in a chemical
species with low variability between replicates.
We have calculated coefficients of variation for each of the
experiments, either lumping all treatments (Table 4.22) or considering
only controls (Table 4.23). Overall examination of the tables reveals
a definite pattern. Calcium, magnesium, and DOC consistently showed
lower variability, and NO-, P, and NH. were consistently higher.
This would indicate that Mg, Ca, and DOC could be expected to be reli-
able indicators of disturbance to the soil core systems.
Variability alone is not an adequate indicator. A chemical species
could be a consistent, invariate parameter; indeed, so invariate that
no effect of a perturbation could be detected at all. Therefore, the
analysis for variability must be supplemented with analysis of
-------
244
Table 4.23. Comparison of variability among elements leached from terrestrial
microcosms. Values are coefficients of variation
(S.D. x 100/mean) for control groups. "R" indicates variability
among replicates (C.V. calculated for each leaching date and
then averaged). "T" indicates variability through time (C.V.
calculated for each core and then averaged).3
Mg
Ca
DOC
NOc
NH4
Grassland I
Grassland II
Treecosm I
Treecosm II
Washburn
O'Neill
( intact)
Dosage Cd
R
T
R
T
R
T
R
T
R
T
R
T
R
T
53
5?
53
5?
20
H
28
M
40
31
63
68
58
83
23
51
50
51
27
H
42
44
56
H
64
80
54
57
35
51
34
93
37
5T
43
71*
61
7?
69
68
46
126*
55
84
55
69
47
38
79*
151*
123*
140*
140*
69*
114*
148*
57*
102*
87*
112
64*
70
89*
165*
85*
114*
61*
109*
70*
95
63*
98*
81
173*
63
159*
91*
115*
aThe two lowest coefficients are underlined in each comparison and the two
largest are indicated by asterisk.
-------
245
sensitivity. In one approach we performed an analysis of variance on
all the experiments. Table 4.24 gives the alpha values for the signifi-
cance of a treatment effect when the experiments are analyzed for each
element separately. By examining the alpha values which are less than
0.01, it can be seen that Mg and Ca show reasonable levels of sensitiv-
ity, probably due to the low variability of these elements. The DOC
appears less reliable than when variability alone is considered, being
quite insensitive in four of the experiments. Phosphorus and NH.
continue to show themselves as less than useful. The most significant
new information gained from this analysis is that NO, now appears as
a sensitive indicator. That is, in spite of the variability of NO.,
data, the response to perturbation was large enough to clearly indicate
a response.
Table 4.25 shows a different indicator of sensitivity. In this
case the three pretreatment leachates were averaged and compared to the
three posttreatment leachates. A measure of sensitivity would then be
the difference in these means divided by the variances, i.e., a calcu-
lated t value, which is presented in the table. It should be emphasized
that this is not proposed as a test of significance, which is the usual
connotation within which the t value is calculated. Here we are simply
interested in the sensitivity of the chemical species. Therefore, the
difference between pre- and posttreatment mean values, weighted by the
total variance, is a reasonable representation of sensitivity. It is
incidental that this measure is equal to a calculated t.
Table 4.25 confirms the impressions of sensitivity which were
gained from Table 4.24. Magnesium and Ca continue to appear as
-------
246
Table 4.24. Alpha values for the significance of treatment (toxic substance)
effect in eight experiments. Results are shown for total element
export (not concentrations).3
Grassland I
Grassland II
Van Voris
Treecosm I
Treecosm II
Washburn
O'Neill
Cd Dosage
Mg
NSb
0.0022
0.0005
0.0005
0.0308
Ca
NS
0.0001
0.0001
0.0520
0.0315
0.0237
0.0114
0.0001
DOC
NS
0.0001
0.0159
P_._P_go_i
NS
NS
NS
K
0.0372
0.0001
NS
0.0010
NS
NS
0.0220
N03
0.0026
0.0001
0.0115
0.0033
(hOOOJ.
0.0006
NS
0.0001
P
NS
0.0001
NS
NS
NS
0.0002
NS
NH4
NS
NS
NS
Underlined values are significant.
^Effects that are not significant at the 0.05 level are indicated by NS.
-------
247
Table 4.25. Average "t" values compared among elements. Values are based
on three pretreatment leaching samples compared with three
posttreatment leaching samples. A value was calculated for
each replicate and averaged across treatments, substrates,
location, etc. The calculated "t" values do not represent
any test of significance but represent an index of sensitivity
to perturbation.3
Mg
Grassland I 1.087
Grassland II 2.33
Van Voris 0.85
Treecosm II
O'Neill
Dosage Cd
Ca
0.63
1.88
0.73*
1.08
1.71
1.67
DOC
0.14*
0.63*
2.59
0.30*
0.23*
2.62
K
0.44
0.17*
0.91
1.55
1.24
N03
1.81
1.12
1.69
2.72
1.92
0.89*
P
0.38*
1.38
0.57*
2.64
0.83
NH4
1.74
0.22*
0.71*
aThe two largest coefficients are underlined and the two lowest are indicated
by an asterisk.
-------
248
reliable and sensitive indicators. The DOC again shows itself as
unreliable in many experiments, and NO, again appears as a sensitive
indicator. Phosphorus and NH. show little to recommend them.
In all of the tables, K appears in an intermediate position. It
is neither the most variable nor the least. It is neither the most
sensitive nor the least. Thus, while there is little evidence to
eliminate K, measurement of this element would certainly be given lower
priority.
On the basis of the variability and sensitivity tests, the follow-
ing conclusions can be reached. Although NO., is highly variable, it
consistently shows itself to be a sensitive indicator and should be
definitely included in screening programs. Magnesium and Ca are
extremely reliable indicators and show sufficient sensitivity to be
useful in screening programs. Phosphorus and NH, can be eliminated
as neither reliable nor sensitive. The DOC appears consistent but
relatively insensitive and cannot be given high priority in any screen-
ing application. Potassium retains an intermediate, low priority posi-
ti on.
One further piece of information can be put forward in the attempt
to reduce the number of chemical species to be measured. Table 4.26
shows that the behavior of Mg and Ca are highly correlated in all the
experiments in which they were both measured. As should be intuitively
obvious from the nature of these cations, it would be sufficient to
measure one or the other in any screening application.
In sumnary, it would appear that NO., plus either Ca or Mg would
serve as a practical pair of chemical species to be recommended for a
-------
249
Table 4.26. Correlation coefficients for Ca and Mg export.
Analysis was based on total export data.
Grassland I 0.87
Grassland II 0.99
Van Voris 0.97
Treecosm I 0.78
Washburn 0.83
-------
250
screening program. Either one or the other, or both would have shown a
significant treatment effect at the 0.05 alpha level (see Table 4.24) in
all the experiments performed. Therefore, the synthesis of the experi-
ments indicates that reduction of the monitoring program to these two
measurements would significantly reduce the cost of a screening program,
yet it would introduce minimal risk that a significant effect would be
overlooked. The addition of K or DOC as another indicator might be
considered, but our data indicate that they are unlikely to show an
effect that would not also be detectable in NOo and Ca or Mg.
4.3.3 Size Effects
Examination of Table 4.21 indicates that comparisions across
microcosms for the effects of size are not feasible. There are too
many confounding variables to be able to select two experiments which
differed in size of microcosm but in no other independent variables.
Therefore, the analysis of the effect of size must be restricted to the
Grassland I and Grassland II experiments, within which size was con-
sidered.
Based on nutrient variability and sensitivity, an analysis of
variance was performed for Ca, Mg, NO.,, and DOC, considering both
total export and concentration. Tables 4.27 and 4.28 show means and
standard deviations for these quantities aggregated across all treat-
ments and then separated by treatment level. For noncontrol microcosms,
the analysis was performed both pre- and posttreatment. Based on the
calculated F-values, the size effect was significant at the P <^ 0.01
level for all of the comparisons indicated by an asterisk. Significant
-------
251
Table 4.27. Microcosm size comparison on intact cores in Grassland I where
size was a variable. Analysis of variance was performed over all
treatments; controls only; lowest dose only; highest dose only.
The last two computations were broken into pre- and posttreatments
to see if there was a different response after As was added.3
As dose
0.0 As9
0.05 mg As
per cm'-
5.0 mg As
per cm?
Nutrient
ExpCa6
ExpNOa
ExpDOC
ExpMg
N03
DOC
Mg
ExpCa
ExpNOs
ExpDOC
ExpMg
Ca
N03
DOC
Mg
ExpCa
ExpN03
ExpDOC
ExpMg
Ca
NOs
DOC
Mg
ExpCa
ExpNOs
ExpDOC
ExpMg
Ca
NO 3
DOC
Mg
Small b
x" S.D.
1493 ±1612
972 ±1369
7694 ±8892
407 +477
8.7 +7.1
4.1 ±6.1
68 +44
2.9 ±1.9
1244 ± 1287
891 ±973
5873 ±5893
371 ±425
8.1 ±7.6
5.1 ±7.8
54 ±34
2.3 ±2.0
_
-
-
-
6.7 ±4.8
0.28 ±0.37
93 ±45
1.9 ±1.5
_
-
-
-
5.7±4.5
0.4 + 0.64
84+ 50
1.4± 1.3
Large0
X" S.D.
1569
697
8969
421
* 14.3
4.2
79
4.4
* 2405
1564
9005
567
* 26.2
10.6
* 142
* 5.1
_
-
-
-
9.2
* 0.16
71
2.2
_
-
-
-
9.9
* 0.11
* 52
2.4
1283
1123
15212
492
15.8
9.1
79
3.1
1948
2309
9857
519
37.5
23.3
160
4.9
4.7
0.34
47
1.4
7.1
0.21
28
1.8
Large
7 S.D.
1508 ±1657
549 + 840
7186 ±6404
420 ±463
10.0±6.3
3.8±4.4
54 ±41
2.8± 1.8
1728 ± 1893
1476 ± 2293
10023 ± 14378
430 ± 453
13. 6± 12.0
13. 1± 15.6
70 ± 58
3.3± 2.6
Small
X S.D.
1173 ±891
* 80 ± 167
8166 ±14810
301 ± 272
* 6.4±3.4
*0. 369 ±0.658
* 35 ±40
* 1.6±1.1
1129 ±1011
446 ± 893
9735 ± 20970
395 ± 684
* 7.6±6.4
* 3.0±6.5
* 34 ± 39
2.9±5.6
Differences (p 5 0.01) are indicated by an asterisk.
b!0 x 10 cm.
C30 x 15 cm.
^Analysis of variance and means across all doses, pretreatment and post-
treatment combined.
eExport values = volume leachate x concentration in ppm leachate.
fConcentration in ppm leachate.
9Controls compared, pretreatment period and posttreatment combined.
-------
252
Table 4.28. Microcosm size comparison on intact cores in Grassland II.
Analysis of variance was performed over all treatments;
controls only; lowest dose only; highest dose only. The last
two computations were broken into pre- and posttreatments to
see if there was a different response after As was added.3
As dose
Alld
0.0 As9
0.05 mg As
per cm?
5.0 mg As
per cm^
Nutrient
ExpCa6
ExpNOs
ExpDOC
ExpMg
Caf
N03
DOC
Mg
ExpCa
ExpNOs
ExpDOC
ExpMg
Ca
N03
DOC
Mg
ExpCa
ExpNOs
ExpDOC
ExpMg
Ca
N03
DOC
Mg
ExpCa
ExpNOs
ExpDOC
ExpMg
Ca
N03
DOC
Mg
Smallb
(X S.D.)
1051 ±1259
130 ±226
1696 ± 1460
123 ±84
3.8 ±1.0
0.395 ±0.45
8.9 ±4.2
0.6 ±0.27
1480 ± 2229
13 ±16
1287 ±996
90 ±53
4.96 ±3.99
0.086 ±0.08
6.87 ±3.0
0.6 ±0.38
710 ±643
13 ±15
1423 ±1042
147 ±138
3.9 ±0.59
0.06 ±0.12
4.0 ±6.5
0.8 ±0.2
712 ±538
17 ±21
1434 ±872
142 ± 86
3.1 ±0.9
0.12 ±0.14
8.45 +2.28
0.8 ±0.4
Large0
(X~ S.D.)
* 5535 ±5533
98 + 107
* 6219 ±4829
* 629 ±473
* 7.8±4.1
* 0.11 ±0.12
* 10.6 ±5.7
* 1.3 ±0.8
9109 ±12582
25 ±38
1965 ±2543
633 ±2543
11.6 ±5.8
0.05 ±0.08
7.6 ±3.0
1.9 ±1.5
* 1755 ±1919
20 ±34
* 3420 ±2922
* 412 ±407
4.5 ±0.8
* 0.12 ±0.08
10.2 ±4.7
* 1.2±0.4
2862 ±4536
19 ±40
2403 ± 2355
294 ± 339
* 10.1 ±9.1
* 0.05 ±0.08
9.4 ±3.8
1.1 ±0.6
Large
(7 S.D.)
809 ±866 *
73 ±74
1714 ±1146*
100 ±51 *
2.2 ±1.0 *
0.23 ±0.20*
5.7 ±2.8 *
0.9 ±0.1 *
1112 + 1049 *
650 ±1215
4567 ±3711 *
168 ±122 *
3.1±1.3 *
1.7 ±2.1 *
14.6±7.7 *
0.5 ±14.5 *
Small
(JTs.D.)
3277 ±3855
77+119
5775 ±4352
348 ± 185
3.3+1.2
0.12 ±0.20
7.8±5.1
0.6 ±0.2
7096 ±4775
424 ± 371
21784 ± 14259
1451 ±1075
5.4 ±1.8
0.36 +0.24
21.1 ±0.2
1.08±0.31
Significant differences (p <_ 0.01) are indicated by an asterisk.
b!5 x 25 cm.
C30 x 25 cm.
dAnalysis of variance and means across all doses, pretreatment and post-
treatment combined.
eExport values = volume leachate x concentration in ppm leachate.
fConcentration in ppm leachate.
SControls compared, pretreatment period and posttreatment combined.
-------
253
size effects could be detected at each treatment level and in both pre-
and posttreatment samples. With few exceptions (see control dosage,
Ca, DOC, and Mg concentrations) the large microcosms showed greater
concentrations of leachates.
The microcosms were also compared to determine if the larger size
was consistently less variable. If this were the case, one might wish
to choose the larger size because smaller sample sizes would be required
to detect significant differences. Table 4.29 shows coefficients of
variation for Ca and NO-, compared between large and small microcosms.
No unambiguous pattern can be detected; indeed, in many cases, the
larger microcosms showed greater variability both through time and
between replicates.
Size effects also appeared to play a significant role in CO,,
measurements. Table 4.30 indicates that CO,, measured from large
microcosms showed smaller coefficients of variation with great consist-
ency. This would indicate that the larger microcosms would be more
reliable for this parameter.
The most important test for the significance of the size effect
must be whether the larger size microcosms were more sensitive in
detecting dosage effects. Table 4.31 compares the large and small
microcosms in the Grassland I and II experiments. The value shown in
the table is the calculated F value for treatment effect. The signifi-
cant F values at P <^ 0.01 are indicated by an asterisk. This comparison
shows that the larger size microcosms showed a clearer treatment effect
and can therefore be judged as more sensitive for determining an effect.
The larger the F value, the larger the ratio of treatment sum of squares
-------
254
Table 4.29. Comparison of variability of Ca and N03 between small and
large microcosms. Values are coefficients of variation
(S.D. x 100/mean) for pre- and posttreatment of control,
0.5 mg As per cms and 5.0 mg As per cm?. Coefficients of
variation are then tabulated among replicates (C.V. calculated
for each leaching date and then averaged) and through time
(C.V. calculated for each core and then averaged). C.V. between
replicates is indicated by "R" and C.V. through time by "T."a
Ca
N03
Grassland I
Pretreatment
Control - smallb
largec
0.5 mg As/cm2 - small
large
5.0 mg As/cm^ - small
large
Post
Control - small
large
0.5 mg As/cm' - small
large
5.0 mg As/cm2 - small
large
Grassland II
Pretreatment
Control - smalld
largee
0.5 mg As/err/ - small
large
5.0 mg As/cm2 - small
large
Post
Control - small
large
0.5 mg As/cm^- - small
large
5.0 mg As/cm^ - small
large
Grassland I R small
large
T small
large
Grassland II R small
large
T small
large
78
149
~7T
5T
78
77
101
137
5T
88
W
25
25
15
20
90
110
~5F
45
37
43
3?
58
92
ST
5?
43
71
FT
61
194
T5T
133
210
T4T
189
149
219
178
in
220
91
129
"97
141
TTT
168
107
186
"58"
169
~5?
96
140
T4T
187
107
TUO"
104
144
aHigher coefficients are underlined.
bSmall = 10 x 10 cm.
cLarge = 30 x 15 cm.
dSmall = 15 x 25 cm.
eLarge = 30 x 25 cm.
-------
255
Table 4.30. Comparison of C02 efflux of small (10x10 cm) and large
(30x15 cm) intact core microcosms. Coefficient of
variation (S.D. x 100/mean) were calculated after
treatment for all treatments combined, controls only,
0.5 mg As/cm2; 5.0 mg As/cm2. Values among replicates
(C.V. calculated for each leaching date and then averaged)
are designated by "R" and values through time (C.V.
calculated for each core and then averaged) are designated
by "T."a
Allb
Controls6
0.5 mg/cm2
5.0 mg/cm2
Grassland I
Grassland I
Size
Small0
Large°
Small
Large
Small
Large
Small
Large
C.V.
35
TJ
26
T5
45
17
34
TT
R
T
Small
Large
Small
Large
57
T®
39
15
aHighest coefficients are underlined.
bC.V. averaged over all microcosms.
C10 x 10 cm.
d30 x 15 cm.
eNo arsenic added.
-------
256
Table 4.31. Calculated F values for treatment effect in two microcosm
experiments. The significant values at P <. 0.01 are
indicated by an asterisk. The values show that the
larger microcosms were able to detect a clearer
treatment effect.
Large Small
Grassland I
Ca
N03
ExpCa
ExpNOa
11.5*
4.5*
7.6*
7.9*
0.5
3.2
0.6
2.6
Grassland II Ca 12.8* 4.0
N03 8.7* 8.4*
ExpCa 3.5 1.7
ExpNOa 11.3* 4.2
-------
257
to error sum of squares; therefore, the larger F value indicates that,
compared to the residual error, more of the variance in the large
microcosm measurements could be attributed to a dosage effect. This
result can be interpreted to mean that the large cores are more sensi-
tive to the heavy metal treatment, or that at any given treatment level
fewer samples of the larger size would be required to detect a signifi-
cant effect.
However, the greater sensitivity of the larger microcosms does not
unambiguously answer the question of whether the larger sizes should be
used in screening programs. Table 4.24 shows very clearly that treat-
ment effects could be detected in all of our experiments in either Ca
or NOo leachate. It is clear, therefore, that smaller size microcosms
have sufficient sensitivity to detect an effect. The question of small
versus large microcosm size must be answered, based on the practicality
and expense of the screening operation. The answer would therefore
involve greater experience with the problems of large-scale screening
programs and the economic balance of using a few large microcosms versus
a larger number of small cores. This question seems best answered as
more experience is gained with microcosms as screening tools.
4.3.4 Substrate Effects
The significant effects of microcosm substrate have been shown in
several of our studies. Jackson et al. (in press) showed that forest
soils were more sensitive to arsenic disturbance, particularly evident
in greater losses of calcium compared to grassland soils. O'Neill et
al. (unpublished manuscript) showed that there were significant
-------
258
differences between an artificial sand-soil substrate and either intact
soil cores or homogenized soil cores. There were only small differences
between the intact and homogenized microcosms, and it was argued that
the intact substrate should be chosen because it retained a more natural
array of interacting biological populations. The additional work
involved in preparing the homogenized substrate did not appear to be
justified. Because much of the evidence for substrate effect has
already been presented in manuscripts generated from individual experi-
ments, little remains except to complete the analysis by presenting
evidence from the Grassland II experiment which has not yet reached the
literature.
Analysis of variance of the Grassland II experiment indicates that
there was only a single significant difference (P <_ 0.01) among sub-
strate types (i.e., intact versus homogenized). Table 4.32 presents
means and standard deviations for total export and concentration of Ca
and NOg. At the highest dose level, the intact cores lost signifi-
cantly more nitrates. None of the other measurements showed significant
differences. The scarcity of significant differences is in keeping
with the findings of O'Neill et al. (unpublished manuscript). The
argument would again seem to hold that because the intact cores require
less preparation time and, are less expensive, they should be the sub-
strate of choice for large-scale screening programs.
Table 4.33 compares the variability of the intact and homogenized
cores for Grassland II. The major argument for using homogenized soil
substrates is that homogenization tends to decrease variability. This
decreased variability for the homogenized substrate might be arguable
-------
259
Table 4.32. Homogenized grassland soil cores and intact grassland cores were
compared in Grassland II. Analysis of variance was performed over
all treatments; controls only; lowest dose only; highest dose
only. The last two computations were broken into pre- and
posttreatments to see if there was a different response after As
was added.
Pretreatment
As dose
Alla
Controlsd
0.5 mg As
per cm^
5.0 mg As
per cm^
Nutrient
ExpCa3
ExpN03
Cac
N03
ExpCa
ExpN03
Ca
N03
ExpCa
ExpNOs
Ca
N03
ExpCa
ExpNOa
Ca
N03
Intact
(X+S.D.)
5535 ±5533
98 ± 107
7.8 ±4.1
0.12 ±0.13
9109 ±12582
25 ±38
11.6 ±5.8
0.05 ±0.08
1755 ±1919
20 ±34
4.4 ±0.8
0.12 ±0.07
2862 ±4536
19 ±40
10.2 ±9.1
0.05 ±0.08
Homogenized
(X+S.D.)
7098 ±28317
395 ± 570
8.1 + 5.9
0.41 ±0.42
16281 ±24310
20 + 39
15.7 ±13.7
0.06 ±0.06
666 ±419
17 ±21
3.7 ±0.7
0.1 ±0.1
938 ±1112
6 ±5
4.4 ±1.4
0.08 +0.09
Posttreatment
Intact
(I+S.D. )
3277 ±3855
77 ±119
3.3 ±1.2
0.12 ±0.20
7096 ±4775
424 ±371
0.36 ±0.239
21.1 ±0.20
Homogenized
("X+S.D. )
3935 ±5124
230 ± 562
3.8±2.0
0.13 ±0.20
4487 ±4556
2064 ±2751
5.2 ±3.8
5.2 ±2.0
aAnalysis of variance and means across all doses, pretreatment and post-
treatment combined.
^Export values = volume leachate x concentration in ppm leachate.
Concentration in ppm leachate.
As added, pretreatment and posttreatment combined.
-------
260
Table 4.33. Comparison of variability of Ca and N03 between intact
and homogenized microcosms. Values are coefficients of
variation (S.D. x 100/mean) for pre- and posttreatment
of control 0.5 mg As per cm2, and 5.0 mg As per cm2.
Coefficients of variation are then tabulated among
replicates (C.V. calculated for each leaching date and
then averaged) and through time (C.V. calculated for each
core and then averaged). C.V. between replicates is
indicated by "R" and C.V. through time by "T."a
Ca
Grassland II
Pretreatment
Control Intact 25 129
Homogenized IT 92
0.5 mg As/cm2 Intact 20 141
Homogenized IS 107
5.0 mg As/cm2 Intact 90 168
Homogenized 3? 110
Posttreatment
Control Intact 56 186
Homogenized 97 126
0.5 mg As/cm2 Intact 17 169
Homogenized 5Ł 152
5.0 mg As/cm2 Intact 32 69
Homogenized 7_3 Ł3
Grassland II R Intact 71 100
Homogenized ITS' 100
T Intact 61 144
Homogenized 75. 130
aHigher coefficients are underlined.
-------
261
from the coefficients of variation for pretreatment codes in Table 4.33.
The intact cores consistently had higher coefficients for both Ca and
NOo. However, examination of the posttreatment data indicates that
the pattern is not consistent. For NOg, the intact core was more
variable for two out of three dose levels. For Ca, the homogenized
cores were more variable in every case. Lumping across all the data
(bottom of Table 4.33) fails to show either the intact or the homoge-
nized substrate as consistently less variable.
Because the major argument for using the homogenized substrate is
the potential for decreased variability, and because the data from our
experiments indicate that there is not a consistent trend for lower
variability in the homogenized cores, there seems little argument for
the homogenized substrate. In the small cores suggested for large-scale
screening of toxic substances, the homogenized cores are actually more
expensive to produce. Therefore, small intact cores should be the
recommended basis for the screening programs.
4.4 SIMULATION MODELING AND APPLICATION TO MICROCOSM
AND FIELD DATA (R. J. Luxmoore)
Understanding of the chemical, physical, and physiological pro-
cesses involved in chemical transport in terrestrial ecosystems can be
aided with simulation models based on these processes. The complexity
of process interactions within plants and the litter system, however,
may require that the phenomenological behavior of chemical transport be
modeled with simplified empirically derived functions. Nevertheless,
the continuing aim of process modeling is to represent the principles
of chemistry, physics, and physiology. One attempt at modeling the
-------
262
carbon, water, and chemical fluxes and pools in the soil-plant-litter
system was made in the development of the Unified Transport Model (Baes
et al. 1976). Five models were developed and coupled together and
executed with hourly resolution of photosynthesis, translocation,
respiration, transpiration, solute and water uptake, litter decomposi-
tion, and chemical mineralization (Table 4.34). The sensitivity of the
model output to changes in input parameters is critical information in
model applications because it defines the precision with which input
data should be obtained for a given precision of output results.
Sensitivity analysis also serves to identify important parameters and
important functions (from the model's viewpoint) that may aid in evalua-
tion of real-world phenomena.
Evaluation of the environmental hazard from chemical releases into
terrestrial ecosystems often needs to be made by regulatory agencies
without full information about the nature of the chemical-environment-
biota interactions. Modeling can facilitate the evaluation process at
the quantitative, semi quantitative, or qualitative levels of resolution.
The first part of this study was conducted as an aid in evaluation of
chemical transport in the soil-plant-litter system by providing sensi-
tivity information about model parameters and seeking the identification
of critical steps in chemical transport in which water, carbon, and
chemicals are represented as coupled components.
Several experimental techniques are available for assessing the
short-term effects of chemical releases into the environment. These
testing protocols include the use of basic chemical characterization
(volatility, solubility, pH), animal and plant toxicity (LD), and
-------
263
S-
d)
1 »
• 1—
^—
1
4->
C
rO
O.
r—
o
0)
c
(/)
o
E
rO
c:
-o
r—
ro
O
1
O
XJ
C
rO
S-
d)
ro
^
«,
c:
0
JO
S-
ro
O
CT)
.^
r^
., —
S-
§in
>^
co in
•
ro
•
**"
o>
^~
JD
rO
1—
(/)
O
•r-
c
ro
^
>i
XJ
d)
^
^—
O
to
^^
c
fO
c
o
.^
in
3
4-
• r—
Q
X!
C
rO
C
0
4_>
rO
4->
C
O S-
•r- d)
4-> 4J
4_J *r—
d) i —
cn
d) "O
> c
ro
c
•i—
in
BO)
4->
r — 3
«i ,_—
O
in t/i
in
ro 4-
E 0
C
o
S- U
re ••-
O E
S- C
d) >>
d> 4-> XI
C7)4->
O)
>
.,—
1 —
^^
s-
4->
i — in
•r— '|~~
0 E
to d)
U
d)
S-
O)
O-
l/l
O 3
E 0
1
4J S-
C 0)
i —
e~
CO i—
C QJ
Q. 0
E E
O
to
f^
^
^_
f>x
Q
to
s:
u_
1 — 1
Q
to
UJ
ry
UJ
O
g^
LU
^J^
O
^— «^
s:
UJ
x_x
fV
UJ
Q.
to
§
Q-
Ol
E
r3
Z.
c:
E
O
r
o
LO
r-H
C
E
O
10
S-
o
in
r-4
C
E
O
10
c
E
o
IO
t
o
LT)
t-H
C
•r—
O
ID
S-
O
ir>
1—4
O.
d)
in
d)
E
1 —
•
>> in
_O CO
>
d) rO
>x CO
ro r—
4_>
Q. XJ
^2 C
rO
QJ
3 4-*
.— O
O O
to s-
•— -o
o> e
^^ fo
o
E c:
0
in ••—
4-> (/I
C 3
CO 4-
d) -f—
•— XJ
O-
C «t
•-. 0
c
°fe
l/l 'I
3
4— C
M- 0
•1— •!
XJ 4J
ro
CM 3
0 CT
rO
O C
•r- O
S- •!-
•i- 4J
a. 3
QJ •!-•
S-
(/» 4->
co m
in T-
Z3 T3
^
o
•^
4->
rO C
S- O
O-4J
in ro
c c
10 -i-
S- -Q
"o o
o. o
ro
UJ -O
in
(J
•^
4_>
in
s_
O)
(J
ro
S-
ro
JZ
C_3
1 C 1
O. O CO
3 • T- C
in 4-> rO
10 0) rO s- C S-
ro > 3 4-> O O
O) (O CT "f" 4—
CO d) E •*"*
d) ^~ d) • rO *o
> 1 * o C~ ^_ d)
•r— >} d r^ O •!— in
in -Q Q) JZ -i— Q. 3
3 *r— f^ 4_> (/)
t[ d) ^3 rO c X
4- ^^ rO s» O rO 3
•i- rO S- O O S- •—
Q 4-> C34- •— r— 4-
in i —
4-> ro
O
O 4->
4- S- CO
O
O C
§4-> -r-
3
i — in xi •
i| dj ^— '^^
4-> ro CO
in 3 CO r-^
i/) i — CT>
rO o ^>»— <
E in J3^--
in
I S- d)
•r- O 3 4-J
XI 4- • i— C
C S- c: O > ro i—
>i 0> O -r- ••- O.
t/1 'l~~ 1 * t % I^J
O d) 4-> ro 3 C 4-
4-> 4_) rO O Q. Q; o
O • rO 3 O C 4->
jr in s_ err- -i- o j=
Q.M- 4-> d) in O.4J
in in c in 3
4->dl -Q4-> ~~i 4— CTi
,
•r— •
+J O +->
c in in
d) d)
•r- S- S-
O O
tf— t^ ^_>
4- C
4— ' -'r—
d) XJ
O i^ 4-
(_) s O
3 CD
O i — O XI
1—3 ro ro c
4- O O Q.4- >, C Q. C C
• 'OOO EOCrtJ
^O S S— *r~ QJ *i~" QJ I **
Q ro 4_> | ^ QJ (/)
.C i — Q*O V)
QJ O >•> O"
S-
O
o.
V)
c
fO
i-
4.)
cz
QJ
r_
>-)
x
•
in
4J
c
dJ
o
Q.
°
S-
d)
4-J
i-~
rO
O
$-
•r—
O.
E
t
t—
03
.,_
• ^
^
QJ
-M
Q.
QJ
1 -t->
S_ 13
QJ r- C •
+J O O tO
QJ (/» «r- QJ
"O +-> 3
^— fO »~~
C ^> rt3 ^ n3
O _Q *r— 4J >
•r- 4-> C
) ^ ^j f d) 4-9
(j d) QJ O 3
C C 4-> C Q-
3-r- O O C
4- E CL O -r-
•
C I/)
o a.
4-> JZ
•r- 10
in c
o o
Q.-I-
§4->
rO
O i—
co QJ
o
^«
3
ro ro
-(-> S-
rO -o
jr in
r — d)
ro ,— ••-
U ••- 4->
•^ O S—
S- in co
•r- O.
O. S- O
E O S-
LU 4- Q.
IO
CM
1
O
r—
LU
*^
1 1
to
Z.
^^
1
•^r
l>^
o
10
CM
1
O
1—
UJ
^^
U-
to
py
-x^
— J
2^
fy*
o
If)
CM
1
o
1—
LU
^^
U-
to
-^>^
_1
z.
g
^^^
LL.
t/1 vo
2: i-l
1 (_}
2: 1—
Ci <
O UJ
[^
CM
1
O
3
^s^_
U-
to
•^^
_J
z.
Q
J_
CO
C JD
0 E
•r- 3
4-> C
^D f^^
C CT!
ro ,-1
•*^f
LC*
0 C
•r- O
> VI
0 _*
O) O
QJ fO
CO «-D
^ — v
f*^.
f-^
CTt
r— 1
*— '
r^
rO
jj
QJ
*+—
4-
3
T"
1/1
s»
0
^_J
3
-------
264
soil and sediment adsorption. Chemical movement, accumulation, and
effects in terrestrial and aquatic microcosms can also give short-term
evaluations (months) of possible hazards from chemical releases into
the environment. It is not clear, however, that short-term evaluations
can reveal longer-term hazards to an ecosystem with its complicated
component interactions, turnover rates, accumulations, and seasonality.
Thus it is possible that a chemical may be viewed as "safe" according
to testing protocols but prove to be hazardous at some longer time after
release. It is not always feasible to conduct long-term (several years)
ecosystem testing of new chemicals. Simulation models based on physi-
cal, chemical, and physiological principles offer one means of looking
into the future and may provide valuable assistance in testing protocols
as the models gain in sophistication and their input data gain in reli-
ability.
The second part of this study evaluates the efficacy of using
microcosm data to calibrate models for the prediction of long-term
chemical transport in terrestrial ecosystems.
4.4.1 Parameter Sensitivity Studies
4.4.1.1 Materials and Methods
The parameters selected for investigation (Table 4.35) represent a
range of soil and plant properties that may influence chemical trans-
port. The primary model where the parameter is used is given along
with a description and the range of values used in the analysis. The
soil, vegetation, litter, and climatic data set used in the Crooked
Creek Watershed simulation (Munro et al. 1976) were chosen as the
-------
265
Table 4.35. Standard parameter values and tested values for sensitivity analyses
Soil and chemical parameters (SCEHM)
Distribution coefficient (ml/g)
for upper and lower soil layers
Solubility (ng/ml )
Diffusion coefficient (cm2/sec)
Litter parameter (CERES)
Decomposition rate constant
(9 9"1 hr-1)
Plant parameters (CERES)
Phloem resistances (hr)
Leaf to stem
Stem to root
Stem to fruit
Respiration rates (g g"1 hr-1)
Root
Stem
Leaf
Fruit
Leaf area to weight ratio (cm^/g)
Water potential effects (bars)
Root radius (cm)
Maximum leaf storage (g/m2)
Ambient COg concentration (ml /ml)
Plant uptake parameters (DRYADS)
Leaf cuticle permeability
(cm/ sec)
Cuticle thickness (cm)
Root solute conductivity (cm/sec)
External S02 concentration (ml /ml)
Variable
name
Kd
SP
DL
DMAX
LSPHLO
SRPHLO
SFPHLO
RRESTD
SRESTD
LRESTD
FRESTD
ART
El
R
L MAX
C02X
PERM
FILM
CONDUC
GASEX
Standard
values
500, 600 (lead)
10, 10 (zinc)
3 (lead)
2 (zinc)
10-5
3.5 x ID'4 (fruit)
5.0 x 10-5 (leaf)
1.8 x 10-4 (stem)
1.3 x 10-4 (root)
20
30
500
1.8 x 10-4
1.9 x 10-4
1.8 x 10-4
1 x 10-4
100
4
0.05
420
3.2 x 10-4
lo-n
10-4
10-8
2 x 10-8
Tested
values3
2500,3000
100, 125 (lead)
1000,1000
100, 100 (zinc)
1, 1
300, 30 (lead)
200, 20 (zinc)
10-7, 10-3, io-l
x 10, x 10-1
x 10, x 10-1
x 10, x 10-1
x 10, x 10-1
100, 1
100, 1
1000, 100
x 10, x 10-1
x 10, x 10-1
x 10, x 10-1
x 10, x 10-1
200, 50
-2, +2
5, 0.5, 0.005
210, 42
x2, x3, x4
10-7 io-9
10-lfl, 10-13
10-2, io-5> 10-6
10-6, 10-105 io-12
2 x 10-7, 2 x 10-9
axn or +n implies that the standard value was multiplied by n or added to n,
respectively.
-------
266
standard set of input values. Each parameter was varied around this
standard, and the results were related back to this standard set of
results. The output parameter list used in the analysis (Table 4.36)
is extensive and provides many points at which the system sensitivity
may be evaluated.
The method for testing the parameters consisted of running the
model over a three-month period, May-July, with one particular value
varied in the standard set of input. No data were recorded for the
initial two-month period of the simulations to allow the model to
stabilize. Daily values were output on file for the entire month of
July, and hourly values were recorded on file for a period of eight
days (16th-24th) during July for sensitivity and statistical analysis.
Two indices of variation were computed for each response over time
for all sensitivity runs. The first, termed a relative output coeffic-
ient (r), measures the output response with respect to the standard
response and has similarities with the X statistical test.
, (k).
1-1
where e.. = response of the j-th output parameter for i-th time step
in the standard run,
a.. = response of the j-th output parameter for the i-th time
' J
step in sensitivity run k,
T = total number of time steps, and
r. = relative output coefficient for the j-th output parameter
J
in the k-th sensitivity run.
-------
267
to
CO
gŁ
03 03X51
E m Q x —
XXX Q „ >•
Q Q << X OS .c
x z j< —3 —a
< < _1 OS X 03 X
j j a
H H CC 03 DC 1 |
M M
* iJ l» E B B W o Q
* «UWBulOB:C|c.<«
MDfc«OEBWoOO*JWW»MW
H 03 O U
OOEnOOHOfe
fc. SH pu CU
o O o e> m
OfWMUtUMfUDZnDX
(jZSEBZZBXOH >* Q •< >-* O
PSK <<\Q- W 038 fNfN**(N#
M •< >* K< »*E**E
IM _q < *<»q „ * # \ » « X.
• ^
MWO *» V » V O *C«<— *-**03— 'O5^«>-
O« E\ o^o** ^^jfNjrg w « w •< <
tnz; S, ts — o P P *^.*«(HE-»Wo3E-»op
OB t?«-os ^ac •=•» — ,S rM#JlJa: »jHK3K
lini XQMttW^iaC3-,t-M«EH«t-) # *J 15 — WO*- W H O\\
6i-fl QZH- o O w oa
~aa«s#[naR«3 »M uo uo\w— \_iM>j(-3M*i*:«:e*:«WM
uoo<>4»*H«MU05's-' U>— < OfefezfefeZH"
9OD(N«\C9H pM«Ct>H DH D CdH hMCf^n O O -J J
» E — ee« «?3wai«:1J«''— »SSEJ*.-3V)Z,J«HM H M Me^Htfl^HNMH
WWHO^OMOV) W U O W H M _3 t-< H Z Zas •<4«xOCBOMHe-1JOMt»JDM*WDM'^DM^PH**
wiuWaj^m^cnw^KJOLi^tnyiWKaiWNtJtnNiJuiN^Jtnt^hJtrie^NjtntOt-s
03 \m 3= Q
m OS x*
VZ XO <
0 *0 Z j
3 "^ 53 OS03a9HH<
^M <JHN
O
O O Cw
-------
268
The second index (S) relates the change in output to the change in input
and is called a sensitivity coefficient,
S W - v "iJ " eiJ
j 1=1 \-\ '
where v. = value of input parameter that was varied for the
k-th sensitivity run,
v. = value of standard parameter for v, used in the
standard run, and
(k)
S. ' = sensitivity coefficient for k-th sensitivity run and
J
the j-th parameter.
Values of these coefficients for each variation of input parameters are
documented by Begovich and Luxmoore (1979).
The overall effect of change in each of the parameters is presented
in graphs of the heavy metal contaminant proportions in each of the
soil, litter, and plant components being modeled. The data can also be
plotted as the variation of one response over time for different input
values. The amount of data is too extensive for all of the output to
be examined in this manner and some selected plots provide insight into
the responses.
4.4.1.2 Results
Soil and Chemical Parameters
Distribution coefficient (Kd). The distribution coefficient is
the ratio of the amount of contaminant exchanged onto the soil surface
per unit soil mass to the amount of contaminant in the soil solution
-------
269
per unit solution volume. Increasing this ratio increased the amount
on the soil exchange, reduced the soil solution concentration, and thus
reduced the total plant uptake of lead (Fig. 4.9). A similar result
was obtained for zinc. A large variation of the fraction of contaminant
in the plant parts was also found. The relative contaminant fraction
in the leaves increased along with a decrease in both the root and stem
contaminant fraction as the K. increased. The mineralized contaminant
decreased with increase in zinc K,, whereas there was little change
with increase in lead K^.
Hourly plots of phloem (Fig. 4.10a) and xylem (Fig. 4.10b) trans-
location of zinc show greater transport with lower K, and that phloem
translocation can remain at high rates at night. Both transports vary
indirectly with K ,. The most sensitive outputs include the chemical
content in litter and roots and in the plant xylem transport pathway.
Solubility (SP). The amount of heavy metal contaminant which dis-
solves in the infiltrating soil water depends on the chemical solubility
(SP) and this usually changes for different chemical compounds. The
fraction of zinc and lead in the soil increased with increasing solu-
bility, causing the fraction of chemical in the litter to decrease
(Fig. 4.11). The fraction in each of the separate soil layers and in
the biomass was basically unchanged. The overall fraction of lead in
the plant increased slightly because of increased root uptake. The
fraction of zinc in the plant did not vary due to the plant having
attained near maximum concentrations in each of its tissues. The
increase in the litter-mineralized fraction resulted from both the
decrease in total litter content and a mineralization rate proportional
-------
270
ORNL-DWG-78-12294R
LEflD KD
TRflCTION OF LEflD IN MODEL COMPflRTMENTS
OVERflLL
FRACTION
l.O-i
o.s
l.o
o.s-
l.O-i
gLRNTS
5.00x10° l.OOxlO2 5-OOxlO2
LITTER
5.00x10° l.OOxlO2
SOIL
5.00x10^
5.00x10° l.OOxlO2 S.OOxlO2
FRUITS
LEHVES
STEMS
ROOTS
2.50xl(T
STORflfiE
niNERflUZED
2.50x10°
0-3 CM
3-15 Cfl
15-90 CM
2.50xlOJ
Figure 4.9. Relative lead distribution in plant, litter, and soil
components with change in lead distribution coefficient
(Kd).
-------
271
ORNL-OWG. 78-12360
16 17
ZINC KD
ours IN
a 34
Mrs IN JULI
Figure 4.10. Influence of zinc K,j on zinc transport in the vegetation:
(a) leaf-to-stem phloem, and (b) stem-to-leaf xylem.
-------
272
ORNL-DWG-78-12297R
SP
FRflCTION OF ZINC IN MODEL COMPflRTMENTS
OVERflLL
FRRCTION
PLRNTS_ „ FRUITS
0.5-
0.0-
LERVES
STEMS
ROOTS
2.00x10° 2-OOxlO1 2.00xl02
1.0-
0.5-
0.0-
LITTER STORflK.
/MlhERflLIZED
2.00x10° 2.00X101 2.00xl02
1.0-
0.5-
SOIL
0-3 Ctl
3-15 CH
2.00x10° 2.00X101 2.00xl02
Figure 4.11. Relative zinc distribution in plant, litter, and soil
components with change in zinc solubility (SP).
-------
273
to litter decomposition. As for the distribution coefficient, the most
sensitive outputs to changes in SP include the chemical content in
litter compartments and in the plant tissue xylem.
Diffusion coefficient (PL). The diffusion coefficient of the heavy
metals is used in determining root uptake. Diffusive uptake increases
for increasing values of a a/(DL)b, where a is the root absorbing power,
a is the root radius, DL is the diffusion coefficient, and b is the
buffer power (Baldwin et al. 1973). Thus as DL increases, diffusive
uptake decreases. The overall fraction of contaminant in the plant
decreased for increasing diffusion coefficient for lead (Fig. 4.12),
but was unchanged for zinc because the vegetation attained a maximum
internal zinc concentration. This explained the differing rates of
chemical uptake for zinc and lead. Hourly root uptake during the day
showed greater rates for lead with decrease in DL whereas zinc uptake
was a reverse response. However, maximum root uptake for zinc occurred
at the lowest value of DL. Zinc content in litter and plant xylem
tissues showed the greatest sensitivity to change in DL. The most
sensitive response for lead was in the lead concentration of roots.
Litter Parameter
Decomposition rate constant (DMAX). The rate of material decom-
posed in each of the litter compartments is a function of temperature
and an input value for the decomposition rate constant. Chemical
mineralization from litter is proportional to the decomposition rate.
Thus, increase in the rate constant increased the amount of lead and
zinc mineralization (Fig. 4.13). There was no change in overall
-------
274
ORNL-DWG-78-12298
DL
PRflCTION OF LEflD IN MODEL COMPflRTMENTS
OVERflLL
FACTION
o.s-
o.o -i
l.O-i
o.s-
o.o J
0.3-
0.0 J
10.00x10"
I I I I
gLflNTS
lO.OOxlO"6 lO.OOxlO'4
LITTER
SOIL
FRUITS
LEflVES
STEMS
ROOTS
10.00x10
-2
STORflCE
niNERRLIZED
lO.OOxlO"8 lO.OOxlO"6 lO.OOxlO"4 lO.OOxlO"2
lO.OOxlO"8 lO.OOxlO"6 lO.OOxlO"4 10.00x10
0-3 CM
3-15 CM
e 15-90 CM
-2
Figure 4.12. Relative lead distribution in plant, litter, and soil
components with change in diffusion coefficient of lead
(DL).
-------
275
F
OVERfl
FRflCT
1.0-
D.S-,
0.0 -J
L.D-
O.S-
0.0-
1.0-
0.5-
0.0-
ORNL-DWG- 78-12301
DMflX
RRCTION OF ZINC IN MODEL COMPflRTMENTS
LL
ION
PLflNTS_ „ rRVJITS
LEfWES
STEMS
1 1 """
STDxlO"1 STDxl STDxlO
LITTER
STORflGE
^ — ^^"^ nrNERRLIZED
Q "&
STDxlO"1 STDxl STDxlO
SOIL
0-3 cn
3-15 CM
STDxlO"1 STDxl STDxlO
Figure 4.13. Relative zinc distribution in plant, litter, and soil
components with change in litter decomposition rate
constant (UMAX).
-------
276
chemical content in litter. The litter mineralization of chemical was
the most sensitive output response.
Internal Plant Parameters
Phloem resistances (LSPHLO, SRPHLO. SFPHLO). The rate of transport
of sugar substrate between plant compartments is proportional to the
substrate gradient and a phloem resistance factor (Dixon et al. 1976).
Leaf-to-stem (LSPHLO), stem-to-root (SRPHLO), and stem-to-fruit (SFPHLO)
resistances are all used in the CERES model. Increase in the leaf-to-
stem phloem resistance reduced total growth and changed the relative
biomass distribution of the plant. At higher resistances, there was
greater leaf growth and lower stem growth. The differences in biomass
resulted in corresponding changes in contaminant uptake (Fig. 4.14).
Lower resistance caused greater phloem chemical movement which was
indicated by a higher sulfur content in roots supplied from leaf uptake
and also in the hourly zinc translocation rate.
Increase in the stem-to-root phloem resistance slightly reduced
total growth and changed the relative biomass distribution. Total
contaminant uptake was highest at the lowest resistance, and this was
achieved through greater relative leaf uptake. Rather high leaf uptake
(due to high standard value of PERM) and reduced transport to other
plant components at high phloem resistance, caused leaf concentration
to be high. Mortality of these leaves increased the litter contaminant
level and the rate of zinc mineralization. Thus a stem-root property
influenced chemical dynamics in litter. The phloem translocation of
chemicals was directly influenced by change in phloem resistance and
this was shown in the output responses. Sulfur dioxide gas uptake was
-------
277
LSPHLO
or LLPD IN ncoEL CGHPPRTMENTS
2 OOilO1
LITTER
2 OOilO1
SOIL
__ a
ORNL-DyG-78-123S7
LSPHLO
PRRCTION GP ZINC IN nODEL COMPF1RTMENTS
FLflNTSn
1 00.10U 2 00.101
LITTER
2 00110
SOIL
STDtS
ROOTS
Figure 4.14. Relative chemical distribution in plant, litter, and soil
components with change in leaf-to-stem phloem resistance
(LSPHLO) for (a) lead, and (b) zinc.
-------
278
influenced by LSPHLO. The zinc content in the root litter was very
sensitive to both LSPHLO and SRPHLO.
Standard respiration rates (RRESTD. SRESTD, LRESTD, FRESTD).
Respiration rates determine the loss of sugar substrates from the plant
parts. The total respiration of each plant compartment varied directly
as a Q..Q function of temperature and the compartment biomass. Input
values for tissue respiration rates at given temperatures were used for
the Q,Q functions. Total biomass decreased with increase in the
standard root respiration rate and the relative proportion of roots
decreased.
The root contaminant uptake increased as the root biomass increased
for decreased standard root respiration rate (Fig. 4.15). At the two
lower standard respiration rate values, zinc uptake became limited by
the plant capacity for more zinc, and the plant compartments approached
maximum zinc content. Lead uptake did not approach maximum capacity
and the roots showed a preferential retention of lead for the highest
standard respiration rate case. The large proportion of lead in leaves
was due to leaf uptake.
Zinc uptake definitely decreased with greater stem respiration
rates whereas there was little change for lead. Because of the decrease
in stem biomass, a greater fraction of the zinc stayed in the roots and
leaves. The mortality of leaves with high contaminant levels increased
the litter contaminant mineralization.
Both leaf and fruit respiration rates had little effect on the
overall fraction of contaminant in the plant. There was a small
decrease in amount of zinc in the plant for higher leaf respiration
-------
279
TRRCTIO
OVOWLL
nwcriOK
"1
"1
«DJ
...
1
ORM-DHG- 78-12354
RRESTD
M OF LEflD IN MODEL COMPflRTMENTS
PL«TSD D rmm
LDWCS
^^^_jj 51D1S
1 8>KT5 1 SxlCT* 1.8xlO"3
LITTER
STORflGt
1 3»10"5 1 8xlO"4 1.8xlO"3
son.
0-3 cn
3-LS Cn
1 8xlO"5 1 8«10'4 1 8xllT3
ORNL-DWG-78-12355
RRESTD
^RflCTION OF ZINC IN MODEL COMPflRTMENTS
OVOWLL
FWCT10N
PLflNTSQ n r"UIIS
1 D ° D
LOWES
B * " STWS
1 i ^—-o ^^^
1 B»10'5 1 SxlO"4 1 SxlO"3
LITTER
05"| STORAGE
1 1
oo-l 1 1 D 0 o niKBflLITED
1.8«10'5 1 8«10"4 1 8xlO"3
,BI SO'1-
0-3 Oi
' 3-LS CM
1 8xl(T5 1 SxlO"4 1 8xlO~3
Figure 4.15. Relative chemical distribution in plant, litter, and soil
components with change in standard root respiration rate
(RRESTD) for (a) lead, and (b) zinc.
-------
280
rate. The difference in respiration rates influenced a large number of
plant and chemical output responses. The most significant responses to
the RRESTD, SRESTD, LRESTD, and FRESTD factors are root leaf concentra-
tion, zinc content in leaf xylem, S0? uptake, and fruit zinc concen-
trations respectively. Changes in stem-to-root translocation of zinc
was induced by changes in biomass for the different root respiration
rates. Stem-to-root sugar translocation was highest for the largest
standard stem respiration rate (Fig. 4.16).
Leaf area-to-weight ratio (ARI). The leaf area index is propor-
tional to the area-to-weight ratio and is given by the product of leaf
biomass and ARI. Therefore, the rates of net photosynthesis and trans-
piration are directly affected by varying ARI. Total plant growth
increased with increase in ARI and there were changes in the proportion
of plant parts. This, in turn, affected the concentration of the con-
taminants in the plant tissues. Relative lead and zinc concentrations
in the plant were affected differently (Fig. 4.17). A ratio value of
50 is representative of thickened sun leaves, 100 is representative of
mesic sun leaves, and 200 is representative of mesic shade leaves.
Sulfur dioxide uptake, leaf sulfur concentration, zinc uptake by leaves,
and leaf lead concentration were some of the most responsive outputs
from the models.
Water potential effects (El). A growth coefficient is used to
simulate water stress effects on tissue (leaf, stem, root) growth. The
coefficient is the value of an exponential function with a range from
0 to 1 which represents no growth or unaffected growth respectively.
Input water potential values determining the sensitivity function
-------
281
ORNL-DWG-78-12310R
SRESTD
LEGEN
o- 1.9 E -4
- 1.9 E -3
- 1.9 E -5
to
19 20 21
DflYS IN JULY
Figure 4.16.
Influence of standard stem respiration rate (SRESTD) on
sugar translocation in stem-to-root phloem.
-------
282
r
OTOW
nwcr
...
1.0-
„.
,
ORNL-DHG-78- 12352
flRI
RRCTION OF LEflO IN MODEL COMPflRTMENTS
LL
lot
„ PLflNTS. „ rwiTS
LEftVCS
"~ ' ROOTS
5 OOKlD1 1 Q*102 2 OOxlO2
L I TIER
5TQWE
* " * mmwLiar
5 OOnW1 t OxlO2 J.OOxlO2
SOIL
0-3 CH
s-is en
5 OOxIO1 1.0»10Z e.00«102
r
twiwi
nwcT
...
...
ORNL'WS-78-12353
flRI
ACTION Of ZINC IN MODEL COMPflRTMENTS
L
PLflNTS.
LŁ«VES
IP — —
— w
srtre
* ROOTS
S.OOxlO1 1 OxlO2 3.00xlO?
LITTER
snmet
B . n rtiftWLiani
5 OOxlO1 1 0*10Z 2 OOnlQ2
SOIL
D-? cn
3-lS CH
5 OOxlO1 1 OxlO2 Z.OOxlO2
Figure 4.17. Relative chemical distribution in plant, litter, and soil
components with change in leaf area weight ratio (ARI) for
(a) lead, and (b) zinc.
-------
283
include an initial potential at which tissue growth is reduced, the
potential where tissue growth is reduced by one-half, and the potential
where no further growth occurs (Dixon et al. 1976). Incremental values
are added to each of these potentials in this test of water stress
effects on overall plant growth, such that the greater the input poten-
tial the less growth is affected by the water potential stress. Thus
total productivity was higher with reduced sensitivity to water effects.
Stem growth was proportionally increased for decreased sensitivity,
whereas the fraction of roots was not changed. Water potential influ-
enced leaf growth more readily than other tissues and S0~ uptake was
the most sensitive output response. Stem biomass and chemical content
in xylem tissues were also greatly influenced.
Root radius (R). The average root radius is used to determine
root density and is a factor in the mass flow and diffusive uptake of
soil contaminants. The direct dependence of root radius on uptake is
difficult to determine because many of the factors in the uptake
equations are dependent on R. For zinc, uptake during the sunlight
period is increased with greater root radius. The root uptake for lead
during the sunlight period, however, varies in exactly the opposite way
(Fig. 4.18); there is more root uptake for the decreased root radius.
These responses depend on the interaction of plant demand and the supply
from the soil. The proportional contaminant distribution in the plant,
litter, and soil components shows relatively'little change with change
in root radius. The most sensitive output responses to change in root
radius were the chemical content in litter and the mineralization rates
of chemical in the litter.
-------
284
ORNL-DWG. 78-12314
R RRDIUS
—. - —•
°
a a>
°*
QL
QJ
Q_ «^
CS
1-1 <0
r"
'".n
I d'
a
ac ^
UJ ..
_)°
O r>
cu =
^
CE
e- CM
Q^o
o -
Sd~
LEGEND
o- .05
A - . DOS
x-5.
-i—
25
20
DRY'S IN JULY
30
Figure 4.18. Influence of root radius (R) on root uptake of lead
during the sunlight period.
-------
285
Maximum leaf storage (L MAX). Each plant compartment has an upper
limit to its storage value and the tissue growth rate is proportional
to the difference between the current storage and maximum storage. The
leaf maximum storage was reduced to test the sensitivity on overall
plant response.
A reduction in total biomass as leaf maximum decreases is associ-
ated with a greater proportion of the biomass being located in the roots
and stems. Lead uptake into the stem and root compartment changes
reciprocally with change in L MAX (Fig. 4.19). The main lead uptake is
primarily through the leaves, whereas root uptake accounts for the
smaller lead content in the roots. At the lower biomass levels with
lower L MAX values, the leaf lead content approaches a maximum. S0?
uptake was the most significant model response to change in maximum
leaf storage and surprisingly the zinc content of root litter was also
very responsive.
Plant Boundary and Atmosphere Parameters
Leaf cuticle permeability (PERM). The rate of solute uptake
across the leaf surface is dependent upon the solute concentration
gradient, leaf cuticle thickness, and the leaf cuticle permeability.
The permeability factor is one of the most sensitive parameters tested,
especially in relation to contaminant movement.
The change in leaf uptake causes a change in the amount deposited
on the litter surface. The increase in the overall fraction of con-
taminant in plants is accompanied by an overall decrease of the amount
in the litter (Fig. 4.20). Zinc has a higher solubility than lead and
this is shown in the reduced litter zinc content and greater soil
content.
-------
286
OKNL-OWG-78-12348R
L MflX
FRACTION Of LEflD IN MODEL COMPARTMENTS
DVERH.L
fRflCTION
PLflNTS rmjITS
LEflVES
—SI STEHS
I _-— — -~* fttXJTb
0J B-— -~~"~
42 210 420
BITTER
' I
STDRflSt
J | | ° ° " niicm.izo>
42 210 420
i> SOIL
D-3 C?1
•M
3-15 W
42 210 420
ORNL-DUG-78-1Z349R
L MflX
PRflCTIQN OF ZINC IN MODEL COMPRRTMENTS
OVCRflLL
rRHCTION
PLflNTS ™1IS
B— — — — ~^___^, LEWES
a s -1 B — -— ______^ STEMS
,J i 1
42 2»0 420
LITTER
o i j a STORflCC
I ^^
. 0J 1 1 1 ^--^0 0 fl!«JWL]ZCD
42 3(0 420
SOIL
0-3 cn
" !
i 3-15 O1
42 210 420
Figure 4.19. Relative chemical distribution in plant, litter, and soil
components with change in maximum leaf storage (L MAX) for
(a) lead, and (b) zinc.
-------
287
PERM
PRflCTION OP LERD IN MODEL COMPflRTMENTS
10 OOxlO"14 10 OOxlO"12 lO.OOxlO"11 10 OOxlO"10 IQ.OOxlO"8
LITTER
10 OOxlO"1* 10 OOxlO'1? 10.00.10'11 lO.OOxlQ'10 10.OOxlO'8
SOIL
10 00.10"1* 10 00x10"U 10 00x10"U 10.00x10" 10 00x10'
ORKL-DWG-78-U347
PERM
PRflCTION OP ZINC IN MODEL COMPflRTMENTS
OVDtf
fRRC
LL
„ PLflNT| m"7S
LEWtS
SIOB
1 " m
lO.OOxlO"1* 10 OOxlO"12 10 OOxlO"11 10-OOxlO"10 lO.OOxlO"8
,.
at
LITTER
sTomce
1 1 // rmewLUCD
..1 II 0 ^
10 On.10'14 10 00«10"1Z 10. OOxlO"11 10 OOxlO"10 10 OOxlO"8
I 0
...
SOIL
0-3 CM
3-1 s cn
lO.OOxlO"1* 10 OOxlO"12 10 OOxlO"11 lO.OOxlO"10 10 OOxlO"9
Figure 4.20. Relative chemical distribution in plant, litter, and soil
components with change in leaf permeability to chemicals
(PERM) for (a) lead, and (b) zinc.
-------
288
Leaf uptake of zinc with increasing permeability contrasts with a
corresponding decrease in root uptake of zinc. Both the lead and zinc
leached from leaves were greatly influenced by change in PERM. Chemical
transport within the vegetation and chemical accumulation in litter
were responsive to the changes in leaf permeability.
Cuticle thickness (FILM). The solute uptake from the leaf surface
is inversely proportional to leaf cuticle thickness. Therefore, the
relationships depicting the fraction of contaminant in each of the com-
partments (Fig. 4.21) is almost completely opposite of Fig. 4.20. At
high cuticle thicknesses, leaf uptake is diminished and the majority of
the lead remains in the root system. Significant increases in zinc
litter mineralization occurs at the low cuticle thickness values. The
lead content of stem xylem, lead leached from leaves, and lead xylem
transport from stem to leaf were highly influenced by change in cuticle
thickness. The responses of zinc transport in the plant were suppressed
because the plant had almost saturated its zinc uptake capacity.
Root solute conductivity (CONDUC). The actual plant solute uptake
depends on the plant tissue capacity for additional solutes and the
root solute conductivity. With increase in root solute conductivity,
the overall fraction of contaminant in the plant increased (Fig. 4.22).
The amount of zinc uptake through leaves increased with reduced root
conductivity, causing greater accumulation in the leaf compartment.
Alternatively when zinc uptake by roots was large, leaf uptake was
reduced, allowing more contaminant to reach the litter surface and
causing an increase in the amount mineralized. The model is very sensi-
tive to the conductivity parameter. Zinc and lead content in the litter
-------
289
ORNL-DWG-78-12342
run
FRRCTION OF LEHD IN MODEL COMPflRTMENTS
OVOWLL
nWCTION
10 00x10 10 00«10 10 00x10'^ 10 ODxlO*J
LITTER
10 OOxlO"7 10 OOxlO"6 10 00*10"^ 10 OOxlO"3
SOIL
10 OOxlO'7 10 OO.lO"6 10.OOxlO'5 10.00»10'3
QHNt -DWG- 78- 13343
FILM
PRflCTION OP ZINC IN MODEL COMPRRTMENTS
ovow
nwcr
oo.
LL
ION
PLflNTS nwiTS
LEflVES
STETtS
poors
10 OOilO"7 10 OOilO"6 10 OOxlO"5 10 OOxlO"3
"1
...
LITTER
1 ^ B niNDW.ian
10 OOxlO*7 10 OOxlO"6 10 OOxlO"5 10 OOxlO"3
1 0
...
SOIL
D-3 CT1
3-15 Cfl
0 OOxlO"7 10 OOxlO"6 10 OOxlO"5 10 OOxlO"3
Figure 4.21. Relative chemical distribution in plant, litter, and soil
components with change in leaf cuticle thickness (FILM) for
(a) lead, and (b) zinc.
-------
290
ORNL-OWG-78-12320
CONDUC
FRRCTION OF ZINC IN MODEL COMPflRTMENTS
OVERfllL
FRflCTION
o.s-\
o.o J
i l
o.sH
o.o-i
1.0 1
0.0-1
gLRNTS
lO.OOxlO"13 lO.OOxlO"11
SOIL
lO.OOxlO
lO.OOxlO"13 lO.OOxlO"11 lO.OOxlO"9 10.00x10
FRUITS
LEflVES
STEMS
ROOTS
lO.OOxlO"13 lO.OOxlO"11 lO.OOxlO"9 lO.OOxlO"7
LITTER
STORRCE
MINERflUZED
"9 lO.OOxlO"7
0-3 CM
3-15 W1
15-90 CM
-,-7
Figure 4.22. Relative zinc distribution in plant, litter, and soil
components with change in root chemical conductivity
(CONDUC).
-------
291
and the heavy metal mineralization rate were the most responsive outputs
to change in root conductivity. The root lead concentration was the
most responsive plant property.
External gas concentration (GASEX). The diffusive uptake of gases
(SCL in this case) is calculated by determining the gas concentration
differences between the leaf and the atmosphere. There was greater
S02 uptake and greater sulfur translocation from leaves to stems with
an increase in S02 concentration. The leaf compartment was the domi-
nant sulfur sink for the conditions examined. The SO,, uptake and
sulfur dynamics in the vegetation were very responsive to change in
external S02 concentration.
COg concentration in ambient air (C02X). In the plant growth
model, net photosynthesis depends on the C02 gradient from the
chloroplast to the ambient air (Dixon et al. 1976). Increasing the
ambient C02 concentration affects photosynthesis and causes an
increase in plant growth. The greatest proportion of the increase is
divided between the roots and stems. As a result, a slight increase in
the overall fraction of zinc in the plants was obtained (Fig. 4.23).
Chemical uptake and transport was influenced by change in atmospheric
C02 concentration. The uptakes of S02 and zinc were the most
responsive model outputs. Stem biomass was also significantly
influenced.
4.4.1.3 Discussion
A subjective summary of parameter effects on the model output
(Table 4.37) identifies five of the fifteen parameters as highly sensi-
tive factors in the model behavior. These are the chemical distribution
-------
292
ORNL-DWG-78-12317
C02X
FRflCTION OF ZINC IN MODEL COMPflRTMENTS
OVERflLL
FIWCTION
3.20x10
"4
3.20x10
-4
ELflNTS
3.20xlO"4 6.40xlO'4 9.60xlO"4 1.28xl(T3
LITTER
6.40xK
4
SOIL
6.40xlO'4 9.60xlO"4
rmjiis
LCHVCS
STEI1S
roots
sroraacc
ru NDWC IZED
1.28x10
-3
0-3 cn
3-is on
is-ao cn
1.28x10
-3
Figure 4.23. Relative zinc distribution in plant, litter, and soil
components with change in atmospheric C02 (CC^X).
-------
293
Table 4.37. Subjective evaluation of parameter effects on total chemical level and
distribution in soil, litter, and plant componentsa
Total chemical level
Increase
in
parameter
Kd
SP
DL
DMAX
PHLO
RESTD
ARI
El
R
LMAX
PERM
FILM
CONDUC
GASEX
C02X
Qualitative
+++ large
++ fair
+ small
Soil
Pb Zn
+ ++
+++ ++
+ +
0 0
0 0
0 0
0
0
0 0
0
0 ++
0
_
0 0
0
ranking:
increase
increase
increase
Litter Plant
Pb Zn Pb Zn
0 - - -
— + 0
0 0 - -
0000
000-
0000
000 +
000 +
0000
0 - 0 ++
+++ o
++ ++ — 0
00+++
0000
000 +
Soi
Pb
+
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Chemical distribution
1
Zn
;;
0
0
0
0
0
+
+
0
±
0
0
0
0
0
Litter Plant
Pb Zn Pb Zn
o +•+-+ +++ ++
ttt tii + 0
0 it ii it
+ + 00
0 ++ + +
00++ +
00 + +
00 + +
0 + + 0
0 ++ + ++
+ +++ +++ 0
+ +++ _+++ 0
0 +++ +++ ++
0000
00 + +
0 no change
small decrease
-- fair
— large
decrease
decrease
-------
294
coefficient in soil (Kj)» the chemical solubility (SP), the leaf
permeability to chemical (PERM), leaf cuticle thickness (FILM), and the
root conductivity to chemical (CONDUC). The K., SP, and FILM terms
can be fairly readily measured; however, this is not the case for the
plant characteristics that determine the rate of chemical upake at its
leaf and root boundaries (PERM, CONDUC). Estimates of these character-
istics can be obtained by tuning the models to results from experimental
uptake studies (Luxmoore and Begovich 1979).
Two parameters (DL, R) showed surprising results by generating
opposite responses for zinc (soluble, mobile) and lead (less soluble
and less mobile) movement within the plant.
A consistent pattern obtained in the model output was an increase
in litter mineralization in response to factors that increased the
chemical content in the plant leaf tissue. These include the phloem
resistance (LSPHLO, SRPHLO), standard tissue respiration rate (SRESTD),
maximum leaf weight (LMAX), leaf permeability (PERM), leaf cuticle
thickness (FILM), and root conductivity (CONDUC). Chemical dynamics in
the litter system also showed responses to the chemical distribution
coefficient (KJ, chemical solubility (SP), chemical diffusion coef-
ficient (DL), litter decomposition characteristic (DMAX), and root
radius (R), as shown by the response and sensitivity coefficients.
Some of the indirect relationships would not have been expected but
became apparent through modeling methods. The study also suggests that
monitoring of litter systems (chemical content, chemical mineralization)
can provide one detection method of plant response to chemical perturba-
tion.
-------
295
Sulfur dioxide uptake from a chronic atmospheric concentration
Q
(2 x 10" ml/ml) was found to be very responsive to change in several
plant properties. Reduced leaf-to-stem phloem resistance (LSPHLO),
reduced stem-to-root phloem resistance (SRPHLO), reduced leaf respira-
tion (LRESTD), increased leaf area weight ratio (ARI), reduced water
stress sensitivity (El), increased leaf storage (L MAX), and increased
atmospheric CCL (CCLX) all resulted in an increased S0? uptake by
the vegetation. These effects were all the result of an increased sink
for sulfur by increased growth associated with the change in parameter
value.
Much information is contained in the recent report by Begovich and
Luxmoore (1979) and has not been commented on; nevertheless, the data
can be used as a reference to examine further any model responses of
interest.
4.4.2 Simulation of Chemical Transport in a Deciduous
Forest Using Microcosm Calibration
This work compares the results from a six-year simulation of heavy
metal transport in a deciduous forest with field data collected during
the sixth year after commencement of lead mining and smelting at the
new lead belt in southeastern Missouri. The study does not attempt to
"validate" a model per se, but seeks to identify a role for simulation
in testing protocols and to evaluate some of the advantages and dis-
advantages of a combined microcosm and modeling approach through a test
case application.
-------
296
4.4.2.1' Tree Microcosm Study and Model Parameterization
Six intact soil blocks (45 x 45 x 25 cm depth), each containing
one red maple (Acer rubrum L.) sapling, ground flora, and litter, were
taken from a mesic forest location and reestablished in a greenhouse
(Jackson et al. 1978). Following a two-month adjustment period, heavy
metals (Pb, Cd, Zn, Cu) were added to the soil surface in an initial
dose and in weekly treatments for 12 weeks. After 20 months the heavy
metal content of the tree parts, soil, and litter were analyzed in a
final harvest.
A 24-month simulation (October 1974-September 1976) was conducted
with the coupled models for carbon, water, and heavy metal dynamics
(Table 4.34). The soil system was characterized with representative
hydraulic properties (Peters et al. 1970) and soil chemical properties
(Sweeton, unpublished data). Initial conditions for the litter and
plant dry weight (Jackson, personal communication) and heavy metal
levels in plants, litter, and soil were obtained (Jackson et al. 1978).
Meteorological conditions within the greenhouse were derived by modifi-
cation of Oak Ridge records for the study period, according to inside-
to-outside weighting factors determined experimentally. Daily maximum
air temperature was increased by 11.7°C and the minimum was increased
by 2.2°C. Daily average dew point temperature was increased by 2.2°C
and solar radiation was reduced by 50%. The watering regime of the
microcosms was represented by equivalent precipitation amounts in the
model, adjusted to eliminate canopy interception.
Initial model calibration was conducted on parameters and meteoro-
logical data that influenced total water loss. Watering records were
-------
297
used to calculate annual evapotranspiration (ET) as the difference
between the water applied and the total drainage from the microcosms.
The estimated experimental ET was 148 cm/year. The initial simulations
with parameter adjustments gave ET in the range 59 to 75 cm/year. Some
of the adjustments were extreme and outside the range of published
physiological data. Modification of the temperature and irradiation
data for the greenhouse was also examined. The lack of agreement
between the microcosm and the model in terms of water budget was prob-
ably related to poorly characterized microclimate in the greenhouse and
physiological changes in an outdoor plant adjusting to a greenhouse
environment. Advective heat may have been a major source of energy for
ET that was not possible to characterize because the microcosms were
located near an exhaust fan in the greenhouse. There was no way of
meaningfully tuning the models to the microcosm water budget. The
similation gave results (Table 4.38) representative of outside water
budgets for the Oak Ridge area (Luxmoore et al. 1978b).
Adjustment of the carbon budget of the plant and litter system was
undertaken next, and reasonable agreement was obtained with the final
harvest data (Table 4.38). The main parameters adjusted were the maxi-
mum stem and root storage which determined the overall biomass of the
tree. The litter decomposition rate was also adjusted to give a total
mass that was comparable with the measured value (Table 4.38).
The final phase of adjustment was to obtain agreement between the
observed and simulated chemical data. The chemical solubility was
adjusted such that the amount of observed and simulated chemical in the
litter layer was in close agreement (Table 4.38). The chemical
-------
298
l/>
o>
•r-
•o
3
+J
t/)
cn
c
•r—
r—
(1)
TJ
0
E
T3
C
ro
E
O
O
o
0
•r~
E
re)
^
i^.
O
c
o
•r-
S-
(TJ
a.
J
•
oo
CO
^f
01
p—
X)
ns
r-
O)
-o
O
y
E
(^
O
O
O
S-
o
•1 —
^r
o cn ^~ cn * — i
• • oo tn in
CO O CM I— 1 r-l
cn 10 t— i r-i
^
0
c
rvi
TJ
0
XI
Q-
omom in Or-io <&
o • • • * — t ' — i • • in
r- 1 o r-i co in o o
CO
Ocooom r~ i^^oo o
O » — ' c\j cn CM oo CM • CM
O <— i «— i r>»
•
r--
f~) ^j~ r^ r^ r^« m i — 1 1 — i in
o • • • o i — ' • • i — t
COOOCM i— 1 CMO i— 1
*
CM
omiooo i— i cMioio o
ocMsj-cn oo r^-oo* r^
O i— 1 ID r-l i— 1 CM
«t n I— 1 f\
CM in cn
lŁ) r-l
r-l
in co i — i oo cn
• ^ oo oo m
<Ł> i— 1 CM r-l i-H
r— 1 r-l
3
t ^
C
fsj
^D
0
/*^
Qu
o o o CM in CM r^s i^» o
in r-l CO r-l OO
O CO CM
«
CO
or^r--oo in Oioco r^
om^i-CM oo oo co •— i O
C^ ^d" 10
«
in
00
oc\jinio o cnooco cn
o • • oo i— i
*r o o •— i
*
CM
o CM cn i— i o m CM i— < o
o co o o CM «*• ^i-
f\ t\ •* *i
co cn •— ( in
in •— i
E
E u E
E o o
o in
o CM m
in i— i i CM
1 1 O 1
o in r-i o
*— ^ ^-^
$_ ^^ • — -CM
^-^ ^ ^ 0
>> "••••, E O) CJ) O> CD CD ^*^
^» cn nj ^,^,-v.^. 3- ^ cn
E *~-^ ^"^ cn cn cn cn *• — * cn ;n
O +•> 3^ 3. 3. 3- 3- *^-^
4-> 5 CM 0) O r—
• — • c in - — - E >coccc-r-c n)
1. o OJCM >>^, t. ±e o o o +* o o
fO *f~" ^ C t CT) fQ (^ «|— »|— «^- «. i- x: — a>4J i— rss-s-j-cs- x:
f= -r- 1—5 n3 +-> +J 4-> O) -rJ O
o a. r— 4J o c 4-> c c c u c:
*^^» ^ (O !5 O ^^ «^ J~ QJ fl) Oi) C ^ ^~
CC s. , + -o — T— c c c o c o
cni-<4-j_ Q.OOO O in
It) 4J — 'T3 XTS- r— OOOS-O
CO (JO) >dr— 0> r—
•i- a. C 4- C 4-> Ort3+JE4-+->i — ja a> s- -i- EOO4-> I— o: i^ _J — i oo t—
(O tG JC
30 0
1/5
^_
o
40
c
0
o
c
c
o
•*->
1-
+J
c:
O)
o
c
o
u
01
•4-)
O)
o
r^
TJ
0)
in
«3
Cl)
o
c
CD
x:
+1
m
c
0)
in
ai
Q.
O)
t.
c
o
• r—
4->
i-
^_>
C
01
u
o
o
r~-
»r—
0
in
-o
c
A
OJ
1 <
-M
•r-
•~
*
-M
C
51
A3
-------
299
distribution coefficient (K .) influenced the soil solution concentra-
tion and thus the rate of chemical transport between soil layers. This
parameter was adjusted to modify the pattern of heavy metal transport
through the soil system. The total plant uptake of heavy metals was
adjusted by the appropriate change in the root solute conductivity.
Finally the distribution of the chemical between plant parts was
obtained through the introduction of parameters for chemical chelation
within the plant. These parameters maintained differential levels of
heavy metals in the soluble (thus transportable) form for each plant
tissue (root, stem, leaf, fruit). The final agreements and disagree-
ments in the chemical levels (Table 4.38) were accepted as adequate for
the purposes of this study.
The tuning process proceeded by the adjustment of parameters
sequentially from the independent to the more and more dependent vari-
ables, providing a stepwise process of adjustment. This process is
such that it is difficult to complete unless some predefined optimiza-
tion routine can be used. Parameter adjustment did not result in mini-
mum statistical deviation between observed and simulated results.
The tuned parameter values may be viewed as apparent values in
which the complexity of phenomena in the microcosm system is synthesized
into a set of apparent values as far as the relatively simple model
structure is concerned. Because the water and carbon fluxes in the
microcosms are not related to those at the Crooked Creek Forest, none
of the water and carbon parameter values for the microcosm were trans-
ferred to the Crooked Creek application. Four chemical parameters tuned
to the microcosm were used in the final choice of chemical parameters
-------
300
for the Crooked Creek test application. These parameters included the
distribution coefficient (K^), chemical solubility, root solute con-
ductivity, and plant chelation factors. Both the distribution coeffic-
ient and chemical solubility can be readily measured by laboratory
procedures; however, the tuned values can differ from this measurement,
due to inadequacy of models to represent all significant phenomena.
The tuning process derived values that represent the apparent value for
microcosm phenomena according to the model representation. Root solute
conductivity is a highly sensitive parameter and not readily determined
experimentally. The microcosm data are ideal for estimation of this
parameter. Initial tuning runs showed that there were significant
differences in the uptake of each heavy metal, and the model was changed
to allow a specific conductivity value for each chemical. The original
model structure resulted in preferential chemical accumulation in vege-
tation in the tissue chemical uptake (e.g., roots). The microcosm study
showed that chemicals moved readily in plants, with highest concentra-
tion occurring in leaves. A chelation parameter was introduced into
the model and tuned to reflect the experimental data.
4.4.2.2 Field Study and Test Case Simulation
Heavy metals (P, Cd, Zn, Cu) released from a lead mine and smelter
complex into a deciduous forest in southeastern Missouri have resulted
in a reduction in litter decomposition rate (Watson et al. 1976) and a
change in the nutrient cycling fluxes (Jackson and Watson 1977). These
field studies conducted on the Crooked Creek Watershed during 1974 and
early 1975 report effects of continued heavy metal deposition since the
industrial facilities commenced operation on the watershed boundary in
-------
301
late 1968. Simulation of heavy metal transport and accumulation in the
soil-plant-litter system at 0.4 km from the industrial source was con-
ducted for the test case comparison.
Meteorological and precipitation records for the seven-year period,
October 1968 to September 1975, were coded into the model format from
available records at surrounding meteorological stations. Hourly pre-
cipitation was mainly obtained for the Viburnum station which is 10 km
north of the study area. Average wind speed, maximum and minimum daily
temperature, and average dewpoint temperature were obtained from both
onsite records and from Salem, Missouri (38 km west).
Soil hydraulic properties were determined onsite with the instan-
taneous profile method by Doll (1976), and these data were used as
inputs to the soil-piant-water flow component model. Parameters repre-
senting second growth oak (Quercus alba L., Q.. velutina Lam.) forest
were the same as previously used in simulations by Dixon et al. (1976)
in their application of the CERES forest growth model to Crooked Creek
Watershed. The mean deposition rate of heavy metals from the industrial
-2 -1
operation was 189, 1.26, 16.2, and 7.2 ^g cm month for Pb, Cd,
Zns and Cu, respectively (Begovich and Jackson 1975). Data for the
exchange capacity and bulk density for the soil profile were also taken
from Begovich and Jackson (1975).
The simulation results for chemical transport obtained with the
data set tuned to the microcosm results showed almost no agreement with
the experimental data obtained from Crooked Creek Watershed (Table 4.39).
The only exception was the cadmium and copper concentration results for
the Al soil layer. The 2.4-fold difference in evapotranspiration and
-------
302
Table 4.39. A comparison of initial simulation, experimental data, and
simulation results from this study for Crooked Creek Watershed
Simulated3 Experimentalb
This study
Total litter mass (g/m?)
Litter concentration (ppm)
Lead
Cadmium
Zinc
Copper
Leaf tissue concentration
Lead
Cadmium
Zinc
Copper
Root tissue concentration
Lead
Cadmium
Zinc
Copper
Stem tissue concentration
Lead
Cadmium
Zinc
Copper
2448.0
51000.0
5.3
121.0
375.0
(ppm)
170.0
10.0
50.0
10.0
(ppm)
1150.0
52.0
52.0
10.0
(ppm)
27.0
9.0
35.0
18.0
3600.0
72000.0
150.0
2350.0
1400.0
467.0
3.9
47.0
10.0
2400.0
46.0
48.0
6.0
680.0
3.9
30.0
13.0
4972.5
2.76
0.017
0.23
570
0.436
0.004
0.151
0.061
3.35
0.04
1.28
0.50
3.39
0.04
1.38
0.55
Soil concentration Al layer (ppm)
Lead
Cadmium
Zinc
Copper
271.0
4.4
52.0
46.0
334.0
3.55
25.0
18.0
2561
4.03
55.3
17.5
aFrom Munro et al. 1976.
bFrom Bondietti et al. 1976.
-------
303
drainage in the microcosm tuning exercise cannot account for the differ-
ences in the field test case because values of root solute conductivity
would have been lower at higher water fluxes into roots to give the
same total chemical uptake. It is evident that the chemical parameters
obtained for a seedling tree in a greenhouse environment cannot be
directly applied to mature forest and that the applicability of models
and microcosms in testing protocols requires careful evaluation.
The heavy metal concentration in litter was much lower for Pb, Cd,
and Zn in this study, which was mainly the result of high chemical
solubility values. However, Cu had a lower solubility and the litter
concentration was somewhat closer to the experimental result than that
obtained in the Munro study. The heavy metal concentration in plant
tissues in this simulation study was very much lower than the earlier
simulation which gave results that were in closer agreement with the
experimental data. The root solute conductivity was three to five
orders of magnitude smaller in the tuned data set than was that used in
the Munro et al. (1976) study. The large difference between the experi-
mental and simulated plant chemical concentrations shows that the root
conductivity parameter must also be determined for the real world con-
ditions.
The distribution coefficients used in this simulation were much
lower than those used by Munro et al. 1976, resulting in greater chemi-
cal mobility through the soil. There was closer agreement between the
simulated and experimental data for the soil comparisons than between
those for the plant. The inference is made that the apparent K.
values derived from microcosm studies provide a better representation
-------
304
of the effective chemical distribution between solid and liquid phases
than a laboratory determination of the K , for the soil of interest.
This makes sense if one considers that other phenomena (e.g., chelation)
influence chemical mobility, and these are not accounted for in the
laboratory determination of K. but are included in the derived value
from microcosm studies.
4.4.2.3 Discussion
A recent workshop proceedings on terrestrial microcosms (Gillett
and Witt 1979) recommended that mathematical models using "basic physio-
chemical and biologic data" be tested in microcosm studies with the
results in turn being used to develop models for environmental behavior
of chemicals. This has been done in the present study and modifications
to the models are believed to increase the real world utility of the
codes. The same workshop identified the need to relate quantitatively
the data from microcosms to chemical behavior in natural ecosystems.
This need was addressed in this study with the simulation models provid-
ing the means of quantitatively relating microcosm data to real world
chemical behavior. Our test case application showed that we failed
completely in this task, not because of inadequate model algorithms but
by the choice of parameter values derived from the microcosm data.
Clearly the simulation is limited by the field relevance of the micro-
cosm data set. It was a gross extrapolation to expect that parameter
vaTues derived for a sapling tree grown under greenhouse conditions
should be relevant to a Missouri deciduous forest. The tuned model
data set could perhaps be more readily used to evaluate heavy metal
behavior in a tree nursery-greenhouse facility.
-------
305
Several advantages of the combined use of modeling and microcosm
methodologies may be identified:
(1) Microcosm data can be used to evaluate model algorithms and
structure.
(2) Models may also be validated by microcosm data sets in which
the model is first tuned to one data set and then the model
simulation compared with another data set.
(3) Microcosm data provides one means of model parameter estima-
tion; however, the parameter values may not be representative
of natural ecosystems.
(4) The combined methodologies reduce biological complexity
(microcosm) to a simplified representation (model), allowing
the generation of time extrapolation insights.
The limitations of combined microcosm and modeling methodologies
in protocols for assessing real world phenomena relate to the relevance
of the methods to the real world behavior. The modeling exercise could
be at a double disadvantage by being limited by both algorithm simplifi-
cation of the real world and errors in input parameter estimation.
The exercise undertaken in this study is a first step in the
development of a combined microcosm-modeling protocol. The study showed
that the model could be tuned to microcosm systems, and we believe that
simulation of real world phenomena is possible with suitable sources of
input parameter values. Field studies or "in situ" microcosms are more
likely to provide relevant input data for models than indoor microcosm
studies. Further, the manual stepwise tuning process used in this study
should be replaced with some computer parameter optimization procedure
-------
306
to reduce subjectivity. This has been done for a parametric hydrology
model (Fields and Watson 1975) and the approach can be readily adapted
to the models used in this study.
4.5 DISCUSSION
4.5.1 The Role of Terrestrial Microcosms in Contaminant Research*
(B. S. Ausmus, D. R. Jackson, and P. Van Voris)
4.5.1.1 Introduction
Need for Accuracy of Screening Techniques
Microcosm experimental units are confined, replicable tools which
have been developed in response to recent legislative requirements con-
tained within the Toxic Substances Control Act (P.L. 94-469) and the
Resource Conservation and Recovery Act (P.L. 94-580). These Acts
require testing of potentially hazardous substances resulting from
manufacturing, processing, and waste disposal. Criteria for testing
apply to \those substances (1) which may present an unreasonable risk to
health or environment, (2) for which insufficient data exist for assess-
ment of risk to be made, and (3) for which testing is required to
develop data needed for assessment of risk. Substances deemed
potentially hazardous to health or environment, or produced in
substantial quantities (thereby increasing the risk of exposure), can
be required to be tested.
The criteria for testing remain unclear (P.L. 94-469) and must be
approved by the Environmental Protection Agency (EPA) before use.
*Taken from Ausmus, B. S., Jackson, D. R., and P. Van Voris. 1979. The
Accuracy of Screening Techniques, pp. 123-130. IN James M. Watt and
James W. Gillett (eds.), Industrial Microcosms and Environmental
Chemistry. National Science Foundation (NSF), Washington, D. C.
-------
307
Procedures for use as the initial screening, termed Level 1 Environ-
mental Assessment: Biological Tests (Duke et al. 1977), have three
goals. These are: (1) provision of preliminary environmental assess-
ment data; (2) identification of problem areas; and (3) collation of
data needed to prioritorize further assessment activities. Two addi-
tional levels of assessment remain to be developed. Level 2 studies
will allow more specific studies investigating rates of transport and
levels of effects. Level 3 studies will be designed to discern the
exact mechanisms disrupted by the substances, allowing appropriate
management control or elimination of problems.
Under legislation, research to develop these screening tools can
be sponsored by EPA. Specific testing of the substances, however, are
the responsibility of industry. The initial screening reports accepted
by EPA are expected to be the 'pacesetters' for following reports. The
EPA will require only enough data to reach regulatory decisions and
will update acceptable screening procedures as the science (or art)
progresses. These updates will not jeopardize the acceptability of the
earlier-prescribed screening procedures (P.L. 94-469).
Microcosm Screening Techniques
Microcosms have been used to address several levels of biological
complexity, including cellular, tissue, organ, organismal, population,
community, and ecosystem (Witherspoon et al. 1976). The gamut of
experimental units, then, includes pure cultures to excised, intact
portions of terrestrial landscapes (Table 4.40).
Microcosm studies are being used as experimental tools to assess
chemical transport, fate, and effects upon ecosystem components,
-------
308
Table 4.40. Uses that various microcosm units have for specific
screening applications
Microcosm type
Screening
Application Bioassay
Total transport
Total fate
Distribution
Accumulation
Strata transport
Accumulation
Within strata
Trophic transfer
Target species X
Target communities
Process disruption
Process sensitivity
System disruption
System sensitivity
Mechanisms of
disruption X
Biomag-
nifica- Strata
tion cores
X
X
X X
X X
X X
X
X
X X
Excised
systems
X
X
X
X
X
X
X
X
X
X
X
X
X
In situ
studies
X
X
X
X
X
X
X
X
X
X
X
X
-------
309
specifically populations, and trophic web members (Metcalf et al. 1973,
Booth et al. 1973, Focht and Alexander 1970, Francis et al. 1975).
Within these studies, a variety of artificial and homogenized substrates
have been used to construct experimental units (Witherspoon et al.
1976). However, these studies address ecosystem components rather than
ecosystem processes, such as nutrient cycling and carbon metabolism.
The effects upon an ecosystem's ability to compensate for and to recover
from chemical stress cannot be inferred from studies of ecosystem com-
ponents since the interconnections among components to form an ecosystem
are not necessarily additive (O'Neill et al. 1975, O'Neill et al. 1977).
The criteria for microcosm acceptance as screening tools include
their demonstrated replicability, reproducibility, and accuracy or
realism of the data produced. These will provide initial quality con-
trol for screening procedures, and quantitative background information
for regulatory decisions.
Witherspoon et al. (1976) discuss the complexities, applications,
measured parameters, reproducibility, and realism in their review of
microcosm research. They surmarized the suitability of various designs
for assessment of chemical accumulation sites, mutualistic and antago-
nistic processes, transport of the substances, and identification of
critical problems in microcosm use. Problems listed included: (1) lack
of demonstrated applicability of transfer rates and distribution of sub-
stances within microcosms to ecosystems; (2) lack of comparability or
reproducibility of microcosm experiments; and (3) lack of estimated con-
fidence intervals around microcosm measurements of several parameters.
-------
310
Objectives
We have examined several types of microcosm experimental units
with the ultimate purpose of assessing their utility as analogs of ter-
restrial ecosystem processes. We had three research goals; (1) to con-
struct terrestrial microcosms which can be quantitatively equilibrated,
replicated, and reproduced; (2) to identify and nondestructively measure
parameters in microcosms which are indicative of ecosystem response to
and recovery from stress; and (3) to quantitatively relate microcosm
results to ecosystems (Table 4.41).
This discussion focuses on three themes. These are: (1) param-
eters which can be nondestructively monitored in microcosms to assess
the transport rates of a substance and concomitantly its effect upon
ecosystem processes; (2) the criteria for and comparison of the replic-
ability and reproducibility among microcosm types tested; and (3) the
criteria for and comparison of the accuracy of results among the micro-
cosm types tested.
4.5.1.2 Measured Parameters
Criteria
There are five criteria which should be satisfied by the parameters
chosen to measure in microcosm screening of potentially hazardous sub-
stances. First, the parameters measure should be indicative of terres-
trial ecosystem function. That is, the measurements should be holistic
rather than specific to ecosystem components, such as suspected target
populations. Second, parameters should be used which can be nondestruc-
tively monitored throughout the experiment. This will allow detection
-------
311
Table 4.41. Microcosm test systems, substances tested, and citations
Microcosm test system
Contaminant
Citation
Soil cores, no plants
Soil core with plants
Watershed, field
Cd
Pu, Cu, Cd, Zn
As
HCB
Cd
Pu, Cu, Cd, Zn
As
Pu, Cu, Cd, Zn
Ausmus and O'Neill, 1978
O'Neill et al., 1977
Jackson and Levin, 1979
Ausmus et al., in press
Van Voris, 1978
Jackson et al., 1978
In experimentation
Jackson and Watson, 1977
Watson et al., 1976
-------
312
of changes in parameters through time without the confounding which
results from repeated removal of components of the microcosm units.
Third, parameters monitored should allow rapid data acquisition at
several intervals during the experiment. Daily or weekly measurements
of parameters in experimental and control microcosms will increase
statistical rigor of the screening experiment. Fourth, parameters
should be chosen which allow both transport of the tested substance and
effects on ecosystem processes to be simultaneously determined.
Concentrations of dissolved organic carbon (DOC) and of the substance
under study, for example, could both be monitored in leachates
periodically taken from microcosms. Finally, parameters which are
needed but which cannot be monitored can be determined on replicate
microcosms sacrificed at time zero, and on all microcosms at time
final. These include distribution of tested substance among ecosystem
components and estimation of microbial, invertebrate, and plant biomass.
Monitored Parameters: Indices of Ecosystem Response to Stress
A natural candidate for terrestrial monitoring would be the rate
at which macronutrients and carbon are exported from microcosms.
Shugart et al. (1976) suggest that dissolved nutrients within soil water
may be indicative of total ecosystem cycling of calcium. Microcosm
studies, which treat these experimental units as model ecosystems, have
shown that carbon and nutrient efflux from these systems are affected
by chemical stress (O'Neill et al. 1977). Soil, because it contains
the largest number of interacting components within terrestrial ecosys-
tems, becomes the subsystem focused upon to provide initial screening
-------
313
procedures. Because of the large number of interacting components in
soil, detrimental effects are exhibited as increases in gaseous .and
dissolved carbon loss, and in increased dissolved macronutrient loss.
This loss can be detected irrespective of the specific organisms and
enzymatic pathways which were disrupted by chemical stress. The exact
mode of effects may not be intuitively perceived. Therefore, focus
upon a holistic index of ecosystem response to such an effluent would
be desirable. Three studies, using soil only, intact soil-tree sapling
units, and a watershed illustrate this point (Table 4.42). In each
case disruption of nutrient cycling processes was detected prior to, or
under less stress than, changes in population or community parameters.
A recent aquatic microcosm study (Strange 1976) showed similar increased
nutrient mobility in response to herbicide treatment.
Monitoring nutrient loss rates may not be necessary to detect eco-
system response to chemical stress. An index of nutrient status of soil
under stressed conditions shows depletion of nutrient pools (Jackson
and Hall 1978, Jackson et al. 1978, Ausmus et al. 1978). Figure 4.1
illustrates depletion of extractable nutrients soil following heavy
metal application.
Measurements at Harvest: Distribution and Fate
While the rate of nutrient efflux and of tested substance efflux
can be monitored throughout microcosm experiments, the distribution of
the tested substance through microcosm components must be determined at
the experiment's conclusion. Coupling monitored data with those at
harvest, researchers can determine the mass balance of the tested sub-
stance and of the macronutrients of interest. The distribution of lead
-------
314
Table 4.42 Population and system level parameters measured in soil microcosms, forest microcosms,
and forest under control and chemical stress conditions (after O'Neill et al., 1977)
Level control
and treatment
Arsenic
(NaHAs04)
Arsenic
Heavy metals
Distance from
lead smelter
0
100 ppm
0
100 ppm
0
11 mg Pb/cm2
21 km
2.0 km
Population
parameter
ATP (ppm)
Bacterial
density (106/g)
Annual branch
growth (cm)
Litter invertebrate
diversity (Shannon
Index)
Values
1.58
1.52
2.75
3.18
347
346
2.3
2.5
± 0.33
± 0.30
± 0.13
± 0.41
± 43
± 68
+ 0.15
± 0.23
System
parameter
Leached Ca (ug} '
Leached P (ug)
Leached Ca (ppm)
0.2 litter
mass (g/m<-)
Values
560
714
2.8
21.0
6.6
10.0
1008
1595
± 36
± 50
± 0.5
± 6.0
± 0.4
± 0.8
± 110
± 171
-------
315
in forest microcosm components at the end of a 20-month experiment is
summarized in Table 4.3.
Measurements at Harvest: Biological Composition
Detection of hazardous substance effect upon biota can also be
made at experiment conclusion. Biomass differences between treated and
control microcosms may indicate population shifts or community degrada-
tion due to chemical stress. Table 4.43 shows three indices of soil
microbial populations measured at the conclusion of an experiment.
Control microcosms had significantly greater fungal populations than
microcosms which had been contaminated with a complex effluent of Pb,
Zn, Cd, and Cu (Ausmus et al. 1978).
4.5.1.3 Replicability, Reproducibility
Criteria
In order to establish quality control of screening procedures three
criteria must be met. First, the parameters must be predictable through
the experimental period. Usually, this requirement necessitates equi-
libration of the experimental units prior to treatment with test sub-
stance. Second, confidence limits must be established for the measured
parameters. Unique microcosms for each substance, or each dose rate
are inapplicable. Third, reproducibility of parameters between control
microcosms between comparable experiments must be reasonable.
Comparison of Reproducibility
Several comparable experiments have been conducted using the intact
soil profile with and without aboveground vegetation. These experiments
-------
316
Table 4.43. Indices of microbial activity measured at termination of a
microcosm experiment using complex heavy metal emissions
from a lead smelter (Ausmus et al., 1978)
Microbial parameter
ATP concentration (ppm)
Bacterial density (l()8/g)
Fungal lengths (m/g soil)
Treatment
Contaminated
Control
Contaminated
Control
Contaminated
Control
Soil
0.5 cm
13.?a
5.6
21. 8a
8.0
75*
1300
5-10 cm
4.2b
2.1
10.23
5.1
1353
750
Significantly different from controls at P Ł 0.01.
Significantly different from controls at P < 0.05.
-------
317
allow us to estimate reproducibility between experiments. An example
of the findings is included as Table 4.44. We believe these microcosm
units yield reproducible results as evidenced by the overlapping error
terms for each of the three parameters among experiments.
Comparison of Replicability
Replicability among experimental units within a treatment class
has been calculated during experiments using intact, homogenized, and
sand-soil substrates. We have found that replicability is dependent
upon substrate type and upon the parameter estimated. For example,
C02 efflux was replicable in all substrates, but DOC efflux varied
greatly among homogenized and sand-soil replicates at each sampling
date and through time (Fig. 4.24). Similar results were obtained for
nine nutrients (Table 4.43). Results from grass microcosms established
using these substrates were similar. We conclude that replicability of
intact soil microcosms is greater than other substrates.
4.5.1.4 Accuracy
Criteria
The criteria for determining microcosm result accuracy are dual.
First, parameters indicative of ecosystem processes should be similar
to those measured within comparable strata of the intact ecosystem.
Second, the distribution and the transport rates of the test substance
should mimic those which would occur in the source ecosystem. These
criteria are the most difficult of those needed to establish the utility
of microcosms in screening potentially hazardous substance. Only
limited data have been collected to address this issue. The comparisons
below represent our findings to date.
-------
318
Table 4.44. Comparison of three parameters measured in control
replicates of three soil core microcosm experiments
Parameters
C02 efflux
Experiment (yg/day)
la 5.73 ± 0.19
2b 4.21 ± 0.35
3C 3.17 ± 0.99
Concentrations
N03-N
(pg/ liter)
32.4 ± 4.8
30.1 ± 8.2
27.3 ± 9.2
in soil water
Ca
(vtg/ml)
20.3 ± 1.7
15.5 ± 5.6
18/6 ± 6.1
aJackson et al., 1978.
DAusmus and O'Neill, 1978; O'Neill et al., 1977.
cAusmus et al., unpublished data.
-------
319
103
5 - 0 HOMOGENIZED (
@ HOMOGENIZED (
~ 0 SAND- SOIL
2
/ ^ ^
1
IV — ^ - — - — -pj-
^
^ , ^5
r ~ ? s " ^ ~
40 48 54
SUBGROUP 1 )
SUBGROUP 2
rn
^
/• /• r
t i P3
f 1 ^
In ! i
1~ - : }
''> i ^
j 5
^ | 5 \ :
M 111-
L y H |
II ll II 11
7]
-
-
-
61 68 75 82 89
YEAR DAY
ORNL-DWG 76-13428
(1) HOMOGENIZED SUBGROUP 1
Ł=1591.9 +138.14* (/?2=0.996)
(2) SAND SOIL
Ł=-385.27+112.98* (/?z = 0.914)
(3) POOLED HOMOGENIZED SOIL
(4) INTACT SOIL
Y= 63.5 + 24.96* (/?2= 0.987)
(5) HOMOGENIZED SUBGROUP 2
8000
'E 7000
o
o
o
| 6000
o>
u 3 500°
cj
d
<2 4000
- -1 3000
_ uj
r H 2000
o 1000
;=-37.Ł
•17.69* (A":=0.969)
^Ti i i i r
Y= CUMULATIVE D.O.C.
EXPORT
*=DATE WITHIN INTERVAL
OF LEACHING
7.
//_
-------
320
Comparison of Accuracy
Export of essential elements in water leached from intact, homoge-
nized and sand-soil substrate microcosms were extrapolated to one year
and compared. Table 4.45 shows the relative loss from homogenized and
sand-soil substrates compared to intact soil. If we assume that the
standard is intact soil, the comparability of other substrate types is
poor. No ion is similarly lost from the three substrate types.
Intact coil microcosms can be quantitatively related to ecosystems.
Using data from the source ecosystem, transport of nutrients below the
rooting zone of the forest (60-cm depth) was related to transport
through intact soil microcosms (5-cm depth) by the expressions.
YI = 0.184 Xi ,
where Y^ = DOC, NH3-N, N03-N, PO^-P, or K transport to 60 cm depth
2
(g/m /year), and
X1 = DOC, NH3-N, N03-N, P04~P, or K transport to 5 cm depth
2
in intact soil microcosms extrapolated to g/m /year.
YZ = 0.680 X2 ,
2
where Y2 = Ca, Na transport to 60 cm depth (g/m /year), and
X~ = Ca, Na transport to 5 cm depth in intact soil microcosms
2
extrapolated to g/m /year.
While these relationships alone are not conclusive proof of the accuracy
of intact soil microcosm results, they are strong evidence that such is
the case (Ausmus and O'Neill 1978, O'Neill et al. 1977).
-------
321
Table 4.45 Comparison of extrapolated annual export among substrates
(O'Neill et al., 1977)
Percent extrapolated
Substrate
Homogenized
Block 1
Block 2
Sand-soil
Ca
108
168*
130*
K
74*
76*
20**
DOC
76*
366**
266**
P04-P
100
400**
20**
intact soil microcosm transport
NH3-N
35**
43**
35**
N03-N
180*
30**
18**
Mg
100
138*
138*
Na
128
450**
200*
transport extrapolated from intact soil (g/m2/year); Ca 5.8, K 5.4, DOC
6.2, P04-P 0.03, NH3-N 0.66, NOs-N 1.4, Mg 0.9, and Na 0.6.
* Significantly different from intact soil PŁ 0.05.
**Signif icantly different from intact soil P Ł 0.01.
-------
322
More conclusive evidence was obtained from forest microcosm
experiments (Jackson et al. 1978). Using the data shown in Table 4.3
and comparable data for cadmium, zinc, and copper, microcosm data were
compared to the forest ecosystem under stress from the heavy metal
deposition (see Watson et al. 1976, Jackson and Watson 1977). Enrich-
ment ratios, ER, were computed by dividing the heavy metal concentration
of the treated microcosm components by the same metal concentration of
the control microcosm component. Greatest enrichment of metals occurred
in litter strata. Cadmium had greater ER for soil than did other
metals, although Pb was greatly enriched. Values compared to those
near the heavy metal source were similar to the sampled ecosystem com-
ponents (Jackson et al. 1978).
The most conclusive proof of accuracy to date has been the non-
significant difference in field and microcosm data taken on similar
dates.
4.5.1.5 Problems
Several problems remain in the evaluation of the utility of micro-
cosms as screening tools to assess transport, fate, and effects of
potentially hazardous substances. Four problems seem important. These
are: (1) the rate of ecosystem degradation cannot be calculated from
loss rates of nutrient elements as yet, although in theory it is pos-
sible to do so (O'Neill et al., in press); (2) the microcosms have not
been satisfactorily tested using organic effluents, singly or in combi-
nation; (3) absolute accuracy of microcosm results has not been proven;
and (4) reproducibility among microcosm studies conducted using intact
profiles excised during different seasons has not been established.
-------
323
Some information is needed to conduct adequate screening of poten-
tially hazardous substances. The mode of entry of the substance into
the system is needed. The direct toxicity of the substance to research
workers is needed in order to practice safe laboratory experiments.
The potentially targeted ecosystems should be identified if possible
and used as sources of microcosm units. Lacking this information,
forest or pasture systems should be used because more than 80% of the
non-urban United States is one of these ecosystems.
High mobile invertebrates, macroinvertebrates, and vertebrates are
not well represented within these small microcosm units. Studies
directed toward the potential effects of these substances on these, due
to their economic importance or trophic position, should be conducted
as biomagnification experiments.
4.5.1.6 Recommendations
Based on the several experiments conducted to date, we recommend
the use of intact soil microcosms excised from targeted ecosystems and
encased with or without aboveground vegetation, depending upon experi-
mental purpose. We cannot recommend the use of artificial substrates,
or homogenized soil reinoculated with biota for several reasons,
including: (1) instability of these substrates with respect to param-
eters monitored; (2) limited replicability and accuracy of these units;
and (3) proportionally higher cost in experimental set-up.
There is no substitute for choice of the appropriate experimental
design for the experiment to be performed. We have found that a ran-
domized incomplete block design with incomplete factorial treatment
arrangement is cost-effective and information rich. Analysis of
-------
324
parameter behavior through time under control and stressed conditions
may furnish additional insights into the mode of effects on ecosystem
processes and on response and recovery times which can be expected to
levels of the stressor (Van Voris et al. 1978).
4.5.2 Protocol for Construction and Operation of Terrestrial
Microcosms (B. S. Ausmus and D. R. Jackson)
4.5.2.1 Objectives
The purpose of these protocols is to provide an accurate, rapid,
cost-effective method to assess contaminant transport, fate, and effects
upon terrestrial ecosystems likely to be impacted. The assessment is
two-tiered. The first protocol provides a rapid, binary (yes/no) solu-
tion to assessment questions and emphasizes the relationship between
dose and contaminant transport and effects. The second protocol is
more extensive, and allows quantitative assessment of rates of trans-
port, accumulation, and nutrient cycling disruption.
4.5.2.2 Rationale
The initial question when faced with a potentially hazardous sub-
stance is whether it is mobile, toxic, or disruptive of ecological pro-
cesses at various dosages. After determining whether these do occur,
and if so, at what doses, then one must ask a battery of questions
(within a second protocol), including at what rates transport and accu-
mulation occur.
Previous studies have shown that excised, intact soil blocks with
(protocol 2) or without (protocol 1) plants are the most accurate,
-------
325
replicable, and cost-effective microcosm units for contaminant
assessment (O'Neill et al, submitted; Ausmus and O'Neill, submitted).
Protocol 1 uses intact soil cores without plants to assess dose/mobility
and dose/effects relationships, exploiting the fact that terrestrial
ecosystem processes are predominantly belowground. Protocol 1 should
answer specifically:
(1) Is the contaminant (or some by-product) mobile?
(2) Is the primary mode of soil export gaseous, dissolved, or
particulate?
(3) Is nutrient cycling disruption indicated?
(4) Are soil biota affected?
(5) Do questions 1 to 4 change as a function of dose?
Previous studies have indicated that assessment of whole system
transport, fate, and effects is more time-consumptive and expensive
than assessment of soil processes. Protocol 2 addresses quantitative
assessment of transport, fate, and effect of contaminant under study
throughout the ecosystem. Specifically, we will use protocol 2 to
answer:
(1) What is the rate of contaminant transport?
(2) What transformation/degradation products result and how does
their mobility, etc., compare with that of the parent compound?
(3) How fast does nutrient/carbon depletion occur? Is this effect
sustained or does recovery occur?
(4) Where are contaminant accumulation sites?
(5) What is the rate of plant uptake?
-------
326
(6) How does the contaminant partition among biotic and abiotic
components?
(7) What is the effect upon: (a) plant biomass, (b) microbial
community dynamics, and (c) invertebrate community dynamics?
In order to obtain the most useful information from either proto-
col, it is preferable to use cores extracted from those terrestrial
systems most likely to be the target of contaminants under assessment.
These protocols were developed for forest and grassland, since ^ 80% of
U.S. landscape is one or the other of these. We believe, lacking iden-
tification of target ecosystems, that use of cores from convenient
unimpacted forests or grasslands will yield results representative of
these ecosystem types, regardless of the specific microclimatic regime.
We cannot recommend use of artificial substrates or homogenized
soil that have been reinoculated with biota, for several reasons. Most
important of these are: (1) these substrates are unstable with respect
to the parameters being monitored with/without perturbation; (2) they
are less replicable and reproducible than intact systems; and (3) their
accuracy for assessing environmental transport rates, accumulation
rates, and effects (especially on ecosystem processes) is unpredictable
(Ausmus and O'Neill, submitted; Jackson et al., submitted)
Careful core excision, encasement, and equilibration will allow
users high quality control over contaminant assessments. Units have
been maintained for protocols 1 and 2 under microcosm conditions for
long intervals (three months to two years). Therefore, many units could
be prepared and equilibrated and used as needed.
-------
327
4.5.2.3 Protocol 1
(1) Obtain soil cores from representative terrestrial ecosystems,
5-cm diam x 5-cm depth; remove aboveground plant parts.
(2) Encase the cores in shrinkable polyvinyl chloride (or teflon,
2/1000 in., within shrinkable PVC if contaminant is organic) and gently
heat shrink until a tight bond with the core (minimum boundary flow) is
achieved. Leave enough lining above the soil surface to use gaseous
export traps, if necessary. Mount on glass funnels in test tubes.
Blacken core wrapping to negate abnormal algae growth. Place in envi-
ronmental chamber under as near the field conditions as possible (see
Fig. 4.25).
(3) Equilibrate two to three weeks. Leach with rainwater (known
water chemistry) three to five times (enough to obtain 20-30 ml/date
during equilibration. If possible, determine dissolved organic carbon
(DOC) concentrations in these samples. If possible use alkali traps to
determine daily CO^ efflux eight to ten days during equilibration.
Use these data to discard dissimilar replicate soil cores.
(4) Experimental procedure: randomized complete block and, if
possible, factorial treatment arrangement of dosages with a minimum of
three cores per dose per terrestrial ecosystem tested. Dosages of a
wide range should be used for this rapid screening test to maximize the
clarity of dose dependent observations.
(a) Add contaminant, if possible, as a uniformly labeled
radiolabelled compound so that transport of any derivative
can be determined with minimal chemical analysis. If
possible, add contaminant in the mode in which it would be
expected to enter the ecosystem, e.g., particulate deposition.
-------
328
MM1 TTTTrT' I ' I ' I ' I ' 1 "• T ' 1
t 2 3
OAK RIDGE NATIONAL LABORATORY
Figure 4.25.
Protocol 1 soil core excised intact from a forest
ecosystem.
-------
329
(b) Set gas traps to monitor C0? efflux and gaseous
emission of contaminant. Standard alkali traps, such as
0.2 N^ KOH can be used for CCL recovery. Contaminant
volatility, solubility, and chemical detection methods will
determine methods specifically used for each contaminant
tested. Collect data at 24-hr intervals.
(c) On Day 6, add sufficient rainwater (known water
chemistry) to collect 20-30 ml leachate per core. Analyze
DOC concentration. Determine contaminant concentration (if
possible by radioassay techniques).
(d) On Day 9, repeat step (c). After leachate collection,
harvest microcosm units. Cut cores into 1-cm depth inter-
vals. Within each depth, determine contaminant concentra-
tion, biotic activity by ATP assay, and/or adenylate energy
charge (Ausmus, in press; Ching and Ching 1972).
(5) Interpretation of results from protocol 1 will allow user
to discern: (1) dose relationship to transport and effect measurements;
and (2) major modes of transport (gaseous, dissolved), if transport is
significant.
4.5.2.4 Protocol 2
(1) Obtain intact soil blocks containing plants (grasslandrcircular:
15-cm diam x 10-cm depth; forest:square 40- x 20-cm depth, containing a
shallow-rooted plant such as red maple).
(2) Encase as in protocol 1 with this modification. Grassland:
use a base of Plexiglas (inorganic contaminant or glass, if organic
-------
330
contaminant) ported in the center with plastic or glass. Place encased
microcosm in container and fill around microcosm with ignited acid plus
distilled water-washed sand. Place on platform and attach leachate
bottle (see Fig. 4.26); maintain in an environmental chamber. Forest
microcosms should be encased in polyethylene or teflon (2/1000 in.) and
sealed tightly outside by placing in a plywood box and filling with
concrete (Fig. 4.27). Leachate ports can be placed at the side. If a
suitable site is available, maintain these microcosms out-of-doors.
(3) Equilibrate three to five weeks. Leach weekly as in
protocol 1. If possible, monitor (XL efflux daily. Determine con-
centrations of four nutrients in leachate: (a) grass land:DOC, NHL,
P04, Ca; (b) forest:DOC, N03, P04, Ca. Use these data to deter-
mine individual microcosm response to treatment after contaminant
addition.
(4) Experimental procedure: randomized complete block with a
minimum of three replicates and three doses per ecosystem.
(a) Add contaminant stable and (if feasible) radiotracer.
(b) Set gas monitoring traps as protocol 1.
(c) Leach weekly. Determine nutrient (four elements above)
and contaminant concentrations.
(d) Take plant tissue samples for determination of contami-
nant concentration weekly.
-------
331
Figure 4.26.
Protocol 2 grassland microcosm excised from a
Festuca-dominated pasture. Encased microcosm showing
heat-shrunk PVC, Plexiglass disk and leachate port.
Photograph taken nine months after encasement.
-------
332
Figure 4.27.
Protocol 2 forest microcosm containing an Acer rubrum
sapling excised from a mesic hardwood forest.
-------
333
(e) At eight to ten weeks, harvest microcosm units; analyze:
(i) extractable nutrients using 1 M KC1, 1 M NaHC03 intact
extraction procedure (Jackson and Hall 1978); (ii) above and
belowground plant contaminant concentration, biomass; (iii)
biotic activity (ATP and/or adenylate energy charge); (iv)
bacterial density (dilution plating); (v) fungal biomass
(Jones-Mollison slides); (vi) nematode and microarthropod
density and species composition (density gradient flotation)
(Parkinson et al. 1971; Plillipson 1971); (vii) contaminant
and daughter compounds concentrations in 1 cm depth of soil.
(5) Interpretation of results from protocol 2 will allow us to
(a) quantify transport rates, (b) determine longevity of nutrient
depletion, (c) calculate plant contaminant uptake rates; and (d) quan-
tify soil and plant populations affected by contaminant.
-------
5.0 LITERATURE CITED
Abbott, W. 1966. Microcosm studies on estuarine waters. I. The
replicability of microcosms. J. Water Pollut. Control Fed.
38:258-270.
Allen, H. L. 1969. Chemo-organotrophic utilization of dissolved
organic compounds by planktonic algae and bacteria in a pond.
Int. Rev. Gesamten. Hydrobiol. 54:1-33.
Ausmus, B. S., N. T. Edwards, and M. Witkamp. 1976. Microbial
immobilization of carbon, nitrogen, phosphorus and potassium:
Implications for forest ecosystem processes, pp. 347-416.
IN D. Parkinson, T. Berthet, and A. Macfayden (eds.), The Role of
Terrestrial and Aquatic Organisms in Decomposition Processes.
Blackwell Scientific Publications, Oxford, England.
Ausmus, B. S., and E. G. O'Neill. 1978. Comparison of carbon dynamics
of three microcosm substrates. Soil Biol. Biochem. 10:425-429.
Ausmus, B. S., G. J. Dodson, and D. R. Jackson. 1978. Behavior of
heavy metals in forest microcosms: III. Effects on litter-soil
carbon metabolism. Water Air Soil Pollut. 10:19-26.
Ausmus, B. S., Susan Kimbrough, D. R. Jackson, and S. E. Lindberg.
The behavior of hexachlorobenzene in pine forest microcosms:
Transport and effects on soil processes. Environ. Pollut. (in
press).
335
-------
336
Baes, C. F., C. L. Begovich, W. M. Culkowski, K. R. Dixon, D. E. Fields,
J. T. Holdeman, D. D. Huff, D. R. Jackson, N. M. Larson, R. J.
Luxmoore, J. K. Munro, M. R. Patterson, R. J. Raridon, M. Reeves,
0. C. Stein, J. L. Stolzy, and T. C. Tucker. 1976. The unified
transport model, pp. 13-62. IN R. I. Van Hook and W. D. Shults
(eds.), Ecology and Analysis of Trace Contaminants Progress
Report, October 1974 - December 1975. QRNL/NSF/EATC-22. Oak
Ridge National Laboratory, Oak Ridge, Tennessee. 200 pp.
Baldwin, J. P. 1976. Competition for plant nutrients in soil. A
theoretical approach. J. Agric. Sci. 87:341-356.
Baldwin, J. P., P. H. Nye, and P. B. Tinker. 1973. Uptake of solutes
by multiple root systems from soil. III. A model for calculating
the solute uptake by a randomly dispersed root system developing
in a finite volume of soil. Plant Soil 38:621-635.
74
Ball, R. C., and F. H. Hooper. 1966. Use of As-tagged sodium
arsenite in a study of effects of a herbicide on pond ecology.
pp. 149-163. IN Isotopes in Weed Research. International Atomic
Energy Agency, Vienna.
Becker, C. D., and T. 0. Thatcher. 1973. Toxicity of power plant
chemicals to aquatic life. WASH-1249. Battelle Pacific Northwest
Laboratories, Richland, Washington.
Begovich, C. L., and D. R. Jackson. 1975. Documentation and applica-
tion of SCEHM. A model for soil chemical exchange of heavy
metals. ORNL/EATC/NSF-16. Oak Ridge National Laboratory, Oak
Ridge, Tennessee. 67 pp.
-------
337
Begovich, C. 1., and R. 0. Luxmoore. 1979. Some sensitivity studies
of chemical transport simulated in models of soil-piant-litter
system. ORNL/TM-6791. Oak Ridge National Laboratory, Oak Ridge,
Tennessee.
Beyers, R. J. 1962. Relationship between temperature and the
metabolism of experimental ecosystems. Science 135:980-982.
Beyers, R. J. 1963. The metabolism of twelve aquatic laboratory micro-
ecosystems. Ecol. Monogr. 33:281-306.
Blackman, R. B., and J. W. Tukey. 1958. The Measurement of Power
Spectra, From the Point of View of Communication Engineering.
Dover Publications, Inc., New York. 190 pp.
Bolton, N. E., J. A. Carter, J. F. Emery, C. Feldman, W. Fulkerson,
L. D. Hulett, and W. S. Lyon. 1975. Trace element mass balance
around a coal-fired steam plant, pp. 175-187. IN S. P. Babu
(ed.), Trace Elements in Fuel. Advances in Chemistry Series 141,
American Chemical Society, Washington, D.C.
Bondietti, E. A., K. R. Dixon, S. 6. Hildebrand, J. W. Huckabee,
D. R. Jackson, S. E. Lindberg, F. H. Sweeton, R. R. Turner,
R. I. Van Hook, and A. P. Watson. 1976. Ecological research.
pp. 63-110. IN R. I. Van Hook and W. D. Shults (eds.), Ecology
and Analysis of Trace Contaminants Progress Report, October 1974 -
December 1975. ORNL/NSF/EATC-22. Oak Ridge National Laboratory,
Oak Ridge, Tennessee. 200 pp.
Booth, G. M., C. Yu, and D. 0. Hansen. 1973. Fate, metabolism, and
toxicity of 3-isopropyl-lH-2,l,3-benzothiadiazin-4(3H)-l-2,
2-dioxide in a model ecosystem. J. Environ. Qual. 2:408-411.
-------
338
Boothe, P. N., and G. A. Knauer. 1972. The possible importance of
fecal material in the biological amplification of trace and heavy
metals. Limnol. Oceanogr. 17:270-274.
Braunstein, H. M., E. D. Copenhaver, and H. A. Pfuderer. 1977.
Environmental, health, and control aspects of coal conversion: An
information overview. ORNL/EIS-94. Oak Ridge National
Laboratory, Oak Ridge, Tennessee.
Carrow, R. N., P. E. Rieke, and B. G. Ellis. 1975. Growth of
turf-grasses as affected by soil phosphorus and arsenic. Soil
Sci. Soc. Am. Proc. 39:1121-1124.
Copeland, B. J. 1965. Evidence for regulation of community metabolism
in a marine ecosystem. Ecology 46:563-564.
Cushman, R. M., S. G. Hildebrand, R. H. Strand, and R. M. Anderson.
1977. The toxicity of 35 trace elements in coal to freshwater
biota: A data base with automated retrieval capabilities.
ORNL/TM-5793. Oak Ridge National Laboratory, Oak Ridge, Tennessee.
Dixon, K. R., R. J. Luxmoore, and C. L. Begovich. 1976. CERES - A
model of forest stand biomass dynamics for predicting trace
contaminant, nutrient, and water effects. ORNL/NSF/EATC-25. Oak
Ridge National Laboratory, Oak Ridge, Tennessee. 102 pp.
Dixon, K. R., R. J. Luxmoore, and C. L. Begovich. 1978a. CERES - A
model of forest stand biomass dynamics for predicting trace
contaminant, nutrient, and water effects. I. Model description.
Ecol. Model. 5:17-38.
-------
339
Dixon, K. R., R. J. Luxmoore, and C. L. Begovich. 1978b. CERES - A
model of forest stand biomass dynamics for predicting trace
contaminant, nutrient, and water effects. II. Model application.
Ecol. Model. 5:93-114.
Doll, J. C. 1976. Horizon hydraulic conductivities determined in situ
for the wilderness, Mexico, and Menfro soils and used to predict
soil water content. Ph.D. Thesis. University of Missouri,
Columbia. 179 pp.
Dudzik, M., J. Harte, A. Jassby, E. Lapan, D. Levy, and J. Rees. 1979.
Some considerations in the design of aquatic microcosms for
plankton studies. Int. J. Environ. Stud. 13:125-130.
Duke, K. M., M. E. Davis, and A. J. Dennis. 1977. IERL-RTP procedures
manual: Level 1 Environmental Assessment Biological Tests. Draft
Final Report to EPA, January 14, 1977. 128 pp.
Feldman, C. 1977. Determination of traces of arsenic in silicious
materials. J. Anal. Chem. 49:825-828.
Ferguson, 0. F., and J. Gavis. 1972. A review of the arsenic cycle in
natural waters. Water Res. 6:1259-1274.
Fields, D. E., and S. B. Watson. 1975. OPTRM - A hydrologic transport
model with parameter optimization. ORNL/NSF/EATC-14. Oak Ridge
National Laboratory, Oak Ridge, Tennessee. 125 pp.
Focht, D. D., and M. Alexander. 1970. Bacterial degradation of
diphenyl methane, a DDT model substrate. Appl. Microbiol.
20:608-611.
Forbes, S. A. 1887. The lake as a microcosm. Bull Sc. A. Peoria.
Reprinted: Forbes, S. A. 1925. 111. Nat. Hist. Surv. Bull.
15:537-550.
-------
340
Francis, A. J., J. M. Duxbury, and M. Alexander. 1975. Formation of
volatile organic products in soils under anaerobiosis. II. Soil
Biol. Biochem. 7:51-56.
Fulkerson, W., W. D. Shults, and R. I. Van Hook. 1974. Ecology and
analysis of trace contaminants. ORNL/NSF/EATC-11. Oak Ridge
National Laboaratory, Oak Ridge, Tennessee.
Gass, S. I. 1977. Evaluation of complex models. Comput. Oper.
Res. 4:27-35.
Giddings, J. M., and G. K. Eddlemon. 1977. The effects of microcosm
size and substrate type on aquatic microcosm behavior and arsenic
transport. Arch. Environ. Contam. Toxicol. 6:491-505.
Giddings, J. M., and G. K. Eddlemon. 1978. Photosynthesis/respiration
ratios in aquatic microcosms under arsenic stress. Water Air Soil
Pollut. 9:207-212.
Giddings, J. M., B. T. Walton, G. K. Eddlemon, and K. G. Olson.
1979. Transport and fate of anthracene in aquatic microcosms.
pp. 312-320. IN A. W. Bourquin and P. H. Pritchard (eds.),
Microbial Degradation of Pollutants in Marine Environments.
EPA-600/9-79-012. Environmental Protection Agency, Washington,
D.C.
Gillett, J. W., and J. M. Witt (eds.). 1979. Proceedings of Symposium
on Terrestrial Microcosms and Environmental Chemistry. Report
prepared for National Science Foundation - Research Applied to
National Needs, Washington, D.C.
Gissel-Nielsen, G., and B. Bisbjerg. 1970. The uptake of applied
selenium by agricultural plants. Plant Soil 32(2):382-396.
-------
341
Gray, T. R. G., and S. T. Williams. 1971. Microbial productivity in
soil. pp. 255-287. IN Symposium of the Society for General
Microbiology, XXI. Microbes and Biological Productivity.
Cambridge University Press, Cambridge, England.
Hayes, F. R., J. A. McCarter, M. L. Cameron, and D. A. Livingston.
1952. On the kinetics of phosphorus exchange in lakes. J. Ecol.
40:202-216.
Herbes, S. E., G. R. Southworth, and C. W. Gehrs. 1976. Organic
contaminants in aqueous coal conversion effluents: environmental
consequences and research priorities, pp. 295-303. IN D. D.
Hemphill (ed.), Trace Substances in Environmental Health.
University of Missouri, Columbia.
Herbes, S. E., L. R. Schwall, and G. A. Williams. 1977. Rate of
microbial transformation of polycyclic aromatic hydrocarbons: A
chromatographic procedure. Appl. Environ. Microbiol. 32:244-246.
Herbes, S. E., and L. R. Schwall. 1978. Microbial transformation of
polycyclic aromatic hydrocarbons in pristine and petroleum-
contaminated sediments. Appl. Environ. Microbiol. 35(2):306-316.
Hildebrand, S. G., R. M. Cushman, and J. A. Carter. 1977. The poten-
tial toxicity and bioaccumulation in aquatic systems of trace
elements present in coal conversion effluents, pp. 305-313. IN
D. D. Hemphill (ed.), Trace Substances in Environmental Health.
University of Missouri Press, Columbia, Missouri.
Huff, D. D., R. J. Luxmoore, J. B. Mankin, and C. L. Begovich. 1977.
TEHM: A terrestrial ecosystem hydrology model. ORNL/NSF/EATC-27.
Oak Ridge National Laboratory, Oak Ridge, Tennessee. 152 pp.
-------
342
Isensee, A. R. 1976. Variability of aquatic model ecosystem-derived
data. Int. J. Environ. Studies 10:35-41.
Isensee, A. R., P. C. Kearney, E. A. Woolson, 6. E. Jones, and
V. P. Williams. 1973. Distribution of alkyl arsenicals in model
ecosystem. Environ. Sci. Technol. 7:841-845.
Jackson, D. R. Ecology and Ecosystem Analysis Section, Battelle
Columbus Laboratories, Columbus, Ohio (personal communication).
Jackson, D. R., and J. M. Hall. 1978. Extraction of nutrients from
intact soil cores to assess the impact of chemical toxicants on
soil. Pedobiologia Bd. 18, S:272-278.
Jackson, D. R., and M. Levin. 1979. Transport of arsenic in grassland
microcosms and field plots. Water Air Soil Pollut. 11:3-12.
Jackson, D. R., W. J. Selvidge, and B. S. Ausmus. 1978. Behavior of
heavy metals in forest microcosms. I. Transport and disruption
among Components. II. Effects on nutrient cycling processes.
Water Air Soil Pollut. 10:3-11.
Jackson, D. R., K. Washburne, and B. S. Ausmus. Loss of Ca and NO,-N
from terrestrial microcosms as an indicator of soil pollution.
Unpublished Manuscript.
Jackson, D. R., and A. P. Watson. 1977. Disruption of nutrient pools
and transport of heavy metals in a forested watershed near a lead
smelter. J. Environ. Qua!. 6:331-338.
Jones, P. C. T., and J. E. Mollison. 1948. A technique for the
quantitative estimation of soil micro-organisms. J. Gen.
Microbiol. 2:54-69.
-------
343
Jordan, M., and G. E. Likens. 1975. An organic carbon budget for an
oligotrophic lake in New Hampshire, U.S.A. Verh. Int. Verein.
Limnol. 19:994-1003.
Kaushik, N. K., and H. B. N. Hynes. 1971. The fate of the dead leaves
that fall into streams. Arch. Hydrobiol. 68:465-515.
Leo, A. J. 1975. Calculation of partition coefficients useful in the
evaluation of the relative hazards of various chemicals in the
environment. IN G. D. Veith and D. E. Konascwich (eds.), Symposium
on Structure-Activity Correlations in Studies of Toxicity and
Bioconcentration with Aquatic Organisms. International Joint
Commission, Burlington, Ontario.
Likens, G. E., F. H. Bormann, N. M. Johnson, D. W. Fisher, and R. S.
Pierce. 1970. Effects of cutting and herbicide treatment on
nutrient budgets in the Hubbard Brook Watershed ecosystem. Ecol.
Monogr. 40:23-47.
Lu, P-Y, R. L. Metcalf, and E. M. Carlson. 1978. Environmental fate of
five radiolabelled coal conversion by-products evaluated in a
laboratory model ecosystem. Environ. Health Perspect. 24:201-208.
Luxmoore, R. J., C. L. Begovich, and K. R. Dixon. 1976. DRYADS and
DIFMAS. FORTRAN models for investigating solute uptake and
incorporation into vegetation and litter. ORNL/NSF/EATC-26. Oak
Ridge National Laboratory, Oak Ridge, Tennessee. 133 pp.
Luxmoore, R. J., C. L. Begovich, and K. R. Dixon. 1978a. Modelling
solute uptake and incorporation into vegetation and litter. Ecol.
Model. 5:137-171.
-------
344
Luxmoore, R. J., D. D. Huff, R. K. McConathy, and B. E. Dinger. 1978b.
Some measured and simulated plant water relations of yellow
poplar. For. Sci. 24(3):327-341.
Luxmoore, R. J., and C. L. Begovich. Simulated heavy metal fluxes in
tree microcosms and a deciduous forest. Int. Soc. Ecol. Model. J.
(in press).
MacArthur, R. H. 1955. Fluctuations of annual populations and a
measure of community stability. Ecology 36:533-536.
Mackay, D., and W. Y. Shin. 1977. Aqueous solubility of polynuclear
aromatic hydrocarbons. J. Chem. Eng. Data 22:399-402.
May, R. M. 1973. Stability and Complexity in Model Ecosystems.
Princeton University Press, Princeton, New Jersey.
McConnell, W. J. 1962. Productivity relations in carbon microcosms.
Limnol. Oceanogr. 7:335-343.
Mclntire, C. D., R. L. Garrison, H. K. Phinney, and C. E. Warren. 1964.
Primary production in laboratory streams. Limnol. Oceanogr.
9:92-102.
Metcalf, R. L., I. P. Kapoor, Po-Yung, Lu, C. K. Schuth, and P. Sherman.
1973. Model ecosystem studies of the environmental fate of six
organochlorine pesticides. Environ. Health Perspect. 35-44, June.
Metcalf, R. L., G. K. Sangha, and I. P. Kapoor. 1971. Model ecosystem
for the evaluation of pesticide biodegradability and ecological
magnification. Environ. Sci. Technol. 5:709-713.
Metcalf, R. L. 1977. Model ecosystem approach to insecticide
degradation: A critique. Ann. Rev. Entomol. 22:241-261.
-------
345
Munro, J. K., R. J. Luxmoore, C. L. Begovich, K. R. Dixon, A. P. Watson,
M. R. Patterson, and D. R. Jackson. 1976. Application of the
unified transport model to the movement of Pb, Cd, In, Cu, and S
through the Crooked Creek Watershed. ORNL/NSF/EATC-28. Oak Ridge
National Laboratory, Oak Ridge, Tennessee. 92 pp.
National Academy of Sciences (NAS). 1974. Chromium. National Academy
of Sciences, Washington, D.C. 155 pp.
National Academy of Sciences (NAS). 1976. Selenium. Committee on
Medical and Biologic Effects of Environmental Pollutants.
Division of Medical Sciences, National Research Council,
Washington, D.C.
National Academy of Sciences (NAS). 1977. Arsenic. National Academy
of Sciences, Washington, D.C. 332 pp.
Odum, E. P. 1969. The strategy of ecosystem development. Science
164:262-270.
Odum, H. T. 1956. Primary production in flowing waters. Limnol.
Oceanogr. 1:102-117.
Odum, H. T. 1957. Trophic structure and productivity of Silver
Springs, Florida. Ecol. Monogr. 27:55-112.
Odum, H. T., and C. M. Hoskin. 1958. Comparative studies on the
metabolism of marine waters. Publ. Inst. Mar. Sci., Univ. Tex.
5:16-46.
O'Neill, R. V. 1976. Ecosystem persistence and heterotrophic
regulation. Ecology 57:1244-1253.
O'Neill, E. G.s B. S. Ausmus, and R. V. O'Neill. The effect of
substrate on elemental transport through terrestrial microcosms.
Unpublished Manuscript.
-------
346
O'Neill, R. V., W. F. Harris, B. S. Ausmus, and D. E. Reichle. 1975.
A theoretical basis for ecosystem analysis with particular
reference to element cycling, pp. 28-40. IN F. 6. Howell,
J. B. Gentry, and M. H. Smith (eds.), Mineral Cycling in
Southeastern Ecosystems. CONF-750513. Technical Information
Center, Oak Ridge, Tennessee.
O'Neill, R. V., B. S. Ausmus, D. R. Jackson, R. I. Van Hook,
P. Van Voris, K. Washburne, and A. P. Watson. 1977. Monitoring
terrestrial ecosystems by analysis of nutrient export. Water Air
Soil Pollut. 8:271-277.
Parkinson, D., T. R. Gray, and S. T. Williams. 1971. Methods for
Studying the Ecology of Soil Micro-Organisms. Blackwell
Scientific Publishers, Oxford, England. 116 pp.
Penrose, W. R., R. Black, and M. J. Hayward. 1975. Limited arsenic
dispersion in sea water, sediments, and biota near a continuous
source. J. Fish. Res. Board Can. 32:1275-1281.
Peters, L. N., D. F. Grigal, J. W. Curlin, and W. J. Selvidge. 1970.
Walker Branch Watershed Project: Chemical, physical, and
morphological properties of the soils of Walker Branch Watershed.
ORNL/TM-2968. Oak Ridge National Laboratory, Oak Ridge,
Tennessee. 96 pp.
Reichle, D. E., R. V. O'Neill, and W. F. Harris. 1975. Principles of
energy and material exchange in ecosystems, pp. 27-43. IN W. H.
Van Dobben and R. H. Lowe-McConnell (eds.), Unifying Concepts in
Ecology. W. Junk, the Hague. 302 pp.
-------
347
Riley, G. A. 1956. Oceanography of Long Island Sound, 1952-1954. IX.
Production and utilization of organic matter. Bull. Bingham
Oceanogr. Coll. 15:324-344.
Ritchie, J. A. 1961. Arsenic and antimony in some New Zealand thermal
waters. N. Z. J. Sci. 4:218-229.
Ruttner, F. 1963. Fundamentals of Limnology, 3rd ed. Toronto:
University of Toronto Press, Toronto, Canada. 295 pp.
Seydel, S. 1972. Distribution and circulation of arsenic through
water, organisms and sediments of Lake Michigan. Arch. Hydrobiol.
71:17-30.
Shugart, H. H., D. E. Reichle, N. T. Edwards, and J. R. Kercher.
1976. A model of calcium-cycling in an East Tennessee
Liriodendron forest: Model structure, parameters, and frequency
response analysis. Ecology 57:99-109.
Sohacki, L. P. 1968. Dynamics of arsenic in the aquatic environment.
Ph.D. thesis, Michigan State University, East Lansing, Michigan.
Soil Science Society of America. 1967. Soil Testing and Plant
Analysis, Part 1. Special Publication No. 2. Soil Science Society
of America, Madison, Wisconsin.
Southworth, 6. R. 1977. Transport and transformations of anthracene in
natural waters: Process rate studies. CONF-7710120-1. National
Technical Information Service, Springfield, Virginia.
Strange, R, J. 1976. Nutrient release and community metabolism
following application of herbicide to macrophytes in microcosms.
J. Appl. Ecol. 13(3):889-897.
-------
348
Sweeton, F. H. ORNL; unpublished distribution coefficients (K.) for
Walker Branch Watershed soils.
Toxic Substances Control Act of 1976 (TSCA): House of Representatives
14032, Senate 3149.
Trappe, J. U., E. A. Stahly, N. R. Benson, and P. U. Duff. 1973.
Mycorrhizal deficiency of apple trees in high arsenic soil.
BioScience 8:52-53.
Triska, F. J., J. R. Sedell, and B. Buckley. 1975. The processing of
conifer and hardwood leaves in two coniferous forest streams: II.
Biochemical and nutrient changes. Verh. Int. Verein. Limnol.
19:1628-1639.
Van Hook, R. I., B. S. Ausmus, S. Draggan, J. M. Giddings, D. R.
Jackson, and M. Witkamp. 1976. Evaluation of microcosms as
potential tools for estimating environmental transport of toxic
materials. Progress Report to the Environmental Protection
Agency, Athens, Georgia (mimeo).
Van Voris, P., R. V. O'Neill, H. H. Shugart, and W. R. Emanuel. 1978.
Functional complexity and ecosystem stability: An experimental
approach. ORNL/TM-6199. Oak Ridge National Laboratory, Oak
Ridge, Tennessee.
Walter, M. T., and H. E. Johnson. 1977. A model system to study the
desorption and biological availability of PCB in hydrosoils.
pp. 178-195. IN F. L. Mayer and J. L. Hamelink (eds.), Aquatic
Toxicology and Hazard Evaluation. ASTM STP 634. American Society
for Testing Materials.
-------
349
Warren, C. E., and W. J. Liss. 1977. Design and evaluation of
laboratory ecological system studies. EPA-600/3-77-022.
Environmental Protection Agency, Washington, D.C.
Watson, A. P., R. I. Van Hook, D. R. Jackson, and D. E. Reichle. 1976.
Impact of a lead mining-smelting complex on the forest floor litter
arthropod fauna in the New Lead Belt Region of southeast Missouri.
ORNL/NSF/EATC-30. Oak Ridge National Laboratory, Oak Ridge,
Tennessee. 163 pp.
Wetzel, R. G. 1964. Primary productivity of aquatic macrophytes.
Verh. Int. Verein. Limnol. 15:426-436.
Whittaker, R. H. 1961. Experiments with radiophosphorus tracer in
aquarium microcosms. Ecol. Monogr. 31:157-188.
Whitworth, W. R., and T. H. Lane. 1969. Effects of toxicants on
community metabolism in pools. Limnol. Oceanogr. 14:53-58.
Witherspoon, J. P., E. A. Bondietti, S. Draggan, F. B. Taub, N. Pearson,
and 0. R. Trabalka. 1976. State-of-the-art and proposed testing
for environmental transport of toxic substances. ORNL/EPA-1. Oak
Ridge National Laboratory, Oak Ridge, Tennessee.
-------
352
EXPERIMENT 1. MICROCOSM SIZE AND SUBSTRATE TYPE; ARSENIC TRANSPORT
Objectives
(1) To establish basic microcosm techniques (set-up, sampling, and
monitoring procedures).
(2) To identify parameters of interest for the measurement of
ecosystem properties, such as dissolved oxygen, water chemistry, and phi.
(3) To determine the effect of size [7-liter (1.85 gal) vs
70-liter (18.5 gal)] on ecosystem properties, microcosm stability and
persistence, and replicability.
(4) To determine the effect of sediment type (sand vs lake
sediment) on ecosystem properties, microcosm stability and persistence,
and replicability.
(5) To measure the accumulation of As (in tracer amounts) in
various ecosystem components, and to assess effects of size and
sediment type on arsenic transport.
Timetable
January 12-February 19, 1976 Microcosms assembled.
February 19 Began biological and chemical
sampling, and routine monitoring of
pH, temperature, conductivity, and
dissolved oxygen.
February 23 One 70-liter sand microcosm burst
and was replaced.
-------
353
April 12 Sodium arsenate added to all
microcosms. Began monitoring
74
As in water.
May 18-19 Final sampling; termination of
experiment.
Microcosm Construction
Six 7-liter and six 70-liter microcosms were assembled.
Approximately 20 kg of sediment was collected from Watts Bar and Melton
Hill reservoirs (Tennessee), dried at 105°C in a forced-air drying oven,
and ground with a mortar and pestle. The same amount of Ottawa sand
was rinsed in distilled water and oven-dried. On January 12, 1976,
approximately 3.5 cm of sand or sediment was placed in each aquarium,
with three replicates of each size-substrate combination. Spring water
was then added slowly. A minor catastrophe occurred after three of the
large aquaria had been filled—the supporting rack in the environmental
chamber collapsed under the weight, and the completion of the microcosms
was delayed for four weeks while stronger tables were constructed. On
February 11, filling was resumed. The lake sediment tanks were murky
at first due to suspended clay particles, but gradually cleared over
the next seven days.
On February 19, filamentous algal communities were collected
from a concrete fish holding tank outside the Environmental Sciences
Laboratory. The algae were drained in a sieve; six 30-mg and six
6-mg portions were weighed out and placed in the 70-liter and 7-liter
microcosms, respectively. Ten large Physa (total live wt approximately
10 g) were placed in each large tank, and two in each small tank.
-------
354
On February 23, one of the large sand microcosms burst, and was
reconstructed following the same procedure.
Parameters Measured
PH
temperature
conductivity
dissolved oxygen
concentrations of major ions (Na, K, Ca, Mg, HCO,, Cl, SO.)
nutrient concentrations (N03-N, NH^-N, PO.-P, Total P, Fe)
dissolved organic carbon
final algal biomass
final chlorophyll (in water)
final sediment ATP
dominant organisms (qualitative)
arsenic-74 in water, substrate, algae, and snails
Further Information
For a description of sampling and radioassay techniques, see
Section 3.2.2.1. This section also presents results on arsenic
transport. Other results of Experiment 1 are discussed in Sections
3.2.1.3., 3.2.1.4., and 3.2.1.5. The experiment was the subject of a
paper presented at the 1976 AIBS Annual Meeting in New Orleans, and
later published in the open literature (Giddings and Eddlemon 1977).
-------
355
EXPERIMENT 2: DOSAGE EFFECTS—ARSENIC
Objectives
(1) To determine the effect of dosage level on the transport and
distribution of arsenic.
(2) To determine the effects of three arsenic dosages on
biological and chemical parameters in microcosms.
(3) To repeat a portion of Experiment 1 for assessment of the
reproducibility of microcosm results.
Timetable
June 29-July 6, 1976 Twelve microcosms assembled.
July 6 Began biological and chemical
sampling, and routine monitoring of
pH, temperature, conductivity, and
dissolved oxygen.
July 8 One tank burst and was replaced.
August 18-19
August 25-26
24-hr production, respiration
experiments.
September 16-17
September 17 Sodium arsenate added to 9
microcosms (3 concentrations x 3
replicates; plus 3 controls).
74
September 20 Began monitoring As in water
and sediment.
September 16-17 Began weekly production and
respiration experiments.
-------
356
November 7-12 Final sampling.
November 24 Experiment terminated.
Microcosm Construction
Sediment was shoveled from the bottom of a shallow pond (mean
depth approximately 1 m) behind the ORNL Aquatic Ecology Laboratory, on
June 29, 1976, and carried into the environmental chamber in tubs. On
July 1, an 8-cm layer of sediment was placed into each of 12 70-liter
aquaria. Spring water was added to three of the tanks on July 1, and
to the other nine on July 6. A mixed Elodea-Potamogeton community was
collected from the pond, drained, and 40-g aliquots placed in each
microcosm on July 6. As in Experiment 1, one of the aquaria burst a
few days after the experiment began, and was replaced.
Parameters Measured
PH
temperature
conductivity
dissolved oxygen
concentrations of major ions (see Experiment 1)
concentrations of nutrients (see Experiment 1)
dissolved organic carbon
net primary production
total ecosystem respiration
dominant organisms (qualitative)
arsenic-74 in water, sediment, and biota
-------
357
EXPERIMENT 3: POND ENCLOSURES
Objectives
(1) To compare the ecological characteristics of pond communities
enclosed in situ with conditions in the surrounding pond and in
microcosms.
(2) To obtain field data on arsenic transport for comparison with
Experiment 3.
Timetable
July 15, 1976 Six Plexiglas enclosures placed in
pond.
July 19 Began chemical sampling, and
routine monitoring of pH,
temperature, conductivity, and
dissolved oxygen.
September 17 Sodium arsenate added to 3
enclosures.
74
September 21 Began monitoring As in water.
October 26-28 Final sampling.
Description of Enclosures
Six enclosures were constructed out of Plexiglas. The enclosures
measured 60 x 120 cm and were 91.4 cm high. The corners were
reinforced with square Plexiglas rods, and two Plexiglas strips were
fastened between the long sides at the upper edge of each enclosure for
added stability. Each enclosure had two handles at each end.
-------
358
The enclosures were positioned in a shallow pond (the source of
the microcosm components for Experiment 2) by two men wearing waders.
The pond had an irregular rocky bottom, overlain by 5 to 10 cm of
sediment. Because of the uneven bottom, it was difficult to push the
enclosures evenly into the sediment, and many possible locations were
tried before six suitable sites were found. The enclosures extended
about 2 to 5 cm below the sediment surface and above the waterline.
Parameters Measured
PH
temperature
conductivity
dissolved oxygen
concentrations of major ions (see Experiment 1)
nutrient concentrations (see Experiment 1)
dissolved organic carbon
arsenic-74 in water, sediment, and biota
-------
359
EXPERIMENT 4: FIRST PROTOCOL TEST—CHROMIUM
Objectives
(1) To test our aquatic microcosm protocol with a second inorganic
contaminant.
(2) To evaluate the production/respiration ratio as an indicator
of damage to aquatic ecosystems.
(3) To continue development of techniques for studying trace
element transport in microcosms.
Timetable
January 12, 1977 Eight microcosms assembled.
January 19-20 Began monitoring production and
respiration with Dissolved Oxygen
Monitoring System.
January 25 Began routine monitoring of pH,
temperature, and conductivity.
February 9 Snails added to microcosms.
April 121
April 27
May 11
May 25
Sodium chromate added to 6
microcosms.
(3 concentrations x 2 replicates;
plus 2 controls).
April 12 Began monitoring chromium in water
and sediment.
June 9-10 Final sampling; experiment
terminated.
-------
360
Microcosm Construction
Sediment was collected from the Clinch River, just below Melton
Hill Dam, on January 12, 1977. Within hours, a 5-cm layer was placed
in each of eight 7-liter aquaria. The aquaria were then filled with
spring water. A 6.0-g (drained wet wt) aliquot of a filamentous algal
community from Target Range Pond was added to each microcosm. On
February 9, five adult snails (Physa) from a large fish culture tank
were placed into each microcosm (total live wt 0.77-0.85 g).
Parameters Measured
PH
temperature
conductivity
dissolved oxygen
net primary production
total ecosystem respiration
chromium in water, sediment, and biota
-------
361
EXPERIMENT 5: FISH MICROCOSMS
Objective
To investigate the consequences of adding fish to the standard
microcosm design.
Timetable
June 30, 1977
July 7
July 15
October 3
October 14
November 22
December 12
March 17, 1978
Eight microcosms assembled.
Fish added.
Began monitoring pH, temperature,
conductivity, anmd dissolved
oxygen; chemical and biological
sampling.
Bluegill and Gambusia microcosms
treated with Se (Experiment 6).
Began monitoring net primary
production and total ecosystem
respiration with dissolved Oxygen
Monitoring System.
Bluegill and Gambusia microcosms
terminated (Experiment 6).
Control microcosms treated with
?rn
Hg (Experiment 7).
Remaining microcosms terminated.
-------
362
Microcosm Construction
All microcosm components (except fish) were collected on June 30,
1977, from Target Range Pond, a shallow (1-m) spring-fed pond on the
Oak Ridge Reservation. Water was collected in large polyethylene
containers near the outlet of the pond. Sediment and an El odea
community were collected from the center of the pond. The sediment was
distributed among eight 70-liter aquaria; the sediment depth was 5 cm.
Pond water was then slowly siphoned into each aquarium, and 100 g
(drained wet wt) of the El odea community was added.
Fish were introduced on July 7. Gambusia were collected from a
stocked pond, and bluegill and fathead minnows were selected from
laboratory cultures. The fish additions were as follows:
Microcosm number Fish added Total wet wt
Ml 10 Gambusia 0.36 g
M2 " 0.63 g
Fl 5 fathead minnows 0.68 g
F2 " 0.68 g
Bl 2 bluegill 7.8 g
B2 " 8.1 g
Cl None (controls)
C2 "
Parameters Measured
PH
temperature
conductivity
dissolved oxygen
concentrations of major ions (see Experiment 1)
-------
363
nutrient concentrations (see Experiment 1)
dissolved organic carbon
net primary production
total ecosystem respiration
dominant organisms (qualitative)
-------
364
EXPERIMENT 6: SELENIUM TRANSPORT
Objectives
(1) To examine the movement of selenium in pond microcosms
containing fish.
(2) To increase the amount of information on contaminant transport
which can be obtained from a microcosm experiment.
Timetable
June 30, 1977 Microcosms assembled as part of
Experiment 5 (Microcosms Bl, B2,
Ml, and M2).
October 3 Bluegill and Gambusia microcosms
from Experiment 5 treated with
sodium selenate. Began monitoring
Se in water, sediment, Elodea,
and snails.
November 18-22 Final sampling.
November 22 Experiment terminated.
Microcosms
See Experiment 5.
Parameters Measured
selenium-75 in water, sediment, and biota
(see also Experiment 5)
-------
365
EXPERIMENT 7: MERCURY TRANSPORT
Objectives
(1) To examine the movement of mercury in pond microcosms.
(2) To compare the behavior of mercury with that of the other
trace elements tested (arsenic, chromium, and selenium).
Timetable
June 30, 1977 Microcosms assembled as part of
Experiment 5 (Microcosms Cl and C2),
December 12 Control microcosms from Experiment
5 treated with mercuric nitrate.
203
Began monitoring Hg in water,
sediment, Elodea, and snails.
January 18-25, 1978 Final sampling.
March 17 Experiment terminated.
Microcosm Construction
See Experiment 5.
Parameters Measured
mercury-203 in water, sediment, and biota
elimination of mercury by snails
(see also Experiment 5)
-------
366
EXPERIMENT 8: ANTHRACENE TRANSPORT AND FATE
Objectives
(1) To test the standard pond microcosm design with an organic
contaminant.
(2) To study the transport and chemical fate of anthracene in pond
microcosms.
(3) To compare microcosm results with data collected by
conventional contaminant testing methods.
Timetable
January 7-12, 1977 Two microcosms assembled
January 19-20 Began monitoring production and
respiration with Dissolved Oxygen
Monitoring System.
January 25 Began routine monitoring of pH,
temperature, and conductivity.
July 13 Anthracene added. Began monitoring
anthracene and derivatives in
water, sediment, and biota.
October 5-6 Final sampling.
October 6 Experiment terminated.
Microcosm Construction
Sediment and a filamentous algal community were collected from
Target Range Pond on January 7, 1977. On January 10, approximately
6 cm of sediment were placed in each of two 70-liter aquaria. The
-------
367
aquaria were then filled with spring water. Two days later, unweighed
portions of the algal community (approximately 10 g) were added to each
microcosm.
Parameters Measured
PH
temperature
conductivity
dissolved oxygen
net primary production
total ecosystem respiration
dominant organisms (qualitative)
anthracene and derivatives in water, sediment, and biota
-------
368
EXPERIMENT 9: SUCCESSION AND NUTRIENT CYCLING
Objective
To observe changes in community structure, ecosystem metabolism,
and nutrient cycling during the first two months following construction
of a pond microcosm.
Timetable
November 14, 1977 Two microcosms assembled. Began
biological and chemical sampling
and routine monitoring of pH,
temperature, and conductivity.
November 15 Began monitoring net primary
production and total ecosystem
respiration with the Dissolved
Oxygen Monitoring System.
March 17, 1978 Experiment terminated.
Microcosm Construction
Water, sediment, and an El odea community were collected from
Target Range Pond on November 14, 1977, as in Experiment 5. About 4 cm
D
of sediment was placed in each of two 70-liter aquaria. One Ami con
hollow fiber interstitial sampling unit was placed on the sediment in
each microcosm, and covered with an additional 2 cm of sediment. Pond
water was then slowly siphoned in. A portion of the Elodea community
(100 g drained wet wt) was added to each microcosm.
-------
369
Parameter Measured
PH
temperature
conductivity
dissolved oxygen
nutrient concentrations (see Experiment 1) in water and
interstitial water
net primary production
total ecosystem respiration
dominant organisms (qualitative)
-------
372
Microcosm Biota
Blue-greens
I.
II
III
Green algae
IV.
VI,
Chroococcales
1. Aphanocapsa
2. Coelosphaerium
3. Gomphosphaeria
4. Merismopedia
5. Chroococcus
6. Microcystis
Oscillatoriales
7. Oscillatoria (Lyngbya)
Nostocales
8. Anabaena
9. Cylindrospermum
Volvocales
10. Carteria
11. Chlamydomonas
12. Gonium
13. Pandorina
Tetrasporales
14. Gloeocystis
Chlorococcales
15. Planktosphaeria
16. Tetraedron
17. Palmella
18. Sphaerocystis
19. Ankistrodesmus
20. Chiorella
21. Nephrocytium
22. Oocystis
23. Selenastrum
24. Golenkinia
25. Actinastrum
26. Coelastrum
27. Scenedesmus
28. Pediastrum
VII. Ulotrichales
29. Geminella
30. Ulothrix
VIII. Chaetophorales
31. Aphanochaete
32. Coleochaete
IX. Oedogoniales
33. Oedogonium
-------
373
X. Cladophorales
34. Cladophora
35. Pithophora
36. Rhizoclonium
XI. Zygnematales
37. Mougeotia
38. Spirogyra
39. Zygnema
40. Closterium
41. Cosmarium
42. Pleurotaenium
43. Staurastrum
Euglenophytes
XII. Euglenales
44. Euglena
45. Phacus
46. Menoidium
Dinoflagellates
XIII. Dinokontae
47. Gymnodinium
Cryptophytes
XIV. Cryptomonadaceae (family)
48. Cryptomonas
Chrysophytes
XV. Ochromonadales
49. Dinobryon
50. Mallomonas
51. Synura
Diatoms
XVI. Centrales
52. Melosira
XVII. Pennales
53. Asterione!la
54. Fragilaria
55. Meridion
56. Synedra
57. label1 aria
58. Eunotia
59. Cocconeis
60. Gomphonema
61. Amphora
62. Epithemia
63. Rhopalodia
64. Unknown A
65. Unknown B
-------
374
PROTOZOA
XVIII. Sarcodina
66. Amoeba
67. Difflugia
68. Heliozoa
XIX. Mastigophora
69. Oikomonas
70. Other flagellates
XX. Infusoria
71. Coleps
72. Lacrymaria
73. Prorodon
74. Loxophyllum
75. Chilodon
76. Urocentrum
77. Colpidium
78. Uronema
79. Glaucoma
80. Paramecium
81. Lembadion
82. Cyclidium
83. Spirostomum
84. Stentor
85. Stichotricha
86. Uroleptus
87. Oxytricha
88. Stylonychia
89. Euplotes
90. Vorticella
91. Other ciliates
MACROFAUNA
XXI. Hydra
XXII. Turbellaria
XXIII. Nematoda
XXIV. Rotatoria
XXV. Gastrotricha
XXVI. Oligochaeta
XXVII. Cladocera
XXVIII. Copepoda
XXIX. Ostracoda
XXX. Hydrocarina
XXXI. Gastropoda
XXXII. Insecta
XXXIII. Vertebrata
-------
376
PROJECT PUBLICATIONS
Ausmus, B. S., D. R. Jackson, and P. Van Voris. 1977. The accuracy of
screening techniques, pp. 123-130. IN J. M. Witt, and J. W.
Gillett (eds.), Terrestrial Microcosms and Environmental Chemistry
Proceedings, June 13-14, 1977, Oregon State University, Corvallis,
Oregon.
Ausmus, B. S., G. J. Dodson, and D. R. Jackson. 1978. Behavior of
heavy metals in forest microcosms. III. Effects on litter-soil
carbon metabolism. Water Air Soil Pollut. 10:19-26.
Ausmus, B. S., and E. G. O'Neill. 1978. Comparison of carbon dynamics
of three microcosm substrates. Soil Biol. Biochem. 10:425-429.
Ausmus, B. S., S. Kimbrough, D. R. Jackson, and S. E. Lindberg. The
behavior of hexachlorobenzene in pine forest microcosms: Transport
and effects on soil processes. Environ. Pollut. (in press).
Begovich, C. L., and R. J. Luxmoore. 1979. Some sensitivity studies
of chemical transport simulated in models of soil-plant-litter
system. ORNL/TM-6791. Oak Ridge National Laboratory, Oak Ridge,
Tennessee.
Draggan, S., and J. M. Giddings. 1978. Testing toxic substances for
protection of the environment. Sci. Total Environ. 9:63-74.
Giddings, J. M., and G. K. Eddlemon. 1977. The effects of microcosm
size and substrate type on aquatic microcosm behavior and arsenic
transport. Arch. Environ. Contam. Toxicol. 6:491-505.
Giddings, J. M., and G. K. Eddlemon. 1978. Photosynthesis/respiration
ratios in aquatic microcosms under arsenic stress. Water Air Soil
Pollut. 9:207-212.
-------
377
Giddings, J. M. 1978. Types of aquatic microcosms and their research
applications. Symposium: Microcosms in Ecological Research,
Augusta, Georgia, November 8-10, 1978, Savannah River Ecology
Laboratory. 32 pp.
Giddings, J. M., and G. K. Eddlemon. 1979. Some ecological and
experimental properties of complex aquatic microcosms. Int. J.
Environ. Stud. 13:119-123.
Jackson, D. R., W. J. Selvidge, and B. S. Ausmus. 1978. Behavior of
heavy metals in forest microcosms. I. Transport and distribution
among components. Water Air Soil Pollut. 10:3-11.
Jackson, D. R., W. J. Selvidge, and B. S. Ausmus. 1978. Behavior of
heavy metals in forest microcosms. II. Effects on nutrient
cycling processes. Water Air Soil Pollut. 10:13-18.
Jackson, D. R., and J. M. Hall. 1978. Extraction of nutrients from
intact soil cores to assess the impact of chemical toxicants on
soil. Pedobiologia Bd. 18, S:272-278.
Jackson, D. R., and B. S. Ausmus. 1979. Effects of arsenic on nutrient
dynamics of grassland microcosms and field plots. Water Air Soil
Pollut. 11:13-21.
Jackson, D. R., and M. Levin. 1979. Transport of arsenic in grassland
microcosms and field plots. Water Air Soil Pollut. 11:3-12.
Levin, M. J., and D. R. Jackson. 1977. A comparison of in situ
extractors for sampling soil water. Soil Sci. Soc. Am. J.
41 (3): 535-536.
-------
378
Luxmoore, R. J., and C. L. Begovich. 1979. Simulated heavy metal
fluxes in tree microcosms and a deciduous forest. Int. Soc. Ecol.
Model. J. (in press).
O'Neill, R. V., B. S. Ausmus, D. R. Jackson, R. I. Van Hook, P. Van
Voris, C. Washburne, and A. P. Watson. 1977. Monitoring
terrestrial ecosystems by analysis of nutrient export. Water Air
Soil Pollut. 8:271-277.
Van Voris, P., R. V. O'Neill, W. R. Emanuel, and H. H. Shugart, Jr.
Functional complexity and ecosystem stability. (Submitted to
Ecology).
-------
379
ORNL/EPA-4
EPA-600/3-80-042
INTERNAL DISTRIBUTION
1-8.
9.
10-12.
13.
14-33.
34-36.
37.
38.
39.
40.
S.
G.
J.
A.
W.
R.
E.
R.
D.
B.
I.
K.
M.
S.
F.
0.
G.
V.
E.
M.
Auerbach
Eddlemon
Giddings
Harmons
Harris
Luxmoore
O'Neill
O'Neill
Reichle
Ross-Todd
41. 0. S. Shriner
42. W. Van Winkle
43. J. B. Waide
44. ESD Library
45-46. Central Research Library
47-48. Laboratory Records Department
49. Laboratory Records, ORNL-RC
50. ORNL Y-12 Technical Library
51. ORNL Patent Office
EXTERNAL DISTRIBUTION
52. B. S. Ausmus, Ecology and Ecosystems Analysis Section,
Biological, Ecological and Medical Department, Battelle
Columbus Laboratories, 505 King Ave., Columbus, OH 43201
53. W. R. Bibb, Research Support Programs Branch, DOE-ORO,
Oak Ridge, TN 37830
54-553. Donald L. Brockway, Project Officer, Environmental Research
Laboratory, U.S. Environmental Protection Agency,
Athens, GA 30605
554. D. C. Coleman, Natural Resource Ecology Laboratory, Colorado
State University, Fort Collins, CO 80521
555. J. W. Gillette, Terrestrial Toxics and Hazardous Materials
Branch, U.S. Environmental Protection Agency, Corvallis
Environmental Research Laboratory, 200 SW 35th Street,
Corvallis, OR 97330
556. D. R. Jackson, Ecology and Ecosystems Analysis Section,
Battelle Columbus Laboratories, 505 King Ave.,
Columbus, OH 43201
557. P. Van Voris, Ecology and Ecosystems Analysis Section,
Battelle Columbus Laboratories, 505 King Ave.,
Columbus, OH 43201
558. Zillioux, Division of Environmental Review, Office of Toxic
Substances, U.S. Environmental Protection Agency,
401 M St., SW, Washington, DC 20460
559. Office of Assistant Manager, Energy Research and Development
DOE-ORO
560-586. Technical Information Center, Oak Ridge, TN 37830
587-677. Special EPA Distribution
------- |