A Primer on Trace Metal-

Sediment Chemistry

United States
Geological
Survey
Water-Supply
Paper 2277
A Primer on Trace Metal
Sediment Chemistry

By ARTHUR J. HOROWITZ

U.S. GEOLOGICAL SURVEY WATER-SUPPLY PAPER 2277

DEPARTMENT OF THE INTERIOR
DONALD PAUL MODEL, Secretary

U.S. GEOLOGICAL SURVEY

Dallas L. Peck, Director

UNITED STATES GOVERNMENT PRINTING OFFICE: 1985

For sale by the Distribution Branch, U.S. Geological Survey,

604 South Pickett Street, Alexandria, VA 22304

Library of Congress Cataloging-in-Publication Data

Horowitz, Arthur J.
A primer on trace metal-sediment chemistry.
(United States Geological Survey water-supply paper ; 84-
2277)
Bibliography: p. 61
Supt. of Docs, no.: I. 19.13:2277
1. Trace elements. 2. Metals. 3. Geochemistry. 4. Sediments
(Geology) I. Title. II. Series.
U.S. Geological Survey water-supply paper ; 2277.
QE516.T85H67 1985 551.9 84-600347
 PREFACE
This Primer was originally developed as a companion volume to a short course on
trace metal-sediment chemistry. The purpose of the course was to expose the participants
to the basic principles and techniques that control and are used to elucidate trace metal
sediment interactions. Participants were not expected to walk away from the course as
experts in sediment chemistry; however, they were expected to have a good knowledge
of the basic principles governing this complex subject.
This Primer, organized in the same way as the instructional session, contains all the
diagrams and tables used for the course, along with a descriptive text. It also includes a
large list of selected references on the subject of trace metal-sediment chemistry. This list
is by no means exhaustive. Many of the references were used in developing the material
in the course and in the Primer. The other references are provided for information on
techniques and methods, and as examples of how sediment chemistry can be used to deal
with various types of environmental studies.

Ill
CONTENTS
Preface iii
Abstract 1
Introduction 1
Glossary 3
1.0 Importance of sediments to aquatic trace metal chemistry
1.1 Monitoring studies 4
1.2 Comparison of dissolved and suspended trace metal concentrations 6
1.3 Riverine transport of trace metals by suspended sediments 8
1.4 Comparison of trace metal concentrations in suspended and bottom sediments
versus dissolved levels 10
1.5 Historical records 12
2.0 Physical factors affecting sediment-trace metal concentrations
2.1 Introduction 14
2.2 Grain-size ranges and effect of grain size 14
2.3 Chemical analysis of various grain sizes in bottom sediments 16
2.4 Chemical analysis of various grain sizes in suspended sediments 18
2.5 Effect of grain size on trace metal concentrations in samples collected from the
same basin 20
2.6 Comparison of samples having similar bulk chemistries but differing grain-
size distributions 22
2.7 Effect of grain size on sediment-associated chemical transport at differing
discharge rates 24
2.8 Effect of surface area 26
2.9 Importance of surface area to sediment-trace metal concentrations 28
2.10 Cation exchange capacity 30
2.11 Significant sedimentary trace metal collectors 32
2.12 Iron and manganese oxides 32
2.13 Organic matter 34
2.14 Clay minerals 34
3.0 Chemical factors affecting sediment-trace metal concentrations
3.1 Introduction to chemical partitioning 36
3.2 Utility of chemical partitioning data 38
3.3 Chemical partitioning methods 38
3.4 Chemical partitioning instrumental methods 40
3.5 Chemical partitioning partial extraction methods 42
3.6 Chemical partitioning of suspended sediments by partial extraction 44
3.7 Chemical partitioning of bottom sediments by partial extraction 46
3.8 Chemical partitioning by density gradient and mineralogy 48
3.9 Chemical partitioning by mathematical manipulation or modeling 50
4.0 Sediment-trace metal data manipulations
4.1 Corrections for grain-size differences 52
4.2 Carbonate corrections 54
4.3 Normalization to "conservative" elements 58
5.0 Sampling and data handling general considerations 60
Selected references 61
Metric conversion factors 67
A PRIMER ON TRACE METAL-
SEDIMENT CHEMISTRY

By Arthur J. Horowitz

Abstract INTRODUCTION

In most aquatic systems, concentrations of trace The basic goal of most chemically oriented
metals in suspended sediment and the top few centime- water-quality studies is to describe or evaluate exist-
ters of bottom sediment are far greater than concentra- ing environmental conditions and to attempt to iden-
tions of trace metals dissolved in the water column. Con- tify the source or sources of the constituents under
sequently, the distribution, transport, and availability of investigation. An equally important goal is to attempt
these constituents can not be intelligently evaluated, nor to predict or determine potential impacts. This head-
can their environmental impact be determined or pre- ing could accommodate such subjects as bioavailabil-
dicted solely through the sampling and analysis of dis- ity, amount of constituent transport, location of chem-
solved phases. This Primer is designed to acquaint the
ical sinks, ultimate fate, and potential toxic effects.
reader with the basic principles that govern the concen-
tration and distribution of trace metals associated with
Historically, the U.S. Geological Survey has
bottom and suspended sediments. attempted to assess trace metals in aquatic systems by
The sampling and analysis of suspended and bot- analyzing water samples. This assessment has en-
tom sediments are very important for monitoring studies, tailed determining concentrations of total and dis-
not only because trace metal concentrations associated solved elements and compounds through the collec-
with them are orders of magnitude higher than in the tion and analysis, respectively, of unfiltered and
dissolved phase, but also because of several other fac- filtered water. Concentrations associated with sus-
tors. Riverine transport of trace metals is dominated by pended sediment (particulates, seston) are determined
sediment. In addition, bottom sediments serve as a indirectly by the difference between total and dis-
source for suspended sediment and can provide a histor- solved concentrations. It is recognized that this ap-
ical record of chemical conditions. This record will help
proach casts doubt on the reliability of reported
establish area baseline metal levels against which exist-
ing conditions can be compared.
suspended-sediment chemical analyses. As a result,
Many physical and chemical factors affect a sedi-
water quality tends to be evaluated on the kinds and
ment's capacity to collect and concentrate trace metals. concentration of various constituents found in solution
The physical factors include grain size, surface area, (Feltz, 1980). However, in most aquatic systems, the
surface charge, cation exchange capacity, composition, concentration of trace metals in suspended sediment
and so forth. Increases in metal concentrations are and the top few centimeters of bottom sediment is far
strongly correlated with decreasing grain size and in- greater than the concentration of trace metals dis-
creasing surface area, surface charge, cation exchange solved in the water column. The strong association of
capacity, and increasing concentrations of iron and numerous trace metals (for example, As, Cd, Hg, Pb,
manganese oxides, organic matter, and clay minerals. Zn) with seston and bottom sediments means that the
Chemical factors are equally important, especially for
distribution, transport, and availability of these con-
differentiating between samples having similar bulk
chemistries and for inferring or predicting environmental
stituents can not be intelligently evaluated solely
availability. Chemical factors entail phase associations through the sampling and analysis of the dissolved
(with such sedimentary components as interstitial water, phase.
sulfides, carbonates, and organic matter) and ways in Additionally, because bottom sediments can act
which the metals are entrained by the sediments (such as as a reservoir for many trace metals, they must, for
adsorption, complexation, and within mineral lattices). several reasons, be given serious consideration in the
planning and design of any water-quality study. First, not the major, source of suspended sediment. There-
an undisturbed sediment sink contains a historical fore, they must be investigated to determine transport
record of chemical conditions. If a sufficiently large potential. Under changing hydrologic conditions
and stable sink can be found and studied, it will allow (such as a heavy storm or spring runoff), a localized
the investigator to study changes over time and, possi- pollution problem can suddenly become widespread
bly, to establish area baseline levels against which and result in significant environmental impact.
existing conditions can be compared and contrasted. The foregoing discussion indicates that data on
Second, under changing environmental or physico- suspended and bottom sediments, as well as on the
chemical conditions (like pH, Eh, dissolved oxygen, dissolved phase, are a requisite for the development of
bacterial action), sediment-bound trace metals can a comprehensive understanding of the impact of trace
dissolve into the water column, possibly enter the metals on water quality. Through the use of such
food chain, and have a significant environmental im- additional data, it may be possible to begin to identify
pact. Third, several relatively inert or otherwise envi- sources and sinks and the fate and potential effects of
ronmentally harmless inorganic constituents can de- toxic or environmentally necessary metals. Similarly,
grade, or react with others, to form soluble and sediment-chemical data are a requisite for transport
potentially toxic forms (for example, the conversion modeling, for estimating geochemical cycles, and for
of elemental mercury to methyl-mercury). Finally, inferring the availability of various trace metals in an
bottom sediments should be regarded as a major, if ecological system.
GLOSSARY

Sediment Particles derived from rocks or biological materials that have been transported
by a fluid, or solid material suspended in or settled from water.
Suspended sediment, seston, particulates Material actively transported by a fluid.
Bottom sediment, bed material Material that is temporarily stationary in the bottom of
a water course.
Dissolved load, dissolved constituent Operationally defined as that material that passes
through a 0.45-jxm (micrometer) filter.
Suspended load, suspended constituent Operationally defined as that material that is
retained by a 0.45-|xm (micrometer) filter during the filtration of a whole water
sample.
Trace metal, heavy metal, metal Terms used interchangeably throughout this Primer.
Total concentration Refers to the true concentration of metal in a sediment sample as
determined by the complete dissolution of the sample. Unless otherwise stated,
concentrations in this Primer are totals.
Total recoverable concentration Refers to the concentration of metal in a sediment
sample that has been extracted with a solution that does not completely dissolve the
sediment. In other words, this is an operationally defined concentration, the defini-
tion being dependent upon the extraction solution.
1.0 IMPORTANCE OF SEDIMENTS TO AQUATIC TRACE METAL CHEMISTRY

/. / Monitoring Studies

Table 1.1-1 has been extracted from an article stituent on the basis of its chemical behavior and bio-
by Chapman and others (1982) dealing with the design logical significance. These factors, in turn, determine
of monitoring studies for priority pollutants. The au- what type of sample(s) should be collected (water,
thors point out that, to design an effective monitoring sediment, biota). All the inorganic priority pollutants,
program, an investigator must decide two basic ques- with the exception of antimony, are persistent and can
tions: what samples should be collected, and for what bioaccumulate, and all tend to be found in either sed-
should they be analyzed? This decision requires an iment (bottom and suspended) or biota (table 1.1-1).
understanding of the relative importance of each con-
Table 1.1-1. Recommendations for types of environ-
mental sampling for monitoring purposes
[Data from Chapman and others, 1982]

Category
Metal rank* Water Sediment Biota
Antimony 3 X
Arsenic 1 X X
Beryllium 1 X X
Cadmium 1 X X
Chromium 1 X X
Copper 1 X X
Lead 1 X X
Mercury 1 X X
Selenium 1 X X
Silver 1 X X
Thallium 1 X X
Zinc 1 X X
*Category rank: 1 and 2 - persistent and may bioaccumulate
3 and 4 - persistent and nonaccumulative
5 - nonpersistent
1.0 IMPORTANCE OF SEDIMENTS TO AQUATIC TRACE METAL CHEMISTRY Continued

7 .2 Comparison of Dissolved and Suspended Trace

Metal Concentrations

As stated previously, both bottom and sus- upon the analysis of dissolved and suspended loads for
pended sediments contain significantly more trace selected trace metals in the Amazon and Yukon Rivers
metals than are found in the dissolved phase. Hence, and graphically displays this phenomenon (adapted
riverine transport of these constituents can be domi- from Gibbs, 1977).
nated by suspended sediment. Figure 1.2-1 is based
 Chromium Manganese

Iron Cobalt

Nickel Copper

AMAZON RIVER YUKON RIVER AMAZON RIVER YUKON RIVER

EXPLANATION
Dark areas: percentage in suspended sediments
Light areas: percentage in solution

Figure 1.2-1 . Transport phases of selected trace metals in rivers. (Data from Gibbs, 1977)
1.0 IMPORTANCE OF SEDIMENTS TO AQUATIC TRACE METAL CHEMISTRY Continued

1.3 Riverine Transport of Trace Metals by Suspended

Sediments

As a further indication that riverine transport of suspended sediment can involve significant weights of
metals may be dominated by the suspended-sediment constituent transport. For example, the arsenic con-
load, it is worthwhile to determine, in relatively sim- centration in Willamette River suspended sediment is
ple terms, what order of magnitude of material is only 0.002 mg/kg (2 ppb), and yet this means that
being transported. Examine the data in table 1.3-1. 0.9 t/d of As is transported by the river. At the other
The discharge equation used to generate the data end of the spectrum, examine the iron data for the
comes from Porterfield (1972), and the basic data for Mississippi River. Suspended-sediment iron concen-
the actual calculations come from the annual reports trations are 46 mg/kg (46 ppm); thus some 75,000 t/d
of Pennsylvania, Oregon, and Louisiana. Note that of Fe is transported.
even relatively low concentrations of trace metals in
Table 1.3-1. Transport of metals associated with suspended sediment in selected rivers

Element Susquehanna River 03/08/79 Willamette River 01/16/80 Mississippi River 01/18/81

Discharge 1 346,000 189,000 671,000

Concentration 2 Transport3 Concentration 2 Transport3 Concentration 2 Transport*

Sediment4 406 344,200 106 49,000 641 1,050,000

Iron 12 10,200 4.4 2,040 46 75,600

Manganese 0.590 430 0.100 46 1.3 2,150

Zinc 0.070 59 0.030 14 0.090 148

Lead 0.018 15 0.005 2.3 0.030 49

Copper 0.015 13 0.006 2.8 0.018 30

Cobalt 0.010 8.5 - 0.005 8.2

Chromium 0.003 6.8 0.010 4.6 0.028 46

Nickel 0.004 1.9 0.035 58

Arsenic 0.002 0.9 0.012 20

'Discharge, in cubic feet per second

Concentration, total recoverable, in milligrams per liter
3Sediment and chemical transport in metric tons per day, calculated from Porterfield, 1972: (concentration, in milligrams per liter)
(discharge, in cubic feet per second) 0.00245
4Sediment concentration, in milligrams per liter
1.0 IMPORTANCE OF SEDIMENTS TO AQUATIC TRACE METAL CHEMISTRY Continued

7 .4 Comparison of Trace Metal Concentrations in Sus-

pended and Bottom Sediments Versus Dissolved
Levels

Bottom sediment and suspended sediment con- for example, lead levels for water and sediment for
tain significantly higher concentrations of trace metals Hamburg). During the Schuylkill River Basin assess-
than are found in solution. What is meant by signifi- ment study, similar results were obtained for dis-
cantly higher? Examination of the data presented in solved and total recoverable levels (Table 1.4-2, from
the accompanying tables provides an approximate an- Feltz, 1980). Table 1.4-3 shows data from the Ama-
swer to this question. Table 1.4-1 contains data from zon and Yukon Rivers and compares and contrasts
the Elbe River, FRG (Federal Republic of Germany), dissolved metal concentrations with those in sus-
and compares and contrasts dissolved metal concen- pended sediment. The suspended-sediment concentra-
trations with those associated with bottom sediments. tions for the Amazon and the Yukon are approxi-
The bottom-sediment concentrations can be more than mately 10,000 and 7,000 times higher, respectively,
100,000 times higher than the dissolved levels (see, than the dissolved load.

10
Table 1.4-1. Comparison of metal concentration in river
water and bottom sediment from the Elbe River, FRG. (All
concentrations in milligrams per kilogram)
[Data from Forstner and Wittmann, 1979]

Location Cadmium Chromium Copper Lead Zinc

Stade (W)1 0.0005 0.003 0.005 0.005 0.025
(S)2 5 100 100 200 600

Hamburg (W) 0.0007 0.010 0.012 0.005 0.120

(S) 32 500 450 500 1,600

Hitzacker (W) 0.0008 0.020 0.020 0.008 0.140

(S) 23 250 200 200 750
1 (W)-concentration in water
2(S)-concentration in sediment

Table 1.4-2. Comparison of metal content in river water

and bottom sediment from the Schuylkill Basin. (All concen-
trations in milligrams per kilogram)
[Data from Feltz, 1980]

Location Cadmium Chromium Copper Lead Zinc

Phoenixville (W) 1 0.013 0.001 0.005 0.190 0.080
(S)2 30 130 190 250 1,000

Norristown (W) 0.010 <0.001 0.007 0.110 0.120

(S) 10 30 70 400 1,000

Philadelphia (W) 0.002 <0.001 0.004 0.001 <0.001

(S) 10 50 80 180 170
'(WJ-concentration in water
2(S)-concentration in sediment, totally recoverable

Table 1.4-3. Comparison of metal content in river water

and suspended sediment from the Amazon and 'Yukon
Rivers. (All concentrations in milligrams per kilogram)
[Data from Gibbs, 1977]

Location Chromium Manganese Iron Cobalt Nickel Copper

Amazon (W) 1 0.019 0. 110 5.1 0.004 0.010 0.026
(S)2 195 1, 100 56,000 41.4 105 265

Yukon (W) 0.019 0. 200 8.9 0.006 0.020 0.060

(S) 115 1, 300 63,000 40.6 135 415
1 (W)-concentration in water
2(S)-concentration in suspended sediment

11
1.0 IMPORTANCE OF SEDIMENTS TO AQUATIC TRACE METAL CHEMISTRY Continued

7.5 Historical Records

An undisturbed sediment sink can contain a his- around 1800 to 1820. That activity, in fact, diluted the
torical record of chemical conditions. If a sufficiently mercury levels below normal background for the area
large and stable sink (one in which neither physical through increased erosion and deposition of non-
nor chemical alteration has occurred) can be found mercury-bearing material. The levels found below 30
and studied, the investigator can establish the occur- cm are believed to reflect natural background levels of
rence of chemical changes over time and may estab- 0.3 to 0.4 mg/kg (300 to 400 ppb). Industrial input is
lish area baseline levels with which existing condi- believed to have begun around 1900 (~9 cm); there
tions can be compared. When chemical analyses are ensued a steady rise until around 1940 (~5 cm), when
used together with radiometric dating techniques concentrations leveled off at about four times natural
(Pb210 , C 14), historical changes in water quality can be levels. A similar pattern can be seen in sediments
elucidated. Figure 1.5-1 gives two examples of histor- from Lake Windermere in England. Since 1400, mer-
ical records detailing changes in mercury concentra- cury levels have risen steadily as a result of erosion,
tion through time. The data from Lake Ontario indi- industrialization, mining, quarrying, use of fossil
cate that high concentrations of mercury occur in the fuels, and sewage disposal. The onset of industrial
top 6 cm; below 8 cm, the concentration decreases input occurred around 1880 ( 24 cm), and today's
significantly. At about 25-cm depth, the concentration levels are about four to five times higher than natural
reaches a low of 0.140 mg/kg (140 ppb), which corre- background concentrations.
sponds to a period of active deforestation by settlers

12
 MERCURY IN SEDIMENT, IN MILLIGRAMS PER KILOGRAM

0.2 0.4 0.4 0.6 0.8 1.0 1.2

I______I______I______I______|_

10

15 -

20-

24-
1880 26

1860

AMBROSIA HORIZON 1840

20-
1820

1800
1440-1800 34

44
40
520-1260 70

Lake Ontario, Canada Lake Windermere, England

Figure 1.5-1. Examples of mercury concentration changes through time. (Data from Thomas, 1972; Aston and others, 1973)

13
2.0 PHYSICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS

2.1 Introduction

The preceding material is based on the determi- size, surface area, specific gravity, magnetic proper-
nation of total bottom or suspended sediment chemical ties, mineralogy, and so forth). Chemical partitioning
concentrations. This type of chemical information is a is the separation of various trace metals with respect
typical and standard starting point for most water- to their association with other materials within the
quality studies. However, to understand the factors sediment (organic matter, carbonates, manganese ox-
affecting sediment trace metal chemistry, to enable an ides, sulfides, or silicates) and to the type of associa-
investigator to address such topics as trace metal tion that exists (adsorption, complexation, substitu-
transport and potential environmental impact intelli- tion within mineral lattices, and so forth). The
gently, and to permit the comparison of sample data separation into physical and chemical categories is
from the same or different areas, a number of physical somewhat arbitrary; for example, composition could
and chemical factors must be examined. They fall just as easily be placed under chemical partitioning.
under the general heading of the various components For convenience, the two types of partitioning are
or properties of a sediment. Physical partitioning is discussed separately, but they are interrelated. The
the separation of a sediment sample (bottom or sus- factors involved in partitioning are listed in
pended) into various physical categories (by grain table 2.1-1.

2.2 Grain-Size Ranges and Effect of Grain Size

One of the most significant factors controlling square meters per gram as opposed, for example, to
both suspended- and bottom-sediment capacity for re- sand-sized particles with surface areas commonly
taining trace metals is grain size (Goldberg, 1954; measured in tens of square centimeters per gram
Krauskopf, 1956; Goldberg and Arrhenius, 1958; (Grim, 1968; Jones and Bowser, 1978). Surface
Hirst, 1962; Jenne, 1968; Kharkar and others, 1968; chemical reactions are extremely important to aquatic
Gibbs, 1973; Horowitz, 1974; Stoffers and others, trace metal sediment interactions; thus, fine-grained
1977; Gibbs, 1977; Jones and Bowser, 1978; Filipek sediments, because of their large surface areas, are the
and Owen, 1979; Jenne and others, 1980; Thorne and main sites for the collection and transport of inorganic
Nickless, 1981). Table 2.2-1 lists the sediment constituents (Krauskopf, 1956; Jenne, 1968; Gibbs,
particle-size classes (names) and the size ranges the 1973; Jones and Bowser, 1978; Jenne and others,
names represent. In lakes, rivers, estuaries, and 1980). However, Jenne (1976) indicates that clay-
oceans and in sediment chemistry in general most sized particles may be viewed simply as mechanical
seidments tend to be composed of materials smaller substrates upon which trace metals can concentrate
than 2,000 |xm (very coarse sand). (without chemical interaction). Bear in mind that
There is a very strong positive correlation be- metal concentrations can and do accumulate on many
tween decreasing grain size and increasing trace metal substrates, including sand, pebbles, cobbles, and
concentrations. This correlation results from numer- boulders (Filipek and others, 1981; Robinson, 1982);
ous factors that are both physical and mineralogical nevertheless, high concentrations are more commonly
(compositional). Clay-sized sediments (less than 2 to associated with fine-grained material.
4 |jim, table 2.2-1) have surface areas measured in

14
Table 2.1-1. Examples of types of physical and chemical partitioning

Chemical Chemical
Physical Mechanistic Phase

Grain size Adsorption Interstitial water

Surface area Precipitation Carbonates
Composition Co-precipitation Clay minerals
Cation exchange Organometallic Organic matter
capacity bonding Hydrous iron and
Specific gravity Incorporation in manganese oxides
Magnetic properties crystalline minerals Sulfides
Surface charge Silicates

Table 2.2-1. Sediment particle size and classes

[Data from Guy, 1969]

Class Name Millimeters Micrometers

Boulders >256
Large cobbles 256 - 128
Small cobbles 128- 64
Very coarse gravel 64-32
Coarse gravel 32- 16
Medium gravel 16- 8
Fine gravel 8- 4
Very fine gravel 4- 2
Very coarse sand 2,000- 1,000
Coarse sand 1,000- 500
Medium sand 500- 250
Fine sand 250- 125
Very fine sand 125- 62
Coarse silt 62- 31
Medium silt 31- 16
Fine silt 16- 8
Very fine silt 8- 41

Coarse clay 1 4- 2
Medium clay 2- 1
Fine clay 1- 0.5
Very fine clay 0.5- 0.25
!Many sedimentologists consider that the clay/silt break occurs at
2 tun rather than at 4 fun.

15
2.0 PHYSICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

2.3 Chemical Analysis of Various Grain Sizes in

Bottom Sediments

How significant is the effect of grain size on though the less than 2-jjum fraction represents only
chemical composition in bottom sediments? Table some 20 percent of the bulk sediment, its copper con-
2.3-1 contains data for a marine sample that was bro- tribution amounts to 75 percent of the total copper in
ken down into various size fractions; each fraction the sample. The 10 to 20-jjLm fraction copper contri-
was then subjected to chemical analysis. The sediment bution amounts to only 9 percent of the total copper,
displays bimodal grain-size distribution with peaks in although it is the largest single size fraction in the
the ranges of less than 2 jjum and 10 to 20 jjum; this is sample. This type of pattern is similar to ones for other
typical for marine material (Rex and Goldberg, 1958). metals such as Cd, Ni, Co, Zn, and Pb.
The less than 2-jjum fraction has, by far, the
highest copper concentration of all the fractions. Al-

16
Table 2.3-1. Distribution of copper by size fraction in a
bottom sediment sample

Chemical Contribution
Size Fraction Percent Concentration to Sample1
(M-m) of Sample (mg/kg) (mg/kg)
20'
<2 750 150
2- 6 15 60 9
6- 10 18 110 19.8
10-20 30 60 18
20-32 10 25 2.5
32-64 7 20 1.4
100 200.7
Calculated by multiplying chemical concentration by percent of sample.

17
2.0 PHYSICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

2.4 Chemical Analysis of Various Grain Sizes in

Suspended Sediments

Are the patterns just described for bottom sedi- metal transport can be attributed to the less than
ments also typical of suspended matter? Data pre- 2-(Jim fraction. These data also imply that most chem-
sented in figure 2.4-1 should clarify this point. All the ical transport may occur under relatively low dis-
inorganic constituents investigated Mn, Fe, Cr, Ni, charge because high metal concentrations are associ-
Cu, and Co have their highest concentrations in the ated with the fine-grained material, which requires
less than 2-micrometer fraction (fig. 2.4-1). Further, relatively low discharge conditions for suspension and
this illustration also shows that the majority of the transport.

18
 EXPLANATION

Mn manganese Ni nickel
Fe iron Cu copper
Cr chromium Co cobalt

Relative mass transport calculated from size distribution

data of material transported by each river and the
metal concentrations for each size fraction:
YUKON RIVER
as a relative term, it has no units
36 - 1800

0.1 1.0 10 100 250 0.1 1.0 100 250

PARTICLE DIAMETER. IN MICROMETERS

Figure 2.4-1. Distribution of metals by particle size, and relative mass transport by suspended sediment. (Data from Gibbs, 1977)

19
2.0 PHYSICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

2.5 Effect of Grain Size on Trace Metal Concentra-

tions in Samples Collected From the Same Basin

The effect of grain size on sediment-trace metal concentrations. However, if the chemical data are
concentrations should be fairly evident now, in light plotted against grain-size information, very distinct
of the preceding discussions. Further, it should be patterns emerge. There is a strong positive correlation
apparent that the finer grain sizes contain the majority between increasing chemical concentration and the
of the metals associated with sediments. However, it increasing percentage of fine-grained material. This
is important to understand the significant implications relation, which is common in sediments, can provide
such relations have for such common practices as try- a means for tracing the extent of pollutant transport or
ing to trace the extent of pollutant discharge from a dispersion from a point source because, as material
point source, and for comparing sediment-chemical moves away from the source, it is usually diluted by
concentrations within and between depositional other constituents and other grain sizes. As a result,
basins. the use of bulk chemical data would make it difficult
The data in figure 2.5-1 come from bottom sed- to trace the extent of dispersion because of the effects
iments collected from the River Ems in Europe of dilution. However, these effects could be reduced,
(de Groot and others, 1982). The manganese concen- and the extent of transport determined, either by sep-
trations vary over a wide range, from some 7,000 to arating out the constituent-bearing size fraction or by
about 25,000 mg/kg; cobalt, iron, and mercury also determining the percentage of the constituent-bearing
show wide concentration ranges. Solely on the basis size fraction in a bulk sediment sample and recalculat-
of bulk chemical analysis, there appears to be no inter- ing the bulk data accordingly. Recalculation methods
relation between the various samples, and all that an are discussed in section 4.0 of this Primer.
investigator could report would be the wide range of

20
 32

28

24

Manganese x 100, in mg/kg

20

s 16
LU
O
z
o
o
12

Mercury X|, in mg/kg

-A
A A

0 - I I I
0 20 40 60 80 100

GRAIN SIZE LESS THAN 16 /Ltm, IN PERCENT

Figure 2.5-1. Relation between metal concentration and grain size for the River Ems, FRG. (Data
from deGroot and others, 1982)

21
2.0 PHYSICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

2.6 Comparison of Samples Having Similar Bulk

Chemistries But Differing Grain-Size Distributions

Keep in mind that two sediments with the same ples in the distribution of Mn. Sample AD 600 has two
bulk chemical concentrations are not necessarily the major Mn peaks, one in the less than 2 (Jim fraction,
same. The sources for the constituents in question and the other in the 10-to-20 (xm fraction. Sample AD
may be different, the constituents may be held in 500 has similar Mn concentrations for all size frac-
different chemical forms, they may be associated with tions ranging from 2 (Jim to 32 (Jim. Subsequent min-
different size fractions, and so forth. As an example, eralogical and chemical analyses have indicated that
examine the data in table 2.6-1. Both samples are manganese oxide coatings and micronodules caused
marine in origin. Sample AD 600 comes from a hy- the Mn peaks in AD 600, while the Mn in AD 500
drothermally active area, while sample AD 500 comes came from volcanic ash and glass. Differentiation of
from a nonhydrothermal area. Both samples in this the two samples could not have been made solely on
table have similar manganese concentrations the basis of either grain-size distribution or bulk
( 3500 mg/kg) and similar grain-size distributions. chemical data, but can be made on the basis of size-
However, the chemical data for each grain-size range fraction chemical analysis.
indicate significant differences between the two sam-

22
Table 2.6-1. Comparison of manganese distributions in two samples having similar grain-size distributions
and bulk concentrations

Sample AD 600 Sample AD 500

Size Fraction Percent Concentration Contribution 1 Percent Concentration Contribution 1

(jim) of Sample (mg/kg) (mg/kg) of Sample (mg/kg) (mg/kg)

<2 20 5,000 1,000 18 4,000 720

2-6 12 1,000 120 13 3,800 494
6 - 10 14 1,000 140 15 3,800 570
10- 20 30 7,100 2,130 25 4,000 1,000
20- 32 10 1,000 100 15 4,000 600
32-64 7 500 35 5 2,000 100
64- 125 4 300 12 5 1,000 50
125 - 250 3 300 9 4 500 20

100 3,550 100 3,550

Contribution calculated by multiplying chemical concentration by percent of sample.

23
2.0 PHYSICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

2.7 Effect of Grain Size on Sediment-Associated

Chemical Transport at Differing Discharge Rates

It has now been reasonably well established that trace metals, is diluted with metal-poor, larger sized
trace metals tend to be concentrated in the finer grain material, suspended-sediment concentration de-
sizes of both suspended and bottom sediments. It has creases, as it does for chromium (fig. 2.7-1). Remem-
also been shown that larger sized material can act as ber that the weight of chromium or other metals trans-
a diluent for the constituent-rich finer grain sizes. ported will not decrease and may, in fact, increase
What, then, is the implication for suspended-sediment (see table 1.3-1 for calculations concerning transport);
constituent concentrations under differing discharge only the concentration in the suspended load will de-
conditions? The immediate assumption might be that crease. (This relation is applicable to large rivers;
as both discharge and the suspended-sediment load caution should be exercised in applying it to small
increase, associated constituent concentrations also rivers.) The data presented in figure 2.7-1, and as also
increase. However, a little thought and an examina- shown by figure 2.4-1, reemphasize the point that
tion of the data presented in figure 2.7-1 should dispel most chemical transport may occur under relatively
this assumption. As discharge increases, the size of low flow conditions, because of the high concentra-
sediment grains being suspended and transported will tions of metals associated with small-size fractions
also increase. When fine-grained material, rich in that do not require high velocities to move them.

24
 CONCENTRATION, IN MILLIGRAMS PER KILOGRAM
gf _* -*
ro o> o £
£ Y1 8
5 *l 018 i8 r§
a> -L
3 S
Q- (t

I!
OJ in

-§Sa.

OJ

5
Q.
ii
o Q
:r
OJ m
«
O" z
n
^
O §
E' O

5' mH
c m
33
CO
| o
CL
8. m
I CO
m
0) O
ut
O c*>
O" Z Q
0 8
2.0 PHYSICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

2.8 Effect of Surface Area

The preceding discussions have demonstrated areas on the order of tens of square meters per gram.
the importance of grain size to sediment-trace metal Bear in mind that these numbers are for spheres.
concentrations. A number of other factors also have Obviously, real sediments are composed not of
significant effects on metal concentrations (see table spheres, but of irregular grains. Jackson (1979) has
2.1-1); there is a "common thread" for all these fac- calculated that the surface area of montmorillonite (a
tors: grain size. As grain size decreases, these addi- platey clay mineral) can theoretically be more than
tional factors become more important, and a positive 800 square meters per gram (if it were a sphere of
correlation usually exists between them and increas- 2 (Jim in diameter, its surface area would only be 1.13
ing metal concentration. In some instances, it is im- square meters per gram). Actual measurements of real
possible to differentiate between effects caused by materials, including montmorillonite, indicate that
factors like surface area, cation exchange capacity, surface area can be and usually is less than the theoret-
and surface charge, and effects due to grain size. ical value but is still much larger than that of a sphere
Hence, if only one physical parameter of a sediment of equivalent diameter (Forstner and Wittmann,
sample is to be determined, in order to aid in interpret- 1979). These surface areas were measured for selected
ing chemical data, grain size is by far the parameter of materials with less than 2-|jim diameter:
choice, because it integrates all the others.
Surface area is almost indistinguishable from Material Surface area
grain size; as grain size decreases, surface area in-
creases sharply. As an example, examine the data in Calcite 12.5 m2/g
table 2.8-1. The surface area for simple spheres of Clay minerals
differing diameters (grain size) has been calculated Kaolinite 10 to 50 m2/g
(Jackson, 1979). Sand-sized particles have surface ar- Illite 30 to 80 m2/g
eas on the order of tens to hundreds of square centime- Montmorillonite 50 to 150 m2/g
ters per gram, silt-sized particles have surface areas on Iron hydroxide 300 m2/g
the order of hundreds to thousands of square centime- Organic matter I,900m2/g
ters per gram, and clay-sized particles have surface (humic acids)

26
Table 2.8-1. Calculated surface areas assuming sphericity
[Data from Jackson, 1979]

Surface Area
Size Class Diameter ((Jim) (m2/g) (cm2/g)
Very coarse sand 2,000 0.00113 11.3
Very fine sand/coarse silt 62 0.036 360
Very fine silt/coarse clay 4 0.57 5,700
Medium clay 2 1.13 11,300
Fine clay 1 2.26 22,640
Very fine clay 0.5 4.52 45,280
Colloids 0.1 22.6 226,400

27
2.0 PHYSICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

2.9 Importance of Surface Area to Sediment-Trace

Metal Concentrations

Why is surface area so important in controlling constituents can concentrate without any chemical in-
sediment-trace metal concentrations? It is because all teraction between the material and the constituent.
the various means by which sediment tends to collect Thus, deposited materials like organic matter and hy-
metals fall into the general category of so-called sur- drous iron and manganese oxides, rather than the orig-
face reactions or surface chemistry. It follows that inal surface, may act as metal collectors. Obviously,
materials with large surface areas (small grain size) as surface area increases, so does the amount of these
are the main sites for the transport and collection of collectors, thus increasing the metal-concentrating ca-
these constituents (Krauskopf, 1956; Jenne, 1968; pacity of the surface.
Gibbs, 1973; Jenne, 1976; Jones and Bowser, 1978; An example of how surface area affects inor-
Jenne and others, 1980). The major mechanism for ganic concentrations is shown in figure 2.9-1. The
the collection of trace metals on surfaces is adsorp- data come from the Ottawa and Rideau Rivers and
tion. This process entails the condensation of atoms, clearly show that as surface area increases, the con-
ions, or molecules on the surface of another sub- centrations of both nickel and copper increase. This
stance. Materials having large surface areas are good pattern is also typical of many other metals like Fe,
adsorbers. Adsorption can occur with or without Mn, Zn, Pb, and Cd. Bear in mind that the surface
cation exchange (see section 2.10), and should not be areas plotted on these graphs (fig. 2.9-1) are actual
confused with absorption, which involves penetration measurements they were not calculated. Do not as-
of a substance into the body, or inner structure, of sume, for example, that a surface area of 22.6 m2/g
another material. Additionally, Jenne (1976) indicates implies a sediment particle size of some 0.1 (xm,
that materials with large surface areas can be viewed which would be the case if the particle were a sphere
simply as mechanical substrates upon which inorganic (see section 2.8).

28
 CONCENTRATION, IN MILLIGRAMS PER KILOGRAM

ro o> oo fO o> oo
o o o o o o
0)
£ r- C
=a. fI*
73 'i
O
m
0 >»
fo §
i ^T m
n
6) Q
"
CO
O

sr 3
o3 sr~ m

0§
m
m
x
CO

m
| o
2.0 PHYSICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

2.70 Cation Exchange Capacity

Many different materials that are common com- (iron hydroxides); and COOH and OH (organic mat-
ponents of sediment and that have large surface areas, ter) (Forstner and Wittmann, 1979). CEC may also
such as clay minerals, iron hydroxides, manganese occur between the layers of certain silicates, like ex-
oxides, and organic matter, are capable of sorbing pandable clays (montmorillonite, smectites), which,
cations from solution and releasing equivalent depending on how one chooses to define the layers,
amounts of other cations back into solution. This might not be viewed as "surfaces" (Grim, 1968).
process is called ion exchange. Since most trace The data in table 2.10-1 include measured
metals are cations (have a positive ionic charge), and cation exchange capacities for selected materials com-
the surfaces of most materials with a capacity for this monly associated with this phenomenon. The data
process have a net negative charge, the process is also show that as grain size decreases and surface area
termed cation exchange. The ability of a particular increases, CEC increases significantly (tables 2.10-2
material to sorb cations can be measured and is called and 2.10-3). Note that CEC results from the availabil-
CEC (cation exchange capacity). The actual mecha- ity of "unbalanced" negatively charged sites; thus, it
nism by which cation exchange occurs is a matter of could readily be stated that it is a result of a net
some debate, but appears to be due to the sorptive negative surface charge (another important physical
properties of negatively charged anionic sites such as factor see table 2.1-1).
SiOH, A1(OH)2 , and A1OH (clay minerals); FeOH

30
Table 2.10-1. Cation exchange capacity of selected materials

Exchange Capacity
Material (mEq/100 g) Reference
Kaolinite 3- 15 1
Illite 10- 40 1,2
Chlorite 20- 50 1
Montmorillonite 80- 120 1, 2
Smectites 80- 150 2
Vermiculite 120- 200 2
Iron hydroxide 10- 25 1
Soil humic acids 170- 590 1
Manganese oxides 200- 300 1
1 - From Forstner and Wittmann (1979)
2 - From Drever (1982)

Table 2.10-2. Variation of cation exchange capacity with particle size for kaolinite
[Data from Grim, 1968]

Particle size 10-20 5-10 2-4 1-0.5 0.5-0.25 0.25-0.1 0.1-0.05

Cation exchange capacity

(mEq/100 g) 2.4 2.6 3.6 3.8 3.9 5.4 9.5

Table 2.10-3. Variation in cation exchange capacity with particle size for illite
[Data from Grim, 1968]

Particle size ((Jim) 1-0.1 0.1-0.06 <0.06

Cation exchange capacity
(mEq/100 g) Sample A 18.5 21.6 33

Sample B 13.0 20.0 27.5

Sample C 20.0 30.0 41.7

31
2.0 PHYSICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

2.77 Significant Sedimentary Trace Metal Collectors

As has been shown, a number of physical fac- Table 2.11-1 lists the various important features
tors affect the ability of a sediment to collect and of metal sedimentary collectors. Forstner (1982a) has
concentrate trace metals. With these factors as a basis, listed these materials in descending order according to
plus additional chemical work, it is possible to iden- their ability to collect and concentrate trace metals:
tify the compounds and substances that are most sig- manganese oxides, humic substances (organic mat-
nificant in providing a sediment with the capacity to ter), hydrous iron oxides, and clay minerals. These
concentrate metals. These compounds and substances results are based upon sequential extraction studies
have large surface areas, cation exchange capacities, (see section 3.5). Within each category there is sub-
and high surface charges, and they tend to be found in stantial variability depending upon the geochemical
the finer size fractions. Further, sedimentary metal character of the environment (type of clay minerals
collectors tend to be thermodynamically unstable and present, concentration of organic matter, pH, Eh) and
are amorphous or cryptocrystalline (Jones and the various metals involved.
Bowser, 1978). The most common materials meeting
these criteria are clay minerals, organic matter, hy-
drous manganese oxides, and hydrous iron oxides.

2.72 Iron and Manganese Oxides

The characteristics of iron and manganese ox- 1978; Forstner, 1982a,b;). Micronodules have been
ides are outlined in table 2.12-1. These substances found in a variety of size ranges but, as their name
have long been known as excellent scavengers of trace implies, tend to be found in the smaller (less than 20
metals from solution (Goldberg, 1954; Krauskopf, |xm) size ranges (Goldberg and Arrenius, 1958;
1956). Although the most spectacular demonstration Chester and Hughes, 1969; Dymond and others, 1973;
of this effect is the manganese nodules located at the Horowitz, 1974). In soils, suspended sediment, and in
sediment-water interface on deep ocean floors or on bottom sediment, iron and manganese oxides also
lake beds (Mero, 1962; Moore and others, 1973), the commonly occur as coatings on various minerals and
separation and identification of manganese micronod- finely dispersed particles (Forstner and Wittmann,
ules in core and grab samples indicate that they are 1979). Those forms most capable of concentrating
ubiquitous and play a significant role throughout the metals range from amorphous to microcrystalline to
sediment column as metal collectors in aquatic envi- crystalline and have large surface areas on the order
ronments (Goldberg and Arrhenius, 1958; Chester of 200 to 300 m2/g (Fripiat and Gastuche, 1952; Buser
and Hughes, 1967; Jenne, 1968; Cronan and Garrett, and Graf, 1955). Regardless of form, whether mi-
1973; Duchart and others, 1973; Dymond and others, cronodules or coatings, hydrous iron and manganese
1973; Moore and others, 1973; Horowitz, 1974; Lee, oxides are significant concentrators for trace metals in
1975; Horowitz and Cronan, 1976; Jones and Bowser, aquatic systems.

32
Table 2.11-1. Compositional controls on metal concentra-
tion

Important Features

1. Thermodynamically unstable
2. Amorphous or cryptocrystalline
3. Capable of extensive substitution
4. High cation exchange capacity
5. Large surface area
6. High surface charge
7. Small particle size
Most Significant Collectors

1. Hydrous iron oxides

2. Hydrous manganese oxides
3. Organic matter
4. Clay minerals
Relative Capacity of Collectors1

Manganese oxides > organic matter > iron oxides

> clay minerals
lFrom Forstner (1982a)

Table 2.12-1. Iron and manganese oxides

Characteristics

1. Fine grained
2. Amorphous or poorly crystallized
3. Large surface area
4. High cation exchange capacity
5. High negative surface charge

33
2.0 PHYSICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

2.73 Organic Matter

The characteristics of aquatic and soil organic stituent and the type of organic matter (Swanson and
matter are outlined in table 2.13-1. The ability of others, 1966; Saxby, 1969; Rashid, 1974; Bunzl and
organic matter to concentrate trace metals in and on others, 1976; Jonasson, 1977). Organic matter can
soils as well as suspended and bottom sediments is concentrate between 1 percent and 10 percent dry
well recognized (Goldberg, 1954; Krauskopf, 1956; weight of Co, Cu, Fe, Pb, Mn, Mo, Ni, Ag, V, and
Kononova, 1966; Swanson and others, 1966; Saxby, Zn (Swanson and others, 1966). The ability to con-
1969; Schnitzer and Kahn, 1972; Gibbs, 1973; Bunzl centrate various trace metals appears to be related to
and others, 1976; Jenne, 1976; Singer, 1977; Stoffers a number of factors, including large surface area, high
and others, 1977; Nriagu and Coker, 1980; Ghosh and cation exchange capacity, high negative surface
Schnitzer, 1981; Forstner, 1982a,b). Aquatic organic charge, and physical trapping (table 2.13-1). It is also
matter, generally termed humic substances, has been related to the stability of the organic-metal constituent
subdivided by Jonasson (1977) into four categories: complex. In soils, the sequence in descending order is
humins, humic acids, fulvic acids, and yellow organic Pb, Cu, Ni, Co, Zn, Cd, Fe, Mn, Mg the so-called
acids (see table 2.13-1). Gibbs (1973), among others, Irving-Williams Series (Irving and Williams, 1948)
has indicated the importance of organic molecules in (table 2.13-1). Similar results have also been found
controlling trace metal concentrations on and in sus- for aquatic organic matter (Swanson and others, 1966;
pended and bottom sediments, and in sediment-water Saxby, 1969; Rashid, 1974; Bunzl and others, 1976;
interactions. Saxby (1969) has shown that the relative Jonasson, 1977). The concentration of aquatic organic
attraction between metals with colloidal, suspended, matter, as indicated by such constituents as total or-
and bottom sediment-associated organic matter can ganic carbon and total organic nitrogen, tends to show
range from weak and readily replaceable (adsorption) a strong positive correlation with decreasing grain size
to strong (chemically bonded). The ability of organic and increasing surface area (Kuenen, 1965; Forstner
matter to concentrate metals varies with the con- and Wittmann, 1979).

2.74 Clay Minerals

The characteristics of clay minerals are outlined tens of minutes) and depends upon a number of
in table 2.14-1. These materials can act as significant physicochemical factors like the valence, ionic radii,
collectors and concentrators of trace metals in aquatic and concentration of the constituent, the type of clay,
systems (Goldberg, 1954; Krauskopf, 1956; Goldberg the pH of the solution, and the nature and concentra-
and Arrhenius, 1958; Hirst, 1962; Grim, 1968; tion of competing substrates (Forstner and Wittmann,
Kharkar and others, 1968; Gibbs, 1973; Jenne, 1976; 1979). An empirical affinity sequence for some heavy
Jones and Bowser, 1978; Forstner and Wittmann, metals and clay minerals gives, in descending order,
1979; Forstner, 1982a,b). Clay mineral capacity for lead, nickel, copper, and zinc (Mitchell, 1964). Re-
cation exchange is governed by broken chemical cent studies in some European rivers appear to indi-
bonds around the edges of mineral grains, the substi- cate that clay minerals may not play a significant role
tution of Al+3 (aluminum) for Si +4 (silicon) with the in the direct chemical concentration of inorganic con-
associated charge imbalance, and the presence of ex- stituents by, for example, adsorption (Forstner and
pandable lattices (Grim, 1968). Hirst (1962) has Wittmann, 1979). That would support the view of
pointed out the importance of different clay minerals Jenne (1976) that the major role of clay minerals as
in controlling the background levels of metals in bot- metal concentrators is to act as mechanical substrates
tom sediments and has evaluated the capacity of vari- for the precipitation and flocculation of organic matter
ous types of clays with respect to their ability as trace and secondary minerals (for example, hydrous iron
metal concentrators in descending order, as follows: and manganese oxides). In other words, the clays are
montmorillonite, vermiculite, illite equal to chlorite, coated with material that, rather than the clays them-
kaolinite. The actual process by which clay minerals selves, actually carries out the concentration of trace
concentrate constituents is not well understood; how- metals.
ever, laboratory studies indicate that it can be rapid (in

34
Table 2.13-1. Organic matter

Characteristics
1. Concentration increases with decreasing grain size
2. Large surface area
3. High cation exchange capacity
4. High negative surface charge
5. Capable of physical trapping
Types 1

1. Humins
2. Humic acids
3. Fulvic acids
4. Yellow organic acids
Metal Affinities 2

Lead > copper > nickel > cobalt > zinc > cadmium
> iron > manganese > magnesium
'From Jonasson (1977)
2From Irving and Williams (1948)

Table 2.14-1. Clay minerals

Characteristics

1. Fine grained
2. Large surface area
3. Moderate to high cation exchange capacity
4. High negative surface charge
a. Broken bonds on mineral edges
b. Substitution of aluminum(III) for silicon(IV)
Capacity for Metal Concentration 1
Montmorillonite > vermiculite > illite=chlorite > kaolinite
Hirst (1962)

35
3.0 CHEMICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS

3.1 Introduction to Chemical Partitioning

Now that we have discussed physical partition- Adsorption was discussed in section 2.9. Pre-
ing and elucidated the major physical factors affecting cipitation and coprecipitation are readily understand-
interactions between trace metals and sediments, it is able terms. Organometallic bonding is the attachment
time to examine chemical partitioning. Just as aquatic of a metal directly to carbon; such bonds can be
systems are physically dynamic (water moving in a formed by almost any metal and by several nonmetals
streambed, wind-created waves in a lake) they are also (silicon, phosphorus). Examples are diethylzinc,
chemically dynamic. The key to understanding and methyl- and dimethyl mercury, and tetraethyl lead.
predicting metal transport and environmental Incorporation in crystal minerals is also called substi-
availability, as well as to identifying sources and sinks tution. This entails the substitution of one element for
for aquatic metal constituents, is the identification and another within a fixed crystal structure; substitution is
quantification of the metal associations in sediments governed by ionic radius and charge. In general and
(suspended and bottom) and the reactions among sed- there are exceptions substitution may take place
iment, water, and biota (see table 3.1-1). when ionic radii are within ± 15 percent and when the
A search of the literature on sedimentary chem- charge is the same, or differs by no more than ±1.
ical partitioning shows that two approaches have been The second approach seeks to determine where
used. The first approach aims at determining how inorganic constituents are retained on or by sediments
trace metals are retained on or by sediments the (phase or site) the so-called phase approach (see
so-called mechanistic approach. According to Gibbs table 3.1-2). This has been attempted because individ-
(1977), there are five major mechanisms for inorganic ual constituents (like Fe, Mn, Cd, Pb, Zn, Ni, Co, and
accumulation in or on sediments (suspended and Cu) may be, and usually are, associated with several
bottom): phases. The term phase is used in the thermodynamic
sense and incorporates categories like interstitial
1. Adsorption on fine-grained material water, clay minerals, sulfides, carbonates, humic
2. Precipitation of metal compounds acid, manganese oxides, and so forth (table 3.1-3).
3. Co-precipitation with hydrous iron and manganese Despite this relatively simple division into two ap-
oxides and carbonates proaches, very few attempts to chemically partition
4. Association either by adsorption or organometallic complex sediment samples entail a purely mechanistic
bonding with organic matter or phase approach; rather, they combine aspects of
5. Incorporation in crystalline minerals (table 3.1-2) both.

36
Table 3.1-1. Chemical partitioning

Aquatic systems are chemically dynamic. The key to under-

standing and predicting metal transport and environmental
availability, as well as to identifying sources and sinks, lies in
identifying and quantifying the metal associations in sediments
and the reactions that occur between sediment, water, and biota.

Table 3.1-2. Mechanisms for metal accumulation in sedi-

ments
[Data from Gibbs, 1977]

1. Adsorption on fine-grained material

2. Precipitation of metal compounds
3. Co-precipitation with hydrous iron and manganese oxides
and carbonates
4. Association either by adsorption or organometallic bond-
ing with organic matter
5. Incorporation in crystalline minerals

Table 3.1-3. Examples of phases where metal accumulates

in sediments

1. Interstitial water
2. Carbonates
3. Clay minerals
4. Organic matter
5. Hydrous iron and manganese oxides
6. Sulfides
7. Silicates

37
3.0 CHEMICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

3.2 Utility of Chemical Partitioning Data

The preceding discussion outlined the concept bioavailability, and thus the concept remains subject
of chemical partitioning. It is important, at the same to much interpretation and argument.
time, to understand the practical implications of the Because aquatic systems are both physically
concept and to grasp its importance to the environ- and chemically dynamic, both aspects must be consid-
ment and to water-quality studies. Because aquatic ered when discussing sediment-associated chemical
systems are chemically dynamic, the goal of water- transport. Physical transport of sediment is relatively
quality projects should be not only to describe existing straightforward and involves hydrodynamics. A num-
conditions at some point in time and infer their effect ber of existing physical-transport models have already
on the environment, but also to attempt to predict been tested and used. However, in order to predict
likely occurrences later, or when physicochemical chemical transport adequately, an effective model
conditions change. must be able to take into account chemical changes
Table 3.2-1 points out some of the potential that result from altered environmental conditions. The
uses of chemical partitioning and indicates some of first step toward creating such a model is an under-
the potential changes that can result from altered standing of the chemical reactions that can take place
physicochemical conditions. The table is divided into in a system and requires input regarding the existing
three major categories: bioavailability, transport mod- chemical phases and the way in which trace metals are
eling, and remobilization. Bioavailability is a com- held. This type of data can only be obtained by chem-
mon term encompassing a concept that most individu- ical partitioning.
als can grasp. However, determining actual Under the heading, "Remobilization," table
bioavailability is another matter entirely. Bio- 3.2-1 gives practical examples of how changing
availability is always operationally defined, usually as physicochemical conditions affect sediment-trace
the result of some chemical test or toxicity experi- metal concentrations. The types of changes are typical
ment. Unfortunately, no one has yet come up with a of many common environments and should be famil-
universally accepted method of determining iar occurrences to most water-quality investigators.

3.3 Chemical Partitioning Methods

Table 3.3-1 outlines the most common methods methods and their practical uses and limitations are
by which chemical partitioning is ascertained. The discussed in the following pages.

38
Table 3.2-1. Rationale for the determination of chemical partitioning

I. Bioavailability
II. Transport modeling
III. Remobilization
A. Elevated salt levels-alkali and alkaline earth displacement
1. Deicing salts
2. Saline effluents
3. Saltwater intrusion
4. Estuarine, fresh, and saltwater mixing
B. Redox changes-dissolution of iron and manganese oxides and hydroxides with subsequent metal release
1. Eutrophication
2. Organic inputs
C. pH reduction-dissolution of carbonates and hydroxides, desorption by H+
1. Acid mine drainage
2. Acid rain
3. Acidic industrial effluents
D. Increases in complexing agents-form stable soluble metal complexes
1. Tannic acid
2. Peat and lignite
3. Nitriloacetic acid-detergent additive in place of phosphate
E. Microbial activity-solubilization of metals by direct and indirect action
1. Alkylation of metals (mercury, arsenic, lead, selenium, and so forth) for example, methyl mercury
2. Destruction of organic matter
3. Metabolic activity decreasing pH or oxygen

Table 3.3-1. Methods for determining chemical partitioning

1. Manual selection of phases and analysis

2. Instrumental selection of phases and analysis
3. Partial chemical extractions
4. Density gradients and analysis
5. Statistical manipulations of bulk chemical data
6. Mathematical modeling

39
3.0 CHEMICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

3.4 Chemical Partitioning Instrumental Methods

Ideally, chemical partitioning can be accom- 1981). Such techniques as x-ray photoelectron spec-
plished by separating out specific mineral or organic troscopy have been used on many types of geological
sediment phases, whether instrumentally (for exam- materials, including sediments, with mixed success
ple, magnetic separators, heavy mineral flotation) or (Jones and Bowser, 1978; Johnson and Maxwell,
manually (picking out mineral grains with the aid of a 1981). Many of these procedures permit the user to
microscope), and subjecting them to chemical analy- select the analytical site visually, thus making possi-
sis. In actual practice, some separation techniques ble the determination of phase association for the
may change the sediment chemistry, and most are derived chemical data. Although these techniques are
extremely time consuming. Moreover, it is virtually promising, several problems limit their use. Foremost
impossible to separate the individual constituents in is the problem of determining correction factors for
the finer size range (e.g., clay minerals) although that the various counting systems used and for background
is likely to be the size range of maximum interest. interference. Because of the complexity of these cor-
Relatively new chemical analytical techniques rection factors, detection limits are usually higher than
may be used to determine chemical partitioning in 100 mg/kg (100 ppm) and can be significantly higher
sediments directly (table 3.4-1). These methods per- (Johnson and Maxwell, 1981). This problem tends to
mit the quantitative analysis of various minerals or limit the usefulness of these procedures to major ele-
mineral assemblages in natural mixtures, without sep- ments such as Fe, Mn, and Na, and could exclude
aration and sometimes nondestructively, from areas as such environmentally important trace metals as Pb,
small as 1 square micrometer (Johnson and Maxwell, Cd, Ag, and Cu.

40
Table 3.4-1. Instrumental techniques for determining chemical parti-
tioning

1. Electron microprobe-Electron beam producing x-ray emissions

2. X-ray photoelectron spectroscopy (ESCA, XPS)-Photon producing elec-
trons
3. U.V. photoelectron spectroscopy-Photon beam producing ionization and
electrons
4. Auger electron spectroscopy (AES)-Electron beam producing ionization
and secondary electrons
5. Secondary ion mass spectroscopy (SIMS)-Positive ion beam producing
chemical fragments
6. Ion scattering spectroscopy (ISS)-Ion beam producing primary ions

41
3.0 CHEMICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

3.5 Chemical Partitioning Partial Extraction

Methods

One of the oldest and most commonly used Forstner and Wittmann, 1979; Tessier and others,
methods of chemically partitioning sediments in- 1979; Nriagu and Coker, 1980; Diks and Alien,
volves the use of partial chemical extractions. Much 1983). These studies used sequential chemical extrac-
of the original work in this area was carried out on tions in an attempt to differentiate between anthropo-
marine material (Goldberg and Arrhenius, 1958; Hirst genic and natural inorganic pollutants and to try to
and Nicholls, 1958; Arrhenius and Korkish, 1959; predict or estimate bioavailability. The sequences be-
Chester, 1965; Lynn and Bonatti, 1965; Chester and came more and more complex as attempts to clarify
Hughes, 1966; 1967; 1969; Chester and Messiah- chemical partitioning increased. Some procedures in-
Hanna, 1970; Cronan and Garrett, 1973; Horowitz, volved six, seven, and even nine steps (Schmidt and
1974; Horowitz and Cronan, 1976). These early pro- others, 1975; Brannon and others, 1976; Summer-
cedures usually involved two- or, at most, three-step hayes and others, 1976; Stoffers and others, 1977;
sequential extractions plus a total analysis, used in Forstner, 1982a,b).
increasing strength, which attempted to partition trace Sequential extractions have been used widely
metals. The concept of sequential extractions is based and on many types of material, but they are not a
on the idea that a particular reagent is either phase- panacea. The major problem is that the extraction
specific or mechanistic-specific (for example, acetic reagents are not as "selective" as many users assert
acid will only attack and dissolve carbonates; ammo- (Malo, 1977; Pilkington and Warren, 1979). Also, as
nium acetate at pH 7 will only remove adsorbates). Forstner (1982a) points out, extraction efficiencies
Table 3.5-1 lists most of the extractants that have been vary according to length of treatment and sediment-to-
used over the years to partition sediments and soil extractant ratio. However, sequential procedures also
chemically. have advantages. They permit differentiation between
After the initial work on marine sediments, fur- samples having similar bulk chemistries. Also, they
ther advances in chemical partitioning using sequen- represent one of the few practical methods for the
tial partial extractions have been made by workers in determination of concentration mechanisms and thus
many diverse fields (Bruland and others, 1974; Gupta provide a possible means for estimating bioavailabil-
and Chen, 1975; Brannon and others, 1976; Chen and ity (see Diks and Alien, 1983). Finally, they offer a
others, 1976; Gambrell and others, 1977; Luoma and potential means of providing the requisite chemical
Jenne, 1977; Malo, 1977; Stoffers and others, 1977; partitioning data needed for transport modeling.

42
Table 3.5-1. Reagents employed in sequential partial chemical extractions

Classification Reagent Reference

Adsorbates and Exchangeables 0.2 BaCl2-triethanolamine, pH 8.1 Jackson, 1958
1 M NH4OAc, pH 7 Jackson, 1958
Distilled deionized water Nissenbaum, 1972
1 M NH4OAc Gupta and Chen, 1975
1 M MgCl2 , pH 7 Tessier and others, 1979
1 M NaOAc, pH 8.2 Tessier and others, 1979
Carbonates 1 M HOAc (25% v/v HOAc) Chester and Hughes, 1967
1 M NaOAc, pH 5 w/ HOAc Tessier and others, 1979
CO2 treatment Pachineelam, 1975
Exchange columns Deurer and others, 1978
Detrital/Auth igenic EDTA treatment Goldberg and Arrhenius, 1958
Hydrogenous/Lithogenous 0.1 M HC1 Piper, 1971
0.3 M HC1 Piper, 1971
Reducible 1 M NH2OH-HC1 w/ 25% v/v HOAc Chester and Hughes, 1967
Moderately Reducible Oxalate buffer Schwertmann, 1964
(hydrous iron oxides) Dithionate/citrate buffer Holmgren, 1967
Easily Reducible
0.1 M NH2OH-HC1 w/ 0.01 M HNO3 Chao, 1972
(Mn and amorphous oxides)
Organic s Na hypochlorite w/ dithionate-citrate Gibbs, 1973
H2O2 at 95°C, pH 2 Engler and others, 1974
30% H2O2 w/ 0.02 N HNO3 , pH 2 Gupta and Chen, 1975
extracted w/ 1 M NH4OAc in 6% HNO3
30% H2O2 w/ 0.02 N HNO3 , pH 2 Gupta and Chen, 1975
extracted w/ 0.01 M HNO3
1:1 methanol-benzene (bitumen) Summerhayes and others, 1976
0.1 N NaOH (humic acids) Schnitzer and Kahn, 1978
0.02 M HNO3 w/ H2O2 , pH 2 w/ HNO3 at Tessier and others, 1979
85°C, w/ 3.2 M NH4OAc in 20% HNO3
30% H2O2 in 0.5 N HC1, heat Kitano and others, 1980
Sulfides 30% H2O2 at 95°C, extracted w/ Gupta and Chen, 1975
1 N NH4OAc
0.1 N HC1 w/ air Kitano and others, 1980
Detrital Silicates HF/HC1O4/HNO3 Johnson and Maxwell, 1981
Borate fusions w/ HNO3 Johnson and Maxwell, 1981

43
3.0 CHEMICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

3.6 Chemical Partitioning of Suspended Sediments by

Partial Extraction

For an example of the type of information that With these results as a basis, it is possible to
can be obtained using partial chemical extractions, discuss potential environmental availability. Metals
examine the data in figure 3.6-1. The information was associated with crystal lattices are essentially unavail-
generated by a four-step chemical extraction sequence able. Metals associated with ferromanganese coatings
performed on suspended sediment from the Amazon probably will be stable or environmentally unavail-
and Yukon Rivers (Gibbs, 1977). Most of the able unless there is a significant decrease in dissolved
chromium, cobalt, and copper is concentrated in min- oxygen or a significant increase in biological activity.
eral lattices; most of the manganese and nickel is Ferromanganese-associated metals may be released in
associated with ferromanganese coatings. The iron the digestive systems of certain organisms (Diks and
was evenly divided between mineral lattices and ferro- Alien, 1983). Organics-associated metals could be
manganese coatings. Adsorption was not a significant available, particularly if ingested by an organism. Ad-
contributor to the concentration of any of the metals sorbed metals are readily available but, as can be
studied, and organic matter was a significant collector seen, represent a relatively small percentage of the
only in the case of cobalt and nickel. metals present in the suspended matter.

44
organic organic exchange
~^*^^j^^^~ __
exchange coating coating exchange exchange organic solution organic

Chromium Manganese

crystal

^-.r-
coating fiVorganic

organic solution solution exchange

solution
Iron Cobalt

organic

organic
coating
solution exchange exchange solution

Nickel Copper

AMAZON RIVER YUKON RIVER AMAZON RIVER YUKON RIVER

Solid phases, in percent

Phase iromium Manganese Iron Cobalt Nickel Copper

Absorbed 3.5 0.8 0.007 6.5 3.2 3.5
Coatings 5.2 54.8 43.6 28.8 47.2 5.7
Organics 11.0 6.7 9.0 16.4 15.0 4.5
Crystalline 80.3 37.8 47.4 483 34.6 86.2

Figure 3.6-1. Chemical partitioning of selected metals in the Amazon and Yukon Rivers. (Data from Gibbs, 1977)

45
3.0 CHEMICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

3.7 Chemical Partitioning of Bottom Sediments by

Partial Extraction

The data in table 3.7-1 provide another example break down manganese oxides. For this extraction, it
of the utility of partial chemical extractions. This in- is obvious that the samples are dissimilar. The sample
formation was generated by carrying out a four-step from the hydrothermally active area has three times as
sequential extraction on two marine bottom sediment much Mn associated with oxides as has the sample
samples, both having similar Mn concentrations. from the inactive area. The hot HC1 (hydrochloric
Sample 1 comes from a hydrothermally active region, acid) attack is designed to break down all but the most
and sample 2 comes from a hydrothermally inactive resistant silicates and aluminosilicates. Here, too, the
region. Ordinarily, hydrothermally active areas dis- samples are dissimilar; much more manganese is held
play high Mn concentrations. Two questions about in this fraction in the hydrothermally inactive sample
these samples need to be addressed: Are the samples than in the active one. These results show that the
geochemically similar, as implied by the bulk chemi- sequential partial extractions provided a means of de-
cal results? If they are dissimilar, why? termining how manganese was held in the samples;
The acetic acid extraction is designed to remove they also permitted the differentiation of the two sam-
adsorbates and metals associated with carbonate min- ples. This differentiation would have been impossible
erals. For this extraction, the two sediments are simi- solely on the basis of bulk chemical data.
lar. The hydroxylamine extraction is designed to

46
Table 3.7-1. An example of manganese partitioning in bottom sediment
samples using partial chemical extractions

Active Hydrothermal Area Inactive Hydrothermal Area

Concentration Percentage Concentration Percentage

Remaining of Total Remaining of Total
Extraction (mg/kg) Removed (mg/kg) Removed

Bulk 1,400 1,200

Acetic acid 1,300 7 1,100 8

Hydroxylamine 600 50 900 17

Hot HC1 200 29 200 58

47
3.0 CHEMICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

3.8 Chemical Partitioning by Density Gradient and

Mineralogy

Another method employed to determine chemi- highest concentrations of the metals investigated,
cal partitioning uses physical separations, determina- while the heaviest (most dense) fraction of sample D
tion of mineralogy, and subsequent chemical analysis contains the highest concentrations. This is true de-
(Muller and Burton, 1965; Francis and others, 1972; spite the fact that in both cases, these density subfrac-
Pilkington and Warren, 1979; Dossis and Warren, tions constitute only minor components of each sam-
1980). Table 3.8-1 shows an example of the results of ple (8.5 percent for A, and 1.3 percent for D).
this type of procedure. Each sample was broken down Because the mineralogy of each density subfraction is
into four size fractions by settling in acetone (Pilking- known, establishing chemical partitioning is much
ton and Warren, 1979). Each size fraction was further more direct and certain than with partial chemical
subdivided into density subfractions by heavy liquid extractions.
flotations using tetrabromoethane (bromoform). Each This type of procedure also has drawbacks. It is
density subfraction was subjected to mineralogical not very useful when there is a large percentage of fine
analysis by x-ray diffraction or differential thermal particles, because of difficulties involved in carrying
analysis. Finally, each size or density subfraction was out density separations on material finer than 10 (Jim
chemically analyzed. The data in table 3.8-1 are the (Pilkington and Warren, 1979). There is no assurance,
chemical results from two samples, A and D, and without individual testing, that the acetone or bromo-
represent the six density subfractions of the 1,000-to- form will not alter the chemistry. Finally, the method
10-(Jim size separation of each. It is apparent that the is extremely labor-intensive and consequently expen-
lighter (least dense) fraction of sample A contains the sive to carry out.

48
Table 3.8-1. Density gradients and chemical analysis
[Data from Dossis and Warren, 1980]
[Procedure: 1. Sizing in acetone (> 1,000, 1,000 to 10, 10 to 1, and <1 |xm)
2. Density separations using bromoform (<2.20, 2.20 to 2.55, 2.55 to 2.66, 2.66 to
2.75, 2.75 to 2.95, and >2.95 g/cc) 3. Phase identification by x-ray diffraction or
differential thermal analysis 4. Chemical analysis of each density subfraction]

Concentration of Metals by Density Subfraction

Weight Zinc in Lead in Cadmium in
Density Distribution Zinc Lead Cadmium Fraction Fraction Fraction
Sample Subfraction (percent) (p-g/g) (M-g/g) (M-g/g) (percent) 1 (percent) 1 (percent) 1
AS2 Dl 8.5 486 436 22 37 43 38
D2 6.5 204 166 9.4 12 13 13
D3 45.7 74 45 1.9 30 24 18
D4 20.3 85 61 5.4 15 15 23
D5 19.0 35 23 1.9 6 5 8
D6 0.1

DS2 Dl 2.3 1,580 1,370 25.3 3 4 4

D2 27.5 1,170 1,140 22.7 13 43 38
D3 61.6 183 168 1.7 5 14 7
D4 3.6 648 992 4.1 1 5 1
D5 3.6 740 1,100 7.3 1 6 2
D6 1.3 148,000 15,000 607 79 28 49
'Percent in fraction calculated by multiplying the concentration by the weight distribution for each
subfraction, summing, and then dividing each fractional contribution by the total.

49
3.0 CHEMICAL FACTORS AFFECTING SEDIMENT-TRACE METAL CONCENTRATIONS Continued

3.9 Chemical Partitioning by Mathematical ments from a number of estuaries. Their statistical
Manipulation or Modeling work assessed the utility of the various extractions in
determining phase associations and bioavailability.
Thus far, the discussion of chemical partition- Substrate (phase) characterization seems best accom-
ing has concentrated on direct determination by either plished by iron and manganese, and humic acid ex-
chemical or instrumental means. Attempts also have tractions. Extractable Fe phases seem more important
been made to determine chemical partitioning indi- than total Fe for the concentration of Ag, Cd, Cu, Pb,
rectly, using various statistical manipulations and by and Zn, while humic material is also important for the
constructing mathematical models. concentration of Ag and Cu. The results of this study
The statistical manipulation of chemical data support the view that the concentration of metals is a
can be used to clarify the processes by which various dynamic process in which various substrates
trace metals are partitioned on or in seston and bottom "compete" for different constituents, and that the rela-
sediments, especially when dealing with large data tive concentrations of differing substrates strongly in-
sets (Jones and Bowser, 1978). Statistical treatment of fluence partitioning (Oakley and others, 1980; 1981;
chemical data, or chemical and mir^ralogical data, Benjamin and Leckie, 1981). Although not specifi-
ranges from the calculation of relatively simple corre- cally stated, partitioning can change with varying
lation coefficients through highly complex cluster and physicochemical conditions (changes in Eh, pH, sol-
multivariate techniques. For example, Rossman and ute concentrations, solute form, and so forth) and with
others (1972) used correlation coefficients to partition transport and mixing processes, because of the addi-
bulk chemical data for ferromanganese nodules from tion, dilution, or removal of available substrates.
Green Bay. Because iron and manganese are the Statistical manipulations of sediment-chemical
major elements in the samples and appear to form two data pose some problems. For example, when the
distinct phases, they show strong negative correla- procedures used successfully by Leinen and others
tions. Additional calculations, using chemical data on (1980), Pisias and Leinen (1980), and Leinen and
other elements, indicate that they partition with either Pisias (1984) on the Nazca Plate sediments were ap-
the iron or manganese phases. The Fe phase appears plied to relatively homogeneous sediments from other
to concentrate Si and PO4 while the Mn phase appears Pacific areas, they produced much more ambiguous
to concentrate Ba, Ni, Co, Mo, Mg, and Sr. results. Another problem associated with this type of
Leinen and others (1980), Pisias and Leinen technique is also associated with partial chemical ex-
(1980), and Leinen and Pisias (1984) used an ex- tractions: a lack of sufficient specificity to identify
tended version of Q-mode factor analysis (a method phases positively (accomplish partitioning); identifi-
for determining associations among samples) and lin- cation is by inference. However, the greatest diffi-
ear programming to partition (identify phases of) culty arising from the interpretation of the statistical
chemical data from sediments collected on the Nazca treatment of sediment-chemical data is that it must be
Plate in the Pacific Ocean. The procedure identified made with a view to real processes and relations
five geochemically and mineralogically separate end (Jones and Bowser, 1978). This becomes increasingly
members (groups), which were termed (1) detrital, important as the statistical manipulations become
(2) hydrothermal, (3) biogenic, (4) authigenic hy- more complex. The manipulations may produce
drogenous, and (5) biogenic residual. Gibbs (1967; statistically significant results, but they can not be
1977) used multiple regression analysis to evaluate the ascribed to a rational environmental setting or process;
major transporting mechanisms for metal constituents in such cases they serve no useful purpose. Remem-
associated with seston in the Amazon River. The ber, applied statistics are an empirical tool used to
major transport phases and their relative importance clarify and evaluate processes and do not represent an
for carrying transition metals are, in decreasing order, end in themselves (Jones and Bowser, 1978).
metals held in crystal lattices equal to metals associ- According to Luoma and Bryan (1981), mathe-
ated with hydroxide coatings; metals associated with matical models designed to predict sediment chemical
organic matter; and metals held as adsorbates on partitioning may represent the most viable approach to
minerals. the problem. Others share this view (Leinen and
Luoma and Bryan (1981) applied statistical others, 1980; Swallow and Morel, 1980; Oakley and
techniques (linear regressions on log-transformed others, 1980; Pisias and Leinen, 1980; Benjamin and
data) to assess the "competition" between various sub- Leckie, 1981; Oakley and others, 1981). The develop-
strates for different metals using data obtained from ment of mathematical models and their application to
various partial chemical extractions on bottom sedi- natural environments for determining and predicting

50
 sediment-trace metal partitioning depend on a number substances, and hydrous iron and manganese oxides,
of factors like thermodynamic data, results of partial in seawatei. Laboratory experiments showed that clay
extraction studies, ability to identify and quantify sub- was the dominant sink for copper except where humic
strates, development of constants that describe the substances constituted more than 90 percent of the
strength and stability range of metal-substrate bind- substrate; clay was also the dominant sink for cad-
ing, quantification of metal speciation in various mium, except where hydrous iron and manganese ox-
strength solutions, Eh, pH, and so forth. ides constituted more than 90 percent of the substrate.
Swallow and Morel (1980) studied the behavior Also, clay was the dominant sink for both copper and
of copper and lead in the presence of hydrous iron cadmium when it constituted more than 60 percent of
oxides in artificial seawater and under varying pH the substrate. On the basis of these laboratory studies,
conditions. It was found that adsorption of both metals an attempt was made to predict bioavailability through
increased with increasing pH. Further, it seemed that the construction of a mathematical model; it was
other factors such as the ionic strength and back- tested in the laboratory using a polychaete worm that
ground electrolyte levels of the solution, the aging of fed on the substrate. The results were somewhat am-
the iron oxide substrate, and the precipitation of the biguous; nevertheless, the authors stated that they
oxides in the presence or absence of the metals in were sufficiently satisfied to conclude that the
question, had no significant effect. Although the be- bioavailability of copper and cadmium from geo-
havior of copper and lead with oxides in seawater is chemical phases (clay, humic acid, hydrated oxide)
somewhat analogous to surface complexation, the ox- could be more significant than the bioavailability of
ides are better viewed as a three-dimensional, metal- copper and cadmium dissolved in the seawater, at
absorbing "sponge" that accommodates metal ions natural concentrations.
into the solid as they hydrolize. These conclusions Mathematical modeling shows some promise
were based on experiments carried out in the labora- for determining the processes involved and for pre-
tory; a mathematical model (expression) was devel- dicting metal partitioning in aqueous-sedimentary sys-
oped to predict copper and lead behavior experimen- tems. However, a number of problems are associated
tally determined in the presence of iron oxides. with the models. The experiments and calculations on
Davis and Leckie (1978) also studied the uptake which the models were based were carried out using
of copper and silver by iron oxides. As did Swallow laboratory-generated or purified substrates to elimi-
and Morel (1980), they found that uptake varied with nate interpretational ambiguities; even in multiphase
pH. They also determined that complexing ligands systems (Oakley and others, 1980; 1981), the sub-
strongly affect trace metal adsorption and may en- strates were relatively simple compared to the natural
hance or depress the concentration capabilities of the environment. Also, we lack adequate data on the iden-
iron oxides. Some ligands, which bind to the surface tity and structure of organic constituents and on the
of the oxides, increase their collection efficiency, stability constants for many metal-ligand, ligand-
while other ligands remain in solution and compete substrate, and metal-substrate interactions in aquatic
with oxide surfaces for metals, thus decreasing collec- systems (Vuceta and Morgan, 1978). The results of
tion efficiency. These results suggest that the distribu- some modeling studies appear to contradict each
tion of metals in aquatic systems may be controlled other: for example, the results from Swallow and
more by surface binding on sediments coated with Morel (1980) compared with those of Vuceta and
organic matter than by those coated with iron oxides. Morgan (1978) and Benjamin and Leckie (1981). Fi-
Using purely mathematical models, this result has nally, some of the reported results from modeling
been confirmed for several metals including Cu, Pb, appear to contradict direct observations. For example,
Cd, Co, Ni, and Zn (Vuceta and Morgan, 1978; Swallow and Morel (1980) assert that adsorption of
Benjamin and Leckie, 1981). This is also in agree- metals by hydrous iron oxides is independent of the
ment with partial extraction studies (Forstner, presence or absence of the metals when oxide precip-
1982a,b). Further, Benjamin and Leckie (1981) indi- itation occurs. The suggested mechanism for metal
cate that metal adsorption from aqueous solutions also concentration by hydrous iron oxides, when studied in
depends on adsorbent and adsorbate concentration and the natural environment (manganese nodules), has al-
on the speciation of metals in solution. most invariably been coprecipitation (Cronan, 1974).
Using copper and cadmium, Oakley and others This conclusion is supported by partial extraction
(1980; 1981) attempted to develop a mathematical studies (Chester and Hughes, 1966; 1967; Forstner,
model capable of predicting partitioning and, ulti- 1982a) and by direct examination of nodule layers
mately, bioavailability. The study employed a ternary (Cronan, 1974).
substrate system composed of clay (bentonite), humic

51
4.0 SEDIMENT-TRACE METAL DATA MANIPULATIONS

4. 7 Corrections for Grain-Size Differences 100

100-percent of size fraction greater than 20 |j,m

The discussion of physical and chemical parti- Then this product gives the normalized chemi-
tioning showed that many factors affect the distribu- cal data (in mg/kg):
tion and concentration of trace metals associated with
both suspended and bottom sediments. Hence it can (dilution factor) X (chemical concentration in mg/kg)
be very difficult to determine the presence or absence
of significant distribution patterns from a point As an example, assume a sediment has 20 per-
source, for example. However, certain mathematical cent less than 20-|xm grain size and a zinc concentra-
or graphic manipulations of chemical data may help tion of 100 mg/kg.
the investigator determine if significant patterns do Dilution factor:
exist.
Figure 4.1-1 shows the relation between grain 100 = 100
size and chemistry. As pointed out earlier, metals tend 100-80 20
to concentrate in the finer size ranges (lesfthan 16 or
20 |xm in diameter), and the addition of larger grained Then the normalized chemical data are
material can be viewed as a diluent. Such diluents can
readily hide a significant pattern (de Groot and others, (5)x( 100)-500 mg/kg Zn
1982). One way to deal with the dilution problem is to
separate out the less than 20-|xm (16-|xm) fraction and
chemically analyze it. However, a reasonably good Remember that the underlying assumption for
result may be obtained by determining the less than this type of correction is that all of the constituent of
20-|xm (16-(Jim) percentage in the bulk sample of each interest is found in the less than 20-|xm (16-(JLm) frac-
sediment and normalizing the chemical data to it. The tion. This is almost certainly not true, but it does help
mathematics are very simple and straightforward. to clarify a distribution pattern that has been obscured
The dilution factor is calculated thus: because of the presence of varying amounts of diluent.

52
 32

28

24

Manganese x 100, in mg/kg ^

20

1 16
LU
O
Z
o
o
12

Mercury x± in mg/kg _

20 40 60 80 100
GRAIN SIZE LESS THAN 16 yum, IN PERCENT

Figure 4.1-1. Relation between metal concentration and grain size for the River Ems, FRG. (Data
from deGroot and others, 1982)

53
4.0 SEDIMENT-TRACE METAL DATA MANIPULATIONS Continued

4.2 Carbonate Corrections

Another commonly used correction factor is the The dilution factor is then
normalization of data to a carbonate-free basis. As
with the grain-size correction, the assumption is made 100 = 100 =1
that the carbonate fraction does not contain any of the 100-25 75
inorganic constituents of interest. Data in figure
4.2-1, for a marine bottom-sediment sample, certainly The normalized chemical data are thus:
support this view, with the possible exception of lead
and cadmium. As with the grain-size example, the (1.33)x(50)=66.7 mg/kg Zn
carbonate fraction could be physically separated and
chemically analyzed, but with extreme difficulty. The As an example of how a carbonate correction
usual procedure is to determine the carbonate content can clarify a pattern, examine the data in figure 4.2-2.
and then normalize the data accordingly. The mathe- The dashed line represents uncorrected data and the
matics for normalizing chemical data are simple and solid line the corrected data. The difference between
straightforward. the two patterns is striking. However, bear in mind
The dilution factor is given by: that this type of correction assumes that all of the
constituents of interest are concentrated in the noncar-
_____100______
100 percent carbonate bonate fraction. This is almost certainly not true for
every situation or trace metal, as the data in figure
Then the normalized chemical data are the 4.2-3 show. The pattern for calcium which represents
product (in mg/kg): the dominant form of carbonate (CaCO3 ), is the same
as those for other constituents. Thus, before applying
(dilution factor) X (chemical concentration in mg/kg) any correction to chemical data, an investigator must
be certain that the underlying assumption upon which
As an example, assume a sediment has 25 per- the correction is based is valid or, at least, fairly rea-
cent carbonate, and a zinc concentration of 50 mg/kg. sonable.

54
 106

10 5

oi 104
o
o
DC
LLJ
Q.
CO

DC
g 103

o
o

10° _L I _L
< 3.5 3.5-6.3 6.3-10 10-19.5 19.5-32 32-63 63-125 125-250
GRAIN SIZE. IN MICROMETERS
EXPLANATION

AI 2 O 3 aluminum TiO2 trtanium Cr chromium

Fe iron Mn manganese Cu copper

Ca x - calcium Zn zinc Li lithium

Mg magnesium Pb lead Cd cadmium

Figure 4.2-1. Variation in metal content with grain size from the same bottom-sediment sample.

55
 CONCENTRATION, IN MILLIGRAMS PER KILOGRAM
-k -k N>
*» 00 N> O> O
O O O O O
O O O O O .8
1 "
H K
?B ^J
^J
OQ
-I Ii ~
f\i I

"i- ^
1/1 TO
K)
I" I I
hJ _k
Q 01

ffi (D CO
(D O
a~ Z
c O
5' m i
"_ 01
- O S
T3 D O o^ z
(W ^ * N3'

5.1 3 n
^D (JQ
- c
9
HK O
QJ ~^ fl0
-; fg CU
fD ^ .' g
c
(D CO _^
HK 01
O
3-
(D

I
<
 105

104

CC
O
o
CC
LLI
10 3

QC
O

cc
H-
Z
LLI
O
o
o

< 3.5 35-6.3 6.3-10 10-19.5 19.5-32 32-63 63-125

GRAIN SIZE, IN MICROMETERS

EXPLANATION

Al 2 03 aluminum TiO 2 titanium Cr chromium

Fe iron Mn manganese Cu copper

Ca x r calcium Zn zinc Li lithium

Mg magnesium Pb lead Cd cadmium

Figure 4.2-3. Variation in metal content with grain size from the same bottom-sediment sample.

57
4.0 SEDIMENT-TRACE METAL DATA MANIPULATIONS Continued

4.3 Normalization to "Conservative" Elements

One final commonly employed method of nor- However, unlike the grain-size and carbonate
malizing bulk sediment data entails a comparison with corrections, the resulting values are simply ratios
so-called conservative elements. These elements are rather than chemical concentrations. This makes com-
assumed to have had a uniform flux from crustal-rock parison with data from other areas difficult, unless the
sources to the sediments over a long period of time; data are similarly recalculated; also, a ratio rather than
consequently, compensation for changes in levels of a concentration is conceptually more difficult to
various diluents can be made (Forstner and Wittmann, grasp. Nevertheless, normalizing to a conservative
1979). The most commonly used conservative ele- element can significantly clarify distribution patterns,
ments are aluminum (Piper, 1973; Bruland and others, as shown in figure 4.3-1. The data in that figure are
1974), and titanium (Forstner and Wittmann, 1979). from a series of marine sediment samples. The dashed
The mathematics of this type of correction is a simple line represents bulk copper data corrected to a carbonate-
ratio: free basis, while the solid line depicts the same bulk
copper data normalized to aluminum (as A12O3).
Concentration of inorganic constituent
Concentration of conservative element

58
 500

g
O

ft*»

8!
CO
300
oc
o Carbonate Free Basis

". 200
z
o

Z
o 100
8 Uncorrected Data

_L
-100 -80 -60 -40 -20 0 20 40 60 80 100 120 140

DISTANCE, IN ARBITRARY UNITS

Figure 4.3-1. Distribution of copper from a point source. (Data from Piper, 1973)

59
5.0 SAMPLING AND DATA HANDLING GENERAL CONSIDERATIONS

Several final points need to be discussed. Some 4. Fixed transect sampling at fixed and predeter-
are not solely related to sediment chemistry but apply mined sites that need not be at constant intervals.
to the subject, and they should be considered when This method of sample design, while extremely
planning, carrying out, and interpreting the results of simple, has this drawback: Because not all possible
water-quality investigations. sampling sites have an equal chance of being se-
Sampling and program design are extremely lected, any statistical inferences or conclusions are
important to the ultimate success of a study. More associated only with the selected sites, and areal
time, effort, and money have been wasted in environ- conclusions may not be valid.
mental studies because of poor program design than
for any other reason (Keith and others, 1983; The computer has been a boon to scientific in-
Hakanson, 1984). The decision on how, where, and vestigators because it permits the handling of large
when to sample must be predicated on what an inves- data sets and the application of many complex statisti-
tigator is trying to discover and how the data are to be cal manipulations. This can also be viewed as a curse,
used. For example, if a large areal study is needed, a because many investigators use statistical manipula-
type of random sampling should be used to character- tions as a replacement for thinking about what their
ize distributions. If random sites are not carefully se- data are telling them. Remember that statistically sig-
lected but just picked, it is possible that only interpre- nificant results must be matched by realistic environ-
tations based on that specific set of sites can be made, mental interpretations; otherwise, all the work that has
and generalizations about the entire area will be diffi- gone into a study can be wasted. Before embarking on
cult or impossible. It is a good idea to obtain statistical a major statistical program, understand what the im-
help in developing a program plan if an investigator plications of each manipulation are. In fact, it is more
lacks expertise in this area. It is fairly safe to say, advisable to develop a theory or hypothesis and test its
although statistical proof is lacking, that more errors statistical validity than to manipulate data statistically
are introduced into a study by improper sampling de- and then try to fashion a theory to fit the statistics.
sign and handling than could ever be attributed to Statistics is a tool, not an answer.
analytical errors in the laboratory. Understand and appreciate the limits of labora-
Of the many types of sampling programs, four tory analytical data. Try to develop a feeling for ana-
generally used procedures are discussed here. lytical precision and accuracy. Precision means repro-
ducibility; accuracy means how close the number is to
the true value. Do not overinterpret analytical results.
1. Simple random sampling the selection of sites For example, if analytical accuracy is ±10 percent
such that every possible sample has an equal op- and all the results differ by ±5 percent, then to all
portunity of being chosen. This method is very intents and purposes, the results are the same.
efficient in homogeneous study areas. However, it Use and build upon the work of others. Al-
may be ineffective in heterogeneous areas and though no one receives credit for re-inventing the
could lead to overlooking important sites or data. wheel, there is nothing wrong with using a successful
2. Stratified random sampling dividing a heteroge- investigator's sampling and interpretation program
neous area into homogeneous subareas within that has been applied elsewhere. You might save time,
which sampling locations are randomly selected. initially, by ignoring previous work, during the plan-
This often permits the elucidation of subtle but real ning of a study or during data analysis and report
differences. preparation. However, most studies do not produce
3. Systematic sampling establishing a constant in- new advances. Good science tends to be built upon
terval of distance between sampling sites deter- much retesting and repetition; occasionally, advances
mined by the number of samples to be collected. are made, and, rarest of all, a giant step occurs. Re-
The initial sampling site is randomly determined. member, the purposes of good science and of most
The advantage of such a scheme is its ease of studies are to answer specific questions as capably as
application; the disadvantage is that it can produce possible within time and financial constraints not to
very biased results. find the ultimate solution.

60
SELECTED REFERENCES

Ackermann, F., Bergmann, H., and Schleichert, U., 1983, 1974, History of metal pollution in the southern Cali-
Monitoring of heavy metals in coastal and estuarine fornia coastal zone: Environmental Science and Tech-
sediments a question of grain size: <20 jxm versus nology, v. 8, p. 425-432.
<60 jjim: Environmental Technology Letters, v. 4, *Bunzl, K., Schmidt, W., and Sanson, B., 1976, Kinetics
p. 317-328. of ion exchange in soil organic matter, IV, adsorption
Armstrong, F., 1958, Inorganic suspended matter in sea- and desorption of Pb +2 , Cu+ 2 , Cd +2 , Zn +2 , and Ca+ 2 ,
water: Journal of Marine Research, v. 17, p. 23-25. by peat: Journal of Soil Science, v. 27, p. 32-41.
*Arrhenius, G., and Korkish, J., 1959, Uranium and tho- Burton, J., 1978, The modes of association of trace metals
rium in marine minerals: First International Oceano- with certain compounds in the sedimentary cycle, in
graphic Congress, American Association for the Ad- Goldberg, E., ed., Biogeochemistry of estuarine sedi-
vancement of Science, preprints, 497 p. ments, Paris, UNESCO Press, p. 33-41.
Ashley, H., 1979, Particle size analysis by elutriation and *Buser, W., and Graf, F., 1955, Differenzierung von man-
centrifugation as exemplified by the Bahco Microparti- gan (Il)-manganit and 8-MnC>2 durch oberflachenmes-
cle Classifier, in Stockham, J., and Fachtman, E., sung nach brunauer-emmet-teller: Helvetica Chimica
eds., Particle size analysis: Ann Arbor, Mich., Ann Acta, v. 38, p. 830-842.
Arbor Press, p. 101-110. Calmano, W., and Forstner, U., 1983, Chemical extraction
* Aston, S., Bruty, D., Chester, R., and Padgham, R., of heavy metals in polluted river sediments in central
1973, Mercury in lake sediments. A possible indicator Europe: The Science of the Total Environment, v. 28,
of technological growth: Nature (London), v. 241, p. 77-90.
p. 450-451. Calmano, W., Wellershaus, S., and Forstner, U., 1982,
Aulio, K., 1980, Accumulation of copper in fluvial sedi- Dredging of contaminated sediments in the Weser estu-
ments and yellow water lilies (Nuphar lutea ) at varying ary: chemical forms of some heavy metals: Environ-
distances from a metal processing plant: Bulletin of mental Technology Letters, v. 3, p. 199-208.
Environmental Contamination and Toxicology, v. 25, *Chao, T., 1972, Selective dissolution of manganese oxides
p. 713-717. from soils and sediments with acidified hydroxylamine
Benjamin, M., Hayes, K., and Leckie, J., 1982, Removal hydrochloride: Soil Science Society of America Pro-
of toxic metals from power-generation waste streams ceedings, v. 36, p. 764-768.
by adsorption and coprecipitation: Journal of the Water *1984, Use of partial dissolution techniques in geo-
Pollution Control Federation, v. 54, p. 1472-1481. chemical exploration: Journal of Geochemical Explo-
*Benjamin, M., and Leckie, J., 1981, Conceptual model ration, v. 20, 101-136.
for metal-ligand-surface interactions during adsorp- Chao, T., and Zhou, L., 1983, Extraction techniques for
tion: Environmental Science and Technology, v. 15, selective dissolution of amorphous iron oxides from
p. 1050-1057. soils and sediments: Soil Science Society of America
*1982, Effects of complexation by Cl, SO4 , and S2O3 Journal, v. 47, p. 225-232.
on adsorption behavior of Cd on oxide surfaces: Envi- Chapman, P., Romberg, G., and Vigers, G., 1982, Design
ronmental Science and Technology, v. 16, p. 162-170. of monitoring studies for priority pollutants: Journal of
Berner, R., 1981, A new geochemical classification of sed- the Water Pollution Control Federation, v. 54, p. 292-
imentary environments: Journal of Sedimentary Petrol- 297.
ogy, v. 51, p. 359-365. Charm, W., 1967, Freeze drying as a rapid method of disag-
Bostrom, K., Burman, J., Ponter, C., and Ingri, J., 1981, gregating silts and clays for particle size analysis: Jour-
Selective removal of trace elements from the Baltic by nal of Sedimentary Petrology, v. 37, p. 970-971.
suspended matter: Marine Chemistry, v. 10, p. 335- *Chen, K., Gupta, S., Sycip, A., Lu, J., Knezevic, M., and
354. Choi, W., 1976, The effect of dispersion and resedi-
*Brannon, J., Engerl, R., Rose, J., Hunt, P., and Smith, I., mentation on migration of chemical constituents during
1976, Distribution of Mn, Ni, Zn, Cd, and As in sed- open water disposal of dredged material: Contract Re-
iments and the standard elutriate: U.S. Army Engineers port, U.S. Army Engineers Waterways Experimental
Waterways Experimental Station, Environmental Ef- Station, Vicksburg, Miss. 221 p.
fects Laboratory, Miscellaneous Papers, D 76-18, *Chester, R., 1965, Geochemical criteria for differentiating
38 p. reef from non-reef facies in carbonate rocks: American
Breward, N., and Peachey, D., 1983, The development of Association of Petroleum Geologists Bulletin, v. 49,
a rapid scheme for the elucidation of the chemical p. 253-276.
speciation of elements in sediments: The Science of the *Chester, R., and Hughes, M., 1966, The distribution of
Total Environment, v. 29, p. 155-162. manganese, iron, and nickel in a North Pacific deep sea
*Bruland, K., Bertine, K., Koide, M., and Goldberg, E., clay core: Deep Sea Research, v. 13, p. 627-634.
* 1967, A chemical technique for the separation of
ferromanganese minerals, carbonate minerals, and ad-
*References cited in the text or diagrams are marked with an aster- sorbed trace elements from pelagic sediments: Chemi-
isk(*). cal Geology, v. 2, p. 249-262.

61
* 1969, The trace element geochemistry of a north trifugation; consequences for the determination of
Pacific pelagic clay core: Deep Sea Research, v. 16, some heavy metals: Estuarine and Coastal Marine Sci-
p. 639-654. ence, v. 11, p. 701-707.
*Chester, R., and Messiah-Hanna, R., 1970, Trace element *Feltz, H., 1980, Significance of bottom material data in
partition patterns in North Atlantic deep sea sediments: evaluating water quality, in Baker, R., ed., Contami-
Geochimica et Cosmochimica Acta, v. 34, p. 2212- nants and sediments, v. 1, Ann Arbor, Mich., Ann
1128. Arbor Science Publishers, Inc., p. 271-287.
*Cronan, D., 1974, Authigenic minerals in deep sea sedi- *Filipek, L., Chao, T., and Carpenter, J., 1981, Factors
ments, in Goldberg, E., ed., The Sea, v. 5, New York, affecting the partitioning of Cu, Zn, and Pb in boulder
John Wiley, p. 491-525. coatings and stream sediments in the vicinity of a poly-
*Cronan, D., and Garrett, D., 1973, The distribution of metallic sulfide deposit: Chemical Geology, v. 33,
elements in metalliferous Pacific sediments collected p. 45-64.
by the DSDP: Nature Physical Science, v. 242, p. 88- *Filipek, L., and Owen, R., 1979, Geochemical associa-
89. tions and grain size partitioning of heavy metals in
*Davis, J., and Leckie, J., 1978, Effect of adsorbed com- lacustrine sediments: Chemical Geology, v. 26,
plexing ligands on trace metal uptake by hydrous ox- p. 105-117.
ides: Environmental Science and Technology, v. 12, *Forstner, U., 1982a, Chemical forms of metal enrichment
p. 1309-1315. in recent sediments, in Amstutz, G., and others, eds.,
*de Groot, A., Zshuppe, K., and Salomons, W., 1982, Ore genesis, New York, Springer-Verlag, p. 191-199.
Standardization of methods of analysis for heavy * 1982b, Accumulative phases for heavy metals in
metals in sediments: Hydrobiologia, v. 92, p. 689-695. limnic sediments: Hydrobiologia, v. 91, p. 269-284.
*Deurer, R., Forstner, U., and Schmoll, G., 1978, Selec- Forstner, U., andPatchineelam, S., 1980, Chemical associ-
tive chemical extraction of carbonate-associated metals ations of heavy metals in polluted sediments from the
from recent lacustrine sediments: Geochimica et Cos- Lower Rhine River, in Kavanaugh, M., and Leckie, J.,
mochimica Acta, v. 42, p. 415-427. eds., Particulates in water, Advances in chemistry se-
*Diks, D., and Alien, H., 1983, Correlation of copper dis- ries, v. 189, American Chemical Society, Washington,
tribution in a freshwater-sediment system to D.C., p. 177-193.
bioavailability: Bulletin of Environmental Contamina- Forstner, U., and Stoffers, P., 1981, Chemical fractionation
tion and Toxicology, v. 30, p. 37-43. of transition elements in Pacific pelagic sediments:
*Dossis, P., and Warren, L., 1980, Distribution of heavy Geochimica et Cosmochimica Acta, v. 45, p. 1141-
metals between the minerals and organic debris in a 1146.
contaminated marine sediment, in Baker, R., ed., Con- *Forstner, U., and Wittmann, G., 1979, Metal pollution in
taminants and sediments, v. 1, Ann Arbor, Mich., Ann the aquatic environment: New York, Springer-Verlag,
Arbor Science Publishers, Inc., p. 119-142. 486 p.
*Drever, J., 1982, The geochemistry of natural waters: En- *Francis, C., Bonner, W., and Tamura, T., 1972, An eval-
glewood Cliffs, N.J., Prentice-Hall, Inc., 388 p. uation of zonal centrifugation as a research tool in soil
*Duchart, P., Calvert, S., and Price, N., 1973, Distribution science, I, methodology: Soil Science Society of
of trace metals in the pore waters of shallow water America Proceedings, v. 36, p. 366-376.
marine sediments: Limnology and Oceanography, Francis, C., and Brinkley, F., 1976, Preferential adsorption
v. 18, p. 605-610. of Cs 137 to micaceous minerals in contaminated fresh-
Durum, W., and Haffty, J., 1963, Implications of minor water sediment: Nature (London), v. 260, p. 511-513.
element content of some major streams of the world: *Fripiat, J., and Gastuche, M., 1952, Etude physio-
Geochimica et Cosmochimica Acta, v. 27, p. 1-11. chimique des surfaces des argiles Les combinaisons
*Dymond, J., Corliss, J., Heath, G., Field, C., Dasch, J., de la kaolinite avec des oxides de fer trivalent:
and Veeh, H., 1973, Origin of metalliferous sediments Publications de L'lnstitute National Pour L'Etude
from the Pacific ocean: Geological Society of America Agronomique du Congo Beige, v. 54, p. 7-35.
Bulletin, v. 84, p. 3355-3372. *Gambrell, R., Khalid, R., Verloo, M., and Patrick, W.,
*Engler, R., Brannon, J., and Rose, J., 1974, A practical 1977, Transformations of heavy metals and plant nutri-
selective extraction procedure for sediment characteri- ents in dredged sediments as affected by oxidation-
zation: American Chemical Society, v. 168, Atlantic reduction potential and pH: U.S. Army Corps of Engi-
City, N.J., 17 p. neers, Vicksburg, Miss., Report D 77-4, 309 p.
Etchebar, H., and Jouanneau, J., 1980, Comparison of dif- *Ghosh, K., and Schnitzer, M., 1981, Fluorescence excita-
ferent methods for the recovery of suspended matter tion spectra and viscosity behavior of a fulvic acid and
from estuarine waters: deposition, filtration, and cen- its copper and iron complexes: Soil Science Society of
America Journal, v. 45, p. 25-29.
*Gibbs, R., 1967, Amazon River: environmental factors
*References cited in the text or diagrams are marked with an aster- that control its dissolved and suspended load: Science,
isk (*). v. 156, p. 1734-1737.

62
* 1973, Mechanisms of trace metal transport in *Irving, H., and Williams, R., 1948, Order of stability of
rivers: Science, v. 180, p. 71-73. metal complexes: Nature (London), v. 162, p. 746-
* 1977, Transport phases of transition metals in the 747.
Amazon and Yukon Rivers: Geological Society of *Jackson, M., 1958, Soil chemical analysis: Englewood
America Bulletin, v. 88, p. 829-843. Cliffs, N.J. Prentice-Hall, 498 p.
*Goldberg, E., 1954, Marine geochemistry I-chemical * 1979, Soil chemical analysis advanced course,
scavengers of the sea: Journal of Geology, v. 62, 2nd edition, Madison, Wis., published by the author,
p. 249-265. 898 p.
*Goldberg, E., and Arrhenius, G., 1958, Chemistry of *Jenne, E., 1968, Controls of Mn, Fe, Co, Ni, Cu, and Zn
Pacific pelagic sediments: Geochimica et Cosmochim- concentrations in soils and water: the significance of
ica Acta, v. 13, p. 153-212. hydrous Mn and Fe oxides: Advances in Chemistry
*Grim, R., 1968, Clay mineralogy, 2nd edition, New York, Series, v. 73, p. 337-387.
McGraw-Hill, 596 p. * 1976, Trace element sorption by sediments and
*Gupta, S., and Chen, K., 1975, Partitioning of trace soils sites and processes, in Chappell, W., and
metals in selective chemical fractions of near-shore Peterson, K., eds., Symposium on molybdenum, v. 2,
sediments: Environmental Letters, v. 10, p. 129-158. New York, Marcel-Dekker, p. 425-553.
*Guy, H., 1969, Laboratory theory and methods for sedi- *1981, Speciation of aqueous contaminants role of
ment analysis; U.S. Geological Survey Techniques of the geochemical model, in Proceedings of the NBS
Water Resources Investigations, book 5, chapter Cl, Workshop on Aqueous Speciation of Dissolved Con-
58 p. taminants, Gaithersburg, Md., CONF 810588-4,
*Hakanson, L., 1984, Sediment sampling in different NTIS, 14 p.
aquatic environments statistical aspects: Water Re- *Jenne, E., Kennedy, V., Burchard, J., and Ball, J., 1980,
sources Research, v. 20, p. 41-46. Sediment collection and processing for selective ex-
Hart, B., 1982, Uptake of trace metals by sediments and traction and for total metal analysis, in Baker, R., ed.,
suspended particulates: a review: Hydrobiologia, Contaminants and sediments, v. 2, Ann Arbor, Mien.,
v. 91, p. 299-313. Ann Arbor Science Publishers, Inc., p. 169-189.
Heath, G., and Dymond, J., 1977, Genesis and transforma- Jenne, E., and Luoma, S., 1977, Forms of trace elements in
tion of metalliferous sediments from the East Pacific soils, sediments, and associated waters: an overview of
Rise, Bauer Deep, and Central Basin, northwest Nazca their determination and biological availability, in
Plate: Geological Society of America Bulletin, v. 88, Wildung, R., and Drucker, H., eds., Biological impli-
p. 723-733. cations of metals in the environment, CONF-750929,
Hem, J., and Robertson, C., 1967, Form and stability of NTIS, Springfield, Va., p. 110-143.
aluminum hydroxide complexes in dilute solution: *Johnson, W., and Maxwell, J., 1981, Rock and mineral
U.S. Geological Survey Water Supply Paper, 1827-A, analysis, 2nd edition, New York, John Wiley, 489 p.
p. A24. *Jonasson, I., 1977, Geochemistry of sediment/water inter-
Hem, J., Robertson, C., Lind, D., and Polzer, W., 1973, actions of metals, including observations on availabil-
Chemical interactions of aluminum with aqueous silica ity, in Shear, H., and Watson, A., eds., The fluvial
at 25°C: U.S. Geological Survey Water Supply Paper, transport of sediment-associated nutrients and contami-
1827-E. nents, IJC/PLUARG, Windsor, Ontario, p. 255-271.
*Hirst, D., 1962, The geochemistry of modern sediments * Jones, B., and Bowser, C., 1978, The mineralogy and
from the Gulf of Paria: Geochimica et Cosmochimica related chemistry of lake sediments, in Lerman, A.,
Acta, v. 26, p. 1147-1187. ed., Lakes: chemistry, geology, physics, New York,
*Hirst, D., and Nicholls, G., 1958, Techniques in sedimen- Springer-Verlag, p. 179-235.
tary geochemistry, I, separation of the detrital and non- *Keith, L., Crummett, W., Deegan, J., Libby, R., Taylor,
detrital fractions of limestones: Journal of Sedimentary J., and Wentler, G., 1983, Principles of environmental
Petrology, v. 28, p. 468-481. analysis: Analytical Chemistry, v. 55, p. 2210-2218.
*Holmgren, G., 1967, A rapid citrate-dithionate extractable Kennedy, V., Zellwager, G., and Jones, B., 1974, Filter
iron procedure: Soil Science Society of America Pro- pore-size effects on the analysis of Al, Fe, Mn, and Ti
ceedings, v. 31, p. 210-211. in water: Water Resources Research, v. 10, p. 785-
*Horowitz, A., 1974, The geochemistry of sediments from 790.
the northern Reykjanes Ridge and the Iceland-Faeroes Keulder, P., 1982, Particle size distribution and chemical
Ridge: Marine Geology, v. 17, p. 103-122. parameters of the sediments of a shallow turbid im-
*Horowitz, A., and Cronan, D., 1976, The geochemistry of poundment: Hydrobiologia, v. 91, p. 341-353.
basal sediments from the North Atlantic Ocean: Marine *Kharkar, D., Turekian, K., andBertine, K., 1968, Stream
Geology, v. 20, p. 205-228. supply of dissolved Ag, Mo, Sb, Se, Cr, Ca, Rb, and
Cs to the oceans: Geochimica et Cosmochimica Acta,
*References cited in the text or diagrams are marked with an aster- v. 33, p. 285-298.
isk (*).

63
*Kitano, Y., Sakata, M., and Matsumoto, E., 1980, Parti- in Jenne, E., ed., Chemical modeling in aqueous sys-
tioning of heavy metals into mineral and organic frac- tems, ACS Symposium Series 93, American Chemical
tions in a sediment core from Tokyo Bay: Geochimica Society, Washington, D.C., p. 577-609.
et Cosmochimica Acta, v. 44, p. 1279-1285. 1979, Trace metal bioavailability: modeling chemi-
*Kononova, M., 1966, Soil organic matter, 2nd ed., cal and biological interactions of sediment-bound lead
Nowakowski, T., and Newman, A., translators, New to the estuarine bivalve Scrobicularia plana: Journal of
York, Pergamon Press, p. 377-419. the Marine Biological Association of the United King-
*Krauskopf, K., 1956, Factors controlling the concentration dom, v. 58, p. 793-802.
of thirteen rare metals in sea water: Geochimica et 1981, A statistical assessment of the form of trace
Cosmochimica Acta, v. 9, p. 1-32. metals in oxidized estuarine sediments employing
*Kuenen, Ph., 1965, Geological conditions of sedimenta- chemical extractants: The Science of the Total Envi-
tion, in Riley, J., and Skirrow, G., eds., Chemical ronment, v. 17, p. 165-196.
oceanography, v. 2, New York, Academic Press, Luoma, S. and Davis, J., 1983, Requirements for modeling
p. 1-22. trace metal partitioning in oxidized estuarine sedi-
Laxen, D., and Chandler, I., 1983, Size distribution of iron ments: Marine Chemistry, v. 12, p. 159-181.
and mangenese species in freshwaters: Geochimica et Luoma, S., and Jenne, E., 1976, Factors affecting the
Cosmochimica Acta, v. 47, p. 731-741. availability of sediment-bound cadmium to the estua-
*Lee, G., 1975, Role of hydrous metal oxides in the trans- rine deposit-feeding clam, Macoma balthica, in Radi-
port of heavy metals in the environment, in Krenkel, ological problems associated with the development of
P., ed., Heavy metals in the aquatic environment, New energy sources, Proceedings of the Fourth National
York, Pergamon Press, p. 137-147. Symposium on Radioecology, Corvallis, Oreg.,
*Leinen, M., and Pisias, N., 1984, An objective technique ERDA Report CONF-759503, Stroudsburg, Pa.,
for determining end member compositions and for par- Dowden, Hutchinson, and Russ Publishers, 16 p.
titioning sediments according to their sources: * 1977, Estimating bioavailability of sediment-
Geochimica et Cosmochimica Acta, v. 48, p. 47-62. bound metals with chemical extractants, in Hemphill,
*Leinen, M., Pisias, N., Dymond, J., and Heath, G., 1980, D., ed., Trace substances in environmental health,
Geochemical partitioning: application of an objective v. 10, Columbia, Mo., University of Missouri Press,
technique for end-member characterization: Geological p. 343-351.
Society of America Abstracts with Programs, v. 93, 1977, The availability of sediment-bound cobalt, sil-
p. 470 (abstract). ver and zinc to a deposit-feeding clam, in Wilding, R.,
Lion, L., Altmann, R., and Leckie, J., 1982, Trace metal and Drucker, H., eds., Biological implications of
adsorption characteristics of estuarine particulate mat- metals in the environment, CONF-750929, NTIS,
ter: evaluation of contributions of Fe/Mn oxide and Springfield, Va., p. 213-230.
organic surface coatings: Environmental Science and *Lynn, D., and Bonatti, E., 1965, Mobility of manganese
Technology, v. 16, p. 660-666. in diagenesis of deep sea sediments: Marine Geology,
Liu, B., Raabe, O., Smith, W., Spencer, H., Ill, and v. 3, p. 457-474.
Kuykendal, W., 1980, Advances in particle sampling *Malo, B., 1977, Partial extraction of metals from aquatic
and measurement: Environmental Science and Tech- environments: Environmental Science and Technol-
nology, v. 14, p. 392-397. ogy, v. 11, p. 277-282.
Loring, D., 1981, Potential bioavailability of metals in east- Martin, H., Wilhelm, E., Laville-Timsit, L., and Lecomte,
ern Canadian estuarine and coastal sediments: Rapports P., 1984, Enhancement of stream-sediment geochemi-
et Proces-Verbaux des Reunions Conseil International cal anomalies in Belgium and France by selective ex-
Pour 1'Exploration de la Mer, v. 181, p. 93-101. tractions and mineral separations: Journal of Geochem-
Luoma, S., 1983, Bioavailability of trace metals to aquatic ical Exploration, v. 20, p. 179-205.
organisms a review: The Science of the Total Envi- *Mero, J., 1962, Occurrence of Mn nodules: Economic
ronment, v. 28, p. 1-22. Geology, v. 57, p. 747-767.
Luoma, S., and Bryan, G., 1978, Factors controlling the *Mitchell, R., 1964, Trace elements in soil, in Bear, F.,
availability of sediment-bound lead to the estuarine ed., Chemistry of the soil, New York, Reinhold Pub-
bivalve Scrobicularia plana: Journal of the Marine lishing Corp., p. 320-368.
Biological Association of the United Kingdom, v. 58, *Moore, R., Meyer, R., and Morgan, C., 1973, Investiga-
p. 793-802. tion of the sediments and potential manganese nodule
*1979, Trace metal bioavailability: modeling chemi- resources of Green Bay, Wisconsin: University of Wis-
cal and biological interactions of sediment-bound zinc, consin Technical Report WIS-SG-73-213, Madison,
Wis., 144 p.
Mora, S., and Harrison, R., 1983, The use of physical
separation techniques in trace metal speciation studies:
*References cited in the text or diagrams are marked with an aster- Water Research, v. 17, p. 723-733.
isk (*). Mouvet, C., and Bourg, A., 1983, Speciation (including

64
 adsorbed species) of copper, lead, nickel, and zinc in *Piper, D., 1971, The distribution of Co, Cr, Cu, Fe, Mn,
the Meuse River: Water Research, v. 17, p. 641-649. Ni, and Zn in Framvaren, a Norwegian anoxic fjord:
*Muller, L., and Burton, C., 1965, The heavy liquid den- Geochimica et Cosmochimica Acta, v. 35, p. 531-550.
sity gradient and its application in ore dressing mineral- * 1973, Origin of metalliferous sediments from the
ogy: Proceedings of the eighth commonwealth mining East Pacific Rise: Earth and Planetary Science Letters,
and metallurgical congress, v. 6, Melbourne, Aus- v. 19, p. 75-82.
tralia, p. 1151-1163. *Pisias, N., and Leinen, M., 1980, Geochemical partition-
Musani, Lj., Valenta, P., Nurnberg, H., Konrad, Z., and ing of deep-sea sediments using an extended version of
Branica, M., 1980, On the chelation of toxic trace Q-mode factor analysis and linear programming: Geo-
metals by humic acid of marine origin: Estuarine and logical Society of America Abstracts with Programs,
Coastal Marine Science, v. 11, p. 639-649. v. 93, p. 500 (abstract).
Nembrini, G., Rapin, F., Garcia, J., and Forstner, U., *Porterfield, G., 1972, Computation of fluvial sediment
1982, Speciation of Fe and Mn in a sediment core of discharge: U.S. Geological Survey Techniques of
the Baie de Villefrance: Environmental Technology Water Resources Investigations, book 3, chapter C-3,
Letters, v. 3, p. 545-552. p. 43-47.
Nishida, H., Miyai, M., Tada, F., and Suzuki, S., 1982, *Rashid, M., 1974, Adsorption of metals on sedimentary
Computation of the index of pollution caused by heavy and peat humic acids: Chemical Geology, v. 13,
metals in river sediment: Environmental Pollution (Se- p. 115-123.
ries B), v. 4, p. 241-248. *Rex, R., and Goldberg, E., 1958, Quartz contents of
*Nissenbaum, A., 1972, Distribution of several metals in pelagic sediments of the Pacific Ocean: Tellus, v. 10,
chemical fractions of a sediment core from the Sea of p. 153-159.
Okhotsk: Israeli Journal of Earth Science, v. 21, Robinson, G., 1981, Adsorption of Cu, Zn, and Pb near
p. 143-154. sulfide deposits by hydrous manganese-iron oxide
*Nriagu, J., and Coker, R., 1980, Trace metals in humic coatings on stream alluvium: Chemical Geology,
and fulvic acids from Lake Ontario sediments: Envi- v. 12, p. 233-266.
ronmental Science and Technology, v. 14, p. 443-446. * 1982, Trace metal adsorption potential of phases
*Oakley, S., Nelson, P., and Williamson, K., 1981, Model comprising black coatings on stream pebbles: Journal
of trace-metal partitioning in marine sediments: Envi- of Geochemical Exploration, v. 17, p. 205-219.
ronmental Science and Technology, v. 15, p. 474-480. -1983, Heavy-metal adsorption by ferromanganese
*Oakley, S., Williamson, K., and Nelson, P., 1980, The coatings on stream alluvium: natural controls and im-
geochemical partitioning and bioavailability of trace plications for exploration: Chemical Geology, v. 38,
metals in marine sediments: Project completion report, p. 157-174.
OWRR Project No. A-044-ORE-WF3, October 1, *Rossman, R., Callender, E., and Bowser, C., 1972, In-
1977-September 31, 1979, Oregon State University, terelement geochemistry of Lake Michigan ferroman-
Corvallis, Oreg., 84 p. ganese nodules: Proceedings of the 24th international
*Oliver, B., 1973, Heavy metal levels of Ottawa and geological congress, Montreal, Section 10, p. 336-
Rideau River sediment: Environmental Science and 341.
Technology, v. 7, p. 135-137. Saar, R., and Weber, J., 1981, Lead (II) complexation by
Olsen, C., Cutshall, N., and Larsen, I., 1982, Pollutant- fulvic acid: how it differs from fulvic acid complexa-
particle associations and dynamics in coastal marine tion of copper (II) and cadmium (II): Geochimica et
environments: a review: Marine Chemistry, v. 11, Cosmochimica Acta, v. 44, p. 1381-1384.
p. 501-533. *Saxby, J., 1969, Metal-organic chemistry of the geochem-
*Patchineelam, S., 1975, Untersuchungen uber die haupt- ical cycle: Reviews of Pure and Applied Chemistry,
bindungsarten und die mobilisierburkeit von schwer- v. 19, p. 131-150.
metallen in fluviatilen sedimenten: unpublished Ph.D. Schalscha, E., Morales, M., Vergara, I., and Chang, A.,
dissertation, University of Heidelberg, 136 p. 1982, Chemical fractionation of heavy metals in
Patchineelam, S., and Forstner, U., 1977, Bildungsformen wastewater affected soils: Journal of the Water Pollu-
von schwermetallen in marinene sedimenten: Sencken- tion Control Federation, v. 54, p. 175-180.
bergiana Maritima, v. 9, p. 75-104. *Schmidt, R., Garland, T., and Wiluung, R., 1975, Copper
Pickering, W., 1981, Selective chemical extraction of soil in Sequim Bay sediments: Batelle Pacific Northwest
components and bound metal species: Critical Reviews Laboratory Annual Report, part 2, p. 136.
in Analytical Chemistry, v. 12, p. 233-266. Schnitzer, M., and Kerndorff, H., 1981, Reactions of fulvic
*Pilkington, E., and Warren, L., 1979, Determination of acid with metal ions: Water, Air, and Soil Pollution,
heavy metal distribution in marine sediments: Environ- v. 15, p. 97-108.
mental Science and Technology, v. 13, p. 295-299. *Schnitzer, M., and Kahn, S., 1972, Humic substances in
the environment, New York, Marcel-Dekker, 327 p.
*References cited in the text or diagrams are marked with an aster- * 1978, Soil organic matter, New York, Elsevier,
isk (*). p. 1-64.

65
Schultz, L., 1964, Quantitative interpretation of mineralog- *Summerhayes, C., Ellis, J., Stoffers, P., Briggs, S., and
ical composition from x-ray and chemical data for the Fitzgerald, M., 1976, Fine-grained sediment and in-
Pierre Shale: U.S. Geological Survey Professional dustrial waste distribution in New Bedford Harbor and
Paper 391-C, 31 p. western Buzzards Bay, Mass.: Woods Hole Oceano-
*Schwertmann, U., 1964, Differenzierung der eisenoxide graphic Institution Technical Report WHOI-76-115,
des bodens durch photochemische extraktion mit sauer 110 p.
ammonium-oxalat-losung: Zeitschrift fur Pflanzener- *Swallow, K., and Morel, F., 1980, Adsorption of trace
ahrung, Dungung mit Bodenkunde, v. 105, p. 194- metals by hydrous ferric oxides in seawater: Environ-
202. mental Protection Agency Report No. EPA 600/3-80-
*Singer, P., 1977, Influence of dissolved organics on the 011, Environmental Research Laboratory, Narragan-
distribution, transport, and fate of heavy metals in sett, R.I., 52 p.
aquatic systems, in Suffet, I., ed., Fate of pollutants in *Swanson, V., Frist, L., Rader, R., Jr., and Huffman, C.,
the air and water environment, part I, New York, John Jr., 1966, Metal sorption by northwest Florida humate:
Wiley, p. 155-182. U.S. Geological Survey Professional Paper 550-C,
Skei, J., and Paus, P., 1979, Surface metal enrichment and p. 174-177
partitioning of metals in a dated sediment core from a *Tessier, A., Campbell, P., and Bisson, M., 1979, Sequen-
Norwegian fjord: Geochimica et Cosmochimica Acta, tial extraction procedure for the speciation of particu-
v. 43, p. 239-246. late trace metals: Analytical Chemistry, v. 51, p. 844-
Skoog, D., and West, D., 1981, Principles of instrumental 851.
analysis, 2nd edition, Philadelphia, Pa., Saunders Col- *1982, Paniculate trace metal speciation in stream
lege Press, p. 452-457. sediments and relationships with grain size: implica-
Slavek, J., and Picketing, W., 1981, The effect of pH on tions for geochemical exploration: Journal of Geo-
the retention of Cu, Pb, Cd, and Zn by clay-fulvic acid chemical Exploration, v. 16, p. 77-104.
mixtures: Water, Air, and Soil Pollution, v. 16, p. 209- *Thomas, R., 1972, The distribution of mercury in the
221. sediment of Lake Ontario: Canadian Journal of Earth
Sposito, G., Lund, L., and Chang, A., 1982, Trace metal Science, v. 9, p. 636-651.
chemistry in arid-zone field soils phases: Soil Science *Thorne, L., and Nickless, G., 1981, The relation between
Society of America Journal, v. 46, p. 260-264. heavy metals and particle size fractions within the
*Stoffers, P., Summerhayes, C., Forstner, U., and Severn estuary (U.K.) inter-tidal sediments: The Sci-
Patchineelam, S., 1977, Copper and other heavy metal ence of the Total Environment, v. 19, p. 207-213.
contamination in sediments from New Bedford Harbor, van Valin, R., and Morse, J., 1982, An investigation of
Massachusetts: a preliminary note: Environmental Sci- methods commonly used for the selective removal and
ence and Technology, v. 11, p. 819-821. characterization of trace metals in sediments: Marine
Stover, R., Sommers, L., and Silviera, D., 1976, Evalua- Chemistry, v. 11, p. 535-564.
tion of metals in wastewater sludge: Journal of the *Vuceta, J., and Morgan, J., 1978, Chemical modeling of
Water Pollution Control Federation, v. 48, p. 2165- trace metals in fresh waters: role of complexation and
2175. adsorption: Environmental Science and Technology,
v. 12, p. 1302-1309.
Young, T., Depinto, J., and Seger, E., 1982, Transport and
fate of heavy metals in Onondaga Lake, N.Y.: Bulletin
*References cited in the text or diagrams are marked with an aster- of Environmental Contamination and Toxicology,
isk (*). v. 29, p. 554-561.

66
CONVERSION FACTORS
Multiply SI unit By To obtain inch-pound units

cm (centimeter) 0.3937 inch (in)

mm (millimeter) 0.03937 inch (in)
jxm (micrometer) 0.00003937 inch (in)
fxm2 (square micrometer) 0.00154997 in2 (square inch)
t (metric ton) 1.102 ton, short (t)
kg (kilogram) 2.2046226 pound, avdp (Ib)
g (gram) 0.035273962 ounce, avdp (oz)
mg (milligram) 0.000035273962 ounce, avdp (oz)
Hg (microgram) 3.5273962 + 10~ ounce, avdp (oz)
m2/g (meter squared per gram) 305.186 ft2/oz (foot squared per ounce)
cm2/g (centimeter squared per gram) 3.05186 in2/oz (inch squared per ounce)
L (liter) 33.81497 ounce, fluid (oz)
m3/s (cubic meter per second) 35.31 ft3/s (foot cubed per second)
fxg/kg (microgram per kilogram) 1.000 part per billion (ppb)
(microgram per liter) 1.000 part per billion (ppb)
(microgram per gram) 1.000 part per million (ppm)
mg/kg (milligram per kilogram) 1.000 part per million (ppm)
concentration of 1 percent 10,000 part per million (ppm)

>U.S. GOVERNMENT PRINTING OFFICE: 1986-491-405:20082

67