Reprinted from the Journal qf Environmental Quality
Volume 28, no. 4, July-Aug. 1999. Copyright © 1999, ASA, CSSA, SSSA
677 South Segoe Rd., Madison, WI 53711 USA
Baseline Concentrations of 15 Trace Elements in Florida Surface Soils
Ming Chen, Lena Q. Ma,* and Willie G. Harris
ABSTRACT
The objective of this study was to establish baseline concentrations
for 15 potentially toxic elements (Ag, As, Ba, Be, Cd, Cr, Cu, Hg,
Mn, Mo, Ni, Pb, Sb, Se, and Zn) based on 448 representative Florida
surface soils using microwave assisted HNOrHCI-HF digestion.
Baseline concentrations of those elements were (mg kg-I): Ag 0.072.50, As 0.02-7.01, Ba 1.67-112, Be 0.04-4.15, Cd 6-0.33, Cr 0.89-80.7,
Cu 0.22-21.9, Hg 0.00075--0.0396, Mo 0.13-6.76, Ni 1.70-48.5, Pb 0.6942.0, Sb 0.06-0.79, Se 0.01-1.11, and Zn 0.89-29.6, respectively. Upper
baseline values for most elements corresponded with these reported
in literature, except Ba, Hg, Mn, Sb, and Zn, which were 3 to 23 times
lower. Soil properties, including pH, organic carbon (OC), particle
size, cation-exchange capacity (CEC), available water, extractable
base, extractable acidity, total Ca, Mg, P, K, Fe, and AI concentrations,
were related to metal concentrations using factorial analysis. Eight
factors were identified (total Fe and AI, CEC, pH, clay, OC, total Ni
and Mo, total Sb and Pb, and total Hg) and accounted for 87% of
the total variance, suggesting that metal concentrations were primarily
controUed by soil compositions. Multiple regression of elemental concentrations against total Fe, total AI, clay, OC, CEC, and pH was
significant for aU elements. Partial correlation coefficients indicated
that total Fe and/or AI explained most of the variance for Mn, Ni,
Ba, Be, Hg, As, Cd, Cr, Cu, Mo, Pb, and Zn concentrations. Most
of the variance in Se was related to clay, whereas those of Ag and
Sb related to clay and total AI.
T
HE PRESENCE of potentially toxic metals in landapplied waste materials is of public concern. Federal
and state regulations list Ag, As, Ba, Be, Cd, Cr, Cu,
Hg, Mn, Mo, Ni, Pb, Sb, Se, and Zn as potentially toxic
elements (Florida Department of Environmental Protection, 1995; U.S. Environmental Protection Agency,
1996). The U.S. Environmental Protection Agency
(1996) has established risk-based soil screening levels
as a reference for site-specific cleanup for trace metals.
However, no federal regulation specifies the maximum
metal concentrations in non-hazardous wastes for land
Soil and Water Sciences Dep., Univ. of Florida, Gainesville, FL 326110290. Approved for publication as the Florida Agricultural Experiment Station Journal Series No. R-06229. Received 19 Mar. 1998.
*Corresponding author (Qma@gnv.ifas.ufl.edu).
Published in J. Environ. Qual. 28:1173-1181 (1999).
application, except in the case of sewage sludge (U.S.
Environmental Pt;otection Agency, 1995). Natural background concentrations of trace elements in soils where
these materials are to be applied can be used as a reference (Kabata-Pendias and Pendias, 1992). Unless a reliable database on concentrations of trace metals in soils
is available, inaccurate or unrealistically low mandatory
guideline levels may be set by regulators (Davies, 1992;
McGrath, 1986; Pierce et aI., 1982). Thus, it is important
to establish background concentrations of trace metals
for soils occurring within a region, and to document
systematic variation in concentrations according to soil
classes and properties.
Background measurement represents natural elemental concentrations in soils without human influence
(Kabata-Pendias et aI., 1992; Gough, 1993). This measurement depicts an idealized situation. Due to longrange transport of contaminants, truly pristine ecosystems may no longer exist, making establishing background concentrations a difficult task. For example,
background levels for Pb are commonly elevated due
to long-term usage of Pb-based gasoline and paint. Thus,
it is almost impossible to find a surface soil sample
completely free of Pb contamination (Fergusson, 1990).
The term geochemical baseline concentration is often
used to express an expected range of element concentrations around a mean in a normal sample medium. It is
not generally a true background concentration and is
defined as 95% of the expected range of background
concentration (Kabata-Pendias et aI., 1992; Dudka,
1993; Gough, 1993). Based on log- normal distribution
theory, the expected range can be expressed as the average of logarithms ::!:: 2 standard deviations (Dudka et
aI., 1995). Since it is becoming more and more difficult
to determine background levels of certain elements, the
baseline values have been recognized as the only means
to establish reliable worldwide elemental concentraAbbreviations: AM, arithmetic mean; ASD, arithmetic standard deviation; CEC, cation exchange capacity; FCSSP, the Florida Cooperative
Soil Survey Program; GM, geometric mean; GSD, geometric standard
deviation; OC, organic carbon.
1174
J. ENVIRON. QUAL., VOL. 28, JULY-AUGUST 1999
tions in natural materials (Gough et aI., 1988; KabataPendias and Pendias, 1992).
Baseline concentrations of many elements can be obtained for soils of the USA (Shacklette and Boerngen,
1984; Gough et aI., 1988, 1994; Ames and Prych, 1995),
China (Wei et aI., 1990), Great Britain (McGrath, 1986),
and other European countries (Dudka, 1993). Researchers pointed out that baseline concentrations were a better measure of the variation in trace element concentrations than the observed ranges (i.e., ranges of background concentrations) since the distorting effects of a
few high values were minimized by log-transformation
of the data (Dudka et aI., 1995). They recommended
the use of baseline concentrations as alternative criteria
for assessing possible trace element contamination in
soils (Gough et aI., 1994), or the use of the upper limit
of the baseline concentration range to assess the background concentration with an acceptable degree of confidence (Dudka et aI., 1995).
Unfortunately, existing data on baseline concentrations of trace elements in Florida soils are inadequate
for determining the issue of how clean is clean for cleaning up contaminated soils and how dirty is dirty for land
application of waste materials. Since only 40 soil samples
were used in a previous study (Ma et aI., 1997), a larger
soil sample pool and more systematic sampling strategy
is necessary to establish a comprehensive database for
baseline concentrations of potentially toxic elements in
Florida soils.
The present investigation was conducted to (i) establish baseline concentrations of the 15 potentially toxic
trace elements in 448 representative Florida surface soil
horizons; and (ii) investigate relationship among elements and between soil properties and elemental concentrations. Results of this research can be used as a
reference in assessing anthropogenic vs. natural levels
of trace elements in Florida soils.
MATERIAL AND MEmODS
Sample Selection and Characterization
Soils used in this study were sampled and characterized as
part of the Florida Cooperative Soil Survey Program (FCSSP)
conducted jointly by the University of Florida Soil and Water
Science Department and the USDA Natural Resources Conservation Service. Soil horizons were delineated and sampled
using USDA soil survey conventions and procedures (Soil
Survey Division Staff, 1993) as guidelines. Based on the mean
coefficient of variations from a previous study (Ma et aI.,
1997), a minimum of 214 soil samples are required to establish
a statistically valid database for Florida soils (with 95% confidence level and 20% accepted error).
In the present study, a total of 448 archived soil samples
were selected to assure both taxonomic and geographic representation. The overall taxonomic representation was achieved
by weighting the number of samples for each soil order by
their estimated areal occurrences in Florida. The total mapped
area is 11 265 532 ha and covers as much as 80% of the total
land area of Florida. Most of the mapped areas in Florida
are represented in the current study. Seven soil orders were
identified from 51 counties in Florida and their approximate
coverages are: Spodosols (28%), Entisols (22%), Ultisols
(19%), Alfisols (14%), Histosols (10%), Mollisols (4%), and
Inceptisols (3%). Based on the areal occurrence of each soil
order, the samples included surface horizons from 122 Spadosols, 107 Entisols, 90 Ultisols, 60 Alfisols, 39 Histosols, 17
Mollisols, and 13 Inceptisols.
Physical, chemical, and mineralogical analyses were previously determined through the FCSSP, include taxonomic
class, morphological information and mineralogy (Sadek et
aI., 1990). A statistical summary of selected properties for the
448 soils samples used in this study is presented in Table 1.
Sample Preparation and Trace Element Analysis
All soil samples were air dried, ground, and passed through
a 6O-mesh sieve. The screened samples were stored in polyethylene containers before analysis. Approximately 1 g of soil
sample was weighed into a 120-mL teflon pressure digestion
vessel; 9 mL of concentrated HN03, 4 mL of concentrated
HF, and 1 mL of concentrated HCI were then added. Samples
and reagents were well mixed, sealed, and digested in a CEM
MDS-2000 digestion microwave oven (CEM, Matthews, NC)
for 20 min at 120 psi. After cooling, 2 g of boric acid were
added to the digested solution to neutralize excess HF. For
Histosols rich in organic matter, only 0.5 g of sample was used
and 1.0 mL of HzO z was added prior to digestion. The fmal
volume of the digested solution was 100 mL after filtration
Table 1. Statistical summary of selected properties for the 448 soil samples used in this study.
Sand
Silt
Clay
Organic C
Cation exchange
capacity
pH-H2O
pH-KO
(emol kg-I)
0.28--375
21.9
51.0
10.2 3.35
2.70-8.10
5.04 0.97
4.94 1.21
2.08--8.30
4.24 :!: 1.11
4.11
1.28
Total AI
Total Fe
Range
AM
ASDt
GM
GSD:j:
- - - - - - - - - - gkg- I - - - - - - - - - 12.0-999
0-734
0-820
0.80-559
41.7 :!: 83.3
893=164
64.4 102
51.0 111
36.2 :!: 3.05
84S:!: 1.62
19.9 3.12
3.34
17.1
Range
AM
ASDt
GM:!: GSD:j:
- - - - - - - - - - - - - - - - - g kg- I - - - - - - - - - - - - - - - - 0.01-383
0-14.9
0-5.57
0-2.87
0.13--26.5
0.09-34.2
8.30 :!: 45.2
0.40 1.50
0.40 0.80
0.20 0.40
2.20 3.00
2.30
4.SO
0.308 :!: 6.76
4.98
3.33
0.045
0.185
3.87
0.088
1.41
2.41
1.01
3.24
Range
AM
ASDt
GM
GSD:j:
- - - - - - - - - - - - - cmol kg-I - - - - - - - - - - - - 0-271
0-10.2
0.01-254
0-221
0-84.3
2.48 :!: 18.0
1.01 :!: 6.43
0.84
2.18 :!: 8.07
0.22
13.8 26.8
0.08:!: 6.18
19.9:!: 3.12
4.42
0.06
5.56
0.29
3.13
6.32
=
=
=
Total Ca
=
=
=
Extr-Na
=
=
t Arithmetic mean
:j: Geometric mean
Total Mg
Extr-K
=
=
=standard deviation.
=geometric standard deviation.
=
Total K
=
=
Extr-Ca
=
=
Total P
=
=
Extr·Mg
=
=
=
=
=
=
=
=
=
Total acid
Avail-H2O
(em em-I)
0.02--0.72
0.16 0.12
0.82
1.96
=
=
=
=
=
1175
CHEN ET AL.: BASELINE CONCENTRATIONS OF TRACE ELEMENTS IN FLORIDA SOILS
based on the loading of the variables. Each factor contains
all variables but only variable with loadings above 0.50 was considered to be important for interpreting a factor (Dudka,
1992).
In addition, multiple regression analysis was used to regress
the concentrations of trace elements against clay, OC, pH,
CEC, and total concentrations of Al and Fe of the soils based
on the factorial analysis. If regression against the six independent variables was significant, partial correlation coefficients
were calculated to show the contribution of individual variable
to the total explained variance. Because concentration of trace
elements showed a log-normal distribution (data not shown),
the data were log transformed before analysis to meet the
assumption of normality required for the regression model.
(Whatman 42) and was stored in precleaned polyethylene
bottles in a refrigerator before analysis.
Concentrations of Ag, Ba, Be, Cr, Cu, Mn, Mo, Ni, Sb, and
Zn were analyzed on a Perkin-Elmer ELAN 6000 ICP-MS
unit (Norwalk, Cf), whereas those of As, Se, Pb, and Cd were
analyzed on a Perkin-Elmer SIMAA 6000 atomic absorption
spectrophotometer. Mercury was analyzed on Perkin-Elmer
2380 atomic absorption spectrophotometer equipped with a
Perkin-Elmer MHS-lO mercury-hydride system.
Data Analysis
All element concentrations are presented on a dry matter
basis. Both arithmetic and geometric means were used to
describe the central tendency and variation of the data. The
arithmetic mean (AM) and arithmetic standard deviation
(ASD) are best used as estimates of geochemical abundance
of an element. The geometric mean (GM) and geometric standard deviation (GSD), however, are better maximum likelihood estimators for most geochemical data (Gough et aI.,
1988). Baseline concentrations of 15 trace elements were calculated using GM/GSD2 and GM X GSD2 of the samples,
which include 95% of sample population (Dudka et aI., 1995).
All statistical analyses were performed using SAS (SAS
Institute, 1987). Analysis of variance was used to assess significant differences between different parameters. The confidence
interval for the Student (-test was calculated at <X = 0.05.
Simple correlation analysis was used to relate element concentrations to soil properties and among elements themselves.
R-mode factorial analysis was employed to associate elemental
concentrations with soil properties using 34 variables (Davis,
1986). The 34 variables include the total concentrations of 15
trace elements (Ag, As, Ba, Be, Cd, Cr, Cu, Hg, Mn, Mo, Ni,
Pb, Sb, Se, and Zn) and 19 quantitative soil properties (Table
1). These properties include pH measured in water and KCl;
organic carbon (OC); particle size distribution (clay, silt, sand);
cation-exchange capacity (CEC); N~OAC-extractablebases
(K, Na, Ca, Mg); available water, BaCIz-triethanolamineextractable acidity (total acid); total Ca, Mg, P, K, Fe, and
AI. A factor was interpreted as a physical or chemical process
RESULTS AND DISCUSSION
Baseline Concentrations of 15 Trace Elements
in 448 Florida Surface Soils
Florida soils formed primarily from well-weathered
sandy marine sediment (Brown et aI., 1990), thus they
are very sandy with a mean sand concentration of 89.3%
(Table 1) and contain little weatherable primary minerals. The small amount of resistant secondary minerals
present in Florida soils occurs mainly as sand-grain coatings. The coatings are dominated by minerals such as
kaolinite, hydroxy-interlayered vermiculite, gibbsite,
and quartz, as cemented by lesser amounts of metal
oxides (Harris et aI., 1995). These minerals have relatively low CEC compared with other secondary soil
minerals such as smectite (Brady and Weil, 1990). The
dominance of quartz sand in Florida soils along with
the low activity and small amount of clay, contributed
to their extremely low elemental concentrations.
Concentrations of most elements in Florida soils were
significantly lower than those reported for other regions
(Table 2). Concentrations of Cd, Cu, Ni, and Pb in the
Table 2. Concentration of trace elements in Florida surface soils (mg kg-I except for "g, which is p.g kg-I) with comparison data from
different sources.
Florida soils in this study
Elements
Ag
As
Ba
Be
Cd
Cr
Co
Hg
Mn
Mo
Ni
Pb
Sb
Se
Zn
No. of
samplest
448
445
444
417
439
444
444
443
436
442
444
439
353
445
448
Range
Median
0.16-6.00
0.01-SO.6
1.0-1990
0.01-5.92
0.004-2.80
0.02-447
0.1-318
0.62-430
1.40-1642
0.04-14.1
0.04-375
0.18-290
0.02-3.2
0.01-4.62
0.90-169
0.42
0.35
11.3
0.46
0.004
8.40
L90
4.31
18.8
1.00
8.55
4.89
0.15
0.082
4.60
Comparison chlta
AM ± ASO:j:
0.50 ±
1.34 ±
30.7 ±
0.67 ±
0.07 ±
15.9 ±
6.10 ±
U.6 ±
48.8 ±
1.52 ±
13.0 ±
11.2 ±
0.28 ±
0.25 ±
8.35 ±
0.38
3.77
108
0.76
0.23
30.6
22.1
34.4
123
1.81
23.0
26.3
0.30
0.50
13.8
GM ± GSO§
0.42
0.42
13.7
0.40
0.01
8.45
2.21
5.45
20.3
0.95
9.08
5.38
0.22
0.10
5.U
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
2.44
4.10
2.86
3.24
5.26
3.09
3.15
2.70
3.41
2.67
2.31
2.79
0.19
3.32
2.40
Florichl
soilsl
California
soW
NAn
0.41
2.8
1.1
NA
NA
0.21
3.9
3.7
4.2
25
NA
6.5
4.1
NA
NA
U
468
1.14
0.26
76
24.0
200
592
0.90
36
21.7
0.50
0.028
145
U.s. soiIstt
NA
5.2
440
0.63
NA
37
17
58
330
0.59
13
16
0.48
0.26
48
t Number of samples above the detection limits.
:j: Arithmetic mean ± arithmetic standard deviation.
§ Geometric mean ± geometric standard deviation.
11 Geometric mean reported by Ma et aI., 1997; N = 40.
# Geometric mean reported by Bradford et aI., 1996; N = SO.
tt Geometric mean reported by Shacklette and Boemgen, 1984; N = U18.
:j::j: Geometric mean reported by Wei et aI., 1990; N = 4095.
§§ Geometric mean reported by Berrow and Reaves, 1984, and arithmetic mean reported by Fergusson, 1990 (in bracket).
111 NA; Data not available.
China
soilsH
0.105
9.2
450
1.82
0.074
53.9
20.0
40
482
1.2
23.4
23.6
1.06
0.22
67.7
World
soiIs§§
NA
(11.3)
NA
6
0.4 (0.62)
SO
U
60 (98)
450
1.5
25
15 (29.2)
«0.9)
(0.4)
40
1176
J. ENVIRON. QUAL., VOL. 28, JULY-AUGUST 1999
current study were lower than these trace elements in
Florida agricultural soils reported by Holmgren et aI.
(1993). This is possible because many of the soils used
in this study were collected through the FCSSP from
uncultivated, or minimally cultivated sites (Sodek et aI.,
1990). Concentrations of 15 trace elements in this study
agreed with previously published values for 11 metals
in 40 Florida soils (Ma et aI., 1997). Concentrations of
Zn, As, and Cd, however, were much lower and those
of Cr were much greater in the current study than those
of Ma et aI. (1997). Discrepancies may relate to the
greater variety of soils analyzed in the current study.
Soil contamination may be considered when concentrations of an element in soils were two- to three times
greater than the mean background levels (Logan and
Miller, 1983). In the current study, the observed concentration ranges of 15 trace elements (Table 2) were significantly greater than their upper baseline concentration
limits (Table 3), which may suggest either contamination
in these soils (Dudka, 1993), or influence from pedogenic factors (Ma et aI., 1997). The fact that the GMs
were much closer to the medians than were AMs for
all elements in Florida surface soils confirms that the
data were strongly positively skewed (Table 2). The
calculated baseline concentrations of trace elements in
Table 3, therefore, better represent their natural concentrations in the soils because the distorting effects of
a few high concentrations are minimized (Dudka, 1993).
The upper baseline concentration limits of Cd, Co,
Cr, Cu,- Fe, Mn, Ni, S, and Zn were used by Dudka
et aI. (1995) to assess possible metal contamination in
Ontario soils. McGrath (1986) reported that the upper
baseline concentration range for Pb concentrations in
topsoils from England and Wales fell well within most
soil protection guidelines. In the current study, upper
baseline ranges for 10 elements (Ag, As, Be, Cd, Cr,
Cu, Mo, Ni, Pb, Se) corresponded well with ,~he upper
baseline values reported in literature (Table 3). In instances where significant differences were found (i.e.,
Ba, Hg, Mn, Sb, and Zn), Florida soils generally showed
lower baseline concentrations (Table 3).
Correlation Analysis for Concentrations of 15
Trace Elements in 448 Florida Surface Soils
Correlation analysis is a useful tool for analyzing similarities between paired data and is widely used in trace
metal data analyses (Bradford et aI., 1996; Dudka et aI.,
1995; Lee et aI., 1997). In the current study, correlation
analysis between elemental concentrations and soil
properties (total Fe, total AI, pH, clay, OC, and CEC)
of 448 surface soils and among trace elements was conducted (Tables 4 and 5).
Correlation between Trace Element Concentrations
and Soil Properties
Soil pH significantly correlated with concentrations
of As, Cd, Cr, Cu, Mn, Se, and Zn (Table 4). This is
consistent with the fact that their concentrations were
the lowest in Spodosols (Ma et aI., 1997), which had
the lowest pH among the seven soil orders. No such
correlation, however, was reported by Ma et aI. (1997),
possibly due to the limited sample numbers in that study.
Clay content is highly correlated with concentrations
of all trace elements except for Ba, Hg, and Ni (Table
4). This is consistent with previously published data by
Ma et aI. (1997). They reported that concentrations of
AI, As, Cr, Cu, Fe, Hg, Ni, Pb, and Zn were strongly
correlated with clay content in 40 Florida surface soils.
Correlations between clay content and concentrations
of Zn and Ni in Oklahoma soils (Lee et at, 1997), and
between clay content and concentrations of As, Cd, Cr,
Cu, Mo, Ni, Pb, Sb, and Zn in Canada soils (Mermut
et aI., 1996) were reported. Soon and Abboud (1990)
Table 3. Calculated baseline concentrations of trace elements in Florida surface soils (mg kg-I except for "g, which is fLg kg-I) compared
with published baseline concentrations in soils.
This study
Element
Ag
As
Ba
Be
Cd
Cr
Cu
Hg
Mn
Mo
Ni
Pb
Sb
Se
Zn
Soil from other studies
Florida
Floridat
0.07-2.50
0.02--7.01
1.67-112
0.04-4.15
NAn
0.03-37
NA
NA
~.33
0.11~.41
0.89-80.7
0.22--21.9
0.75-39.6
1.74-236
0.13-6.76
1.76-48.5
0.69--42.0
0.06--0.79
0.01-1.11
0.89--29.6
0.88-17.2
0.84-16.3
0.62--28.4
3.2--196
NA
4.5-9.4
0.93-18.1
NA
NA
7.1-20.0
Bull Island§
Alaska'll
California#
USAtt
ChinaH
NA
NA
NA
1.2&-35.8
213-1659
0.6S-3.33
NA
12.5-200
7.33-78.6
NA
76-3718
0.14-5.29
5.1-113
0.06--2.86
0.63-12.3
197-1110
0.36--3.65
0.05-1.34
14.8-392
7.41-77.8
44.5-899
263-1332
0.181-4.48
6.25-207
9.64-48.8
0.154-1.62
O.OOH.23
NA
NA
1.05-25.9
96.1-2015
0.11-3.57
NA
6.59--208
2.86--101
9.1-368
43-2532
0.08-4.37
2.44-69.4
4.62-55.4
0.093-2.47
0.043-1.57
12.&-183
0.OH.41
2.5-33.6
266-761
0.85-3.9
1~380
NA
NA
6.8-29
0.35-5~2
NA
45-500
NA
0.96-4.6
5.7-15
NA
NA
2.8-12
3.~36.3
NA
NA
2&-188
t Based on 95% confidence intervals (GMlGSDt to GM X GSDt); N = 448.
t Calculation based on data reported by Ma et aI., 1997; N = 40.
§ Calculation based on data reported by Gough et aI., 1994; N = 16.
'II Calculation based on data reported by Gough et al., 1988; N = 437.
# Calculation based on data reported by Bradford et al., 1996; N = 50.
tt Calculation based on data reported by Shaddette and Boemgen, 1984; N = 1218.
H Calculation based on data reported by Wei et al., 1990; N = 4095.
§§ Data reported by Dudka, 1992, 1993; N = 127.
I'll NA; Data not available.
PoJand§§
NA
0~.1
19.3-154)
7.3-55
5.9--270
1343--1740
0.15-9.8
7.73-70.9
9.95-56.0
0.38-3.0
125-409
NA
0.1-1.7
3.7-75.3
2.6-18.0
NA
83-1122
NA
2.6-27.0
NA
2.6-27.0
0.047~.99
0.07~.30
28.5-161
10.5-154
0.0~.33
1177
CHEN ET AL.: BASELINE CONCENTRATIONS OF TRACE ELEMENTS IN FLORIDA SOILS
Table 4. Correlation coefficient (r) of element concentrations with soil properties in 448 Florida surface soils.
Element
Clay
NCt
pH
0
0
0
0
0
1
1
1
1
1
1
2
2
3
4
0.14**
0.10§
0.27'11
0.10§
0.221
-0.03 NS:j:
0.10§
0.271
-8.07 NS
0.03 NS
-0.01 NS
0.03 NS
0.00 NS
-0.01 NS
0.00 NS
As
Cd
Cu
Se
Zn
All
Cr
Mn
Me
Pb
Sb
Ba
Hil
Be
Ni
0.331
0.12§
0.14**
0.28'1[
0.461
0.401
0.3411
0.3411
0.441
0.131
0.20***
0.05 NS
0.02 NS
0.58'11
0.04 NS
OC
CEC
Total Fe
Total AI
0.581
0.13**
0.15**
'-0.551
0.191
0.631
0.06 NS
0.06 NS
8.571
8.16***
0.581
0.191
0.331
-0.02 NS
0.02 NS
8.391
0.13**
0.11§
0.451
0.231
8.511
8.13**
8.13**
0.661
0.421
8.571
0.431
0.741
8.511
0.671
0.751
0.421
0.571
0.2S1
0.611
8.38'1
0.471
0.271
0.601
0.391
0.511
0.421
0.621
0.571
0.581
0.571
0.451
0.541
0.301
0.591
0.281
0.451
0.231
tt.6lIt
0.17***
0.391
0.13**
0.241
0.07 NS
-0.01 NS
**,*** Significantly different at levels of a = 0.01 and 0.001, respectively.
t NC, number of correlation that are not significant among trace elements and soil properties.
:j: NS, not significant.
§ Significantly different at levels of a = 0.10.
1 Significant different at levels of a = 0.0001.
found that Cr, Pb, and Cu were correlated with clay
content. Strong correlations between concentrations of
As, Cr, Cu, Mn, and Ni and the amounts of particles
<0.02 mm in surface soils of Poland also were reported
by Dudka (1993). They suggested that clay content was
important in controlling the level and distribution of
trace metal concentrations in soils (Soon and Abboud,
1990; Dudka, 1993). Mermut et ai. (1996) indicated that
Pb was likely adsorbed on the 2:1 silicate clay minerals,
and therefore Pb concentrations would be expected to
increase with increasing clay content. High correlation
coefficients (P < 0.001) between background values of
trace elements (As, Ba, Cr, Mn, Sb, and Zn) and clay
content in Dutch topsoils were attributed to the phenomenon that soils in the Netherlands had been developed from sediments (Edelman and de Bruin, 1986).
As such, elemental concentrations were used with clay
and OC contents to establish guideline values for contaminant levels in Dutch soils (Forstner, 1995).
In contrast to previously published data (Ma et aI.,
1997), OC showed significant correlation with trace element concentrations except for Be, Cr, Mn, and Ni (Table 4). This may be due to the fact that no organic
soil was included in the former study. Ames and Prych
(1995) reported that concentrations of most transition
metals (Co, Cu, Hg, Mn, Ni, and Zn) in soils of Washington showed significant correlation with Oc. This finding
was attributed to the strong absorption of transition
metals by soil organic matter (Stevenson, 1982). Significant correlation (P < 0.05) was also found between
certain metal concentrations and OC contents in Dutch
topsoils (Edelman and de Bruin, 1986) and agricultural
soils of northwestern Alberta (Soon and Abboud, 1990).
In Florida, surface oxidation of OC in the aerobic layers
of Histosols may concentrate Cd in the surface of these
soils (Holmgren et aI., 1993). Furthermore, humic substances in organic soils can serve as strong reducing and
complexing agents and influence the processes controlling mobilization of many toxic metals including Hg
(Gough et aI., 1996). A recent study on the sawgrass
(Cladium jamaicense Crantz) prairie wetlands in south
Florida demonstrated that nonessential trace elements
(such as Cr, Co, Pb, and Hg) were generally not being
cycled but were concentrated in the organic-rich sediments (U.S. Geological Survey, 1996).
CEC showed significant positive correlation with
most trace elements except Be and Ni (Table 4). This
result agrees well with Holmgren et ai. (1993), who
Table 5. Correlation coefficients (r) among elemental concentrations in 448 Florida surface soils.
Ag
As
Mo
Zn
Se
Sb
Mn
Cu
Cr
Cd
Pb
Ba
Be
Hg
Ni
NCt
Ag
As
0
1
1
1
1
2
2
2
3
3
3
3
4
6
10
1
0.18
0.67
0.39
0.49
0.75
0.16
0.25
0.17
0.17
0.15
0.41
0.18
0.56
0.09
***
1
0.23
0.21
0.28
0.18
0.35
0.15
0.17
0.32
0.45
0.09
0.13
0.13
0.01
Mo
'II
'II
1
0.30
0.52
0.53
0.10
0.37
0.50
0.13
0.07
0.16
0.13
0.20
0.31
Zn
'II
'II
'II
1
0.25
0.22
0.53
0.70
0.31
0.37
0.20
0.20
0.46
0.12
0.07
Se
Sb
Mn
'II
t
'II
**
***
'II
'II
§
11
1
0.60
0.12
0.63
0.16
0.24
0.10
0.22
0.03
0.38
0.08
'II
'II
1
0.12
0.15
0.15
0.17
0.15
0.35
-0.01
0.65
0.05
§
11
Cu
Cr
Cd
Pb
1
**
11
***
***
***
***
11
..
**
**
§
NS
I
§
**
**
§
NS
§
§
**
1
0.39
0.25
0.24
0.15
0.12
0.45
0.01
0.08
1
0.23
0.44
0.12
0.13
0.21
0.07
0.05
'I
'II
1
***
**
***
'II
I
'I
'I
1
0.13
0.05
0.07
0.36
0.07
0.56
**
1
0.34
0.07
0.17
0.08
0.01
**,*** Significantly different at levels of a = 0.01 and 0.001, respectively.
t NC, number of correlation that are not significant among trace elements with the maximum being 14.
:j: NS, Not significant.
§ Significantly different at levels of a = 0.10.
'II Significant different at levels of a = 0.0001.
'I
'I
'1\
1
0.10
0.08
0.09
-0.04
BA
Be
Hg
Ni
1
§
***
NS
***
***
**
'I
**
'I
§
§
NS:j:
1
I
NS
NS
'I
'I
***
§
§
1
-0.04
0.07
NS
NS
NS
NS
§
**
NS
1
0.01
1
§
**
NS
NS
§
1
0.08
0.11
-0.01
,,
,
'I
§
NS
NS
NS
,
NS
NS
NS
II(S
NS
1
1178
J. ENVIRON. QUAL., VOL. 28, JULY-AUGUST 1999
reported that trace elements show good correlation with
both CEC and Oc. This is understandable since CEC
is simply correlated with clay containing trace elements
and shows significant positive correlation with OC and
pH (Stevenson, 1982).
All 15 trace elements were highly correlated to both
total Fe and Al concentrations in Florida surface soils
(Table 4). These correlation coefficients were the
strongest among all six variables tested in Table 4. A
similar but a slightly weaker correlation was reported by
Ma et aI. (1997). Dudka (1993) found good correlation
between concentrations of As, Co, Cr, Cu, Ga, Mn, Ni,
and Se and concentrations of Al and Fe in surface soils
of Poland. He concluded that levels of most elements
were mainly controlled by minerals present in those
soils (Dudka, 1992). Total Fe and Al concentrations
(2300 and 2200 mg kg-I) in Florida soil are 16 to 32
times lower than the average concentrations reported
for other soils (38000 and 71 000 mg kg-I; Lindsay,
1979). Apparently, total Fe and total AI, even at such
low concentrations, are significant in controlling metal
concentrations in Florida soils. The capacity of Fe and/
or Al oxides in sorbing and/or co-precipitating trace
elements has been widely studied (Zachara et aI., 1993;
Karthikeyan, 1997). We hypothesize that trace elements
may have co-precipitated with Fe-AI oxides during their
formation in soils, existing as structural components of
Fe-AI oxides instead as exchangeable ions on Fe-AI
oxide surface. This is supported by the fact that trace
element concentrations correlated better with total Fe
and total Al than with CEC (Table 4).
(1997). They reported that concentrations of AI, As,
Cr, Cu, Fe, Mn, Pb, and Zn in 40 Florida soils positively
correlated with each other. Among the 15 trace elements tested in the present study, Ag was correlated
with all other elements, whereas Ni correlated only with
Ag, Cr, Mo, and Se. Good correlation between concentrations of Ni and Cr has been reported, however, for
surface soils of California (r = 0.95, P <0.01; Bradford
et aI., 1996) and Minnesota (r = 0.90, P < 0.01; Pierce
et aI., 1982). Correlation between concentrations of Ni
and Cr and concentrations of Ti and Al in Washington
soils was reported by Prych et aI. (1995). They suggested
that Cr and Ni were associated mostly with the mineral
phase in the soils.
Lead and Ba displayed significant correlation with
most elements, excluding Mo, Ni, Cr, and Cd; beryllium
displayed significant positive correlation with most elements, excluding Ni, Se, and Sb; and Hg showed significant correlation with most elements except for Be, Cd,
Cr, Cu, Mn, and Ni (Table 5). High correlation among
trace elements in Florida soils suggests that similar processes control element associations in parent materials
(Bradford et aI., 1996). To separate natural from anthropogenic factors influencing trace element concentrations in soils, however, normalization of the data based
on weather- or leach- resistant reference elements, such
as AI, Ti, and Zr, is needed.
Correlation among Trace Elements
Factorial Analysis
Significant correlation was found among most trace
elements, especially Ag, As, Cd, Cr, Cu, Mn, Mo, Sb,
Se, and Zn (Table 5). This may occur because they have
similar ionic radii, with the exception of Ag (KabataPendias and Pendias, 1992; Dean, 1992). This result is
consistent with previously published data by Ma et aI.
Factorial analysis is an extension of correlation analysis. It can divide variables into groups that are consistent
with anthropogenic or pedogenic processes (Davies and
Wixson, 1987; Dudka, 1992). In the present study, eight
factors satisfactorily described distributions of trace element concentrations in Florida surface soils. These fac-
Factorial Analysis and Multiple Regression
of Trace Element Concentrations
and Soil Properties
Table 6. Results of R-mode factorial analysis for Florida surface soils showing relative loading from element concentrations and soil
properties on eight factors derived by varimax rotation.
Factor number
Factor
loading
Factor I
28.9t
Factor II
14.8t
Factor III
10.3t
Factor IV
8.1t
Factor V
7.5t
Factor VI
6.7t
Factor VII
6.4t
Factor
VIII 3.8t
Percent
variance
explained
(%)86.6:1:
100
1.00
Extr-Mg,
Extr-Na, CEC
90
80
0.90
Fe,AI
Extr-Ca
OC
Total-acid
pH-KCI
70
Ni
0.80
K, Mn, Ba
Extr-K
pH-H,O, Ca
Silt, Clay
Hg
Mo
0.70
Avail-H,O
Zn,P,Be
Cr
60
50
Sb
40
0.60
Pb, As
Mg
Cr, Co
Co
Pb,Ag
30
0.50
0.40
t Percentage of tbe total variance explained by a factor.
t Percentage of the total variance explained by all factors.
Se, Avail-H,O
Ag
Cd,As
Mg
20
1179
CHEN ET AL.: BASELINE CONCENTRATIONS OF TRACE ELEMENTS IN FLORIDA SOILS
tors explained 87 % of the variance using the 34 variables
in the analysis (Table 6). The resulting varimax factors
were not correlated and different variables generally
had different loadings on different factors. Apparently
some of the elements (As, Ag, Cr, Cu, and Pb) were
controlled by more than one factor.
Factor 1 (total Fe and AI) explained 29% of the total
variance and was the most important factor, which was
consistent with correlation data in Table 4. It had large
loadings from total concentrations of Fe and Al and
moderate loadings from total concentrations of K and
P in the soils. This factor may represent Fe and Al
oxides, perhaps combining with K and P minerals as
well as some resistant metals. Four trace elements Mn,
Ba, Zn, Be showed strong associations with Factor 1.
Relatively low loadings from Cr, Cu, Pb, and As on the
total Fe and Al factor suggest that other factors influence these elements in Florida soils. It was reported
that in acidic soils, the main forms of As were Al and
Fe arsenates (AIAs0 4and FeAs04), whereas in alkaline
and calcareous soils the main form was Ca3(As04)2 (Fergusson, 1990).
Factors 2, 3, 4, and 5 were related to major soil properties; namely CEC (Factor 2), pH (Factor 3), clay (Factor
4), and OC (Factor 5), which together explained 41 %
of the total variations. The large loadings from the extractable bases on Factor 2 (soil CEC factor) means
that these extractants were highly correlated with CEC.
Factor 3 (soil pH factor) had a large loading from pH.
Total concentrations of Ca and Mg had moderate loadings on Factor 3, implying that total concentrations of
both base elements contributed to the soil pH factor.
Total concentrations of Cu had some loading on the
soil pH factor, confirming the significant associations
between Cu and soil pH (Table 4). Factor 4 (soil clay
factor) was primarily due to particle-size distribution
(positively associated with silt and clay, negatively associated with sand), and to a lesser extent, with available
moisture content of the soils. The loading of Se on
Factor 4 was consistent with the significant correlation
coefficients between concentrations of Se with soil clay
content (Table 4). Factor 5 (soil OC) is clearly due to
OC and total acid, and to a lesser extent, available
moisture content of the soils. The lack of significant
loading of any trace metals on Factor 2 and Factor 5
implies that both OC and CEC did not have major
influence on occurrences of these trace elements in Florida surface soils.
Factor 6 (Ni and Mo factor) explains 7% of the total
variances. It had large loadings from Ni and Mo and to
a lesser extent of Cr and Ag. This is consistent with the
relationships between concentrations of Ni with Mo, Cr
and Ag (Table 5). Chromium and Mo are transition
elements that locate at Group 6b of the Periodic Table
and have strong lithophile tendencies. Though they have
variable oxidation states, they are preferably hexavalent
in their oxygen compounds (Kabata-Pendias and Pendias, 1992). Nickel and Cr in soils are mostly from pedogenic sources (Kabata-Pendias et aI., 1992). During magnetic fractionation, Cr is closely associated with Ni and
accumulates in ultrabasic rocks (Davies and Wixson,
1987).
Factors 7 and 8 were of minor importance; together
they explained ~ 10% of the total variations. They may
be regarded as anthropogenic factors based on information described previously. These factors included Sb,
Pb, Ag, Cd, As (Factor 7), and Hg (Factor 8). Factor 7
(Sb and Pb) confirmed significant associations among
total concentrations of these elements (Table 5). Accumulations of Cd in organic soils were attributed to application of phosphate fertilizers containing Cd (Holmgren
et aI., 1993). The composition of Factor 8 (Hg) indicates
an obvious relationship of total concentration of Mg
with Hg in the soils. It was reported atmospheric deposition was more important for Pb, As, and Hg, whereas
phosphate fertilizer is marginally more important for Cd
Table 7. Multiple and partial correlation coefficients for regression of concentrations of trace metals in florida surface soils against key
soil properties.
I
l
Trace
elementt
Number of
variable
Mn
Ni
Ba
Be
Hg
Ag
Sb
Se
As
Cd
Cr
Cu
Pb
Zn
Mo
Number of elements
3
2
2
2
2
2
2
1
1
1
1
1
1
1
1
Multiple correlation
coefficient
Partial correlation coefficients (r)
Total Fet
Total Alt
0.597
0.125
0.240
0.234
0.187
NS
NS
NS
0.252
0.139
0.407
0.169
0.203
0.358
0.116
12
NS:j:
NS
0.119
0.270
-0.128
0.196
0.130
NS
NS
NS
NS
NS
NS
NS
NS
5
*** Significantly different at levels of a = 0.0001.
t Concentration after log-transformation.
:j: NS, not significant at levels of a = 0.05 using the student '-test.
Clay
-0.134
NS
NS
NS
NS
0.130
0.199
0.245
NS
NS
NS
NS
NS
NS
NS
4
OC
CEC
pH-HzO
R'
F-test
~0.157
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
0
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
0
0.65
0.09
0.33
0.44
0.10
0.29
0.10
0.25
0.42
0.23
0.46
0.33
0.30
0.52
0.13
***
***
***
***
***
***
***
***
***
***
***
***
***
***
***
-0.126
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
2
1180
J. ENVIRON. QUAL., VOL. 28, JULY-AUGUST 1999
contamination in soils (Fergusson, 1990). Very recent
studies suggest that Hg deposition in South Florida is
generally driven by large-scale regional or hemispheric
processes as opposed to local emission/deposition processes (Gough et aI., 1996).
Multiple Regression Analysis
Multiple regression of concentrations of trace elements against clay, OC, pH, CEC, and total concentrations of Al and Fe supported the relationships of trace
elements with important soil properties (Table 4) and
the results of factorial analysis (Table 6). Regressions of
log-transformed concentrations of the 15 trace elements
against six soil variables were all significant, explaining
between 9 to 65% of the total variance (Table 7).
Partial correlation coefficients between individual element and soil property provided the relative importance of each soil property on elemental distributions.
The number of variables and magnitude of partial correlation coefficients (r) confirmed that total Fe and total
Al were the two major variables controlling concentrations and distributions of most trace elements in Florida
surface soils as demonstrated previously using simple
correlation analysis (Table 4) and factorial analysis (Table 6). Total Fe concentrations are important for all 15
trace elements except for Ag, Sb, and Se, which were
apparently related to clay content, whereas total Al
concentrations were important for elements Ba, Be, Hg,
Ag, and Sb. Though CEC and pH are critical soil properties, their importance on the distribution of trace elements was not significant because they depend highly
on soil components (i.e., clay, OC, Fe, and Al oxides),
which also had significant positive correlation with concentrations of most trace elements (Tables 4 and 6).
Our results demonstrated the importance of Fe and/or
Al oxides and clay in controlling trace element concentrations in Florida soils (Table 7).
Among the 15 trace elements, partial correlation coefficients of Fe with Mn (r = 0.60), Cr (r = 0.41), and Zn
(r = 0.36) were the highest (Table 7). Iron and Mn
have similar redox chemistry and geochemical behavior,
which explain their high covariance in soils. Actually,
hydrous oxides of Fe and Mn in soils were reportedly the
most important compounds in sorption of trace metallic
pollutants, and they exhibit diverse affinities to NiH,
CuH , ZnH , CdH , Pb4+, and Ag+, which have approximately the same physical dimensions as Mn and Fe ions
(Fergusson, 1990). In addition, oxidation of As, Cr, and
Hg by Mn oxides is likely to control the redox behavior
of these three elements in soils (Kabata-Pendias and
Pendias, 1992). In general, Cr closely resembles Fe and
Al in ionic size and in geochemical properties. The association between Cr and Fe may reflect the fact that most
of the Cr in soils is present as chromite (FeCrz04) or in
other spinel structures, substituting for Fe. However,
under conditions induced by fluctuating water tables,
Fe is depleted and Al is rendered less crystalline and
more prone to organo-complexation. Aluminum hydroxides can thus adsorb a variety of trace elements and
be more important than that of Fe oxides in remaining
certain trace elements like Ag, Be, and Sb (Table 7).
CONCLUSIONS
The GM concentration levels of 15 potentially toxic
trace elements in Florida surface soils were lower than
the average of USA and world soils. The upper limit of
baseline concentrations for most trace elements (Ag,
As, Be, Cd, Cr, Cu, Mo, Ni, Pb, Se), however, corresponded well with those reported in the literature. Baseline concentrations of 15 trace elements in 448 representative Florida soil samples were proposed as reference
concentrations in Florida. Due to possible influence
from anthropogenic factors on concentrations of Sb,
Pb, Ag, Cd, As, and Hg (Table 6), however, baseline
concentrations estimated for Ba, Be, Cu, Mn, Mo, Ni,
Cr, Se, and Zn are probably a better measure of the
natural levels of these elements in Florida surface soils.
Total Fe and Al showed the strongest relationship
with concentrations of most trace elements, based on
simple correlation analysis, factorial analysis and multiple regression (Tables 4, 6, and 7). Other important
variables include clay and Oc. The importance of CEC
and pH on the distribution of trace elements was diluted
by these component variables. Significant correlation
coefficients were also found among Ag, As, Cd, Cr, Cu,
Mn, Mo, Sb, Se, and Zn (Table 5).
ACKNOWLEDGMENTS
This research was sponsored in part by the Florida Center
for Solid and Hazardous Waste Management (Contract no.
96011017). The authors would like to thank the Chemistry
Laboratory of the Florida Department of Environmental Protection for determining concentrations of 10 trace elements.
We are also indebted to those who participated in the Florida
Cooperative Soil Survey. Their collection and characterization
of a large number of Florida soil samples made this study
possible. The helpful suggestions made by Dr. Nick Basta and
three anonymous reviewers are gratefully acknowledged.
REFERENCES
Ames, K.c., and E.A. Prych. 1995. Background concentrations of
metals in soils from selected regions in the State of Washington.
USGS Water Resour. Invest. Rep. 95-4018. U.S. Geol. Surv., Tacoma, WA.
Berrow, M.L., and G.A. Reaves. 1984. Background levels of trace
elements in soils. p. 333-340. In Proc. 1st Int. Conf. on Environmental Contamination. CEP Consultants, Edinburgh, Scotland.
Bradford, G.R, A.c. Chang, A.L. Page, D. Bakhtar, J.A. Frampton,
and H. Wright. 1996. Background concentrations of trace and major
elements in California soils. Kearney Foundation Spec. Rep., Univ.
of California, Riverside.
Brady, N.C., and RR Weil. 1996. The nature and properties of soils.
p. 261. Prentice Hall, Upper Saddle River, NJ.
Brown, RB., E.L. Stone, and V.W. Carlisle. 1990. Soil. p. 35--69. In
RL. Myers and J.J. Ewel (ed.) Ecosystems of Florida. Univ. of
Central Florida Press, Orlando.
Davies, B.E., and B.G. Wixson. 1987. Use of factor analysis to differentiate pollutants from other trace metals in surface soils of the
mineralized area of Madison County, Missouri, USA. Water Air
Soil Pollut. 33:339-348.
Davies, B.E. 1992. Trace metals in the environment: Retrospect and
prospect. p. 1-18. In D.C. Adriano (ed.) Biogeochemistry of trace
metals. CRC Press, Boca Raton, FL.
i
(
CHEN ET AL.: BASELINE CONCENTRATIONS OF TRACE ELEMENTS IN FLORIDA SOILS
Davis, J.C 1986. Statistics and data analysis in geology. 2nd ed. John
Wiley & Sons, New York.
Dean, J.A (ed.) 1992. Lange's handbook of chemistry, 14th ed.
McGraw-Hill, New York.
Dudka, S. 1992. Factor analysis of total element concentrations in
surface soils of Poland. Sci. Total Environ. 121:39-52.
Dudka, S. 1993. Baseline concentrations of As, Co, Cr, Cu, Ga, Mn,
Ni, and Se in surface soils, Poland. Appl. Geochem. 2:23-28.
Dudka, S., R Ponce-Hernandez, and T.C Hutchinson. 1995. Current
levels of total element concentrations in the surface layer of Sudbury's soils. Sci. Total Environ. 162:161-172.
Edelman, T., and M. de Bruin. 1986. Background values of 32 elements
in Dutch topsoils, determined with non-destructive neutron activation analysis. p. 88-98. In J.W. Assink and J. van den Brink (ed.)
Contaminated soil. Martinus Nijhoff Publishers, Dordrecht, the
Netherlands.
Florida Department of Environmental Protection. 1995. Domestic
wastewater residuals. FDEP Rep. 62-640. Florida Dep. of Environ.
Protection, Tallahassee.
Fergusson, J.E. (ed.) 1990. The heavy elements: Chemistry, environmental impact and health effects. Pergamon Press, Oxford.
Forstner, U. 1995. Land contamination by metals: global scope and
magnitude of problem. p. 1-33. In H.E. Allen et al. (ed.) Metal
speciation and contamination of soil. Lewis Publ., Boca Raton, FL.
Gough, L.P. 1993. Understanding our fragile environment, lessons
from geochemical studies. USGS Circular 1105. U.S. Gov. Print.
Office, Washington, DC
Gough, L.P., RC. Severson, and H.T. Shacklette. 1988. Element concentrations in soils and other surficial materials of Alaska. USGS
Prof. Pap. 1458. U.S. Gov. Print. Office, Washington, DC
Gough, L.P., RC Severson, and L.L. Jackson. 1994. Baseline element
concentrations in soils and plants, Bull Island, Cape Romain National Wildlife Refuge, South Carolina, USA Water Air Soil Pollut. 74:1-17.
Gough, L.P., RK. Kotra, CW. Holmes, P.H. Briggs, J.G. Crock, D.L.
Fey, P.L. Hageman, and AL. Meier. 1996. Chemical analysis results
for mercury and trace elements in vegetation, water, and organicrich sediments, south Florida. USGS Openfile Rep. 96-091. U.S.
Geol. Surv., Denver Federal Ctr., Denver, CO.
Harris, W.G., S.H. Crownover, and N.B. Comerford. 1995. Experimental formation of aquod-like features in sandy coastal plain soils.
Soil Sci. Soc. Am. J. 59:877-86.
Holmgren, G.S., M.W. Meyer, RL. Chaney, and RB. Daniels. 1993.
Cadmium, lead, zinc, copper, and nickel in agricultural soils of the
United States of America. J. Environ. Qual. 22:335-348.
Kabata-Pendias, A, S. Dudka, and A Chlopecha. 1992. Background
levels and environmental influences on trace metals in soils of the
temperate humid zone of Europe. p. 61-84. In D.C. Adriano. (ed.)
Biogeochemistry of trace metals. CRC Press, Boca Raton, FL.
Kabata-Pendias, A, and H. Pendias (ed.) 1992. Trace elements in
soils and plants. 2nd ed. CRC Press, Boca Raton, FL.
Karthikeyan, K.G., H.A Elliott, et al. 1997. Adsorption and coprecipitation of copper with the hydrous oxides of iron and aluminum.
Environ. Sci. Technol. 31:2721-2725.
1181
Lee, B.D., B.J. Carter, NT. Basta, and B. Weaver. 1997. Factors
influencing heavy metal distribution in six Oklahoma benchmark
soils. Soil Sci. Soc. Am. J. 61:218-233.
Lindsay, W.L. (ed.). 1979. Chemical equilibria in soils. John Wiley &
Sons, New York.
Logan, T.J., and RH. Miller. 1983. Background levels of heavy metals
in Ohio farm soils. Soil contamination analysis. Res. Circ. Ohio
Agric. Res. Dev. Ctr. Wooster 275:3-15.
Ma, L.Q., F. Tan, and W.G. Harris. 1997. Concentrations and distributions of eleven elements in Florida soils. J. Environ. Qual. 26:
769-775.
McGrath, S.P. 1986. The range of metal concentrations in topsoils of
England and Wales in relation to soil protection guidelines. p.
242-252. In D.D. Hemphill (ed.) Trace substance in environmental
health. Univ. of Missouri, Columbia.
Mermut, AR, J.C Jain, L. Song, R Kerrich, L. Kozak, and S. Jana.
1996. Trace element concentrations of selected soils and fertilizers
in Saskatchewan, Canada. J. Environ. Qual. 25:845-853.
Pierce, F.J., RH. Dowdy, and D.F. Grigal. 1982. Concentrations of
six trace metals in some major Minnesota soil series. J. Environ.
Qual. 11 :416-422.
Prych, E.A, D.L. Kresch, J.C Ebbert, and G.L. Turney. 1995. Data
and statistical summaries of background concentrations of metals in
soils and streambed sediments in part of Big Soos Creek Drainage
Basin, King County, Washington. USGS Water Resour. Invest.
Rep. 94-4047. U.S. Geol. Surv., Tacoma, WA.
SAS Institute. 1987. SAS user's guide: Statistics. SAS Inst., Gary, NC
Shacklette, H.T., and J.G. Boerngen. 1984. Element concentrations
in soils and other surficial materials of the conterminous United
States. USGS Prof. Pap. 1270. U.S. Gov. Print. Office. Washington, DC
Sodek, F. III., V.W. Carlisle, M.E. Collins, L.C Hammond. and W.G.
Harris. 1990. Characterization data for selected Florida soils. Soil
Sci. Res. Rep. 90-1. Soil and Water Science Dep.. Univ. of Florida, Gainesville.
Soil Survey Division Staff. 1993. Soil survey manual. USDA Handb.
no. 18. U.S. Gov. Print. Office, Washington, DC
Soon, Y.K., and S. Abboud. 1990. Trace elements in agricultural soils
of northwestern Alberta. Can. J. Soil Sci. 70:277-288.
Stevenson, F.J. (ed.) 1982. Humus chemistry: genesis. composition,
reactions. John Wiley & Sons. New York.
U.S. Environmental Protection Agency. 1995. Standards for the use
or disposal of sewage sludge. 40 CFR Parts 403 and 503. USEPA,
Washington, DC
U.S. Environmental Protection Agency. 1996. Soil screening guidance:
User's guidance. USEPA 540/R-96/018. USEPA, Washington, DC
U.S. Geological Survey. 1996. South Florida ecosystems: The role of
peat in the cycling of metals. U.S. Geol. Surv. Fact Sheet. FS-16196. USGS, Washington, DC
Wei, F.S., J.S. Chen, CJ. Zheng, and D.Z. Jiang (ed.) 1990. Elemental
background concentrations of soils of China. China Environ. Sci.
Publ. Ltd., Beijing (In Chinese).
Zachara, J.M., S.C Smith, J.P. McKinley, and CT. Resch. 1993. Cadmium sorption on specimen containing layer silicates and iron and
aluminum oxides. Soil Sci. Soc. Am. J. 57:1491-1501.