Cases and solutions
Bed sediment-associated trace
metals in an urban stream, Oahu,
Hawaii
R.A. Sutherland
Abstract Of the 117 stream and lake systems sam-
pled nationwide, fish from Manoa Stream on Oahu,
Hawaii, have consistently shown the highest Pb
concentrations. Therefore a detailed study was con-
ducted to examine total metal contents in bed sedi-
ments from a 5.8-km stretch of Manoa Stream. A
total of 123 samples (~63 mm) were examined for
18 elements and 14 samples for 21 elements. Se-
lected samples were also examined using different
leach solutions to examine metal phase associa-
tions. All trace metal data, computations of enrich-
ment ratios and the modified index of geoaccumu-
lation point to mineralogical control for Cr and Ni;
minor anthropogenic contamination for Ba, Cd, Cu,
Hg and Zn; and a very strong contamination signal
for Pb. Maximum Pb contents (up to 1080 mg kg –1)
were associated with anthropogenic material dump-
ing in minor tributaries, storm sewer sediments
and sediments in the “lower” section of the basin.
Proportionally Pb had the highest non-residual
component of elements examined; dominantly in
the reducible phase associated with Mn and amor-
phous Fe oxyhydroxides. The contamination signal
was typically lowest in the “undisturbed” headwater
reach of the basin (above 5.1 km) with significant
increases throughout the “residential” and “com-
mercial-institutional” zones of the mid-basin. The
spatial pattern of bed sediment contamination and
evidence from storm sewer-outlet sediments strong-
ly indicates that Pb, and to a lesser degree some
other metals, is still being transported to the
stream and the primary agent is soil erosion and
transport of metals sorbed to sediments. The pri-
mary source of sediment-associated metals is con-
sidered to be the automobile, though other minor
sources can not be ruled out.
Received: 3 November 1998 7 Accepted: 26 January 1999
R.A. Sutherland
University of Hawaii, Geomorphology Laboratory, Department
of Geography, 2424 Maile Way, Honolulu, Hawaii 96822, USA
e-mail: Sutherla6Hawaii.edu
Key words Lead 7 Bed sediments 7 Enrichment
ratios 7 Modified index of geoaccumulation 7
Manoa Stream 7 Trace metals 7 Spatial variation 7
Phase associations
Introduction
Anthropogenic metal pollution of aquatic environments
is a critical area of inquiry. Significant advances have
been made over the last decade in detecting metal pollu-
tion sources, locating metal sinks and deciphering chron-
ologies of pollution inputs. In addition, significant strides
have been made in assessing bio-toxicity of a variety of
organic and inorganic compounds concentrated in the
environment as a result of human activities. Most recent
research has focused on aquatic sediments. Horowitz
(1991) stated that the strong association of numerous
trace elements with sediment (suspended plus bed sedi-
ments) indicates that the distribution, transportation and
availability of these can not be evaluated intelligently
solely through the sampling and analysis of the dissolved
phase. Feltz (1980) summarized the significance and use
of bed sediments as follows:
1. As an historical water quality integrator
2. As a reconnaissance tool
3. In planning analytical schedules
4. In conducting short-lived studies
5. For deriving short- and long-term trends
6. For identification of problem areas
Additionally, bed sediments are not only a sink for trace
metals, but also a source of resuspended sediment, and
under changing environmental conditions, sediment
bound pollutants may be chemically remobilized and ent-
er the water column or food chain (Dickinson and others
1996).
Another important medium that is used to diagnose pol-
lution signals in aquatic ecosystems is tissue analysis, in
the US this research has been conducted by the National
Contaminant Biomonitoring Program (NCBP) under the
auspices of the US Fish and Wildlife Service (Schmitt and
Brumbaugh 1990). The focus of the NCBP has been on
trace metals (and organic pollutants) in whole freshwater
fish samples. The choice of sample media was based on
Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag 611
 Cases and solutions
the realization that trace species can be absorbed from
the water column or ingested through sediments. As
Breault and Harris (1997) note, chemistry of streambed
sediment influences the biotic quality of a stream as
aquatic organisms ingest particulate matter (suspended
sediment and bed sediments) and accumulate trace ele-
ments and organic compounds. It is now widely recog-
nized that concentrations of trace metals in sediments
usually exceed those in the water column by three to five
orders of magnitude and that the bioavailability of even a
small fraction of this assumes considerable importance
(Bryan and Langston 1992). Furthermore, these authors
state that several trace metals, including Pb and Hg, may
be transformed in sediment to organo-metallic com-
pounds having increased bioavailability and toxicity.
Once sediments are ingested by benthic animals, they can
be expected to undergo a unique chemical environment
controlled by the digestive chemistry of the organism
(Mayer and others 1996). In fact, solubilization of sedi-
ment sorbed Cu, and probably other transition metals, by
complexation reactions in fish-gut fluids rich in proteins
would account for enhanced metal bioavailability and
bioaccumulation (Chen and Mayer 1998). Additionally,
there is the possibility of food chain biomagnification of
some elements with increasing trophic level, e.g. Hg.
Thus, in summary it is apparent that aquatic bed sedi-
ments are a critical environmental media for assessing
anthropogenic trace metal signatures.
Trace metals in the environment
of Oahu
A summary of anthropogenic sources for selected trace
metals in the environment are given in Table 1. Hawaii is
unusual in that it is far removed from any major indus-
trial point sources such as coal combustion, non-ferrous
metal production and iron and steel manufacturing. The
primary inputs of trace metals, above baseline levels, in-
clude volcanic emissions, vehicle emissions (gaseous and
fluid), vehicle-associated wear and agricultural inputs
(fertilizers and pesticides).
A limited number of trace metal studies have been con-
ducted on the island of Oahu, Hawaii. Studies by Jerni-
gan and others (1971), Eshleman (1973), and Fu and oth-
ers (1989) have shown that automobile exhaust emissions
have significantly accentuated Pb accumulations in the
air and in soils. In fact, Jernigan and others (1971) con-
cluded that Honolulu (Oahu) had a very significant prob-
lem with automobile air pollution. More recent studies
on coastal sediments (McMurtry and others 1995) and a
chronological study of an artificial estuary in Honolulu
(Ala Wai Canal) by De Carlo and Spencer (1995 , 1997)
and Spencer and others (1995) have confirmed that auto-
mobile emissions have contributed significant amounts of
Pb to the environment. Thus, the legacy of automobile
612 Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag
Table 1
Potential anthropogenic sources of selected trace elements in
the environment
Trace Potential anthropogenic sources a
element
Ba Rubber production, lubricating oil additives, fuel
synthesis, fuel combustion, phosphate fertilizers,
sewage sludges
Cd Lubricating oils, diesel oils, tires, phosphate
fertilizers, sewage sludge, insecticides, electroplating,
pigments, batteries, coal and oil combustion,
non-ferrous metal production, refuse incineration,
iron and steel manufacturing
Cu Metal plating, bearing and brushing wear, moving
engine parts, brake-lining wear, fungicides and
insecticides, anti-foulants, corrosion of Cu plumbing,
algaecides, concrete and asphalt, rubber, phosphate
fertilizers, sewage sludges
Hg Insecticides, fungicides, electrical equipment, paint,
plastics, cosmetics, anti-fouling and mildew-proofing
paints, phosphate fertilizers, batteries, fireworks
Ni Diesel fuel and vehicle exhaust, lubricating oil, metal
plating, brushing wear, brake lining wear, asphalt
paving, phosphate fertilizers, storage batteries
Pb Leaded gasoline, automobile exhaust, tire wear,
lubricating oil and grease, bearing wear, brake
linings, rubber, concrete, paint manufacturing, battery
manufacturing, insecticides, phosphate fertilizers,
sewage sludges
Zn Vulcanization of rubber and tire wear, motor oil,
grease, batteries, galvanizing, plating, air-conditioning
ducts, pesticides, phosphate fertilizers, sewage
sludges, transmission fluid, under coating, brake
linings, asphalt, concrete, coal combustion, smelting
operations, incineration and wood combustion
a
Sources: Lagerwerff and Specht (1970); Frank and others
(1976); Wigington and others (1983); Moore and Ramamoorthy
(1984); Harned (1988); Kabata-Pendias and Pendias (1992); Lee
and others (1994); Alloway (1995); Raine and others (1995);
Monaci and Bargagli (1997)
pollution remains on Oahu, but this is not the only
source. In specific locations military activities are the ma-
jor source of trace metals, including Pb (Char 1977; Ash-
wood and Olsen 1988; Ashwood and others 1989).
Many of the trace metal studies cited above have focused
on soil or sediment. The work of the NCBP has examined
metals in whole-fish composite samples from two streams
in Hawaii both on Oahu, Waikele and Manoa. Consis-
tently, since fish Pb-monitoring began in 1971 Manoa has
ranked number one every year this stream was sampled
out of 117 streams in the US. No statistically significant
change in Pb concentration was noted in fish between the
1976–1977 measurement period and 1984 (Schmitt and
Brumbaugh 1990). Mean Pb values in whole tilapia (Tila-
pia mossambica) samples collected in 1997 from the Ala
Wai Canal (1.5 km downstream of the Manoa basin outlet
and a sink for sediment from Manoa Stream; Fig. 1) mea-
sured 3.0 mg kg –1 (maximum 3.7 mg kg –1) and 0.5 mg

kg –1 for fillets (Brooks and Ling 1998 unpublished Hawaii
Department of Health document). The whole-fish Pb data
compare favorably to the mean value of 4.0 mg kg –1 for
the 1984 sampling of Tilapia in Manoa Stream. Though it
is difficult to assess the relative health hazard from eating
fish from Manoa Stream (or the Ala Wai Canal), it is
clear that levels of Pb remain high within the drainage
basin particularly within soils and aquatic bed sediments.
Objectives
The objective of this paper is to examine in detail the
geochemical signature and spatial distribution patterns of
selected trace metals in the ~63-mm fraction of bed sedi-
ment from an urban river system, Manoa Stream, Oahu,
Hawaii. A secondary objective is to provide some prelim-
Cases and solutions
Fig. 1
Study area location map,
Manoa drainage basin, Oahu,
Hawaii. Note every tenth
sample site is indicated by 0;
this represents a distance of
approximately 500 m
inary data on phase associations of selected metals in se-
diments, i.e. organic bound, reducible, etc. It is assumed
that such a study will answer questions related to poten-
tial source areas of contamination and provide an expla-
nation for the highest Pb concentrations in fish of any
US stream or lake sampled by the NCBP.
Study area
Manoa drainage basin geomorphology and
hydrology
Manoa basin is located in southeast Oahu (Fig. 1) and is
drained by a third-order perennial stream; with a plani-
metric area of about 15 km 2. Two headwater sub-basins,
here termed Waihi (west) and Waiakeakua (east), drain
Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag 613
 Cases and solutions
areas of 2.75 and 2.90 km 2, respectively. The Waihi sub-
basin is the least disturbed, and closest to a “pristine”
area in this part of Honolulu with about 98% of the area
covered by forest.
A distinct rainfall gradient exists in the basin, and over a
distance of less than 8 km the annual rainfall, based on
National Oceanic and Atmospheric Administration
(NOAA 1980–1996) data, increases from 97B24 cm (at
24 m above sea level, ASL), to 407B72 cm (at 152 m ASL)
in the Waihi sub-basin. Manoa Stream is designated In-
land Waters-Class 2 in the Hawaii Administration Rules,
Title 11, as waters that are to be protected for recreation-
al purposes, propagation of fish, shellfish and aquatic life,
agricultural and industrial water supplies, shipping, and
navigation (Hawaii Department of Health 1990).
Typically Manoa is a confined channel, either naturally
by steep valley side-walls or in some locations artificially
by channelization, and thus has little associated flood-
plain. Above the US Geological Survey (USGS) gauging
stations (Fig. 1) the channels have low sinuosity and are
best described as having a cascade-type channel morpho-
logy (Montgomery and Buffington 1997) composed typi-
cally of gravel and boulders. Within the cascade portion
of the stream, depositional locations of fines are asso-
ciated with large obstacles, but their spatial extent is lim-
ited. Below the USGS stations the channel is more sin-
uous and is typically composed of a heterogeneous mix
of sediment sizes with large basalt boulders spalled from
the surrounding valley walls. Depositional bars including
point bars and obstacle related mid-channel bars are fre-
quent in the lower stream sections.
Manoa soils and geology
Manoa Valley is carved into the remnants of the Koolau
Shield Volcano that forms the eastern spine of Oahu.
Koolau Volcano is composed primarily of shield-type
tholeiite basalts (Koolau Basalt) with the subaerial part
constructed between 1.8 and 2.7 Ma ago (Lanphere and
Dalrymple 1980; Frey and others 1994). The Honolulu
Volcanic Series represents ~1% of the Koolau Shield
Volcano, and comprises the post-erosional volcanism on
the shield, with the lavas and tuffs ~1 Ma (Clague and
Frey 1982). These more recent volcanics are characterized
by a nephelinitic suite of rocks (Clague and Frey 1982).
The Sugarloaf vent (37; ca. 67 ka) is the one that is most
relevant to this study since lava flows spread out on the
floor of Manoa Valley to form a broad, nearly level sur-
face from the basin outlet to about 2.8 km upstream.
The soils that have developed in Manoa Valley have
formed on rocks of very different ages. Six soil orders
have been identified from maps of Manoa Valley pub-
lished in Foote and others (1972); the areal abundance of
soil orders in Manoa Valley follows: Entisol (45%) 1 In-
ceptisol (28%) 1 Ultisol (20%) 1 Oxisol (5%) 1 Mollisol
(1%) 1 Vertisol (~1%).
Manoa land use and population
The population of Manoa Valley in the 1930s was esti-
mated by Coulter and Serrao (1932) to be 5000 with 1000
614 Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag
homes. In 1990 the population was 20834 with 6420
households (State of Hawaii 1996). Ikeno (1996) exam-
ined a series of aerial photos and noted that there had
been limited land use change in Manoa since 1953, with
residential housing, parks and communities formed and
set by this date. Chun and others (1972) found less than
an 8% change in urban development in Manoa between
1960 and 1971, and commented that urbanization had al-
most reached its maximum capacity. Conservation and
undeveloped lands account for about 55–60% of the ba-
sin, commercial developments about 2–3%, education
and parks about 10% and the remainder is residential
area.
Based on land use, traffic intensity, anthropogenic activi-
ties, size and frequency of storm sewer outlets draining
directly into the stream, Manoa was divided into four
separate stream zones. Moving downstream the first zone
is termed “undisturbed” as it is the least impacted by hu-
man activity and this encompasses 15 sample sites (num-
bers 102–116) in the headwaters of Waihi sub-basin, i.e.
5.1–5.8 km from the basin outlet. The second zone is de-
signated residential and includes sites 54–101 (np49, in-
cluding tributary and storm drain samples) between 2.7
and 5.1 km above the basin outlet. This zone starts at the
junction of Waihi and Waiakekua Streams and stops up-
stream of a gasoline station in the valley and the major
commercial development in the basin (Manoa Market-
place). This zone is characterized by increased residential
land use, increased traffic and an increased number of
storm drains entering the stream banks or draining from
bridges. The upper two zones are dominated almost ex-
clusively by the Koolau Basalts. The next zone is termed
“commercial-institutional” (comm-inst) and includes 42
samples (sites 16–53) from 0.8–2.7 km above the outlet.
This zone is characterized by increased traffic, the most
significant commercial development in the basin, and a
variety of educational institutes including the University
of Hawaii. The final segment is termed the “lower” zone,
and includes 17 samples (sites 0–15) and includes loca-
tions below Dole Street bridge and the main portion of
the University of Hawaii. Dole Street is a major traffic ar-
tery in the basin and the largest storm outlet that drains
the highly-travelled western portion of the valley near the
university debouches just upstream of the bridge and in-
fluences the lower zone.
Materials and methods
Sample collection and processing
Systematic sampling was used to collect bed sediment at
50-m intervals from the outlet of Manoa Stream during
baseflow on three successive days. At six randomly se-
lected sites pH was measured in the field using a digital
meter. A total of 117 sites were included in the bed-sedi-
ment sampling array for a total distance of 5.8 km. Sam-
ples from depositional zones in the main third-order
channel and second-order segment of Waihi Stream were
 Cases and solutions

collected. Samples from three locations could not be col-
lected due to concrete (box-type) channelization. In addi-
tion to the main stream channel sites, samples from two
minor ephemeral tributaries were collected as they en-
tered the main channel, and samples from the mouth of
seven storm drains entering the main channel.
A clear 10-cm internal diameter Plexiglas tube sampler
was used to collect three separate 5-cm cores at each
sampling site. Cores were composited, double-bagged and
mixed thoroughly prior to transport to the laboratory. All
samples were cold stored at about 5 7C prior to proc-
essing. Several splits were made using a Teflon spatula in
Teflon pans from each sample, for particle size analysis,
organic and inorganic carbon (see Sutherland 1999) and
metal analyses. The split for element analyses was placed
in acid-washed beakers and dried for 48 h at 40 7C. Sam-
ples were lightly crumbled with latex gloves to break up
weak aggregates and dry-sieved through acid washed ny-
lon sieves. Data for the ~63-mm fraction are reported
here as this is the most common fraction analyzed for
trace metals in the literature (e.g. Shelton and Capel
1994).
Analytical techniques
Bed sediment samples ~63 mm (mass about 0.5 g) were
ground using a mixer-mill with tungsten carbide bowl
and balls for 5 min. Hickson and Juras (1986) indicate
tungsten carbide surfaces can significantly contaminate
analyses for cobalt (Co) and tungsten (W). Concentra-
tions of W in the analyzed samples were below detection
limits (~10 mg kg –1) and are thus not reported. Data for
Co are not analyzed in detail due to potential for con-
tamination. Ground samples were digested to dryness
overnight on a hot plate using trace metal grade nitric
(HNO3), perchloric (HClO4) and hydrofluoric (HF) acids.
The residue was solubilized with hydrochloric (HCl) acid
and diluted to volume. Trace metals (except Pb) were de-
termined by inductively coupled plasma-atomic emission
spectrometry (ICP-AES) and Pb by flame atomic-absorp-
tion spectrometry (FAAS). Cadmium and Se were deter-
mined for 14 randomly selected stream samples using ul-
trasonic nebulization ICP after the extraction into a solu-
tion of Aliquat 336 (a quaternary alkyl ammonium chlo-
ride) in MIKB (methyl isobutyl ketone). Mercury was de-
termined for 14 samples following aqua regia digestion
and measured with cold-vapor AAS (CVAAS).
Six bed sediment samples were selected from the lower
three stream zones and subjected to single leaches with
different reagents (Table 2) and elements were deter-
mined by ICP-mass spectrometry (ICP-MS). The single
leach philosophy was considered as a viable alternative to
sequential leaches and has been previously used by
Young and others (1992), Tack and others (1996) and
Tack and Verloo (1999). The “operationally defined”
phases included: (1) soluble, exchangeable, adsorbed and
coprecipitated with carbonate (SAEC phase; cf. Hall and
others 1996); (2) labile organic bound (organic phase);
(3) Mn and amorphous Fe oxyhydroxides (reducible
phase) and this was estimated by subtracting the NaOAc-
extractable contents from the NH2OH0HCl-extractable
contents (Tack and others 1996); and (4) the residual
phase was estimated from the total digestion contents
minus SAEC phasecorganic phasecreducible phase.
Quality control procedures for elements in bed sediment
included analyses of two standard reference materials
(SRMs), National Institute of Standards and Technology

Table 2
Single leach extractions and derived phase associations

Reagent Experimental conditions Target phase(s) released

1 M sodium acetate (NaOAc) in 3% 1.0 g sample, 15 ml reagent solution, Soluble, exchangeable, adsorbed
acetic acid (CH3COOH) at pH 5 shake for 2 h, centrifuge (2800 rpm) coprecipitated with carbonates (SAEC)
5 min, add 100 ml of 5 mg kg –1 Rh as
internal standard, dilute supernatant to
volume (10 ml)
0.1 M sodium pyrophosphate (Na4P2O7) 1.0 g sample, 25 ml reagent solution, Labile organic component (organic)
shake for 1 h, remaining protocol as
above
0.25 M hydroxylamine hydrochloride 1.0 g sample, 20 ml of reagent solution, Mn and amorphous Fe oxyhydroxides
(NH2OH 0 HCl) in 0.25 M HCl 60 7C water bath for 2 h, remaining (reducible) a
protocol as above
Concentrated HNO3, HCl, HF and HClO4 1.0 g sample, 2 ml of HNO3 and heat at Total digestion (all phases) b
200 7C, cool, 2 ml of HCl, heat in water
bath 90 7C, cool, add 5 ml HF, 3 ml
HClO4 and 2 ml HNO3, evaporate at
70 7C, add 3 ml HNO3 and dilute to
20 ml
a b
Reducible phase calculated as the difference between the Residual phase was determined as the difference between the
hydroxylamine hydrochloride leach and the sodium acetate total digestion and the summation of SAECc organicc
leach reducible phases. Therefore the residual phase includes
crystalline and residual oxides, sulfides and silicates

Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag 615

 Cases and solutions

Table 3
Elemental analyses, precision and accuracy for geochemical standard reference samples measured in this study

Element Units STSD-2 a STSD-2 STSD-2 STSD-2 SRM 2704 b SRM 2704 SRM 2704 SRM 2704
X̄ c X̄ d Precision e Accuracy f X̄ g X̄ h Precision e Accuracy f
This study Certified CV (%) (%) This study Certified CV (%) (%)

Al % 8.69 8.52 2.2 c 2.0 6.16 6.11 3.2 c 0.8

Ba mg kg –1 480 540 5.5 P10.7 380 414 7.1 P 8.0
Be mg kg –1 4.0 P 3.1 P 1.5 1.6 6.7 P 6.3
Ca % 2.90 2.86 3.0 c 1.4 2.48 2.60 3.4 P 4.6
Cd mg kg –1 P P P P 3.46 3.45 3.7 c 0.3
Co mg kg –1 16 19 5.6 P14.2 11 14.0 5.8 P22.1
Cr mg kg –1 107 116 4.7 P 7.8 127 135 5.4 P 5.9
Cu mg kg –1 49 47 5.1 c 3.2 96 98.6 4.1 P 2.5
Fe % 4.97 5.2 3.2 P 4.4 3.96 4.11 3.8 P 3.6
Hg mg kg –1 P P P P 1330 1470 7.5 P 9.5
K % 1.98 1.76 3.5 c12.5 2.12 2.00 4.7 c 6.0
Mg % 2.02 1.88 4.2 c 7.4 1.25 1.20 4.1 c 4.2
Mn mg kg –1 970 1060 3.1 P 8.9 530 555 4.0 P 4.7
Ni mg kg –1 60 53 4.2 c14.0 45 44.1 5.9 c 2.7
P mg kg –1 1330 1400 3.1 P 5.0 870 998 3.7 P13.0
Pb mg kg –1 63 66 3.7 P 5.3 149 161 4.8 P 7.5
Se mg kg –1 P P P P 1.2 1.12 4.3 c 7.1
Sr mg kg –1 499 400 6.0 c24.8 148 135 i 5.7 c 9.6
Ti % 0.45 0.49 4.1 P 7.6 0.41 0.457 4.9 P10.3
V mg kg –1 104 101 5.0 c 2.8 95 95 5.4 c 0.3
Zn mg kg –1 240 246 3.3 P 2.4 421 438 3.3 P 3.9
a f
STSD-2 is a geochemical stream sediment reference sample Accuracy (bias) is estimated using the equation of Yeh and
from the CCRMP others (1996) and Gill and Ramsey (1997):
b
SRM 2704 is a standard reference material (Buffalo River X̄this studyPX ¯ Certified
sediment) issued by the NIST
c
3 X̄Certified4 0 100
X̄ represents the arithmetic mean for 17 duplicates g
Based on 18 duplicates
d
Certified values as reported by Lynch (1990) h
Represents the arithmetic mean as tabulated by Epstein and
e
Precision is estimated by the coefficient of variation others (1989)
CVp[SD/X̄] 0 100 i
Arithmetic mean reported by Wilson and others (1994) using
ICP-AES for 18 samples

(NIST) SRM 2704 collected from Buffalo River, New York
(Epstein and others 1989) and the Canadian Certified
Reference Materials Project (CCRMP) STSD-2 collected
from Hirok Stream, British Columbia. Accuracy and pre-
cision for all elements are shown in Table 3, and these
are based on 18 duplicate samples of SRM 2704 and 17
duplicates of STSD-2, with the exception of Cd, Hg and
Se where only four duplicates of SRM 2704 were in-
cluded. Additionally, 24 field duplicate samples (about
5% of the total sample sites) were interspersed randomly
within sample batches. Samples were analyzed in a ran-
dom order, and not in sequence from outlet to headwat-
ers to reduce any possibility of instrument drift.
Precision for all elements from the SRM samples was ex-
cellent with median coefficients of variation of 3.9% for
STSD-2 (maximum 6.0% for Sr) and 4.7% for SRM 2704
(maximum 7.5% for Hg). The maximum coefficient of
variation (CV) for field duplicates was 6.0% (Sr) for
STSD-2, and 7.8% (Ti) for SRM 2704. Instrumental accu-
racy was also very good for most elements with a median
value of –3.4% for STSD-2 (range –14.2% Co to c24.8%
for Sr); and –3.9% for SRM 2704 (range from –22.1% for
Co to c9.6% for Sr).
Baseline element concentrations, min–max
standardization, and modified index of
geoaccumulation
For each sample site up to 21 elements were analyzed.
Only a select number of these will be discussed in detail
in this study since many are not anthropogenically en-
hanced in the Manoa system. Concentration statistics
(dry mass basis) will be presented for comparison to oth-
er studies, and for comparisons to estimated baseline
data. Selected element concentrations were standardized
using minimum and maximum values for a particular
element, such that several elements with widely differing
concentrations could be compared on the same scale with
a range between 0 and 1. The following equation was
used to constrain elemental data:
CnPCn Minimum
MMSn p 3C
n MaximumPCn Minimum
4 (1)
where MMSn represents the min–max standardized value
of element n in the ~63 mm bed sediment fraction, Cn is
the concentration of element n, Cn Minimum is the mini-
mum concentration of element n, and Cn Maximum is

616 Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag

 Cases and solutions

Table 4 3–4pstrongly polluted; 4–5pstrong to very strongly pol-

Best estimates of baseline metal concentrations (mg kg –1, luted; and 1 5pvery strongly polluted.
except Alp%) for the Koolau Basalts and the Honolulu
Volcanics (Sugarloaf flow) in Manoa drainage basin
Trace metal normalization and enrichment ratios
Element Koolau Honolulu a
References Loring (1990) states that grain size and provenance are
Basalts Volcanics two of the most significant parameters controlling the se-
dimentary variability of trace metals which must be com-
Al 7.72 5.83 1, 2, 3, 4, 5, 6 pensated for by any normalization procedure. Grain size
Ba 100 250 (?) b 1, 3, 7, 8 is controlled in this study since all data reported are
Cd 0.22 0.32 6, 8, 9
Cu 100 100 7, 10, 11, 12 ~63 mm. Geochemical normalization of trace metal data
Hg 0.05 0.05 13, 14 to a variety of “conservative” elements is common. These
Ni 190 240 1, 2, 3, 4, 8, 12 elements are assumed to have had a uniform flux from
Pb 6 7 7, 8, 11, 12, 15, 16 crustal-rock sources, from the time sediment particles
Zn 110 180 1, 3, 4, 7, 12, 15 were eroded until the time sediment particles were de-
a
posited, over a long period of time; consequently, com-
1 Roden and others (1984); 2 Leeman and others (1980); 3 pensation for changes in levels of various diluents can be
Frey and others (1994); 4 Clague and Frey (1982); 5 Macdonald
(1968); 6 Macdonald and others (1973); 7 Taylor (1964); 8 Li made (Horowitz 1991). Previously used conservative ele-
(1996); 9 Turekian and Wedepohl (1961); 10 De Carlo and ments include Al (Trefry and others 1985; Schropp and
Spencer (1995); 11 Ashwood and others (1989); 12 Wilkinson others 1990); Cr (McMurtry and others 1995); Fe (White
and Stolz (1983); 13 Adriano (1986); 14 Raine and others (1995); and Tittlebaum 1985); Li (Loring 1990); Mn (Loska and
15 Heinrichs and others (1980); 16 Spencer and others (1995) others 1997); Sc (Lee and others 1994); and Ti (Horowitz
b
Literature values for Ba concentrations for the Honolulu and others 1988). Horowitz (1991) states that normaliza-
Volcanics are unreliable and represent potential contamination
tions of this type entail the determination of a simple ra-
by felsic, pegmatoidal segragations in mililitite (Wilkinson and
Stolz 1983); thus, a value reported by Taylor (1964) is tio in bed sediment:
substituted
Concenration of trace metal
(3)
Concentration of conservative element
Aluminum was used as a conservative element in this
the maximum concentration of element n. A wide array study since it was the most accurately and the most pre-
of literature was reviewed to establish best estimates of cisely measured major element, and also it has been the
baseline concentrations for selected elements for Koolau most commonly used normalizing element in the geo-
Basalt and Honolulu Volcanics (Sugarloaf flow) are pre- chemical literature.
sented in Table 4. Eq. 3 was adapted to form an enrichment ratio (ERn) for
The index of geoaccumulation, a quantitative measure of a given element n following the work of Lee and others
the degree of pollution in aquatic sediments, was devel- (1994) and Lee and others (1997) and is defined as:
oped in the late 1970s and has been widely used in Euro-
pean trace metal studies (cf. Förstner and others 1990). A Cn Sample
modified index of geoaccumulation (MIgeo) is presented
ERn p
3C
Al Sample
4 (4)
in this study: BEn Background
Cn 3
BEAl Background 4
MIgeoplog2 3 1.5 0 BE 4
n
(2)
where Cn is as before; CAl is the aluminum concentration
where Cn is as defined previously; BEn is the best esti- in the ~63 mm bed sediment fraction; BEn Background is
mate of element n in the Koolau Basalts or the area in- as before and is based on values shown in Table 4; and
fluenced by the Sugarloaf flow. The BEn values for the BEAl Background is the best estimate of the concentration
Koolau Basalts in Table 4 were applied to bed sediment of Al in the Koolau Basalts or Honolulu Volcanics. As
sample sites 56–116 (from 2.8 to 5.8 km upstream of the with the MIgeo computation, sites downstream of 56 rep-
basin outlet), below this the Sugarloaf flow outcropped. resent a potential mix of Koolau Basalts and Sugarloaf
Therefore for sites 0–55 (0–2.8 km) the BEn values were flow, thus the Al background for these sites was the
weighted according to an assumed 50% contribution of weighted average of both rock types (i.e. 6.78%).
Koolau Basalts and 50% Honolulu Volcanics. The factor There is no accepted pollution ranking system or catego-
1.5 in Eq. 2 is used because of the possible variations of rization of degree of pollution based on the enrichment
the baseline (BEn) data due to lithogenic effects (cf. Salo- ratio methodology so a preliminary five-category system
mons and Förstner 1984). The following descriptive clas- is proposed:
sification is given for the index of geoaccumulation by 1. ER~2 Depletion to minimal enrichment suggestive of
Förstner and others (1990): ~0ppractically unpolluted; no or minimal pollution
0–1punpolluted to moderately polluted; 1–2pmoderate- 2. ER 2 – 5 Moderate enrichment, suggestive of moderate
ly polluted; 2–3pmoderately to strongly polluted; pollution

Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag 617

 Cases and solutions

3. ER 5 – 20 Significant enrichment, suggestive of a sig-

nificant pollution signal
4. ER 20 – 40 Very highly enriched, indicating a very
strong pollution signal
5. ER 1 40 Extremely enriched, indicating an extreme pol-
lution signal

Quantitative analyses
A variety of descriptive statistics will be presented to de-
scribe concentration data and ratio data (i.e. MMSn, MI-
geo and ERn). This will allow comparisons to be made be-
tween data in this study, those previously published, and
those to be conducted in the future. Spearman’s non-pa-
rametric rank order correlation coefficient (rs) was used
to describe the degree of bivariate monotonic association,
since distributions of a number of metals was non-nor-
mal. Fig. 2
Spatial patterns in selected trace metal data were Minor element geometric mean concentrations (B one
smoothed using robust locally weighted regression esti- standard deviation) for bed sediment (~63 mm) samples from
mates, lowess (Cleveland and Devlin 1988). Lowess is typ- Manoa Stream. The four zones are: L Lower; C-I
Commercial-Institutional; R Residential, and U Undisturbed.
ical of smoothers that emphasize large scale patterns in For a given element bars with the same letter are not
data by reducing the ‘noise’ associated with individual significantly different at ap0.05
sampling points. This was considered to be a useful ap-
proach for examining trends in trace metal bed sediment
data for the four zones defined for Manoa Stream.
One-way analysis of variance (ANOVA) was used to test
for differences in metal concentrations or pollution in- icantly lower values in the undisturbed headwater section
dices on log10 transformed data for the four zones. If a and highest in the lower zone near the basin outlet
significant F-ratio (Prob ^ ap0.05) was found, post-hoc (Fig. 2 and Table 5). Such a pattern may reflect increased
multiple comparison testing was performed using Fisher’s vehicle traffic, and increased road runoff of vehicle-asso-
protected least significant difference (PLSD) test. ciated metals as we proceed from the undisturbed head-
water area to the highly impacted lower stream reach.
Though other sources can not be totally disregarded it is
likely their importance would be localized and could not
account for the general trends identified.
Results and discussion Several studies have found strong associations between
bed sediment metal concentrations (Pb and Zn) and fish
During the three-day measurement period baseflow was metal concentrations with vehicle traffic density (e.g.,
alkaline with pH values between 7.5 and 7.9, with a me- Van Hassel and others 1980; Heiny and Tate 1997). Addi-
dian of 7.6. These values are similar to others previously tionally, studies establishing links between Pb and Zn
reported for Manoa Stream by Ching (1972) and Ikeno from automobiles are multitudinous. McMurtry and oth-
(1996). In aquatic environments with alkaline geochemis- ers (1995) estimated that motor vehicle sources account
try it is common to find low dissolved trace metal con- for at least 97% of the Pb burden of Oahu coastal sedi-
centrations, and this is supported by the data available ments with the remaining sources (volcanic emissions
for Manoa Stream (Yim and Dugan 1975; DOH 1980; and exogenic atmospheric transport) at most contribut-
Hasan 1991). ing 3%. Previous contributions of Pb to the environment
from automobiles is not debatable, and numerous sup-
Concentration data and min–max standardized porting data exist (cf. Jernigan and others 1971; Pierson
values (MMSn) and Brachaczek 1983; Lowenthal and others 1994; Cadle
A detailed statistical summary for selected trace metals and others 1997; Lee and others 1997).
and Al for Manoa bed sediments are given in Table 5. A The median Pb concentration found in the undisturbed
plot of geometric mean concentrations by zone (Fig. 2) zone was 10 mg kg –1, or about 10 times less than that
indicates that Ni is the only trace element that exhibits found in the lower zone (Table 5). Concentrations in the
the highest concentration in the undisturbed section of mid-portion of the basin were intermediate between the
Manoa. This may suggest that Ni concentrations in Ma- upper and lower reaches. Lead data from the bed sedi-
noa bed sediments are primarily controlled by mineralo- ments in the undisturbed zone were only slightly greater
gy; and this would support De Carlo and Spencer’s (1995) than the baseline value adopted for the Koolau Basalts at
contention for a similar control for nearby Ala Wai Canal 6 mg kg –1 (Table 4). This minor increase, and also that
sediments. Barium, Pb and Zn exhibit a pattern of signif- displayed by Zn, may be accounted for by one or more of

618 Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag

 Cases and solutions

Table 5
Descriptive statistics for selected metal concentrations in Manoa
Stream bed sediments for four zones (all concentrations are
mg/kg, except Hgpmg/kg, and Alp%) The number of samples
analysed per zone for Al, Ba, Cu, Ni, Pb and Zn are: Lower 17;
Commercial-Institutional 42; Residential 49; and Undisturbed
15. The number of samples analysed per zone for Cd and Hg
are: Lower 2; Commercial-Institutional 7; Residential 3; and
Undisturbed 2

Metal / Zone MeanBSD Geometric Median 25% 75% Minimum Maximum

Mean

Al
Lower 8.94B0.35 8.93 8.98 8.82 9.10 8.05 9.47
Comm-Inst 9.24B0.57 9.22 9.16 8.74 9.67 8.26 10.65
Residential 9.48B0.54 9.47 9.44 9.24 9.76 8.05 10.90
Undisturbed 9.60B0.44 9.59 9.56 9.42 9.86 8.72 10.65
Ba
Lower 277B52 273 265 240 309 221 430
Comm-Inst 226B71 220 210 199 221 177 541
Residential 202B35 200 199 188 210 154 408
Undisturbed 184B41 180 188 146 229 121 243
Cd
Lower 0.97B0.01 0.97 0.97 0.96 0.97 0.96 0.97
Comm-Inst 0.73B0.22 0.69 0.78 0.68 0.81 0.30 1.01
Residential 0.86B0.29 0.82 1.01 0.65 1.03 0.53 1.04
Undisturbed 0.40B0.17 0.38 0.40 0.28 0.52 0.28 0.52
Cu
Lower 188B29 186 181 177 198 133 258
Comm-Inst 193B20 192 191 183 199 170 300
Residential 206B22 205 199 189 220 172 267
Undisturbed 174B17 174 173 162 187 150 212
Hg
Lower 176B6 176 176 172 180 172 180
Comm-Inst 103B31 98 113 76 125 56 138
Residential 117B41 111 137 87 142 70 144
Undisturbed 51B32 45 51 28 73 28 73
Ni
Lower 317B44 314 307 292 315 278 439
Comm-Inst 316B29 314 315 303 329 216 408
Residential 324B15 324 323 317 331 289 360
Undisturbed 341B24 340 341 320 360 302 379
Pb
Lower 97B31 93 95 72 104 59 175
Comm-Inst 53B24 49 50 45 59 10 175
Residential 84B153 56 52 40 71 17 1078
Undisturbed 13B7 11 10 10 14 5 28
Zn
Lower 274B42 271 276 242 286 210 368
Comm-Inst 246B40 243 242 226 248 180 444
Residential 262B48 258 252 240 267 208 510
Undisturbed 202B29 200 196 181 223 162 268

the following sources: atmospheric deposition of locally
derived (anthropogenic) exhaust emissions, natural depo-
sition from volcanic emanations, or from deposition of
anthropogenically derived exogenic aerosols transported
from distant locations to the Hawaiian islands. Regardless
of source, Pb contributions to the undisturbed portion of
the basin were minimal and inconsequential when com-
pared to the concentrations in sediments in the more
heavily impacted downstream zones of Manoa Valley.
Limited Pb and Zn data are available for soils in Manoa
Valley. Data in Li (1996; 0–10 cm depth) indicate that the
lowest concentrations for these metals occurred in areas
associated with somewhat limited anthropogenic distur-
bance, i.e. Waiakekua sub-basin (Pbp68 mg kg –1;
Znp182 mg kg –1) and Waahila Ridge State Recreation
Area (Pbp38 mg kg –1; Znp185 mg kg –1; see Fig. 1). High
concentrations of Pb (838 mg kg –1) and Zn (972 mg kg –1)
were reported for surface soils associated with significant
automobile traffic in the middle of Manoa Valley. Lead
and Zn bed-sediment concentration data have been re-
ported for only three sample sites in Manoa Stream by
De Carlo and Spencer (1997); however, they support the
more detailed trends identified in this study. Lead data
from the State of Hawaii DOH (Eugene Akazawa, person-
al communication 1991) indicated 10 bed sediment sam-
ples (grain size fraction not specified) collected in 1991

Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag 619

 Cases and solutions

from Manoa Stream had a geometric mean of 41B3 mg

kg –1, with a range of values from 4 –198 mg kg –1. Con-
centrations for Pb and Zn in road-related runoff sedi-
ment samples have been reported by the Hawaii DOH
(1980). Sediment collected in roadside runoff boxes from
an Oahu freeway had Pb and Zn concentrations of 4250
and 338 mg kg –1, respectively. These can be compared to
those reported for a residential area in Manoa Valley,
Pbp1730 mg kg –1 and Znp159 mg kg –1.
All available data supports a strong anthropogenic signal
for Pb and to a lesser degree Zn. The spatial patterns are
strongly suggestive of vehicle contributions even though
Pb in gasoline has recently been phased out in the US. It
is clear that the Pb and Zn levels present in Manoa
Stream sediment reflect physical remobilization through
erosion of polluted soils and sediments from basin slopes
and transport via surface runoff to the stream channel.
Once sediments enter Manoa Stream the sediment sorbed
metals undergo a circuitous path downstream as material
is deposited, reworked and transported to eventually be
deposited in the Ala Wai Canal.
Lead data for the residential section showed a rather
large discrepancy between the arithmetic mean of 84 mg
kg –1 and the median of 52 mg kg –1 (Table 5). The differ-
ence resulted from the inclusion of a very high value for Fig. 3
sediment associated with a small tributary near site 100 Spatial variation in min-max standardized (MMSn) bed
(5.0 km upstream of the basin outlet). This “outlier” rep- sediment concentration values for (a) Cr, Ni and Ba; and (b)
resents a real value since two subsequent analyses on the Cu, Pb and Zn
remaining sample split agreed to within B7% of the val-
ue used in this study (i.e. 1078 mg kg –1). This large value ever, detailed Cu data reported in this study indicates a
greatly influenced the arithmetic mean since it is not ro- more complicated pattern, and argues against using such
bust to large (or small values); thus a clearer picture of small sample sizes in deciphering potential metal trends
central tendency of Pb in the four zones is gained from in streambed sediments.
examining the medians or geometric means. Field notes Copper data indicate some degree of anthropogenic en-
indicated that the extremely high Pb concentration for hancement in Manoa basin, but on their own they do not
the tributary site was associated with corroding pipes, appear to represent a major environmental concern.
probably Pb-based, dumped into the tributary. Another However, synergistic relationships of Cu with Ba, Pb and
potential source of the high Pb may be the leaking un- Zn may be worthy of further study.
derground storage tank located in Paradise Park (Facility Despite the limited number of analyses conducted to
IDp9–101887, Leak IDp860001) that is just upstream of characterize Cd and Hg concentrations, there appeared to
the tributary. be higher Hg concentrations in the lower zone of Manoa
Copper exhibited an unusual concentration pattern Stream (176 mg kg –1) compared to the 51 mg kg –1 in the
(Fig. 3, Table 5) with its lowest values in the undisturbed undisturbed zone (Table 5). It is interesting to note that
and lower zones, while its highest concentrations oc- the concentration of Hg in the undisturbed sediments is
curred in the residential and commercial-institutional almost identical to the baseline value for basalts extracted
zones. Such a pattern may reflect previous use of Cu- from the literature (cf. Table 4). McMurtry and others
based insecticides, fungicides and corrosion of Cu plumb- (1995) state that Hg in Oahu coastal sediments is pre-
ing in the residential area of Manoa Valley. These sources dominantly of volcanic origin, however maximum Hg
are in addition to those contributed by automobiles, contributions of motor vehicles to the total Hg burden is
since Cu concentrations in brake linings can reach 3%, about 43% for urban areas. These Hg data indicate that
tire rubberp247 mg kg –1 (Wigington and others 1983), even if the Hg concentrations reported in this study are
and Cu is also used as an anti-oxidant in gasoline. De associated with anthropogenic sources, their contribution
Carlo and Spencer (1997) argued that roughly half of the is minimal, and localized in the lower section of the ba-
Cu in the sediments of the Ala Wai Canal can be attri- sin.
buted to anthropogenic sources. They also suggested, Median Cd concentrations were higher downstream of
based on only three Manoa Stream bed sediment sam- the undisturbed zone by up to 2.5 times (Table 5). Con-
ples, that there was a trend of increasing Cu from the up- centrations in the three downstream zones (medians
per basin (146 mg kg –1), through the middle portion 0.78–1.01 mg kg –1) were similar in magnitude to the Cd
(186 mg kg –1), to the lower reaches (298 mg kg –1). How- concentrations in the upper 20 cm of sediment from two

620 Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag


cores in the Ala Wai Canal, i.e. range of Cd from
0.83–1.39 mg kg –1, medianp1.09 mg kg –1 for 9 samples;
De Carlo and Spencer (1997). These Cd data indicate only
minor enrichment above baseline values (Table 4), and
may reflect, at least, an automobile source.
Smoothed min-max standardized (MMSn) values (see
Eq. 1) indicate some interesting spatial patterns (Fig. 3a,
b). The pattern for MMSNi was unique and supports mi-
neralogical control, with possible dilution in the lower
section of the basin. Chromium is displayed since its pat-
tern is also unique among the trace metals, and it was
also mineralogically controlled with no apparent dilution
downstream. Barium, Pb and Zn display very similar
large scale patterns, with low values in the upper zone,
significant increases in the upstream portion of the resi-
dential zone and a general leveling out throughout the
commercial-institutional zone. The highest MMSn values
were located within the “lower” zone for Ba, Pb, and Zn.
These patterns and data in Fig. 2 and Table 5 seem to
suggest a common association for these three elements
and possibly a common source, i.e. the automobile (Ta-
ble 1). Barium is usually not discussed in the pollution
literature, the focus is typically on Pb and Zn. This may
stem from the lower toxicity of Ba in vertebrata com-
pared to Pb and Zn but, as Nieboer and Richardson
(1980) note, Ba is considered to be slightly more toxic to
mammalia than Zn. No soil or sediment Ba concentration
data have been reported for sediments in Manoa Valley
or the Ala Wai Canal. Thus, despite the complications ar-
ising from the problems in constraining the Ba content of
the Sugarloaf flow, Ba appears to have a minor anthro-
pogenic signal.
Correlation analyses for bed sediment elements
Correlation analysis was used to explore strength of asso-
ciation between a variety of elements and two important
sedimentary trace element collectors Mn oxides (actually
elemental Mn in this study) and organic matter (actually
organic C in this study). Two hachured zones are de-
picted in Fig. 4 and all elements within the vertical zone
are not significantly (ap0.05) correlated with organic C
content (Cr, Cu, and Zn); all elements within the hori-
zontal band are not significantly correlated with Mn (i.e.
Cr, Fe, K, Mg, Ni, P, Sr, Ti, and V). Additionally, Cd and
Hg, though they fall outside the vertical and horizontal
bands, are not significantly correlated with either organic
C or Mn (they are based on small sample sizes thus crit-
ical bands are much larger for these elements). Trace ele-
ments so far suggested as having at least a minimal an-
thropogenic signal (i.e. Ba, Cu, Pb and Zn) are positively
associated with Mn; Ba and Pb are negatively associated
with organic C. Despite significant correlations between
elements and Mn and organic C the coefficients of deter-
mination (r 2s) are all below 30%, possibly suggesting in-
fluences from other components not measured such as
amorphous Fe oxides, reactive Fe, extractable Fe, surface
area, and clay mineralogy (Horowitz 1991).
Inter-element nonparametric correlation coefficients are
shown in Table 6 only for those bivariate associations
Cases and solutions
Fig. 4
Scatterplot of element Spearman correlation coefficients with
Mn and organic C. Elements that fall within the horizontally
stippled area are not significantly correlated with Mn at
ap0.05. Elements that fall within the vertically line-stippled
area are not significantly correlated with organic C at ap0.05.
Cadmium and Hg are bracketed because they are based on a
sample size of only 14 (all other elements np123), and neither
is significantly correlated at ap0.05 with Mn or organic C
that are statistically significant (ap0.05). It is interesting
to note that all elements suggested to have a minimal an-
thropogenic signal are positively associated with Pb, in-
cluding in order of significance (i.e. the lower the proba-
bility value (Prob) the stronger the evidence against the
null hypothesis, and thus a stronger monotonic (bivar-
iate) statistical association): Zn (Prob~0.0001) 1 Ba (Prob
~0.0001) 1 Cu (Probp0.0006) 1 Cd (Probp0.0047) 1 Hg
(Probp0.0106).
Modified index of geoaccumulation (MIgeo)
Eq 2 was used to estimate MIgeo for seven metals and it is
clear from Table 7 that the Pb MIgeo values were the most
distinct. About 55% of the sample sites were classified as
moderately or strongly polluted (i.e. MIgeo 2–3). Four of
the five samples classified as strong to very strongly pol-
luted (MIgeo 4–5, and 1 5) were associated with storm
drain outlets and a tributary outlet influenced by metal
pipes dumped within the stream net. These Pb MIgeo data
support previously presented concentration and MMSn
values and are suggestive of a continued flux of Pb
sorbed to sediments entering Manoa Stream from urban
runoff.
The smoothed Pb MIgeo pattern is shown in Fig. 5, with
storm drain outlet and tributary samples identified sepa-
rately with arrows. Median Pb MIgeo values for each of
the four zones were 3.3 (lower) 1 2.5 (residential)p2.4
(commercial-institutional) 1 0.1 (“undisturbed”), with
ANOVA indicating no statistically significant differences
between the two mid-basin zones.
Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag 621
 Cases and solutions

Table 6
Spearman correlation coefficient matrix for concentrations of
selected elements for Manoa bed sediment ~63 mm. Only
correlations that where significant at a^0.05 are shown. For all
elements, except Cd and Hg, with correlations ~–0.23 or
1 c0.23 are statistically significant at ap0.01; and those with
correlations ~–0.30 or 1 c0.30 are statistically significant at
ap0.001. For Cd and Hg correlations ~–0.70 or 1 c0.70 are
statistically significant at ap0.01; and those with correlations
~–0.82 or 1 c0.82 are statistically significant at ap0.001,
since only 14 samples were analysed

Ba –0.23
Ca –0.58 0.31
Cd NS a NS NS
Cu 0.25 NS NS NS
Fe 0.40 NS –0.47 NS NS
Hg NS NS NS NS NS NS
K –0.41 0.48 0.68 NS NS –0.50 0.59
Mn 0.18 0.25 0.18 NS 0.41 NS NS NS
Ni 0.62 –0.20 –0.42 NS NS 0.30 –0.74 –0.42 NS
Pb –0.46 0.43 0.47 0.78 0.32 –0.19 0.71 0.58 0.21 –0.44
Ti 0.49 –0.21 –0.58 NS NS 0.86 NS –0.63 NS 0.43 –0.42
Zn NS 0.48 0.32 0.64 0.52 0.22 0.76 0.38 0.46 –0.27 0.60 NS
Org C 0.40 –0.25 –0.51 NS NS 0.35 NS –0.47 NS 0.28 –0.38 0.49 NS
Al Ba Ca Cd Cu Fe Hg K Mn Ni Pb Ti Zn
a
Spearman correlation coefficients are not significant at ap0.05

Table 7
Modified index of geoaccumulation percentages in each
pollution category for trace metals in Manoa Stream bed
sediments (~63 mm). Modified Index of Geoaccumulation
(MIgeo) values were determined using Eq. (2), with values ~0
practically unpolluted; 0–1 unpolluted to moderately polluted;
1–2 moderately polluted; 2–3 moderately or strongly polluted;
3–4 strongly polluted; 4–5 strong to very strongly polluted; 5
very strongly polluted (Förstner and others 1990)

Element ~0 0–1 1–2 2–3 3–4 4–5 15 Median a Minimum Maximum

Ba 43.1 (53) b 55.3 (68) 1.6 (2) – – – – 0.15 –0.57 1.44

Cd 14.2 (2) 42.9 (6) 42.9 (6) – – – – 0.97 –0.43 1.66
Cu 1.6 (2) 97.6 (120) 0.8 (1) – – – – 0.35 –0.17 1.00
Hg 35.7 (5) 50.0 (7) 14.3 (2) – – – – 0.65 –1.42 1.26
Ni 32.5 (40) 67.5 (83) – – – – – 0.11 –0.58 0.44
Pb 3.3(4) 9.7 (12) 12.2 (15) 55.3 (68) 15.4 (19) 3.3 (4) 0.8 (1) 2.42 –0.94 6.90
Zn 5.7 (7) 90.2 (111) 4.1 (5) – – – – 0.38 –0.27 1.63
a b
Median, minimum and maximum represent actual MIgeo Values in parentheses represent the actual number of samples,
values and not percentages with a total of 123 for Ba, Cu, Ni, Pb and Zn; and 14 for Cd
and Hg

Enrichment ratios (ERn)
Enrichment ratios for seven metals were calculated using
Eq. 4 and the data are summarized in Table 8. It is again
apparent that Pb contamination of bed sediments in Ma-
noa Stream is the most pronounced amongst the trace
metals investigated. Approximately 67% of the sites sam-
pled were classified as having a significant degree of Pb
pollution (ERPb 5–20), and three sites were extremely pol-
luted (ERPb 1 40). The median ERPb value for all bed se-
diments from Manoa stream was 6.2. These sediment
data, when considered with literature data indicating low
concentrations of Pb in Manoa Stream water, suggest that
the occurrence of high levels of Pb in fish from this sys-
tem (cf. Schmitt and Brumbaugh 1990) can be attributed
to the sediment that acts as an effective pollutant sink.
The spatial variation in bed sediment ERPb values is
shown in Fig. 6, with arrows identifying sediments asso-
ciated with storm drain outlets and small ephemeral tri-
butaries. Analysis of variance indicated that ERPb values
for each of the four zones where statistically different at
ap0.05, with the following median values: 11.4 (low-
er) 1 6.7 (residential) 1 5.8 (commercial-institutional) 1 1.3
(“undisturbed”). As with all other data presented there is
strong support for Pb enrichment, and the anthropogenic
signal is considered to be related with urban runoff pri-
marily carrying automobile associated pollutants. Howev-
er, other potential sources can not be ruled out, though
they are considered to be much less important.
Phase associations
Several reviews have been published on the “speciation”
or more accurately “operationally” defined phase associa-
tions of metals in soils and sediments (e.g. Chao 1984;
Pickering 1986; Kersten and Forstner 1995; Hall and oth-
ers 1996; Kennedy and others 1997). Many agree that
meaningful comparisons of results from separate studies

622 Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag

 Cases and solutions

Table 8
Enrichment ratio percentages in each pollution category for
trace metals in Manoa Stream bed sediments (~63 mm).
Enrichment ratio (ERn) values were determined using Eq. (4),
with ER values ~2 representing sites with no or minimal
pollution; 2–5 represents sites moderately polluted; 5–20
represents sites with significant pollution; 20–40 represents sites
with very strong pollution; and 1 40 represents those sites that
are extremely polluted

Element ER~2 ER 2–5 ER 5–20 ER 20–40 ER 1 40 Median a Minimum Maximum

Ba 95.1 (117) b 4.9 (6) – – – 1.32 0.74 3.68

Cd 42.9 (6) 57.1 (8) – – – 2.13 0.75 3.87
Cu 92.7 (114) 7.3 (9) – – – 1.45 1.06 2.46
Hg 64.3 (9) 35.7 (5) – – – 1.69 0.46 2.61
Ni 100.0 (123) – – – – 1.29 0.81 1.62
Pb 10.6 (13) 19.5 (24) 66.7 (82) 0.8 (1) 2.4 (3) 6.23 0.63 163.8
Zn 87.8 (108) 12.2 (15) – – – 1.48 0.81 4.18
a b
Median, minimum and maximum represent actual ER values Values in parentheses represent the actual number of samples,
and not percentages with a total of 123 for Ba, Cu, Ni, Pb and Zn; and 14 for Cd
and Hg

Fig. 5
Spatial variation in the lead modified index of geoaccumulation
(MIgeo) for 123 Manoa Stream bed sediment samples. The major
line through the data represents a lowess smoothed trace.
Arrows identify either storm drain outlet or minor tributary
outlet samples
Fig. 6
Spatial variation in the lead enrichment ratio (ERPb) for 123
Manoa Stream bed sediment samples. The major line through
the data represents a lowess smoothed trace. Arrows identify
either storm drain outlet or minor tributary outlet samples

is questionable because experimental conditions are

usually very different, such as reagents used, sequencing
of reagent application and preparation techniques (espe-
cially drying and storage protocols). But most would
agree that single or sequential leach studies provide valu-
able information on potential availability of metals to
biota, and also they are useful in deciphering the anthro-
pogenic “overprint” in a given area.
Figure 7 displays the median percentages for each of the
four phases examined for Manoa bed sediments ranked
from highest percent in the residual phase to the lowest.
The mean Pb concentration for the six samples was
45 mg kg –1, with a mean ERPb of about 7 times (range of
3–11 times). As expected the four elements commonly
used in normalization procedures (Ti, Cr, Al and Fe) had
high residual phase values ( 1 91%), with Ti about 99.5%.
For the trace elements discussed in detail in this paper
(Ba, Cu, Ni, Pb and Zn) their ordering in each phase was
as follows,
1. soluble, adsorbed, exchangeable and carbonate phase
(SAEC): Ba 1 Zn 1 Pb 1 Cu 1 Ni;
2. labile organic phase: Cu 1 Zn 1 Ba 1 Pb 1 Ni;
3. reducible phase: Pb 1 Zn 1 Ba 1 Cu 1 Ni; and
4. residual phase: Ni 1 Cu 1 Zn 1 Ba 1 Pb
These data are informative because they add further sup-
port to the ERn, MIgeo data and to the Pb data in whole
fish samples collected by NCBP in Manoa Stream. Pro-
portionally Ni was the least available of the five elements,
supporting the interpretation that it is mineralogically
controlled. Lead of all the elements examined had the
lowest percentage in the residual phase, confirming a
dominant anthropogenic signal. The non-residual phase
in the sediments would be potentially bioavailable when

Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag 623

 Cases and solutions
Fig. 7
Metal phase partitioning for 14 elements in the ~63 mm bed
sediment samples from the lower three zones of Manoa Stream.
SAEC represents the soluble, exchangeable, adsorbed
coprecipitated with carbonates phase
processed through the acidic gut fluids of benthic feeders; ment sorbed Pb (and other metals) to the stream is un-
since Pb had the highest percentage in the non-residual
phase 87% (SAECp3%, organicp1%, reduciblep83%) it zone, which in turn is a function of depth distribution
would be proportionally the most significant contributor and surface erosion rates. Nevertheless, it is clear that the
to the anthropogenic pollution signal in Manoa Stream.
The data from the leach experiment can also be used as
an independent check on the best estimates for baseline
values compiled from the literature for Manoa drainage
basin (Table 4). The median residual concentrations from ment. The non-residual phase of sediment ingested by
the leach experiment were Bap60, Cup105, Nip280,
Pbp6 and Znp110 mg kg –1. Values for Cu, Pb and Zn
were extremely close to those listed in Table 4 for the
Koolau Basalts and the Honolulu Volcanics. The leach
value for Ni was slightly higher than the best estimates
compiled in Table 4 (i.e. 190–240 mg kg –1). The compari- vanized metal associated with buildings, etc.) but their
son is weakest for Ba since the leach value was signifi-
cantly below the best estimates of 100–250(?) mg kg –1,
but is closest to the Koolau Basalt and within the range
of values reported by Frey and others (1994), 34–149 mg
–1
kg .
Conclusion
Stream bed sediments provide an important archive for
examining trace metals and identifying anthropogenic
pollution signals when detailed spatial investigations are
conducted. Bed sediment data (~63 mm) from Manoa
drainage basin, an urbanized system, in southeastern
Oahu, Hawaii, were analyzed for a variety of major and
624 Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag
trace elements. Concentration data, min-max standar-
dized data, the modified index of geoaccumulation, en-
richment ratios and single leach data all point to Manoa
bed sediments being primarily contaminated by Pb. Sedi-
ments with the highest Pb concentrations were located in
the lower zone of the basin (0–0.8 km upstream of the
basin outlet) and also associated with storm drain outlets
and minor tributaries. The lowest Pb concentrations, near
background levels, were measured in bed sediments of
the undisturbed zone in the Waihi sub-basin, 5.1–5.8 km
upstream of the basin outlet. Data for other trace metals
(Ba, Cd, Cu, Hg and Zn) indicate some degree of anthro-
pogenic enhancement for samples sites between 0 and
5.1 km upstream of the basin outlet, but none show the
magnitude of enhancement exhibited by Pb.
The primary source of Ba, Cd, Pb and Zn contamination
in Manoa Stream is considered to be the automobile, and
includes vehicle wear, tire wear, fluid output, and exhaust
emissions. Many of these sources exist today in Manoa
Valley, but the contemporaneous contribution of Pb from
vehicle emissions is negligible. However, the high Pb ve-
hicle fluxes associated with emissions in the 1960–1980s
have been stored in soils and sediments of the basin and
are now being remobilized by surface erosion processes
and contributed to the stream channel network. The ma-
jor sink for Pb in Manoa sediments is the reducible
phase which is dominated by Mn and amorphous Fe oxy-
hydroxides. The time frame for continued input of sedi-
known, but will depend on the residence time in the soil
high Pb concentrations (highest recorded in 117 rivers
and lakes sampled nationwide in the US) found in whole
fish samples collected from Manoa Stream as part of the
NCBP reflect ingestion of Pb-associated with bed sedi-
benthic feeders will undergo desorption and result in
trace metal bioaccumulation, especially for Pb. There is
potential for anthropogenic sources other than the auto-
mobile to contribute trace metals to Manoa Stream (i.e.
erosion of soil containing Pb-based paint, fertilizers, gal-
contributions are considered to be spatially limited.
Acknowledgements I would like to thank Mr. D. Ko and Dr. C-
T. Lee for providing invaluable assistance during the field-col-
lection phase of this project. The data supplied by Mr. E. Aka-
zawa and Dr. B. Brooks were greatly appreciated. Without the
facilitation of chemical analyses by Mr. J. Gravel, Mr. M. LeGeyt
and Mr. G. Marshall this project would not have been possible.
Constructive comments from anonymous reviewers have signifi-
cantly improved this manuscript, and their efforts are greatly
appreciated. Financial support for this work was provided by
the Department of Geography, University of Hawaii, and from
University of Hawaii Research Council Grant 3–89947.

References
Adriano DC (1986) Trace elements in the terrestrial environ-
ment. Springer-Verlag, Berlin Heidelberg
Alloway BJ (1995) Heavy metals in soils (2nd edn). Blackie
Academic and Professional, Glasgow
Ashwood TL, Olsen CR (1988) Pearl Harbor bombing attack a
contamination legacy revealed in the sedimentary record. Mar
Pollut Bull 19 : 68–71
Ashwood TL, Olsen CR, Larsen IL (1989) Sources and areal
distribution of trace metals in recent sediments of Middle
Loch, Pearl Harbor (Hawaii). (Environ Sci Div Pub Nr 3274,
ORNL/TM-11135), Oak Ridge National Laboratory, Oak
Ridge, Tennessee
Breault RF, Harris SL (1997) Geographical distribution and
potential for adverse biological effects of selected trace ele-
ments and organic compounds in streambed sediment in the
Connecticut, Housatonic, and Thames River Basins, 1992–94.
US Geol Surv Water Res Invest Rep 97–4169, Marlborough,
Massachusetts
Bryan GW, Langston WJ (1992) Bioavailability, accumulation
and effects of heavy metals in sediments with special refer-
ence to United Kingdom estuaries: a review. Environ Pollut
76 : 89–131
Cadle SH, Mulawa PA, Ball J, Donase C, Weibel A, Sage-
biel JC, Knapp KT, Snow R (1997) Particulate emission rates
from in-use high-emitting vehicles recruited in Orange Coun-
ty, California. Environ Sci Technol 31 : 3405–3412
Chao TT (1984) Use of partial dissolution techniques in geo-
chemical exploration. J Geochem Explor 20 : 101–135
Char AL (1977) Heavy metals survey and an estimation of se-
diment yield: Makua Military Reservation, Oahu, Hawaii.
(MSc Thesis) University of Hawaii
Chen Z, Mayer LM (1998) Mechanisms of Cu solubilization
during deposit feeding. Environ Sci Technol 32 : 770–775
Ching RKY (1972) Water quality effects of urban runoff. (MSc
Thesis) University of Hawaii
Chun MJ, Young RHF, Anderson GK (1972) Wastewater ef-
fluents and surface runoff quality. Water Resources Research Hawaii Department of Health (DOH, 1980) Storm water
Center, University of Hawaii, Tech Rep 63
Clague DA, Frey FA (1982) Petrology and trace element geo-
chemistry of the Honolulu Volcanics, Oahu: implications for
the mantle below Hawaii. J Petrol 23 : 447–504
Cleveland WS, Devlin SJ (1988) Locally weighted regression:
an approach to regression analysis by local fitting. J Am Stat Heinrichs H, Schulz-Dobrick B, Wedepohl KH (1980) Ter-
Assoc 83 : 596–610
Coulter JW, Serrao AG (1932) Manoa Valley, Honolulu: a
study in economic and social geography. Bull Geogr Soc Phil Heiny JS, Tate CM (1997) Concentration, distribution, and
30 : 109–130
De Carlo EH, Spencer KJ (1995) Records of lead and other
heavy metal inputs to sediments of the Ala Wai Canal, O’ahu,
Hawai’i. Pac Sci 49 : 471–491
De Carlo EH, Spencer KJ (1997) Retrospective analysis of an-
thropogenic inputs of lead and other heavy metals to the Ha- Horowitz AJ (1991) A primer on sediment-trace element
waiian sedimentary environment. Appl Organometall Chem
11 : 415–437
Dickinson WW, Dunbar GB, McLeod H (1996) Heavy metal
history from cores in Wellington Harbour, New Zealand. En-
viron Geol 27 : 59–69
Epstein MS, Diamondstone BI, Gills TE (1989) A new river
sediment standard reference material. Talanta 36 : 141–150
Eshleman AM (1973) A preliminary survey of lead and mercu- Jernigan EL, Ray BJ, Duce RA (1971) Lead and bromine in
ry in the Hawaiian environment. (PhD Diss) University of
Hawaii
Feltz HR (1980) Significance of bottom material data in evalu- Kabata-Pendias A, Pendias H (1992) Trace elements in soils
ating water quality. In: Baker RA (ed) Contaminants and se-
Cases and solutions
diments (vol 1), fate and transport, case studies, modelling,
toxicity. Ann Arbor Science Publishers, Michigan, pp 271–287
Foote DE, Hill EL, Nakamura S, Stephens F (1972) Soil sur-
vey of the Islands of Kauai, Oahu, Maui, Molokai, and Lanai,
State of Hawaii. USDA, Soil Conservation Service, US Govern-
ment Printing Office, Washington, DC
Förstner U, Ahlf W, Calmano W, Kersten M (1990) Sedi-
ment criteria development – contributions from environmen-
tal geochemistry to water quality management. In: Heling D,
Rothe P, Forstner U, Stoffers P (eds) Sediments and environ-
mental geochemistry: selected aspects and case histories.
Springer-Verlag, Berlin Heidelberg, pp 311–338
Frank R, Ishida K, Suda P (1976) Metals in agricultural soils
of Ontario. Can J Soil Sci 56 : 181–196
Frey FA, Garcia MO, Roden MF (1994) Geochemical charac-
teristics of Koolau Volcano: implications of intershield geo-
chemical differences among Hawaiian volcanoes. Geochim
Cosmochim Acta 58 : 1441–1462
Fu S-L, Hashimoto H, Siegel BZ, Siegel SM (1989) Varia-
tions in plant and soil lead and mercury content in a major
Honolulu park, 1972–1987, a period of significant source re-
duction. Water Air Soil Pollut 43 : 109–118
Gill R, Ramsey MH (1997) What a geochemical analysis
means. In: Gill R (ed) Modern analytical geochemistry an in-
troduction to quantitative chemical analysis techniques for
Earth, environmental and materials scientists. Addison Wes-
ley Longman, Essex, pp 1–11
Hall GEM, Vaive JE, Beer R, Hoashi M (1996) Selective
leaches revisited, with emphasis on the amorphous Fe oxyhy-
droxide phase extraction. J Geochem Explor 56 : 59–78
Harned DA (1988) Effects of highway runoff on streamflow
and water quality in the Sevenmile Creek Basin, a rural area
in the Piedmont Province of North Carolina, July 1981 to July
1982. (US Geol Surv Water Supply Paper 2329) US Govern-
ment Printing Office, Washington, DC
Hasan W (1991) Assessment of toxic metal and microbiological
contaminants of nonpoint sources along Manoa Stream. (Un-
published report) Public Health, University of Hawaii
runoff and urban street contaminants: intensive survey re-
port. State of Hawaii, USA
Hawaii Department of Health (DOH, 1990) Water Quality
Standards (Hawaii administrative rules title 11, chapter 54).
State of Hawaii, USA, 54–1 to 54–64
restrial geochemistry of Cd, Bi, Tl, Pb, Zn and Rb. Geochim
Cosmochim Acta 44 : 1519–1533
comparison of selected trace elements in bed sediment and
fish tissue in the South Platter River Basin, USA, 1992–1993.
Archiv Environ Contam Toxicol 32 : 246–259
Hickson CJ, Juras SJ (1986) Sample contamination by grind-
ing. Can Mineral 24 : 585–589
chemistry (2nd edn). Lewis Publishers, Michigan
Horowitz AJ, Elrick K, Callender E (1988) The effect of
mining on the sediment-trace element geochemistry of cores
from the Cheyenne River arm of Lake Oahe, South Dakota,
USA. Chem Geol 67 : 17–33
Ikeno DE (1996) Urban runoff in Manoa and Palolo Streams.
(MSc Thesis) University of Hawaii
atmospheric particulate matter on Oahu, Hawaii. Atmos En-
viron 5 : 881–886
and plants (2nd edn). CRC Press, Boca Raton, Florida
Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag 625
 Cases and solutions
Kennedy VH, Sanchez AL, Oughton DH, Rowland AP
(1997) Use of single and sequential chemical extractants to
assess radionuclide and heavy metal availability from soils for
root uptake. Analyst 122 : 89R–100R
Kersten M, Forstner U (1995) Speciation of trace metals in
sediments and combustion waste. In: Ure AM, Davidson CM
(eds) Chemical speciation in the environment. Chapman &
Hall, Glasgow, pp 234–275
Lagerwerff JV, Specht AW (1970) Contamination of road-
side soil and vegetation with cadmium, nickel, lead and zinc.
Environ Sci Technol 4 : 583–586
Lanphere MA, Dalrymple GB (1980) Age and strontium iso-
topic composition of the Honolulu Volcanic Series, Oahu,
Hawaii. Am J Sci 280-A:736–751
Lee DS, Garland JA, Fox AA (1994) Atmospheric concentra-
tions of trace elements in urban areas of the United King-
dom. Atmos Environ 28 : 2691–2713
Lee P-K, Touray J-C, Baillif P, Ildefonse J-P (1997) Heavy
metal contamination of settling particles in a retention pond
along the A-71 motorway in Sologne, France. Sci Total Envi-
ron 201 : 1–15
Leeman WP, Budahn JR, Gerlach DC, Smith DR, Powell
BN (1980) Origin of Hawaiian tholeiites: trace element con-
straints. Am J Sci 280-A:794–819
Li H (1996) Baseline study of heavy metal concentrations and
human activity impact analysis in the Hawaiian environment.
(MSc Thesis) University of Hawaii
Loring DH (1990) Lithium – a new approach for the granu-
lometric normalization of trace metal data. Mar Chem
29 : 155–168
Loska K, Cebula J, Pelczar J, Wiechula D, Kwapulinski J
(1997) Use of enrichment, and contamination factors together
with geoaccumulation indexes to evaluate the content of Cd,
Cu, and Ni in the Rybnik water reservoir in Poland. Water
Air Soil Pollut 93 : 347–365
Lowenthal DH, Zielinska B, Chow JC, Watson JG, Gau-
tam M, Ferguson DH (1994) Characterization of heavy-duty
diesel vehicle emissions. Atmos Environ 28 : 731–743
Lynch J (1990) Provisional elemental values for eight new geo-
chemical lake sediment and stream sediment reference mate-
rials LKSD-1, LKSD-2, LKSD-3, STSD-1, STSD-2, STSD-3 and
STSD-4. Geostandards Newslett 14 : 153–167
Macdonald GA (1968) Composition and origin of Hawaiian
lavas. In: Coats RR, Hay RL, Anderson CA (eds) Studies in
volcanology – a memoir in honor of Howel Williams. Geol
Soc Am Mem 116, Colorado, USA, pp 477–522
Macdonald GA, Powers HA, Katsura T (1973) Interlabora-
tory comparison of some chemical analyses of Hawaiian vol-
canic rocks. Bull Volcanol 36 : 1–13
Mayer LM, Chen Z, Findlay RH, Fang J, Sampson S, Self
RFL, Jumars PA, Quetel C, Donard OFX (1996) Bioavaila-
bility of sedimentary contaminants subject to deposit-feeder
digestion. Environ Sci Technol 30 : 2641–2645
McMurtry GM, Wiltshire JC, Kauahikaua JP (1995) Heavy
metal anomalies in coastal sediments of Oahu, Hawaii. Pac
Sci 49 : 452–470
Monaci F, Bargagli R (1997) Barium and other trace metals
as indicators of vehicle emissions. Water Air Soil Pollut
100 : 89–98
Montgomery DR, Buffington JM (1997) Channel-reach mor-
phology in mountain drainage basins. Geol Soc Am Bull
109 : 596–611
Moore JW, Ramamoorthy S (1984) Heavy metals in natural
waters – applied monitoring and impact assessment. Spring-
er-Verlag, Berlin Heidelberg
National Oceanic and Atmospheric Administration
(NOAA, 1980–1996) Climatological data – annual summary –
626 Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag
Hawaii and Pacific. National Climate Data Center, Asheville,
North Carolina
Nieboer E, Richardson DHS (1980) The replacement of the
nondescript term ‘heavy metals’ by a biologically and chemi-
cally significant classification of metal ions. Environ Pollut
1 : 3–26
Pickering WF (1986) Metal ion speciation – soils and sedi-
ments (a review). Ore Geol Rev 1 : 83–146
Pierson WR, Brachaczek WW (1983) Particulate matter asso-
ciated with vehicles on the road, II. Aerosol Sci Technol
2 : 1–40
Raine LM, Siegel BZ, McMurtry GM (1995) Mercury accu-
mulation in sediments of the Ala Wai Canal and in the soils
and stream sediments of the Central Honolulu Watershed.
Pac Sci 49 : 511–525
Roden MF, Frey FA, Clague DA (1984) Geochemistry of tho-
leiitic and alkalic lavas from the Koolau Range, Oahu, Hawaii:
implications for Hawaiian volcanism. Earth Planet Sci Lett
69 : 141–158
Salomons W, Förstner U (1984) Metals in the Hydrocycle.
Springer-Verlag, Berlin Heidelberg
Schmitt CJ, Brumbaugh WG (1990) National contaminant
biomonitoring program: concentrations of arsenic, cadmium,
copper, lead, mercury, selenium, and zinc in US freshwater
fish, 1976–1984. Archiv Environ Contam Toxicol 19 : 731–747
Schropp SJ, Lewis FG, Windom HL, Ryan JD, Calder FD,
Burney LC (1990) Interpretation of metal concentrations is
estuarine sediments of Florida using aluminum as a reference
element. Estuaries 13 : 227–235
Shelton LR, Capel PD (1994) Guidelines for collecting and
processing samples of stream bed sediment for analysis of
trace elements and organic contaminants for the National
Water-Quality Assessment Program. US Geol Surv Open-File
Rep 94–458, Sacramento, CA
Spencer KJ, De Carlo EH, McMurtry GM (1995) Isotopic
clues to sources of natural and anthropogenic lead in sedi-
ments and soils from O’ahu, Hawai’i. Pac Sci 49 : 492–510
State of Hawaii (1996). State of Hawaii data book 1996 – a
statistical abstract. Dept Business, Economic Development
and Tourism, State of Hawaii, USA
Sutherland RA (1999) Distribution of organic carbon in bed
sediments of Manoa Stream, Oahu, Hawaii. Earth Surf Proc
Landforms (in press)
Tack FMG, Verloo MG (1999) Single extractions versus se-
quential extraction for the estimation of heavy metal fractions
in reduced and oxidised dredged sediments. Chem Spec
Bioavail (in press)
Tack FMG, Vossius HAH, Verloo MG (1996) A comparison
between sediment metal fractions, obtained from sequential
extraction and estimated from single extractions. Int J Envi-
ron Anal Chem 63 : 61–66
Taylor SR (1964) Abundance of chemical elements in the con-
tinental crust: a new table. Geochim Cosmochim Acta
28 : 1273–1285
Trefry JH, Metz S, Trocine RP (1985) A decline in lead
transport by the Mississippi River. Science 230 : 439–441
Turekian KK, Wedepohl KH (1961) Distribution of the ele-
ments in some major units of the Earth’s crust. Geol Soc Am
Bull 72 : 175–192
Van Hassel JH, Ney JJ, Garling DL (1980) Heavy metals in a
stream ecosystem at sites near highways. Trans Am Fish Soc
109 : 636–643
White KD, Tittlebaum ME (1985) Metal distribution and
contamination in sediments. J Environ Eng 111 : 161–175
Wigington PJ, Randall CW, Grizzard TJ (1983) Accumula-
tion of selected trace metals in soils of urban runoff deten-
tion basins. Water Res Bull 19 : 709–718

Wilkinson JFG, Stolz AJ (1983) Low-pressure fractionation of
strongly undersaturated alkaline ultrabasic magma: the oliv-
ine-melilite-nephelinite at Moiliili, Oahu, Hawaii. Contrib Mi-
neral Petrol 83 : 363–374
Wilson SA, Briggs PH, Mee JS, Siems DF (1994) Determina-
tion of 32 major and trace elements in three NIST soil SRMs
using ICP-AES and WDXRF. Geostandards Newslett 18 : 85–89
Yeh C-Y, Chiou H-Y, Chen R-Y, Yeh K-H, Jeng W-L, Han B-
C (1996) Monitoring lead pollution near a storage battery re-
cycling plant in Taiwan, Republic of China. Archiv Environ
Contam Toxicol 30 : 227–234
Cases and solutions
Yim SK, Dugan GL (1975) Quality and quantity of nonpoint
pollution sources in rural surface water runoff on Oahu, Ha-
waii. Water Res Res Center, Tech Rep Nr 93, University of
Hawaii
Young LB, Dutton M, Pick FR (1992) Contrasting two meth-
ods for determining trace metal partitioning in oxidized lake
sediments. Biogeochemistry 17 : 205–219
Environmental Geology 39 (6) April 2000 7 Q Springer-Verlag 627