Full text of DOI:10.1002/etc.5620210104

Environmental Toxicology and Chemistry, Vol. 21, No. 1, pp. 24–30, 2002
Printed in the USA
0
730-7268/02 $9.00 ϩ .00
COMPARISON OF COPPER LEVELS IN RUNOFF FROM FRESH-MARKET VEGETABLE
PRODUCTION USING POLYETHYLENE MULCH OR A VEGETATIVE MULCH
P
AMELA J. RICE, LAURA L. MCCONNELL, LYNNE P. H EIGHTON, ALI M. SADEGHI, ALLAN R. ISENSEE,
J
OHN R. TEASDALE, AREF A. ABDUL-BAKI, JENNIFER A. HARMAN-FETCHO, and CATHLEEN J. HAPEMAN*
U.S. Department of Agriculture, Agricultural Research Service, Beltsville Agricultural Research Center, 10300 Baltimore Avenue, Beltsville,
Maryland 20705
(
Received 13 December 2000; Accepted 5 June 2001)
Abstract—Runoff from tomato (Lycopersicon esculentum Mill.) production with polyethylene mulch has been implicated in the
failure of commercial shellfish farms in the Mid-Atlantic Region of the United States. Copper, applied in the form of copper
hydroxide, is the most widely used fungicide–bactericide for control of tomato diseases and recently has been detected in the
Chesapeake Bay (USA) watershed. Elevated levels of copper have been shown to have adverse effects on shellfish, finfish, and
other aquatic organisms. This research evaluates the off-site movement of copper with the dissolved phase and the particulate phase
of runoff from controlled field plots containing tomato plants grown in either polyethylene mulch or a vegetative mulch, hairy
vetch (Vicia villosa Roth.). Overall, runoff collected from polyethylene mulch plots contained significantly (
of dissolved- and particulate-phase copper than runoff from hairy vetch mulch plots. However, the loss of copper associated with
the particulate phase was significantly greater ( 0.05) than that associated with the dissolved phase of runoff from both mulch
p Յ 0.05) greater loads
p
Յ
treatments, with the particulate phase accounting for more than 80% of the copper loads. The reported toxicity of copper to aquatic
organisms and the greater runoff volume, soil loss, and off-site loading of copper measured in runoff from the polyethylene mulch
suggests that this management practice is less sustainable and may have a more harmful impact on aquatic ecosystems.
Keywords—Runoff
Copper
Polyethylene mulch
Vegetative mulch
Soil erosion
INTRODUCTION
pv. tomato) and bacterial spot (Xanthomonas campestris pv.
vesicatoria). The estimated annual use of copper for fresh-
market tomato production throughout the United States is
Tomatoes (Lycopersicon esculentum Mill.) are one of the
most economically important vegetables grown in the United
States, with annual average yields (from 1989 to 1993) of
fresh-market and processing tomatoes valued at $1.1 billion
and $529 million, respectively [1]. Fresh-market vegetable
production requires multiple applications of agrochemicals and
generally utilizes polyethylene (plastic) mulch to control
weeds and preserve soil moisture. Use of this impervious
mulch on 50 to 75% of the surface area of a field produces
greater volumes of runoff with larger sediment loads than from
areas not covered with this mulch [2–4].
The transport of runoff with high sediment levels into ad-
jacent surface waters may have adverse effects on nontarget
organisms as a result of increased turbidity and degraded water
quality [5,6]. In addition, runoff from agriculture has also been
shown to contaminate surface waters with pesticides that also
have adverse affects on aquatic organisms [7–9]. Agrochem-
icals, including those used in vegetable production, have been
detected in the Chesapeake Bay (USA) watershed [10–12].
Fungicides, bactericides, and insecticides required to protect
vegetable crops are known to have adverse effects on shellfish,
finfish, and other aquatic organisms at environmentally rele-
vant levels [13–16].
7
54,000 pounds of active ingredient [1]. In Virginia and South
Carolina, USA, copper is applied to tomato plants 10 and 15
times during a growing season, resulting in the use of 28,000
to 107,000 pounds of active ingredient per year [1]. Copper
has been shown to have adverse effects on aquatic organisms,
including a reduction in the survival of macroinvertebrates [17]
and structural and functional effects on the nervous systems
of fish [18].
As a means of achieving long-term sustainable agriculture
and maintaining environmental quality, management practices
should minimize soil erosion and reduce off-site transport of
agrochemicals and soil. Depending on field slope, management
practices, weather conditions, and chemical properties of the
pesticide, as much as 5% of applied pesticides may be trans-
ported in runoff from agricultural fields [19]. Before imple-
menting a management practice, vegetable producers consider
production costs and gross returns to determine the economic
feasibility of the management practice. To assess the true cost
of vegetable production, one must also consider the hidden
costs associated with lost topsoil and reduced soil fertility, as
well as the environmental impact of soil and pesticide loads
to surrounding sensitive ecosystems [6].
Copper, applied in the form of copper hydroxide, is the
most widely used fungicide–bactericide for control of tomato
diseases [1]. Tomato production areas on the East Coast rely
on copper to control bacterial speck (Pseudomonas syringae
Scientists from the U.S. Department of Agriculture Belts-
ville Agricultural Research Center (Beltsville, MD) have eval-
uated the use of vegetative cover crops and plant residues as
an alternative to plastic mulch. Vegetative cover crops, in-
cluding hairy vetch (Vicia villosa Roth.), have been shown to
reduce soil erosion, increase soil organic matter, improve soil
tilth, increase water infiltration, act as a slow-release fertilizer,
and suppress weeds [20–23]. Tomato production with hairy
*
To whom correspondence may be addressed
hapemanc@ba.ars.usda.gov).
Mention of specific products or supplies is for identification and
does not imply endorsement by the U.S. Department of Agriculture
to the exclusion of other suitable products or suppliers.
(
2
4

Transport of copper in runoff from vegetable production
Environ. Toxicol. Chem. 21, 2002
25
vetch residue as a mulch has been shown to be economically
feasible and to produce greater yields than obtained with black
polyethylene mulch or bare soil without such mulch [24]. Stud-
ies have demonstrated that crop residues and vegetative filter
strips can reduce runoff, soil erosion, and the off-site transport
of agrochemicals [25–29]. A three-year field study comparing
runoff from tomatoes indicated seasonal losses of two to four
times more water and at least three times as much sediment
with polyethylene mulch compared to hairy vetch residues [4].
This manuscript describes a field plot experiment carried
out to measure the off-site movement of copper with the dis-
solved phase and particulate phase of runoff from tomato plots
containing polyethylene mulch or a vegetative residue, hairy
vetch, as a mulch. The increased runoff volume and soil loss
measured in runoff from the polyethylene mulch suggests that
this management practice is less sustainable [4]. Quantifying
off-site copper loads with runoff from plots containing poly-
ethylene or hairy vetch mulch will provide valuable infor-
mation concerning the environmental impact of each manage-
ment practice. Determining the distribution of copper in the
dissolved phase and particulate phase of runoff will also pro-
vide information that can be used to select management strat-
egies that will be more effective for controlling off-site trans-
port of this chemical of concern.
corn plots [30,31]. The following May, the hairy vetch was
flail-mowed leaving a finely chopped layer of mulch residue
over the soil surface. At this time, corn was planted in the
vetch–corn and bare soil–corn plots, beds were constructed in
polyethylene–tomato plots, and drip lines (for trickle irriga-
tion) were installed 8 to 10 cm from the plant row in poly-
ethylene–tomato plots before the installation of the polyeth-
ylene.
During the last week of May, Sunbeam tomato plant seed-
lings were transplanted with a no-tillage planter in the center
of each vetch–tomato or polyethylene–tomato bed and drip
lines were added to the surface of the mulch in the vetch–
tomato beds approximately 8 to 10 cm from the plants. Tomato
plants grown in both the polyethylene and hairy vetch mulch
were watered and fertilized through the trickle irrigation sys-
tem. Approximately equal quantities of water (1,167
Ϯ 73 L/
plot in 1998; 1,522 112 L/plot in 1999) were applied to
Ϯ
each plot, at a rate that did not produce surface runoff, between
one and three times each week depending on the soil moisture
and weather conditions. Urea fertilizer was dissolved in water
and applied to each of the eight tomato plots on Julian day
181, 211, 225, and 245 in 1998 and day 160, 180, 202, and
216 in 1999. Vetch mulch plots received one half of the urea
applied to the plastic mulch (26 kg/ha/application for vetch
and 53 kg/ha/application for polyethylene) because of the ni-
trogen-fixing capability of the hairy vetch and previously es-
timated nitrogen rates for hairy vetch and black polyethylene
systems [30,31].
MATERIALS AND METHODS
Site description and management practices
The study site, located at the Henry A. Wallace Beltsville
Agricultural Research Center, Beltsville, Maryland, was cho-
sen for its soil and slope characteristics, to facilitate the col-
lection of runoff and to simulate grading, which is often uti-
lized in tomato production to improve drainage and reduce
Pesticides
Kocide௡ 101 fungicide–bactericide (Griffin, Valdosta, GA,
USA) containing 77% copper hydroxide was applied at rec-
ommended rates (3.36 kg/ha) three times during the growing
season (Julian day 215, 219, and 232 in 1998; Julian day 225,
231, and 239 in 1999) for a total of 10 kg/ha/season. In ad-
disease levels (http://pestdata.ncsu.edu/cropprofiles). The
2
2
,500-m field, composed of Mattapex silt loam (fine-silty,
mixed, mesic Aquic Hapludult soil with 1.3–1.6% organic
carbon content), was divided into 16 plots. Each plot contained
four raised beds (15 cm high, 27 m long, 0.9 m wide, and 1.5
m center-to-center) prepared in a north–south direction (av-
dition, the organic pesticides Bravo
sciences, Mentor, OH, USA), containing 40.4% chlorothalonil
(tetrachloroisophthalonitrile); Lexone DF herbicide (Du Pont,
Wilmington, DE, USA), containing 75% metribuzin (4-amino-
6-(1,1-dimethylethyl)-3-(methylthio)-1,2,4-triazin-5(4H)-one);
௡ 720 fungicide (ISK Bio-
௡
erage field slope
ϭ 5.8% north to south, 2.6% east to west).
Earthen berms were constructed around each plot to prevent
water movement between plots and to guide runoff from the
three central rows within each four-bed plot to a fiberglass H-
flume. Runoff was measured and collected with an automated
flow meter and runoff sampler containing 24, 350-ml glass
bottles (ISCO model 6700, Lincoln, NE, USA).
Four treatments were assigned to four replicate plots by
means of a randomized complete block design. These treat-
ments included tomatoes grown in polyethylene mulch, to-
matoes grown in hairy vetch residue mulch, corn grown in
bare soil, and corn grown in hairy vetch residue mulch. Tomato
and corn plots were rotated annually (vetch–tomato with
vetch–corn and polyethylene–tomato with bare soil–corn) to
minimize pest pressure. Runoff water was collected from the
eight plots containing tomato plants during each of the three
growing seasons (1997–1999). Copper concentrations and
loads were analyzed only for the runoff samples collected
during the 1998 and 1999 seasons.
Asana
erate ((
methylethyl) benzeneacetate); and Thiodan
cide (FMC, Philadelphia, PA, USA), containing 50% endo-
sulfan (hexachlorohexahydromethano-2,4,3-benzodioxathie-
pin 3-oxide) were also applied at recommended rates and times
during the growing season. Runoff concentrations and loads
of the organic pesticides are discussed elsewhere [4].
௡ XL insecticide (Du Pont), containing 8.4% esfenval-
s
)-cyano(3-phenoxyphenol)methyl(
s
)-4-chloro-
␣-(1-
௡
50 WP insecti-
Precipitation events
Runoff events described in this study were the result of
natural precipitation with the exception of the events occurring
on Julian day 244 of 1998 and day 246 of 1999. On these
days, artificial rain events were created with an overhead sprin-
kler system. The sprinkler system used, described in greater
detail elsewhere [4], delivered relatively uniform rain to the
plots with an average intensity of 10.4 mm/h. The duration of
each artificial rain event was chosen to be sufficient enough
to produce runoff to fill 24 sample bottles. At least 12 runoff
events were monitored each year.
Plant cultivation
Runoff water sample collection and processing
In September, beds were constructed in vetch–tomato plots
(
n
ϭ
4) and hairy vetch seed, inoculated with the appropriate
Automated samplers were programed to collect runoff on
a flow-weighted basis with the date and time recorded as the
Rhizobium strain, was planted in the vetch–tomato and vetch–

2
6
Environ. Toxicol. Chem. 21, 2002
P.J. Rice et al.
sample was delivered to each 350-ml bottle. Each sampler
collected runoff at the beginning of the event and at programed
volumes throughout the event for a total of 24, 300-ml samples.
A detailed description of the automated samplers and sample
processing method is given elsewhere [4]. Briefly, after a rain–
runoff event, water samples were removed from the samplers,
placed on ice, and transported to the laboratory for immediate
processing. Individual and integrated water samples, 50 ml
from each 300-ml sample combined into a single integrated
sample from each plot for each event, were characterized in
terms of total suspended solids and total copper or dissolved-
phase and particulate-phase copper.
Unfiltered runoff—Total copper
Fig. 1. Precipitation and runoff volume from polyethylene and hairy
vetch mulch plots for the 1998 and 1999 runoff events after the ap-
plication of copper hydroxide (Julian day 215, 219, and 232 of 1998;
Julian day 225, 231, and 239 of 1999). The difference in runoff volume
Unfiltered runoff water was processed for total copper with
nitric acid (HNO ) and hydrochloric acid (HCl) digestion in
3
acid-washed glassware [32]. One hundred fifty milliliters of
each runoff sample was stored in a plastic container and pre-
served with 3.0 ml of 16 N HNO (TraceMetal, Fisher, Fair-
between polyethylene and hairy vetch mulch was significant (p ϭ
0.01) for each runoff event. Runoff events occurring on Julian day
244 of 1998 and day 246 of 1999 were the result of artificial rain
events created with an overhead sprinkler system.
3
lawn, NJ, USA). A 100.0-ml aliquot from each sample was
digested with 3 ml of 16 N HNO and this was reduced to a
3
5
-ml volume over low heat. An additional 5 ml of 16 N HNO3
of a diethylenetriamine pentaacetic acid extracting solution
was added and the samples were covered with a watch glass
and allowed to reflux for 2 h until the solution became clear
and the particles were bleached. Samples were then cooled
and the glassware was rinsed with 0.1 N HCl before a final
reduction to 5 ml over low heat. The final extract was filtered
(
0
0.005 M diethylenetriamine pentaacetic acid, 0.01 M CaCl2,
.1 M triethanolamine, adjusted to pH 7.30 with 1.0 N HCl)
was added to each flask and the flasks were shaken for 2 h.
Extracts were filtered by gravity through Whatman No. 42
filter paper and analyzed by atomic absorption spectrometry
for copper (Varian SpectrAA 300/400, wavelength: 324.8 nm,
flame: air acetylene). Untreated matrix samples (filters con-
taining untreated field soil) showed interfering absorbance,
which was subtracted from the sample concentrations. Air-
with Whatman No. 41 ashless filter paper (20- to 25-␮m pore
size, Clifton, NJ, USA) into an acid-washed 25-ml volumetric
flask to facilitate analysis by atomic absorption. Ten milliliters
of 3 N HCl was added to the flask along with the 0.1 N HCl
used to rinse the reflux glassware and filter paper. Additional
dried matrix samples were spiked with 20
␮g of copper and
0
.1 N HCl was added to bring the final extract volume to 25.0
ml. Analyses were carried out with a Varian SpectrAA 300/
00 atomic absorption spectrophotometer (wavelength: 324.8
extracted to determine the extraction efficiency. Recoveries of
copper from spiked samples averaged 105
imum detection limit for copper was 0.03
Ϯ
9.6%. The min-
4
␮
g/ml.
nm, flame: air acetylene, Walnut Creek, CA, USA). Deionized
water samples were processed with field samples as controls
and no contamination was observed. Collection efficiency of
the digestion process was assessed by including deionized wa-
Statistical analysis
Analysis of variance was used to determined the differences
in runoff volume, soil loss, and copper loading from plots
containing tomatoes grown in polyethylene and hairy vetch
ter samples spiked with copper (160
field runoff samples. Copper recoveries averaged 95.7
10). The minimum instrumental detection limit for copper
was 0.03 g/ml.
␮
g/L) with each batch of
Ϯ
3.8%
mulch [34]. Geometric means (Ϯ standard deviation) were
(n ϭ
calculated to provide a single value for the comparison of
overall trends between the two mulch systems [35]. Geometric
means were used rather than arithmetic means because one or
several runoff events, within a growing season, were one or
two order of magnitude greater than the other runoff events.
␮
Dissolved-phase copper
To determine dissolved-phase (filtrate
ϭ dissolved solids
ϩ
particles
tions, integrated runoff samples were filtered, immediately af-
ter field collection and integration, through a 1- m graded-
Ͻ 1 ␮m) and particulate-phase copper concentra-
RESULTS AND DISCUSSION
␮
Runoff volume
density glass microfiber filter (Whatman multigrade glass mi-
crofiber filter 150) to remove large particulate matter from the
runoff. Copper was extracted from the filtered water samples
by means of the HNO –HCl digestion process previously de-
scribed for the unfiltered runoff.
Significantly (p Յ 0.05) larger volumes of runoff were col-
lected from polyethylene mulch than from hairy vetch mulch
treatment plots (Fig. 1). The seasonal water losses for the 1998
3
2
and 1999 growing seasons were 55.4 and 146 L/m per growing
2
season for polyethylene and 13.7 and 75.7 L/m per growing
Particulate-phase copper
season for hairy vetch, respectively. The rate of rainfall infil-
tration compared to rainfall rate and quantity influences the
volume of surface runoff. In plots containing polyethylene
mulch, 50 and 75% of the soil surface is covered with the
impermeable polyethylene, which reduced the infiltration ca-
pacity of the plots and resulted in increased runoff volume
compared to the plots covered with hairy vetch residues.
McCall et al. [2] and Wan and El-Swaify [3] reported similar
Particles captured on the multigrade filter papers from in-
tegrated water samples were extracted with a method modified
from Lindsay and Norvell [33]. Filters were folded to protect
particle material, weighed, and divided into four sections. Each
filter section then was weighed and frozen at
extraction. One quarter of each filter was air-dried and placed
into an acid-washed 125-ml conical flask. Fifteen milliliters
Ϫ20ЊC until

Transport of copper in runoff from vegetable production
Environ. Toxicol. Chem. 21, 2002
27
Fig. 3. Total copper load in unfiltered runoff from polyethylene mulch
Fig. 2. Quantity of soil lost from polyethylene and hairy vetch mulch
plots for the 1998 and 1999 runoff events after the application of
copper hydroxide (Julian day 215, 219, and 232 of 1998; Julian day
and hairy vetch mulch plots after the application of copper hydroxide
2
(
active ingredient of copper, 0.169 g copper/m plot times three ap-
plications each growing season). Copper hydroxide was applied on
Julian day 215, 219, and 232 of 1998 and Julian day 225, 231, and
2
25, 231, and 239 of 1999). The difference in quantity of soil lost
between polyethylene and hairy vetch mulch was significant (
.01) for each runoff event with the exception of the event occurring
on Julian day 251 for 1998. Runoff events occurring on Julian day
44 of 1998 and day 246 of 1999 were the result of artificial rain
p
ϭ
2
39 of 1999. Numbers in parentheses represent days between copper
hydroxide application and runoff. Asterisks (**) represent a significant
difference ( 0.01) in the copper load between mulch treatments
0
p
ϭ
2
for the individual runoff event.
events created with an overhead sprinkler system.
Unfiltered runoff—Total copper
observations in which greater runoff volumes were associated
with the use of impermeable plastic mulch relative to bare soil.
Equal portions of individual runoff subsamples collected
from each plot were combined into eight samples, one inte-
grated sample per plot, for each runoff event and were analyzed
to compare copper concentrations and loads in runoff from the
two mulch treatments. Significantly greater concentrations of
copper were measured in unfiltered runoff from polyethylene
plots (0.47–5.84 mg/L) than hairy vetch plots (0.15–3.60 mg/
L) after the application of copper hydroxide (p Յ 0.05). Copper
loads ranged from 0.26 to 70.4 mg/m per runoff event for
polyethylene and 0.04 to 15.0 mg/m per runoff event for hairy
Soil erosion
2
Soil loss (g/m ) was calculated based on the total volume
of runoff water collected per plot per runoff event, the mass
of filterable suspended solids per volume of runoff, and the
size of each plot. Overall, runoff from polyethylene plots con-
tained more than four times the concentration (mg/L) of sus-
pended solids than runoff from hairy vetch plots (geometric
2
2
mean
ethylene plots and 692
When runoff volumes were considered, significantly (
Ϯ
standard deviation
ϭ
3,334
1.5 mg/L for hairy vetch plots).
0.01)
Ϯ
0.87 mg/L for poly-
Ϯ
vetch plots (Fig. 3). At the conclusion of the 1998 and 1999
field seasons, the measured total copper load from polyeth-
ylene plots was five times greater than from hairy vetch plots.
p
ϭ
greater loads of soil were lost with runoff from polyethylene
mulch plots (Fig. 2). The average soil loss for the 1998 and
Correlation analysis (r) revealed that runoff volume and quan-
1
999 growing seasons was six and four times greater from
tity of suspended particles in the runoff were more important
to the difference in copper loads than the concentration of
2
polyethylene than hairy vetch mulch (247 and 535 g/m per
season for polyethylene; 38.7 and 118 g/m per season for
hairy vetch).
2
copper in the unfiltered runoff sample (polyethylene:
for volume, 0.89 for particulate concentration, and
0.43 for copper concentration; hairy vetch: 0.90 for vol-
ume, 0.76 for particulate concentration, and 0.07 for
r ϭ 0.84
r
ϭ
r ϭ
The difference in soil loss between the two mulch systems
was the result of both the greater concentration of suspended
sediments in the runoff from the polyethylene mulch plots and
the increased runoff volume from the polyethylene mulch
plots. Reduced rainfall infiltration in the polyethylene mulch
plots produced increased runoff volumes with flow rates that
were 1.2 to 7.5 times greater than runoff from hairy vetch
plots. The increased volume and flow velocity of the runoff
from the polyethylene mulch plots was able to displace larger
quantities of soil, whereas the slower runoff velocity and
smaller runoff volume from plots containing hairy vetch res-
idues displaced less soil and allowed additional time for sus-
pended sediments to settle out of the runoff water. Similar to
our findings with the hairy vetch mulch, other researchers also
have reported reduced runoff volume, runoff velocity and soil
loss when plant residues were used as an organic mulch. Man-
nering and Meyer [25] and Foster and Meyer [26] reported
that crop residues dissipate the energy of rain drops and ef-
fectively reduce the velocity and amount of surface runoff.
Zuzel and Pikul [28] noted that straw mulch significantly re-
duced runoff and soil loss from agricultural plots.
r ϭ
r
ϭ
r ϭ
copper concentration). The copper load, in the unfiltered runoff
for events occurring between Julian day 228 and 250, repre-
sents 36.48 and 6.18% of the copper applied during the 1999
season for polyethylene and hairy vetch mulch plots, respec-
tively.
Individual runoff subsamples, from a single runoff event
on Julian day 244 in 1998, were analyzed and coupled with
flow data to estimate loads of copper with time (Fig. 4). The
difference in the loading profiles for the two treatments is
evident, with greater loads of copper associated with runoff
from the polyethylene mulch throughout the measured event.
The steep slope observed within the first minute of the hy-
drograph for runoff from the polyethylene mulch plots indi-
cates that copper is readily washed off the plastic mulch 12 d
after its application.
The automated runoff samplers were programed to collect
runoff subsamples with the first measured runoff, at each gal-
lon of runoff up to 5 gallons, and at additional programed
volumes throughout the event (polyethylene
ϭ every 10 gal-

2
8
Environ. Toxicol. Chem. 21, 2002
P.J. Rice et al.
Fig. 4. Hydrograph of total copper load in unfiltered runoff from
polyethylene mulch and hairy vetch mulch plots (Julian day 244 for
2
1
998, 12 d after copper hydroxide application, 0.169 g copper/m
plot). The average standard deviation for copper loads in (a) poly-
2
2
ethylene (15.00 ng/m ) and (b) hairy vetch (0.44 ng/m ) runoff are
shown on the left side of the graph.
lons from gallon 10 to 180; hairy vetch
ϭ every 5 gallons
from gallon 10 to 95). Because prediction of the duration and
intensity of rainfall is impossible, programing the automated
sampler to collect samples throughout the entire runoff event
is difficult when the number of bottles is limited. As a result,
runoff samples collected during each event represented a per-
centage of the total runoff measured during that event, which
will vary depending on the total runoff volume measured
Fig. 5. Dissolved-phase (a) and particulate-phase (b) loads of copper
in the 1999 runoff from polyethylene mulch and hairy vetch mulch
plots after the application of copper hydroxide (active ingredient of
copper, 0.169 g copper/m plot times three applications). Copper hy-
droxide was applied on Julian day 225, 231, and 239 of 1999. Numbers
in parentheses represent days between copper hydroxide application
(
polyethylene
deviation 57.52
metric mean standard deviation
ϭ
10.84–93.33%, geometric mean
4.29%; hairy vetch 11.11–100%, geo-
63.09 4.34%). The
Ϯ standard
2
ϭ
Ϯ
ϭ
Ϯ
ϭ
Ϯ
hydrograph in Figure 4 represents 58 and 75% of the total
runoff measured from polyethylene and hairy vetch plots for
the runoff event. According to this hydrograph, we underes-
timated the event’s total copper loss in runoff from polyeth-
ylene plots because the loading rate of copper for the poly-
ethylene plots was still declining at the end of the sample
period. Therefore, based on this argument, we can conclude
that the total copper load in runoff from polyethylene plots
may have been more than five times the load measured in
runoff from the hairy vetch plots.
and runoff. Asterisks (**) represent a significant difference (p ϭ 0.01)
in the copper load between mulch treatments for the individual runoff
event.
sured in the runoff for events occurring between Julian day
228 and 250, represent 1.39 and 0.39% of the copper applied
during the 1999 season for polyethylene and hairy vetch
mulch, respectively.
Particulate-phase copper
Dissolved-phase copper
Runoff collected from polyethylene mulch contained sig-
Greater loads of copper were measured in the operationally
defined dissolved-phase runoff from polyethylene plots (0.12–
nificantly (
p Յ 0.05) greater loads of particulate-phase copper
than runoff from hairy vetch mulch treatment plots (0.99–60.6
2
2
2
2
1
.75 mg/m ) than from hairy vetch plots (0.06–0.84 mg/m )
in all runoff events after the application of copper hydroxide
in 1999 ( 0.01; Fig. 5a). The geometric mean standard
deviation of individual dissolved-phase copper loads from
mg/m for polyethylene; 0.06–8.65 mg/m for hairy vetch; Fig.
5b). The geometric mean of individual particulate-phase cop-
p
ϭ
Ϯ
per loads was an order of magnitude greater for polyethylene
runoff than for hairy vetch runoff (10.66
2
Ϯ
0.002 mg/m for
2
2
polyethylene plots (0.542
times greater than the geometric mean for dissolved-phase
copper loads from the hairy vetch plots (0.192 0.001 mg/
Ϯ
0.001 mg/m ) was almost three
polyethylene; 1.053
pended particles from both mulch treatments contained com-
parable concentrations of copper. Correlation analysis ( ) of
the initial runoff event after the application of copper hy-
droxide revealed that particulate-phase loads of copper were
attributed more to the quantity of soil lost with runoff than
the concentration of copper associated with the particulates
Ϯ 0.002 mg/m for hairy vetch). Sus-
Ϯ
r
2
m ). Copper concentrations in filtered runoff from the two
mulch treatments were comparable in more than one half of
the evaluated runoff events, which suggests that the greater
load of copper from polyethylene mulch is primarily the result
of the increased runoff volume associated with this manage-
(polyethylene:
concentration; hairy vetch:
r
ϭ
0.88 for soil loss and
r
ϭ
0.00 for copper
0.19
ment practice. Correlation analysis (
r
) has shown that runoff
r
ϭ
0.75 for soil loss and r ϭ
volume contributed more to the increased copper load than the
concentration of copper measured in the filtered runoff water
for copper concentration). The copper load in the particulate-
phase of runoff for events occurring between Julian day 228
and 250 represents 34.55 and 3.76% of the copper applied
during the 1999 season for polyethylene and hairy vetch mulch
plots, respectively.
(
polyethylene:
concentration; hairy vetch:
for copper concentration). Dissolved-phase copper loads, mea-
r
ϭ
0.66 for volume and
r
ϭ
0.46 for copper
r
ϭ
0.90 for volume and
r ϭ
0.03

Transport of copper in runoff from vegetable production
Environ. Toxicol. Chem. 21, 2002
29
Phase distribution in relation to total copper load
ganisms. Analyses of results from the assays have shown that
the runoff and sediment samples collected from the polyeth-
ylene mulch plots were significantly more toxic to the test
organisms compared to the samples collected from the hairy
vetch mulch plots. Concentrations of the applied pesticides
were determined for both runoff water and sediment collected
from the field plots. In addition, laboratory bioassays with
individual pesticides generated median lethal concentrations
Copper loads associated with both the dissolved phase and
the particulate phase of runoff were compared to determine
which phase contributed more to the total copper load. Greater
(p Յ 0.05) loads of copper were measured in the particulate
phase than in the dissolved phase of runoff from both the
polyethylene and hairy vetch plots, with the particulate phase
accounting for 86.7
Ϯ 14.3% and 81.7 Ϯ 13.0% of the copper
(
LC50s) and median effective concentrations for each of the
load, respectively (Fig. 5a and b). The total copper load, the
sum of dissolved- and particulate-phase loads, ranged from
applied compounds. Copper concentrations detected in the run-
off were above the LC50 for E. affinis and T. pseudonana.
The hard clam M. mercenaria, used in bioassays to determine
the effect of runoff on bivalves, was the most sensitive or-
ganisms, with runoff impacting survival and D-hinge forma-
tion. The concentration of copper detected in both the runoff
water and sediment correlated well with toxic effects observed
with the clams, diatoms, copepods, and amphipods [16].
2
2
1
.45 to 62.3 mg/m for polyethylene and 0.18 to 9.24 mg/m
for hairy vetch. The geometric mean for the total copper load
in runoff from polyethylene plots was nine times greater than
2
runoff from hairy vetch plots (12.1
Ϯ
0.01 mg/m for poly-
2
ethylene; 1.28
loads calculated from the sum of dissolved-phase and partic-
ulate-phase loads (polyethylene 35.94% of applied copper;
hairy vetch 4.15% of applied copper) were comparable to
copper loads measured in the unfiltered runoff water (poly-
ethylene 36.48% of applied copper; hairy vetch 6.18%
of applied copper).
Ϯ 0.01 mg/m for hairy vetch). Total copper
ϭ
ϭ
CONCLUSIONS
The off-site transport of agrochemicals in runoff is a func-
tion of the volume of the carrier, water or suspended particles,
and the concentration of agrochemicals associated with the
carrier. Analyses of our results demonstrated that the most
important factor affecting the total load of copper in the man-
agement systems evaluated was the quantity of the carrier
rather than the concentration of copper associated with the
carrier. This was observed in both the dissolved phase and
particulate phase of runoff from the polyethylene and hairy
vetch mulch plots, with the greatest loss of copper associated
with the particulate phase for both mulch systems. Therefore,
management practices that can reduce runoff volume and, more
importantly, control soil erosion with runoff will substantially
reduce the off-site movement of copper.
The greater runoff volume, soil loss, and off-site loading
of copper measured in runoff from the polyethylene mulch
along with the toxicologic concerns involving adverse impact
of copper exposure to aquatic organism [16–18,40] suggests
that this management practice is less sustainable and may have
a more harmful impact on aquatic ecosystems. To reduce the
environmental impact of fresh-market tomato production, run-
off volumes and the quantity of soil leaving agricultural areas
must be reduced. Replacement of the impermeable polyeth-
ylene mulch with a more permeable vegetative mulch did in-
crease rainfall infiltration, which resulted in significantly re-
duced runoff volumes and soil erosion. This suggests that the
adoption of management practices that replace impermeable
plastic mulch with a mulch that allows greater rainfall infil-
tration or the implementation of additional management strat-
egies that will retard runoff and reduce soil erosion from plastic
mulch systems should reduce the environmental impact of
fresh-market vegetable production.
ϭ
ϭ
The physical and chemical properties of a pesticide and its
interaction with the mulch and soil apparently will influence
the availability of the pesticide to be transported with runoff.
Although copper hydroxide is applied to the tomato plants, a
percentage of this fungicide inevitably is either washed off the
foliage onto the mulch or is applied directly to the mulch
during foliar application. Based on the water solubility of cop-
per hydroxide (2.9 mg/L at pH 7 and 25ЊC) [36] and our
observations, copper seems to be readily washed off polyeth-
ylene mulch (Fig. 4) and transported in surface runoff long
after application (Ͼ30 d after application; Fig. 3). Copper
adsorbs to soil [37,38] and our field results have also shown
that more than 80% of the copper load is transported with
suspended particulates in runoff water, which suggests that
erosion control will significantly reduce copper loading. To
fully understand the availability and transport of copper with
runoff in these two mulch systems, additional studies of copper
hydroxide–vetch residue and copper hydroxide–polyethylene
sorption interactions are needed.
Toxicologic concerns
In 1996, polyethylene mulch was utilized on an estimated
1
1,724 acres of farmland in Virginia (USA), with 9,100 acres
located on the Eastern Shore [39]. The National Agricultural
Pesticide Impact Assessment Program reported that in 1993
copper was applied to 90% of the fresh-market tomatoes grown
in Virginia, which involved 10 applications throughout the
growing season and resulted in a seasonal application of 28,980
pounds of copper hydroxide [1]. The use of polyethylene
mulch along with the large application of copper hydroxide
increases the potential off-site transport of copper with runoff
and the exposure of sensitive aquatic ecosystems to a com-
pound that has been shown to have adverse effects on estuarine
organisms [16–18].
Acknowledgement—We would like to thank Bob Wevadu, Rachel
Herbert, Cumberland Dugan, and Alex Richeal for their assistance in
the processing of field samples and maintaining the tomato plots.
Hetzer et al. [16] collected runoff water and sediment sam-
ples from our polyethylene and hairy vetch mulch plots to
evaluate potential adverse effects on estuarine organisms, in-
cluding the copepod Eurytemora affinis, the hard clam Mer-
cenaria mercenaria, the diatom Thalassiosira pseudonana,
and the amphipod Leptochierus plumulosus. Toxicity assays
evaluated both lethal and sublethal effects of diluted runoff
and sediment samples on these representative estuarine or-
REFERENCES
1
2
. Davis RM, Hamilton G, Lanini WT, Screen TH, Osteen C. 1998.
The importance of pesticides and other pest management practices
in U.S. tomato production. Document I-CA-98. U.S. Department
of Agriculture National Agricultural Pesticide Impact Assessment
Program, Washington, DC.
. McCall EC Jr, Scott GI, Hurley JM. 1988. A comparison of ag-
ricultural non-point source runoff from black plastic and con-
ventional-till tomato test plots. Project Agreement 14-08-0001-

3
0
Environ. Toxicol. Chem. 21, 2002
P.J. Rice et al.
G1251. Technical Completion Report G1251-07. South Carolina
Water Resources Research Institute, Columbia, SC, USA.
. Wan Y, El-Swaify SA. 1999. Runoff and soil erosion as affected
by plastic mulch in a Hawaiian pineapple field. Soil Tillage Res
ing from agricultural fields: A review. J Environ Qual 7:459–
472.
20. Smith MS, Frye WW, Varco JJ. 1987. Legume winter cover crops.
3
4
Adv Soil Sci 7:95–139.
2
2
2
2
1. McVay KA, Radcliffe DE, Hargrove WL. 1989. Winter legume
effects on soil properties and nitrogen fertilizer requirements. Soil
Sci Soc Am J 53:1856–1862.
2. Ranells NN, Wagger MG. 1996. Nitrogen release from grass and
legume cover crop monocultures and bicultures. Agron J 88:777–
5
2:29–35.
. Rice PJ, et al. 2001. Runoff loss of pesticides and soil: A com-
parison between vegetative-mulch and plastic-mulch in vegetable
production systems. J Environ Qual 30:1808–1821.
. Holy M, Ondrackova J. 1980. Erosion and Environment. Per-
gamon, New York, NY, USA.
5
6
7
7
82.
3. Teasdale JR. 1996. Contribution of cover crops to weed man-
. Clark EH II. 1985. The off-site costs of soil erosion. J Soil Water
Conserv 40:19–22.
agement in sustainable agricultural systems. J Prod Agric 9:475–
4
79.
. Scott GI, et al. 1990. Agricultural insecticide runoff effects on
estuarine organisms: Correlating laboratory and field toxicity test-
ing with ecotoxicological biomonitoring. CR-813138-02-1. Final
Project Report. U.S. Environmental Protection Agency, Gulf
Breeze, FL.
. Chandler GT, Scott GI. 1991. Effects of sediment-bound endo-
sulfan on survival, reproduction and larval settlement of meioben-
thic polychaetes and copepods. Environ Toxicol Chem 10:375–
4. Kelly TC, Lu Y-C, Abdul-Baki AA, Teasdale JR. 1995. Econom-
ics of a hairy vetch mulch system for producing fresh-market
tomatoes in the Mid-Atlantic region. J Am Soc Hortic Sci 120:
8
54–860.
2
5. Mannering JV, Meyer LD. 1963. The effect of various rates of
8
9
surface mulch on infiltration and erosion. Soil Sci Soc Am Proc
2
7:84–86.
26. Foster GL, Meyer LD. 1972. Erosion mechanics of mulches.
ASAE Paper 72-754. American Society of Agricultural Engineers,
St. Joseph, MI, USA.
3
82.
. Clark JR, Lewis MA, Pait AS. 1993. Pesticide inputs and risks
in coastal wetlands. Environ Toxicol Chem 12:2225–2233.
27. Ghadiri H, Shea PJ, Wicks GA. 1984. Interception and retention
of atrazine by wheat (Triticum aestivum) stubble. Weed Sci 32:
24–27.
28. Zuzel JF, Pikul JL Jr. 1993. Effects of straw mulch on runoff and
erosion from small agricultural plots in northeastern Oregon. Soil
Sci 156:111–117.
1
0. Johnson WE, Kroll RB, Plimmer JR, Pait AS. 1994. The occur-
rence and distribution of pesticides in Chesapeake Bay. In Nelson
S, Elliott P, eds, Perspectives on Chesapeake Bay, 1994: Ad-
vances in Estuarine Sciences. CRC Publication 147. Chesapeake
Research Consortium, Edgewater, MD, USA, pp 105–145.
1. Harman-Fetcho JA, McConnell LL, Baker JE. 1999. Agricultural
pesticides in the Patuxent River, a tributary of the Chesapeake
Bay. J Environ Qual 28:928–938.
2. Lehotay SJ, Harman-Fetcho JA, McConnell LL. 1999. Agricul-
tural pesticide residues in oysters and water from two Chesapeake
Bay tributaries. Mar Pollut Bull 37:32–44.
3. Scott GI, Baughman DS, Trim AH, Dee JC. 1987. Lethal and
sublethal effects of insecticides commonly found in nonpoint
source agricultural runoff to estuarine fish and shellfish. In Vern-
berg WB, Calabrese A, Thurberg FP, Vernberg FJ, eds, Pollution
Physiology of Estuarine Organisms. University of South Carolina
Press, Columbia, SC, USA, pp 251–273.
4. Baughman DS, Moore DW, Scott GI. 1989. A comparison and
evaluation of field and laboratory toxicity tests with fenvalerate
on an estuarine crustacean. Environ Toxicol Chem 8:417–429.
5. Pait AS, De Souza AE, Farrow DRG. 1992. Agricultural pesticide
use in coastal areas: A national summary. National Oceanographic
and Atmospheric Administration, U.S. Department of Commerce,
Rockville, MD.
2
3
3
3
9. Webster EP, Shaw DR. 1996. Impact of vegetative filter strips on
1
1
1
herbicide loss in runoff from soybean (Glycine max). Weed Sci
4
4:662–671.
0. Abdul-Baki AA, Teasdale JR, Korcak R, Chitwood DJ, Huettel
RN. 1996. Fresh-market tomato production in a low-input alter-
native system using cover-crop mulch. Hortic Sci 31:65–69.
1. Abdul-Baki AA, Teasdale JR, Korcak RF. 1997. Nitrogen re-
quirements of fresh-market tomatoes on hairy vetch and black
polyethylene mulch. Hortic Sci 32:217–221.
2. American Public Health Association, American Water Works As-
sociation, Water Pollution Control Federation. 1989. Standard
Methods for the Examination of Water and Wastewater, 17th ed.
American Public Health Association, Washington, DC.
3. Lindsay WL, Norvell WA. 1978. Development of a DTPA test
for zinc, iron, manganese, and copper. Soil Sci Soc Am J 42:421–
428.
1
1
3
34. Steel RGD, Torrie JH. 1980. Principles and Procedures of Sta-
tistics: A Biometrical Approach. McGraw-Hill, New York, NY,
USA.
35. Campbell GS. 1985. Soil Physics with Basic: Transport Models
for Soil–Plant Systems. Elsevier, New York, NY, USA.
1
6. Hetzer PR, Brown SS, Rice PJ, Baker JE, Harmon-Fetcho JA.
3
6. Tomlin C. 1994. The Pesticide Manual: Incorporating the Ag-
rochemicals Handbook. Crop Protection, Farnham, UK.
2
000. Comparative effects of cultivation practices (vegetative vs.
polyethylene mulch) on pesticide-related ambient toxicity in es-
tuarine habitats. Abstracts, 220th American Chemical Society Na-
tional Meeting, Division of Agrochemicals of the American
Chemical Society, Washington, DC, August 20–24, Issue 59, Ab-
stract 141.
3
7. Basta NT, Tabatabai MA. 1992. Effect of cropping systems on
adsorption of metals by soils III. Competitive adsorption. Soil Sci
1
53:331–337.
3
3
8. Aringhieri R, Carrai P, Petruzzelli G. 1985. Kinetics of copper
and cadmium adsorption by an Italian soil. Soil Sci 139:197–204.
9. Interagency Plasticulture Task Force. 1998. The study of plasti-
culture and related water quality management issues. House Doc-
ument 44. Department of Agriculture and Consumer Services,
Richmond, VA, USA.
1
1
1
7. Clements WH, Cherry DS, Cairns J Jr. 1990. Macroinvertebrate
community responses to copper in laboratory and field experi-
mental streams. Arch Environ Contam Toxicol 19:361–365.
8. Baatrup E. 1991. Structural and functional effects of heavy metals
on the nervous system, including sense organs, of fish. Comp
Biochem Physiol C 100:253–257.
40. Hall LW Jr, Scott MC, Killen WD. 1998. Ecological risk assess-
ment of copper and cadmium in surface waters of Chesapeake
Bay watershed. Environ Toxicol Chem 17:1172–1189.
9. Wauchope RD. 1978. The pesticide content of surface water drain-