Water Research 168 (2020) 115139

Contents lists available at ScienceDirect

Water Research
journal homepage: www.elsevier.com/locate/watres

Glyphosate and aminomethylphosphonic acid (AMPA) are commonly

found in urban streams and wetlands of Melbourne, Australia
Elena Okada a, b, *, 1, Mayumi Allinson c, María P. Barral b, Bradley Clarke d,
Graeme Allinson a
a
Centre for Environmental Sustainability and Remediation, School of Science, RMIT University, GPO Box 2476, Melbourne, Victoria, 3001, Australia
b
n Experimental Agropecuaria Balcarce, Route 226 Km 73.5, 7620, Balcarce, Argentina
INTA Centro Regional Buenos Aires Sur. Estacio
c
Department of Chemical Engineering, The University of Melbourne, Victoria, 3010, Australia
d
School of Chemistry, The University of Melbourne, Victoria, 3010, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Glyphosate is the most widely-used pesticide for weed control in agriculture and in urban and residential
Received 25 June 2019 areas. This is the first study to quantify glyphosate and AMPA levels in surface water in Australia from
Received in revised form different land uses. Glyphosate and AMPA was measured in surface water from 10 rural streams, 30
13 September 2019
urban stormwater wetlands and 9 urban streams located in and around the city of Melbourne, Australia
Accepted 27 September 2019
on five occasions between October 2017 and February 2018. Glyphosate and AMPA were present in most
Available online 30 September 2019
of the urban surface water samples. The frequency of detection of glyphosate was 77% in wetlands and
79% in urban streams, whereas it was only detected in 4% of the rural streams. Similarly, AMPA detection
Keywords:
Glyphosate
was 91% in wetlands and 97% in urban streams, whereas it was only present in 6% of the rural stream
Wetlands samples. In both urban streams and wetlands, the highest average glyphosate concentrations occurred in
Stormwater November (1.8 ± 2.2 mg L
1). Overall, wetlands and streams associated with urban land use are vulnerable
Streams to glyphosate contamination. These results highlight the importance of screening for contaminants in
Urban land use urban stormwater to identify the source of pollutants that may end up in aquatic ecosystems and the
risks therefrom.
© 2019 Elsevier Ltd. All rights reserved.

1. Introduction on pavements, sidewalks, garage driveways, residential lawns, and

The study of glyphosate (N-(phosphonomethyl) glycine) occur-
rence in the environment is of major relevance due to its wide-
spread use and its potential negative impacts on ecosystems.
Glyphosate is the most commonly used herbicide in agricultural
areas throughout the world for the removal of unwanted weeds
(Benbrook, 2016). It is also applied to eliminate competing weeds in
the growing stage of genetically modified glyphosate-resistant
crops, such as soybean, cotton, and maize. Furthermore, glypho-
sate is widely used on non-agricultural land to control noxious
weeds, invasive woody species and, in urban areas for weed control
* Corresponding author. Centre for Environmental Sustainability and Remedia-
tion, School of Science, RMIT University, GPO Box 2476, Melbourne, Victoria, 3001,
Australia.
E-mail address: okada.elena@inta.gob.ar (E. Okada).
1
n Experimental
Current address: INTA Centro Regional Buenos Aires Sur. Estacio
Agropecuaria Balcarce. Route 226 Km 73.5, 7620 Balcarce, Argentina.
golf courses (Skark et al., 2004). It is estimated that by the year 2014
more than 800 million kg of glyphosate had been used globally, of
which 90% was for agricultural use and 10% for non-agricultural
purposes (Benbrook, 2016) (Fig. 1), making glyphosate the most
used pesticide worldwide. In this context, glyphosate is the most
commonly used herbicide in Australia and in the State of Victoria,
where there are more than 500 glyphosate-based products regis-
tered for use (APVMA, 2018). However, there are no available data
on the amounts of glyphosate-based herbicides that are sold or
applied in the region.
Glyphosate is a systemic, post-emergence, broad-spectrum,
foliar herbicide. It is targeted to the plant leaves but the foliar spray
may reach the soil beneath by endo-drift processes (Okada et al.,
2018). Glyphosate is found in its anionic form (pKa ¼ 2, 2.6, 5.8
and 10.8, Sprankle et al., 1975) within the pH range of most soils
(pH z 4e8). The negatively charged glyphosate molecule sorbs
strongly to the soil matrix because it has a high affinity to cations
exchanged from the clay fraction and to Fe and Al oxides (Morillo

https://doi.org/10.1016/j.watres.2019.115139
0043-1354/© 2019 Elsevier Ltd. All rights reserved.
2 E. Okada et al. / Water Research 168 (2020) 115139
Fig. 1. Global glyphosate usage. Adapted from Benbrook (2016).
et al., 2000; Borggaard and Gimsing, 2008). The strong sorption to
the soil matrix reduces glyphosate bioavailability and increases its
persistence in soil (Okada et al., 2017).
Glyphosate is degraded by the microbiota into amino-
methylphosphonic acid (AMPA) or sarcosine (Borggaard and
Gimsing, 2008). AMPA is the most commonly studied metabolite
because it is highly persistent in the environment (Simonsen et al.,
2008), whereas sarcosine, is easily degradable and does not accu-
mulate in soil. The detection of glyphosate and AMPA in surface
water is usually concomitant (Battaglin et al., 2014; Van Stempvoort
et al. 2016). However, AMPA is also a degradation product of
household and industrial phosphonate detergents and therefore it
can also be found in wastewater treatment plant outlets (Grandcoin
et al., 2017).
Glyphosate and AMPA share similar physico-chemical charac-
teristics; they are polar compounds, highly water soluble, and non-
volatile. Both molecules are zwitterions and are found negatively
charged at pH > 6. Glyphosate and AMPA may be transported off-
site by exo-drift processes, such as spray drift during application,
and via particle-bound residues that are wind eroded or by surface
runoff (Yang et al., 2015; Bento et al., 2017). As a consequence, these
compounds may end-up far from zones of application in aquatic
ecosystems and can be found in the water column (Battaglin et al.,
2014; Pe
rez et al., 2017a; Okada et al., 2018), streambed sediments
(Okada et al., 2018; Tush et al., 2018), and aquatic plants (Pe
rez
et al., 2017b).
Melbourne is the largest city in the state of Victoria, Australia.
The metropolitan area of this sprawling city is home to 5 million
people, mostly housed in low-density suburbs that stretch over
40 km west to east, and between around 25 to 45 km north to
south, as the city extends along the northern and eastern shorelines
of Port Philip Bay. Several studies have investigated the presence of
pesticides in streams and rivers of the Yarra River catchment, an
important source of drinking water for Melbourne, with simazine
the most commonly observed agrochemical (e.g. Scha €fer et al.,
2011; Wightwick et al., 2012; Allinson et al., 2014). Screening has
also been undertaken for herbicides in constructed stormwater
wetlands in some of Melbourne’s urban areas (e.g. see Allinson
et al., 2015; Allinson et al., 2017). In the latter study, Allinson
et al. (2017), also found simazine was the most frequently detec-
ted herbicide (100% detection), followed by 2,4-diphenoxy acetic
acid (2,4-D; 93%), and 2-methyl-4-chlorophenoxy acetic acid
(MCPA; 83%). Rippy et al. (2017) also found that the most detected
pesticides in Australian stormwater were herbicides (i.e. diuron,
MCPA, 2,4-D, simazine, and triclopyr). However, none of the above-
mentioned studies included the analytical screening of glyphosate
or its metabolite, AMPA. Therefore, in this work, we present for the
first time, data on non-target environmental concentrations and
seasonal variations of glyphosate and AMPA in aquatic environ-
ments from Australia.
The objectives of the study were to: 1) compare glyphosate and
AMPA occurrence in waterbodies subject to different land uses, 2)
study the temporal variation of glyphosate and AMPA levels in
constructed stormwater wetlands and streams within urban and
rural land uses.
2. Materials and methods
2.1. Study area and sample collection
This study was conducted in the urban area of greater Mel-
bourne, and the rural area of the Upper Yarra River valley to Mel-
bourne’s northeast (Fig. 2A). Five major catchments that discharge
into Port Phillip Bay were sampled: Moorabool River, Werribee
River, Maribyrnong River, Yarra River, and Bunyip River Catchments
(Fig. 2B, Table A1). Surface water sampling sites were grouped ac-
cording to source (wetland or stream) and main type of land use
(rural or urban) (Fig. 2C). The spatial information to characterize
the sample sites (limits of the catchments, streams and land use
data) were obtained from Victoria State Government web site (State
Government of Victoria, 2019) and the riparian vegetation width
was estimated using Google Earth images. Water samples were
taken during dry periods (daily rainfall <10 mm 7 days prior to time
of sampling), except for the samples taken in December, which
were taken after a heavy rainfall (>60 mm, see Appendix A).
Although stream flow was not measured at each site, samples were
taken at base flow during October, November, January and
February. During the December sampling time, most of the streams
had bank over flow or evidence that the bank had overflowed as a
consequence of the heavy rainstorm.
2.1.1. Rural streams
“Rural streams” correspond to ten sites from several streams
from the Upper and Middle Yarra Catchment (Fig. 2C, Appendix B).
Land use in this watershed varies from protected forests and rural
areas in the Upper and Middle Yarra valley, to peri-urban devel-
opment (Fig. 2C). The waterways located in the Upper Yarra sys-
tem provide around 70% of Melbourne’s drinking water
(Melbourne Water, 2017a). The Middle Yarra Valley has many
small-scale rural activities, although it is mainly used for horti-
culture (i.e. grapevine, fruit, and vegetable production) and
grasslands for beef and dairy cattle. All the sites sampled for the
rural streams were located within agricultural land, except for site
S14, located in a protected forest area at the Maroondah Reservoir
Park (see Appendix B, Fig. B1). This site was used as a reference
site, since we expected to find no glyphosate or AMPA residues at
this site.
2.1.2. Urban streams
Nine sites located within urban and industrial land uses were
sampled as “urban streams” (Fig. 2C, Appendix B). These sites were
in the Werribee, Maribyrnong, Lower Yarra, and Bunyip Catch-
ments, which flow through the city’s sub-urban and urban areas
into Port Phillip Bay. The vast majority of Melbourne’s suburban
households are connected to a reticulated sewerage system, with
stormwater directed into a separated system that directs the water
 E. Okada et al. / Water Research 168 (2020) 115139
Fig. 2. A) Area of study in the State of Victoria, Australia, B) major catchments and streams, C) sampling sites and land use description.
into local streams and rivers. Consequently, surface water in this
area receives large amounts of urban runeoff, which reduces the
water quality and changes the water flow rates. In general, the
metropolitan area of the catchment has poor quality vegetation due
to urbanization and industrial impact. In residential areas, the
waterway channels are significantly altered by urbanization (i.e.
straightening, channeling, and concrete-lining) which reduces the
amount of riparian vegetation. The sampling sites of the urban
streams were located within sub-catchments that receive mostly
runoff water from urban areas (see Appendix B; Fig. B1), although
influence from upstream cannot be disregarded, since the water
samples were taken from the stream and not exclusively from run-
off water.
2.1.3. Stormwater wetlands
Thirty stormwater treatment wetlands in urban and peri-urban
developments located in the Moorabol, Werribee, Maribyrnong,
3
Lower Yarra, and Bunyip Catchments were sampled (Fig. 2C,
Appendix B). New residential developments incorporate storm-
water wetland systems built to deliver effective stormwater flow
management, and pollutant reduction, as well as a secondary
community amenity and habitat outcomes (Melbourne Water,
2017b); some are also built to provide a source for irrigation wa-
ter for nearby parks and sports fields. These constructed wetlands
help retard the water flow of around 540 GL/year of stormwater
that enters natural drainage systems (Allinson et al., 2017).
Stormwater wetlands are diverse in size and design. They are built
site-specifically to capture and treat the run-off generated by
impervious or hard surfaces, and to reduce nutrient, sediment
loading of natural waterways. In this sense, wetlands are buffer
zones that can intercept and partly retain run-off related pesticides
(Imfeld et al., 2013). The selected sampled wetlands were located
within sub-catchments that receive mostly runoff water from ur-
ban areas (see Appendix B, Fig. B1).
4 E. Okada et al. / Water Research 168 (2020) 115139
2.2. Sampling
Samples were collected monthly from October 2017 to February
2018 using 500 mL polypropylene bottles, rinsed three times with
the sampled water. The temperature, pH, electrical conductivity
(EC) and dissolved oxygen (DO) of each sample was measured in
situ with a field meter (556 MPS Multi Probe System, YSI Inc., Yel-
low Springs, OH, USA) (Appendix C, Fig. C1). Samples were refrig-
erated immediately after sampling and kept in a cooler for a
maximum of 6 h until arrival to the laboratory. Then, samples were
stored at
18  C until analysis.
2.3. Glyphosate and AMPA analysis
2.3.1. Chemicals
Glyphosate (>98%), AMPA (99%), isotope-labelled glyphosate
(1,2e13C,15N) (>99%), ammonium acetate (NH4Ac; 99.99%), and
ammonium hydroxide (NH4OH; >99%) were purchased from
Sigma-Aldrich. Ultrapure deionized water (resistivity
16.7 mU cm
1) was obtained from a Milli-Q® water purification
system (Millipore Ltd., Bedford, MA, USA). A standard stock solution
of 1000 mg L
1 was prepared for each analyte. Standard working
solutions for LC-MS/MS analysis and for samples fortification were
prepared by dilution with ultrapure water.
2.3.2. LC-MS/MS analysis
Glyphosate and AMPA analyses were carried out using the direct
injection method of Okada et al. (2019). The chromatographic
analysis was performed on an Agilent Technologies 1290 Infinity II
liquid chromatograph (LC) using a Metrosep A Supp 4® 9 mm,
250  4.0 mm ion exchange column (Metrohm, Gladesville, NSW,
Australia). The target molecules were detected with an Agilent®
6495 B Triple Quadrupole mass spectrometer (MS/MS). Target
compounds were monitored according to retention time and two
ion transitions and quantified using Agilent® MassHunter®
Quantitative Analysis software (v. B.08.00).
Procedural blanks that consisted of ultrapure water spiked with
internal standard were used to monitor any background concen-
tration. In addition, blank injections of ultrapure water were run
after the highest level of the calibration curve and every six samples
to ensure the lack of carryover. For quality assurance, spiked matrix
samples were checked for recovery at the end of every run. The
limit of detection (LOD) was set as the lowest concentration that
could be reliably differentiated from the background levels with a
signal-to-noise ratio  3, was 0.25 mg L
1. The method limit of
quantification (LOQ) was 0.5 mg L
1. Samples with concentra-
tions < LOD were considered non-detectable and were set to zero.
Samples that had concentrations < LOQ were set to the LOD value in
figures and statistical analyses (Struger et al., 2015). Selected
samples were also analysed by the enzyme linked immunosorbent
assay (ELISA) using a commercial ELISA kit for glyphosate to
compare the results obtained by the LC-MS/MS method (Appendix
D, Fig. D1).
2.4. Data analysis
Glyphosate and AMPA results from the LC-MSMS analysis were
compared among source (rural stream, urban stream or wetland)
and month of sampling (November, December, January, February)
using a linear mixed model (PROC MIXED). Data from the October
samples was not used in the statistical analysis because no urban
streams were sampled in that month. Mean comparisons were
evaluated with a significance level of 0.0001 using LSMEANS.
3. Results and discussion
3.1. Glyphosate and AMPA occurrence
Glyphosate and AMPA were present in the majority of the urban
samples from wetlands and streams, but rarely in stream samples
taken from rural areas (Fig. 3). In samples from rural streams,
glyphosate and AMPA frequency of detection (FOD) was 4 and
10.2%, respectively (Fig. 3). Glyphosate concentrations were
0.3 mg L
1 and AMPA levels were 0.8 mg L
1. On the other hand,
streams and wetlands situated within urban land use had a much
higher FOD of both glyphosate and AMPA compared to that of the
rural landuse water samples. Glyphosate was present in 79% of the
urban stream samples (mean: 1.1 mg L
1 range: 0.25e4.8 mg L
1)
and in 77% of the stormwater wetland samples (mean: 1.1 mg L
1
range: 0.25e14.2 mg L
1). AMPA was detected in 91% of urban
streams (mean: 1.6 mg L
1 range: 0.25e4.3 mg L
1) and 97% of the
sampled wetlands (mean: 1.3 mg L
1 range: 0.25e10 mg L
1). For
more details on the spatial and temporal variations of glyphosate
and AMPA levels see Fig. E1 (Appendix E).
The streams in the upper, forested, mountainous area of the
Yarra River catchment feed water storages that are used for
drinking water. The sample taken at site S14 is located upstream of
the agricultural area in the Maroondah Reservoir Park, a protected
water supply catchment. This sample had no detectable levels of
glyphosate or AMPA throughout the studied period. The Yarra River
then flows through the Yarra Valley, characterized by horticultural
and viticulture production and livestock. Previous studies have
shown that the application of glyphosate in agricultural soils can
cause the non-point source pollution of groundwater, surface wa-
ter, and sediments (Okada et al., 2018). However, in this study the
detection of glyphosate and AMPA in the rural streams was low,
despite glyphosate being widely used throughout the growing
season of intensive cultures such as vine and fruit orchards
(Preston, 2018). In this case, glyphosate dispersion into aquatic
environments may have been attenuated by the soil capacity to
strongly sorb both glyphosate and AMPA (Okada et al., 2016; Sidoli
et al., 2016). Also, the presence of riparian vegetation (Syversen and
Bechmann, 2004) reduces the probability of glyphosate reaching
surface waters. For instance, several studies have found that
Fig. 3. Glyphosate and AMPA levels and frequency of detection according to source.
 E. Okada et al. / Water Research 168 (2020) 115139 5

riparian vegetation of 5 m or more have a pesticide retention effi-

ciency higher than 40% (Reichenberger et al., 2007). In our study,
the average riparian vegetation width at the rural stream sites was
15 m (except for two sampling points where the width was
>400 m), so pesticide surface runoff may have been strongly
attenuated. Other factors that may have reduced the occurrence of
glyphosate and AMPA in rural streams are i) correct glyphosate
management practices, ii) organic farming, iii) no applications near
time of sampling since the sampling scheme did not cover the
paddock preparation period prior to sowing (April to early June).
The highest glyphosate levels were detected in wetland sam-
ples, with concentrations as high as 11.5 and 14.2 mg L
1. The
highest AMPA concentrations were also detected in wetland sam-
ples with values of 6.5 and 10 mg L
1. The average concentrations in
urban streams and wetlands are consistent with previous reports
on glyphosate and AMPA levels in urban catchments from other
countries (Table 1). Although non-agricultural use represents less
than 10% of the total glyphosate usage, urban contributions of
glyphosate into surface waters should not be underrated. The
application of glyphosate for weed control in residential/industrial
areas can be a source of contamination of surface waters (Skark
et al., 2004). In this sense, a significant amount of glyphosate and
AMPA is washed away in urban stormwater (Botta et al., 2009; Tang
et al., 2015; Mauffrey et al., 2017), and then accumulated in the
wetlands built to retain excess runoff. . In addition, sub-urban en-
vironments increase the area of impermeable surfaces, such as Fig. 4. Bars represent monthly average glyphosate and AMPA concentrations and tri-
asphalt and concrete, compared to pre-development conditions, angles correspond to the frequency of detection (FOD) according to source. Error bars
while at the same time reducing the extent of vegetation. Hence, represent standard error. No samples from urban streams were taken in October.
evapotranspiration and infiltration are reduced while stormwater
run-off increases (Allinson et al., 2017). Also, glyphosate loss from
hard surfaces is higher than from agricultural fields because and wetlands, the highest glyphosate concentrations were detected
glyphosate sorption to soil is stronger than to concrete or asphalt in early spring (November) and then concentrations decreased in
(Ramwell et al., 2014). However, more than 75% of glyphosate can mid-summer (February) (Fig. 4). Average glyphosate concentra-
be lost in run-off from asphalt (Spanoghe et al., 2005), because tions in urban wetlands and urban streams were broadly corre-
asphalt paving material is composed of a mixture of asphalt lated, except in January where glyphosate concentrations in
cement, sand, and crushed rock, with limited clay or mineral con- wetlands were observed at similar values as in November, whereas
tent, thus with fewer sites for glyphosate sorption. concentrations in urban streams remained almost constant. Over-
all, the FOD of glyphosate was similar in urban streams and wet-
lands, being higher than 80% in November, December, and January,
3.2. Seasonal variations and decreasing to less than 60% in February (Fig. 4). In a similar
manner to glyphosate, AMPA was detected at very low concentra-
Rural streams only had detectable levels of glyphosate in early tions in rural streams throughout the sampled months. On the
summer (December), although trace levels of AMPA were detected other hand, AMPA FOD in urban wetlands and urban streams from
in November, December, and February (Fig. 4). In urban streams October to February was between 80 and 100% (Fig. 4). In the case

Table 1
Glyphosate and AMPA occurrence in urban surface waters around the world and in this study.

Reference Country Source n Compound FOD (%) Average (mg L
1) Max. (mg L
1)

Hanke et al. (2010) Switzerland River 23 Gly 86.9 e 3.3

8 AMPA 100 e 0.5
Glozier et al. (2012) Canada River 150 Gly 50 0.1 11.8
AMPA 55 0.1 2.8
Kolpin et al. (2006) US Stream and wastewater treatment plants effluent 40 Gly 67.5 e 2.2
AMPA e 3.9
Ramwell et al. (2014) UK Stormwater outlet e Gly e 2.0 9.0
AMPA 0.5 1.2
Tang et al. (2015) Belgium Stormwater outlet 23 Gly 100 2.3 6.1
AMPA 100 1.9 5.8
Mauffrey et al. (2017) France Stormwater wetland inlet 10 (Spring) Gly e 4.1 11.0
18 (summer) 5.9 15.0
10 (Spring) AMPA 1.4 2.5 0.2 21.0
18 (summer) 2.5 21.0
This study Australia Urban streams 38 Gly 79 1.1 4.8
AMPA 97 1.6 4.3
Stormwater wetlands 130 Gly 77 1.1 14.2
AMPA 91 1.3 10.0

Gly: glyphosate, AMPA: aminomethylphosphonic acid, n: total number of samples; FOD: frequency of detection.
6 E. Okada et al. / Water Research 168 (2020) 115139
of wetlands, the highest concentrations were detected in
November, whereas in urban streams AMPA levels tended to be
higher in January (Fig. 4).
No significant differences were found in glyphosate concentra-
tions between urban streams and wetlands using the mixed linear
model and there was no interaction between source of sample and
month (p > 0.05, data not shown). When glyphosate data is ana-
lysed altogether there is a significant seasonal variation in which
glyphosate has its highest value at the beginning of spring
(November) (p < 0.05; Fig. 5). In the case of AMPA, there were no
significant differences between urban streams and wetlands (data
not shown) or between time of sampling (p > 0.05; Fig. 5). -It is
important to note that samples were taken during base flow, except
for the samples taken in December, which were taken after a heavy
rainfall (see Appendix A; Fig. A1). In December, most of the streams
had bankfull flow, bank over flow, or evidence that the bank had
overflowed as a consequence of the heavy rainstorm. However, we
did not find a clear association between levels of glyphosate and
AMPA and precipitation events.
In average, the highest glyphosate concentrations were detected
in November (springtime), which may result from recent herbicide
applications due to spring weed control in amenity areas, pave-
ments and gardens. We observed that the urban councils per-
formed herbicide sprayings during spring and summer in parks and
near creeks and riverbanks during our sampling trips. This type of
weed management is relevant in the studied area since there are 87
species of declared invasive plants in the State of Victoria, which are
considered noxious weeds (CaLP Act, 1994), of which each Council
has the responsibility to control and manage. From December to
February, glyphosate concentration decreased, reaching its lowest
Fig. 5. Average glyphosate and AMPA variations in urban streams and wetlands, only
taking into account data from November to February. Error bars represent standard
error. Different letters indicate significant differences (P < 0.05).
values towards the end of the study (Fig. 5). AMPA levels tended to
be higher than glyphosate concentrations, and remained steady
throughout the sampled period (Fig. 5). This can be a result of
several factors. First, there are fewer weed controls at the end of the
dry Victorian summer. Secondly, during the warmer months,
glyphosate concentration decreases because biological degradation
is favoured at higher environmental temperatures (Bento et al.,
2016). At the same time, the AMPA that is produced as a degrada-
tion product is more persistent than glyphosate (Wang et al., 2016;
la Cecilia et al., 2018; Sun et al., 2019). Third, glyphosate may also be
removed from the water column because it is strongly sorbed to
sediments (Coupe et al., 2012) and can also undergo fast chemical
degradation (Paudel et al., 2015; la Cecilia and Maggi, 2018). Lastly,
household and industrial detergents, such as diamine tetra-
methylene phosphonate (EDTMP) or diethylenetriamine-penta-
methylene phosphonic acid (DTPMP), also degrade into AMPA
(Nowack, 2003). In this study, the detection of AMPA in urban
streams cannot be attributed to the discharge of wastewater
treatment plant effluents since in Melbourne, sewerage system and
stormwater are separated except on rare occasions. However, it is
possible that phosphonate detergents end up in the stormwater
from general household use, e.g. washing down of patios, decks,
cars in driveways etc.
3.3. Risks
Urban wetlands not only act as water retention and depuration
areas but also offer an appealing environment for diverse biota,
such as endemic birds, frogs, and fish. These constructed aquatic
ecosystems are mainly quiescent water bodies that only have a
major water input during storm events. They are, therefore, a sink
for any chemicals that may be washed away from houses or in-
dustrial developments (Allinson et al., 2015) and can be affected by
the accumulation of pollutants, such as glyphosate. There are
guidelines for the construction of these stormwater wetlands in
Victoria (Melbourne Water, 2017a,b), but not for chemicals in
entrained stormwater other than for total nitrogen and phos-
phorus. The water in and/or extracted from stormwater retention
ponds and wetlands may pose a (real or perceived) risk to soil and
aquatic organisms from chemicals contained therein. In this
context, the concentrations of glyphosate detected in this study
were well below the strictest level of protection for freshwater
species (99% in pristine ecosystems; 370 mg L
1) established by the
Australian and New Zealand Guidelines for Fresh and Marine Water
Quality (ANZECC and ARMCANZ, 2000). The Canadian Council of
Ministers of the Environment Water Quality Guidelines has set a
value of 800 mg L
1 as the benchmark for protection of aquatic life
from glyphosate (Canadian Council of Ministers of the
Environment, 2012). In the USA, the chronic Aquatic Life Bench-
marks for glyphosate are 25,700 and 49,900 mg L
1 for fish and for
aquatic invertebrates, respectively (EPA, 2016). In the EU, there are
no defined environmental quality standards for glyphosate or
AMPA in surface water (Sze
k

acs and Darvas, 2018). However, the EU
has a very strict maximum allowed level of glyphosate in drinking
water of 0.1 mg L
1 (European Parliament and Council, 1998).
Currently, there is no agreement among countries’ regulation
agencies on the human health-based limits for glyphosate in
drinking water. For example in Australia, the guideline value for
glyphosate in drinking water is 1000 mg L
1 (HMRC and NRMMC,
2011), while in the USA the limit is 700 mg L
1 (US EPA 2002) and
in Canada the maximum allowed concentration is 280 mg L
1
(Guidelines for Canadian Drinking Water Quality, 2014). Although
the sampled wetlands and streams are not used as a source of
drinking water, none of the samples exceeded the guideline value
established by the Australian Drinking Water Guidelines (HMRC
 E. Okada et al. / Water Research 168 (2020) 115139
and NRMMC, 2011).
Although glyphosate in surface water can disturb aquatic com-
munities, the concentrations detected in this study do not suggest
an environmental risk from this chemical alone. For example,
glyphosate concentrations of 360 mg L
1 are required to increase
infection of the freshwater fish, Galaxias anomalus, by a flatworm
parasite (Kelly et al., 2010). Even higher chronic exposure levels of
glyphosate (3600 mg L
1) are needed to cause deleterious effects in
the reproductive cycle of silver catfish (Rhamdia quelen) (Soso et al.,
2007). Constructed stormwater wetlands are habitat for many
types of frog in Victoria, however, the risks to these organisms also
appears low since glyphosate concentrations as high as
300,000 mg L
1 do not cause histopathologic lesions in the liver of
tadpoles of the frog Leptodactylus latrans (Bach et al., 2018). In the
case of re-using stormwater as a source of irrigation, the concen-
trations that we have detected in this study likely do not pose a risk
to crops (Comes and Kelley, 1979). However, other pesticides may
be present in aquatic environments, and can have an additive,
synergistic or antagonist effect. As described earlier in Allinson
et al. (2017), Melbourne’s stormwater wetlands contain many
other herbicides, such as simazine, MCPA, diuron, atrazine and 2,4
D. Some of these compounds increase their toxicity when in the
presence of glyphosate, for example, diuron and glyphosate have
been shown to have a potential synergic sublethal effect in tropical
fish (Rossi et al., 2011). Another commonly detected herbicide in
Melbourne’s urban waters, 2,4 D, also has a synergistic effect on fish
mortality when combined with glyphosate (Ruiz de Arcaute et al.,
2018). Consequently, the risk to aquatic ecosystems from glypho-
sate may be higher than predicted when considering glyphosate
alone, especially considering its high FOD in urban surface waters.
4. Conclusions
This is the first report of glyphosate and AMPA occurrence and
seasonal variations in surface waters associated with urban land-
use in Australia. Glyphosate and AMPA were ubiquitous in storm-
water wetlands and streams from sub-urban and urban areas of
Melbourne. In this case, the sources of environmental contamina-
tion of glyphosate and AMPA were associated with urban uses,
instead of agricultural practices. Glyphosate and AMPA should be
considered priority substances in urban stormwater and included
in routine water quality monitoring studies, due to their potential
of contaminating surface waters.
Declaration of competing interest
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
Acknowledgments
E.Okada thanks the Department of Education and Training of the
Australian Commonwealth Government for her Endeavour
Research Fellowship [5911e2017]. The authors would like to thank
T. Coggan and Agilent Technologies for technical assistance with
instrument setup and operation.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.watres.2019.115139.
7
References
Allinson, G., Bui, A., Zhang, P., Rose, G., Wightwick, A.M., Allinson, M., Pettigrove, V.,
2014. Investigation of 10 herbicides in surface waters of a horticultural pro-
duction catchment in southeastern Australia. Arch. Environ. Contam. Toxicol. 67
(3), 358e373. https://doi.org/10.1007/s00244-014-0049-z.
Allinson, G., Zhang, P., Bui, A., Allinson, M., Rose, G., Marshall, S., Pettigrove, V., 2015.
Pesticide and trace metal occurrence and aquatic benchmark exceedances in
surface waters and sediments of urban wetlands and retention ponds in Mel-
bourne, Australia. Environ. Sci. Pollut. Control Ser. 22 (13), 10214e10226.
https://doi.org/10.1007/s11356-015-4206-3.
Allinson, M., Zhang, P., Bui, A., Myers, J.H., Pettigrove, V., Rose, G., Salzman, S.A.,
Walters, R., Allinson, G., 2017. Herbicides and trace metals in urban waters in
Melbourne, Australia (2011 e 12): concentrations and potential impact. Environ.
Sci. Pollut. Control Ser. 24, 7274e7284. https://doi.org/10.1007/s11356-017-
8395-9.
ANZECC, ARMCANZ (Australian and New Zealand Environment and Conservation
Council, Agriculture and Resource Management Council of Australia and New
Zealand), 2000. Australian and New Zealand Guidelines for Fresh and Marine
Water Quality, pp. 1e103.
APVMA, 2018. Final pesticide and veterinary medicines product sales 2016-17
Financial Year. In: Commonwealth of Australia Gazette No APVMA 6, Tuesday 27
March 2018. Australian Pesticides and Veterinary Medicines Authority, Can-
berra, Australia.
Bach, N.C., Marino, D.J., Natale, G.S., Somoza, G.M., 2018. Effects of glyphosate and
its commercial formulation, Roundup® Ultramax, on liver histology of tadpoles
of the neotropical frog, Leptodactylus latrans (amphibia: Anura). Chemosphere
202, 289e297.
Battaglin, W.A., Meyer, M.T., Kuivila, K.M., Dietze, J.E., 2014. Glyphosate and its
degradation product AMPA occur frequently and widely in US soils, surface
water, groundwater, and precipitation. J. Am. Water Resour. Assoc. 50 (2),
275e290.
Benbrook, C.M., 2016. Trends in glyphosate herbicide use in the United States and
globally. Environ. Sci. Eur. 28, 3. https://doi.org/10.1186/s12302-016-0070-0.
Bento, C.P., Yang, X., Gort, G., Xue, S., van Dam, R., Zomer, P., Mol, H.G., Ritsema, C.J.,
Geissen, V., 2016. Persistence of glyphosate and aminomethylphosphonic acid
in loess soil under different combinations of temperature, soil moisture and
light/darkness. Sci. Total Environ. 572, 301e311.
Bento, C.P., Goossens, D., Rezaei, M., Riksen, M., Mol, H.G., Ritsema, C.J., Geissen, V.,
2017. Glyphosate and AMPA distribution in wind-eroded sediment derived from
loess soil. Environ. Pollut. 220, 1079e1089.
Borggaard, O.K., Gimsing, A.L., 2008. Fate of glyphosate in soil and the possibility of
leaching to ground and surface waters: a review. Pest Manag. Sci. 64, 441e456.
https://doi.org/10.1002/ps.1512.
Botta, F., Lavison, G., Couturier, G., Alliot, F., Moreau-Guigon, E., Fauchon, N.,
Guery, B., Chevreuil, M., Blanchoud, H., 2009. Transfer of glyphosate and its
degradate AMPA to surface waters through urban sewerage systems. Chemo-
sphere 77 (1), 133e139. https://doi.org/10.1016/j.chemosphere.2009.05.008.
Canadian Council of Ministers of the Environment, 2012. Canadian water quality
guidelines for the protection of aquatic life: Glyphosate. In: Canadian envi-
ronmental quality guidelines. Canadian Council of Ministers of the Environ-
ment, Winnipeg.
Catchment and Land Protection, 2011. Act 1994. Version No. 041. No. 52 of 1994.
Presence and Levels of Priority Pesticides in Selected Canadian Aquatic Eco-
systems. Water Science and Technology Directorate. Environment Canada.
https://www.ec.gc.ca/eaudouce-freshwater/5EEF7087-5907-4E79-8EFE-
37E6541E6FCF/
PresenceAndLevelsOfPriorityPesticidesInSelectedCanadianAquaticEcosystems%
20FINAL-s.pdf.
Comes, R.D., Kelley, A.D., 1979. Response of certain crops to glyphosate in irrigation
water. Weed Sci. 27 (6), 658e660.
Coupe, R.H., Kalkhoff, S.J., Capel, P.D., Gregoire, C., 2012. Fate and transport of
glyphosate and aminomethylphosphonic acid in surface waters of agricultural
basins. Pest Manag. Sci. 68 (1), 16e30. https://doi.org/10.1002/ps.2212.
EPA, 2016. Aquatic Life Benchmarks and Ecological Risk Assessments for Registered
Pesticides. United States Environmental Protection Agency, Washington, DC.
https://www.epa.gov/pesticide-science-and-assessing-pesticide-risks/aquatic-
life-benchmarks-and-ecological-risk#ref_2.
European Parliament and Council, 1998. Council Directive of 98/83/EC of 3
November 1998 on the Quality of Water Intended for Human Consumption,
pp. 32e54. OJ L330. http://eur-lex.europa.eu/legal-content/EN/TXT/?uri¼CELEX
%3A31998L0083G.
Glozier, N.E., Struger, J., Cessna, A.J., Gledhill, M., Rondeau, M., Ernst, W.R.,
Sekela, M.A., Cagampan, S.J., Sverko, E., Murphy, C., Murray, J.L., 2012. Occur-
rence of glyphosate and acidic herbicides in select urban rivers and streams in
Canada, 2007. Environ. Sci. Pollut. Control Ser. 19 (3), 821e834. https://doi.org/
10.1007/s11356-011-0600-7.
Grandcoin, A., Piel, S., Baures, E., 2017. AminoMethylPhosphonic acid (AMPA) in
natural waters: its sources, behavior and environmental fate. Water Res. 117,
187e197.
Guidelines for Canadian Drinking Water Quality e Summary Table, 2014. Federal-
Provincial-Territorial Committee on Drinking Water of the Federal-Provincial-
Territorial Committee on Health and the Environment. http://wwwhc-scgcca/
ewh-semt/pubs/water-eau/sum_guide-res_recom/index-engphp. (Accessed 29
8 E. Okada et al. / Water Research 168 (2020) 115139
September 2017).
Hanke, I., Wittmer, I., Bischofberger, S., Stamm, C., Singer, H., 2010. Relevance of
urban glyphosate use for surface water quality. Chemosphere 81 (3), 422e429.
https://doi.org/10.1016/j.chemosphere.2010.06.067.
HMRC, NRMMC, 2011. Australian Drinking Water Guidelines Paper 6 National Water
Quality Management Strategy. National Health and Medical Research Council.
National Resource Management Ministerial Council, Commonwealth of
Australia, Canberra.
Imfeld, G., Lefrancq, M., Maillard, E., Payraudeau, S., 2013. Transport and attenuation
of dissolved glyphosate and AMPA in a stormwater wetland. Chemosphere 90
(4), 1333e1339. https://doi.org/10.1016/j.chemosphere.2012.04.054.
Kelly, D.W., Poulin, R., Tompkins, D.M., Townsend, C.R., 2010. Synergistic effects of
glyphosate formulation and parasite infection on fish malformations and sur-
vival. J. Appl. Ecol. 47 (2), 498e504.
Kolpin, D.W., Thurman, E.M., Lee, E.A., Meyer, M.T., Furlong, E.T., Glassmeyer, S.T.,
2006. Urban contributions of glyphosate and its degradate AMPA to streams in
the United States. Sci. Total Environ. 354, 191e197. https://doi.org/10.1016/
j.scitotenv.2005.01.028.
la Cecilia, D., Maggi, F., 2018. Analysis of glyphosate degradation in a soil microcosm.
Environ. Pollut. 233, 201e207.
la Cecilia, D., Tang, F.H., Coleman, N.V., Conoley, C., Vervoort, R.W., Maggi, F., 2018.
Glyphosate dispersion, degradation, and aquifer contamination in vineyards
and wheat fields in the Po Valley. Italy. Water res. 146, 37e54.
Mauffrey, F., Baccara, P.Y., Gruffaz, C., Vuilleumier, S., Imfeld, G., 2017. Bacterial
community composition and genes for herbicide degradation in a stormwater
wetland collecting herbicide runoff. Water, Air, Soil Pollut. 228 (12), 452.
https://doi.org/10.1007/s11270-017-3625-9.
Melbourne Water, 2017a. Yarra Catchment. https://www.melbournewater.com.au/
water/health-and-monitoring/river-health-and-monitoring/yarra-catchment.
Melbourne Water, 2017b. Wetland design manual. https://www.melbournewater.
com.au/planning-and-building/developer-guides-and-resources/standards-
and-specifications/constructed-0.
Morillo, E., Undabeytia, T., Maqueda, C., Ramos, A., 2000. Glyphosate adsorption on
soils of different characteristics. Influence of copper addition. Chemosphere 40
(1), 103e107. https://doi.org/10.1016/S0045-6535(99)00255-6.
Nowack, B., 2003. Environmental chemistry of phosphonates. Water Res. 37,
2533e2546.
Okada, E., Costa, J.L., Bedmar, F., 2016. Adsorption and mobility of glyphosate in
different soils under no-till and conventional tillage. Geoderma 263, 78e85.
Okada, E., Costa, J.L., Bedmar, F., 2017. Glyphosate dissipation in different soils under
no-till and conventional tillage. Pedosphere. https://doi.org/10.1016/S1002-
0160(17)60430-2.
Okada, E., Pe
rez, D., De Gero

nimo, E., Aparicio, V., Massone, H., Costa, J.L., 2018. Non-
point source pollution of glyphosate and AMPA in a rural basin from the
southeast Pampas, Argentina. Environ. Sci. Pollut. Control Ser. 25 (15),
15120e15132.
Okada, E., Coggan, T., Anumol, T., Clarke, B., Allinson, G., 2019. A simple and rapid
direct injection method for the determination of glyphosate and AMPA in
environmental water samples. Anal. Bioanal. Chem. 411, 715e724. https://
doi.org/10.1007/s00216-018-1490-z.
Paudel, P., Negusse, A., Jaisi, D.P., 2015. Birnessite-catalyzed degradation of glyph-
osate: a mechanistic study aided by kinetics batch studies and NMR spectros-
copy. Soil Sci. Soc. Am. J. 79, 815e825.
Pe
rez, D.J., Okada, E., De Gero
nimo, M.L., Menone, V.C., Aparicio, V., Costa, J.L., 2017a.
Spatial and temporal trends and flow dynamics of glyphosate and other pes-
ticides within an agricultural watershed in Argentina. Environ. Toxicol. Chem.
36 (12), 3206e3216. https://doi.org/10.1002/etc.3897.
Pe
rez, D.J., Okada, E., Menone, M.L., Costa, J.L., 2017b. Can an aquatic macrophyte
bioaccumulate glyphosate? Development of a new method of glyphosate
extraction in Ludwigia peploides and watershed scale validation. Chemosphere
185, 975e982. https://doi.org/10.1016/j.chemosphere.2017.07.093.
Preston, C., 2018. The Australian glyphosate sustainability working group. Website;.
http://www.glyphosateresistance.org.au/.
Ramwell, C.T., Kah, M., Johnson, P.D., 2014. Contribution of household herbicide
usage to glyphosate and its degradate aminomethylphosphonic acid in surface
water drains. Pest Manag. Sci. 70 (12), 1823e1830.
Reichenberger, S., Bach, M., Skitschak, A., Frede, H.G., 2007. Mitigation strategies to
reduce pesticide inputs into ground- and surface water and their effectiveness;
A review. Sci. Total Environ. 348, 1e35. https://doi.org/10.1016/
j.scitotenv.2007.04.046.
Rippy, M.A., Deletic, A., Black, J., Aryal, R., Lampard, J.L., Tang, J.Y.M., McCarthy, D.,
Kolotelo, P., Sidhu, J., Gernjak, W., 2017. Pesticide occurrence and spatio-
temporal variability in urban run-off across Australia. Water Res. 115, 245e255.
Rossi, S.C., Da Silva, M.D., Piancini, L.D.S., Ribeiro, C.A.O., Cestari, M.M., de
Assis, H.C.S., 2011. Sublethal effects of waterborne herbicides in tropical fresh-
water fish. Bull. Environ. Contam. Toxicol. 87 (6), 603e607.
Ruiz de Arcaute, C.R., Soloneski, S., Larramendy, M.L., 2018. Opposite effects of
mixtures of commercial formulations of glyphosate with auxinic herbicides on
the ten spotted live-bearer fish Cnesterodon decemmaculatus (Pisces, Poecilii-
dae). Environ. Pollut. 240, 858e866.
Sch€afer, R.B., Pettigrove, V., Rose, G., Allinson, G., Wightwick, A., von der Ohe, P.C.,
Shimeta, J., Kühne, R., Kefford, B.J., 2011. Effects of pesticides monitored with
three sampling methods in 24 sites on macroinvertebrates and microorgan-
isms. Environ. Sci. Technol. 45 (4), 1665e1672. https://doi.org/10.1021/
es2013006.
Sidoli, P., Baran, N., Angulo-Jaramillo, R., 2016. Glyphosate and AMPA adsorption in
soils: laboratory experiments and pedotransfer rules. Environ. Sci. Pollut.
Control Ser. 23 (6), 5733e5742.
Simonsen, L., Fomsgaard, I., Svensmark, B., Spliid, N.H., et al., 2008. Fate and
availability of glyphosate and AMPA in agricultural soil. J. Environ. Sci. Health B
43, 365e375.
Skark, C., Zullei-Seibert, N., Willme, U., Gatzemann, U., Schlett, C., 2004. Contribu-
tion of non-agricultural pesticides to pesticide load in surface water. Pest
Manag. Sci. 60 (6), 525e530.
Soso, A.B., Barcellos, L.J.G., Ranzani-Paiva, M.J., Kreutz, L.C., Quevedo, R.M.,
Anziliero, D., Lima, M., Bolognesi da Silva, L., Ritter, F., Calliari Bedin, A.,
Finco, J.A., 2007. Chronic exposure to sub-lethal concentration of a glyphosate-
based herbicide alters hormone profiles and affects reproduction of female
Jundia
(Rhamdia quelen). Environ. Toxicol. Pharmacol. 23, 308e313.
Spanoghe, P., Claeys, J., Pinoy, L., Steurbaut, W., 2005. Rainfastness and adsorption of
herbicides on hard surfaces. Pest Manag. Sci. 61 (8), 793e798.
Sprankle, P., Meggit, W.F., Penner, D., 1975. Adsorption, mobility and microbial
degradation of glyphosate in the soil. Weed Sci. 23, 229e234.
State Government of Victoria, 2019. Maps and spatial data. https://www2.delwp.
vic.gov.au/maps.
Struger, J., Van Stempvoort, D.R., Brown, S.J., 2015. Sources of amino-
methylphosphonic acid (AMPA) in urban and rural catchments in Ontario,
Canada: glyphosate or phosphonates in wastewater? Environ. Pollut. 204,
289e297.
Sun, M., Li, H., Jaisi, D.P., 2019. Degradation of glyphosate and bioavailability of
phosphorus derived from glyphosate in a soil-water system. Water Res. 163,
114840.
Syversen, N., Bechmann, M., 2004. Vegetative buffer zones as pesticide filters for
simulated surface runoff. Ecol. Eng. 22 (3), 175e184. https://doi.org/10.1016/
j.ecoleng.2004.05.002.

ka
Sze
cs, A., Darvas, B., 2018. Re-registration challenges of glyphosate in the Euro-
pean Union. Front. Environ. Sci. 6, 78. https://doi.org/10.3389/fenvs.2018.00078.
Tang, T., Boe €nne, W., Desmet, N., Seuntjens, P., Bronders, J., Van Griensven, A., 2015.
Quantification and characterization of glyphosate use and loss in a residential
area. Sci. Total Environ. 517, 207e214. https://doi.org/10.1016/
j.scitotenv.2015.02.040.
Tush, D., Maksimowicz, M.M., Meyer, M.T., 2018. Dissipation of polyoxyethylene
tallow amine (POEA) and glyphosate in an agricultural field and their co-
occurrence on streambed sediments. Sci. Total Environ. 636, 212e219.
Van Stempvoort, D.R., Spoelstra, J., Senger, N.D., Brown, S.J., Post, R., Struger, J., 2016.
Glyphosate residues in rural groundwater, nottawasaga river watershed,
ontario, Canada. Pest Manag. Sci. 72, 1862e1872. https://doi.org/10.1002/ps.
4218.
Wang, S., Seiwert, B., K€ astner, M., Miltner, A., Scha€ffer, A., Reemtsma, T., Yang, Q.,
Nowak, K.M., 2016. (Bio) degradation of glyphosate in water-sediment
microcosmseA stable isotope co-labeling approach. Water Res. 99, 91e100.
Wightwick, A.M., Bui, A.D., Zhang, P., Rose, G., Allinson, M., Myers, J.H.,
Reichman, S.M., Menzies, N.W., Pettigrove, V., Allinson, G., 2012. Environmental
fate of fungicides in surface waters of a horticultural-production catchment in
southeastern Australia. Arch. Environ. Contam. Toxicol. 62 (3), 380e390.
Yang, X., Wang, F., Bento, C.P., Xue, S., Gai, L., van Dam, R., Mol, H., Ritsema, C.J.,
Geissen, V., 2015. Short-term transport of glyphosate with erosion in Chinese
loess soilda flume experiment. Sci. Total Environ. 512, 406e414.