Environ Sci Pollut Res DOI 10.1007/s11356-015-4206-3
RESEARCH ARTICLE
Pesticide and trace metal occurrence and aquatic benchmark exceedances in surface waters and sediments of urban wetlands and retention ponds in Melbourne, Australia Graeme Allinson & Pei Zhang & AnhDuyen Bui & Mayumi Allinson & Gavin Rose & Stephen Marshall & Vincent Pettigrove
Received: 21 December 2014 / Accepted: 4 February 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract Samples of water and sediments were collected from 24 urban wetlands in Melbourne, Australia, in April 2010, and tested for more than 90 pesticides using a range of gas chromatographic (GC) and liquid chromatographic (LC) techniques, sample ‘hormonal’ activity using yeastbased recombinant receptor-reporter gene bioassays, and trace metals using spectroscopic techniques. At the time of sampling, there was almost no estrogenic activity in the water column. Twenty-three different pesticide residues were observed in one or more water samples from the 24 wetlands; chemicals observed at more than 40 % of sites were simazine (100 %), atrazine (79 %), and metalaxyl and terbutryn (46 %). Using the toxicity unit (TU) concept, less than 15 % of the Responsible editor: Laura McConnell Electronic supplementary material The online version of this article (doi:10.1007/s11356-015-4206-3) contains supplementary material, which is available to authorized users. G. Allinson School of Applied Sciences, RMIT University, Melbourne, VIC 3001, Australia G. Allinson : M. Allinson Agriculture Research and Development Division, Department of Environment and Primary Industries, DEPI Queenscliff Centre, Queenscliff, VIC 3225, Australia
detected pesticides were considered to pose an individual, short-term risk to fish or zooplankton in the ponds and wetlands. However, one pesticide (fenvalerate) may have posed a possible short-term risk to fish (log10TUf > −3), and three pesticides (azoxystrobin, fenamiphos and fenvalerate) may have posed a risk to zooplankton (logTUzp between −2 and −3); all the photosystem II (PSII) inhibiting herbicides may have posed a risk to primary producers in the ponds and wetlands (log10TUap and/or log10TUalg > -3). The wetland sediments were contaminated with 16 different pesticides; no chemicals were observed at more than one third of sites, but based on frequency of detection and concentrations, bifenthrin (33 %, maximum 59 μg/kg) is the priority insecticide of concern for the sediments studied. Five sites returned a TU greater than the possible effect threshold (i.e. log10TU > 1) as a result of bifenthrin contamination of their sediments. Most sediments did not exceed Australian sediment quality guideline levels for trace metals. However, more than half of the sites had threshold effect concentration quotients (TECQ) values >1 for Cu (58 %), Pb (50 %), Ni (67 %) and Zn (63 %), and 75 % of sites had mean probable effect concentration quotients (PECQ) >0.2, suggesting that the collected sediments may have been having some impact on sediment-dwelling organisms.
G. Allinson (*) : P. Zhang : A. Bui : M. Allinson : G. Rose : S. Marshall : V. Pettigrove
Keywords Pesticides . Trace metals . Two-hybrid yeast recombinant receptor-reportergeneassay activity . Storm water ponds and wetlands . Melbourne
Centre for Aquatic Pollution Identification and Management (CAPIM), The University of Melbourne, Parkville, VIC 3010, Australia e-mail: [email protected]
Introduction
P. Zhang : A. Bui : G. Rose Agriculture Research and Development Division, Department of Environment and Primary Industries, Ernest Jones Drive, Macleod, Victoria 3085, Australia
Increasingly, harvested stormwater is seen as the last unexploited water resource in many Australian cities, one that in turn might be used to augment the security of water supplies. For
Environ Sci Pollut Res
instance, in Melbourne, a city of approximately 4.5 million inhabitants, urban water consumption during the recent drought years of high-to-medium level water restriction, e.g. 2005–2007, was from a minimum of 277 to a maximum of 333 GL/annum, whereas recently, relatively unrestricted water consumption is higher, e.g. 360 GL in 2011 (Melbourne Water 2014). On the other hand, the amount of stormwater runoff from urban catchments in Melbourne is approximately 540 GL/annum (Melbourne Water 2012). This highlights the potential of stormwater to meet a substantial proportion of Melbourne’s water demand provided it is managed correctly. The key principles for urban storm water management include integrating storm water treatment into the urban landscape, e.g. through the use of ponds and wetlands in new urban developments, and the retro-fitting of such amenities into older developments, and reducing potable water demand by using storm water as a resource through capture and reuse for non-potable purposes (CSIRO 1999). This will require protecting water quality while minimising development/utilisation costs. There are guidelines in Victoria for the amount of litter, suspended solids, total nitrogen and phosphorus in storm water but none for other chemical pollutants in storm water and how they might impact water reuse in the urban environment, i.e. in parks and gardens. In that context, urban municipal storm water retention ponds and wetlands may contain a wide structural variety of trace organic contaminants that, for the most part, restricts the applicability of chemical analytical methods that screen for only structurally or physicochemically similar chemicals. Several in vitro assays have been developed to screen the ‘hormonal’ activity of compounds in natural waters, including ligand-binding assays, recombinant receptor-reporter gene assays, assays based on the measurement of cell proliferation, and enzyme-linked immunosorbent assays (ELISA; Streck 2009; Kinnberg 2003). Recombinant receptor-reporter gene assays, such as the yeast two-hybrid bioassays used in this study, measure the activation of receptor and allow for quantification of hormonal activity, without having to know the precise chemical make-up of the sample. They have, however, been little utilised on natural water samples in Australia, with no published Victorian storm water data available at the inception of this study. Pesticides are widely used for the suppression of unwanted plants (weeds) and removal of nuisance insects and fungi in domestic, municipal and business environments as well agricultural production systems. In monetary and volume terms, herbicides are the top-ranked category of pesticides sold in Australia (based on final aggregated 2012–2013 data: 2866 products, A$1262 million), more than thrice that of insecticides (1308 products; A$351 million) and seven times the sales of fungicides (799 products; A$170 million; APVMA 2014a). This data includes pesticides registered and sold for
domestic and commercial use as well as agricultural uses. There are 300 pesticides (active ingredients) registered for use in Victoria (not including repeat registrations and registrations of manufacturing concentrates of an otherwise registered chemical; APVMA 2014b), of which 115 are herbicides, 64 insecticides, 63 fungicides, and 58 ‘other’ chemicals (including acaricides, molluscicides, nematicides, rodenticides and plant growth regulators). The actual number and amount of chemicals used in the state in any year is unknown because, while farmers and other commercial pesticide users must keep written records of pesticide use, there is currently no requirement for any user, whether farmer, licensed chemical user, business or householder, to report pesticide use to either local or centralised authority. Contaminated sediments can act as significant sources of trace elements to the aquatic environment through their release into the overlying water column (Chon et al. 2012), affecting both the immediate environment and aquatic environments further downstream if remobilised through storm events or other disturbances. Moreover, metals in sediments can have adverse effects on aquatic ecosystems, including direct toxicity (e.g. Pettigrove and Hoffmann 2005), and assemblage and community effects (e.g. Carew et al. 2007). There have been many surveys conducted investigating the concentration of metals in aquatic ecosystems that receive urban stormwater runoff, with the interest in sediment quality increasing, as our understanding of sediment chemistry and toxicology has improved. A recent review by the Centre for Aquatic Pollution Identification and Management (CAPIM; K. Townsend, unpublished data) found that the median Zn concentration in urban sediments in Oceania (including Australia; 477 mg/kg) is twice as high as median concentrations found in Asia, Europe and North America (range 107 to 183 mg/kg). The presence of pesticides in the dissolved state in the aquatic environment has been extensively studied internationally and levels in the micrograms per litre and sub-micrograms per litre range reported in drinking and surface water. For instance, in North America the United States Geological Survey (USGS), large, nation-wide surveys of pesticide residues in surface and ground waters produced a wealth of data for use in the environmental risk assessment of herbicides (Gillom et al. 2006). Despite the potential environmental risks, pesticides have received little attention in Victorian waterways compared to other jurisdictions (Wightwick and Allinson 2007). Indeed, prior to this study, there had been no significant studies investigating pesticides in Melbourne’s stormwater wetlands and retention ponds’ waters and sediment. In recognition of the potential risks that chemicals pose to aquatic ecosystems and the lack of robust information on the levels of such compounds in Victoria, samples were collected from 24 urban and peri-urban wetlands around Melbourne in April 2010. One of the challenges in environmental monitoring is the lack of a combined approach using chemical
Environ Sci Pollut Res
measurements of contaminants and bioanalytical tools to investigate health of ecosystems as a parallel study (Sumpter and Johnson 2008). To address this challenge, samples were prepared for a number of chemical and bioanalytical tests, including measurement of more than 90 pesticides using a range of gas chromatographic (GC) and liquid chromatographic (LC) techniques, trace metals using inductively coupled plasma emission spectroscopy, and sample ‘hormonal’ activity using a yeast-based recombinant receptor-reporter gene bioassays into which the genetic coding for the human and medaka (Oryzias latipes) estrogen (hERα and medERα) had been inserted.
Materials and methods Study sites Twenty-four urban and peri-urban wetlands in and around the city of Melbourne were investigated in this study (Fig. 1). In broad terms, these sites were selected as a broad representation of the wide range of urban stormwater treatment wetland designs found in Melbourne, across the major soils types in the region and representing both new developments and wellestablished suburbs. More specifically, the study itself was a first-pass assessment of the risks to biota from pollution in the urban stormwater treatment wetlands, with a specific focus on establishing whether there is widespread pesticide pollution in urban wetlands and streams in the Melbourne area. Consequently, the sites selected included some urban wetlands that are in green-field developments where drainage systems Fig. 1 Approximate location of sampling sites in and around the city of Melbourne, in Victoria, Australia
are/have been constructed using water-sensitive urban designs (WSUDs; sites 300–308 in housing estates no more than 15 years old (and most less than 10 years old)) and older, more established suburbs where the drainage systems were not specifically designed using WSUDs (e.g. sites 309–324).
Water and sediment sampling Water and sediment samples were collected from each of the 24 sampling sites on a single occasion in April 2010. Water temperature, pH and electrical conductivity (EC) were measured in situ at the time of sampling using a field meter (TPS 90FLs+, TPS Australia, Springwood, Queensland). The total organic carbon (TOC) content of water samples was determined according to the American Public Health Association (APHA) method 5301B (APHA 2005). Sediment samples were analysed for organic carbon (OC) using the Walkley and Black method (Rayment and Lyons 2011; Method 6A1). Water samples were collected as a series of ‘grab’ or spot samples from the 24 urban, peri-urban and rural waters (Fig. 1). Samples (1 L) were collected in amber glass bottles (one bottle per subsequent test) and stored on ice at 1–4 °C in a portable ice box or portable fridge and then at 4 °C until processed (within 36 h of collection for bioassays and within 2 weeks for pesticide residues testing). Surficial sediments were sampled with a dip net and transferred to a 63-μm sieve and sieved wet, on site, into a large plastic bucket prior to their subsequent use in Chironomid-based toxicity tests (standardising collected sediments to a fine grain size (e.g.
RESEARCH ARTICLE
Pesticide and trace metal occurrence and aquatic benchmark exceedances in surface waters and sediments of urban wetlands and retention ponds in Melbourne, Australia Graeme Allinson & Pei Zhang & AnhDuyen Bui & Mayumi Allinson & Gavin Rose & Stephen Marshall & Vincent Pettigrove
Received: 21 December 2014 / Accepted: 4 February 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract Samples of water and sediments were collected from 24 urban wetlands in Melbourne, Australia, in April 2010, and tested for more than 90 pesticides using a range of gas chromatographic (GC) and liquid chromatographic (LC) techniques, sample ‘hormonal’ activity using yeastbased recombinant receptor-reporter gene bioassays, and trace metals using spectroscopic techniques. At the time of sampling, there was almost no estrogenic activity in the water column. Twenty-three different pesticide residues were observed in one or more water samples from the 24 wetlands; chemicals observed at more than 40 % of sites were simazine (100 %), atrazine (79 %), and metalaxyl and terbutryn (46 %). Using the toxicity unit (TU) concept, less than 15 % of the Responsible editor: Laura McConnell Electronic supplementary material The online version of this article (doi:10.1007/s11356-015-4206-3) contains supplementary material, which is available to authorized users. G. Allinson School of Applied Sciences, RMIT University, Melbourne, VIC 3001, Australia G. Allinson : M. Allinson Agriculture Research and Development Division, Department of Environment and Primary Industries, DEPI Queenscliff Centre, Queenscliff, VIC 3225, Australia
detected pesticides were considered to pose an individual, short-term risk to fish or zooplankton in the ponds and wetlands. However, one pesticide (fenvalerate) may have posed a possible short-term risk to fish (log10TUf > −3), and three pesticides (azoxystrobin, fenamiphos and fenvalerate) may have posed a risk to zooplankton (logTUzp between −2 and −3); all the photosystem II (PSII) inhibiting herbicides may have posed a risk to primary producers in the ponds and wetlands (log10TUap and/or log10TUalg > -3). The wetland sediments were contaminated with 16 different pesticides; no chemicals were observed at more than one third of sites, but based on frequency of detection and concentrations, bifenthrin (33 %, maximum 59 μg/kg) is the priority insecticide of concern for the sediments studied. Five sites returned a TU greater than the possible effect threshold (i.e. log10TU > 1) as a result of bifenthrin contamination of their sediments. Most sediments did not exceed Australian sediment quality guideline levels for trace metals. However, more than half of the sites had threshold effect concentration quotients (TECQ) values >1 for Cu (58 %), Pb (50 %), Ni (67 %) and Zn (63 %), and 75 % of sites had mean probable effect concentration quotients (PECQ) >0.2, suggesting that the collected sediments may have been having some impact on sediment-dwelling organisms.
G. Allinson (*) : P. Zhang : A. Bui : M. Allinson : G. Rose : S. Marshall : V. Pettigrove
Keywords Pesticides . Trace metals . Two-hybrid yeast recombinant receptor-reportergeneassay activity . Storm water ponds and wetlands . Melbourne
Centre for Aquatic Pollution Identification and Management (CAPIM), The University of Melbourne, Parkville, VIC 3010, Australia e-mail: [email protected]
Introduction
P. Zhang : A. Bui : G. Rose Agriculture Research and Development Division, Department of Environment and Primary Industries, Ernest Jones Drive, Macleod, Victoria 3085, Australia
Increasingly, harvested stormwater is seen as the last unexploited water resource in many Australian cities, one that in turn might be used to augment the security of water supplies. For
Environ Sci Pollut Res
instance, in Melbourne, a city of approximately 4.5 million inhabitants, urban water consumption during the recent drought years of high-to-medium level water restriction, e.g. 2005–2007, was from a minimum of 277 to a maximum of 333 GL/annum, whereas recently, relatively unrestricted water consumption is higher, e.g. 360 GL in 2011 (Melbourne Water 2014). On the other hand, the amount of stormwater runoff from urban catchments in Melbourne is approximately 540 GL/annum (Melbourne Water 2012). This highlights the potential of stormwater to meet a substantial proportion of Melbourne’s water demand provided it is managed correctly. The key principles for urban storm water management include integrating storm water treatment into the urban landscape, e.g. through the use of ponds and wetlands in new urban developments, and the retro-fitting of such amenities into older developments, and reducing potable water demand by using storm water as a resource through capture and reuse for non-potable purposes (CSIRO 1999). This will require protecting water quality while minimising development/utilisation costs. There are guidelines in Victoria for the amount of litter, suspended solids, total nitrogen and phosphorus in storm water but none for other chemical pollutants in storm water and how they might impact water reuse in the urban environment, i.e. in parks and gardens. In that context, urban municipal storm water retention ponds and wetlands may contain a wide structural variety of trace organic contaminants that, for the most part, restricts the applicability of chemical analytical methods that screen for only structurally or physicochemically similar chemicals. Several in vitro assays have been developed to screen the ‘hormonal’ activity of compounds in natural waters, including ligand-binding assays, recombinant receptor-reporter gene assays, assays based on the measurement of cell proliferation, and enzyme-linked immunosorbent assays (ELISA; Streck 2009; Kinnberg 2003). Recombinant receptor-reporter gene assays, such as the yeast two-hybrid bioassays used in this study, measure the activation of receptor and allow for quantification of hormonal activity, without having to know the precise chemical make-up of the sample. They have, however, been little utilised on natural water samples in Australia, with no published Victorian storm water data available at the inception of this study. Pesticides are widely used for the suppression of unwanted plants (weeds) and removal of nuisance insects and fungi in domestic, municipal and business environments as well agricultural production systems. In monetary and volume terms, herbicides are the top-ranked category of pesticides sold in Australia (based on final aggregated 2012–2013 data: 2866 products, A$1262 million), more than thrice that of insecticides (1308 products; A$351 million) and seven times the sales of fungicides (799 products; A$170 million; APVMA 2014a). This data includes pesticides registered and sold for
domestic and commercial use as well as agricultural uses. There are 300 pesticides (active ingredients) registered for use in Victoria (not including repeat registrations and registrations of manufacturing concentrates of an otherwise registered chemical; APVMA 2014b), of which 115 are herbicides, 64 insecticides, 63 fungicides, and 58 ‘other’ chemicals (including acaricides, molluscicides, nematicides, rodenticides and plant growth regulators). The actual number and amount of chemicals used in the state in any year is unknown because, while farmers and other commercial pesticide users must keep written records of pesticide use, there is currently no requirement for any user, whether farmer, licensed chemical user, business or householder, to report pesticide use to either local or centralised authority. Contaminated sediments can act as significant sources of trace elements to the aquatic environment through their release into the overlying water column (Chon et al. 2012), affecting both the immediate environment and aquatic environments further downstream if remobilised through storm events or other disturbances. Moreover, metals in sediments can have adverse effects on aquatic ecosystems, including direct toxicity (e.g. Pettigrove and Hoffmann 2005), and assemblage and community effects (e.g. Carew et al. 2007). There have been many surveys conducted investigating the concentration of metals in aquatic ecosystems that receive urban stormwater runoff, with the interest in sediment quality increasing, as our understanding of sediment chemistry and toxicology has improved. A recent review by the Centre for Aquatic Pollution Identification and Management (CAPIM; K. Townsend, unpublished data) found that the median Zn concentration in urban sediments in Oceania (including Australia; 477 mg/kg) is twice as high as median concentrations found in Asia, Europe and North America (range 107 to 183 mg/kg). The presence of pesticides in the dissolved state in the aquatic environment has been extensively studied internationally and levels in the micrograms per litre and sub-micrograms per litre range reported in drinking and surface water. For instance, in North America the United States Geological Survey (USGS), large, nation-wide surveys of pesticide residues in surface and ground waters produced a wealth of data for use in the environmental risk assessment of herbicides (Gillom et al. 2006). Despite the potential environmental risks, pesticides have received little attention in Victorian waterways compared to other jurisdictions (Wightwick and Allinson 2007). Indeed, prior to this study, there had been no significant studies investigating pesticides in Melbourne’s stormwater wetlands and retention ponds’ waters and sediment. In recognition of the potential risks that chemicals pose to aquatic ecosystems and the lack of robust information on the levels of such compounds in Victoria, samples were collected from 24 urban and peri-urban wetlands around Melbourne in April 2010. One of the challenges in environmental monitoring is the lack of a combined approach using chemical
Environ Sci Pollut Res
measurements of contaminants and bioanalytical tools to investigate health of ecosystems as a parallel study (Sumpter and Johnson 2008). To address this challenge, samples were prepared for a number of chemical and bioanalytical tests, including measurement of more than 90 pesticides using a range of gas chromatographic (GC) and liquid chromatographic (LC) techniques, trace metals using inductively coupled plasma emission spectroscopy, and sample ‘hormonal’ activity using a yeast-based recombinant receptor-reporter gene bioassays into which the genetic coding for the human and medaka (Oryzias latipes) estrogen (hERα and medERα) had been inserted.
Materials and methods Study sites Twenty-four urban and peri-urban wetlands in and around the city of Melbourne were investigated in this study (Fig. 1). In broad terms, these sites were selected as a broad representation of the wide range of urban stormwater treatment wetland designs found in Melbourne, across the major soils types in the region and representing both new developments and wellestablished suburbs. More specifically, the study itself was a first-pass assessment of the risks to biota from pollution in the urban stormwater treatment wetlands, with a specific focus on establishing whether there is widespread pesticide pollution in urban wetlands and streams in the Melbourne area. Consequently, the sites selected included some urban wetlands that are in green-field developments where drainage systems Fig. 1 Approximate location of sampling sites in and around the city of Melbourne, in Victoria, Australia
are/have been constructed using water-sensitive urban designs (WSUDs; sites 300–308 in housing estates no more than 15 years old (and most less than 10 years old)) and older, more established suburbs where the drainage systems were not specifically designed using WSUDs (e.g. sites 309–324).
Water and sediment sampling Water and sediment samples were collected from each of the 24 sampling sites on a single occasion in April 2010. Water temperature, pH and electrical conductivity (EC) were measured in situ at the time of sampling using a field meter (TPS 90FLs+, TPS Australia, Springwood, Queensland). The total organic carbon (TOC) content of water samples was determined according to the American Public Health Association (APHA) method 5301B (APHA 2005). Sediment samples were analysed for organic carbon (OC) using the Walkley and Black method (Rayment and Lyons 2011; Method 6A1). Water samples were collected as a series of ‘grab’ or spot samples from the 24 urban, peri-urban and rural waters (Fig. 1). Samples (1 L) were collected in amber glass bottles (one bottle per subsequent test) and stored on ice at 1–4 °C in a portable ice box or portable fridge and then at 4 °C until processed (within 36 h of collection for bioassays and within 2 weeks for pesticide residues testing). Surficial sediments were sampled with a dip net and transferred to a 63-μm sieve and sieved wet, on site, into a large plastic bucket prior to their subsequent use in Chironomid-based toxicity tests (standardising collected sediments to a fine grain size (e.g.
