Embed Size (px)
Citation preview
Science of the Total Environment 408 (2010) 873–883
Contents lists available at ScienceDirect
Science of the Total Environment
j ourna l homepage: www.e lsev ie r.com/ locate /sc i totenv
Polybrominated diphenyl ethers, perfluorinated alkylated substances, and metalsin tile drainage and groundwater following applications of municipal biosolidsto agricultural fields
N. Gottschall a,1, E. Topp b, M. Edwards a, P. Russell a, M. Payne c, S. Kleywegt d, W. Curnoe e, D.R. Lapen a,⁎a Agriculture and Agri-Food Canada, Ottawa, ON, Canada K1A 0C6b Agriculture and Agri-Food Canada, London, ON, Canada N5V 4T3c Ontario Ministry of Agriculture, Food and Rural Affairs, Stratford, ON, Canada N5A 5T8d Ontario Ministry of the Environment, Toronto, ON, Canada M4V 1M2e University of Guelph-Kemptville, Kemptville, ON, Canada K0G 1G0
⁎ Corresponding author. Tel.: +1 613 759 1537; fax:E-mail addresses: [email protected] (N. Got
(D.R. Lapen).1 Tel.: +1 613 759 1668.
0048-9697/$ – see front matter. Crown Copyright © 20doi:10.1016/j.scitotenv.2009.10.063
a b s t r a c t
a r t i c l e i n f oArticle history:Received 9 June 2009Received in revised form 20 October 2009Accepted 23 October 2009
Keywords:BiosolidsLand applicationTile drainageGroundwaterMetalsPolybrominated diphenyl ethersPerfluorinated alkylated substancesPreferential flow
Polybrominated diphenyl ethers (PBDEs), perfluorinated alkylated substances (PFAS), and metals weremonitored in tile drainage and groundwater following liquid (LMB) and dewatered municipal biosolid (DMB)applications to silty-clay loamagriculturalfieldplots. LMBwasapplied (93,500 Lha−1) in late fall 2005via surfacespreading on un-tilled soil (SSLMB), and a one-pass aerator-based pre-tillage prior to surface spreading (AerWaySSD) (A). The DMB was applied (8 Mgdwha−1) in early summer 2006 on the same plots by injecting DMBbeneath the soil surface (DI), and surface spreading on un-tilled soil (SSDMB). Key PBDE congeners (BDE-47, -99,-100, -153, -154, -183, -209) comprising 97% of total PBDE in LMB, had maximum tile effluent concentrationsranging from 6 to 320 ng L−1 during application-induced tile flow. SSLMB application-induced tile mass loads forthese PBDE congeners were significantly higher than those for control (C) plots (no LMB) (p<0.05), but not Aplots (p>0.05). PBDEmass loss via tile (0–2 h post-application) as a percent ofmass appliedwas ~0.04–0.1% and~0.8–1.7% for A and SSLMB, respectively. Total PBDE loading to soil via LMB and DMB application was 0.0018 and0.02 kg total PBDEha−1yr−1, respectively. Total PBDE concentration in soil (0–0.2 m)after both applicationswas115 ng g−1dw, (sampled 599days and 340 days post LMB and DMB applications respectively). Of all the PFAScompounds, only PFOS (max concentration=17 ng L−1) and PFOA (12 ng L−1) were found above detectablelimits in tile drainage from the application plots. Mass loads of metals in tile for the LMB application-induced tilehydrograph event, and post-application concentrations of metals in groundwater, showed significant (p<0.05)land application treatment effects (SSLMB>A>C for tile and SSLMB and A>C for groundwater for most results).Following DMB application, no significant differences in metal mass loads in tile were found between SSDMB andDI treatments (PBDE/PFAS were not measured). But for manymetals (Cu, Se, Cd, Mo, Hg and Pb) both SSDMB andDI loads were significantly higher than those from C, but only during <100days post DMB application. Clearly,pre-tilling the soil (e.g., A) prior to surface application of LMB will reduce application-based PBDE and metalcontamination to tile drainage and shallow groundwater. Directly injecting DMB in soil does not significantlyincrease metal loading to tile drains relative to SSDMB, thus, DI should be considered a DMB land applicationoption.
Crown Copyright © 2009 Published by Elsevier B.V. All rights reserved.
1. Introduction
Polybrominated diphenyl ethers (PBDEs) (flame retardants),perfluorinated alkylated substances (PFAS) (fluorosurfactants), andmetals can be environmental and human health concerns (Dave andXiu, 1991; Wolfe et al., 1998; Siddiqi et al., 2003; Thibodeaux et al.,2003; Calafat et al., 2006; Fjällborg et al., 2006; Costa and Giordano,
+1 613 759 1701.tschall), [email protected]
09 Published by Elsevier B.V. All rig
2007). Municipal sewage sludge (municipal biosolids), commonlyapplied to land as a soil amendment, can contain PBDEs and PFAS inthe low ppm range (Öberg et al., 2002; North, 2004; Higgins et al.,2005; Loganathan et al., 2007; Eljarrat et al., 2008; Heidler andHalden, 2008). Metals, which are commonly detected in biosolids(Zufiaurre et al., 1998; Ščančar et al., 2000), have led to regulationslimiting their application to land (MOE and OMAFRA, 1996).
Long term accumulation of biosolid-derived metals in field soilsand groundwater has been well documented (Higgins, 1984;Matthews and Davis, 1984; McGrath et al., 1995; McBride et al.,1997; Moolenaar and Beltrami, 1998; Keller et al., 2002; Speir et al.,2003; Rattan et al., 2005; Sukkariyah et al., 2005). For PBDEs and PFAS,
hts reserved.
874 N. Gottschall et al. / Science of the Total Environment 408 (2010) 873–883
concentrations and persistence in soils/sediments exposed either tobiosolids or waste water treatment plant (WWTP) effluent have alsobeen studied (e.g., Sellström et al., 2005; Samara et al., 2006; Eljarratet al., 2008). However, field data on the effects of biosolid landapplicationmethod on the fate/transport of these contaminants to tiledrains (efficient subsurface conduits to surface waters) and ground-water are less well documented; yet such data are required toevaluate the significance of these pollution sources/transport path-ways, with those linked with WWTP effluent (e.g. Karvelas et al.,2003; North, 2004; Heidler and Halden, 2008), landfill leachates (e.g.Kjeldsen et al., 2002; Osako et al., 2004), and/or atmosphericdeposition (e.g. Venier and Hites, 2008), for example.
The properties of PBDEs, PFAS, and metals allow for their accumu-lation in the environment due to their persistence and potential forsorption onto soil and organic matter (Rahman et al., 2001; Litz, 2002;Merrington et al., 2003; Hassanin et al., 2004; Sukkariyah et al., 2005;Zou et al., 2007; Heidler and Halden, 2008). Nevertheless, contaminantsin biosolids applied to land have the potential to rapidly bypass soilsorption sites and contaminate groundwater/tile drainage systems viapreferential flow through macropores (Kay et al., 2004; Jarvis, 2007;Lapen et al., 2008a; Larsboet al., 2009).However, contaminant transportprocesses associated with a land application of liquid municipalbiosolids (LMB) (typically <3% total solids by wet weight; Schut,2005)will differ from those associatedwith anapplication of dewateredbiosolids (DMB) (typically>20% solids bywetweight; Schut, 2005). ForDMB, natural or artificial recharge is typically needed to promotesignificant contaminant transport via surface runoff (Sabourin et al.,2009) and/or leaching in the soil profile (Gottschall et al., 2009; Edwardset al., 2009); whereas for LMB, contaminant transport to depth can berapid and significant at the time of land application (Lapen et al., 2008a,b; Larsbo et al., 2009). The method of biosolid application, and/orproduct placement in soil, can also strongly affect contaminanttransport and degradation/transformation of its constituents (Terry,1979; Gottschall et al., 2009). For instance, aerate tilling the soilimmediately prior to surface application of LMB, relative to not tillingthe soil prior to surface application, can greatly reduce application-induced loads of total Kjeldahl N, NH4–N, PO4–P and E. coli.(>85%reductions) to tile drains, and therefore, stream systems (Lapen et al.,2008a). Lapen et al. (2008b) also found for this same field study, thatsuch pre-tillage reduced, on average, the mass loss of nine pharmaceu-tical/personal care products to tile drains over a 46 day period by ~49%.For DMB, product placement within or on the soil has been shown toaffect nutrient transformations, persistence of bacteria, and transport ofcontaminants to tile drains/groundwater (Sierra et al., 2001; He et al.,2003; Zaleski et al., 2005; Gottschall et al., 2009).
This study investigated PBDE, PFAS andmetal contamination in tilewater, and metal contamination in groundwater resulting from a fallapplication of LMB followed by a summer application of DMB. For LMBtwo application approaches were examined: i) surface spreadingfollowed by shallow (~0.1 m) incorporation (SSLMB), and ii) aeratetilling the soil immediately prior to surface spreading using a one-passAerWay SSD system (A). The A approach can increase LMB storage inthe tillage layer, disrupt continuous macropore networks, and createlarge laterally transmissive soil pockets (Turpin et al., 2007a,b), rela-tive to the SSLMB approach. Thus, it was expected that the A methodwould reduce more significantly contaminant loads to tile drainage/shallow groundwater; especially immediately following land appli-cation, as found in companion studies (Lapen et al., 2008a,b). ForDMB, the two land application approaches consisted of: i) surfacespreading followed by shallow incorporation (SSDMB) (~0.1 m),and ii) directly injecting DMB below the surface (DI) (~0.1 mdepth). Direct injection of DMB could result in a slower release ofcontaminants over time, compared to surface application, due toreduced organic matter decomposition for direct injected material asa result of its more cohesive form (lower surface area) and fullplacement beneath the soil surface (Edwards et al., 2009; Gottschall
et al., 2009). Hence, residual PBDE concentrations in surface soilswhere both LMB and DMB were applied, were also characterized inthis study.
2. Materials and methods
2.1. Site description and history
Water quality monitoring was conducted on experimental plots ineastern Ontario Canada (latitude 45′03″ N, longitude 75′21″ W)during October to December 2005 and July to December 2006. Soilswere classified as North Gower clay loam soils (Orthic Humic Gleysol(Canadian)). Key soil properties include (0–0.20 m depth): 28% clay,52% silt, 20% sand, 20 gkg−1 organic carbon, 1.3 Mg m−3 bulk densityand CEC ~20 me 100 g−1. Crops have consisted of corn (Zea mays L.),soybean (Glycine max L.) and wheat (Triticum spp.) under bothconventional tillage and no-tillage; the former practice being springcultivation and fall mouldboard plowing (maximum plow depth ofapproximately 0.2 m).
2.2. Experimental field setup
Eight experimental field plots (100 m in length×15 m wide),centered over tile lines (100 mm diam. plastic tiles for six plots and100 mm diam. clay tiles for two plots) were utilized in this study(Lapen et al., 2008a). Tiles were approximately 0.8 m below thesurface at their deepest point along the plots. The six plastic-tiledplots were designated for biosolid land application treatments, whilethe two ceramic-tiled plots, being hydrologically isolated from theothers, were used as controls (no biosolid application). There werethree plot replications for each land application treatment alternatingspatially as per Lapen et al. (2008a) and Gottschall et al. (2009); T1,T3, and T5 were replicates for the A and DI treatments, whereas T2, T4and T6 were replicates for SSLMB and SSDMB. For the control plots (C1and C2), tillage and planting operations were conducted to mimicthose performed on the respective application treatment plots usingequipment prior to contact with biosolids used in this study.
2.3. Land application of biosolids
2.3.1. Liquid municipal biosolid (LMB) application methodsThe LMB application occurred in 2005 on October 21 (day of year
(DOY) 294). The anaerobically digested LMB (Tables 1 and 2) wasobtained from a municipal wastewater treatment plant that handlesthe waste from about 20,000 people. The application approaches forthe LMB consisted of an AerWay SSD (subsurface deposition) system(A)(Holland Equipment Limited, Norwich, ON), and surface spreadingfollowed by shallow (~0.1 m) incorporation (within 24h) using aKongskilde vibro-flex cultivator (SSLMB) (Kongskilde, Exeter, ON)(Fig. 1). Briefly, the AerWay SSD system surface-applies slurry close tothe ground immediately following rolling tines that effect aerator-type tillage of the soil (Lapen et al., 2008a; Turpin et al., 2007a). ForSSLMB, surface application was conducted using the AerWay SSDimplement sans tillage action (tines were lifted above ground surfaceduring application passes). Both treatments used an LMB applicationrate of 93,500 Lha−1 (~1 Mgdwha−1; under the commercial ratelimit per 5 years of 8 Mgdwha−1 (MOE and OMAFRA, 1996)). Therewere two application passes per plot, each pass being approximately4 m wide. For control plots, ‘dry’ passes were made prior to contactwith biosolids in a similar manner. Further details can be found inLapen et al. (2008a).
2.3.2. Dewatered municipal biosolid (DMB) application methodsAfter land application of the LMB the field was left uncultivated
until the July 2006 (DOY 188 (July 7)) land application of DMB. TheDMB used in this study was anaerobically digested and dewatered
Table 1Key properties for liquid (LMB) and dewatered (DMB) municipal biosolids used in thestudy.
Parameter LMB (2005) DMB (2006)
Specific conductivity (dsm−1) 5.5 ndpH 7.5 7.8Total solids (% dw) 1.2 30a
Total volatile solids (% dw) 48 ndTotal solids, ash (% dw) 41 ndTotal Kjeldahl N (% dw) 11 4a
(NO3+NO2)–N (% dw) 0.07 0.002a
(NH3+NH4)–N (% dw) 5.6 2a
Total P (% dw) 2.9 3a
PO4–P (% dw) 0.05 ndE. coli (cfu g−1dw) 1500000b 30000C. perfringens (cfu g−1dw) 1 600 000b 7800000Total PBDE (ng g−1dw) 1500 2500BDE-47 280 430BDE-99 360 370BDE-100 68 91BDE-153 35 53BDE-154 32 36BDE-183 7 20BDE-209 700 1500Total PFAS (ng g−1dw) ~80–600c ~80–600c
Of all PBDEs measured, only dominant congener concentrations are given.PBDEs measured: BDE-17, -28, -49, -71, -47, -66, -77, -100, -119, -99, -85, -126, -154,-153, -138, -183, -209.PFAS measured: PFHxS, PFOS, PFDS, PFOSA, PFHpA, PFOA, PFNA, PFDA, PFUA, PFDoA.nd=not determined.
a Average of two samples.b Average of three samples taken when nurse tanker full, half full, and near empty.c Range of values in biosolids from wastewater treatment plants in Ontario
(Kleywegt, 2008, personal communication).
875N. Gottschall et al. / Science of the Total Environment 408 (2010) 873–883
with a high speed centrifuge (Tables 1 and 2). DMB application rateswere 8Mgdwha−1 (max. commercial rate per 5 years (MOE andOMAFRA, 1996)). The application approaches consisted of: i) surfacespreading using a Hydra-spread II spreader (Hagedorn, Paisley, ON)with subsequent (within a few hours post-application) DMBincorporation (~0.1 m depth) using a Kongskilde vibro-shankcultivator (SSDMB) (Kongskilde, Exeter, ON), and ii) Terratec Environ-mental Ltd.'s dewatered biosolids direct injection system (DI)(~0.05 m diam. DMB extruded from injector to ~0.1 m depth in soil)(Fig. 1). Both treatments consisted of two implement passes of ~4 meach over the plots, totaling ~8 m×100 m centered over the tile. Twodays after application, the plots were rolled over once with a roller/packer to prepare the soil surface for direct drilling of short seasonsoybeans. There was no tillage following the DMB application andsoybeans, again, were planted in spring 2007 using “no-till”equipment. Further details are given in Gottschall et al. (2009).
Table 2Metal concentrations in LMB and DMB used in the study (µg g−1dw), and provinciallimits on metals in DMB (µg g−1dw) (MOE and OMAFRA (1996).
Metal LMB (2005) DMB (2006) Metal content limits
Aluminum (Al) 9700 12,000a naArsenic (As) 5.9 2.60a 170Cadmium (Cd) <4.2 2.7a 34Chromium (Cr) 96 230a 2800Cobalt (Co) 8.4 5a 340Copper (Cu) 490 1500a 1700Iron (Fe) 85,000 56,000a naLead (Pb) 38 87a 1100Molybdenum (Mo) 18 15a 94Mercury (Hg) <4.2 <1.7a 11Nickel (Ni) 21 33a 420Selenium (Se) 2.9 4.3a 34Zinc (Zn) 540 1100a 4200
na=not applicable.a Average of two samples.
2.4. Field data collection and sample analyses
2.4.1. Field data collectionTable 3 summarizes PBDE, PFAS, and metal study periods and
sampling matrix. Weather data collected at the site, and tile effluentmonitoring details are given in Lapen et al. (2008a) and Gottschallet al. (2009). Briefly, for PBDE/PFAS and metals on LMB applicationday (2005, DOY 294), manual tile effluent ‘grab’ samples were takenfrom tile every 15 min from 15 min prior to application up to 2 h post-application, to capture at least the rising limb of the LMB application-induced tile hydrograph event; additional application-induced hydro-graph samples for metals were taken after the primary 2 hmonitoringperiod using ISCO 6712 (ISCO Inc., Lincoln, NE) automated samplers,up to 24 h post-application. These grab samples were collected in 1 Lpolypropylene bottles to reduce breakage potential in field. Tile waterfor metals analyses was sampled during tile discharge events, pre-and post-application in 2005 and 2006 using ISCO 6712 samplers, setto trigger when a specified rainfall depth was achieved. These samplebottles (1 L) were Teflon lined. One tile water sample was selected perplot per sampling event for metals. Due to cost and sample collectionlimitations, these ISCO water samples were selected on the basis ofsample turbidity/hydrograph time; based on previously findingrelatively higher pollutant levels during the turbid rising limb of tilehydrograph events. Metal samples collected from each plot duringrainfall-triggered events were temporally collocated with each otherwith respect to sampling time on respective tile hydrographs. Tile‘grab’ samples were also taken periodically over the 2005 and 2006study seasons between rain event collections during, typically,recession limbs of the tile hydrograph.
Shallow groundwater was monitored via piezometers (intakelength=0.15 m) at depths of 1.2 and 2.0 m below the surface.Groundwater samples were taken weekly during 2005, pre- and post-application, and biweekly during the 2006 study period. Piezometerswere first pumped, then allowed to equilibrate, at which time sampleswere extracted using a peristaltic pump fit to Teflon lined extractiontubing set for each piezometer. All field water samples wereimmediately transported from the field in iced coolers to Agricultureand Agri-Food Canada laboratories in Ottawa, ON and split intoanalyte-appropriate bottles (as per the external laboratory instruc-tions) for express shipment to the Laboratory Services Branch of theOntario Ministry of the Environment (MOE), in Etobicoke, Ontario foranalyses. For bulk metal analyses water samples were preserved withnitric acid, whereas for mercury samples, nitric acid and potassiumdichromate were added, prior to shipment, in accordance with MOEprotocols.
Soil was also sampled in the DI plots (0–0.20 m depth; 0.02 mdiam.; five cores per plot, pooled by treatment) in 2007 (DOY 163)and analysed for PBDEs only. Soil samples were collected from thesesame plot locations in July 2005 (pre LMB application) in a similarmanner. These samples were frozen at −20 °C until PBDE analysiscould be completed.
2.4.2. PBDE, PFAS, and metals analysesPBDEs and PFAS water samples were analysed at MOE laboratories
according to methods BDE3430 and PFAS3457. Briefly, for tile waterPBDE analyses, unfiltered aqueous samples were extracted using aC18 Solid Phase adsorption disk and then toluene/ethanol was used toextract the PBDEs from the disk and particulate matter. The extractswere cleaned using a single stage silica (acid/base/AgNO3) cleanup,and then analysed using gas chromatography–high resolution massspectrometry (GC-HRMS). For PFAS, unfiltered aqueous samples weredirectly injected and analysed by C18 reversed phase liquidchromatography-negative electrospray tandem mass spectrometry.Analysis of PBDEs in DMB and soil–biosolid composites, wasconducted by the ALS Laboratory Group (Burlington, ON, Canada)following a modified USEPA method 1614. More specifically, samples
Fig. 1. Stylized schematics of the different biosolid application methods used in this study (not to scale). Arrow showing direction of travel. a) LMB application using AerWay SSD(subsurface deposition) with tillage implement engaged. Profile view of one tine only. b) LMB surface spreading using AerWay SSD with tines lifted above ground surface (no pre-tillage action). Profile view of one tine only. *Subsequent LMB incorporation not shown. c) Terratec Environmental Ltd's DMB direct injection system. Profile view of one tine shown.d) DMB surface spreading of dewatered municipal biosolids using a Hydra-spread spreader. Profile view of entire dry box container and beaters shown. *Subsequent DMBincorporation not shown.
876 N. Gottschall et al. / Science of the Total Environment 408 (2010) 873–883
were extracted by Soxhlet/Dean-Stark using toluene as the extractingsolvent. Compounds of interest were further isolated using acidifiedsilica and alumina column chromatography prior to analysis by GC-HRMS.
Table 3Overview of samples analysed for the 2005 LMB and the 2006 DMB land applicationstudy periods.
Tile water Groundwater
Metals PBDE PFAS Metals
LMB (2005)Pre-application X X X Xa
Entire LMB application-induced tilehydrograph (24 h post app.)
X
2 h post LMB application X X>24 h post-application (DOY 295-340) X X
DMB (2006)Pre-application X<100 day post-application (DOY 188-288) X>100 day post-application (DOY 289-335) X
a Based on controls plots.
Analysis (as totals) for the regulated metals arsenic (As), cadmium(Cd), chromium (Cr), cobalt (Co), copper (Cu), lead (Pb), nickel (Ni),molybdenum (Mo), selenium (Se), and zinc (Zn), as well as non-regulated aluminum (Al) and iron (Fe), was conducted according toMOE methods E3094 and E3302. Both Al and Fe were monitoredbecause of their common use in the wastewater treatment process toprecipitate out phosphorus (Yeoman et al., 1988). Mercury (Hg)analyses were conducted according to MOE method E3060.
2.5. Data treatment and statistical approach
Tile contaminant mass loads in mass per 15 min of flow (calculatedusing tile flow (m3 15 min−1 for consistency due to time frame ofdischarge measurements in the field) and concentration at that time)were required to compare application treatment effects on tile drainagewater quality. Tile water concentrations of PBDEs and PFAS weremeasured only up to 2hpost-application (temporally discretemass loaddata pooled on that basis), as maximum LMB application-induced tileflow occurred within this time period. Metal tile mass load data for the2005 LMB study were compared by treatment for time periodsrepresenting: i) the application-induced tile hydrograph event (up to
877N. Gottschall et al. / Science of the Total Environment 408 (2010) 873–883
24 h post-application); and ii) the remainder of the study period(45days). The 2006 DMB study tile metals data were partitioned on thebasis of an ‘early’ (<100days post-application) and ‘late’ (>100dayspost-application) study period. These temporal data groupings weredefined on the basis of a mid study season dry spell that resulted in notile flow at the site (Fig. 2).We expected based on our experience at thesite to find differences in contaminant transport processes in the soilbefore and after the dry spell as a result of, for instance, soil desiccationcracking (Turpin et al., 2007a,b). Metal concentrations in groundwaterwere pooled by treatment and piezometer depth using all data collectedpost LMB application (over a 46 dayperiod). All statistical analyseswereconducted with SYSTAT v. 10 (SPSS Inc., Chicago, Illinois, USA), usingANOVAs and conservative Bonferroni post-hoc tests. All concentration/tile mass load data were log-transformed prior to statistical analyses tobetter meet assumptions of normality and homogeneity of variance. Analpha level of 0.05 was used as a significance threshold for all tests.
3. Results and discussion
3.1. General background information
Key properties of the LMB and DMB are presented in Tables 1 and2, including initial concentrations of each analyte examined. For the
Fig. 2. Precipitation, air temperature, and tile flow (from T1) f
LMB study (DOY 294–340, 2005), total rain precipitation was 124 mm(Fig. 2). Tiles were flowing at, and prior to, land application. Flow ofland applied LMB to tile occurred between 3 and 39min afterapplication with peak application-induced discharge occurring within~120 min post-application for both SSLMB and A plots. The LMBapplication produced tile discharge hydrographs with total flowvolumes of (mean±SE): 1.06±0.22 m3 for SSLMB and 0.24±0.1 m3
for A treatments (Lapen et al., 2008a). For the DMB study (DOY 188–335, 2006), total precipitation that occurred was 442 mm. There wasno tile flow that occurred between roughly DOY 220 (August 8) andDOY 289 (October 16) (Fig. 2). Due to the lower water content of theDMB, there was no application-induced tile discharge. However, thefirst rainfall post-application stimulated tile discharge on DOY 196.
3.2. PBDE and PFAS in tile drainage (LMB application) and soil(post DMB application)
The PBDE congeners in the raw LMB (BDE-47, -99, -100, -153, -154,-183 and -209) comprised 18, 24, 4.5, 2, 2, 0.5, and 46% of the totalPBDE concentration of 1500 ng g−1dw. For DMB, these respectivepercentages were 17, 15, 3.6, 2, 1, 0.8, and 58%, relative to total PBDEconcentration of 2500 ng g−1dw. These congeners represent themain components of commercial PBDE mixtures: Penta-BDE (47, 99,
or the: a) LMB (2005) and b) DMB (2006) study periods.
878 N. Gottschall et al. / Science of the Total Environment 408 (2010) 873–883
100, 153, 154), Octa-BDE (183) and Deca-BDE (209) (as summarizedby Zhang et al., 2008), and thus, were selected for applicationtreatment comparisons. The pre-application (background) tile watersamples for BDE-47, -99, -100, -153, -154, -183 and -209 in all plotsranged between BDL (below detectable limits) to 24, 28, 6, 2, 2, BDLand 35 ng L−1, respectively. The highest observed post-app tileeffluent concentrations of these respective PBDEs were: 270, 320, 48,28, 23, 6 and 290 ng L−1. All of themaximum concentrations occurredon plot T2 (SSLMB) within 60 min after LMB application (Fig. 3).
Comparisons of PBDE tile mass loads by ANOVA for these PBDEcongeners showed significant land application treatment effects(p<0.05) (Table 4). Post-hoc (Bonferroni) tests indicated that all ofthese congeners had mass loads for SSLMB that were significantlyhigher than those for control plots (p<0.05); however, SSLMB plotsdid not have significantly higher observed tile mass loads than A plots(BDE-47, -99, -100, p<0.1; BDE-153, -154, -183, -209, p>0.1).Nevertheless, mean discrete tile mass loads were indeed higher forall congeners for SSLMB than they were for A (Table 4). In fact, the totalmass of selected PBDEs lost via tile and allowed to enter receivingsurface waters during 2h post-application via A treatment, as apercentage of that applied, was (mean±standard error) 0.12±0.06,0.11±0.06, 0.11±0.06, 0.10±0.06, 0.09±0.05, 0.10±0.06 and0.04±0.03% for BDE-47, -99, -100, -153, -154, -183 and -209, re-spectively. However, for SSLMB, these respective percentages were1.73±0.72, 1.53±0.68, 1.21±0.55, 1.38±0.59, 1.23±0.54, 1.44±0.59 and 0.75±0.30%. PBDEs have been shown to strongly sorb toorganic material in soil and biosolids (Litz, 2002; Hassanin et al., 2004;Song et al., 2006; Zou et al., 2007). However, topsoil and subsoil organiccarbon at the site is ~20 and 4–9 g kg−1 respectively, and the soil ismacroporous with respect to earthworm burrows, and desiccationcracks when dry (Turpin et al., 2007b; Ouellet et al., 2008). Macroporesand the modest degrees of organic matter in the soil can facilitate massloss of LMB-derived contaminants to depth in the soil profile viareducing sorption potential (Camobreco et al., 1996; Moradi et al., 2005).Nevertheless, the observed tile mass loads among treatments areattributed mostly to the impact of LMB interaction with macropores,as found in companion studies regarding other compounds andmicroorganisms during this same land application experiment (Lapenet al., 2008a,b; Larsbo et al., 2009). Sorption potential will be reduced,particularly in the macropores where flow velocities will be larger, andtherefore LMB interaction with soil sorption sites reduced (Jarvis, 2007;Larsbo et al., 2009). LMB, applied via SSLMB, was able to move rapidly totile drains and shallow groundwater via large, undisrupted, continuousmacropores, whereas a greater volume of LMB was retained in the moreporous surface soils for the A treatment due to tillage factors describedabove (Turpin et al., 2007a,b). Although PBDE results showed signifi-cantly higher tile mass loads from SSLMB plots compared to control plots,tile effluent concentrations of some of the most prevalent PBDEs (47 and99) were several orders of magnitude below those necessary to produceacute toxic effects in aquatic invertebrates and algae, which are typicallyin the mg L−1 rather than ng L−1 range (Breitholtz and Wollenberger,2003; Evandri et al., 2003). Thus, pulse exposure to toxic concentrationswould seemingly be unlikely at land application, as the concentrations inthree of six tiles were already decreasing towards the end of the 2h post-application monitoring period; albeit, our data were too temporallylimited to speculate on longer term implications. Moreover, the range oftile drainage concentrations of the PBDE congeners examined (excludingthe highest concentrations from T2(SSLMB), which were an order ofmagnitude higher) were slightly higher, but of a similar scale as somereported WWTP effluent concentrations (ranging from ~1 to 30 ng L−1,with 47 and 99 on the higher end) (e.g. North, 2004; Song et al., 2006;Zhang et al., 2008).
The average total PBDE mass loss to tile drains within 2 h post-application, was roughly an order of magnitude lower for A (~0.04–0.1% of total applied) than SSLMB (~0.8–1.7% of total applied) plots.Notwithstanding evidence for long term PBDE persistence in soil–
water environments (Eljarrat et al., 2008), management practices thatdelay the entry of PBDEs to off-field hydrological transport pathwayssuch as tile drainage systems as long as possible, such as the Atreatment, will increase the probability of in situ degradation–transformation processes (Wania and Dugani, 2003; Vonderheideet al., 2006) that could ultimately reduce PBDE contamination ofecologically more sensitive surface water environments (those thatreceive tile drain effluent for example); first-order degradation half-lives for most of the PBDEs studied have been estimated at 3600 h forsoil and water, and 14,400 h for sediment (Wania and Dugani, 2003).In this study, soil samples (0–0.2 m depth) taken on DOY 163 in 2007(599 and 340 days post LMB and DMB application, respectively) in theDI plots directly where DMB was injected in soil and where LMB wasapplied via A, indicated that BDE-47, -99, -100, -153, -154, -183 and-209 concentrations were 22, 26, 5, 3.3, 2.2, 2.3 and 52 ng g−1dw,respectively. Soil PBDE concentrations pre LMB application (DOY 1862005) at the same plot locations were 0.019±0.0016 for BDE-47,0.015±0.0001 for BDE-99, BDL for BDE-100, -153, -154 and -183, and0.151±0.005 ng g−1dw for BDE-209. Residual PBDE concentrationsobserved in the land applied soils represent some combination ofleaching and degradation losses, but do characterize at least aminimum residual concentration that could be expected from suchbiosolid land applications. However, extraction efficiencies have beenshown to decrease over time in soil so these residual soil concentra-tions may be underestimated (Welsh et al., 2009). Other studies haveshown post-application PBDE concentrations to be generally lowerthan those presented here, especially for BDE-47 and -99, althoughresults were highly variable, representing different soil types andapplication regimes (Sellström et al., 2005; Eljarrat et al., 2008). Thehigher values from the current study may also be due to thedifferences in the types and amounts of PBDEs used in North America,versus Europe, where the comparison studies were conducted (Haleet al., 2003; Song et al., 2006); Eljarrat et al. (2008) show sourcesewage sludge concentrations with considerably lower proportions ofpenta-BDE congeners than the LMB and DMB used here.
Tile water concentrations of all 10 PFAS measured (Table 1) wereBDL with the exception of PFOS (perfluorooctane sulfonate; MDL(method detection limit) 10 ng L−1) and PFOA (perfluorooctanoic acid;MDL 10 ng L−1). Maximum observed tile drainage concentrations of17 ng L−1 for PFOS were found only on T2(SSLMB) and 12 ng L−1 forPFOA on T1(A) treatment plots, not on controls. Low PFAS concentra-tions could have been related to modest concentrations of thesecontaminants in the biosolids (~80–600 ng g−1dw; Kleywegt, 2008,personal communication) (Table 1). Nevertheless, toxicological risksfrom PFAS in the tile drainage were likely minimal as acute and chronictoxicity studies (for PFOA and PFOS) found effect concentrations severalorders of magnitude above the levels observed in tile drainage duringthe LMB-induced hydrograph event, in themg rather than ng L−1 range(Sanderson et al., 2003; Ji et al., 2008).
Although the tile drainage data presented here is limited to thecritical application-induced tile discharge events, this study is one of thefirst to document both PBDE and PFAS direct contributions to agricul-tural tile drainage systems via land applications of municipal biosolids.These tile drainage systems are efficient pathways for agriculturaldrainage to enter adjacent surfacewater. To put broader perspective onour limited findings, total PBDE inputs to the experimental fieldsamounted to ~1.7×105 ng m−2yr−1 from the LMB application, and2×106ng m−2 yr−1 for the DMB application, as contrasted withestimated atmospheric inputs (<1000 to 89,100 ngm−2yr−1; e.g., terSchure and Larsson, 2002; Moon et al., 2007). However, compared toWWTP effluent concentrations noted above, and the considerablevolumes of effluent released from these facilities routinely (e.g., Sinclairand Kannan, 2006), PBDE inputs derived from specific land applicationsof biosolids may be less important in terms of overall environmentalexposure; especially if prudent land application measures are under-taken to reduce contaminantmass loss at application.We estimated for
Fig. 3. Concentrations of selected PBDE congeners in tile drainage (BDE-47, -99, -100, -153, -154, -183, and -209) for -15 min (background) up to 2 h post LMB-application; LMBapplication occurred at 0 min.
879N. Gottschall et al. / Science of the Total Environment 408 (2010) 873–883
Table 4Means±SE of log-transformed mass loads (ng15 min−1) of PBDEs in tile drainage for samples collected up to 2 h post LMB application.
Treat. BDE-47 BDE-99 BDE-100 BDE-153 BDE-154 BDE-183 BDE-209
A 2.7±0.09 2.6±0.12 2.0±0.11 1.5±0.14 1.5±0.12 0.76±0.13 2.7±0.10SSLMB 3.2±0.21⁎,C 3.2±0.22⁎,C 2.4±0.19⁎,C 2.0±0.24⁎,C 2.0±0.23⁎,C 1.3±0.26⁎,C 3.1±0.24⁎,C
C 2.2±0.12 2.1±0.11 1.5±0.61 0.83±0.12 0.79±0.12 0.32±0.13 2.1±0.08
n=24 samples per congener.A=AerWay, SSLMB=surface spread, C=controls.⁎ Significantly higher (p<0.05) than treatment indicated.
880 N. Gottschall et al. / Science of the Total Environment 408 (2010) 873–883
the LMB and DMB application that there was loading to the soil(application) of 0.0018 and 0.02 kg total PBDE ha−1yr−1; in contrast toestimated total PBDE WWTP effluent loads of ~0.9 kg total PBDE yr−1
(Palo Alto WWTP, California) (e.g., North, 2004). For PFOS, concentra-tions in tile water observed here (17 ng L−1) were small, but roughlyequal to those documented for the effluent of sixWWTPs (3–68 ng L−1)(Sinclair and Kannan, 2006). For PFOA, however, our observedconcentrations (12 ng L−1) were smaller than those documented forthose same six WWTPs (58–1050 ng L−1).
3.3. Metals in tile drainage and groundwater following LMB application
Pre-application total metal concentrations in all tiles were BDL,with the exception of Al, Fe, Cu, Se and Zn (Table 5). Tile mass loads of
Table 5Range of total metals concentrations in tile drainage for LMB (2005) and DMB (2006) stud
Metal LMB (2005)
≤24h post-app >24h post-app
Al(mg L−1)
A 0.1–3.78 0.06–17.4SSLMB 0.14–9.48 0.06–32.4C 0.175–0.625 0.2–41
As(mg L−1)
A BDLa−0.0025 BDL-0.002SSLMB BDL-0.006 BDL-0.003C BDL BDL-0.0035
Cd(mg L−1)
A BDL BDLSSLMB BDL BDLC BDL BDL
Cr(mg L−1)
A BDL-0.029 BDL-0.09SSLMB BDL-0.075 BDL-6.39C BDL BDL-0.023
Co(mg L−1)
A BDL BDL-0.007SSLMB BDL-0.01 BDL-0.015C BDL BDL-0.016
Cu(mg L−1)
A 0.014-0.108 BDL-0.059SSLMB 0.019-0.283 BDL-0.072C 0.016-0.032 0.01-0.063
Fe(mg L−1)
A BDL-20.5 BDL-21.3SSLMB 0.03-41.3 BDL-39.8C 0.08-0.66 0.058-51.7
Pb(mg L−1)
A BDL BDLSSLMB BDL-0.095 BDLC BDL BDL
Hg(µg L−1)
A BDL-0.04 BDL-0.03SSLMB BDL-0.12 BDL-0.07C BDL BDL-0.05
Mo(mg L−1)
A BDL BDLSSLMB BDL-0.025 BDLC BDL BDL
Ni(mg L−1)
A BDL BDL-0.344SSLMB BDL-0.038 BDL-0.208C BDL BDL-0.1
Se(mg L−1)
A BDL-0.002 BDL-0.0015SSLMB S 0.001-0.0045 BDL-0.002C BDL BDL
Zn(mg L−1)
A 0.007-0.143 BDL-0.088SSLMB 0.008-0.377 BDL-0.177C 0.006-0.012 BDL-0.161
A=AerWay, SS=surface spread of LMB or DMB, DI=direct injection and C=control.a Below detection limit: 0.02 µg L−1 (Hg), 0.0005 mg L−1 (As, Se), 0.002 mg L−1 (Cu), 0.00
metals associated with the application-induced tile hydrographshowed significant (p<0.05) land application treatment effects forall metals examined (Table 6). Averagemass loss through tile drainage(as a percentage of total applied) within 2 h post LMB application viathe A treatment ranged between (mean±SE) <0.001±0.0002% (forHg) to 0.13±0.03% (for Pb); for the SSLMB treatment, averages rangedfrom <0.005±0.0007% (for Hg) to 0.46±0.17% (for Pb). For theremainder of the study period (>24 h post-application; DOY 295–340), there were only significant land application treatment effects forNi and Se tile mass loads; however, most observed tile mass loads ofmetals during this period were higher than those observed onapplication day. Metal concentrations in groundwater (post-applica-tion) were BDL for most samples, with the exception of Al and Fe;concentration ranges are given in Table 7. There were significant
y periods.
DMB (2006)
<100days post-app >100days post-app
DI 9.92–100 0.135–5.04SSDMB 4.96–42.3 0.065–3.27C 1.37–41.4 0.185–9.03DI BDL-0.0015 BDLSSDMB BDL-0.0025 BDLC BDL-0.006 BDLDI BDL BDLSSDMB BDL BDLC BDL BDLDI 0.018-0.189 BDL-0.01SSDMB 0.008-0.075 BDL-0.018C BDL-0.074 BDL-0.019DI BDL-0.048 BDL-0.002SSDMB BDL-0.019 BDLC BDL-0.017 BDLDI 0.018-0.19 0.004-0.022SSDMB 0.013-0.107 0.003-0.018C 0.006-0.044 0.003-0.017DI 10.5-111 0.028-5.62SSDMB 4.71-47.6 0.025-5.25C 1.29-44.8 0.092-10.5DI BDL BDLSSDMB BDL BDLC BDL BDLDI BDL-0.08 BDLSSDMB BDL-0.05 BDLC BDL-0.05 BDLDI BDL BDLSSDMB BDL BDLC BDL BDLDI 0.012-0.124 BDL-0.344SSDMB 0.006-0.052 BDL-0.528C 0.006-0.042 BDLDI BDL-0.001 BDL-0.0035SSDMB BDL-0.001 0.001-0.0035C BDL-0.001 BDLDI 0.045-0.469 0.006-0.024SSDMB 0.02-0.173 0.003-0.108C 0.006-0.164 0.002-0.036
5 mg L−1 (Cd, Co, Zn), 0.01 mg L−1 (Cr, Ni), 0.02 mg L−1 (Fe, Mo), and 0.05 mg L−1 (Pb).
Table6
Mea
n±
SEforlog-tran
sformed
massload
sof
metalsin
tile
draina
geforsamples
colle
cted
during
theLM
B(2
005)
andDMB(2
006)
stud
ype
riod
s.
LMB(2
005)
LMB(2
005)
DMB(2
006)
DMB(2
006)
(≤24
hpo
st-app
.)(>
24hpo
st-app
.)<10
0da
yspo
st-app
lication
>10
0da
yspo
st-app
lication
n=
81n=
150
n=
16n=
28
ASS
LMB
CA
SSLM
BC
DI
SSDMB
CDI
SSDMB
C
Al
0.97
±0.05
1.48
±0.08
⁎A,C
0.68
±0.10
1.84
±0.13
1.68
±0.12
1.50
±0.17
3.85
±0.24
⁎C
3.64
±0.16
2.91
±0.21
2.33
±0.19
2.10
±0.17
2.49
±0.93
Cd−
0.77
±0.03
−0.65
±0.05
*C−
0.96
±0.05
−0.53
±0.07
−0.59
±0.07
−0.80
±0.07
−0.10
±0.18
⁎C
−0.20
±0.13
⁎C
−1.17
±0.14
−0.71
±0.11
−0.79
±0.08
−0.66
±0.93
Co−
0.77
±0.03
−0.64
±0.06
⁎C
−0.96
±0.05
−0.53
±0.07
−0.59
±0.07
−0.80
±0.07
0.47
±0.25
⁎C
0.31
±0.16
−0.44
±0.14
−0.73
±0.13
−0.82
±0.08
−0.66
±0.93
Cr−
0.45
±0.03
−0.31
±0.06
⁎C
−0.66
±0.05
−0.14
±0.08
−0.15
±0.09
−0.44
±0.08
1.10
±0.24
⁎C
0.89
±0.17
0.14
±0.23
−0.15
±0.16
−0.31
±0.13
−0.19
±0.81
Cu0.10
±0.04
⁎C
0.37
±0.06
⁎A,C
−0.25
±0.05
0.22
±0.09
0.14
±0.08
−0.12
±0.09
1.14
±0.22
⁎C
1.03
±0.15
⁎C
0.08
±0.05
0.27
±0.11
0.13
±0.09
−0.12
±0.78
Fe0.79
±0.11
1.89
±0.13
⁎A,C
0.29
±0.16
1.82
±0.15
1.68
±0.13
1.52
±0.19
3.88
±0.24
⁎C
3.67
±0.17
2.94
±0.23
2.36
±0.21
2.02
±0.20
2.45
±1.01
Pb0.23
±0.03
0.40
±0.06
⁎A,C
0.04
±0.05
0.47
±0.07
0.41
±0.07
0.20
±0.07
1.60
±0.18
⁎C
1.49
±0.13
⁎C
0.53
±0.14
0.86
±0.17
0.90
±0.08
1.03
±0.93
Mo
−0.17
±0.03
−0.05
±0.05
⁎C
−0.36
±0.05
0.07
±0.07
0.01
±0.07
−0.20
±0.07
0.60
±0.18
⁎C
0.49
±0.13
⁎C
−0.47
±0.14
−0.01
±0.11
−0.10
±0.08
0.03
±0.93
Ni
−0.47
±0.03
−0.33
±0.06
⁎C
−0.66
±0.05
−0.17
±0.08
−0.07
±0.07
⁎C
−0.47
±0.08
0.97
±0.22
⁎C
0.81
±0.14
0.17
±0.11
0.40
±0.21
0.31
±0.18
0.11
±0.93
Zn−
0.31
±0.05
⁎C
0.14
±0.08
⁎A,C
−0.77
±0.06
0.01
±0.09
−0.13
±0.10
−0.34
±0.12
1.53
±0.24
⁎C
1.31
±0.15
0.50
±0.19
0.28
±0.14
0.06
±0.10
0.08
±0.83
Hg
−0.16
±0.03
0.29
±0.05
⁎A,C
−0.36
±0.05
0.08
±0.07
0.02
±0.08
−0.20
±0.07
1.10
±0.18
⁎C
0.95
±0.19
⁎C
0.12
±0.13
0.42
±0.10
0.47
±0.09
0.96
±0.00
As
−1.71
±0.04
⁎C
−1.31
±0.06
⁎A,C
−1.96
±0.05
−1.39
±0.07
−1.45
±0.08
−1.63
±0.09
−0.36
±0.18
⁎C
−0.59
±0.13
−1.16
±0.15
−1.11
±0.11
−1.12
±0.08
−1.27
±0.63
Se−
1.46
±0.05
⁎C
−1.14
±0.06
⁎A,C
−1.96
±0.05
−1.35
±0.07
⁎C
−1.36
±0.09
⁎C
−1.80
±0.07
−0.65
±0.13
⁎C
−0.76
±0.14
⁎C
−1.70
±0.23
−0.48
±0.11
−0.54
±0.10
−1.27
±0.63
A=
AerW
ay,S
S LMB=
surfacesp
read
LMB,
SSDMB=
surfacesp
read
,DI=
direct
injection,
C=
controls.
n=
totaln
umbe
rof
samples
forco
mpa
rative
statisticala
nalyses.
⁎Sign
ificantly
high
er(p
<0.05
)than
trea
tmen
tindicated.
Table 7Range of total metals concentrations in groundwater for the LMB (2005) study period at1.2 m and 2.0 m depth.
Metal Treatment 1.2 m 2.0 m
Al(mg L−1)
A 3.36–16.2*C 0.18–0.405SSLMB 2.76–29*C 0.44–2.29*AC 0.295–2.23 0.29–1.61*A
Cd(mg L−1)
A BDLa BDLSSLMB BDL BDLC BDL BDL
Cr(mg L−1)
A BDL–0.04 BDL-0.018SSLMB 0.022–0.055*A BDL-0.096C BDL-0.019 BDL-0.197
Co(mg L−1)
A BDL BDLSSLMB BDL-0.009 BDLC BDL BDL
Cu(mg L−1)
A 0.011–0.04 BDL-0.026SSLMB 0.011–0.06 0.009-0.024C 0.008–0.022 BDL-0.021
Fe(mg L−1)
A 2.8–12.7*C 0.029-0.288SSLMB 2.42–24.1*C 0.319–1.71*AC 0.161–1.82 0.211–2.13*A
Pb(mg L−1)
A BDL BDLSSLMB BDL-0.025 BDLC BDL BDL
Mo(mg L−1)
A BDL-0.058 BDLSSLMB BDL BDLC BDL BDL
Ni(mg L−1)
A BDL BDLSSLMB BDL-0.03 BDL-0.052C BDL BDL-0.082
Zn(mg L−1)
A 0.01–0.057 BDLSSLMB 0.012–0.105*C BDLC BDL-0.022 BDL-0.032
* Significantly higher (p<0.05) than treatment indicated.A=AerWay, SSLMB=surface spread, and C=control.
a Below detection limit: 0.004 mg L−1 (Cd, Co, Cu, Zn), 0.008 mg L−1 (Cr), 0.02 mg L−1
(Pb, Mo), and 0.024 mg L−1 (Ni).
881N. Gottschall et al. / Science of the Total Environment 408 (2010) 873–883
treatment effects for Al, Cr, Fe and Zn concentrations at 1.2 m; andat 2.0 m depth, Al and Fe concentrations showed significant landapplication treatment effects (Table 7).
Overall, metals in tile drainage behaved similarly to othermonitored compounds and microorganisms (Lapen et al., 2008a,b)during the application-induced tile flow event for reasons previouslydescribed. Thus, minimizing metal leaching at land application usingprudent land application options, like A to reduce contaminanttransport potential, will ultimately facilitate net reductions in metallosses to the broader environment where they can accumulate inecologically sensitive areas (e.g. Forstner and Muller, 1973). Thelonger metals can remain in field, the more likely they can be utilizedas micronutrients by standing crops (Ashworth and Alloway, 2004;Jones et al., 1996). Thus, in many regions in Canada, spring LMBapplications may be a better application time with respect tominimizing net metal loss to the environment via plant uptake,since late fall/winter leaching will occur to various degrees (Burns,1980) and plant uptake during this time would be minimal.
3.4. Metals in tile drainage following DMB application
Pre-application tile concentrations of metals were BDL, with theexceptions of Al, Cu, Fe and Zn. For most metals, the highest tileeffluent concentrations observed during the study period occurred asa result of the first significant precipitation event post-application(DOY 206), while for Se and Ni, the highest concentrations occurredduring the first rain events (DOY 293 and 300, respectively) after thesignificant dry period (no tile flow) that occurred between DOY 220and 289 (>100days post-application) (Table 5). For <100days post-application, there were significant land application treatment effectsfor all regulated metals, as well as Al and Fe (p<0.05) (Table 6); yetno significant differences were found between DI and SSDMB metal
882 N. Gottschall et al. / Science of the Total Environment 408 (2010) 873–883
mass loads in tile (p>0.05). For >100days post-application, therewere no land application treatment effects for any metals (p>0.05)(Table 6). However, both SSDMB and DI treatments had significantlyhigher (p<0.05) mass loads during <100days post-application,relative to >100days post-application for all metals except Se. Higherearly study period tile mass loads of metals in this study could beattributed to a combination of the leaching of, first, easily-solublemetal fractions and, soon after, soluble organo-metal complexes asthe biosolid further decomposes (McBride et al., 1999; Al-Wabel et al.,2002; Ashworth and Alloway, 2004). The considerable reduction intile mass loads following these higher loads, could have been due toslower decomposition of the remaining, more recalcitrant, organicmatter (Merrington et al., 2003). Overall, the limited data heresuggested that SSDMB and DI provided, at least, statistically similarmetal tile mass loads over a summer and fall study period;notwithstanding that metal loads in tile were enhanced by DMBapplication. More rapid decomposition of DMB was expected forSSDMB (relative to DI), allowing for amore rapid release of organically-bound metals into the soil–water environment during precipitationevents, whereas DI was expected to show tile metal contaminationmanifesting over a longer period due to retarded biosolid decompo-sition as a result of full product placement in soil and injectioncohesiveness (e.g., Gottschall et al., 2009; Edwards et al., 2009); butsuch did not appear to be the case. Considering these findings, DI,which additionally reduces odour (Gottschall et al., 2009) and,potentially, vector attraction problems (USEPA, 1999) compared toSSLMB (which despite incorporation efforts still leaves biosolids on thesoil surface), is a clear land application option.
4. Conclusions and summary
Several key findings in this study can be summarized below:
1. Land applications of LMB can be a source of total PBDE and metalloading to tile drainage systems and groundwater, and thereforeadjacent receiving surfacewaters. However, loading can be reducedat land application using prudent land application options thatminimize preferential flow of contaminants through soil macro-pores. For key PBDE congeners (BDE-47, -99, -100, -153, -154, -183,-209) for example, mass loss to tile (0–2 h post-application) as apercent ofmass applied, was ~0.04–0.1% and ~0.8–1.7% for AerWaySSD (one-pass surface apply over aerate tilled soil) and surfaceapplication (plus incorporation), respectively; a similar patternwasobserved for mass loss of metals.
2. PBDE concentrations in tile drainage resulting directly from LMBapplication were, on average, consistent with those from WWTPeffluents, but total discharges from tile (see Fig. 1) are considerablylower relative to those from WWTPs.
3. PBDE concentrations in surface soils on plotswhereDMBwas injectedin soil (2006) andwhere LMBwas applied via A (2005) indicated thatBDE-47, -99, -100, -153, -154, -183 and -209 concentrations were 22,26, 5, 3.3, 2.2, 2.3 and52 ng g−1dw, respectively. These residual PBDEconcentrations were generally two to three orders of magnitudegreater than soil with no biosolid application nearly 600days postLMB and 340days post DMB application.
4. Of all the PFAS compounds examined at LMB application, only PFOS(max concentration=17 ng L−1) and PFOA (12 ng L−1) were foundabove detectable limits in tile drainage. Thus, PFAS may not beconsiderable loading risks at LMB application, although, environ-mental fate of these compounds was not rigorously examined.
5. Following the DMB application, no significant differences in metalmass loads in tile were found between SSDMB and DI treatments(PBDE/PFAS were notmeasured) over a 147day summer–fall studyperiod. Thus, directly injecting DMB in soil does not appear toincrease significantly metal loading to tile drains relative to SSDMB,
and therefore, DI, a relatively new land application approach,should be considered a DMB land application option.
Acknowledgments
This research was supported by Agriculture and Agri-FoodCanada's Agricultural Policy Framework GAPs program, and OntarioMinistry of Agriculture and Food and Rural Affairs/Ontario Ministry ofEnvironment's Nutrient Management Joint Research Program. Fieldassistance by Mr. D. Irving (University of Guelph-Kemptville College)is greatly appreciated. A special thanks to Mr. M. Janiec for utilizationof Terratec Environmental Ltd.'s dewatered biosolids direct injectionsystem and in-kind logistical support.
References
Al-Wabel MA, Heil DM, Westfall DG, Barbarick KA. Solution chemistry influence onmetal mobility in biosolids-amended soils. J Environ Qual 2002;31:1157–65.
Ashworth DJ, Alloway BJ. Soil mobility of sewage sludge-derived dissolved organicmatter, copper, nickel and zinc. Environ Pollut 2004;127:137–44.
Breitholtz M,Wollenberger L. Effects of three PBDEs on development, reproduction, andpopulation growth rate of the parpacticoid copepod Nitocra spinipes. Aquat Toxicol2003;64:85–96.
Burns IG. A simple model for predicting the effects of winter leaching of residual nitrateon the nitrogen fertilizer need of spring crops. Eur J Soil Sci 1980;31(2):187–202.
Calafat AM, Kuklenyik Z, Caudill SP, Reidy JA, Needham LL. Perfluorochemicals in pooledserum samples from United States residents in 2001 and 2002. Environ Sci Tech2006;40:2128–34.
Camobreco VJ, Richards BK, Steenhius TS, Peverly JH, McBride MB. Movement of heavymetals through undisturbed and homogenized soil columns. Soil Sci 1996;161(11):740–50.
Costa LG, Giordano G. Developmental neurotoxicity of polybrominated diphenyl ether(PBDE) flame retardants. NeuroToxicology 2007;28:1047–67.
Dave G, Xiu R. Toxicity of mercury, copper, nickel, lead, and cobalt to embryos andlarvae of zebrafish, Brachydanio rerio. Arch Environ Contam Toxicol 1991;21(1):126–34.
Edwards M, Topp E, Metcalfe CD, Li H, Gottschall N, Bolton P, et al. Pharmaceutical andpersonal care products in tile drainage following surface spreading and injection ofdewatered municipal biosolids to an agricultural field. Sci Tot Environ 2009;407:4220–30.
Eljarrat E, Marsh G, Labandeira A, Barceló D. Effect of sewage sludges contaminated withpolybrominated diphenylethers on agricultural soils. Chemosphere 2008;71:1079–86.
Evandri MG, Costa LG, Bolle P. Evaluation of brominated diphenyl ether-99 toxicity withRaphidocelis subcapitata and Daphnia magna. Environ Toxicol Chem 2003;22(9):2167–72.
Fjällborg B, Li B, Nilsson E, Dave G. Toxicity identification evaluation of five metalsperformed with two organisms (Daphnia magna and Lactuca sativa). Arch EnvironContam Toxicol 2006;50:196–204.
Forstner U, Muller G. Heavy metal accumulation in river sediments: a response toenvironmental pollution. Geoforum 1973;14:53–61.
Gottschall N, Edwards M, Topp E, Bolton P, Payne M, Kleywegt S, et al. Nitrogen,phosphorus, and bacteria tile and groundwater quality following direct injection ofdewatered municipal biosolids into soil. J Environ Qual 2009;38:1066–75.
Hale RC, Alaee M, Manchester-Neesvig JB, Stapleton HM, Ikonomou MG. Polybromi-nated diphenyl ether flame retardants in the North American environment.Environ Int 2003;29(6):771–9.
Hassanin A, Breivik K, Meijer SN, Steinnes E, Thomas GO, Jones KC. PBDEs in Europeanbackground soils: levels and factors controlling their distribution. Environ Sci Tech2004;38:738–45.
He Z, Calvert D, Alva A, Li Y, Stoffella P, Banks D. Nitrogen transformation and ammoniavolatilization from biosolids and compost applied to calcareous soil. Compost SciUtil 2003;11:81–8.
Heidler J, Halden RU. Meta-analysis of mass balances examining chemical fate duringwastewater treatment. Environ Sci Tech 2008;42(17):6324–32.
Higgins A. Land application of sewage sludge with regard to cropping systems andpollution potential. J Environ Qual 1984;13:441–8.
Higgins CP, Field JA, Criddle CS, Luthy RG. Quantitative determination of perfluorochem-icals in sediments and domestic sludge. Environ Sci Tech 2005;39(11):3946–56.
Jarvis NJ. A review of non-equilibrium water flow and solute transport in soilmacropores: principles, controlling factors and consequences for water quality.European J Soil Sci 2007;58:523–46.
Ji K, Kim Y, Oh S, Ahn B, Jo H, Choi K. Toxicity of perfluorooctane sulfonic acid anperfluorooctanoic acid on freshwatermacroinvertebrates (Daphniamagna andMoinaMacrocopa) and fish (Oryzias latipes). Environ Toxicol Chem 2008;27(10):2159–68.
Jones DL, Darah PR, Kochian LV. Critical evaluation of organic acid mediated irondissolution in the rhizosphere and its potential role in root iron uptake. Plant Soil1996;180(1):57–66.
Karvelas M, Katsoyiannis A, Samara C. Occurrence and fate of heavy metals in thewastewater treatment process. Chemosphere 2003;53(10):1201–10.
Kay P, Blackwell PA, Boxall ABA. Fate of veterinary antibiotics in a macroporous tiledrained clay soil. Environ Toxicol Chem 2004;2004:1136–44.
883N. Gottschall et al. / Science of the Total Environment 408 (2010) 873–883
Keller C, McGrath SP, Dunham SJ. Trace metal leaching through a soil–grassland systemafter sewage sludge application. J Environ Qual 2002;31:1550–60.
Kjeldsen P, Barlaz MA, Rooker AP, Baun A, Ledin A, Christensen TH. Present and long-term composition of MSW landfill leachate: a review. Crit Rev Environ Sci Tech2002;32(4):297–336.
Lapen DR, Topp E, Edwards M, Sabourin L, Curnoe W, Gottschall N, et al. Effect of liquidmunicipal biosolid application method on tile and ground water quality. J EnvironQual 2008a;37:925–36.
Lapen DR, Topp E, Metcalfe CD, Li H, Edwards M, Gottschall N, et al. Pharmaceutical andpersonal care products in tile drainage following land application of liquidmunicipal biosolids. Sci Tot Environ 2008b;399(1–3):50–65.
Larsbo M, Lapen DR, Topp E, Metcalfe CD, Abbaspour KC, Fenner K. Simulation ofpharmaceutical and personal care product transport to tile drains after biosolidsapplication. J Environ Qual 2009;38.
Litz N. Some investigations into the behaviour of pentabromodiphenyl ether (PeBDE) insoils. J Plant Nutr Soil Sci 2002;165:692–6.
Loganathan BG, Sajwan KS, Sinclair E, Kumar KS, Kannan K. Perfluoroalkyl sulfonatesand perfluorocarboxylates in two wastewater treatment facilities in Kentucky andGeorgia. Water Res 2007;41:4611–20.
Matthews PJ, Davis RD. Control of metal application rates from sewage sludgeutilization in agriculture. Crit Rev Environ Sci Technol 1984;14(3):199–250.
McBride MB, Richards BK, Steenhius T, Russo JJ, Sauvé S. Mobility and solubility of toxicmetals and nutrients in soil fifteen years after sludge application. Soil Sci 1997;162(7):487–500.
McBride MB, Richards BK, Steenhius T, Spiers G. Long-term leaching of trace elementsin a heavily sludge-amended silty clay loam soil. Soil Sci 1999;164(9):613–23.
McGrath SP, Chaudri AM, Giller KE. Long-term effects of metals in sewage sludge onsoils, microorganisms and plants. J Indust Micro Biotech 1995;14(2):94-104.
Merrington G, Oliver I, Smernik RJ, McLaughlin MJ. The influence of sewage sludgeproperties on sludge-borne metal availability. Adv Environ Res 2003;8:21–36.
MOE and OMAFRA (Ministry of the Environment and Energy and Ministry ofAgriculture, Food and Rural Affairs. Guidelines for the utilization of biosolids andother wastes on agricultural land; 1996.
Moolenaar SW, Beltrami P. Heavymetal balances of an Italian soil as affected by sewagesludge and Bordeaux mixture applications. J Environ Qual 1998;27:828–35.
Moon H, Kannan K, Lee S, Choi M. Atmospheric deposition of polybrominated diphenylethers (PBDEs) in coastal areas in Korea. Chemosphere 2007;66(4):585–93.
Moradi A, Abbaspour KC, Afyuni M. Modelling field-scale cadmium transport below theroot zone of a sewage sludge amended soil in an arid region in Central Iran. J ContamHydrol 2005;79(3–4):187–206.
North KD. Tracking polybrominated diphenyl ether releases in a wastewater treatmentplant effluent, Palo Alto, California. Environ Sci Tech 2004;38(17):4484–8.
Öberg K, Warman K, Öberg T. Distribution and levels of brominated flame retardants insewage sludge. Chemosphere 2002;48:805–9.
Osako M, Kim Y, Sakai S. Leaching of brominated flame retardants in leachate fromlandfills in Japan. Chemosphere 2004;57(10):1571–9.
Ouellet G, Lapen DR, Topp E, Sawada M, Edwards M. Tree-structured relationshipsbetween earthworm biomass, land management and soil physical properties;Eastern Ontario, Canada. Appl Soil Ecol 2008;39:35–45.
Rahman F, Langford KH, Scrimshaw MD, Lester JN. Polybrominated diphenyl ether(PBDE) flame retardants. Sci Tot Environ 2001;275:1-17.
Rattan RK, Datta SP, Chhonkar PK, Suribabu K, Singh AK. Long-term impact of irrigationwith sewage effluents on heavy metal content in soils, crops and groundwater—acase study. Agric Ecosyst Environ 2005;109(3-4):310–22.
Sabourin L, Beck A, Duenke PW, Kleywegt S, Lapen DR, Lif H, et al. Runoff ofpharmaceuticals and personal care products following application of dewateredmunicipal biosolids to an agricultural field. Sci Tot Environ 2009;407(16):4596–604.
Samara F, Tsai CW, Aga DS. Determination of potential sources of PCBs and PBDEs insediments of the Niagara River. Environ Poll 2006;139(3):489–97.
Sanderson H, Boudreau TM, Mabury SA, Solomon KR. Impact of perfluorooctanoic acidon the structure of the zooplankton community in indoor microcosms. AquatToxicol 2003;62(3):227–34.
Ščančar J, Milačič R, Stražar M, Burica O. Total metal concentrations and partitioning ofCd, Cr, Cu, Fe, Ni and Zn in sewage sludge. Sci Tot Environ 2000;250:9-19.
Schut L. Sewage biosolids—managing urban nutrients responsibly for crop production;2005. Available online: http://www.omafra.gov.on.ca/english/nm/nasm/info/bro-chure.htm.
Sellström U, DeWit CA, Lundgren N, TysklindM. Effect of sewage-sludge application onconcentrations of higher-brominated diphenyl ethers in soils and earthworms.Environ Sci Technol 2005;39(23):9064–70.
Siddiqi MA, Laessig RH, Reed KD. Polybrominated diphenyl ethers (PBDEs): newpollutants—old diseases. Clinical Medicine and Research 2003;1(4):281–90.
Sierra J, Fontaine S, Desfontaines L. Factors controlling N mineralization, nitrification,and nitrogen losses in an Oxisol amended with sewage sludge. Aust J Soil Res2001;39:519–34.
Sinclair E, Kannan K. Mass loading and fate of perfluoroalkyl surfactants in wastewatertreatment plants. Environ Sci Tech 2006;40:1408–14.
SongM, Chu S, Letcher RJ, Seth R. Fate, partitioning, andmass loading of polybrominateddiphenyl ethers (PBDEs) during the treatment processing of municipal sewage.Environ Sci Tech 2006;40:6241–6.
Speir TW, Van Schaik AP, Percival HJ, Close ME, Pang L. Heavy metals in soil, plants andgroundwater following high-rate sewage sludge application to land. Wat Air SoilPoll 2003;150:319–58.
Sukkariyah BF, Evanylo G, Zelazny L, Chaney RL. Recovery and distribution of biosolids-derived trace metals in a clay loam soil. J Environ Qual 2005;34:1843–50.
Terry RE. NelsonDW, Sommers LE. Carbon cycling during sewage sludge decompositionin soils. Soil Sci Soc Am J 1979;43:494–9.
Ter Schure AFH, Larsson P. Polybrominated diphenyl ethers in precipitation in SouthernSweden (Skåne, Lund). Atmospheric Environ 2002;36(25):4015–22.
Thibodeaux JR, Hanson RG, Rogers JM, Grey BE, Barbee BD, Richards JH, et al. Exposureto perfluorooctane sulfonate during pregnancy in rat and mouse. I: Maternal andprenatal evaluations. Toxicol Sci 2003;74:369–81.
Turpin KM, Lapen DR, Robin MJ, Topp E, Edwards M, Curnoe WE, et al. Slurry-application implement tine modification of soil hydraulic properties underdifferent soil water content conditions for silt–clay loam soils. Soil Tillage Res2007a;95:120–32.
Turpin KM, Lapen DR, Topp E, Robin MJ, Edwards M, Curnoe WE, et al. Tine-influencedinfiltration patterns and informing timing of liquid amendment applications usingbrilliant blue dye tracers. Biosyst Eng 2007b;98:235–47.
USEPA. Environmental regulations and technology. Control of pathogens and vectorattraction in sewage sludge; 1999. EPA/625/R-92/013.
Venier M, Hites RA. Atmospheric deposition of PBDEs to the Great Lakes featuring aMonte Carlo analysis of errors. Environ Sci Tech 2008;42:9058–64.
Vonderheide AP, Mueller-Spitz SR, Meija J, Welsh GL, Mueller KE, Kinkle BK, et al. Rapidbreakdown of brominated flame retardants by soil microorganisms. J Anal AtSpectrom 2006;21:1232–9.
Wania F, Dugani C. Assessing the long-range transport potential of polybrominateddiphenyl ethers: a comparison of four multimedia models. Environ Toxicol Chem2003;22(6):1252–61.
Welsh GL, Mueller KE, Soman RS, Vonderheide AP, Shann JR. Accessibility of polybromi-nated diphenyl ether congeners in aging soil. J Environ Mon 2009;11:1658–63.
Wolfe MF, Schwarzbach S, Sulaiman RA. Effects of mercury on wildlife: a comprehen-sive review. Environ Toxicol Chem 1998;17(2):146–60.
Yeoman S, Stephenson T, Lester JN, Perry R. Removal of phosphorus during wastewatertreatment: a review. Environ Poll 1988;49(3):183–233.
Zaleski KJ, Josephson KL, Gerba CP, Pepper IL. Survival, growth and regrowth of entericindicator and pathogenic bacteria in biosolids, compost, soil, and land appliedbiosolids. J Resid Sci Technol 2005;2(1):49–63.
Zhang J, Tomanek M, Dong H, Arnold RG, Ela WP, Quanrud DM, et al. Fate ofpolybrominated diphenyl ethers, nonylphenol, and estrogenic activity duringmanaged infiltration of wastewater effluent. J Environ Eng 2008;134(6):433–42.
Zou M, Ran Y, Gong J, Mai B, Zeng EY. Polybrominated diphenyl ethers in watershedsoils of the Pearl River delta, China: occurrence, inventory, and fate. Environ SciTech 2007;41(24):8262–7.
Zufiaurre R, Olivar A, Chamorro P, Nerín C, Callizo A. Speciation of metals in sewagesludge for agricultural uses. Analyst 1998;123:255–9.
