Global Distributions, Source-Type Dependencies, and Concentration Ranges of Per- and Polyfluoroalkyl Substances in Groundwater

Abstract

A meta-analysis was conducted of published literature reporting concentrations of per- and polyfluoroalkyl substances (PFAS) in groundwater for sites distributed in 21 countries across the globe. Data for more than 35 PFAS were aggregated from 96 reports published from 1999 to 2021. The final data set comprised approximately 21,000 data points after removal of time-series and duplicate samples as well as non-detects. The reported concentrations ranged over many orders of magnitude, from ng/L to mg/L levels. Distinct differences in concentration ranges were observed between sites located within or near sources versus those that are not. Perfluorooctanoic acid (PFOA), ranging from <0.03 ng/L to ~7 mg/L, and perfluorooctanesulfonic acid (PFOS), ranging from 0.01 ng/L to ~5 mg/L, were the two most reported PFAS. The highest PFAS concentration in groundwater was ~15 mg/L reported for the replacement-PFAS 6:2 fluorotelomer sulfonate (6:2 FTS). Maximum reported groundwater concentrations for PFOA and PFOS were compared to concentrations reported for soils, surface waters, marine waters, and precipitation. Soil concentrations are generally significantly higher than those reported for the other media. This accrues to soil being the primary entry point for PFAS release into the environment for many sites, as well as the generally significantly greater retention capacity of soil compared to the other media. The presence of PFAS has been reported for all media in all regions tested including areas that are far removed from specific PFAS sources. This gives rise to the existence of a “background” concentration of PFAS that must be accounted for in both regional and site-specific risk assessments. The presence of this background is a reflection of the large-scale use of PFAS, their general recalcitrance, and the action of long-range transport processes that distribute PFAS across regional and global scales. This ubiquitous distribution has the potential to significantly impact the quality and availability of water resources in many regions. In addition, the pervasive presence of PFAS in the environment engenders concerns for impacts to ecosystem and human health.

1.0. Introduction

With the ever-increasing global production of organic chemicals including flame retardants, antibiotics, pharmaceuticals, steroids, and per- and polyfluoroalkyl substances (PFAS), it is not surprising that there is continuing focus on contaminants of emerging and critical concern in the natural environment. PFAS in particular have received significant attention over the past several years (e.g., ATSDR, 2021; EPA, 2021). Many of these compounds, of which several thousand have been identified (OECD, 2018), are known to be environmentally persistent, bioaccumulative (especially long-chain homologues), and have potential deleterious toxicological impacts to a variety of organisms (e.g., Buck et al. 2011). The detection, occurrence, and characterization of PFAS in the environment is being investigated and reported at multiple scales, ranging from individual contaminated sites, to regional background studies, to global surveys. It has become evident that PFAS are ubiquitous in environmental media across the globe (e.g., Prevedouros et al. 2006; Rayne and Forest, 2009; Ahrens, 2011; Krafft and Riess, 2015; Brusseau et al. 2020). Their widespread global distribution and accumulation in waters, soils, and biota coupled with their persistence and potential ecological and human-health impacts makes quantitative characterization of the global distribution of PFAS compounds necessary for improved risk-assessment studies.

Many hundreds of studies have been conducted over the past two decades investigating the presence of PFAS in various environmental media, with a major focus on surface water, groundwater, marine waters, and soil. A recent meta-analysis of studies investigating PFAS distributions in soil demonstrated that PFAS are ubiquitous in soil across the globe (Brusseau et al., 2020). Furthermore, it was demonstrated that soil concentrations range over many orders of magnitude, in part as a function of site type. Based on this study and the studies cited therein, it is likely that soil is a significant long-term reservoir of PFAS at many sites. There are multiple potential concerns for sites with PFAS present in soil, including uptake by plants, entrainment in surface runoff, aeolian transport, and leaching to groundwater. The current work is focused on potential impacts to groundwater.

Groundwater has been and continues to be a critical component of water resources for many locales and regions. For example, groundwater is estimated to provide almost half of all drinking water worldwide (UNESCO, 2022). Furthermore, groundwater is estimated to represent 26% of total water withdrawals for combined domestic, industrial, and irrigation uses (Margat and van der Gun, 2013). The apparent widespread distribution of PFAS in the environment may pose a significant threat to groundwater quality and, hence, water-resource sustainability in some areas. While there have been many recent reviews providing anecdotal overviews of PFAS occurrence in various media, a detailed, quantitative analysis of PFAS concentrations and distributions specifically for groundwater at a global scale has yet to be reported.

This study is part of a special-topics issue focused on groundwater quality (Lapworth et al., 2022). The objectives of this study are four-fold. The first objective is to determine the global distributions of PFAS in groundwater. Second, the reported data will be characterized to delineate concentration ranges observed for different PFAS. Third, the concentration ranges as a function of location and source type will be examined. Fourth, the PFAS distributions for groundwater will be compared to those observed for soils, surface waters, marine waters, and precipitation. The results of the study will provide a comprehensive, quantitative characterization of global PFAS distributions in groundwater that is anticipated to be useful for future assessments of potential impacts on water resources.

2.0. Materials and Methods

2.1. General Approach

Published literature was searched to identify studies reporting concentrations of PFAS in groundwater, surface water, marine waters, precipitation, and soil. Multiple search tools were used, including Web of Science, SCOPUS, and Google Scholar. Multiple PFAS-related search terms were used, including “PFAS”, “Perfluor*”, “Polyfluor*”, and “PFC”. In addition, individual PFAS terms such as “PFOS”, “PFOA”, and “GenX” were used. Cited references in all identified publications were also examined for applicable studies. Information including the type of study, nature of the locations surveyed, the number of sampling locations, number of PFAS analyzed, and the PFAS concentrations was recorded.

As noted, this study is part of a special-topics issue on groundwater quality. Therefore, the detailed data analyses and interpretations reported herein are focused specifically on PFAS in groundwater. The full ranges of concentrations reported for all monitored PFAS will be aggregated for groundwater. Conversely, the comparison of concentrations in groundwater to those in other media will focus on the two most-investigated PFAS, PFOA and PFOS. The maximum reported concentrations will be used in the comparison.

All identified groundwater studies were processed following the approach described in section 2.2. For the soil studies, the original study of Brusseau et al. (2020) included 77 reported data sets gleaned from all identified studies. An additional 17 new studies were added for this update. The soil data therefore represent an analysis of all identified studies. The references are presented in the SI. Several hundred studies were identified for surface waters and marine waters. Large percentage samples of these were processed for the data comparison. The reference bibliographies from which the tabulated studies were drawn are reported in the SI. Twenty studies were identified that reported concentrations of PFAS in precipitation. All of these were processed, with the references presented in the SI.

2.2. Groundwater Studies

Following Brusseau et al., (2020), the sites sampled in the groundwater studies were split into three categories, no-known-source sites (sites with no identified nearby PFAS source), primary-source sites (fire-training areas, manufacturing plants), and secondary-source sites (biosolids application, treated wastewater use, landfills). The classification of the sites was based on information reported by the authors of the original studies. The great majority of the studies did not report isomer information for individual PFAS. Thus, the reported concentrations are assumed to represent total concentration of the individual PFAS. For each study used in this work, all duplicate data points for the same sampling location were compiled as the arithmetic average. For investigations reporting PFAS concentrations measured sequentially at a given location, the time-series data were arithmetically averaged to avoid the inclusion of repeat measurements reported for a single groundwater monitoring site.

Sample processing and analysis methods varied somewhat across the studies, and the reported quantitative detection limits (LOQ) also varied among the studies. Weighted-mean concentrations were calculated within source-type categories using weighted averaging scaled by the respective number of groundwater sampling locations at each site or study. In addition, non-weighted arithmetic means were determined using the total number of aggregated concentrations. 95% confidence intervals (CI) were calculated for the means using the respective coefficients of variation determined for the reported PFAS concentrations. Maximum concentrations were tabulated, as well as the minimum reported PFAS concentrations determined as the reported minimum concentration or the reported LOQ. Concentrations reported as detected but not quantifiable were treated using the standard approach of ½ the respective method detection limit to provide specific values. The frequency of concentrations quantified using this approach for each PFAS was calculated for each source-type category. Non-detects were not included in the analyses. Summary results obtained for groundwater are compared to summary data sets generated for soils, surface waters, marine waters, and precipitation. Statistical analyses were conducted using tools in Excel.

3.0. Groundwater Results and Discussion

3.1. Study Parameters

An extensive search for publications, reports, conference proceedings, investigative studies, health advisory directives, and environmental guidance documents was conducted to collect PFAS concentrations measured in groundwater around the world. More than 230 documents were accessed and reviewed. These documents represent those that are publically accessible through standard search means. Each study was reviewed against a set of criteria to determine inclusion or exclusion in the meta-analysis.

Studies were excluded from the meta-analysis based on the following: (1) measured groundwater concentrations were not reported; (2) the data comprised only non-detects; (3) the data were previously reported in other studies; (4) the data represented potable-water samples and not direct groundwater samples; and (5) the data represented benthic porewater samples. Only original-source data were used to avoid repeat data sets (double counting). Potable-water data were excluded even if groundwater was the source of the sampled potable water, given that the impacts of water treatment and water distribution may alter PFAS concentrations. Benthic porewater data were excluded due to the potential impact of the water composition and PFAS levels in the surface-water body on porewater concentrations. In addition, several works report only the range of PFAS concentrations measured and not individual concentrations; those study data are only included in the analysis of overall maximum and minimum PFAS concentrations.

A total of 96 studies dating from 1999 to 2021 were mined for information on PFAS concentrations in groundwater (Table 1). The number of publications reporting groundwater concentration data each year is presented in Figure 1. Inspection of the figure shows that the number of studies increased quickly in the early 2000s and peaked in the mid-late 2000s.

Table 1.

Studies reporting PFAS concentrations in groundwater.

Date	Reference	Source Type	Location	Number of PFAS Reported
2022	Johnson	S	USA (Western) – biosolids land-application	2
2022	Wang et al., 2022	B	China (Jiangxi Province)	18
2021	Cáñez et al., 2021	S	USA (AZ) – reclaimed water recharge	8
2021	Nickerson et al., 2021	P	USA – AFFF impacted; military installation	20
2021	Pepper et al., 2021	B	USA (AZ)	8
2021	Petre et al., 2021	P	USA (NC) – PFAS manufacturing facility	24
2021	Propp et al., 2021	S	Canada – municipal landfill leachate impacted	17
2021	Roostaei et al., 2021	P	USA (NC) – fluorochemical facilty	1
2021	Xu et al., 2021	S	China – municipal landfill leachate impacted	15
2021	Yong et al., 2021	P	South Korea – PFAS leakage incident in 2018	8
2021	Zhou et al., 2021	B	China (Loess Plateau)	34
2020	Barreca et al., 2020	B	Italy	12
2020	EA, 2020	B	England and Wales	14
2020	Harrad et al., 2020	S	Ireland – landfill leachate impacted	5
2019	GHD, 2019b	P	Australia (Mackay Airport) – AFFF impacted	25
2019	GHD, 2019a	P	Australia (Cairns Airport) – AFFF impacted	26
2019	Bao et al., 2019	P	China (Fuxin) – fluorochemical industrial park	10
2019	Bjornsdotter et al., 2019	P	Sweden – firefighting training sites	26
2019	Cao et al., 2019	B	China (Yuqiao Reservoir of Tianjin)	17
2019	Dauchy et al., 2019	S	France – near firefighting training sites	16
2019	Gao, Q. et al., 2019	S	Sweden – wastewater infiltration	14
2019	Gao, Y. et al., 2019	P	China – around a PFSA manufactory	10
2019	Hepburn et al., 2019	S	Australia – legacy landfill leachate impacted	14
2019	Kaserzon et al., 2019	S	Australia – near fire training grounds	17
2019	Liu et al., 2019	B	China (Shijiazhuang City near Hutuo River)	10
2019	Martin et al., 2019	P	Newfoundland and Ontario, Canada – fire equipment testing sites	41
2019	Sammut et al., 2019	B	Maltese	7
2018	GHD, 2018	P	Australia (Adelaide Airport) – AFFF impacted	26
2018	Gobelius et al., 2018 Ahrens et al., 2016	S	Sweden – mix of hotspots and drinking water sources	26
2018	Horsley Witten, 2018	P	USA (MA) – airport fire fighting sites	6
2018	Lapworth et al., 2018	B	England	5
2018	Liu et al., 2018	B	China	14
2018	Steele et al., 2018	P	Alaska (Military Installation and Pease AFB)	11
2018	Sun et al., 2018	B	China (Shanghai)	10
2018	Szabo et al., 2018	S	Australia – irrigation of treated wastewater	13
2018	von der Trenck et al., 2018	B	Germany	9
2018	Wei et al., 2018	P, B	China – industrial site	17
2017	Barber et al., 2014 Weber et al., 2017	P	USA (MA) – fire training site	16
2017	Braunig et al., 2017	P	Australia – AFFF impacted site	11
2017	Dauchy et al., 2017	S	France – oil refinery; fire training area	10
2017	Hongkachok et al., 2017	S	Thailand – municipal and industrial waste disposal sites	5
2017	Munoz et al., 2017	B	Overseas France	17
2017	Procopio et al., 2017	S	USA (NJ) – near industrial business park	10
2017	Woodard et al., 2017	P	USA (New England) – AFFF fire fighting activities (US Air Force Base)	12
2017	Zhu et al., 2017	B	China (Huai River Basin)	12
2016	GHD, 2016	P	Australia (Gold Coast Airport) – AFFF impacted	24
2016	Anderson et al., 2016	P	USA – AFFF release sites (US Air Force Base)	14
2016	Boiteux et al., 2016	P	France –AFFF fire fighting training site	29
2016	Chen et al., 2016	S	China – farms near fluorochemical industrial area and in rural area	17
2016	Kuroda et al., 2016	B	Japan (Tokyo)	13
2016	Li et al., 2016	B	China	7
2016	Liu et al., 2016	S	China – 10 km radius around a mega-fluorochemical industrial park	12
2016	Qi et al., 2016	B	China	13
2016	Sharma et al., 2016	B	India (Ganges River)	19
2016	Shiwaku et al., 2016	P	Japan (Osaka) – fluoropolymer plant	11
2015	Barzen-Hanson and Field, 2015	P	USA – nine military bases with record of fire fighting AFFF activity	8
2015	Castiglioni et al., 2015	S	Italy – industrial and highly urbanized area	1
2015	DMEPA et al., 2015	S	Denmark – beneath a carpet industry	NA
2015	Duong et al., 2015	B	Vietnam	7
2015	Filipovic et al., 2015	P	Sweden – AFB former military airfield	4
2015	Lin et al., 2015	B	Taiwan (Taipei City, Taipei County, Hsinchu)	10
2015	Xiao et al., 2015	P	USA - former PFAS manufacturing and disposal site	2
2014	Boone et al., 2014	P	USA (MS) – AFFF fire training area	9
2014	McGuire et al., 2014	P	USA (SD) – AFFF fire fighting training area	10
2013	Anumol et al., 2013	B	USA (AZ)	6
2013	Backe et al., 2013	P	USA – military bases including TAFB, NASF, WAFB	16
2013	Eschauzier et al., 2013	S, B	The Netherlands – military camps, landfills, and urban areas	8
2013	Houtz et al., 2013	P	USA (SD) – firetraining area US Air Force Base	17
2013	Post et al., 2013	B	USA (NJ)	10
2013	Wagner et al., 2013	P	Germany – AFFF impacted site	9
2012	Boiteux et al., 2012	B	France	10
2012	Weiβ et al., 2012	S	Germany – downstream of fire training area	6
2011	Bao et al., 2011	P	China (NE) – fluorochemical industrial park	7
2011	Hoffman et al., 2011	P	USA (OH and WV) – fluoropolymer production facility	1
2011	Lindstrom et al., 2011	P	USA (Al) – biosolids from fluorochemicalimpacted wastewater treatment plant	10
2011	Meyer et al., 2011	B	Canada	9
2010	Loos et al., 2010	B	Europe (pan-European)	9
2010	Wilhelm et al., 2010	B	Germany	NR
2009	Jin et al., 2009	B	China	2
2009	Murakami et al., 2009	B	Japan (Tokyo)	9
2009	Post et al., 2009	B	USA (NJDEP)	1
2009	Quinones and Snyder, 2009	S	USA (Lake Mead) – WWTP impacted	8
2009	Rumsby et al., 2009	P	Europe (London) - Buncefield Oil Depot	1
2009	Rumsby et al., 2009	B	France (Island of Jersey)	1
2008	Atkinson et al., 2008	P	England – fire fighting training airbase	6
2008	Markall et al., 2008	P	UK (East Anglia) – AFFF activity airbase	6
2008	Plumlee et al., 2008	B	USA	10
2008	SFT, 2008	P	Norway – fire training facilities	19
2006	Oliaei et al., 2006	P	USA (MN) – 3M PFAS production facilities	12
2005	ATSDR, 2005	P	USA (MN) – 3M Cottage Grove PFAS production facility	7
2004	Schultz et al., 2004	P	USA (Wurtsmith AFB)	10
2003	Moody et al., 2003	P	USA (MI) – fire training area Wurtsmith Air Force Base	4
1999	Moody and Field, 1999	P	USA – fire training facility (NAS Fallon and Tyndall Air Force Base)	3

Figure 1.

Number of studies reporting PFAS concentrations in groundwater per year.

The aggregated data set includes more than 35 PFAS reported in groundwater measured for more than 4400 sampling locations. The total number of groundwater-concentration data points is approximately 21,000, after accounting for duplicates and excluding time-series data and non-detects. The total number of individual PFAS reported in each study varies greatly, ranging from 1 to 41 (Table 1). The PFAS that were reported multiple times in the literature are shown in Figure 2. It is observed that PFOA, PFOS, PFHxS, PFHpA, PFHxA, and PFBS are the most frequently reported. All perfluorocarboxylates (PFCA) from C3 to C18 are included, as well as all perfluorosulfonates (PFSA) from C3 to C10. The most frequently reported PFAS replacement and/or precursor constituents include PFOSA and 4:2, 6:2, and 8:2 FTS. GenX was reported in a few studies.

Figure 2.

Numbers of publications reporting groundwater concentrations for different PFAS.

The data compilation includes PFAS concentrations measured in 21 countries (Figure 3). The distribution of the sampling locations are visualized in Map-1 presented in the Supplemental Information (SI). The vast majority of data have been reported for Asia, North America, and Western Europe. Given the widespread presence of PFAS reported for the regions sampled, it would appear prudent to conduct investigations in those regions lacking studies.

Figure 3.

Number of studies reporting PFAS concentrations in groundwater per country.

3.2. PFAS Concentrations in Groundwater

The groundwater sites were split into the three categories of primary-source sites, secondary-source sites, and no-known-source (background) sites. Approximately 40% of the data are reported for sites classified as primary-source sites (Table 2). More than half of these sites are identified as military installations or airports, primarily comprising AFFF-impacted sites. PFAS production and manufacturing facilities comprise nearly one-third of the primary-source sites. Approximately 30% of the data are reported for secondary-source sites, which include those impacted by the disposal of waste products (e.g., municipal and industrial landfills, treated wastewater-impacted sites, and land application of biosolids). Data reported for no-known-source sites comprise approximately 29% of the total.

Table 2.

PFAS concentrations (ng/L) in groundwater for three source types.

PFAS	Primary Sources (n = 8460)			Secondary Sources (n = 6279)			No-Known Sources (n = 5990)
	Total # of Sample Locations	Max	Meana (95% CI)b	Total # of Sample Locations	Max	Meana (95% CI)b	Total # of Sample Locations	Max	Meana (95% CI)b
PFPrA	12	18000	820 (2800)	0	NR	NR	0	NR	NR
PFBA	441	87000	4900 (1100)	380	2000	21 (12)	478	90.3	5.7 (1.2)
PFPeA	447	300000	13000 (3700)	412	902	11 (5)	720	290	5.0 (1.0)
PFHxA	500	875000	25000 (7300)	441	3800	31 (17)	755	299	4.1 (1.1)
PFHpA	485	206000	4700 (1500)	436	4300	22 (18)	774	61	6.3 (1.0)
PFOA	663	6570000	64000 (29000)	485	70000	325 (320)	1051	1800	13 (5)
PFNA	421	66700	890 (380)	409	1034	3.0 (3.2)	677	620	3.4 (1.6)
PFDA	352	143040	570 (890)	404	560	3.6 (2.8)	670	34	0.42 (0.09)
PFUnA	194	760	11 (6)	291	20	0.9 (0.3)	304	28	0.53 (0.21)
PFDoDA	193	1040	6.7 (1.2)	269	59	0.9 (0.5)	308	13.7	0.32 (0.12)
PFTriA	40	<LOQ	3.8 (1.7)	249	50	1.5 (0.8)	115	5.9	0.27 (0.14)
PFTeA	50	4000	6.7 (3.5)	252	50	1.3 (0.9)	291	9.2	0.19 (0.07)
PFHxDA	18	<LOQ	0.90 (0.24)	221	1.0	0.05 (0.01)	192	8.9	0.20 (0.09)
PFOcDA	18	<LOQ	0.34 (0.03)	221	0.40	0.12 (0.01)	191	2.9	0.17 (0.03)
PFPrS	78	176000	30200 (14000)	0	NR	NR	0	NR	NR
PFBS	459	822000	14000 (5900)	429	750	7.1 (3.5)	668	143	8.9 (1.9)
PFPeS	132	374000	25000 (16000)	33	16	2.8 (0.9)	0	NR	NR
PFHxS	506	2380000	65000 (19000)	437	2860	19 (11)	732	56	6.0 (1.0)
PFHpS	158	66700	6400 (2400)	47	1325	56 (50)	67	10	0.70 (0.36)
PFOS	582	4600000	93000 (27000)	421	8350	42 (24)	1043	20000	46 (48)
PFOSA	192	22000	1300 (425)	36	13	0.5 (0.5)	154	95	1.0 (1.0)
PFNS	73	20000	2800 (1400)	0	NR	NR	0	NR	NR
PFDS	176	23800	140 (290)	262	20	0.6 (0.3)	71	6.9	0.66 (0.36)
4:2 FTS	55	12000	2500 (980)	0	NR	NR	30	NR	0.38 (NA)
6:2 FTS	280	14600000	120000 (110000)	102	10	1.1 (0.2)	223	118	1.8 (1.2)
8:2 FTS	251	98000	3100 (1100)	0	NR	NR	30	NR	1.1 (NA)
GenX	1207	4000	108 (NA)	54	1116	459 (85)	30	NR	0.060 (NA)

^aMean concentration calculated as a weighted average scaled to total number of sample locations per PFAS

^b95% Confidence Interval (CI)

NA = Not Available

Information on the lengths of groundwater contaminant plumes associated with identified single sources was provided in a few of the studies. The characterized plume lengths range from approximately 0.5 to >5 km. The sites include AFFF-impacted sources (Moody et al., 2003; Filipovic et al., 2015; Braunig et al., 2017; Weber et al., 2017; Nickerson et al., 2021), a manufacturing plant (Bao et al., 2019), and a treated-wastewater recharge facility (Canez et al., 2021). These relatively large plume lengths are consistent with the relatively low magnitudes of sorption and transformation anticipated for many PFAS in groundwater systems.

The arithmetic mean concentrations for all sites for PFOA and PFOS, the two most reported PFAS in groundwater, are 15,000 (±12000) ng/L and 21,000 (±11000) ng/L, respectively (Table 3). Mean concentrations were also calculated for each category of site (Table 2). These means were calculated as a weighted average scaled to the respective number of sampling locations to account for the large ranges in numbers of locations sampled among the studies. The weighted mean concentrations of PFOA and PFOS at primary source sites are 64,000 (±29000) ng/L and 93,000 (±27000) ng/L, respectively. The weighted means for secondary-source sites and no-known source sites are 325 (±320) ng/L and 13 (±5) ng/L, respectively, for PFOA, and 42 (±24) ng/L and 46 (±48) ng/L, respectively, for PFOS. The concentrations reported for these latter two source types are several orders of magnitude lower than those reported for primary-source sites. This disparity is observed for all reported PFAS with the exception of the longer-chain C11 – C18 PFAS.

Table 3.

Mean, minimum, and maximum reported PFAS concentrations.

PFAS	# Well Locations	Concentration Range (ng/L)
		Min or LOQa	Maximum		Mean
			Reported	Reference	Arithmetic	95%CIb
PFPrA	12	5.1	18,000	Martin et al., 2019	2200	2900
PFBA	1298	0.004	87,000	Houtz et al., 2013	1400	430
PFPeA	1578	0.0211	300,000	Nickerson et al., 2021	3700	1400
PFHxA	1695	0.02	875,000	GHD, 2018	8100	2900
PFHpA	1695	0.02	206,000	ATSDR, 2005	1400	520
PFOA	2199	0.028	6,570,000	Moody and Field, 1999	15000	12000
PFNA	1507	0.02	66,700	McGuire et al., 2014	160	120
PFDA	1426	0.0096	143,040	McGuire et al., 2014	160	250
PFUnA	789	0.0024	760	EA, 2020	3.4	1.5
PFDoDA	769	0.019	1,040	EA, 2020	2.1	0.4
PFTriA	404	0.015	7.0	Propp et al., 2021	1.4	0.5
PFTeA	593	0.020	4,000	EA, 2020	1.2	0.4
PFHxDA	431	0.005	8.9	Wang et al., 2022	1.3	0.1
PFOcDA	430	0.02	2.92	Wang et al., 2022	0.1	0.1
PFPrS	77	0.009	176,000	Nickerson et al., 2021	22000	14000
PFBS	1555	0.01	822,000	GHD, 2018	4900	2300
PFPeS	165	0.2	374,000	GHD, 2018	25000	14000
PFHxS	1674	0.01	2,380,000	GHD, 2018	20000	7800
PFHpS	272	0.022	66,700	GHD, 2018	2600	1100
PFOS	2046	0.01	4,600,000	Nickerson et al., 2021	21000	11000
PFOSA	382	0.004	22,000	Nickerson et al., 2021	293	232
PFNS	72	2	20,000	Nickerson et al., 2021	1800	1400
PFDS	509	0.005	23,800	Anderson et al., 2016	56	99
4:2 FTS	84	0.38	12,000	Nickerson et al., 2021	1600	1000
6:2 FTS	605	0.06	14,600,000	Moody and Field, 1999	56000	58000
8:2 FTS	281	0.25	98,000	Nickerson et al., 2021	1600	1200
GenX	1291	0.060	4,000	Roostaei et al., 2021	440	84

^aReported minimum equal to limit of quantification (LOQ); cells shaded in blue

^b95% Confidence Interval (CI)

Several short-chain PFAS including C3 – C6 PFCA and PFSA were reported in the studies. PFPrA and PFPrS (acronyms are presented in the SI), both C3 compounds, are reported multiple times but only at primary-source sites (Barzen-Hanson and Field, 2015; Bjornsdotter et al., 2019; Martin et al., 2019; Nickerson et al., 2021). All other short-chain PFCA and PFSA (C4 – C6, C7) have been reported for all three source categories. Longer-chain PFAS (C11 – C18) including PFUnDA, PFDoDA, PFTriA, PFTeA, PFHxDA, PFOcDA have been reported at all three source types, with mean concentrations ranging from 0.1 to 3.4 ng/L.

The limits of quantification reported in the studies for each PFAS are presented in Table 4. The LOQs vary by a factor of 2 to a few orders of magnitude for a given PFAS within each source type. They also vary for each PFAS across the various source types. As described previously, concentrations reported as detected but not quantifiable were treated in this study using ½ the respective method detection limit to provide numerical values for the analysis. The frequency of concentrations quantified using this approach for each measured PFAS and in each source-type category is listed in Table 4. The analysis reveals that the frequency of <LOQ concentrations measured varies greatly by PFAS. Only 13% of the PFOA concentrations and 11% of the PFOS concentrations were reported at <LOQ for the primary-source sites. Conversely, from ~50% up to 100% of the reported concentrations for C11 – C18 PFAS were reported at <LOQ. Similarly, nearly 50% or more of the measured concentrations of C3 in groundwater were reported at <LOQ. The differences in the magnitudes of LOQs have the potential to impact statistical analyses of the concentration data sets, with the magnitude of impact depending upon the specific range in LOQs and the relative proportion of the total data set represented by non-quantifiable detections.

Table 4.

Range of limits of quantification (<LOQ) and frequency of reported LOQ.

PFAS	Limit of Quantification Range (ng/L)
	Primary Sites	Secondary Sites	Background Sites
PFPrA	5.1 – 25 (48%)	NR	NR
PFBA	0.1 – 100 (35%)	0.1 – 0.20 (55%)	0.004 – 5 (11%)
PFPeA	0.2 – 50.8 (29%)	0.03 – 40 (52%)	0.0211 – 5 (9%)
PFHxA	0.5 – 50(15%)	0.93 – 10 (40%)	0.02 – 5 (7%)
PFHpA	1 – 50(18%)	0.053 – 40 (42%)	0.02 – 5 (21%)
PFOA	0.5 – 97.6 (13%)	0.05 – 5 (40%)	0.03 – 5 (5%)
PFNA	0.05 – 100 (35%)	0.02 – 40 (57%)	0.027 – 7 (14%)
PFDA	0.1 – 100 (56%)	0.035 – 40 (59%)	0.017 – 5 (54%)
PFUnDA	1 – 50 (51%)	0.03 – 40 (71%)	0.04 – 0.5 (66%)
PFDoDA	0.3 – 50 (67%)	0.03 – 40 (72%)	0.025 – 0.3 (55%)
PFTriDA	0.05 – 50 (78%)	0.05 – 100 (82%)	0.04 – 0.3 (73%)
PFTeDA	0.3 – 120 (75%)	0.041 – 100 (88%)	0.015 – 0.25 (80%)
PFHxDA	0.5 – 1 (100%)	0.05 – 0.2 (99%)	0.006 – 0.5 (91%)
PFOcDA	0.4 – 0.5 (100%)	0.2 – 0.25 (99%)	0.02 – 0.5 (86%)
PFPrS	0.009 – 50 (65%)	NR	NR
PFBS	2.3 – 100 (21%)	0.097 – 40 (47%)	0.02 – 25 (14%)
PFPeS	5 – 30 (19%)	0.1 (9%)	NR
PFHxS	0.1 – 0.5 (9%)	0.15 – 10 (51%)	0.017 – 5 (25%)
PFHpS	2 – 50 (16%)	0.025 – 40 (24%)	0.004 – 0.25 (60%)
PFOS	0.05 – 100 (11%)	0.21 – 40 (55%)	0.01 – 5 (6%)
PFOSA	0.7 – 50 (78%)	0.008 – 0.031 (17%)	0.004 – 11.5 (48%)
PFNS	2 – 20 (34%)	NR	NR
PFDS	1.4 – 50 (64%)	0.25 – 40 (77%)	0.007 – 11.5 (70%)
4:2 FTS	6 – 50 (42%)	NR	NA
6:2 FTS	0.5 – 100 (30%)	0.06 – 2.6 (66%)	0.06 – 7 (NA)
8:2 FTS	0.5 – 100 (52%)	NR	NR
GenX	3.1 (0%)	1.3 (47%)	NA

NR = Not reported

NA = Not available

4.0. Comparison of PFAS Concentrations in Groundwater to Other Media

4.1. Groundwater

The comparison of concentrations in groundwater to those in other media will focus on the two most-investigated PFAS, PFOA and PFOS. The maximum reported concentrations will be used in the analysis. The groundwater sites were split into three categories, no-known-source sites, primary-source sites, and secondary-source sites as noted above.

Concentration frequency plots were created based on 0.5-log unit concentration-interval bins for all three types of sites. They are presented in Figures SI1–3. The concentrations are observed to exhibit multi-modal distributions for the primary- and secondary-source sites. This distribution likely reflects in part the type of source and the nature of the release with which the sampled sites are associated. Conversely, the concentrations for the no-known-source sites exhibit an approximately negatively-skewed distribution for PFOS and normal distribution for PFOA. The ranges of concentrations likely reflect in part the locations of the sites with respect to urban versus rural areas.

The median values of the maximum groundwater concentrations reported for primary-source sites are approximately 45,700 and 7,400 ng/L for PFOS and PFOA, respectively. The maximum concentrations range up to 9.7 (9.7×10⁶ ng/L) and 6.6 mg/L (6.6×10⁶ ng/L) for PFOS and PFOA, respectively. The highest reported concentrations are associated with aqueous film-forming foams (AFFF) application sites. The median of the maximum concentrations reported for secondary-source sites are 95 and 96 ng/L for PFOS and PFOA, respectively. The median concentrations for the secondary-source sites are 2–3 orders-of-magnitude lower than those for the primary-source sites.

The median values of the maximum groundwater concentrations reported for the no-known-source sites are approximately 30 and 15 ng/L for PFOS and PFOA, respectively. These concentrations are several orders-of-magnitude lower than those for the primary source sites. Conversely, they are 3 and 6 times lower, respectively, than those reported for the secondary source sites. The highest concentrations reported for the no-known-source sites are generally associated with metropolitan areas. Hence, it is possible that some of the sites may have been impacted by sources either identified but located relatively far away or unidentified.

4.2. Surface Water

The aggregated surface-water studies used for the comparison of maximum PFOA and PFOS concentrations comprise a total of more than 1,700 sampling sites, with concentration data collected from >30 countries and regions. The sampled sites are divided into two categories, locations within or near an identified PFAS source and locations not located near an identified source. These two sets are referred to as “source” and “no-known-source” sites. Concentration frequency plots for both types of sites are presented in Figures SI4 and SI5, respectively. The concentrations are observed to exhibit a skewed distribution for the source sites, with one or two sites comprising the highest reported concentrations. This distribution is likely to be influenced by several factors, including the type of source, the nature of the release with which the sampled sites are associated, and the relative location of the sampling sites to the point of release. This information is not provided in all of the studies, so it is not possible to determine the impact of each of the factors. Conversely, the concentrations for the no-known-source sites exhibit a normal distribution. This likely reflects the more diffuse nature of the PFAS distributions in regions located away from identified sources, and the associated longer-range transport processes involved for these sites.

The source group consists of sites that include PFAS discharges from large-scale spills, firefighting training facilities, manufacturing facilities, airports, and wastewater treatment plants. The maximum PFOS concentrations range from 2.5 ng/L to 0.3 mg/L, with a median of 22 ng/L. The maximum PFOA concentrations range from approximately 8.8 ng/L to ~0.6 mg/L, with a median of 569 ng/L. The median PFOA concentration is more than 20-times larger than that of PFOS. This may result from the preponderance of industrial/manufacturing facilities in the regions with the highest PFAS concentrations, and the specific profile of PFAS associated with these facilities. For example, Heydebreck et al. (2015) reported high concentrations of PFOA within the Xiaoqing River in China, which is within an highly industrialized region and receives discharges from numerous industrial sources.

The reported maximum PFOS and PFOA concentrations for sites not located near an identified PFAS source were also compiled. The maximum PFOS concentrations range from approximately 0.3 ng/L to ~350 ng/L, with a median of 10 ng/L. The maximum PFOA concentrations range from 0.4 ng/L to ~1,000 ng/L, with a median of 17 ng/L. In contrast to the source sites, the median values are in a similar range for PFOS and PFOA for the no-known-source sites. Notably, the median of the maximum PFOS concentrations for the source sites is only a factor of two larger than the median of the maximum concentration for the no-known-source sites. Conversely, the PFOA median concentration is 30 times larger for the source sites compared to the no-known-source sites.

4.3. Marine Waters

PFAS concentrations in marine waters were reported for many oceanic and seaway water bodies around the world. The PFOS and PFOA maximum concentration frequency plots for the marine waters are presented in Figure SI6. The distributions are approximately positively skewed, with the greatest number of concentrations within a range of 0.01–0.3 ng/L. The median of the maximum PFOS concentrations is 0.2 ng/L, with the maximums ranging from 0.001 to ~760 ng/L. The maximum PFOA concentrations range from 0.001 to ~1,600 ng/L, with a median of 0.3 ng/L. The higher concentrations are generally associated with restricted and semi-closed seas, coastal shorelines, and bays adjacent to high vessel-traffic ports and high population/industrialized centers. Potential PFAS sources for the various marine sampling locations were noted in some studies. However, source information was generally lacking for most marine/oceanic/coastal PFAS surveys. The lower concentrations of PFAS present in open-sea regions were attributed primarily to atmospheric deposition, shipping routes, or general transport and dilution processes from distant sources. Conversely, the higher PFAS concentrations reported for coastal waters were generally attributed to surface runoff and river discharges.

4.4. Precipitation

Twenty studies were accessed that reported concentrations of PFAS in precipitation. The majority of the studies were conducted in China and North America, with a few reported for other locations including western Europe, Japan, and India. The sampling sites were located primarily within metropolitan areas. Most of the studies sampled rainwater, while a few sampled snow.

The median of the maximum PFOS concentrations in rainwater is ~3 ng/L, with the maximums ranging from 0.1 to 50 ng/L. The maximum PFOA concentrations for rainwater range from 0.07 to 121 ng/L, with a median of 10 ng/L. The maximum concentrations of PFOS and PFOA in snow range from ~0.5–545 and ~20–66 ng/L, respectively. These results demonstrate that PFAS are present in precipitation at significant levels. Additional studies are warranted to determine distributions in other regions of the world, with data sets differentiated by population densities, industrial development, and climate to delineate the impact of these factors on PFAS concentrations.

4.5. Soil

For the original meta-analysis of PFAS soil concentrations (Brusseau et al., 2020), the sites were divided into three categories, no-known-source sites (sites with no identified nearby PFAS source), primary-source sites (fire-training areas, manufacturing plants), and secondary-source sites (biosolids application, irrigation water use, landfills). This approach is continued for the updated analysis. The addition of the new studies did not significantly change the outcomes of the original study.

The frequency plots for PFOS and PFOA soil concentrations reported for primary source sites are presented in Figure SI7. Notably, significant separation of the two peaks of the distributions is observed, consistent with the two orders-of-magnitude difference in median values of the maximum concentrations. The median values of the maximum soil concentrations for primary-source sites are 4,000 and 58 μg/kg for PFOS and PFOA, respectively. The medians for secondary source sites are 350 and 38 μg/kg for PFOS and PFOA, respectively. In addition, PFOS and PFOA exhibit opposing skewed distributions. The disparity between the two distributions reflects in part the nature of the sources, with AFFF-impacted sources comprising the great majority of primary source sites (Brusseau et al., 2020).

The distribution for the no-known-source sites is slightly skewed (Figure SI8). This is likely a reflection of the location of the sampling sites with respect to surrounding land use and proximity to PFAS sources. The medians of the maximum soil concentrations for no-known-source sites are 2.7 and 3.1 μg/kg for PFOS and PFOA, respectively. The median PFOS maximum concentrations for the source sites are observed to be two and three orders-of-magnitude larger than the median no-known-source concentration. Conversely, they are approximately one order-of-magnitude larger for PFOA.

An important point to consider with respect to soil concentrations is the relationship between soil and soil porewater concentrations. Soil concentrations are the reporting standard for site investigations. However, there is growing interest in measuring and predicting porewater concentrations, as they mediate mass discharge to groundwater and typically represent the mass that is directly bioavailable (Anderson, 2021; Quinnan et al., 2021; Brusseau and Guo, 2022). Brusseau and Guo (2022) developed a comprehensive model to describe the distribution of PFAS within a soil sample and the relationship between total soil concentrations and soil porewater concentrations. Porewater concentrations are anticipated to range from ng/L to mg/L depending upon soil concentrations, which in turn depend upon the nature of the sites. The results of initial field and mathematical-modeling studies reflect these anticipated concentrations (Guo et al., 2020; Quinnan et al., 2021; Brusseau and Guo, 2022).

4.6. Comparison of Concentrations

Sample processing and analysis methods varied across the studies, as did the quantitative detection limits. Therefore, the data analysis conducted for the comparative assessments presented in this section focused on maximum reported concentrations. The comparisons are presented for PFOS and PFOA, the two most reported PFAS for groundwater. The maximum concentration reported in each study was used, irrespective of the number of sampling locations or number of samples collected. The data for surface waters and marine waters represent samples of the entire set of studies. In addition, the sampling locations are nonuniformly distributed as discussed previously. All of these factors may influence the representativeness of the data sets. This comparative analysis is therefore considered to represent an illustrative overview of the ranges of maximum concentrations reported for PFOS and PFOA in the various media. Source versus non-source site delineations were not used for the marine waters and precipitation data sets.

The maximum PFOS and PFOA concentrations reported for the various media are summarized in Tables S1 and S2 for source and no-known-source sites, respectively. Comparison of the values shows that soil concentrations are typically orders-of-magnitude higher than those reported for the other media. The median values of the maximum PFOS and PFOA soil concentrations are 4,000,000 and 58,000 ng/kg, respectively, for the primary-source sites. The median values for PFOS and PFOA, respectively, are approximately 45,700 and 7,400 ng/L for groundwater and 22 and 570 ng/L for surface water. The significantly higher concentrations observed for PFOS compared to PFOA for soil and groundwater is related to many of the primary-source sites comprising AFFF-impacted sites and that PFOS was a primary PFAS in AFFF.

The observation of PFAS concentrations in soil often being orders of magnitude greater than those in groundwater has been reported in prior studies (Anderson et al. 2019; Brusseau et al., 2020). The much higher concentrations generally reported for soils is expected for multiple reasons. First, soil is the initial point of PFAS release for many sites. Second, many PFAS experience significant magnitudes of retention in soil and the vadose zone (e.g., Brusseau, 2018; Guo et al., 2020). Third, the influence of dispersal and dilution processes are typically much less significant compared to surface and marine waters in particular. The PFOS and PFOA groundwater concentrations are three and one orders-of-magnitude higher, respectively, than those for surface water for the source sites. This disparity is likely related to differences in source concentrations and in the impacts related to dilution and dispersal processes.

The median values of the maximum PFOS and PFOA soil concentrations for no-known-source sites are 2,700 and 3,100 ng/kg, respectively. The no-known-source-site median values for PFOS and PFOA, respectively, are 30 and 15 ng/L for groundwater and 10 and 18 ng/L for surface water. The median values are ~3 and ~18 ng/L, respectively, for PFOS and PFOA in precipitation (combined rainwater and snow data). Similar to the source sites, the soil concentrations for the no-known-source sites are orders-of-magnitude higher than for the other media. Notably, the median values of the maximum PFOA concentrations in precipitation, surface water, and groundwater are very similar for the no-known-source sites. This result may in part reflect large-scale transport of PFOA throughout the hydrologic cycle. Further research is needed to determine if the results are merely coincidence or reflective of an underlying interconnection.

The median values of the maximum concentrations are 0.2 and 0.3 ng/L, respectively, for PFOS and PFOA in the marine waters. Overall, the concentrations of PFOS and PFOA reported for the marine waters are generally much lower than those reported for the other media. This disparity likely reflects the impact of a number of factors. It is probable that PFAS present in marine systems experience greater magnitudes of dilution and dispersal compared to the other systems. In addition, the concentrations of PFAS entering marine systems are anticipated to be generally lower than inputs to other media. For example, one source of PFAS release into marine waters is precipitation. As noted above, the concentrations reported for PFOS and PFOA in precipitation are in the ng/L range.

5.0. Conclusion

A comprehensive, quantitative analysis was conducted of published PFAS concentrations in groundwater. PFAS was determined to be present in groundwater across the globe. The reported concentrations of PFAS in groundwater ranged over many orders of magnitude, with the highest concentrations in part-per-million (ppm) levels. PFOS and PFOA have been the primary focus of many prior studies and are typically present in the highest or near-highest concentrations. However, a range of additional PFAS have been reported, the specific numbers of which vary greatly among the studies. The total number of PFAS reported is expected to continue to rise as analytical methods continue to be developed and refined. The concentrations of PFOA and PFOS in groundwater were compared to those reported in soils, surface waters, marine waters, and precipitation. Concentrations in soil are typically much greater than those in all other media.

A distinction between source and no-known-source sites was employed in this study to assist in characterizing PFAS distributions. Distinct differences in concentration ranges were observed between sites located within or near sources versus those that are not. The delineation of no-known-source sites in particular may in some cases be nebulous, considering that it is quite likely that there are many PFAS sources yet to be discovered. However, given the distinct differences in concentration ranges reported for the different source-type sites, it is important that source type be considered in future studies of PFAS distributions in environmental systems, and that sources be fully described when possible.

The presence of PFAS has been reported for all environmental media in all regions tested including areas that are far removed from specific PFAS sources. This indicates that PFAS are pervasive and widely distributed throughout the environment globally. This gives rise to the existence of a “background” concentration of PFAS. The presence of this background is a reflection of the large-scale use of PFAS, their general recalcitrance, and the action of long-range transport processes that distribute PFAS across regional and global scales. The importance of these background concentrations has been discussed in detail by Bryant et al. (2022). This background must be accounted for in both regional and site-specific risk assessments.

The term “environmentally relevant concentration” has been used in the scientific literature specifically in regard to the presence of PFAS in environmental media. In some cases, the term is used in reference to laboratory investigations of PFAS transport and fate behavior, with investigators for example stating that certain concentrations are “not environmentally relevant”. In other cases, it is employed in comparing PFAS concentration distributions to those of other types of contaminants. Based on the results of this study, PFOS, PFOA, and other PFAS are obviously present in soil and waters at concentrations ranging across many orders of magnitude, from part-per-trillion to ppm levels. Hence, using the term “environmentally relevant” without any context to designate a certain range of concentrations as being the sole representative “relevant” concentration lacks rigor. For example, while the mean concentrations reported for PFOS in soil and groundwater are in the part-per-billion range, ppm-level concentrations have been reported for some sites. These ppm levels are certainly relevant for the sites at which they are present. In addition, the use of this term without context fails to recognize the significant differences in concentration ranges typically observed for source versus no-known-source sites, and for the differences observed between different media. Based on this, it is evident that “environmentally relevant concentration” is a very ambiguous term that has the potential to be misused, and that its use without accompanying context does a disservice to the investigation and characterization of PFAS in the environment. A much preferred and more valid approach would be to recognize the very wide range of PFAS concentrations present in environmental media, and to consider the medium, site, and source-type specificity of the relevant concentrations.

The results presented in this study demonstrate the widespread presence of PFAS in both groundwater and surface water. A critical implication of this is the potential for PFAS to be present in potable water supplies originating from impacted sources. The presence of PFAS in drinking water has been illustrated in multiple studies (e.g., Hu et al., 2016; Guelfo and Adamson, 2018). Inspection of Tables 2 and SI1–2 reveals that the reported concentrations at many of the sites are greater than regulatory or advisory levels that have been implemented for PFAS in drinking water. For example, the median values of the maximum groundwater concentrations are 30 and 15 ng/L for PFOS and PFOA, respectively, for no-known-source sites. These concentrations are below the EPA’s lifetime health advisory (LHA) of 70 ng/L, but are above limits enacted in some U.S. states. For another example, the European Commission has proposed a limit of 100 ng/L for each of 16 individual PFAS in the 2018 revision of the EU Drinking Water Directive (EEA, 2022). The mean values of the reported concentrations are greater than this value for the great majority of the tabulated PFAS (Table 2). The ubiquitous distribution of PFAS in the global hydrosphere poses a threat to drinking-water quality and water-resource sustainability. A large fraction of the groundwater studies were concentrated in a few regions. Given the results of this study, it appears prudent that additional studies be conducted in regions that have not been sufficiently investigated to date.