Subscriber access provided by Kaohsiung Medical University
Characterization of Natural and Affected Environments
Widespread micropollutant monitoring in the Hudson River Estuary reveals spatiotemporal micropollutant clusters and their sources Corey Carpenter, and Damian E. Helbling Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.8b00945 • Publication Date (Web): 09 May 2018 Downloaded from http://pubs.acs.org on May 9, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 34
Environmental Science & Technology
1
Widespread micropollutant monitoring in the Hudson River Estuary
2
reveals spatiotemporal micropollutant clusters and their sources
3
Corey M. G. Carpenter and Damian E. Helbling*
4
School of Civil and Environmental Engineering, Cornell University, Ithaca, NY, USA
5
*Corresponding author:
[email protected], phone: +1 607 255 5146, fax: +1 607 255
6
9004
7 8
WORD COUNT. Title, abstract, and full manuscript (5865) + Figure 1 (300 words) + Figure 2
9
(300 words) + Figure 3 (300 words) + Figure 4 (300 words) = 7065 words
1 ACS Paragon Plus Environment
Environmental Science & Technology
10
Abstract
11
The objective of this study was to identify sources of micropollutants in the Hudson River
12
Estuary (HRE). We collected 127 grab samples at seventeen sites along the HRE over two years
13
and screened for up to 200 micropollutants. We quantified 168 of the micropollutants in at least
14
one of the samples. Atrazine, gabapentin, metolachlor, and sucralose were measured in every
15
sample. We used data-driven unsupervised methods to cluster the micropollutants based on their
16
spatiotemporal occurrence and normalized-concentration patterns. Three major clusters of
17
micropollutants were identified: ubiquitous and mixed-use (core micropollutants); sourced from
18
sewage treatment plant outfalls (STP micropollutants); and derived from diffuse upstream
19
sources (diffuse micropollutants). Each of these clusters was further refined into sub-clusters that
20
were linked to specific sources based on relationships identified through geospatial analysis of
21
watershed features. Evaluation of cumulative loadings of each sub-cluster revealed that the
22
Mohawk River and Rondout Creek are major contributors of most core micropollutants and STP
23
micropollutants and the upper HRE is a major contributor of diffuse micropollutants. These data
24
provide the first comprehensive evaluation of micropollutants in the HRE and define distinct
25
spatiotemporal micropollutant clusters that are linked to sources and conserved across surface
26
water systems around the world.
2 ACS Paragon Plus Environment
Page 2 of 34
Page 3 of 34
27
Environmental Science & Technology
Introduction
28
Data from monitoring studies have routinely confirmed the occurrence of 100s of organic
29
micropollutants in surface water systems around the world.1 The main targets of monitoring
30
studies have been pharmaceuticals,2 personal care products,3 illicit drugs,4 pesticides,5 industrial
31
chemicals,6 or other anthropogenic organic chemicals that have known or putative toxic effects
32
on aquatic ecosystems or exposed human populations.7–9 The potential sources of
33
micropollutants are varied, with much attention focused on sewage treatment plant (STP)
34
outfalls,3 combined sewer overflows,10 industrial discharges,11 stormwater outfalls,12 and diffuse
35
runoff from agricultural and urban landscapes,13 while many other potential sources are being
36
explored.14
37
Recently, long-term monitoring data characterizing micropollutant occurrence at the
38
watershed scale has been used to identify key insights into sources of micropollutants. For
39
example, mass balance and multivariate analyses revealed three types of micropollutant sources
40
in a Minnesota River including diffuse runoff, STP outfalls, and mixed pathways (diffuse runoff
41
and STP outfalls).15 Long-term longitudinal sampling along the Rhine River was used to identify
42
several previously unknown sources of micropollutants, particularly from tributaries and
43
industrial sources.16 A geospatial analysis of poly- and perfluoroalkyl substances (PFASs)
44
revealed that PFASs were found at higher concentrations in more urban areas and different types
45
of PFASs were associated with different point sources such as airports, textile mills, and metal
46
smelting.17 Lastly, the U.S. Geological Survey (USGS) conducted a national-scale
47
micropollutant monitoring survey and used a statistical approach to reveal significant
48
relationships between contaminant summary statistics and wastewater discharge and urban
49
development.18 These examples demonstrate powerful ways in which geospatial data can be
3 ACS Paragon Plus Environment
Environmental Science & Technology
50
combined with micropollutant occurrence data to improve our fundamental understanding of
51
micropollutant sources.
52
The primary goal of this research was to assess the relative contributions of various
53
sources of micropollutants in the Hudson River Estuary (HRE). The HRE provides drinking
54
water to more than 100,000 people as a surface water source and is an important waterway for
55
recreational and commercial activities. A recent study surveyed the occurrence of 16
56
pharmaceutical compounds in the HRE,19 but no previous study has combined a comprehensive
57
micropollutant screening with geospatial analyses to identify the relative contributions of various
58
sources of micropollutants in the HRE. We hypothesized that groups of micropollutants would
59
cluster together based on their spatiotemporal occurrence or concentration patterns, and that
60
those clusters would associate with specific upstream sources. To test this hypothesis, we
61
collected grab samples at seventeen sites along the HRE during the 2016 and 2017 recreational
62
seasons (May – October). Samples were analyzed to quantify the occurrence of up to 200
63
micropollutants identified in surface waters around the world. We used ArcGIS to develop maps
64
of the watershed that include geospatial references for likely micropollutant sources. We used
65
data-driven unsupervised methods to explore the complexity of micropollutant occurrence,
66
including hierarchical clustering to identify groups of micropollutants with similar
67
spatiotemporal occurrence and normalized-concentration patterns. We were able to categorize
68
the resulting micropollutant clusters based on their likely sources, link the clusters to various
69
geospatial features, and assess the relative contributions of specific sources and tributaries to
70
micropollutant occurrence in the HRE. We finally used a statistical approach to discover a
71
contamination event and identify micropollutants that are suitable indicators of overall
72
micropollutant occurrence and concentrations.
4 ACS Paragon Plus Environment
Page 4 of 34
Page 5 of 34
Environmental Science & Technology
73
Material and Methods
74
Study area. The HRE catchment area is a large mixed-use watershed located in eastern New
75
York State with an area of approximately 34,300 km2 and a population of over 2.5 million. A
76
map of the study area, the locations of seventeen sampling sites, and a delineation of tributary
77
watersheds is provided in Figure 1. Samples were collected from sites between the Mohawk
78
River and the Tappan Zee Bridge; specific sites are described in Table S1 of the Supporting
79
Information (SI).
80
Sample collection. The sample locations were selected to target STP outfalls and tributaries that
81
are expected to be major sources of micropollutants in the HRE. Grab samples were collected in
82
collaboration with Riverkeeper,20 an organization dedicated to monitoring and protecting the
83
waters of the HRE, during nine sampling events over the 2016 and 2017 recreational seasons
84
(see Table S2 for a complete list of the sampling dates and times). Samples were collected in 1L
85
amber, trace clean glass bottles and stored in an ice bath on the sampling vessel for up to three
86
days. The samples were then shipped on ice in a cooler to our laboratory at the end of each
87
sampling campaign, and stored at 4°C until sample preparation. The total sample holding time
88
prior to sample preparation and analysis was always between one and four days. A full
89
description of the sampling procedure is provided in the SI. The sample sites included three STP
90
outfalls (Orangetown, O_STP; West Point, W_STP; and Rondout Creek, R_STP), four sites at
91
the mouth of tributaries (Pocantico River, PR_M; Cedar Pond Brook, CB_M; Furnace Brook,
92
FB_M; and Annesville Creek, AC_M), eight sites inside tributaries (Rondout Creek: RC_U
93
(upstream), RC_D (downstream); Esopus Creek: EC_U, EC_D; Catskill Creek: CC_U, CC_D;
94
Normans Kill, NK; and the Mohawk River, MR), and two control sites that were sampled in the
95
mid-channel of the HRE at the northern (Upper Hudson River, UHR) and southern (Lower
5 ACS Paragon Plus Environment
Environmental Science & Technology
96
Hudson River, LHR) ends of the study boundaries. The STP outfall samples were collected from
97
the upwelling or directly adjacent to STP outfall pipes and thus contain a mixture of STP effluent
98
and river water. Two samples were lost during sample shipment, therefore a total of 127 samples
99
were processed and analyzed. It must be noted that data derived from grab samples do not
100
necessarily reflect the expected dynamics of micropollutant occurrence or concentration in
101
surface water systems.21 However, a series of grab samples can be analyzed to provide robust
102
estimates of the likelihood of occurrence and average concentrations of specific micropollutants
103
at a particular sample site. We further note that no field blanks were collected during this study,
104
though the sampling procedure was explicitly designed to limit contamination in the field. We
105
converted measured micropollutant concentrations to loads using river flow data obtained from
106
USGS stream gages under the assumption that the water columns were well-mixed.22 The lower
107
portion of the HRE is a partially-mixed estuary with significant vertical stratification.23
108
Therefore, we only estimated loads from samples collected inside of tributaries that are located
109
in the upper portion of the HRE where vertical stratification is not expected. USGS streamflow
110
rates are provided in Table S3.
111
Sample preparation and analysis. The samples were prepared using a mixed-bed solid-phase
112
extraction (SPE) method to concentrate the samples as previously described.24 We then used high
113
performance liquid chromatography (HPLC) coupled to high-resolution mass spectrometry (MS)
114
to quantify the occurrence of 200 diverse micropollutants which have been previously detected
115
or are likely to occur in surface waters.24,25 The analytical HPLC-MS/MS method was previously
116
developed and validated for a broad range of micropollutants.24,26 These methods are
117
summarized in the SI and the target micropollutants, their respective use-class, structure,
6 ACS Paragon Plus Environment
Page 6 of 34
Page 7 of 34
Environmental Science & Technology
118
physiochemical properties, analytical data, and limits of quantification (LOQs) are provided in
119
Tables S4-S5.
120
Geospatial analysis. Mapping and geospatial analyses were conducted in ArcGIS v10.4. We
121
used publically available data to produce maps of the HRE catchment area for geospatial
122
references including land cover (Figure S1) and STP outfalls (Figure S2). The geospatial data
123
sources are summarized in Table S7.
124
Statistical analysis. Statistical analyses were conducted using R Statistical Software v3.3.
125
Micropollutant clusters were determined using hierarchical clustering with the hclust function,
126
Ward’s agglomeration method, and either binary or Euclidean distance matrixes. Micropollutant
127
data were converted into binary occurrence data (91%) in STP
194
outfall samples and was defined as the STP exclusive sub-cluster of micropollutants. This sub-
195
cluster is also defined as containing micropollutants with the highest concentrations relative to
196
the other spatiotemporal occurrence clusters and sub-clusters (p0.05, paired WRS). This suggests that the sources of sub-
253
cluster A micropollutants, like the diffuse micropollutants cluster, cannot be exclusively
254
attributed to STP outfalls. We conclude that the micropollutants in sub-cluster A are
12 ACS Paragon Plus Environment
Page 12 of 34
Page 13 of 34
Environmental Science & Technology
255
predominantly used in agriculture or on urban landscapes and are attributed to diffuse upstream
256
sources.15,29
257
The normalized concentration patterns of sub-cluster A micropollutants also separated the
258
tributary sample sites into two main groups. The first group of tributary sample sites consists of
259
UHR, EC, and CC and the second group consists of RC, MR, NK, and the five samples sites
260
located in the lower HRE (LHR, PR_M, CB_M, FB_M, and AC_M). We examined the major
261
differences in land cover in the watersheds of these groups of tributaries. Our geospatial analysis
262
revealed that the watersheds of the former group has a significantly (p
