Subscriber access provided by Chalmers Library
Article
Differences in riverine and pond water dissolved organic matter composition and sources in Canadian High Arctic watersheds affected by active layer detachments Jun-Jian Wang, Melissa Lafrenière, Scott Lamoureux, Andre J Simpson, Yves Gélinas, and Myrna J Simpson Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b05506 • Publication Date (Web): 04 Jan 2018 Downloaded from http://pubs.acs.org on January 6, 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 free 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 accessible to all readers and 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.
Environmental Science & Technology 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 35
Environmental Science & Technology
Wang et al. Graphical Abstract
ACS Paragon Plus Environment
Environmental Science & Technology
Page 2 of 35
1
Differences in riverine and pond water dissolved organic matter composition and
2
sources in Canadian High Arctic watersheds affected by active layer detachments
3 4
Jun-Jian Wang†,#, Melissa J. Lafrenière‡, Scott F. Lamoureux‡, André J. Simpson†, Yves
5
Gélinas§, and Myrna J. Simpson†,*
6 7
†
8
University of Toronto Scarborough, 1265 Military Trail, Toronto, Ontario M1C 1A4,
9
Canada
Environmental NMR Centre and Department of Physical and Environmental Sciences,
10
‡
11
Kingston, Ontario K7L 3N6, Canada
12
§
13
7141 Sherbrooke West, Montréal, Quebec H4B 1R6, Canada
14
*Corresponding
15
[email protected] (M.J. Simpson)
16
#
17
Laboratory of Soil and Groundwater Pollution Control, Southern University of Science
18
and Technology of China, Shenzhen, Guangdong 518055, China
Department of Geography and Planning, Queen’s University, 68 University Ave.,
GEOTOP and the Department of Chemistry and Biochemistry, Concordia University,
Author:
Tel:
416-287-7234;
Fax:
416-287-7279;
e-mail:
Present address: School of Environmental Science and Engineering and Shenzhen Key
19
1 ACS Paragon Plus Environment
Page 3 of 35
Environmental Science & Technology
20
Abstract
21
Regional warming has caused permafrost thermokarst and disturbances, such as active
22
layer detachments (ALDs), which may alter carbon feedback in Arctic ecosystems.
23
However, it is currently unclear how these disturbances alter DOM biogeochemistry in
24
rivers and ponds in Arctic ecosystems. Water samples from the main river channel, ALD-
25
disturbed/undisturbed tributaries, and disturbed/undisturbed ponds within a catchment in
26
the Canadian High Arctic were collected and analyzed using carbon isotopes and
27
spectroscopic methods. Both river and pond samples had large variations in dissolved
28
organic carbon (DOC) concentrations. Ponds, particularly ALD-disturbed ponds, had
29
much older 14C DOC ages than rivers. Results from δ13C and absorption and fluorescence
30
analyses indicate higher autochthonous contributions in ponds than rivers and increasing
31
autochthonous contributions from upper to lower reaches of the main channel. The
32
disturbed samples had less carbohydrates but more carboxyl-rich alicyclic molecules in
33
1
34
samples also contained less terrestrial-humic-like but more oxidized-quinone-like
35
components in the fluorescence spectra. Interestingly, the disturbed pond DOM displayed
36
the greatest DOM oxidation with ALDs compared to undisturbed areas. Compared to
37
Arctic rivers, small Arctic ponds have DOM predominantly from permafrost and
38
microbial sources and may have a disproportionally stronger positive feedback on climate
39
warming.
H nuclear magnetic resonance spectra than undisturbed samples. These ALD-impacted
40
2 ACS Paragon Plus Environment
Environmental Science & Technology
41
Keywords: Permafrost thawing; Active layer detachments, Nuclear magnetic resonance
42
spectroscopy, Fluorescence spectroscopy, Parallel factor analysis PARAFAC, Cape
43
Bounty Arctic Watershed Observatory
3 ACS Paragon Plus Environment
Page 4 of 35
Page 5 of 35
Environmental Science & Technology
44
INTRODUCTION
45
The Arctic environment is a critical zone for biogeochemical processes because of
46
its high carbon storage (~1300 Pg; ~50% of global soil organic carbon1,2) and high
47
vulnerability to climate change.2-4 Variations in organic matter storage and cycling in
48
Arctic ecosystems are major sources of uncertainty in modeling of carbon, nutrient, and
49
pollutant cycling and climate projections on local to global scales.5-7 Dissolved organic
50
matter (DOM) represents one of the most mobile but the least characterized carbon pools
51
in Arctic environments because of the high complexity of DOM composition and
52
difficulty to access the Arctic region, particularly the High Arctic.8-10 A better
53
understanding of DOM chemistry and degradation and the controlling factors in Arctic
54
ecosystems will help predict global climate change feedback in the future.7-9,11,12
55
Previous studies have focused on the river, lake, and estuary DOM character and
56
explored the terrigenous carbon transport from Arctic watersheds to lakes, northern seas,
57
and the Arctic Ocean.6,13-16 Arctic rivers are estimated to transport 34–38 Tg yr−1 of
58
dissolved organic carbon (DOC) to the Arctic Ocean and surrounding basins,15 whereas
59
37–84 Tg yr−1 of DOC is delivered to inland waters but degraded to the atmosphere or
60
buried in lakes and streams before entering the ocean.17,18 Characteristics of Arctic river
61
DOM present strong spatial-temporal patterns due to varying DOM sources (e.g., soil and
62
plant materials, permafrost and glaciers, and microbes9,19) and varying degrees of
63
degradation.9,14,20 While large amounts of high molecular weight, relatively fresh, and
64
lignin-rich aromatic DOM was exported in spring freshets to rivers, relatively lower
65
concentrations of DOM with lower molecular weights, less aromaticity, and older ages
66
was exported later in the summer season.10,21-23 Importantly, increasingly severe glacier
4 ACS Paragon Plus Environment
Environmental Science & Technology
67
melt and permafrost thaw may reduce the DOC concentration (proxy for DOM
68
concentration) and specific ultraviolet absorbance at 254 nm (SUVA, proxy for DOM
69
aromaticity), and increase the
70
studies have shown the lack of a permafrost signal in DOC exported via major Arctic
71
rivers to the ocean, likely due to the high lability of the ancient permafrost-derived
72
DOM.26-29
14
C age of DOC in Arctic rivers.24,25 In contrast, some
73
Compared to rivers, Arctic ponds, which are typically a few square meters in area,
74
do not directly contribute to DOM flux from land to ocean, but disproportionally
75
contribute to atmospheric carbon fluxes and biodiversity within Arctic ecosystems.30-32
76
For example, their greenhouse gas emissions as a function of surface area are higher than
77
those of large lakes.33 In addition, ponds have been considered as excellent indicators of
78
climate change because of their hydrological isolation, small water volumes, and large
79
surface area-to-depth ratios.34,35 Significant declines in Arctic pond area and abundance
80
are occurring.30,31 Although the ponds account for about half of the total surface water in
81
permafrost landscapes,36 DOM biogeochemistry within these small ponds has not been
82
studied to the same extent as rivers.32,37 Moreover, most current knowledge of Arctic
83
DOM characteristics has been derived from the Sub-Arctic and Low Arctic environments
84
and few from the High Arctic that has shorter growing seasons, cooler summers, and
85
reduced diversity of flora and fauna.10,14,37,38 Whether and how the DOM characteristics,
86
and their responses to climate change, differ between the rivers and ponds remain
87
unknown in the High Arctic, which is a region subject to increasing warming and
88
precipitation and higher risks of permafrost degradation and disturbance.39
5 ACS Paragon Plus Environment
Page 6 of 35
Page 7 of 35
Environmental Science & Technology
89
As a representative type of permafrost disturbance, active layer detachments
90
(ALDs) are slope failures that displace soil along the upper permafrost boundary due to
91
active layer thickening and ground ice thawing, and are increasing in frequency and
92
severity in Canadian High Arctic.40,41 Such disturbance can alter soil organic matter
93
characteristics by increasing the release of labile compounds and stimulating microbial
94
degradation of previously unavailable soil organic matter.42-44 These processes are
95
expected to alter the terrestrial carbon export to downstream waterbodies. For example,
96
the DOC export was found higher in an ALD-disturbed catchment than an undisturbed
97
catchment in the Canadian High Arctic 2 years post-disturbance.21 Also, DOM in streams,
98
rivers and lakes in recently disturbed catchments appeared to contain more low molecular
99
weight fluorescent components10 and showed higher biological and photochemical
100
lability than those from undisturbed systems.14 Despite this progress, the relatively long-
101
term (>5 years) impact of ALDs (or impact of “historic” ALDs in contrast to the “recent”
102
ALDs (within 2-3 years)) on DOM characteristics is still unclear. To understand the
103
chemical characteristics between pond and river DOM and their responses to ALDs, this
104
study characterized the molecular-level DOM chemistry from river and pond waters in a
105
previously ALD-disturbed catchment using detailed isotopic and spectroscopic analyses.
106
The West River watershed at the Cape Bounty Arctic Watershed Observatory (CBAWO),
107
Nunavut, Canada was selected because of intensive ALD activity in 2007-08.40
108 109
MATERIALS AND METHODS
110
Site information. Water samples from 14 river and 6 pond sites were collected in early
111
August of 2014 and 2015 within the CBAWO, southern Melville Island, Nunavut,
6 ACS Paragon Plus Environment
Environmental Science & Technology
112
Canada (74°54’N, 109°35’W; Fig. 1; Fig. S1; Table S1). This region has a High Arctic
113
climate with a thaw period between June and August. The mean monthly January and
114
July air temperatures at the nearest long-term weather station, Mould Bay (200 km to the
115
west) are -33.1 °C and 4.0°C, respectively.45 The mean annual precipitation recorded at
116
Mould Bay includes 102 mm of snowfall and 16.3 mm of rainfall.46 The 14 river sites
117
included 7 main channel sites (from upper to lower: WR-1 to WR-7), 5 undisturbed
118
tributary sites (WRE-1, WRE-2, WRW-1, WRW-2, and Goose), and 2 ALD-disturbed
119
tributary sites (Ptarmigan-D and ALD-D; Fig. 1; Fig. S1; Table S1). The 6 small (< 30 m2)
120
ephemeral ponds included 4 undisturbed ponds (Pond-2, Pond-4, Pond-5, and Pond-6)
121
and 2 disturbed ponds (Pond-1-D and Pond-3-D; Fig. 1; Fig. S1; Table S1). More
122
information about the CBAWO and the sites is available in the supplementary
123
information.
124 125
Sample analyses. The pH and electrical conductivity of water were measured in the field
126
using Orion Star A215 benchtop meter, with the Orion Ross Ultra pH and DuraProbe
127
conductivity probes. Surface water samples were collected at a depth of 10-20 cm,
128
filtered through precombusted 0.7 µm glass fiber filters, and transported to the lab at 4 oC
129
for analyses of DOC, radiocarbon, and optical properties. Samples for radiocarbon were
130
collected in pre-cleaned certified 500 mL glass amber bottles. Sample aliquots destined
131
for δ13C and 14C analyses were acidified to a pH of 2 in the field, using trace metal grade
132
HCl. Samples for nuclear magnetic resonance (NMR) analyses were collected in 1 L
133
amber HDPE bottles, frozen immediately, and transported to the University of Toronto.
134
The concentrations of DOC and total dissolved nitrogen (TDN) were quantified using a
7 ACS Paragon Plus Environment
Page 8 of 35
Page 9 of 35
Environmental Science & Technology
135
Shimadzu (TOC-V) organic carbon analyzer with a nitrogen detection unit (Shimadzu
136
TMN-1; Kyoto, Japan).10
137
Radiocarbon analyses of the DOC were performed on a 3 MV tandem accelerator
138
mass spectrometer (AMS) at the A.E. Lalonde AMS Laboratory (Ottawa, Canada). After
139
removing the inorganic carbon using 85% phosphoric acid, DOC was oxidized to CO2
140
using a sodium persulfate wet oxidation technique.47,48 The CO2 was converted to
141
elemental C using semi-automated graphitization lines in the presence of hydrogen. The
142
fraction of modern carbon, F14C, is calculated as the ratio of sample
143
standard
144
years. Sample and standard
145
measured 13C/12C ratio.50 Ages are in years before present (years BP; CE 1950).51,52 The
146
stable isotope composition was measured at the Concordia University, using high
147
temperature combustion and continuous flow IRMS, and reported as δ13C (in ‰), relative
148
to the Pee Dee Belemnite (PDB) reference standard according to methods detailed in
149
Lalonde et al.53
14
C/12C using Oxalic II as a reference,49 and the Libby 14
14
14
C/12C and the
C half-life of 5568
C/12C ratios are background-corrected using the AMS-
150
For solution-state NMR analysis, 1 liter of each water sample was filtered, frozen,
151
freeze-dried and dried under vacuum over P2O5. Depending on the abundance of
152
dissolved components, 4-40 mg of dried sample was reconstituted in 60 µL NaOD/D2O
153
solution (pH=12), centrifuged to remove insoluble materials, and transferred into Bruker
154
BioSpin 1.7 mm NMR tubes.54 The analyses were conducted using a Bruker BioSpin
155
Avance III 500 MHz NMR spectrometer (Karlsruhe, Germany) equipped with a 1H-15N-
156
13
157
one-dimensional 1H NMR spectra were collected using presaturation utilizing relaxation
C TXI 1.7 mm microprobe fitted with an actively shielded Z gradient.54 Specifically,
8 ACS Paragon Plus Environment
Environmental Science & Technology
158
gradients and echoes to suppress the resonances from residual water.55 A total of 256
159
scans were collected using a recycle delay of 2 s, and 32K time domain points. Spectra
160
were processed using a zero filling factor of 2 and were apodized by multiplication with
161
an exponential decay corresponding to 0.3 Hz line broadening. Diffusion-edited 1H NMR
162
spectra were collected using a bipolar pulse longitudinal encode–decode sequence.56 A
163
total of 2048 scans were collected using a 2.5 ms, 53.5 gauss cm−1, sine-shaped gradient
164
pulse, a diffusion time of 200 ms, 16K time domain points and 10 Hz line broadening.
165
The diffusion edited 1H NMR experiment highlights signals that do not move position in
166
the NMR tube.57 In this study, the diffusion filter emphasizes molecules that do not move
167
more that (~ 1 µm) within the sample over the diffusion time of 200 ms58 and as such will
168
emphasize molecules with little translation motion including macromolecular and
169
aggregated compounds. The one-dimensional 1H NMR spectra were integrated into four
170
regions: 1) materials derived from linear terpenoids (MDLT), 0.6–1.6 ppm; 2) carboxyl-
171
rich alicyclic molecules (CRAM), 1.6–3.2 ppm; 3) carbohydrates, 3.2–4.5 ppm; and 4)
172
aromatics and phenolics, 6.5–8.4 ppm.14,59 The percentages for the four regions were
173
calculated to indicate the relative abundances of protons (99.98% natural abundance of all
174
H isotopes) in different forms of molecular groups. Therefore, the percentages represent
175
the total 1H signal from specific molecular groups found within DOM. For example,
176
although the aromatic and phenolic constituents accounted for 2%-6% of total 1H NMR
177
resonance, the aromatic and phenolic compounds account for higher absolute abundance
178
in DOM than 2%-6% because the aromatic and phenolic compounds have low H/C ratio.
179
Moreover, to provide better spectral dispersion, two-dimensional 1H-13C heteronuclear
180
single quantum coherence (HSQC) spectra were obtained in digital quadrature detection
9 ACS Paragon Plus Environment
Page 10 of 35
Page 11 of 35
Environmental Science & Technology
181
mode with echo–antiecho gradient selection. A total of 1200 scans were collected with
182
2K and 96 time domain increments in the F2 and F1 dimensions respectively. HSQC
183
spectra were processed using a function corresponding to 25 Hz line broadening in F2
184
and a sine-squared function shifted by 90o in F1 with a zero-filling factor of 2 in both
185
dimensions.
186
Ultraviolet-visible absorbance and fluorescence emission excitation matrices
187
(EEMs) of the filtered samples were collected at room temperature using a Horiba
188
Aqualog spectrometer (NJ, USA) as previously described.10 All EEMs were corrected for
189
blanks and inner filter effects and calibrated to a quinone sulfate unit.60 Optical indices
190
including specific ultraviolet absorbance at 254 nm (SUVA), E2/E3 ratio, fluorescence
191
index (FI), biological index (BIX), and humification index (HIX) were calculated.61,62
192
Specifically, the SUVA is indicative of DOM aromaticity and was calculated as
193
ultraviolet absorbance at 254 nm divided by DOC concentration.63 E2/E3 ratio was
194
calculated as absorbance ratio of 254 nm to 365 nm, and higher E2/E3 ratio indicates
195
lower molecular size of DOM.64 Note that the accuracy of SUVA and E2/E3 ratios as
196
chemical proxies can be impacted by colloidal material65 or dissolved ions such as iron
197
(III),66 which were not measured because of the limited amount of samples. In addition,
198
the aromaticity reflected by SUVA and the relative abundance of aromatic and phenolic
199
constituents in the 1H NMR signal may not be directly comparable because the former
200
estimates the overall aromatic and potential interfering light absorbance at 254 nm and
201
the latter measures the 1H NMR resonance (instead of carbon resonance) in aromatic and
202
phenolic molecular groups. The fluorescence index (FI), the signal ratio of emission 470
203
nm to 520 nm at excitation 370 nm, is an index to differentiate microbial or terrestrial
10 ACS Paragon Plus Environment
Environmental Science & Technology
204
origins (~1.8 for microbial origin and ~1.2 for terrestrial origin).67 The biological index
205
(BIX), an index for the contribution of recently produced autochthonous DOM, was
206
calculated as the signal ratio of emission at 380 nm divided by the maximum emission
207
between 420 and 435 nm for excitation at 310 nm.68 The humification index (HIX), an
208
index of humic substance content, was determined as the area under the emission range
209
435–480 nm divided by that under the emission range 300–345 nm, for excitation at 254
210
nm.69 The fluorescent DOM components for the 20 EEM data were determined using
211
modelling of the parallel factor analysis (PARAFAC)70 and validation of split-half
212
analysis.71 The components were compared to published models using the OpenFluor
213
database.72 Specifically, signals with excitation wavelengths lower than 260 nm, emission
214
wavelengths lower than 300 nm, and those influenced by 1st and 2nd order scattering were
215
removed from the matrices to avoid inaccurate noises.10,71 A five-component model
216
explaining 99.0% of the fluorescence variability was selected.
217 218
Statistical analyses. The significant differences in chemical parameters between river
219
and pond waters (n = 14 and 6) and among main channel, tributary, and pond waters (n =
220
7, 7, and 6, respectively) were examined using one-way analyses of variances (ANOVA)
221
with Tukey post hoc test. Several parameters including electrical conductivity, DOC, and
222
TDN were logarithmically transformed to satisfy the normality assumption of ANOVA.
223
When any assumption of parametric ANOVA (e.g., no outlier, normally-disturbed data,
224
and homogeneity of variance) is violated, the nonparametric Kruskal-Wallis test (or
225
“ANOVA on ranks”) and Dunn’s post hoc multiple comparison with a Bonferroni
226
correction was used instead to compare differences between water source groups. To
11 ACS Paragon Plus Environment
Page 12 of 35
Page 13 of 35
Environmental Science & Technology
227
examine the main effects of ALDs and their possible interactions with impacts of water
228
source (tributary versus pond) on the water chemistry, two-way ANOVA was used on all
229
undisturbed and disturbed tributary and pond waters (n=13). Data related to the general
230
water quality, optical properties, and DOM chemistry were standardized and analyzed
231
using principal component analysis (PCA). Note that the small sample size did not satisfy
232
the PCA requirement and will limit the accuracy of analysis, however, we still use it as an
233
exploratory tool to provide a rough overview of the reduced dimensions of variables and
234
sample clustering. The loading of different parameters on the first two principal
235
components that explained 61.5% of the total variance and the relevant component scores
236
of each water samples were calculated. SPSS 15.0 (IL, USA) was used for all statistical
237
analyses and the level of significance was set at 0.05.
238 239
RESULTS AND DISCUSSION
240
General water quality. The pH, EC, DOC and TDN concentrations differed
241
significantly between the river and pond waters (Table 1; Table S2). Specifically, both
242
the pH and EC values were lower in rivers (pH: 6.2-7.5; EC: 201-361 µS/cm) than ponds
243
(pH: 7.5-8.8; EC: 1104-4310 µS/cm). These pH and EC values are mostly comparable to
244
those previously reported for High Arctic rivers and ponds.73-75 However, the relatively
245
high EC values of the ponds are likely due to water evaporation in summer and the
246
surface and subsurface transfer of inorganic solutes.40,76 The DOC concentration of river
247
waters ranged from 1.1 to 5.9 mg/L and that of the pond waters varied highly from 1.9 to
248
51.3 mg/L. Similarly, the TDN concentration ranged from 0.06 to 0.29 mg/L for river
249
samples and varied markedly from 0.21 to 8.92 mg/L for pond waters. The DOC and
12 ACS Paragon Plus Environment
Environmental Science & Technology
250
TDN concentrations of river waters were similar to the base-flow concentrations of
251
previous years reported10 and also comparable to other Arctic rivers.6 The relatively high
252
DOC and TDN concentrations in most ponds support high microbial activity.32 However,
253
their large variations indicate that the ponds may have various carbon and nutrient inputs
254
and sinks, and different degrees of DOM concentration (e.g., evaporation) and dilution
255
(e.g., by melting ice). In addition, the DOC/TDN ratio was significantly lower in ponds
256
(0.2-18.2) than in river waters (15.6-27.7). As the DOC/TDN ratio commonly decreases
257
with increasing degradation of overall DOM and increasing autochthonous contributions
258
in natural water,77 the lower values in ponds suggest higher contributions from
259
autochthonous sources.
260
Within the river samples, there was no significant difference in pH, EC, DOC or
261
TDN concentrations between the main channel and tributary samples or between the
262
undisturbed and ALD-disturbed tributary waters (P>0.05). However, from WR-1 to WR-
263
7, significant increases in DOC concentrations and DOC/TDN ratios were observed
264
(Table S2), likely due to increasing autotrophic primary production (more detail in the
265
allochthonous versus autochthonous sources section). Within the pond samples, the
266
disturbed ponds showed much lower DOC concentrations (1.87 and 2.44 mg/L) than
267
undisturbed ponds (4.78-51.3 mg/L). One possibility causing low DOC concentrations in
268
disturbed ponds could be the higher degree of DOM degradation stimulated by the labile
269
permafrost inputs.14 Another possibility could be that the ALDs resulted in the
270
translocation of surface soils to areas downslope, leaving previously buried mineral soils
271
at the surface42 and thus reducing the pool of soil organic matter and inputs of DOM
272
from surrounding catchment area.
13 ACS Paragon Plus Environment
Page 14 of 35
Page 15 of 35
Environmental Science & Technology
273 14
274
14
275
1028 to 4975 years BP, much older than those of the river waters (≤ 440 years BP;
276
P