Subscriber access provided by READING UNIV
Article
Riverine export of aged carbon driven by flow path depth and residence time Rebecca T Barnes, David Ellison Butman, Henry Wilson, and Peter A. Raymond Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b04717 • Publication Date (Web): 09 Jan 2018 Downloaded from http://pubs.acs.org on January 14, 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 25
1
Environmental Science & Technology
Riverine export of aged carbon driven by flow path depth and residence time
2 3
Rebecca T. Barnes1*, David E. Butman2, Henry F. Wilson3, Peter A. Raymond4
4 5
1
Environmental Program, Colorado College, Colorado Springs, CO 80903, USA
6
2
School of Environmental and Forest Sciences, University of Washington, Seattle, WA 98195,
7
USA
8
3
9
Canada
Brandon Research Centre, Agriculture and Agri-Food Canada, Brandon, Manitoba R7A 5Y3,
10
4
11
*corresponding author:
[email protected] School of Forestry & Environmental Sciences, Yale University, New Haven, CT 06115, USA
12
ACS Paragon Plus Environment
Environmental Science & Technology
13
Abstract
14
The flux of terrestrial C to rivers has increased relative to pre-industrial levels, a fraction of
15
which is aged dissolved organic C (DOC). In rivers, C is stored in sediments, exported to the
16
ocean, or (bio)chemically processed and released as CO2. Disturbance changes land cover and
17
hydrology, shifting potential sources and processing of DOC. To investigate the likely sources of
18
aged DOC, we analyzed radiocarbon ages, chemical, and spectral properties of DOC and major
19
ions from nineteen rivers draining the coterminous U.S. and Arctic. DOC optics indicated that
20
the majority is exported as aromatic, high molecular weight, modern molecules while aged DOC
21
tended to consist of smaller, microbial degradation products. Aged DOC exports, observed
22
regularly in arid basins and during base flow in arctic rivers, are associated with higher
23
proportion of mineral weathering products, suggesting deeper flows paths. These patterns also
24
indicate potential for production of microbial byproducts as DOC ages in soil and water with
25
longer periods of time between production and transport. Thus, changes in hydrology associated
26
with landscape alteration (e.g. tilling or shifting climates) that can result in deeper flow paths or
27
longer residence times will likely lead to a greater proportion of aged carbon in riverine exports.
28 29
ACS Paragon Plus Environment
Page 2 of 25
Page 3 of 25
30 31
Environmental Science & Technology
Introduction To predict how the carbon (C) cycle will respond to global change drivers, we need to
32
understand how organic matter pools at the land-water interface will shift in response to drivers
33
such as altered precipitation and increased agricultural development1. The terrestrial landscape
34
exports approximately 2.7 Pg C yr-1 to aquatic systems2, 3, a value that has increased by ~0.1 –
35
0.2 Pg C yr-1 due to anthropogenic activities3. Inland waters both process and transport organic
36
matter, storing C in sediment4 and releasing it to the atmosphere as CO2 and CH42, 5-7. The fate of
37
organic carbon within inland waters is largely determined by its chemical composition and the
38
hydrology of the system. For example, the chemical composition of dissolved organic matter will
39
determine, in part, how susceptible it is to microbial metabolism8 and UV oxidation9. Further,
40
studies have illustrated that the majority of annual dissolved organic C (DOC) export occurs
41
during high flow events, in both temperate10, tropical11, and arctic12 systems. During these
42
periods of high flow, it is likely that shorter transit times lead to increased export of less
43
processed DOC to the coastal ocean.
44
Stream DOC is dominated by terrestrial sources in most systems13 and thus if the age and
45
nature of the DOC varies between systems or through time it follows that DOC source regions
46
(or processing) are different or shifting. The majority of this DOC is modern in age, reflecting its
47
dominant sources: terrestrial vegetation and surface soils14. However, rivers also export aged
48
DOC, organic carbon that is stored in terrestrial sinks, including: shales15, peatlands16,
49
permafrost17, as well as non-terrestrial sources such as precipitation18, 19 and petroleum products
50
ranging from soaps to motor oil20. A recent synthesis illustrated that human landscape
51
disturbance, i.e. urban and agricultural development, is associated with the export of aged carbon
52
from streams and rivers21. Further, several studies have documented relationships between flow
ACS Paragon Plus Environment
Environmental Science & Technology
53
conditions and DOC age; though some systems export older carbon during baseflow17 and others
54
during high flow22, 23. This aged carbon, once thought recalcitrant, is biologically available
55
within aquatic systems 24-26, fueling CO2 production and evasion. Given that shifts in
56
precipitation are expected to be a dominant driver in future riverine C fluxes27, 28, it is critical that
57
we understand how this changing hydrology will affect C source areas, export and, cycling.
58
Changes in hydrology are likely to affect flow paths and thus the nature and amount of
59
carbon exported. Kaiser and Kalbitz29 describe a conceptual model that explains the vertical
60
profile observations of soil organic matter (SOM) characteristics. Several studies have
61
documented an increase in the age of SOM with depth30 which cannot be explained by the simple
62
leaching of organic matter decomposition products in surface horizons to depth31. In addition to
63
the physio-chemical stripping of dissolved organic matter (DOM) components (fractional
64
sorption and co-precipitation32), DOM undergoes microbial processing during transport29. Thus
65
given that exported DOM reflects both the sources and cycling of SOM33, riverine chemistry
66
should reflect differences in flow path depth. For example, surficial and shallow flow paths will
67
export large amounts of modern, plant-derived DOM, while deeper flows paths are more likely
68
to have an aged, microbial DOM signature.
69
Deeper portions of the soil column will likely have a greater proportion of organic matter
70
protected via organo-mineral associations, which are key to retaining C within soils34, 35 as
71
organo-mineral associations (cation bridging, ligand exchange, cation-anion exchange, hydrogen
72
bonds, Van der Waal forces, etc.) are more important in predicting stability and turnover of SOM
73
than the molecular structure of the OM35, 36. Tipping et al.37 found that SOM associated with
74
minerals had an average residence time of 100-200 years, while SOM not associated with
75
minerals turned over in 20-30 years. As such, it follows that older DOC should be exported with
ACS Paragon Plus Environment
Page 4 of 25
Page 5 of 25
Environmental Science & Technology
76
relatively greater mineral weathering products (e.g. base cations, Si). In addition, many microbial
77
byproducts are non-aromatic and recalcitrant38 and when in the deeper portion of the soil more
78
likely protected from further decomposition, increasing the likelihood that these molecules will
79
age in place. Weathering and pedogenic processes stratify soils structure into distinct vertical
80
horizons, that create large ranges of physical/chemical environments experiences by microbes.
81
Overtime, changes in soil mineralogy and the release of metal cations like Fe and Al, produce
82
binding sites on the remaining organo-mineral complexes that can control the stabilization of soil
83
organic matter, as well as alter the quality and quantity of DOM. The alteration of soil
84
mineralogy and the release of Fe and Al over time produce binding sites on mineral surfaces and
85
organo-mineral complexes that not only stabilize SOM but alter the quality and quantity of DOM
86
remaining in soil39-41. For example, reducing conditions in soil can result in the export of Fe2+
87
and DOC to streams42. As water moves through this portion of the soil it flushes both particulates
88
and dissolved constituents into ground and/or surface waters, where additional OM may be
89
released from particles43.
90
Given that multiple factors of global change (climate, land cover and land use, etc.)
91
directly and indirectly shift flow paths, and thus the nature of carbon exports, we examine the
92
spatiotemporal variability in the stoichiometry of DOC and mineral weathering products (base
93
cations) within large river basins draining the coterminous United States and Arctic. Given the
94
conceptual model of Kaiser and Kalbitz29, we hypothesize that more labile, older DOC will be
95
exported with relatively greater amounts of weathering products (i.e. a low DOC to base cation
96
ratio). These exports will be associated with more mineral rich and/or deeper flow paths and
97
longer residence times associated with arid climates44, temporal changes in water routing over
98
the hydrograph (e.g. baseflow17), or alteration to the landscape (e.g. agricultural liming45).
ACS Paragon Plus Environment
Environmental Science & Technology
99 100
Materials & Methods To examine these questions, we examined the riverine export of DOC to coastal waters
101 102
from thirteen major watersheds of the U.S.44 and the six largest Arctic watersheds that comprise
103
the Arctic Great Rivers Observatory (Arctic-GRO). The U.S. watersheds are monitored as part of
104
the National Stream Quality Accounting Network (NASQAN) program and in 2009 additional
105
DOC quality and age were measured on 6-11 of the monthly samples. This one year of data,
106
describing the coterminous U.S. watersheds, allows us to examine how large spatial variability
107
(different climates, ecosystems, levels of development) shifts dissolved weathering product and
108
organic carbon fluxes. The Arctic-GRO data (2003-2016) provides a way to examine six systems
109
with strong seasonal shifts in hydrology (i.e. freshet vs. baseflow) draining relatively organic rich
110
soils.
111
The U.S. watershed data describe samples collected across the 2009 hydrograph from 13
112
rivers that drain 79% of the land area and make up 90% of the freshwater discharge in the
113
coterminous U.S.44. Watersheds range in size from 2.9 million km2 (Mississippi) to 29,973 km2
114
(Potomac) and encompass a range of climate regimes with average precipitation ranging from
115
>1400 mm yr-1 in southeastern U.S. watersheds to less than 350 mm yr-1 in the high plains of the
116
Colorado River basin. As such, water yield varies significantly across these systems from
117
monthly lows of 1 mm yr-1 in the Rio Grande (Texas) to highs of 2615 mm yr-1 in the Altamaha
118
River (Georgia). Additional information (e.g. land use, population density) about these sites is
119
available in Tables 1a-b in Butman et al.44
120 121
The six Arctic watersheds (Yenisey, Lena, Ob’, Mackenzie, Yukon, Kolyma) comprise more than 50% of the land area draining to the Arctic Ocean, suggesting that differences in
ACS Paragon Plus Environment
Page 6 of 25
Page 7 of 25
Environmental Science & Technology
122
permafrost extent, microbial community, soil characteristics, and bedrock weathering amongst
123
the basins should be reflected in this dataset. Five of the six basins’ headwaters are within boreal
124
forest, draining a tundra dominated landscape before emptying into the Arctic Ocean. In contrast,
125
the Yukon River follows a different trajectory, with its headwaters in tundra, moving south into a
126
forest dominated landscape. Continuous permafrost coverage varies significantly between
127
systems with the Kolyma almost entirely underlain with permafrost (99%) and the Ob’ having
128
just 1% of its area occupied by permafrost46. In all six systems, water and constituent fluxes
129
associated with organic matter are greatest during the spring freshet12. There is significant
130
variability in the strength of the seasonality in discharge and DOC fluxes, with 50% or greater of
131
annual DOC flux occurring during this ~two month period of time in the Yenisey, Lena,
132
Kolyma, and Yukon while less than 30% of the annual total occurred during this period of time
133
in the Ob’ and the Mackenzie46. The average water yield at the time of sampling ranges from 210
134
± 120 mm yr-1 (Mackenzie) to 460 ± 460 mm yr-1 (Lena) with relatively larger seasonal
135
variability than the NASQAN rivers. The fourteen years of sampling across the hydrograph in
136
these systems, provides a robust way to compare spatial variation amongst the catchments as
137
well as how seasonal shifts in hydrology affect the nature and concentration of dissolved carbon
138
and base cation exports.
139
Sampling protocols were similar between projects allowing us to compare the data.
140
Sampling of these large rivers consisted of integrated grab samples occurring at gauged sites or
141
near gauging stations (in most cases maintained by the USGS), providing concomitant discharge
142
and water quality measurements. The NASQAN and majority of Arctic (2002-2011) samples
143
represent a composite of depth integrated samples from multiple points across the channel, while
ACS Paragon Plus Environment
Environmental Science & Technology
144
more recent sampling efforts in the Arctic (2012-2016) represent surface samples from multiple
145
points across the channel.
146
Sample processing for carbon constituents was similar (all samples filtered through 0.7
147
µm pre-combusted GF/F filters) and for the most part conducted in the same group of labs at the
148
USGS, Yale University, and Woods Hole. There were a few differences in the processing or
149
handling of samples. While all samples were filtered through the same filters, the NASQAN
150
samples were shipped unfiltered on ice within 24 hours of sampling and immediately filtered
151
upon arrival, while the Arctic samples were filtered in the field and then shipped. The largest
152
difference in carbon constituent analysis is the measurement of DOC concentrations. DOC
153
concentration for NASQAN rivers was measured using the persulfate wet oxidation on an OI
154
Analytical Model 700 TOC Analyzer47, while Arctic samples were analyzed using a Shimadzu
155
TOC analyzer12. All carbon quality measurements were made consistently across the two studies.
156
Briefly, radiocarbon isotope (∆14C-DOC) measurements determined the average age of DOC
157
exported and all samples were processed using established methods14, 48, which involve oxidizing
158
the DOC with UV light, converting it to CO2 which was then trapped and cryogenically purified.
159
The sample was then sent to the National Ocean Sciences Accelerator Mass Spectrometry
160
(NOSAMS) for isotopic analysis. The specific UV absorbance at 254 nm (SUVA254) was used as
161
a proxy for the aromaticity of dissolved organic matter, therefore sample absorbance at 254 nm
162
was measured using a UV-Vis Spectrometer and then normalized to DOC concentrations and
163
reported in L/(mg C*m) 49. Organic acid fractions were chromatographically separated using
164
columns packed with Amberlite™ XAD-8 and XAD-4 resins47. These operationally defined
165
dissolved organic matter fractions: larger molecular weight hydrophobic organic acid (HPOA),
166
smaller molecular weight hydrophilic molecules (HPI), and transphilic acids (TPIA)47 can
ACS Paragon Plus Environment
Page 8 of 25
Page 9 of 25
Environmental Science & Technology
167
provide information about carbon quality for reactivity, bacterial mineralization, and
168
photodegradation. The sum of the four major base cations (Ca2+, Na+, K+, Mg2+), as reported in
169
meq L-1, was used as a proxy for mineral soil weathering. In the Arctic river systems, water
170
isotopic (δ18O-H2O) measurements were used as proxy for shifts in flow paths, given systematic
171
seasonal shifts of the system (i.e. freshet always has more depleted δ18O values as compared to
172
later season base flow samples)50. For a more detailed field and lab protocol descriptions
173
associated with Arctic-GRO please see: carbon12, major ion chemistry51, and water isotope
174
analyses50. Carbon sampling protocols for the NASQAN sites are given in Butman et al.44 and
175
major ions and discharge measurements were downloaded from the USGS National Water
176
Inventory System (www.usgs.gov/nwis).
177
All statistical analyses were conducted in R (3.3.2, 2016 R Foundation for Statistical
178
Modeling) and variables were transformed to meet requirements of normality. For example, the
179
ratio of DOC to the sum of base cations was transformed using log10. Statistical relationships
180
were considered significant at p
