BGBiogeosciencesBGBiogeosciences1726-4189Copernicus PublicationsGöttingen, Germany10.5194/bg-14-3743-2017A global hotspot for dissolved organic carbon in hypermaritime watersheds of
coastal British ColumbiaOliverAllison A.aaoliver@ualberta.caTankSuzanne E.https://orcid.org/0000-0002-5371-6577GiesbrechtIanKorverMaartje C.FloydWilliam C.SanbornPaulBulmerChuckLertzmanKen P.University of Alberta, Department of Biological Sciences, CW 405,
Biological Sciences Bldg., University of Alberta, Edmonton, AB, T6G
2E9, CanadaHakai Institute, Tula Foundation, P.O. Box 309, Heriot Bay, BC, V0P 1H0, CanadaMinistry of Forests, Lands and Natural Resource Operations, 2100
Labieux Rd, Nanaimo, BC, V9T 6E9, CanadaVancouver Island University, 900 Fifth Street, Nanaimo, BC, V9R 5S5,
CanadaEcosystem Science and Management Program, University of Northern
British Columbia, 3333 University Way, Prince George, BC, V2N 4Z9, CanadaBC Ministry of Forests Lands and Natural Resource Operations, 3401
Reservoir Rd, Vernon, BC, V1B 2C7, CanadaSchool of Resource and Environmental Management, Simon Fraser
University, TASC 1 – Room 8405, 8888 University Drive, Burnaby, BC, V5A 1S6,
CanadaAllison A. Oliver (aaoliver@ualberta.ca)15August201714153743376210January201719January201729June201710July2017This work is licensed under the Creative Commons Attribution 3.0 Unported License. To view a copy of this licence, visit https://creativecommons.org/licenses/by/3.0/This article is available from https://bg.copernicus.org/articles/14/3743/2017/bg-14-3743-2017.htmlThe full text article is available as a PDF file from https://bg.copernicus.org/articles/14/3743/2017/bg-14-3743-2017.pdf
The perhumid region of the coastal temperate rainforest (CTR) of
Pacific North America is one of the wettest places on Earth and contains
numerous small catchments that discharge freshwater and high concentrations
of dissolved organic carbon (DOC) directly to the coastal ocean. However,
empirical data on the flux and composition of DOC exported from these
watersheds are scarce. We established monitoring stations at the outlets of
seven catchments on Calvert and Hecate islands, British Columbia, which
represent the rain-dominated hypermaritime region of the perhumid CTR. Over
several years, we measured stream discharge, stream water DOC concentration,
and stream water dissolved organic-matter (DOM) composition. Discharge and
DOC concentrations were used to calculate DOC fluxes and yields, and DOM
composition was characterized using absorbance and fluorescence spectroscopy
with parallel factor analysis (PARAFAC). The areal estimate of annual DOC
yield in water year 2015 was 33.3 Mg C km-2 yr-1, with
individual watersheds ranging from an average of
24.1 to 37.7 Mg C km-2 yr-1. This represents some of the highest
DOC yields to be measured at the coastal margin. We observed seasonality in
the quantity and composition of exports, with the majority of DOC export
occurring during the extended wet period (September–April). Stream flow from
catchments reacted quickly to rain inputs, resulting in rapid export of
relatively fresh, highly terrestrial-like DOM. DOC concentration and measures
of DOM composition were related to stream discharge and stream temperature and correlated with watershed attributes, including the extent of lakes and
wetlands, and the thickness of organic and mineral soil horizons. Our discovery
of high DOC yields from these small catchments in the CTR is especially
compelling as they deliver relatively fresh, highly terrestrial organic
matter directly to the coastal ocean. Hypermaritime landscapes are common on
the British Columbia coast, suggesting that this coastal margin may play an
important role in the regional processing of carbon and in linking
terrestrial carbon to marine ecosystems.
Introduction
Freshwater aquatic ecosystems process and transport a significant amount of
carbon (Cole et al., 2007; Aufdenkampe et al., 2011; Dai et al., 2012).
Globally, riverine export is estimated to deliver around 0.9 Pg C yr-1
from land to the coastal ocean (Cole et al., 2007), with typically
> 50 % quantified as dissolved organic carbon (DOC) (Meybeck,
1982; Ludwig et al., 1996; Alvarez-Cobelas et al., 2012; Mayorga et al.,
2010). Rivers draining coastal watersheds serve as conduits of DOC from
terrestrial and freshwater sources to marine environments (Mulholland and
Watts, 1982; Bauer et al., 2013; McClelland et al., 2014) and can have
important implications for coastal carbon cycling, biogeochemical
interactions, ecosystem productivity, and food webs (Hopkinson et al., 1998;
Tallis, 2009; Tank et al., 2012; Regnier et al., 2013). In addition, because
the transfer of water and organic matter from watersheds to the coastal ocean
represents an important pathway for carbon cycling and ecological subsidies
between ecosystems, better understanding of these linkages is needed for
constraining predictions of ecosystem productivity in response to
perturbations such as climate change. In regions where empirical data are
currently scarce, quantifying land-to-ocean DOC export is therefore a
priority for improving the accuracy of watershed and coastal carbon models
(Bauer et al., 2013).
While quantifying DOC flux within and across systems is required for
understanding the magnitude of carbon exchange, the composition of DOC (as
dissolved organic matter, or DOM) is also important for determining the
ecological significance of carbon exported from coastal watersheds. The
aquatic DOM pool is a complex mixture that reflects both source material and
processing along the watershed terrestrial–aquatic continuum, and as a
result it can show significant spatial and temporal variation (Hudson et al.,
2007; Graeber et al., 2012; Wallin et al., 2015). Both DOC concentration and
DOM composition can serve as indicators of watershed characteristics
(Koehler et al., 2009), hydrologic flow paths (Johnson et al., 2011; Helton
et al., 2015), and watershed biogeochemical processes (Emili and Price,
2013). DOM composition can also influence its role in downstream processing
and ecological function, such as susceptibility to biological (Judd et al.,
2006) and physiochemical interactions (Yamashita and Jaffé, 2008).
The coastal temperate rainforests (CTRs) of Pacific North America extend from
the Gulf of Alaska through British Columbia to Northern California and span
a wide range of precipitation and climate regimes. Within this rainforest
region, the “perhumid” zone has cool summers and summer precipitation is
common (> 10 % of annual precipitation) (Alaback, 1996)
(Fig. 1). The perhumid CTR extends from southeast Alaska through the outer
coast of central British Columbia and contains forests and soils that have
accumulated large amounts of organic carbon above and below ground (Leighty
et al., 2006; Gorham et al., 2012). Due to high amounts of precipitation and
close proximity to the coast, this area represents a potential hotspot for
the transport and metabolism of carbon across the land-to-ocean continuum,
and quantifying these fluxes is pertinent for understanding global carbon
cycling.
The location of Calvert Island, British Columbia, Canada, within the
perhumid region of the coastal temperate rainforest (right) and the study
area on Calvert and Hecate islands, including the seven study watersheds,
corresponding stream outlet sampling stations, and location of the rain gauge
(left). Characteristics of individual watersheds are described in Table 1.
Within the large perhumid CTR, there is substantial spatial variation in
climate and landscape characteristics that create uncertainty about carbon
cycling and pattern. In Alaska, for example, riverine DOC concentrations vary
with wetland cover (D'Amore et al., 2015a) and glacial cover (Fellman et al.,
2014). Previous studies have shown that streams in southeast Alaska can
contain high DOC concentrations (Fellman et al., 2009a; D'Amore et al.,
2015a) and produce high DOC yields (D'Amore et al., 2015a, b; Stackpoole et
al., 2016), but no known field estimates have been generated for the perhumid
CTR of British Columbia, an area of approximately 97 824 km2 (adapted
from Wolf et al., 1995). Within the perhumid CTR of British Columbia,
terrestrial ecologists have defined a large (29 935 km2)
“hypermaritime” subregion where rainfall dominates over snow, seasonality is
moderated by the ocean, and wetlands are extensive (Pojar et al., 1991; area
estimated using British Columbia Biogeoclimatic Ecosystem Classification
Subzone/Variant mapping Version 10, 31 August 2016, available at
https://catalogue.data.gov.bc.ca/dataset/f358a53b-ffde-4830-a325-a5a03ff672c3).
Previous work in the hypermaritime CTR showed that DOC concentrations are
high in small streams and tend to increase during rain events (Gibson et al.,
2000; Fitzgerald et al., 2003; Emili and Price, 2013). Taken together, these
conditions should be expected to generate high yields and fluxes of DOC from
hypermaritime watersheds to the coastal ocean.
The objectives of this study were to provide the first field-based estimates
of DOC exports from watersheds in the extensive hypermaritime region of
British Columbia's perhumid CTR, to describe the temporal and spatial
dynamics of exported DOC concentration and DOM composition, and to identify
relationships between DOC concentration, DOM composition, and watershed
characteristics.
Watershed characteristics, discharge, DOC concentrations, and DOC
yields for the seven study watersheds on Calvert and Hecate islands.
Additional details on the methods used to determine watershed characteristics
can be found in the Supplement.
1 Calculated for water year 2015 (WY2015;
1 October 2014–30 September 2015). 2 Wet-period average monthly yield
calculated from October to April and September, WY2015, and October to April,
WY2016. 3 Dry-period average monthly yield calculated from May to August,
WY2015. a Mean ± standard deviation.
b Total ± 95 % confidence interval.
MethodsStudy sites
Study sites are located on northern Calvert Island and adjacent Hecate Island
on the central coast of British Columbia, Canada (lat 51.650, long
-128.035; Fig. 1). Average annual precipitation and air temperature at sea
level from 1981 to 2010 was 3356 mm yr-1 and 8.4 ∘C (average
annual min: 0.9 ∘C; average annual max: 17.9 ∘C)
(available online at http://www.climatewna.com/; Wang et al., 2012),
with precipitation dominated by rain and winter snowpack persisting only at
higher elevations. Sites are located within the hypermaritime region of the
CTR on the outer coast of British Columbia. Soils overlying the granodiorite
bedrock (Roddick, 1996) are usually < 1 m thick and have formed in
sandy colluvium and patchy morainal deposits, with limited areas of coarse
glacial outwash. Chemical weathering and organic-matter accumulation in the
cool, moist climate have produced soils dominated by Podzols and Folic
Histosols, with Hemists up to 2 m thick at depressional sites (IUSS Working
Group WRB, 2015). The landscape is comprised of a mosaic of ecosystem types,
including exposed bedrock, extensive wetlands, “bog forests”, and woodlands,
with organic-rich soils (Green, 2014; Thompson et al., 2016). Forest stands
are generally short with open canopies reflecting the lower productivity of
the hypermaritime forests compared to the rest of the perhumid CTR (Banner et
al., 2005). Dominant trees are western red cedar, yellow cedar, shore pine, and
western hemlock, with composition varying across topographic and edaphic
gradients. Widespread understory plants include bryophytes, salal, deer fern,
and tufted clubrush. Wetland plants are locally abundant including diverse
Sphagnum mosses and sedges. Although the watersheds have no history
of mining or industrial logging, archaeological evidence suggests that humans
have occupied this landscape for at least 13 000 years (McLaren et al.,
2014). This occupation has had a local effect on forest productivity near
habitation sites (Trant et al., 2016) and on fire regimes (Hoffman et al.,
2016). We selected seven watersheds with streams draining directly into the
ocean (Fig. 1). These numbered watersheds (626, 693, 703, 708, 819 844, and
1015) range in size (3.2 to 12.8 km2) and topography (maximum elevation
160 to 1012 m), are variably affected by lakes (0.3–9.1 % lake
coverage), and – as is characteristic of the perhumid CTR – have a high
degree of wetland coverage (24–50 %) (Table 1).
Soils and watershed characteristics
Watersheds and streams were delineated using a 3 m resolution digital
elevation model (DEM) derived from airborne laser scanning (lidar) and flow
accumulation analysis using geographic information systems (GISs) to summarize
watershed characteristics for each watershed polygon and for all watersheds
combined (Gonzalez Arriola et al., 2015; Table 1). Topographic measures were
estimated from the DEM, and lake and wetland cover was estimated from the Province of
British Columbia terrestrial ecosystem mapping (TEM) (Green, 2014), and soil
material thickness was estimated from unpublished digital soil maps (Supplement S1). We recorded the thickness of organic soil material, the thickness of mineral
soil material, and total soil depth to bedrock at a total of 353 field sites.
Mineral soil horizons have ≤ 17 % organic C, while organic soil
horizons have > 17 % organic C, per the Canadian System of
Soil Classification (Soil Classification Working Group, 1998). In addition to
field-sampled sites, 40 sites with exposed bedrock (0 cm soil depth) were
located using aerial photography. Soil thicknesses were combined with a suite
of topographic, vegetation, and remote-sensing (lidar and RapidEye satellite
imagery) data for each sampling point and used to train a random forest model
(randomForest package in R; Liaw and Wiener, 2002) that was used to predict
soil depth values. Soil material thicknesses were then averaged for each
watershed (Table 1). For additional details on field site selection and
methods used for predictions of soil thickness, see Supplement S1.1.
Sample collection and analysis
From May 2013 to July 2016, we collected stream water grab samples from each
watershed stream outlet every 2–3 weeks (ntotal= 402), with
less frequent sampling (∼ monthly) during winter (Fig. 1). All samples
were filtered in the field (Millipore Millex-HP Hydrophilic PES
0.45 µm) and kept in the dark, on ice, until analysis. DOC samples
were filtered into 60 mL amber glass bottles and preserved with 7.5 M
H3PO4. Fe samples were filtered into 125 mL HDPE bottles and
preserved with 8 M HNO3. DOC and Fe samples were analyzed at the BC
Ministry of the Environment Technical Services Laboratory (Victoria, BC,
Canada). DOC concentrations were determined on a total organic carbon (TOC) analyzer (Aurora 1030;
OI Analytical) using wet chemical
oxidation with persulfate followed by infrared detection of CO2. Fe
concentrations were determined on a dual-view inductively coupled plasma optical emission spectrometry (ICP-OES) spectrophotometer (Prodigy;
Teledyne Leeman Labs) using a Seaspray pneumatic nebulizer.
In May 2014, we began collecting stream samples for stable isotopic
composition of δ13C in DOC (δ13C-DOC; n= 173) and the
optical characterization of DOM using absorbance spectroscopy (n= 259).
Beginning in January 2016, we also analyzed samples using fluorescence
spectroscopy (see Sect. 2.6). Samples collected for δ13C-DOC were
filtered into 40 mL EPA glass vials and preserved with H3PO4.
δ13C-DOC samples were analyzed at GG Hatch Stable Isotope
Laboratory (Ottawa, ON, Canada) using high-temperature combustion (TIC-TOC
Combustion Analyser Model 1030; OI Analytical) coupled to a
continuous-flow isotope ratio mass spectrometry (Finnigan Mat DeltaPlusXP;
Thermo Fischer Scientific) (Lalonde et al., 2014). Samples analyzed for
optical characterization using absorbance and fluorescence were filtered into
125 mL amber HDPE bottles and analyzed at the Hakai Institute (Calvert
Island, BC, Canada) within 24 h of collection.
Hydrology: precipitation and stream discharge
We measured precipitation using a TB4-L tipping bucket rain gauge with a
0.2 mm resolution (Campbell Scientific Ltd.) located in watershed 708
(elevation: 16 m a.s.l.). The rain gauge was calibrated twice per year
using a field calibration device, model 653 (HYQUEST Solutions Ltd).
We determined continuous stream discharge for each watershed by developing
stage discharge rating curves at fixed hydrometric stations situated in close
proximity to each stream outlet. Sites were located above tidewater influence
and were selected based on favourable conditions (i.e., channel stability and
stable hydraulic conditions) for the installation and operation of pressure
transducers to measure stream stage. From August 2014 to May 2016
(21 months), we measured stage every 5 min using a pressure level sensor (OTT PLS-L, OTT Hydromet, Colorado, USA) pressure transducer (0–4 m range SDI-12)
connected to a CR1000 (Campbell Scientific, Edmonton, Canada) data logger.
Stream discharge was measured over various intervals using either the
velocity area method (for flows < 0.5 m3 s-1; ISO
Standard 9196, 1992; ISO Standard 748, 2007) or salt dilution (for flows
> 0.5 m3 s-1; Moore, 2005). Rating curves were
developed using the relationship between stream stage height and stream
discharge (Supplement S2).
DOC flux
From 1 October 2014 to 30 April 2016, we estimated DOC flux for each
watershed using measured DOC concentrations (n= 224) and continuous
discharge recorded at 15 min intervals. The watersheds in this region
respond rapidly to rain inputs and as a result DOC concentrations are highly
variable. To address this variability, routine DOC concentration data (as
described in Sect. 2.2) were supplemented with additional grab samples
(n= 21) collected around the peak of the hydrograph during several high-flow events throughout the year. We performed watershed-specific estimates of
DOC flux using the rloadest package (Lorenz et al., 2015) in R
(version 3.2.5, R Core Team, 2016), which replicates functions developed in
the US Geological Survey load-estimator program, LOADEST (Runkel et al.,
2004). LOADEST is a multiple-regression adjusted maximum likelihood (ML) estimation model that calibrates a regression between measured constituent
values and stream flow across seasons and time and then fits it to
combinations of coefficients representing nine predetermined models of
constituent flux. To account for potentially small sample size, the best
model was selected using the second-order Akaike Information Criterion (AICc)
(Akaike, 1981; Hurvich and Tsai, 1989). Input data were log-transformed to
avoid bias and centered to reduce multicollinearity. For additional details
on model selection, see Supplement Table S3.1.
Optical characterization of DOM
Prior to May 2014, absorbance measures of water samples (n= 99) were
conducted on a Varian Cary-50 (Varian, Inc.) spectrophotometer at the BC
Ministry of the Environment Technical Services Laboratory (Victoria, BC,
Canada) to determine specific UV absorption at 254 nm (SUVA254). After
May 2014, we conducted optical characterization of DOM by absorbance and
fluorescence spectroscopy at the Hakai Institute field station (Calvert
Island, BC, Canada) using an Aqualog fluorometer (Horiba Scientific, Edison,
New Jersey, USA). Strongly absorbing samples (absorbance units
> 0.2 at 250 nm) were diluted prior to analysis to avoid
excessive inner filter effects (Lakowicz, 1999). Samples were run in 1 cm
quartz cells and scanned from 220 to 800 nm at 2 nm intervals to determine
SUVA254 as well as the spectral slope ratio (SR).
SUVA254 has been shown to positively correlate with increasing molecular
aromaticity associated with the fulvic acid fraction of DOM (Weishaar et al.,
2003), and it is calculated by dividing the decadic absorption coefficient at
254 nm by DOC concentration (mg C L-1). To account for potential Fe
interference with absorbance values, we corrected SUVA254 values by Fe
concentration according the method described in Poulin et al. (2014).
SR has been shown to negatively correlate with molecular weight
(Helms et al., 2008) and is calculated as the ratio of the spectral slope
from 275 to 295 nm (S275-295) to the spectral slope from 350 to 400 nm
(S350-400).
We measured excitation and emission spectra (as excitation emission matrices,
EEMs) on samples every three weeks from January to July 2016 (n= 63).
Samples were run in 1 cm quartz cells and scanned from excitation
wavelengths of 230–550 nm at 5 nm increments and emission wavelengths of
210–620 nm at 2 nm increments. The Horiba Aqualog applied the appropriate
instrument corrections for excitation and emission, inner filter effects, and
Raman signal calibration. We calculated the fluorescence index and freshness index for each EEM. The fluorescence index is often used to indicate DOM
source, where higher values are more indicative of microbially derived sources
of DOM and lower values indicate more terrestrially derived sources (McKnight
et al., 2001), and is calculated as the ratio of emission intensity at 450 to
500 nm, at an excitation of 370 nm. The freshness index is used to indicate
the contribution of autochthonous or recently microbial-produced DOM, with
higher values suggesting greater autochthony (i.e., microbial inputs), and is
calculated as the ratio of emission intensity at 380 nm to the maximum
emission intensity between 420 and 435 nm, at an excitation of 310 nm (Wilson and
Xenopoulos, 2009).
To further characterize features of DOM composition, we performed parallel
factor analysis (PARAFAC) using EEM data within the drEEM toolbox for Matlab
(Mathworks, MA, USA) (Murphy et al., 2013). PARAFAC is a statistical
technique used to decompose the complex mixture of the fluorescing DOM pool
into quantifiable, individual components (Stedmon et al., 2003). We detected
a total of six unique components and validated the model using core
consistency and split-half analysis (Murphy et al., 2013; Stedmon and Bro,
2008). Components with similar spectra from previous studies were identified
using the online fluorescence repository, OpenFluor (Murphy et al., 2014),
and additional components with similar peaks were identified through
literature review. Since the actual chemical structure of fluorophores is
unknown, we used the concentration of each fluorophore as maximum
fluorescence of excitation and emission in Raman units (Fmax) to derive
the percent contribution of each fluorophore component to total
fluorescence. Relationships between PARAFAC components were also evaluated
using Pearson correlation coefficients in the R package Hmisc (Harrell
et al., 2016).
Evaluating relationships in DOC concentration and DOM composition with
stream discharge and temperature
We used linear mixed-effects (LME) models to assess the relationship between
DOC concentration or DOM composition (δ13C-DOC, SR,
SUVA254, fluorescence index, freshness index, PARAFAC components),
stream discharge, and stream temperature. Analysis was performed in R using
the nlme package (Pinheiro et al., 2016). Watershed was included as a random
intercept to account for repeat measures on each watershed. For some
parameters, a random slope of either discharge or temperature was also
included based on data assessment and model selection. Model selection was
performed using AIC to compare models fit using ML (Burnham and Anderson, 2002;
Symonds and Moussalli, 2010). The final model was fit using restricted
maximum likelihood (REML). Marginal R2, which represents an
approximation of the proportion of the variance explained by the fixed
factors alone, and conditional R2, which represents an approximation of
the proportion of the variance explained by both the fixed and random
factors, were calculated based on the methods described in Nakagawa and
Schielzeth (2013) and Johnson (2014).
Redundancy analysis: relationships between DOC concentration, DOM
composition, and watershed characteristics
We evaluated relationships between stream water DOC and watershed
characteristics by relating DOC concentration and measures of DOM composition
to catchment attributes using redundancy analysis (RDA; type 2 scaling) in
the package rdaTest (Legendre and Durand, 2014) in R (version 3.2.2, R Core
Team, 2015). To maximize the amount of information available, we performed
RDA analysis on samples collected from January to July 2016 and therefore
included all parameters of optical characterization (i.e., all PARAFAC
components and spectral indices). We assessed the collinearity of DOM
compositional variables using a variance inflation factor (VIF) criteria of
> 10, which resulted in the removal of PARAFAC components C2, C3,
and C5 prior to RDA analysis. Catchment attributes for each watershed
included average slope, percent area of lakes, percent area of wetlands,
average depth of mineral soil, and average depth of organic soil.
Relationships between variables were linear, so no transformations were
necessary and variables were standardized prior to analysis. To account for
repeat monthly measures per watershed and potential temporal correlation
associated with monthly sampling, we included sample month as a covariable
(“partial-RDA”). To test whether the RDA axes significantly explained
variation in the dataset, we compared permutations of residuals using ANOVA
(9999 iterations; test.axes function of rdaTest).
ResultsHydrology
We present work for water year 2015 (WY2015;
1 October 2014–30 September 2015) and water year 2016 (WY2016;
1 October 2015–30 September 2016). Annual precipitation for both water years
was lower than historical mean annual precipitation (WY2015: 2661 mm;
WY2016: 2587 mm). It is worth noting that mean annual precipitation at
our rain gauge location (2890 mm yr-1; elevation: 16 m) is
substantially lower than the average amount received at higher elevations,
which from 1981 to 2010 was approximately 5027 mm yr-1 at an elevation
of 1000 m within our study area. This area receives a very high amount of
annual rainfall but also experiences seasonal variation, with an extended wet
period from fall through spring and a much shorter, typically drier period
during summer. In WY2015 and WY2016, 86–88 % of the annual precipitation
on Calvert Island occurred during the 8 months of wetter and cooler weather
between September and April (∼ 75 % of the year), designated the
“wet period” (WY2015 wet: 2388 mm, average air
temp: 7.97 ∘C; WY2016 wet: 2235 mm, average air
temp: 7.38 ∘C). The remaining annual precipitation occurred
during the drier and warmer summer months of May–August, designated the
“dry period” (WY2015 dry: 314 mm, average air
temp: 13.4 ∘C; WY2016 dry: 352 mm, average air
temp: 13.1 ∘C). Overall, although WY2015 was slightly wetter
than WY2016, the two years were comparable in relative precipitation during
the wet versus dry periods.
Hydrological patterns typical of watersheds located in the study
area. (a) The hydrograph and precipitation record from Watershed 708
for the study period of 1 October 2015–30 April 2016. Grey shading indicates
the wet period (1 September–30 April) and the unshaded region indicates the
dry period (1 May–30 August). (b) Correlation of daily (24 h) areal
runoff (discharge of all watersheds combined) to 48 h total rainfall
recorded at watershed 708. For the period of study, comparisons of daily
runoff to 48 h rainfall (runoff : rainfall mean: 0.92; SD ± 0.27)
indicated rapid discharge response to rainfall.
Stream discharge (Q) responded rapidly to rain events and as a result,
closely tracked patterns in total precipitation (Fig. 2). Total Q for all
watersheds was on average 22 % greater for the wet period of WY2015
(total Q: 223.02 × 106;
range: 5.13 × 106–111.51 × 106 m3)
compared to the wet period of WY2016 (total
Q: 182.89 × 106;
range: 4.17 × 106–91.45 × 106 m3).
Stream discharge and stream temperature were significantly different for wet
versus dry periods (Mann–Whitney tests, p<0.0001).
Seasonal (timelines, by date) and spatial (box plots, by watershed)
patterns in DOC concentration and DOM composition for stream water collected
at the outlets of the seven study watersheds on Calvert and Hecate islands.
Boxes represent the 25th and 75th percentile, while whiskers
represent the 5th and 95th percentile. Daily precipitation and
annual temperature are shown in the top left panel. Grey shading indicates
the wet period (1 September–30 April) and the unshaded region indicates the
dry period of each water year.
Temporal and spatial patterns in DOC concentration, yield, and
flux
Stream waters were high in DOC concentration relative to the global average
for freshwater discharged directly to the ocean (average DOC for Calvert and
Hecate islands: 10.4 mg L-1, SD: 3.8; average global
DOC: ∼ 6 mg L-1) (Meybeck, 1982; Harrison et al., 2005)
(Table 1; Fig. 3). Q-weighted average DOC concentrations were higher than
average measured DOC concentrations (11.1 mg L-1, Table 1), and also
resulted in slightly different ranking of the watersheds for highest to
lowest DOC concentration. Within watersheds, Q-weighted DOC concentrations
ranged from a low of 8.4 mg L-1 (watershed 693) to a high of
19.3 mg L-1 (watershed 819), and concentrations were significantly
different between watersheds (Kruskal–Wallis test, p<0.0001).
Seasonal variability tended to be higher in watersheds where DOC
concentration was also high (watersheds 626, 819, and 844) and lower in
watersheds with greater lake area (watersheds 1015 and 708) (Table 1; box
plots, Fig. 3). On an annual basis, DOC concentrations generally decreased
through the wet period and increased through the dry period, and
concentrations were significantly lower during the wet period compared to the
dry period (Mann–Whitney test, p= 0.0123). Results of our LME model (Table S6.1) indicate that DOC concentration was
positively related to both discharge (b= 0.613, p<0.001) and
temperature (b= 0.162, p= 0.011) (model conditional R2= 0.57,
marginal R2= 0.09).
Monthly areal DOC yields and precipitation for water year 2015
(WY2015) and the wet period (1 October–30 April) of water year 2016
(WY2016). Error bars represent standard error. Total rain and DOC yield were
significantly correlated (r2=0.77), and months of higher rain
produced higher DOC yields. In WY2015, the majority of DOC export
(∼ 94 % of annual flux) occurred during the wet period
(∼ 88 % of annual precipitation).
Annual and monthly DOC yields are presented in Table 1. For the total period
of available Q (1 October 2014–30 April 2016; 19 months), areal (all
watersheds) DOC yield was 52.3 Mg C km-2 (95 % CI: 45.7 to
68.2 Mg C km-2) and individual watershed yields ranged from 24.1 to
43.6 Mg C km-2. For WY2015, areal annual DOC yield was
33.3 Mg C km-2 yr-1 (95 % CI: 28.9 to
38.1 Mg C km-2 yr-1). Total monthly rainfall was strongly
correlated with monthly DOC yield (Fig. 4), and average monthly yield for the
wet period (3.35 Mg C km-2 month-1; 95 % CI: 2.94 to
4.40 Mg C km-2 month-1) was significantly greater than average
monthly yield during the dry period (0.50 Mg C km-2 month-1;
95 % CI: 0.41 to 0.62 Mg C km-2 month-1) (Mann–Whitney
test, p<0.0001).
DOC fluxes and yields for the seven study watersheds and the total
area of study (“areal”, all watersheds combined) on Calvert and Hecate
islands for water year 2015 (WY2015; 1 October–30 September), and 1 October–30 April
of the wet period for water year 2015 (WY2015 wet) and water year 2016
(WY2016 wet). Because DOC yields were only available for September in
WY2015, this month was excluded from the wet-period totals in order to make
similar comparisons between years. Error bars represent standard error.
Across our study watersheds, DOC flux generally increased with increasing
watershed area (Fig. 5). In WY2015, total DOC flux for all watersheds
included in our study was 1562 Mg C (95 % CI: 1355 to 1787 Mg C),
and individual watershed flux ranged from ranged from 82 to 276 Mg C. DOC
flux was significantly different in wet versus dry periods (Mann–Whitney
test, p<0.0001). Overall, 94 % of the export in WY2015
occurred during the wet period, and export for the wet period of WY2015 was
lower than export for the wet period of WY2016 (Fig. 5).
Temporal and spatial patterns in DOM composition
The stable isotopic composition of dissolved organic carbon (δ13C-DOC) was relatively tightly constrained over space and time (average
δ13C-DOC: -26.53 ‰, SD: 0.36;
range: -27.67 to -24.89 ‰). Values of SR were
low compared to the range typically observed in surface waters (average
SR= 0.78, SD: 0.04; range: 0.71 to 0.89), and
Fe-corrected SUVA254 values were at the high end of the range compared
to most surface waters (average SUVA254 for Calvert and Hecate
islands: 4.42 L mg-1 m-1, SD: 0.46; range of
SUVA254 in surface waters: 1.0 to 5.0 L mg-1 m-1)
(Spencer et al., 2012). Values for both the fluorescence index (average
fluorescence index: 1.36, SD: 0.04; range: 1.30 to 1.44) and
freshness index (average freshness index: 0.46, SD: 0.02;
range: 0.41 to 0.49) were relatively low compared to the typical range
found in surface waters (Fellman et al., 2010; Hansen et al., 2016).
Differences between watersheds were observed for δ13C-DOC
(Kruskal–Wallis test, p= 0.0043), SR (Kruskal–Wallis test,
p= 0.0001), fluorescence index (Kruskal–Wallis test,
p= 0.0030), and freshness index (Kruskal–Wallis test,
p= 0.0099), but watersheds did not differ in SUVA254
(Kruskal–Wallis test, p= 0.4837).
We observed seasonal variability in δ13C-DOC throughout the period
of sampling (Fig. 3 and our LME model (Table S6.1) indicate that
δ13C-DOC declined with increasing discharge (b=-0.049,
p= 0.014) and stream temperature (b=-0.024, p<0.001)
(model conditional R2= 0.35, marginal R2= 0.10). In contrast,
although SUVA254 appeared to exhibit a general seasonal trend of values
increasing over the wet period and decreasing over the dry period,
SUVA254 was not significantly related to either discharge or stream
temperature in the LME model results. SR also appeared to
fluctuate seasonally, with lower values during the wet season and higher
values during the dry season. SR was negatively related to
discharge (b=-0.026; p<0.001) and positively related to the
interaction between discharge and stream temperature (b= 0.0015;
p<0.001) (model conditional R2= 0.62; marginal
R2= 0.28). The freshness index was negatively related to stream temperature
(b=-0.003; p= 0.008) (model conditional R2= 0.59; marginal
R2= 0.23), while the fluorescence index was not significantly related to
either discharge or stream temperature.
Percent contribution of the six components identified in parallel
factor analysis (PARAFAC) for samples collected every three weeks from
January–July 2016 from the seven study watersheds on Calvert and Hecate
islands. The grey shading indicates the wet period and the unshaded region
indicates the dry period. Note that while the y axis for each panel has a
range of 20 %, the max and min for each y axis varies by panel.
Spectral composition for the six fluorescence components identified
using PARAFAC, including excitation (Ex.) and emission (Em.) peak values,
percent composition across all samples, and likely structure and
characteristics of the fluorescent component based on previous studies.
ComponentEx. (nm)Em. (nm)%composition*Potential structure/characteristicsPrevious studies with comparable resultsC131543634.1 ± 2.2 (31.1–39.3)Humic-like, less processed terrestrial,high molecular weight, widespread buthighest in wetland and forest environmentGarcia et al. (2015) (C1); Graeber et al. (2012) (C1); Walker et al. (2014) (C1); Yamashita et al. (2011) (C1); Cory and McKnight (2005) (C1)C2270/38048420.2 ± 1.9 (16.1–25.6)Humic-like, resembles fulvic acid,widespread, high molecular weight terrestrialStedmon and Markager (2005) (C2); Stedmon et al. (2003) (C3); Cory and McKnight (2005) (C5)C327047817.8 ± 1.8 (12.8–20.8)Humic-like, highly processed terrestrial; suggested as refractoryStedmon and Markager (2005) (C1); Yamashita et al. (2010) (C2)C4305/43552214.8 ± 2.6 (9.4–22.3)Not commonly reported, similarities tofulvic-like, contributed from soilsLochmuller and Saavedra (1986) (E)C53254429.8 ± 3.5 (0.0–15.9)Aquatic humic-like from terrestrialenvironments; autochthonous, microbialproduced; may be photoproducedBoehme and Coble (2000) (Peak C); Coble et al. (1998) (Peak C); Stedmon et al. (2003) (C3)C62853383.4 ± 2.5 (0.0–9.3)Amino acid-like/tryptophan-like. Freshly added from land, autochthonous. Rapidly photodegradableMurphy et al. (2008) (C7); Shutova et al. (2003) (C4); Stedmon et al. (2007) (C7); Yamashita et al. (2003) (C5)
* Mean ± SD (min–max) from all samples.
PARAFAC characterization of DOM
Six fluorescence components were identified through PARAFAC (“C1” through
“C6”) (Table 2). Additional details on PARAFAC model results are provided
in Supplement Table S4.1 and Figs. S4.2 and S4.3. Of the six components,
four were found to have close spectral matches in the OpenFluor database
(C1, C3, C5, C6; minimum similarity score > 0.95), while the
remaining two (C2 and C4) were found to have similar peaks represented in the
literature. The first four components (C1 through C4) are described as
terrestrially derived, whereas components C5 and C6 are described as
autochthonous or microbially derived (Table 2). In general, the rank order of
each component's percent contribution to total fluorescence was maintained
over time, with C1 comprising the majority of total fluorescence across all
watersheds (Fig. 6).
Across watersheds, components fluctuated synchronously over time and
variation between watersheds was relatively low, although slightly more
variation between watersheds was observed during the beginning of the dry
period relative to other times of the year (Fig. 6). The percent
contributions of components C1, C3, C5, and C6 to total fluorescence were not
significantly different across watersheds (for all components Kruskal–Wallis
test, p>0.05); however, the percent composition of both C2 and C4
was different (Kruskal–Wallis test, p= 0.0306 and p= 0.0307,
respectively) and higher for watersheds 819 and 844 relative to the other
watersheds (Fig. S4.4).
PARAFAC components exhibited significant relationships with stream discharge
and stream temperature, although predicted changes (beta, or b) in
fluorescence components with discharge and/or stream temperature were small
(Supplement Table S6.2). C3 increased with discharge (b= 0.006,
p= 0.003), whereas C2, C4, and C5 decreased with discharge (C2:
b=-0.005, p= 0.022; C4: b= -0.008, p= 0.002; C5:
b=-0.008, p= 0.002). C1, C4, and C6 increased with temperature
(C1: b= 0.001, p= 0.050; C4: b= 0.003, p<0.001; C6:
b= 0.005, p= 0.005), while both C3 and C5 decreased with
temperature (C3: b= -0.003, p= 0.003; C5: b=-0.003,
p= 0.027). Conditional R2 values for the models ranged from 0.28
to 0.69, while marginal R2 ranged from 0.20 to 0.46. Overall, greater
changes in component contribution to total fluorescence were observed with
changes in discharge relative to changes in stream temperature.
Results from the partial redundancy analysis (RDA; type 2 scaling)
of DOC concentration and DOM composition versus watershed characteristics.
Angles between vectors represent correlation; i.e., smaller angles indicate
higher correlation. Symbols represent different watersheds, and numbers on
symbols represent the sample month in 2016: 1 – January;
2 – February; 3 – March; 4 – early April; 5 – late April; and
6 – May.
Relationships between watershed characteristics, DOC concentrations, and
DOM composition
Results of the partial RDA (type 2 scaling) were significant in explaining
variability in DOM concentration and composition (semi-partial
R2= 0.33, F= 7.90, p<0.0001) (Fig. 7). Axes 1
through 3 were statistically significant at p<0.001, and the
relative contribution of each axis to the total explained variance was 47,
30, and 22 %, respectively. Additional details on the RDA test are
provided in Figs. S5.1–S5.2 and Tables S5.3–S5.5. Axis 1 described a
gradient of watershed coverage by water-inundated ecosystem types, ranging
from more wetland coverage to more lake coverage. Total lake coverage (area)
and mean mineral soil material thickness showed a strong positive
contribution, and wetland coverage (area) showed a strong negative
contribution to this axis. The freshness index, the fluorescence index,
SR, and the fluorescence component C6 were positively correlated with
Axis 1, while component C4 showed a clear negative correlation. Axis 2
described a subtler gradient of soil material thickness ranging from greater
mean organic soil material thickness to greater mean mineral soil material
thickness. DOC concentration, δ13C-DOC, SUVA254, and
fluorescence component C1 all showed a strong, positive correlation with
Axis 2. Axis 3 described a gradient of watershed steepness, from lower
gradient slopes with more wetland area and thicker organic soil material to
steeper slopes with less developed organic horizons. Average slope
contributed negatively to Axis 3 (see Table S5.5), followed by positive
contributions from both wetland area and the thickness of organic soil material.
δ13C-DOC showed the most positive correlation with Axis 3, whereas
fluorescence components C1 and C4 showed the most negative.
DiscussionDOC export from small catchments to the coastal ocean
In comparison to global models of DOC export (Mayorga et al., 2010) and DOC
exports quantified for southeastern Alaska (D'Amore et al., 2015a, 2016;
Stackpoole et al., 2017), our estimates of freshwater DOC yield from Calvert and Hecate island watersheds are in the upper range predicted for the
perhumid rainforest region. When compared to watersheds of similar size, DOC
yields from Calvert and Hecate island watersheds are some of the highest
observed (see reviews in Hope et al., 1994; Alvarez-Cobelas et al., 2012),
including DOC yields from many tropical rivers, despite the fact that
tropical rivers have been shown to export very high DOC (e.g., Autuna River,
Venezuela, DOC yield: 56 946 kg C km-2 yr-1; Castillo et
al., 2004) and are often regarded as having disproportionately high carbon
export compared to temperate and Arctic rivers (Aitkenhead and McDowell,
2000; Borges et al., 2015). Our estimates of DOC yield are comparable to, or
higher than, previous estimates from high-latitude catchments of similar size
that receive high amounts of precipitation and contain extensive organic
soils and wetlands (e.g., Naiman, 1982 (DOC
yield: 48 380 kg C km-2 yr-1); Brooks et al., 1999 (DOC
yield: 20 300 kg C km-2 yr-1); Ågren et al., 2007 (DOC
yield: 32 043 kg C km-2 yr-1)). However, many of these
catchments represent low- (first- or second-) order headwater streams that drain
to higher-order stream reaches, rather than directly to the ocean. Although
headwater streams have been shown to export up to 90 % of the total
annual carbon in stream systems (Leach et al., 2016), significant processing
and loss typically occurs during downstream transit (Battin et al., 2008).
Over much of the incised outer coast of the CTR, small rainfall-dominated
catchments contribute high amounts of freshwater runoff to the coastal ocean
(Royer, 1982; Morrison et al., 2012; Carmack et al., 2015). Small mountainous
watersheds that discharge directly to the ocean can exhibit
disproportionately high fluxes of carbon relative to watershed size and in
aggregate may deliver more than 50 % of total carbon flux from
terrestrial systems to the ocean (Milliman and Syvitski, 1992; Masiello and
Druffel, 2001). Extrapolating our estimate of annual DOC yield from Calvert and Hecate island watersheds to the entire hypermaritime subregion of British
Columbia's CTR (29 935 km2) generates an estimated annual DOC flux of
0.997 Tg C yr-1 (0.721 to 1.305 Tg C yr-1 for our lowest to
highest yielding watersheds, respectively), with the caveat that this
estimate is rudimentary and does not account for spatial heterogeneity in
controlling factors such as wetland extent, topography, and watershed size.
Regional comparisons estimate that Southeast Alaska (104 000 km2), at
the northern range of the CTR, exports approximately 1.25 Tg C yr-1
(Stackpoole et al., 2016), while south of the perhumid CTR, the wet
northwestern United States and its associated coastal temperate rainforests
export less than 0.153 Tg C yr-1 as DOC (reported as TOC; Butman et
al., 2016). This suggests that the hypermaritime coast of British Columbia
plays an important role in the export of DOC from coastal temperate
rainforest ecosystems of western North America, in a region that is already
expected to contribute high quantities of DOC to the coastal ocean.
DOM composition
The composition of stream water DOM exported from Calvert and Hecate island
watersheds is mainly terrestrial, indicating that the production and overall
supply of terrestrial material is sufficient to exceed microbial demand, and
thus a relatively abundant supply of terrestrial DOM is available for export.
Values for δ13C-DOC suggest that terrestrial carbon sources from C3
plants and soils were the dominant input to catchment stream water DOM
(Finlay and Kendall, 2007). Measures of SR and SUVA254 were
typical of environments that export large quantities of high molecular-weight, highly aromatic DOM such as some tropical rivers (e.g., Lambert et
al., 2016; Mann et al., 2014), streams draining wetlands (e.g., Ågren et
al., 2008; Austnes et al., 2010), or streams draining small undisturbed
catchments comprised of mixed forest and wetlands (e.g., Wickland et al.,
2007; Fellman et al., 2009a; Spencer et al., 2010; Yamashita et al., 2011).
This suggests that the majority of the DOM pool is comprised of larger molecules
that have not been extensively chemically or biologically degraded through
processes such as microbial utilization or photodegradation and therefore
are potentially more biologically available (Amon and Benner, 1996).
Biological utilization of DOM is influenced by its composition (e.g., Judd et
al., 2006; Fasching et al., 2014); therefore, differences in DOM can alter the
downstream fate and ecological role of freshwater-exported DOM. For example,
the majority of the fluorescent DOM pool was comprised of C1, which is
described as humic-like, less processed terrestrial soil and plant material
(see Table 2). In addition, although the tryptophan-like component C6,
represents a minor proportion of total fluorescence, even a small
proteinaceous fraction of the overall DOM pool can play a major role in
overall bioavailability and bacterial utilization of DOM (Berggren et al.,
2010; Guillamette and Giorgio, 2011). These contributions of stream-exported
DOM may represent a relatively fresh, seasonally consistent contribution of
terrestrial subsidy from streams to the coastal ecosystem, which in this
region is relatively lower in carbon and nutrients throughout much of the
year (Whitney et al., 2005; Johannessen et al., 2008).
DOC and DOM export: sources and seasonal variability
On Calvert and Hecate islands, the relationship between DOC concentration and
discharge varied by watershed (see Supplement Fig. S6.1), as might be
expected given the known influence of watershed characteristics (e.g., lake
area, wetland area, soils, etc.) on DOC concentration and export. However,
overall DOC concentrations increased in all watersheds with both discharge
and temperature indicating that the overarching drivers of DOC export are the
hydrologic coupling of precipitation and runoff from the landscape with the
seasonal production and availability of DOC (Fasching et al., 2016).
Precipitation is a well-established driver of stream DOC export
(Alvarez-Cobelas et al., 2012), particularly in systems containing organic
soils and wetlands (Olefeldt et al., 2013; Wallin et al., 2015; Leach et
al., 2016). Frequent, high-intensity precipitation events and short
residence times are expected to result in pulsed exports of stream DOC that
are rapidly shunted downstream, thus reducing time for in-stream processing
(Raymond et al., 2016). Flashy stream hydrographs indicate that hydrologic
response times for Calvert and Hecate island watersheds are rapid,
presumably as a result of small catchment size, high drainage density, and
relatively shallow soils with high hydraulic conductivity (Gibson et al.,
2000; Fitzgerald et al., 2003). Rapid runoff is presumably accompanied by
rapid increases in water tables and lateral movement of water through
shallow soil layers rich in organic matter (Fellman et al., 2009b; D'Amore
et al., 2015b). It appears that on Calvert and Hecate islands, the
combination of high rainfall, rapid runoff, and abundant sources of DOC from
organic-rich wetlands and forests results in high DOC fluxes.
The relationship between DOC, stream temperature, and discharge indicates
that seasonal dynamics play an important role in the variability of DOC
exported from these systems. For example, DOC concentrations decrease in all
watersheds during the wet period of the year; these decreases are associated
with clear changes in DOM composition, such as increasing δ13C-DOC and SUVA254, and decreasing SR. This is in contrast
with patterns observed during the dry period, when DOC concentrations
gradually increase, while δ13C-DOC, SUVA254 decrease.
Fluctuations in DOC and DOM composition occur throughout the wet and the dry
season, suggesting that temperature and runoff – and perhaps other seasonal
drivers – are important year-round controls on DOC concentration as well as
certain measures of DOM composition, such as δ13C-DOC and
SR.
The process of “DOC flushing” has been shown to increase stream water DOC
during higher flows in coastal and temperate watersheds (e.g., Sanderman et
al., 2009; Deirmendjian et al., 2017). Flushing can occur through various
mechanisms. For example, Boyer et al. (1996) observed that during drier
periods, DOC pools can increase in soils and are then flushed to streams when
water tables rise. Rising water tables can establish strong hydraulic
gradients that initiate and sustain prolonged increases in metrics like
SUVA254, until the progressive drawdown of upland water tables
constrains flow paths (Lambert et al., 2013). DOC concentrations can vary
during flushing in response to changing flow paths, which can shift sources
of DOC within the soil profile from older material in deeper soil horizons to
more recently produced material in shallow horizons (Sanderman et al., 2009)
or from changes in the production mechanism of DOC (Lambert et al., 2013).
For example, Sanderman et al. (2009), observed distinct relationships between
discharge and both δ13C-DOC and SUVA254 and postulated that
during the rainy season, hillslope
flushing shifts DOM sources to more aged soil organic material. In addition,
instream production can also provide a source of DOC and therefore affect
seasonal variation in DOC concentration and composition (Lambert et al.,
2013). The extent of these effects can shift seasonally; relationships
between flow paths and DOC export in rain-dominated catchments can vary
within and between hydrologic periods depending on factors such as the degree
of soil saturation, the duration of previous drying and rewetting cycles,
soil chemistry, and DOM source-pool availability (Lambert et al., 2013).
Our observations of changes in DOC and DOM related to discharge and stream
temperature suggest that a variety of mechanisms may be important for
controlling dynamics of seasonal export in Pacific hypermaritime watersheds.
We observed elevated DOC concentrations during precipitation events following
extended dry periods, suggesting DOC may accumulate during dry periods and be
flushed to streams during runoff events. Increased discharge was
significantly related to δ13C-DOC and SR, with higher
discharge resulting in more terrestrial-like DOM. One possible explanation is
that hydrologic connectivity increases during higher discharge as soil
conditions become more saturated, therefore promoting the mobilization of DOM
from across a wider range of the soil profile (McKnight et al., 2001; Kalbitz
et al., 2002). In addition, the mechanisms of DOC production and sources of
DOC appear to shift seasonally. Relationships between increased temperature
and lower values of δ13C-DOC and higher values of the freshness index, C1, and C4 suggest that warmer conditions result in a fresh supply of
DOM exported from terrestrial sources (Fellman et al., 2009a; Fasching et
al., 2016). This may represent a shift in the source of DOM and/or increased
contributions from less aromatic, lower molecular-weight material, such as
DOM derived from increased terrestrial primary production (Berggren et al.,
2010). Further, fine-scaled investigation into the mechanistic underpinnings
of the relationship between discharge, stream temperature, and DOM
represents a clear priority for future research in this region.
Relationships between watershed attributes and exported DOM
Previous studies have implicated wetlands as a major driver of DOM
composition (e.g., Xenopoulos et al., 2003; Ågren et al., 2008; Creed et
al., 2008); however, the analysis of relationships between Calvert and Hecate
island landscape attributes and variation in DOM composition suggests that
controls on DOM composition are more nuanced than being solely driven by the
extent of wetlands. Ågren et al. (2008) found that when wetland area
comprised > 10 % of total catchment area, wetland DOM was the
most significant driver of stream DOM composition during periods of high
hydrologic connectivity. Although wetlands comprise an average of 37 % of
our study area, they do not appear to be the single leading driver of
variability in DOC concentration and DOM composition. Other factors, such as
watershed slope, the depth of organic and mineral soil materials, and the
presence of lakes also appear to be influence DOC concentration and DOM
composition. The presence of cryptic wetlands (Creed et al., 2003) and limitations of the wetland mapping method
could also weaken the link between wetland extent, DOC, and DOM.
In these watersheds, soils with pronounced accumulations of organic matter
are not restricted to wetland ecosystems. Peat accumulation in wetland
ecosystems results in the formation of organic soils (Hemists), where mobile
fractions of DOM accumulate under saturated soil conditions and limited
drainage, resulting in the enrichment of poorly biodegradable, more stable
humic acids (Stevenson, 1994; Marschner and Kalbitz, 2003). Although Hemist
soils comprise 27.8 % of our study area, Folic Histosols, which form
under more freely drained conditions, such as steeper slopes, occur over an
additional 25.7 % of the area (Supplement S1.2). In freely drained
organic soils, high rates of respiration can result in further enrichment of
aromatic and more complex molecules, and this material may be rapidly
mobilized and exported to streams (Glatzel et al., 2003). This suggests the
importance of widely distributed, alternative soil DOM source pools, such as
Folic Histosols and associated Podzols with thick forest floors on
hillslopes, available to contribute high amounts of terrestrial carbon for
export.
Although lakes make up a relatively small proportion of the total landscape
area, their influence on DOM export appears to be important. The proportion
of lake area can be a good predictor of organic carbon loss from a catchment
since lakes often increase hydrologic residence times and thus increase
opportunities for biogeochemical processing (Algesten et al., 2004; Tranvik
et al., 2009). In our study, watersheds with a larger percentage of lake area
exhibited a slower response following rain events (Fig. S2.2) and lower DOC
yields, and lake area was correlated with parameters that represent greater
autochthonous DOM production or microbial processing such as higher freshness index, SR, and fluorescence index and a higher proportions of
component C6. In contrast, watersheds with a high percentage of wetlands
contributed DOM that was more allochthonous in composition. Lakes are known to
be important landscape predictors of DOC, as increased residence time enables
removal via respiration, thus reducing downstream exports from lake outlets
(Larson et al., 2007). The proximity of wetlands and lakes to the watershed
outlet can also play an important role in the composition of DOM exports
(Martin et al., 2006).
Conclusions
Previous work has demonstrated that freshwater discharge is substantial along the
coastal margin of the North Pacific temperate rainforest and plays an
important role in processes such as ocean circulation (Royer, 1982; Eaton and
Moore, 2010). Our finding that small catchments in this region contribute
high yields of terrestrial DOC to coastal waters suggests that freshwater
inputs may also influence ocean biogeochemistry and food web processes
through terrestrial organic-matter subsidies. Our findings also suggest that
this region may be currently underrepresented in terms of its role in global
carbon cycling. Currently, there is no region-wide carbon flux model for the
Pacific coastal temperate rainforest or the greater Gulf of Alaska, which
would quantify the importance of this region within the global carbon budget.
Our estimates point to the importance of the hypermaritime outer-coast zone
of the CTR, where subdued terrain, high rainfall, ocean-moderated
temperatures and poor bedrock have generated a distinctive bog-forest landscape mosaic within the greater temperate rainforest (Banner et al.,
2005). However, even within our geographically limited study area, we
observed a range of DOC yields across watersheds. To quantify regional-scale
fluxes of rainforest carbon to the coastal ocean, further research will be
needed to estimate DOC yields across complex spatial gradients of topography,
climate, hydrology, soils, and vegetation. Long-term changes in DOC flux have
been observed in many places (e.g., Worrall et al., 2004; Borken et al.,
2011; Lepistö et al., 2014; Tank et al., 2016), and continued monitoring
of this system will allow us to better understand the underlying drivers of
export and evaluate future patterns in DOC yields. Coupled with current
studies investigating the fate of terrestrial material in ocean food webs,
this work will improve our understanding of coastal carbon patterns and
increase capacity for predictions regarding the ecological impacts of climate
change.
Chemistry and DOC flux data are available in Oliver et
al. (2017) (10.21966/1.321324). Ecosystem comparison plot data are
available in Giesbrecht et al. (2015) (10.21966/1.56481). Stage and
discharge time series data are available in Floyd et al. (2016) (10.21966/1.243102). Lidar-derived watersheds and associated metrics data
are available in Gonzalez Arriola et al. (2015) (10.21966/1.15311).
The Supplement related to this article is available online at https://doi.org/10.5194/bg-14-3743-2017-supplement.
AAO prepared the paper with contributions from all authors,
designed analysis protocols, analyzed samples and performed the modelling and
analysis for dissolved organic carbon fluxes, parallel factor analysis of
dissolved organic-matter composition, and all remaining statistical
analyses. SET assisted with designing the study and overseeing
laboratory analyses, crafting the scope of the paper, and determining the
analytical approach.
IG led the initial DOC sampling design, helped coordinate the
research team, oversaw routine sampling and data management, and led the
watershed characterization.
MCK developed the rating curves and conducted the statistical
analysis of discharge measurement uncertainties and rating curve
uncertainties. WCF led the hydrology component of this project,
selected site locations, installed and designed the hydrometric stations, and
developed the rating curves and final discharge calculations. CB and
PS collected and analyzed soil field data and prepared the digital
soils map of the watersheds. KPL conceived of and co-led the
overall study of which this paper is a component, helped assemble and guide
the team of researchers who carried out this work and provided input to each
stage of the study.
The authors declare that they have no conflict of
interest.
Acknowledgements
This work was funded by the Tula Foundation and the Hakai Institute. The
authors would like to thank many individuals for their support, including
Skye McEwan, Bryn Fedje, Lawren McNab, Nelson Roberts, Adam Turner, Emma Myers, David Norwell, and Chris Coxson for sample collection and data
management, Clive Dawson and North Road Analytical for sample processing and
data management, Keith Holmes for creating our maps, Matt Foster for database
development and support, Shawn Hateley for sensor network maintenance, Jason Jackson, Colby Owen, James McPhail, and the entire staff at Hakai Energy
Solutions for installing and maintaining the sensors and telemetry network,
and Stewart Butler and Will McInnes for field support. Thanks to Santiago
Gonzalez Arriola for generating the watershed summaries and associated data
products and Ray Brunsting for overseeing the design and implementation of
the sensor network and the data management system at Hakai. Additional thanks
to Lori Johnson and Amelia Galuska for soil mapping field assistance, and
Francois Guillamette for PARAFAC consultation. Thanks to Dave D'Amore for
inspiring the Hakai project to investigate aquatic fluxes at the coastal
margin and for technical guidance. Lastly, thanks to Eric Peterson and
Christina Munck, who provided significant guidance throughout the process of
designing and implementing this study.
Edited by: Steven Bouillon
Reviewed by: three anonymous referees
ReferencesÅgren, A., Buffam, I., Jansson, M., and Laudon, H.: Importance of
seasonality and small streams for the landscape regulation of dissolved
organic carbon export, J. Geophys. Res.-Biogeosci., 112,
10.1029/2006JG000381, 2007.Ågren, A., Buffam, I., Berggren, M., Bishop, K., Jansson, M., and Laudon,
H.: Dissolved organic carbon characteristics in boreal streams in a
forest-wetland gradient during the transition between winter and summer, J.
Geophys. Res.-Biogeosci., 113, 10.1029/2007JG000674, 2008.Akaike, H.: Likelihood of a model and information criteria, J. Econometrics,
16, 3–14, 10.1016/0304-4076(81)90071-3, 1981.Aitkenhead, J. A. and McDowell, W. H.: Soil C : N ratio as a predictor of annual
riverine DOC flux at local and global scales, Global Biogeochem. Cy.,
14, 127–138, 10.1029/1999GB900083, 2000.Alaback, P. B.: Biodiversity patterns in relation to climate: The coastal
temperate rainforests of North America, Ecol. Stud., 116, 105–133,
10.1007/978-1-4612-3970-3_7, 1996.Algesten, G., Sobek, S., Bergström, A., Ågren, A., Tranvik, L., and
Jansson, M.: Role of lakes for organic carbon cycling in the boreal zone,
Glob. Change Biol., 10, 141–147, 10.1111/j.1365-2486.2003.00721.x,
2004.Alvarez-Cobelas, M., Angeler, D., Sánchez-Carrillo, S., and Almendros,
G.: A worldwide view of organic carbon export from catchments,
Biogeochemistry, 107, 275–293, 10.1007/s10533-010-9553-z, 2012.
Amon, R. M. W. and Benner, R.: Bacterial utilization of different size
classes of dissolved organic matter, Limnol. Oceanogr., 41, 41–51, 1996.Aufdenkampe, A., Mayorga, E., Raymond, P., Melack, J., Doney, S., Alin, S.,
Aalto, R., and Yoo, K.: Riverine coupling of biogeochemical cycles between
land, oceans, and atmosphere, Front. Ecol. Environ., 9, 53–60,
10.1890/100014, 2011.Austnes, K., Evans, C. D., Eliot-Laize, C., Naden, P. S., and Old, G. H.:
Effects of storm events on mobilisation and in-stream processing of dissolved
organic matter (DOM) in a Welsh peatland catchment, Biogeochemical, 99, 157–173,
10.1007/s10533-009-9399-4, 2010.Banner, A., LePage, P., Moran, J., and de Groot, A. (Eds.): The HyP3 Project:
pattern, process, and productivity in hypermaritime forests of coastal
British Columbia – a synthesis of 7-year results, Special Report 10, Res.
Br., British Columbia Ministry Forests, Victoria, British Columbia, 142 pp.,
available at: http://www.for.gov.bc.ca/hfd/pubs/Docs/Srs/Srs10.htm
(last access: 11 August 2017), 2005.
Battin, T. J., Kaplan, L. A., Findlay, S., Hopkinson, C. S., Marti, E., Packman,
A. I., Newbold, D., and Sabater, F.: Biophysical controls on organic carbon
fluxes in fluvial networks, Nat. Geosci., 1, 95–100, 2008.Bauer, J. E., Cai, W. J., Raymond, P. A., T. S., Bianchi, Hopkinson, C. S., and
Regnier, P. A. G.: The changing carbon cycle of the coastal ocean, Nature,
504, 61–70, 10.1038/nature12857, 2013.Berggren, M., Laudon, H., Haei, M., Ström, L., and Jansson, M.: Efficient
aquatic bacterial metabolism of dissolved low-molecular-weight compounds from
terrestrial sources, ISME J., 4, 408–416, 10.1038/ismej.2009.120,
2010.Boehme, J. and Coble, P.: Characterization of Colored Dissolved Organic
Matter Using High-Energy Laser Fragmentation, Environ. Sci. Technol.,
34, 3283–3290, 10.1021/es9911263, 2000.Borken, W., Ahrens, B., Schultz, C., and Zimmermann, L.: Site-to-site
variability and temporal trends of DOC concentrations and fluxes in temperate
forest soils, Glob. Change Biol., 17: 2428–2443,
10.1111/j.1365-2486.2011.02390.x, 2011.Borges, A. V., Darchambeau, F., Teodoru, C. R., Marwick, T. R., Tamooh, F.,
Geeraert, N., Omengo, F. O., Guérin, F., Lambert, T., Morana, C., Okuku,
E., and Bouillon, S.: Globally significant greenhouse-gas emissions form
African inland waters, Nat. Geosci., 8, 637–642, 10.1038/ngeo2486,
2015.
Boyer, E. W., Hornberger, G. M., Bencala, K. E., and McKnight, D.: Overview of
a simple model describing variation of dissolved organic carbon in an upland
catchment, Ecol. Modell., 86, 183–188, 1996.
Burnham, K. P. and Anderson, D. R.: Model selection and multimodel inference,
2nd Edn., Springer, New York, 2002.Carmack, E., Winsor, P., and William, W.: The contiguous panarctic Riverine
Coastal Domain: A unifying concept, Prog. Oceanogr., 139, 13–23,
10.1016/j.pocean.2015.07.014, 2015.Castillo, M. M., Allan, J. D., Sinsabaugh, R. L., and Kling, G. W.: Seasonal and
interannual variation of bacterial production in lowland rivers of the
Orinoco basin, Freshwater Biol., 49, 1400–1414,
10.1111/j.1365-2427.2004.01277.x, 2004.Clark, J. M., Lane, S. N., Chapman, P. J., and Adamson, J. K.: Export of
dissolved organic carbon from an upland peatland during storm events:
Implications for flux estimates, J. Hydrol., 347, 438–447, 10.1016/j.jhydrol.2007.09.030, 2007.Coble, P., Castillo, C., and Avril, B.: Distribution and optical properties of
CDOM in the Arabian Sea during the 1995 Southwest Monsoon, Deep-Sea Res. Pt.
II, 45, 2195–2223, 10.1016/S0967-0645(98)00068-X,
1998.Cole, J., Prairie, Y., Caraco, N., McDowell, W., Tranvik, L., Striegl, R.,
Duarte, C., Kortelainen, P., Downing, J., Middelburg, J., and Melack, J.:
Plumbing the Global Carbon Cycle: Integrating Inland Waters into the
Terrestrial Carbon Budget, Ecosystems, 10, 172–185,
10.1007/s10021-006-9013-8, 2007.Cory, R. M. and McKnight, D. M.: Fluorescence spectroscopy reveals
ubiquitous presence of oxidized and reduced quinines in dissolved organic
matter, Environ. Sci. Technol., 39, 8142–8149, 10.1021/es0506962,
2005.Creed, I. F., Beall, F. D., Clair, T. A., Dillon, P. J., and Hesslein, R. H.:
Predicting export of dissolved organic carbon from forested catchments in
glaciated landscapes with shallow soils, Glob. Biogeochem. Cy., 22,
GB4024, 10.1029/2008GB003294, 2008.
Creed, I. F., Sanford, S. E., Beall, F. D., Molot, L. A., and Dillon, P. J.:
Cryptic wetlands: integrating hidden wetlands in regression models of the
export of dissolved organic carbon from forested landscapes, Hydrol. Process., 17, 3629–3648, 2003.D'Amore, D. V., Edwards, R. T., and Biles, F. E.: Biophysical controls on
dissolved organic carbon concentrations of Alaskan coastal temperate
rainforest streams, Aquat. Sci., 2, 381–393,
10.1007/s00027-015-0441-4, 2015a.D'Amore, D. V., Edwards, R. T., Herendeen, P. A., Hood, E., and Fellman, J. B.:
Dissolved organic carbon fluxes from hydropedologic units in Alaskan coastal
temperate rainforest watersheds, Soil Sci. Soc. Am. J., 79, 378–388,
10.2136/sssaj2014.09.0380, 2015b.
D'Amore, D. V., Biles, F. E., Nay, M., and Rupp, T. S.: Watershed carbon budgets in
the southeastern Alaskan coastal forest region, in: Baseline and projected
future carbon storage and greenhouse-gas fluxes in ecosystems of Alaska,
US Geological Survey Professional Paper, 1826, 196 pp., 2016.Dai, M., Yin, Z., Meng, F., Liu, Q., and Cai, W.J.: Spatial distribution of
riverine DOC inputs to the ocean: an updated global synthesis, Curr. Opin.
Sust., 4, 170–178, 10.1016/j.cosust.2012.03.003, 2012.Deirmendjian, L., Loustau, D., Augusto, L., Lafont, S., Chipeaux, C.,
Poirier, D., and Abril, G.: Hydrological and ecological controls on dissolved
carbon concentrations in groundwater and carbon export to surface waters in a
temperate pine forest watershed, Biogeosciences Discuss.,
10.5194/bg-2017-90, in review, 2017.
DellaSala, D. A.: Temperate and Boreal Rainforests of the World, Island
Press, Washington, DC, 2011.Emili, L. and Price, J.: Biogeochemical processes in the soil-groundwater
system of a forest-peatland complex, north coast British Columbia, Canada,
Northwest Sci., 88, 326–348, 10.3955/046.087.0406, 2013.Fasching, C., Behounek, B., Singer, G., and Battin, T.: Microbial degradation
of terrigenous dissolved organic matter and potential consequences for carbon
cycling in brown-water streams, Sci. Rep., 4, 4981, 10.1038/srep04981,
2014.Fasching, C., Ulseth, A., Schelker, J., Steniczka, G., and Battin, T.:
Hydrology controls dissolved organic matter export and composition in an
Alpine stream and its hyporheic zone, Limnol. Oceanogr., 61, 558–571,
10.1002/lno.10232, 2016.Fellman, J., Hood, E., D'Amore, D., Edwards, R., and White, D.: Seasonal
changes in the chemical quality and biodegradability of dissolved organic
matter exported from soils to streams in coastal temperate rainforest
watersheds, Biogeochemistry, 95, 277–293, 10.1007/s10533-009-9336-6,
2009a.Fellman, J., Hood, E., Edwards, R., and D'Amore, D.: Changes in the
concentration, biodegradability, and fluorescent properties of dissolved
organic matter during stormflows in coastal temperate watersheds, J. Geophys.
Res.-Biogeo., 114, 10.1029/2008JG000790, 2009b.Fellman, J., Hood, E., and Spencer, R.: Fluorescence spectroscopy opens new
windows into dissolved organic matter dynamics in freshwater ecosystems: A
review, Limnol. Oceanogr., 55, 2452–2462, 10.4319/lo.2010.55.6.2452,
2010.Fellman, J., Nagorski, S., Pyare, S., Vermilyea, A. W., Scott, D., and Hood,
E.: Stream temperature response to variable glacier cover in coastal
watersheds of Southeast Alaska, Hydrol. Process., 28, 2062–2073,
10.1002/hyp.9742, 2014
Finlay, J. C. and Kendall, C.: Stable isotope tracing of temporal and spatial
variability in organic matter sources and variability in organic matter
sources to freshwater ecosytems, in: Stable Isotopes in Ecology and
Environmental Science, edited by: Michener, R. and Lajtha, K., Blackwell
Publishing Ltd, Oxford, UK, 2, 283–324, 2007.Fitzgerald, D., Price, J., and Gibson, J.: Hillslope-swamp interactions and
flow pathways in a hypermaritime rainforest, British Columbia, Hydrol.
Process., 17, 3005–3022, 10.1002/hyp.1279, 2003.Gibson, J. J., Price, J. S., Aravena, R., Fitzgerald, D. F., and Maloney, D.:
Runoff generation in a hypermaritime bog-forest upland, Hydrol. Process, 14,
2711–2730, 10.1002/1099-1085(20001030)14:15<2711::AID-HYP88>3.0.CO;2-2,
2000.
Glatzel, S., Kalbitz, K., Dalva, M., and Moore, T.: Dissolved organic matter
properties and their relationship to carbon dioxide efflux from restored peat
bogs, Geoderma, 113, 397–411, 2003.Gonzalez Arriola S., Frazer, G. W., and Giesbrecht, I.: LiDAR-derived
watersheds and their metrics for Calvert Island, Hakai Institute,
10.21966/1.15311, 2015.
Gorham, E., Lehman, C., Dyke, A., Clymo, D., and Janssens, J.: Long-term
carbon sequestration in North American peatlands, Quaternary Sci. Rev., 58,
77–82, 2012.Graeber, D., Gelbrecht, J., Pusch, M., Anlanger, C., and von Schiller, D.:
Agriculture has changed the amount and composition of dissolved organic
matter in Central European headwater streams, Sci. Total Environ., 438,
435–446, 10.1016/j.scitotenv.2012.08.087, 2012.
Green, R. N.: Reconnaissance level terrestrial ecosystem mapping of priority
landscape units of the coast EBM planning area: Phase 3, Prepared for British
Columbia Ministry Forests, Lands and Natural Resource Ops., Blackwell and
Associates, Vancouver, Canada, 2014.Guillemette, F. and Giorgio, P.: Reconstructing the various facets of
dissolved organic carbon bioavailability in freshwater ecosystems, Limnol.
Oceanogr., 56, 734–748, 10.4319/lo.2011.56.2.0734, 2011.Hansen, A. M., Kraus, T. E. C., Pellerin, B. A., Fleck, J. A., Downing, B.
D., and Bergamaschi, B. A.: Optical properties of dissolved organic matter
(DOM): Effects of biological and photolytic degradation, Limnol. Oceanogr.,
61, 1015–1032, 10.1002/lno.10270, 2016.Harrell, F. E. and Dupont, C.: Hmisc: Harrell Miscellaneous. R package
version 4.0-2. https://CRAN.R-project.org/package=Hmisc, 2016.Harrison, J., Caraco, N., and Seitzinger, S.: Global patterns and sources of
dissolved organic matter export to the coastal zone: Results from a spatially
explicit, global model, Global Biogeochem. Cy., 19,
10.1029/2005gb002480, 2005.Helms, J., Stubbins, A., Ritchie, J., Minor, E., Kieber, D., and Mopper, K.:
Absorption spectral slopes and slope ratios as indicators of molecular
weight, source, and photobleaching of chromophoric dissolved organic matter,
Limnol. Oceanogr., 53, 955–969, 10.4319/lo.2008.53.3.0955, 2008.Helton, A., Wright, M., Bernhardt, E., Poole, G., Cory, R., and Stanford, J.:
Dissolved organic carbon lability increases with water residence time in the
alluvial aquifer of a river floodplain ecosystem, J. Geophys. Res.-Biogeo.,
120, 693–706, 10.1002/2014JG002832, 2015.Hoffman, K. M., Gavin, D. G., Lertzman, K. P., Smith, D. J., and Starzomski,
B. M.: 13 000 years of fire history derived from soil charcoal in a British
Columbia coastal temperate rain forest, Ecosphere, 7, e01415,
10.1002/ecs2.1415, 2016.Hope, D., Billett, M. F., and Cresser, M. S.: A review of the export of
carbon in river water: Fluxes and processes, Environ. Pollut., 84, 301–324,
10.1016/0269-7491(94)90142-2, 1994.
Hopkinson, C. S., Buffam, I., Hobbie, J., Vallino, J., and Perdue, M.:
Terrestrial inputs of organic matter to coastal ecosystems: An
intercomparison of chemical characteristics and bioavailability,
Biogeochemistry, 43, 211–234, 1998.Hudson, N., Baker, A., and Reynolds, D.: Fluorescence analysis of dissolved
organic matter in natural, waste and polluted waters-a review, River Res.
Appl., 23, 631–649, 10.1002/rra.1005, 2007.Hurvich, C. M. and Tsai, C.: Regression and time series model selection in
small samples, Biometrika, 76, 297–307, 10.2307/2336663, 1989.
IUSS Working Group WRB: World Reference Base for Soil Resources,
International soil classification system for naming soils and creating
legends for soil maps, World Soil Resources Reports No. 106, Food and
Agricultural Organization of the United Nations, Rome, Italy, 2015.ISO Standard 9196: Liquid flow measurement in open channels – Flow
measurements under ice conditions, International Organization for
Standardization, available online at: www.iso.org (last access: 1
November 2016), 1992.ISO Standard 748: Hydrometry – Measurement of liquid flow in open channels
using current-meters or floats, International Organization for
Standardization, available online at: www.iso.org (last access:
1 November 2016), 2007.Johannessen, S. C., Potentier, G., Wright, C. A., Masson, D., and Macdonald,
R. W.: Water column organic carbon in a Pacific marginal sea (Strait of
Georgia, Canada), Mar. Environ. Res., 66, S49–S61,
10.1016/j.marenvres.2008.07.008, 2008.Johnson, M., Couto, E., Abdo, M., and Lehmann, J.: Fluorescence index as an
indicator of dissolved organic carbon quality in hydrologic flowpaths of
forested tropical watersheds, Biogeochemistry, 105, 149–157,
10.1007/s10533-011-9595-x, 2011.Johnson, P. C. D.: Extension of Nakagawa and
Schielzeth's RGLMM2 to random slopes models, Methods Ecol.
Evol., 5, 944–946, 10.1111/2041-210X.12225, 2014.Judd, K., Crump, B., and Kling, G.: Variation in dissolved organic matter
controls bacterial production and community composition, Ecology, 87,
2068–2079, 10.1890/0012-9658(2006)87[2068:VIDOMC]2.0.CO;2, 2006.Kalbitz, K., Schmerwitz, J., Schwesig, D., and Matzner, E.: Biodegradation of
soil-derived dissolved organic matter as related to its properties,
Geoderma, 113, 273–291, 10.1016/S0016-7061(02)00365-8, 2003.Kling, G., Kipphut, G., Miller, M., and O'Brien, W.: Integration of lakes and
streams in a landscape perspective: the importance of material processing on
spatial patterns and temporal coherence, Freshwater Biol., 43, 477–497,
10.1046/j.1365-2427.2000.00515.x, 2000.Koehler, A.-K., Murphy, K., Kiely, G., and Sottocornola, M.: Seasonal
variation of DOC concentration and annual loss of DOC from an Atlantic
blanket bog in South Western Ireland, Biogeochemistry, 95, 231–242,
10.1007/s10533-009-9333-9, 2009.
Lakowicz, J. R.: Principles of Fluorescence Spectroscopy, 2, Kluwer Academic,
New York, 1999.Larson, J. H., Frost, P. C., Zheng, Z., Johnston, C. A., Bridgham, S. D.,
Lodge, D. M., and Lamberti, G. A.: Effects of upstream lakes on dissolved
organic matter in streams, Limnol. Oceanogr., 52, 60–69,
10.4319/lo.2007.52.1.0060, 2007.Leighty, W. W., Hamburg, S. P., and Caouette, J.: Effects of management on
carbon sequestration in forest biomass in Southeast Alaska, Ecosystems, 9,
1051, 10.1007/s10021-005-0028-3, 2006.Lalonde, K., Middlestead, P., Gélinas, Y.: Automation of 13C/12C ratio
measurement for freshwater and seawater DOC using high temperature
combustion, Limnol. Oceanogr.-Meth., 12, 816–829,
10.4319/lom.2014.12.816, 2014.Lambert, T., Bouillon, S., Darchambeau, F., Massicotte, P., and Borges, A.
V.: Shift in the chemical composition of dissolved organic matter in the
Congo River network, Biogeosciences, 13, 5405–5420,
10.5194/bg-13-5405-2016, 2016.Leach, J., Larsson, A., Wallin, M., Nilsson, M., and Laudon, H.: Twelve year
interannual and seasonal variability of stream carbon export from a boreal
peatland catchment, J. Geophys. Res. 121, 1851–1866,
10.1002/2016JG003357, 2016.Legendre, P. and Durand, S.: rdaTest, Canonical redundancy analysis, R
package version 1.11, available at:
http://adn.biol.umontreal.ca/~numericalecology/Rcode/ (last access: 1
January 2017), 2014.Lepistö, A., Futter, M .N., and Kortelainen, P.: Almost 50 years of
monitoring shows that climate, not forestry, controls long-term organic
carbon fluxes in a large boreal watershed, Glob. Change Biol., 20,
1225–1237, 10.1111/gcb.12491, 2014.
Liaw, A. and Wiener, M.: Classification and Regression by randomForest, R
News, 2, 18–22, 2002.
Lochmuller, C. H. and Saavedra, S. S.: Conformational changes in a soil
fulvic acid measured by time dependent fluorescence depolarization, Anal.
Chem., 38, 1978–1981, 1986.Lorenz, D., Runkel, R., and De Cicco, L.: rloadest, River Load Estimation, R
package version 0.4.2, available at:
https://github.com/USGS-R/rloadest, 2015.Ludwig, W., Probst, J., and Kempe, S.: Predicting the oceanic input of
organic carbon by continental erosion, Global Biogeochem. Cy., 10, 23–41,
10.1029/95GB02925, 1996.Mann, P. J., Spencer, R. G. M., Dinga, B. J., Poulsen, J. R., Hernes, P. J.,
Fiske, G., Salter, M. E., Wang, Z. A., Hoering, K. A., Six, J., and Holmes,
R. M.: The biogeochemistry of carbon across a gradient of sreams and rivers
within the Congo Basin, J. Geophys. Res.-Biogeo., 119, 687–702,
10.1002/2013JG002442, 2014.
Marschner, B. and Kalbitz, K.: Controls on bioavailability and
biodegradability of dissolved organic matter in soils, Geoderma, 113,
211–235, 2003.Martin, S. L. and Soranno, P. A.: Lake landscape position: Relationships to
hydrologic connectivity and landscape features, Limnol. Oceanogr., 51,
801–814, 10.4319/lo.2006.51.2.0801, 2006.Masiello, C. A. and Druffel, E. R. M.: Carbon isotope geochemistry of the
Santa Clara River, Global Biogeochem. Cy., 15, 407–416,
10.1029/2000GB001290, 2001.Mayorga, E., Seitzinger, S., Harrison, J., Dumont, E., Beusen, A., Bouwman,
A. F., Fekete, B., Kroeze, C., and Drecht, G.: Global Nutrient Export from
WaterSheds 2 (NEWS 2): Model development and implementation, Environ. Model.
Softw., 25, 837–853, 10.1016/j.envsoft.2010.01.007, 2010.McClelland, J., Townsend-Small, A., Holmes, R., Pan, F., Stieglitz, M.,
Khosh, M., and Peterson, B.: River export of nutrients and organic matter
from the North Slope of Alaska to the Beaufort Sea, Water Resour. Res., 50,
1823–1839, 10.1002/2013WR014722, 2014.McKnight, D., Boyer, E., Westerhoff, P., Doran, P., Kulbe, T., and Andersen,
D.: Spectrofluorometric characterization of dissolved organic matter for
indication of precursor organic material and aromaticity, Limnol. Oceanogr.,
46, 38–48, 10.4319/lo.2001.46.1.0038, 2001.
McLaren, D., Fedje, D., Hay, M. B., Mackie, Q., Walker, I. J., Shugar, D. H.,
Eamer, J. B. R., Lian, O. B., and Neudorf, C.: A post-glacial sea level hinge
on the central Pacific coast of Canada, Quaternary Sci. Rev.., 97, 148–169,
2014.Meybeck, M.: Carbon, nitrogen, and phosphorus transport by world rivers, Am.
J. Sci., 282, 401–450, available from:
http://earth.geology.yale.edu/~ajs/1982/04.1982.01.Maybeck.pdf (last
access: 11 August 2017), 1982.
Milliman, J. D. and Syvitski J. P. M.: Geomorphic tectonic control of
sediment discharge to the ocean: The importance of small mountainous rivers,
J. Geol., 100, 525–544, 1992.
Moore, R. D.: Introduction to salt dilution gauging for streamflow
measurement part III: Slug injection using salt in solution, Streamline
Watershed Management Bulletin, 8, 1–6, 2005.Morrison, J., Foreman, M. G. G., and Masson, D.: A method for estimating
monthly freshwater discharge affecting British Columbia coastal waters,
Atmos.-Ocean, 50, 1–8, 10.1080/07055900.2011.637667, 2012.Mulholland, P. and Watts, J.: Transport of organic carbon to the oceans by
rivers of North America: a synthesis of existing data, Tellus, 34, 176–186,
10.1111/j.2153-3490.1982.tb01805.x, 1982.Murphy, K., Stedmon, C., Graeber, D., and Bro, R.: Fluorescence spectroscopy
and multi-way techniques. PARAFAC, Anal. Methods, 5, 6557–6566,
10.1039/C3AY41160E, 2013.Murphy, K., Stedmon, C., Wenig, P., and Bro, R.: OpenFluor – A spectral
database of auto-fluorescence by organic compounds in the environment, Anal.
Methods, 6, 658–661, 10.1039/C3AY41935E, 2014.Naiman, R. J.: Characteristics of sediment and organic carbon export from
pristine boreal forest watersheds, Can. J. Fish. Aquat. Sci., 39, 1699–1718,
10.1139/f82-226, 1982.Nakagawa, S. and Schielzeth, H.: A general and simple method for
obtaining R2 from generalized linear mixed-effects models, Methods Ecol.
Evol., 4, 133–142, 10.1111/j.2041-210x.2012.00261.x, 2013.Olefeldt, D., Roulet, N., Giesler, R., and Persson, A.: Total waterborne
carbon export and DOC composition from ten nested subarctic peatland
catchments-importance of peatland cover, groundwater influence, and
inter-annual variability of precipitation patterns, Hydrol. Process., 27,
2280–2294, 10.1002/hyp.9358, 2013.Oliver, A. A., Tank, S. E. , Giesbrecht, I., Korver, M. C., Floyd, W. C.,
Sanborn, P., Bulmer, C., and Lertzman, K. P.: Aquatic carbon flux data
package, 10.21966/1.321324, 2017
Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., and R Core Team: nlme:
Linear and Nonlinear Mixed Effects Models, R package version 3.1-128, 2016.
Pojar, J., Klinka, K., and Demarchi, D. A.: Chapter 6, Coastal Western
Hemlock Zone, in: Special Report Series 6, Ecosystems of British Columbia,
edited by: Meidiner, D. and Pojar, J., Ministry of Forests, British Columbia,
Victoria, 330 pp., 1991.Poulin, B., Ryan, J., and Aiken, G.: Effects of iron on optical properties of
dissolved organic matter, Environ. Sci. Technol., 48, 10098–106,
10.1021/es502670r, 2014.R Core Team, R: A language and environment for statistical computing, R
Foundation for Statistical Computing, Vienna, Austria,
http://www.R-project.org/ (last access: 11 August 2017), 2013.Raymond, P., Saiers, J., and Sobczak, W.: Hydrological and biogeochemical
controls on watershed dissolved organic matter transport: pulse-shunt
concept, Ecology, 97, 5–16, 10.1890/14-1684.1, 2016.Regnier, P., Friedlingstein, P., Ciais, P., Mackenzie, F., Gruber, N.,
Janssens, I., Laruelle, G., Lauerwald, R., Luyssaert, S., Andersson, A.,
Arndt, S., Arnosti, C., Borges, A., Dale, A., Gallego-Sala, A., Goddéris,
Y., Goossens, N., Hartmann, J., Heinze, C., Ilyina, T., Joos, F., LaRowe, D.,
Leifeld, J., Meysman, F., Munhoven, G., Raymond, P., Spahni, R.,
Suntharalingam, P., and Thullner, M.: Anthropogenic perturbation of the carbon
fluxes from land to ocean, Nat. Geosci., 6, 597–607, 10.1038/ngeo1830,
2013.
Roddick, J. R.: Geology, Rivers Inlet-Queens Sound, British Columbia, Open
File 3278, Geological Survey of Canada, Ottawa, Canada, 1996.
Royer, T. C.: Coastal fresh water discharge in the northeast, Pacific, J.
Geophys. Res., 87, 2017–2021, 1982.
Runkel, R. L., Crawford, C. G., and Cohn, T. A.: Load Estimator (LOADEST): A
FORTRAN program for estimating constituent loads in streams and rivers, U.S.
Geological Survey Techniques and Methods Book 4, Chapter A5, 65 pp., 2004.Sanderman, J., Lohse, K. A., Baldock, J. A., and Amundson, R.: Linking soils
and streams: Sources and chemistry of dissolved organic matter in a small
coastal watershed, Water Resourc. Res., 45, W03418, 10.1029/2008WR006977,
2009.Spencer, R., Butler, K., and Aiken, G.: Dissolved organic carbon and
chromophoric dissolved organic matter properties of rivers in the USA, J.
Geophys. Res.-Biogeo., 117, G03001, 10.1029/2011JG001928, 2012.Spencer, R. G., Hernes, P. J., Ruf, R., Baker, A., Dyda, R. Y., Stubbins, A.,
and Six, J.: Temporal controls on dissolved organic matter and lignin
biogeochemistry in a pristine tropical river, Democratic Republic of
Congo, J. Geophys. Res., 115, G03013, 10.1029/2009JG001180, 2010.
Stackpoole, S. M., Butman, D. E., Clow, D. W., Verdin, K. L., Gaglioti, B.,
and Striegl, R.: Carbon burial, transport, and emission from inland aquatic
ecosystems in Alaska, in: Baseline and projected future carbon storage and
greenhouse-gas fluxes in ecosystems of Alaska, edited by: Zhiliang, Z., and
David, A., US Geological Survey Professional Paper, 1826, 196 pp., 2016.Stackpoole, S. M., Butman, D. E., Clow, D. W., Verdin, K. L., Gaglioti, B.
V., Genet, H., and Striegl, R. G.: Inland waters and their role in the carbon
cycle of Alaska, Ecol. Appl., 27, 1403–1420, 10.1002/eap.1552, 2017.Stedmon, C. and Bro, R.: Characterizing dissolved organic matter fluorescence
with parallel factor analysis: a tutorial, Limnol. Oceanogr.-Meth., 6,
572–579, 10.4319/lom.2008.6.572b, 2008.Stedmon, C. and Markager, S.: Tracing the production and degradation of
autochthonous fractions of dissolved organic matter by fluorescence analysis,
Limnol. Oceanogr., 50, 1415–1426, 10.4319/lo.2005.50.5.1415, 2005.Stedmon, C., Markager, S., Bro, R., Stedmon, C., Markager, S., and Bro, R.:
Tracing dissolved organic matter in aquatic environments using a new approach
to fluorescence spectroscopy, Mar. Chem., 82, 239–254,
10.1016/S0304-4203(03)00072-0, 2003.
Stevenson, F. J.: Humus Chemistry: Genesis, Composition, Reactions, 2, Jon
Wiley and Sons Inc., New York, United States of America, 1994.Symonds, M. R. E. and Moussalli, A.: A brief guide to model selection,
multimodel inference, and model averaging in behavioural ecology using
Akaike's information criterion, Behav. Ecol. Sociobiol., 65, 13–21,
10.1007/s00265-010-1037-6, 2011.Tallis, H.: Kelp and rivers subsidize rocky intertidal communities in the
Pacific Northwest (USA), Mar. Ecol.-Prog. Ser., 389, 8596,
10.3354/meps08138, 2009.Tank, S., Raymond, P., Striegl, R., McClelland, J., Holmes, R., Fiske, G.,
and Peterson, B.: A land-to-ocean perspective on the magnitude, source and
implication of DIC flux from major Arctic rivers to the Arctic Ocean, Global
Biogeochem. Cy., 26, GB4018, 10.1029/2011GB004192, 2012.Tank, S., Striegl, R. G., McClelland, J. W., and Kokelij, S. V.:
Multi-decadal increases in dissolved organic carbon and alkalinity flux from
the Mackenzie drainage basin to the Arctic Ocean, Environ. Res. Lett., 11,
10.1088/1748-9326/11/5/054015, 2016.Thompson, S. D., Nelson, T. A., Giesbrecht, I., Frazer, G., and Saunders, S.
C.: Data-driven regionalization of forested and non-forested ecosystems in
coastal British Columbia with LiDAR and RapidEye imagery, Appl. Geogr., 69,
35–50, 10.1016/j.apgeog.2016.02.002, 2016.Trant, A. J., Niijland, W., Hoffman, K. M., Mathews, D. L., McLaren, D.,
Nelson, T. A., and Starzomski, B. M.: Intertidal resource use over millennia
enhances forest productivity, Nat. Commun., 7, 12491,
10.1038/ncomms12491, 2016.van Hees, P., Jones, D., Finlay, R., Godbold, D., and Lundström, U.: The
carbon we do not see-the impact of low molecular weight compounds on carbon
dynamics and respiration in forest soils: a review, Soil Biol. Biochem., 37,
1–13, 10.1016/j.soilbio.2004.06.010, 2005.Wallin, M., Weyhenmeyer, G., Bastviken, D., Chmiel, H., Peter, S., Sobek, S.,
and Klemedtsson, L.: Temporal control on concentration, character, and export
of dissolved organic carbon in two hemiboreal headwater streams draining
contrasting catchments, J. Geophys. Res.-Biogeo., 120, 832–846,
10.1002/2014jg002814, 2015.
Wang, T., Hamann, A., Spittlehouse, D. L., and Murdock, T. Q.: ClimateWNA-
High resolution spatial climate data for Western North America, J. Appl.
Meterol. Climatol., 51, 16–29, 10.1175/JAMC-D-11-043.1, 2012.Weishaar, J. L., Aiken, G. R., Bergamaschi, B. A., Fram, M. S., Fujii, R.,
and Mopper, K.: Evaluation of specific ultraviolet absorbance as an indicator
of the chemical composition and reactivity of dissolved organic carbon,
Environ. Sci. Technol., 37, 4702–4708, 10.1021/es030360x, 2003.
Whitney, F. A., Crawford, W. R., and Harrison, P. J.: Physical processes that
enhance nutrient transport and primary productivity in the coastal and open
ocean of the subarctic NE Pacific, Deep-Sea Res. Pt. II, 52, 681–706, 2005.
Wickland, K., Neff, J., and Aiken, G.: Dissolved Organic Carbon in Alaskan
Boreal Forest: Sources, Chemical Characteristics, and
Biodegradability, Ecosystems, 10, 1323–1340, 2007.Wilson, H. F. and Xenopoulos, M. A.: Effects of agricultural land use on the
composition of fluvial dissolved organic matter, Nat. Geosci., 2, 37–41,
10.1038/ngeo391, 2009.Wolf, E. C., Mitchell, A. P., and Schoonmaker, P. K.: The Rain Forests of
Home: An Atlas of People and Place, Ecotrust, Pacific GIS, Inforain, and
Conservation International, Portland, Oregon, 24 pp., available at:
http://www.inforain.org/pdfs/ctrf_atlas_orig.pdf, 1995.Worrall, F., Burt, T., and Adamson, J.: Can climate change explain increases
in DOC flux from upland peat catchements?, Sci. Total. Environ., 326,
95–112, 10.1016/j.scitotenv.2003.11.022, 2004.
Xenopoulos, M. A., Lodge, D. M., Frentress, J., Kreps, T. A., Bridgham, S.
D., Grossman, E., and Jackson, C. J.: Regional comparisons of watershed
determinants of dissolved organic carbon in temperate lakes from the Upper
Great Lakes region and selected regions globally, Limnol. Oceanogr., 48,
2321–2334, 2003.Yamashita, Y. and Jaffeì, R.: Characterizing the Interactions between
Trace Metals and Dissolved Organic Matter Using Excitation–Emission Matrix
and Parallel Factor Analysis, Environ. Sci. Technol., 42, 7374–7379,
10.1021/es801357h, 2008.Yamashita, Y., Kloeppel, B., Knoepp, J., Zausen, G., and Jaffé, R.:
Effects of Watershed History on Dissolved Organic Matter Characteristics in
Headwater Streams, Ecosystems, 14, 1110–1122,
10.1007/s10021-011-9469-z, 2011.