a global hotspot for dissolved organic carbon in

Biogeosciences 14 3743ndash3762 2017httpsdoiorg105194bg-14-3743-2017copy Author(s) 2017 This work is distributed underthe Creative Commons Attribution 30 License

A global hotspot for dissolved organic carbon in hypermaritimewatersheds of coastal British ColumbiaAllison A Oliver12 Suzanne E Tank12 Ian Giesbrecht27 Maartje C Korver2 William C Floyd342Paul Sanborn52 Chuck Bulmer6 and Ken P Lertzman72

1University of Alberta Department of Biological Sciences CW 405 Biological Sciences Bldg University of AlbertaEdmonton AB T6G 2E9 Canada2Hakai Institute Tula Foundation PO Box 309 Heriot Bay BC V0P 1H0 Canada3Ministry of Forests Lands and Natural Resource Operations 2100 Labieux Rd Nanaimo BC V9T 6E9 Canada4Vancouver Island University 900 Fifth Street Nanaimo BC V9R 5S5 Canada5Ecosystem Science and Management Program University of Northern British Columbia3333 University Way Prince George BC V2N 4Z9 Canada6BC Ministry of Forests Lands and Natural Resource Operations 3401 Reservoir Rd Vernon BC V1B 2C7 Canada7School of Resource and Environmental Management Simon Fraser University TASC 1 ndash Room 84058888 University Drive Burnaby BC V5A 1S6 Canada

Correspondence to Allison A Oliver (aaoliverualbertaca)

Received 10 January 2017 ndash Discussion started 19 January 2017Revised 29 June 2017 ndash Accepted 10 July 2017 ndash Published 15 August 2017

Abstract The perhumid region of the coastal temperate rain-forest (CTR) of Pacific North America is one of the wettestplaces on Earth and contains numerous small catchments thatdischarge freshwater and high concentrations of dissolvedorganic carbon (DOC) directly to the coastal ocean How-ever empirical data on the flux and composition of DOCexported from these watersheds are scarce We establishedmonitoring stations at the outlets of seven catchments onCalvert and Hecate islands British Columbia which rep-resent the rain-dominated hypermaritime region of the per-humid CTR Over several years we measured stream dis-charge stream water DOC concentration and stream wa-ter dissolved organic-matter (DOM) composition Dischargeand DOC concentrations were used to calculate DOC fluxesand yields and DOM composition was characterized usingabsorbance and fluorescence spectroscopy with parallel fac-tor analysis (PARAFAC) The areal estimate of annual DOCyield in water year 2015 was 333 Mg C kmminus2 yrminus1 withindividual watersheds ranging from an average of 241 to377 Mg C kmminus2 yrminus1 This represents some of the highestDOC yields to be measured at the coastal margin We ob-served seasonality in the quantity and composition of ex-ports with the majority of DOC export occurring during the

extended wet period (SeptemberndashApril) Stream flow fromcatchments reacted quickly to rain inputs resulting in rapidexport of relatively fresh highly terrestrial-like DOM DOCconcentration and measures of DOM composition were re-lated to stream discharge and stream temperature and corre-lated with watershed attributes including the extent of lakesand wetlands and the thickness of organic and mineral soilhorizons Our discovery of high DOC yields from these smallcatchments in the CTR is especially compelling as they de-liver relatively fresh highly terrestrial organic matter directlyto the coastal ocean Hypermaritime landscapes are commonon the British Columbia coast suggesting that this coastalmargin may play an important role in the regional processingof carbon and in linking terrestrial carbon to marine ecosys-tems

1 Introduction

Freshwater aquatic ecosystems process and transport a sig-nificant amount of carbon (Cole et al 2007 Aufdenkampeet al 2011 Dai et al 2012) Globally riverine export isestimated to deliver around 09 Pg C yrminus1 from land to the

Published by Copernicus Publications on behalf of the European Geosciences Union

3744 A A Oliver et al A global hotspot for dissolved organic carbon

coastal ocean (Cole et al 2007) with typically gt 50 quan-tified as dissolved organic carbon (DOC) (Meybeck 1982Ludwig et al 1996 Alvarez-Cobelas et al 2012 Mayorgaet al 2010) Rivers draining coastal watersheds serve as con-duits of DOC from terrestrial and freshwater sources to ma-rine environments (Mulholland and Watts 1982 Bauer et al2013 McClelland et al 2014) and can have important im-plications for coastal carbon cycling biogeochemical inter-actions ecosystem productivity and food webs (Hopkinsonet al 1998 Tallis 2009 Tank et al 2012 Regnier et al2013) In addition because the transfer of water and organicmatter from watersheds to the coastal ocean represents an im-portant pathway for carbon cycling and ecological subsidiesbetween ecosystems better understanding of these linkagesis needed for constraining predictions of ecosystem produc-tivity in response to perturbations such as climate change Inregions where empirical data are currently scarce quantify-ing land-to-ocean DOC export is therefore a priority for im-proving the accuracy of watershed and coastal carbon models(Bauer et al 2013)

While quantifying DOC flux within and across systemsis required for understanding the magnitude of carbon ex-change the composition of DOC (as dissolved organic mat-ter or DOM) is also important for determining the ecologi-cal significance of carbon exported from coastal watershedsThe aquatic DOM pool is a complex mixture that reflectsboth source material and processing along the watershedterrestrialndashaquatic continuum and as a result it can show sig-nificant spatial and temporal variation (Hudson et al 2007Graeber et al 2012 Wallin et al 2015) Both DOC con-centration and DOM composition can serve as indicators ofwatershed characteristics (Koehler et al 2009) hydrologicflow paths (Johnson et al 2011 Helton et al 2015) and wa-tershed biogeochemical processes (Emili and Price 2013)DOM composition can also influence its role in downstreamprocessing and ecological function such as susceptibility tobiological (Judd et al 2006) and physiochemical interac-tions (Yamashita and Jaffeacute 2008)

The coastal temperate rainforests (CTRs) of Pacific NorthAmerica extend from the Gulf of Alaska through BritishColumbia to Northern California and span a wide rangeof precipitation and climate regimes Within this rainforestregion the ldquoperhumidrdquo zone has cool summers and sum-mer precipitation is common (gt 10 of annual precipita-tion) (Alaback 1996) (Fig 1) The perhumid CTR extendsfrom southeast Alaska through the outer coast of centralBritish Columbia and contains forests and soils that have ac-cumulated large amounts of organic carbon above and be-low ground (Leighty et al 2006 Gorham et al 2012) Dueto high amounts of precipitation and close proximity to thecoast this area represents a potential hotspot for the trans-port and metabolism of carbon across the land-to-ocean con-tinuum and quantifying these fluxes is pertinent for under-standing global carbon cycling

Within the large perhumid CTR there is substantialspatial variation in climate and landscape characteristicsthat create uncertainty about carbon cycling and patternIn Alaska for example riverine DOC concentrations varywith wetland cover (DrsquoAmore et al 2015a) and glacialcover (Fellman et al 2014) Previous studies have shownthat streams in southeast Alaska can contain high DOCconcentrations (Fellman et al 2009a DrsquoAmore et al2015a) and produce high DOC yields (DrsquoAmore et al2015a b Stackpoole et al 2016) but no known fieldestimates have been generated for the perhumid CTR ofBritish Columbia an area of approximately 97 824 km2

(adapted from Wolf et al 1995) Within the perhumid CTRof British Columbia terrestrial ecologists have defined alarge (29 935 km2) ldquohypermaritimerdquo subregion where rain-fall dominates over snow seasonality is moderated by theocean and wetlands are extensive (Pojar et al 1991 areaestimated using British Columbia Biogeoclimatic Ecosys-tem Classification SubzoneVariant mapping Version 1031 August 2016 available at httpscataloguedatagovbccadatasetf358a53b-ffde-4830-a325-a5a03ff672c3) Previ-ous work in the hypermaritime CTR showed that DOC con-centrations are high in small streams and tend to increaseduring rain events (Gibson et al 2000 Fitzgerald et al2003 Emili and Price 2013) Taken together these condi-tions should be expected to generate high yields and fluxes ofDOC 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 ex-tensive hypermaritime region of British Columbiarsquos perhu-mid CTR to describe the temporal and spatial dynamicsof exported DOC concentration and DOM composition andto identify relationships between DOC concentration DOMcomposition and watershed characteristics

2 Methods

21 Study sites

Study sites are located on northern Calvert Island and adja-cent Hecate Island on the central coast of British ColumbiaCanada (lat 51650 long minus128035 Fig 1) Average annualprecipitation and air temperature at sea level from 1981 to2010 was 3356 mm yrminus1 and 84 C (average annual min09 C average annual max 179 C) (available online athttpwwwclimatewnacom Wang et al 2012) with pre-cipitation dominated by rain and winter snowpack persist-ing only at higher elevations Sites are located within thehypermaritime region of the CTR on the outer coast ofBritish Columbia Soils overlying the granodiorite bedrock(Roddick 1996) are usually lt 1 m thick and have formedin sandy colluvium and patchy morainal deposits with lim-ited areas of coarse glacial outwash Chemical weatheringand organic-matter accumulation in the cool moist climate

Biogeosciences 14 3743ndash3762 2017 wwwbiogeosciencesnet1437432017

A A Oliver et al A global hotspot for dissolved organic carbon 3745

Tabl

e1

Wat

ersh

edch

arac

teri

stic

sdi

scha

rge

DO

Cco

ncen

trat

ions

and

DO

Cyi

elds

for

the

seve

nst

udy

wat

ersh

eds

onC

alve

rtan

dH

ecat

eis

land

sA

dditi

onal

deta

ilson

the

met

hods

used

tode

term

ine

wat

ersh

edch

arac

teri

stic

sca

nbe

foun

din

the

Supp

lem

ent

Wat

ersh

edA

rea

Avg

L

akes

Wet

land

sA

vgd

epth

Avg

dep

thTo

tal

DO

C1

aQ

-wei

ghte

dD

OC

DO

CD

OC

slop

eor

gani

cm

iner

alQ

yiel

d1av

gD

OC

1an

nual

mon

thly

mon

thly

soils

soils

yiel

dbyi

eldb

yiel

db

WY

2015

1w

etse

ason

2dr

yse

ason

3

(km

2 )(

)(

Are

a)(

Are

a)(c

m)

(cm

)(m

m)

(mg

Lminus

1 )(m

gLminus

1 )(M

gC

kmminus

2 )(M

gC

kmminus

2 )(M

gC

kmminus

2 )

626

32

217

47

480

394plusmn

243

308plusmn

83

3673

110plusmn

35

153

377

359

062

(31

9ndash44

2)

(30

5ndash4

18)

(04

9ndash0

77)

1015

33

342

91

238

395plusmn

172

337plusmn

86

3052

112plusmn

16

129

247

256

027

(23

6ndash25

8)

(24

5ndash2

78)

(02

5ndash0

28)

819

48

301

03

502

379plusmn

191

298plusmn

57

3066

140plusmn

35

193

357

380

057

(31

7ndash40

2)

(33

7ndash5

10)

(04

8ndash0

67)

844

57

325

03

352

354plusmn

180

291plusmn

64

4129

131plusmn

36

159

436

424

054

(34

2ndash54

9)

(33

6ndash5

30)

(03

6ndash0

77)

708

78

285

75

463

362plusmn

197

299plusmn

60

3805

95plusmn

24

109

241

267

038

(22

2ndash26

0)

(24

6ndash4

07)

(03

4ndash0

43)

693

93

302

44

428

354plusmn

161

302plusmn

64

5866

77plusmn

25

84

297

319

041

(25

9ndash34

0)

(27

9ndash4

94)

(03

2ndash0

52)

703

128

403

19

243

373plusmn

165

358plusmn

134

6058

63plusmn

26

90

370

348

064

(32

5ndash42

0)

(30

7ndash4

02)

(05

2ndash0

77)

All

469

327

37

371

374plusmn

177

322plusmn

92

4730

104plusmn

38

111

333

335

050

(28

9ndash38

1)

(29

4ndash4

40)

(04

1ndash0

62)

1C

alcu

late

dfo

rwat

erye

ar20

15(W

Y20

151

Oct

ober

2014

ndash30

Sept

embe

r201

5)2

Wet

-per

iod

aver

age

mon

thly

yiel

dca

lcul

ated

from

Oct

ober

toA

pril

and

Sept

embe

rW

Y20

15a

ndO

ctob

erto

Apr

ilW

Y20

163

Dry

-per

iod

aver

age

mon

thly

yiel

dca

lcul

ated

from

May

toA

ugus

tW

Y20

15a

Mea

nplusmn

stan

dard

devi

atio

nb

Tota

lplusmn95

co

nfide

nce

inte

rval

wwwbiogeosciencesnet1437432017 Biogeosciences 14 3743ndash3762 2017


Figure 1 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 andlocation of the rain gauge (left) Characteristics of individual watersheds are described in Table 1

have produced soils dominated by Podzols and Folic His-tosols with Hemists up to 2 m thick at depressional sites(IUSS Working Group WRB 2015) The landscape is com-prised of a mosaic of ecosystem types including exposedbedrock extensive wetlands ldquobog forestsrdquo and woodlandswith organic-rich soils (Green 2014 Thompson et al 2016)Forest stands are generally short with open canopies reflect-ing the lower productivity of the hypermaritime forests com-pared to the rest of the perhumid CTR (Banner et al 2005)Dominant trees are western red cedar yellow cedar shorepine and western hemlock with composition varying acrosstopographic and edaphic gradients Widespread understoryplants include bryophytes salal deer fern and tufted clu-brush Wetland plants are locally abundant including diverseSphagnum mosses and sedges Although the watersheds haveno history of mining or industrial logging archaeological ev-idence suggests that humans have occupied this landscapefor at least 13 000 years (McLaren et al 2014) This occupa-tion has had a local effect on forest productivity near habita-tion sites (Trant et al 2016) and on fire regimes (Hoffman etal 2016) We selected seven watersheds with streams drain-ing directly into the ocean (Fig 1) These numbered water-sheds (626 693 703 708 819 844 and 1015) range in size(32 to 128 km2) and topography (maximum elevation 160 to1012 m) are variably affected by lakes (03ndash91 lake cov-erage) and ndash as is characteristic of the perhumid CTR ndash havea high degree of wetland coverage (24ndash50 ) (Table 1)

22 Soils and watershed characteristics

Watersheds and streams were delineated using a 3 m reso-lution digital elevation model (DEM) derived from airbornelaser scanning (lidar) and flow accumulation analysis usinggeographic information systems (GISs) to summarize wa-tershed characteristics for each watershed polygon and forall watersheds combined (Gonzalez Arriola et al 2015 Ta-ble 1) Topographic measures were estimated from the DEMand lake and wetland cover was estimated from the Provinceof British Columbia terrestrial ecosystem mapping (TEM)(Green 2014) and soil material thickness was estimatedfrom unpublished digital soil maps (Supplement S1) Werecorded the thickness of organic soil material the thicknessof mineral soil material and total soil depth to bedrock at atotal of 353 field sites Mineral soil horizons have le 17 or-ganic C while organic soil horizons have gt 17 organic Cper the Canadian System of Soil Classification (Soil Classi-fication Working Group 1998) In addition to field-sampledsites 40 sites with exposed bedrock (0 cm soil depth) werelocated using aerial photography Soil thicknesses were com-bined with a suite of topographic vegetation and remote-sensing (lidar and RapidEye satellite imagery) data for eachsampling point and used to train a random forest model (ran-domForest package in R Liaw and Wiener 2002) that wasused to predict soil depth values Soil material thicknesseswere then averaged for each watershed (Table 1) For ad-ditional details on field site selection and methods used forpredictions of soil thickness see Supplement S11



23 Sample collection and analysis

From May 2013 to July 2016 we collected stream water grabsamples from each watershed stream outlet every 2ndash3 weeks(ntotal = 402) with less frequent sampling (sim monthly) dur-ing winter (Fig 1) All samples were filtered in the field(Millipore Millex-HP Hydrophilic PES 045 microm) and kept inthe dark on ice until analysis DOC samples were filteredinto 60 mL amber glass bottles and preserved with 75 MH3PO4 Fe samples were filtered into 125 mL HDPE bottlesand preserved with 8 M HNO3 DOC and Fe samples wereanalyzed at the BC Ministry of the Environment TechnicalServices Laboratory (Victoria BC Canada) DOC concen-trations were determined on a total organic carbon (TOC) an-alyzer (Aurora 1030 OI Analytical) using wet chemical oxi-dation with persulfate followed by infrared detection of CO2Fe concentrations were determined on a dual-view induc-tively 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 forstable isotopic composition of δ13C in DOC (δ13C-DOCn= 173) and the optical characterization of DOM us-ing absorbance spectroscopy (n= 259) Beginning in Jan-uary 2016 we also analyzed samples using fluorescencespectroscopy (see Sect 26) Samples collected for δ13C-DOC were filtered into 40 mL EPA glass vials and preservedwith H3PO4 δ13C-DOC samples were analyzed at GG HatchStable Isotope Laboratory (Ottawa ON Canada) using high-temperature combustion (TIC-TOC Combustion AnalyserModel 1030 OI Analytical) coupled to a continuous-flowisotope ratio mass spectrometry (Finnigan Mat DeltaPlusXPThermo Fischer Scientific) (Lalonde et al 2014) Samplesanalyzed for optical characterization using absorbance andfluorescence were filtered into 125 mL amber HDPE bot-tles and analyzed at the Hakai Institute (Calvert Island BCCanada) within 24 h of collection

24 Hydrology precipitation and stream discharge

We measured precipitation using a TB4-L tipping bucket raingauge with a 02 mm resolution (Campbell Scientific Ltd)located in watershed 708 (elevation 16 m asl) The raingauge was calibrated twice per year using a field calibrationdevice model 653 (HYQUEST Solutions Ltd)

We determined continuous stream discharge for each wa-tershed by developing stage discharge rating curves at fixedhydrometric stations situated in close proximity to eachstream outlet Sites were located above tidewater influenceand were selected based on favourable conditions (ie chan-nel stability and stable hydraulic conditions) for the installa-tion and operation of pressure transducers to measure streamstage From August 2014 to May 2016 (21 months) we mea-sured stage every 5 min using a pressure level sensor (OTTPLS-L OTT Hydromet Colorado USA) pressure trans-

ducer (0ndash4 m range SDI-12) connected to a CR1000 (Camp-bell Scientific Edmonton Canada) data logger Stream dis-charge was measured over various intervals using eitherthe velocity area method (for flows lt 05 m3 sminus1 ISO Stan-dard 9196 1992 ISO Standard 748 2007) or salt dilution(for flows gt 05 m3 sminus1 Moore 2005) Rating curves weredeveloped using the relationship between stream stage heightand stream discharge (Supplement S2)

25 DOC flux

From 1 October 2014 to 30 April 2016 we estimated DOCflux for each watershed using measured DOC concentrations(n= 224) and continuous discharge recorded at 15 min inter-vals The watersheds in this region respond rapidly to raininputs and as a result DOC concentrations are highly vari-able To address this variability routine DOC concentrationdata (as described in Sect 22) were supplemented with ad-ditional grab samples (n= 21) collected around the peak ofthe hydrograph during several high-flow events throughoutthe year We performed watershed-specific estimates of DOCflux using the rloadest package (Lorenz et al 2015) in R(version 325 R Core Team 2016) which replicates func-tions developed in the US Geological Survey load-estimatorprogram LOADEST (Runkel et al 2004) LOADEST is amultiple-regression adjusted maximum likelihood (ML) esti-mation model that calibrates a regression between measuredconstituent values and stream flow across seasons and timeand then fits it to combinations of coefficients representingnine predetermined models of constituent flux To accountfor potentially small sample size the best model was se-lected using the second-order Akaike Information Criterion(AICc) (Akaike 1981 Hurvich and Tsai 1989) Input datawere log-transformed to avoid bias and centered to reducemulticollinearity For additional details on model selectionsee Supplement Table S31

26 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 EnvironmentTechnical Services Laboratory (Victoria BC Canada) to de-termine specific UV absorption at 254 nm (SUVA254) AfterMay 2014 we conducted optical characterization of DOMby absorbance and fluorescence spectroscopy at the HakaiInstitute field station (Calvert Island BC Canada) using anAqualog fluorometer (Horiba Scientific Edison New JerseyUSA) Strongly absorbing samples (absorbance units gt 02at 250 nm) were diluted prior to analysis to avoid excessiveinner filter effects (Lakowicz 1999) Samples were run in1 cm quartz cells and scanned from 220 to 800 nm at 2 nmintervals to determine SUVA254 as well as the spectral sloperatio (SR) SUVA254 has been shown to positively correlatewith increasing molecular aromaticity associated with the



fulvic acid fraction of DOM (Weishaar et al 2003) and itis calculated by dividing the decadic absorption coefficientat 254 nm by DOC concentration (mg C Lminus1) To accountfor potential Fe interference with absorbance values we cor-rected SUVA254 values by Fe concentration according themethod described in Poulin et al (2014) SR has been shownto negatively correlate with molecular weight (Helms et al2008) and is calculated as the ratio of the spectral slope from275 to 295 nm (S275minus295) to the spectral slope from 350 to400 nm (S350minus400)

We measured excitation and emission spectra (as excita-tion emission matrices EEMs) on samples every three weeksfrom January to July 2016 (n= 63) Samples were run in1 cm quartz cells and scanned from excitation wavelengthsof 230ndash550 nm at 5 nm increments and emission wavelengthsof 210ndash620 nm at 2 nm increments The Horiba Aqualog ap-plied the appropriate instrument corrections for excitationand emission inner filter effects and Raman signal calibra-tion We calculated the fluorescence index and freshness in-dex for each EEM The fluorescence index is often used toindicate DOM source where higher values are more indica-tive of microbially derived sources of DOM and lower valuesindicate more terrestrially derived sources (McKnight et al2001) and is calculated as the ratio of emission intensity at450 to 500 nm at an excitation of 370 nm The freshness in-dex is used to indicate the contribution of autochthonous orrecently microbial-produced DOM with higher values sug-gesting greater autochthony (ie microbial inputs) and iscalculated as the ratio of emission intensity at 380 nm to themaximum emission intensity between 420 and 435 nm at anexcitation of 310 nm (Wilson and Xenopoulos 2009)

To further characterize features of DOM composition weperformed parallel factor analysis (PARAFAC) using EEMdata within the drEEM toolbox for Matlab (Mathworks MAUSA) (Murphy et al 2013) PARAFAC is a statistical tech-nique used to decompose the complex mixture of the fluo-rescing DOM pool into quantifiable individual components(Stedmon et al 2003) We detected a total of six uniquecomponents and validated the model using core consistencyand split-half analysis (Murphy et al 2013 Stedmon andBro 2008) Components with similar spectra from previ-ous studies were identified using the online fluorescencerepository OpenFluor (Murphy et al 2014) and additionalcomponents with similar peaks were identified through lit-erature review Since the actual chemical structure of fluo-rophores is unknown we used the concentration of each flu-orophore as maximum fluorescence of excitation and emis-sion in Raman units (Fmax) to derive the percent contributionof each fluorophore component to total fluorescence Rela-tionships between PARAFAC components were also evalu-ated using Pearson correlation coefficients in the R packageHmisc (Harrell et al 2016)

27 Evaluating relationships in DOC concentration andDOM composition with stream discharge andtemperature

We used linear mixed-effects (LME) models to assess the re-lationship between DOC concentration or DOM composition(δ13C-DOC SR SUVA254 fluorescence index freshness in-dex PARAFAC components) stream discharge and streamtemperature Analysis was performed in R using the nlmepackage (Pinheiro et al 2016) Watershed was included asa random intercept to account for repeat measures on eachwatershed For some parameters a random slope of eitherdischarge or temperature was also included based on dataassessment and model selection Model selection was per-formed using AIC to compare models fit using ML (Burn-ham and Anderson 2002 Symonds and Moussalli 2010)The final model was fit using restricted maximum likelihood(REML) Marginal R2 which represents an approximationof the proportion of the variance explained by the fixed fac-tors alone and conditional R2 which represents an approx-imation of the proportion of the variance explained by boththe fixed and random factors were calculated based on themethods described in Nakagawa and Schielzeth (2013) andJohnson (2014)

28 Redundancy analysis relationships between DOCconcentration DOM composition and watershedcharacteristics

We evaluated relationships between stream water DOC andwatershed characteristics by relating DOC concentration andmeasures of DOM composition to catchment attributes usingredundancy analysis (RDA type 2 scaling) in the packagerdaTest (Legendre and Durand 2014) in R (version 322 RCore Team 2015) To maximize the amount of informationavailable we performed RDA analysis on samples collectedfrom January to July 2016 and therefore included all param-eters of optical characterization (ie all PARAFAC compo-nents and spectral indices) We assessed the collinearity ofDOM compositional variables using a variance inflation fac-tor (VIF) criteria of gt 10 which resulted in the removal ofPARAFAC components C2 C3 and C5 prior to RDA anal-ysis Catchment attributes for each watershed included av-erage slope percent area of lakes percent area of wetlandsaverage depth of mineral soil and average depth of organicsoil Relationships between variables were linear so no trans-formations were necessary and variables were standardizedprior to analysis To account for repeat monthly measures perwatershed and potential temporal correlation associated withmonthly sampling we included sample month as a covariable(ldquopartial-RDArdquo) To test whether the RDA axes significantlyexplained variation in the dataset we compared permutationsof residuals using ANOVA (9999 iterations testaxes func-tion of rdaTest)



3 Results

31 Hydrology

We present work for water year 2015 (WY2015 1 Oc-tober 2014ndash30 September 2015) and water year 2016(WY2016 1 October 2015ndash30 September 2016) Annualprecipitation for both water years was lower than historicalmean annual precipitation (WY2015 2661 mm WY20162587 mm) It is worth noting that mean annual precipitationat our rain gauge location (2890 mm yrminus1 elevation 16 m)is substantially lower than the average amount received athigher elevations which from 1981 to 2010 was approxi-mately 5027 mm yrminus1 at an elevation of 1000 m within ourstudy area This area receives a very high amount of an-nual rainfall but also experiences seasonal variation with anextended wet period from fall through spring and a muchshorter typically drier period during summer In WY2015and WY2016 86ndash88 of the annual precipitation on CalvertIsland occurred during the 8 months of wetter and coolerweather between September and April (sim 75 of the year)designated the ldquowet periodrdquo (WY2015 wet 2388 mm av-erage air temp 797 C WY2016 wet 2235 mm averageair temp 738 C) The remaining annual precipitation oc-curred during the drier and warmer summer months of MayndashAugust designated the ldquodry periodrdquo (WY2015 dry 314 mmaverage air temp 134 C WY2016 dry 352 mm average airtemp 131 C) Overall although WY2015 was slightly wet-ter than WY2016 the two years were comparable in relativeprecipitation during the wet versus dry periods

Stream discharge (Q) responded rapidly to rain eventsand as a result closely tracked patterns in total precip-itation (Fig 2) Total Q for all watersheds was on av-erage 22 greater for the wet period of WY2015 (totalQ 22302times 106 range 513times 106ndash11151times 106 m3) com-pared to the wet period of WY2016 (total Q 18289times 106range 417times 106ndash9145times 106 m3) Stream discharge andstream temperature were significantly different for wet ver-sus dry periods (MannndashWhitney tests plt00001)

32 Temporal and spatial patterns in DOCconcentration yield and flux

Stream waters were high in DOC concentration relativeto the global average for freshwater discharged directly tothe ocean (average DOC for Calvert and Hecate islands104 mg Lminus1 SD 38 average global DOC sim 6 mg Lminus1)

(Meybeck 1982 Harrison et al 2005) (Table 1 Fig 3)Q-weighted average DOC concentrations were higher thanaverage measured DOC concentrations (111 mg Lminus1 Ta-ble 1) and also resulted in slightly different ranking ofthe watersheds for highest to lowest DOC concentrationWithin watersheds Q-weighted DOC concentrations rangedfrom a low of 84 mg Lminus1 (watershed 693) to a high of193 mg Lminus1 (watershed 819) and concentrations were sig-

Figure 2 Hydrological patterns typical of watersheds located in thestudy area (a) The hydrograph and precipitation record from Wa-tershed 708 for the study period of 1 October 2015ndash30 April 2016Grey shading indicates the wet period (1 Septemberndash30 April) andthe unshaded region indicates the dry period (1 Mayndash30 August)(b) Correlation of daily (24 h) areal runoff (discharge of all water-sheds combined) to 48 h total rainfall recorded at watershed 708For the period of study comparisons of daily runoff to 48 h rainfall(runoff rainfall mean 092 SDplusmn 027) indicated rapid dischargeresponse to rainfall

nificantly different between watersheds (KruskalndashWallis testplt00001) Seasonal variability tended to be higher in wa-tersheds where DOC concentration was also high (water-sheds 626 819 and 844) and lower in watersheds withgreater lake area (watersheds 1015 and 708) (Table 1 boxplots Fig 3) On an annual basis DOC concentrations gener-ally decreased through the wet period and increased throughthe dry period and concentrations were significantly lowerduring the wet period compared to the dry period (MannndashWhitney test p= 00123) Results of our LME model (Ta-ble S61) indicate that DOC concentration was positively re-lated to both discharge (b = 0613 plt0001) and temper-ature (b = 0162 p= 0011) (model conditional R2

= 057marginal R2

= 009)Annual and monthly DOC yields are presented in Ta-

ble 1 For the total period of available Q (1 Octo-ber 2014ndash30 April 2016 19 months) areal (all water-sheds) DOC yield was 523 Mg C kmminus2 (95 CI 457 to682 Mg C kmminus2) and individual watershed yields rangedfrom 241 to 436 Mg C kmminus2 For WY2015 areal annualDOC yield was 333 Mg C kmminus2 yrminus1 (95 CI 289 to381 Mg C kmminus2 yrminus1) Total monthly rainfall was strongly



o

Figure 3 Seasonal (timelines by date) and spatial (box plots by watershed) patterns in DOC concentration and DOM composition for streamwater collected at the outlets of the seven study watersheds on Calvert and Hecate islands Boxes represent the 25th and 75th percentile whilewhiskers represent the 5th and 95th percentile Daily precipitation and annual temperature are shown in the top left panel Grey shadingindicates the wet period (1 Septemberndash30 April) and the unshaded region indicates the dry period of each water year

correlated with monthly DOC yield (Fig 4) and averagemonthly yield for the wet period (335 Mg C kmminus2 monthminus195 CI 294 to 440 Mg C kmminus2 monthminus1) was sig-nificantly greater than average monthly yield duringthe dry period (050 Mg C kmminus2 monthminus1 95 CI041 to 062 Mg C kmminus2 monthminus1) (MannndashWhitney testplt00001)

Across our study watersheds DOC flux generally in-creased with increasing watershed area (Fig 5) In WY2015total DOC flux for all watersheds included in our study was1562 Mg C (95 CI 1355 to 1787 Mg C) and individualwatershed flux ranged from ranged from 82 to 276 Mg CDOC flux was significantly different in wet versus dry peri-ods (MannndashWhitney test plt00001) Overall 94 of theexport in WY2015 occurred during the wet period and ex-

port for the wet period of WY2015 was lower than export forthe wet period of WY2016 (Fig 5)

33 Temporal and spatial patterns in DOM composition

The stable isotopic composition of dissolved organic carbon(δ13C-DOC) was relatively tightly constrained over spaceand time (average δ13C-DOC minus2653 permil SD 036 rangeminus2767 to minus2489 permil) Values of SR were low comparedto the range typically observed in surface waters (averageSR = 078 SD 004 range 071 to 089) and Fe-correctedSUVA254 values were at the high end of the range com-pared to most surface waters (average SUVA254 for Calvertand Hecate islands 442 L mgminus1 mminus1 SD 046 range ofSUVA254 in surface waters 10 to 50 L mgminus1 mminus1) (Spenceret al 2012) Values for both the fluorescence index (av-erage fluorescence index 136 SD 004 range 130 to



Figure 4 Monthly areal DOC yields and precipitation for wateryear 2015 (WY2015) and the wet period (1 Octoberndash30 April)of water year 2016 (WY2016) Error bars represent standard er-ror Total rain and DOC yield were significantly correlated (r2

=

077) and months of higher rain produced higher DOC yields InWY2015 the majority of DOC export (sim 94 of annual flux) oc-curred during the wet period (sim 88 of annual precipitation)

144) and freshness index (average freshness index 046SD 002 range 041 to 049) were relatively low com-pared to the typical range found in surface waters (Fellmanet al 2010 Hansen et al 2016) Differences between wa-tersheds were observed for δ13C-DOC (KruskalndashWallis testp= 00043) SR (KruskalndashWallis test p= 00001) fluores-cence index (KruskalndashWallis test p= 00030) and freshnessindex (KruskalndashWallis test p= 00099) but watersheds didnot differ in SUVA254 (KruskalndashWallis test p= 04837)

We observed seasonal variability in δ13C-DOC through-out the period of sampling (Fig 3 and our LME model(Table S61) indicate that δ13C-DOC declined with increas-ing discharge (b =minus0049 p= 0014) and stream tempera-ture (b =minus0024 plt0001) (model conditional R2

= 035marginal R2

= 010) In contrast although SUVA254 ap-peared to exhibit a general seasonal trend of values increas-ing over the wet period and decreasing over the dry pe-riod SUVA254 was not significantly related to either dis-charge or stream temperature in the LME model resultsSR also appeared to fluctuate seasonally with lower val-ues during the wet season and higher values during the dryseason SR was negatively related to discharge (b =minus0026plt0001) and positively related to the interaction betweendischarge and stream temperature (b = 00015 plt0001)(model conditional R2

= 062 marginal R2= 028) The

freshness index was negatively related to stream tempera-ture (b =minus0003 p= 0008) (model conditional R2

= 059marginal R2

= 023) while the fluorescence index was notsignificantly related to either discharge or stream tempera-ture

Figure 5 DOC fluxes and yields for the seven study watersheds andthe total area of study (ldquoarealrdquo all watersheds combined) on Calvertand Hecate islands for water year 2015 (WY2015 1 Octoberndash30 September) and 1 Octoberndash30 April of the wet period for wateryear 2015 (WY2015 wet) and water year 2016 (WY2016 wet) Be-cause DOC yields were only available for September in WY2015this month was excluded from the wet-period totals in order to makesimilar comparisons between years Error bars represent standarderror

34 PARAFAC characterization of DOM

Six fluorescence components were identified throughPARAFAC (ldquoC1rdquo through ldquoC6rdquo) (Table 2) Additional de-tails on PARAFAC model results are provided in SupplementTable S41 and Figs S42 and S43 Of the six componentsfour were found to have close spectral matches in the Open-Fluor database (C1 C3 C5 C6 minimum similarity scoregt 095) while the remaining two (C2 and C4) were foundto have similar peaks represented in the literature The firstfour components (C1 through C4) are described as terrestri-ally derived whereas components C5 and C6 are describedas autochthonous or microbially derived (Table 2) In gen-eral the rank order of each componentrsquos percent contribu-tion to total fluorescence was maintained over time with C1comprising the majority of total fluorescence across all wa-tersheds (Fig 6)

Across watersheds components fluctuated synchronouslyover time and variation between watersheds was relatively



Table2Spectralcom

positionforthe

sixfluorescence

components

identifiedusing

PAR

AFA

Cincluding

excitation(E

x)andem

ission(E

m)

peakvaluespercentcom

positionacross

allsamplesand

likelystructure

andcharacteristics

ofthefluorescentcom

ponentbasedon

previousstudies

Com

ponentE

x(nm)

Em

(nm)

composition

lowast

Potentialstructurecharacteristics

Previousstudies

with

comparable

results

C1

315436

341plusmn

22(311ndash393)

Hum

ic-likelessprocessed

terrestrialhigh

molecularw

eightwidespread

buthighestin

wetland

andforestenvironm

ent

Garcia

etal(2015)(C1)G

raeberetal(2012)

(C1)W

alkeretal(2014)

(C1)

Yam

ashitaetal(2011)

(C1)C

oryand

McK

night(2005)(C1)

C2

270380484

202plusmn

19(161ndash256)

Hum

ic-likeresembles

fulvicacid

widespreadhigh

molecularw

eightterrestrial

Stedmon

andM

arkager(2005)

(C2)

Stedmon

etal

(2003)(C

3)C

oryand

McK

night(2005)(C5)

C3

270478

178plusmn

18(128ndash208)

Hum

ic-likehighlyprocessed

terrestrialsuggested

asrefractory

Stedmon

andM

arkager(2005)

(C1)

Yam

ashitaetal(2010)(C

2)

C4

305435522

148plusmn

26(94ndash223)

Notcom

monly

reportedsimilarities

tofulvic-likecontributed

fromsoils

Lochm

ullerandSaavedra

(1986)(E)

C5

325442

98plusmn

35(00ndash159)

Aquatic

humic-like

fromterrestrial

environmentsautochthonousm

icrobialproducedm

aybe

photoproduced

Boehm

eand

Coble

(2000)(Peak

C)

Coble

etal(1998)

(PeakC

)Stedmon

etal(2003)(C3)

C6

285338

34plusmn

25(00ndash93)

Am

inoacid-liketryptophan-like

Freshlyadded

fromlandautochthonous

Rapidly

photodegradable

Murphy

etal

(2008)(C

7)Shutova

etal(2003)(C

4)Stedm

onetal(2007)

(C7)Y

amashita

etal(2003)(C5)

lowastM

eanplusmn

SD(m

inndashmax)from

allsamples



con

tribu

tion

to to

tal f

luor

esce

nce

Figure 6 Percent contribution of the six components identified inparallel factor analysis (PARAFAC) for samples collected everythree weeks from JanuaryndashJuly 2016 from the seven study water-sheds on Calvert and Hecate islands The grey shading indicates thewet period and the unshaded region indicates the dry period Notethat while the y axis for each panel has a range of 20 the maxand min for each y axis varies by panel

low although slightly more variation between watershedswas observed during the beginning of the dry period rela-tive to other times of the year (Fig 6) The percent contribu-tions of components C1 C3 C5 and C6 to total fluorescencewere not significantly different across watersheds (for allcomponents KruskalndashWallis test pgt005) however the per-cent composition of both C2 and C4 was different (KruskalndashWallis test p= 00306 and p= 00307 respectively) andhigher for watersheds 819 and 844 relative to the other wa-tersheds (Fig S44)

PARAFAC components exhibited significant relation-ships with stream discharge and stream temperature al-though predicted changes (beta or b) in fluorescence com-ponents with discharge andor stream temperature weresmall (Supplement Table S62) C3 increased with discharge(b = 0006 p= 0003) whereas C2 C4 and C5 decreased

with discharge (C2 b =minus0005 p= 0022 C4 b = -0008p= 0002 C5 b =minus0008 p= 0002) C1 C4 and C6 in-creased with temperature (C1 b = 0001 p= 0050 C4b = 0003 plt0001 C6 b = 0005 p= 0005) while bothC3 and C5 decreased with temperature (C3 b = -0003p= 0003 C5 b =minus0003 p= 0027) Conditional R2 val-ues for the models ranged from 028 to 069 while marginalR2 ranged from 020 to 046 Overall greater changes incomponent contribution to total fluorescence were observedwith changes in discharge relative to changes in stream tem-perature

35 Relationships between watershed characteristicsDOC concentrations and DOM composition

Results of the partial RDA (type 2 scaling) were significantin explaining variability in DOM concentration and composi-tion (semi-partial R2

= 033 F = 790 plt00001) (Fig 7)Axes 1 through 3 were statistically significant at plt0001and the relative contribution of each axis to the total ex-plained variance was 47 30 and 22 respectively Addi-tional details on the RDA test are provided in Figs S51ndashS52and Tables S53ndashS55 Axis 1 described a gradient of water-shed coverage by water-inundated ecosystem types rangingfrom more wetland coverage to more lake coverage Totallake coverage (area) and mean mineral soil material thick-ness showed a strong positive contribution and wetland cov-erage (area) showed a strong negative contribution to thisaxis The freshness index the fluorescence index SR andthe fluorescence component C6 were positively correlatedwith Axis 1 while component C4 showed a clear negativecorrelation Axis 2 described a subtler gradient of soil ma-terial thickness ranging from greater mean organic soil ma-terial thickness to greater mean mineral soil material thick-ness DOC concentration δ13C-DOC SUVA254 and fluores-cence component C1 all showed a strong positive correlationwith Axis 2 Axis 3 described a gradient of watershed steep-ness from lower gradient slopes with more wetland area andthicker organic soil material to steeper slopes with less devel-oped organic horizons Average slope contributed negativelyto Axis 3 (see Table S55) followed by positive contributionsfrom both wetland area and the thickness of organic soil ma-terial δ13C-DOC showed the most positive correlation withAxis 3 whereas fluorescence components C1 and C4 showedthe most negative

4 Discussion

41 DOC export from small catchments to the coastalocean

In comparison to global models of DOC export (Mayorgaet al 2010) and DOC exports quantified for southeasternAlaska (DrsquoAmore et al 2015a 2016 Stackpoole et al2017) our estimates of freshwater DOC yield from Calvert



1

2

3

4

5

6

1

2

34

6

1

2 34

5

6

1

2

34

56

1

2

3

4

5

6

1

23

45

6

1 2

3

45

6

DOC

DO13C

Sr

FI

SUVA FRESHC1

C4

C6

Lakes

Slope

Wetlands

MinThick

OrgThick

-1

0

1

2

-2 -1 0 1 2RDA1

RD

A2

Watershed626

693

703

708

819

844

1015

Figure 7 Results from the partial redundancy analysis (RDA type 2 scaling) of DOC concentration and DOM composition versus watershedcharacteristics Angles between vectors represent correlation ie smaller angles indicate higher correlation Symbols represent differentwatersheds and numbers on symbols represent the sample month in 2016 1 ndash January 2 ndash February 3 ndash March 4 ndash early April 5 ndash lateApril and 6 ndash May

and Hecate island watersheds are in the upper range pre-dicted for the perhumid rainforest region When comparedto watersheds of similar size DOC yields from Calvert andHecate island watersheds are some of the highest observed(see reviews in Hope et al 1994 Alvarez-Cobelas et al2012) including DOC yields from many tropical rivers de-spite the fact that tropical rivers have been shown to ex-port very high DOC (eg Autuna River Venezuela DOCyield 56 946 kg C kmminus2 yrminus1 Castillo et al 2004) and areoften regarded as having disproportionately high carbon ex-port compared to temperate and Arctic rivers (Aitkenheadand McDowell 2000 Borges et al 2015) Our estimatesof DOC yield are comparable to or higher than previousestimates from high-latitude catchments of similar size thatreceive high amounts of precipitation and contain exten-sive organic soils and wetlands (eg Naiman 1982 (DOCyield 48 380 kg C kmminus2 yrminus1) Brooks et al 1999 (DOCyield 20 300 kg C kmminus2 yrminus1) Aringgren et al 2007 (DOCyield 32 043 kg C kmminus2 yrminus1)) However many of thesecatchments represent low- (first- or second-) order headwa-ter streams that drain to higher-order stream reaches ratherthan directly to the ocean Although headwater streams havebeen shown to export up to 90 of the total annual carbonin stream systems (Leach et al 2016) significant processingand loss typically occurs during downstream transit (Battinet al 2008)

Over much of the incised outer coast of the CTR smallrainfall-dominated catchments contribute high amounts of

freshwater runoff to the coastal ocean (Royer 1982 Mor-rison et al 2012 Carmack et al 2015) Small mountainouswatersheds that discharge directly to the ocean can exhibitdisproportionately high fluxes of carbon relative to water-shed size and in aggregate may deliver more than 50 oftotal carbon flux from terrestrial systems to the ocean (Mil-liman and Syvitski 1992 Masiello and Druffel 2001) Ex-trapolating our estimate of annual DOC yield from Calvertand Hecate island watersheds to the entire hypermaritimesubregion of British Columbiarsquos CTR (29 935 km2) gener-ates an estimated annual DOC flux of 0997 Tg C yrminus1 (0721to 1305 Tg C yrminus1 for our lowest to highest yielding wa-tersheds respectively) with the caveat that this estimate isrudimentary and does not account for spatial heterogene-ity in controlling factors such as wetland extent topogra-phy and watershed size Regional comparisons estimate thatSoutheast Alaska (104 000 km2) at the northern range of theCTR exports approximately 125 Tg C yrminus1 (Stackpoole etal 2016) while south of the perhumid CTR the wet north-western United States and its associated coastal temperaterainforests export less than 0153 Tg C yrminus1 as DOC (re-ported as TOC Butman et al 2016) This suggests that thehypermaritime coast of British Columbia plays an importantrole in the export of DOC from coastal temperate rainforestecosystems of western North America in a region that is al-ready expected to contribute high quantities of DOC to thecoastal ocean



42 DOM composition

The composition of stream water DOM exported fromCalvert and Hecate island watersheds is mainly terrestrial in-dicating that the production and overall supply of terrestrialmaterial is sufficient to exceed microbial demand and thus arelatively abundant supply of terrestrial DOM is available forexport Values for δ13C-DOC suggest that terrestrial carbonsources from C3 plants and soils were the dominant input tocatchment stream water DOM (Finlay and Kendall 2007)Measures of SR and SUVA254 were typical of environmentsthat export large quantities of high molecular-weight highlyaromatic DOM such as some tropical rivers (eg Lambertet al 2016 Mann et al 2014) streams draining wetlands(eg Aringgren et al 2008 Austnes et al 2010) or streamsdraining small undisturbed catchments comprised of mixedforest and wetlands (eg Wickland et al 2007 Fellman etal 2009a Spencer et al 2010 Yamashita et al 2011) Thissuggests that the majority of the DOM pool is comprised oflarger molecules that have not been extensively chemicallyor biologically degraded through processes such as microbialutilization or photodegradation and therefore are potentiallymore biologically available (Amon and Benner 1996)

Biological utilization of DOM is influenced by its compo-sition (eg Judd et al 2006 Fasching et al 2014) there-fore differences in DOM can alter the downstream fate andecological role of freshwater-exported DOM For examplethe majority of the fluorescent DOM pool was comprisedof C1 which is described as humic-like less processed ter-restrial soil and plant material (see Table 2) In addition al-though the tryptophan-like component C6 represents a mi-nor proportion of total fluorescence even a small proteina-ceous fraction of the overall DOM pool can play a ma-jor role in overall bioavailability and bacterial utilizationof DOM (Berggren et al 2010 Guillamette and Giorgio2011) These contributions of stream-exported DOM mayrepresent a relatively fresh seasonally consistent contribu-tion of terrestrial subsidy from streams to the coastal ecosys-tem which in this region is relatively lower in carbon andnutrients throughout much of the year (Whitney et al 2005Johannessen et al 2008)

43 DOC and DOM export sources and seasonalvariability

On Calvert and Hecate islands the relationship betweenDOC concentration and discharge varied by watershed (seeSupplement Fig S61) as might be expected given theknown influence of watershed characteristics (eg lake areawetland area soils etc) on DOC concentration and exportHowever overall DOC concentrations increased in all water-sheds with both discharge and temperature indicating that theoverarching drivers of DOC export are the hydrologic cou-pling of precipitation and runoff from the landscape with the

seasonal production and availability of DOC (Fasching et al2016)

Precipitation is a well-established driver of stream DOCexport (Alvarez-Cobelas et al 2012) particularly in sys-tems containing organic soils and wetlands (Olefeldt et al2013 Wallin et al 2015 Leach et al 2016) Frequenthigh-intensity precipitation events and short residence timesare expected to result in pulsed exports of stream DOC thatare rapidly shunted downstream thus reducing time for in-stream processing (Raymond et al 2016) Flashy stream hy-drographs indicate that hydrologic response times for Calvertand Hecate island watersheds are rapid presumably as a re-sult of small catchment size high drainage density and rela-tively shallow soils with high hydraulic conductivity (Gibsonet al 2000 Fitzgerald et al 2003) Rapid runoff is presum-ably accompanied by rapid increases in water tables and lat-eral movement of water through shallow soil layers rich in or-ganic matter (Fellman et al 2009b DrsquoAmore et al 2015b)It appears that on Calvert and Hecate islands the combina-tion of high rainfall rapid runoff and abundant sources ofDOC from organic-rich wetlands and forests results in highDOC fluxes

The relationship between DOC stream temperature anddischarge indicates that seasonal dynamics play an importantrole in the variability of DOC exported from these systemsFor example DOC concentrations decrease in all watershedsduring the wet period of the year these decreases are asso-ciated with clear changes in DOM composition such as in-creasing δ13C-DOC and SUVA254 and decreasing SR Thisis in contrast with patterns observed during the dry periodwhen DOC concentrations gradually increase while δ13C-DOC SUVA254 decrease Fluctuations in DOC and DOMcomposition occur throughout the wet and the dry seasonsuggesting that temperature and runoff ndash and perhaps otherseasonal drivers ndash are important year-round controls on DOCconcentration as well as certain measures of DOM composi-tion such as δ13C-DOC and SR

The process of ldquoDOC flushingrdquo has been shown to increasestream water DOC during higher flows in coastal and tem-perate watersheds (eg Sanderman et al 2009 Deirmend-jian et al 2017) Flushing can occur through various mech-anisms For example Boyer et al (1996) observed that dur-ing drier periods DOC pools can increase in soils and arethen flushed to streams when water tables rise Rising wa-ter tables can establish strong hydraulic gradients that initi-ate and sustain prolonged increases in metrics like SUVA254until the progressive drawdown of upland water tables con-strains flow paths (Lambert et al 2013) DOC concentra-tions can vary during flushing in response to changing flowpaths which can shift sources of DOC within the soil pro-file from older material in deeper soil horizons to more re-cently produced material in shallow horizons (Sandermanet al 2009) or from changes in the production mechanismof DOC (Lambert et al 2013) For example Sanderman etal (2009) observed distinct relationships between discharge



and both δ13C-DOC and SUVA254 and postulated that dur-ing the rainy season hillslope flushing shifts DOM sourcesto more aged soil organic material In addition instream pro-duction can also provide a source of DOC and therefore af-fect seasonal variation in DOC concentration and composi-tion (Lambert et al 2013) The extent of these effects canshift seasonally relationships between flow paths and DOCexport in rain-dominated catchments can vary within and be-tween hydrologic periods depending on factors such as thedegree of soil saturation the duration of previous drying andrewetting cycles soil chemistry and DOM source-pool avail-ability (Lambert et al 2013)

Our observations of changes in DOC and DOM relatedto discharge and stream temperature suggest that a varietyof mechanisms may be important for controlling dynamicsof seasonal export in Pacific hypermaritime watersheds Weobserved elevated DOC concentrations during precipitationevents following extended dry periods suggesting DOC mayaccumulate during dry periods and be flushed to streamsduring runoff events Increased discharge was significantlyrelated to δ13C-DOC and SR with higher discharge result-ing in more terrestrial-like DOM One possible explanationis that hydrologic connectivity increases during higher dis-charge as soil conditions become more saturated thereforepromoting the mobilization of DOM from across a widerrange of the soil profile (McKnight et al 2001 Kalbitzet al 2002) In addition the mechanisms of DOC produc-tion and sources of DOC appear to shift seasonally Rela-tionships between increased temperature and lower valuesof δ13C-DOC and higher values of the freshness index C1and C4 suggest that warmer conditions result in a freshsupply of DOM exported from terrestrial sources (Fellmanet al 2009a Fasching et al 2016) This may represent ashift in the source of DOM andor increased contributionsfrom less aromatic lower molecular-weight material such asDOM derived from increased terrestrial primary production(Berggren et al 2010) Further fine-scaled investigation intothe mechanistic underpinnings of the relationship betweendischarge stream temperature and DOM represents a clearpriority for future research in this region

44 Relationships between watershed attributes andexported DOM

Previous studies have implicated wetlands as a major driverof DOM composition (eg Xenopoulos et al 2003 Aringgrenet al 2008 Creed et al 2008) however the analysis ofrelationships between Calvert and Hecate island landscapeattributes and variation in DOM composition suggests thatcontrols on DOM composition are more nuanced than beingsolely driven by the extent of wetlands Aringgren et al (2008)found that when wetland area comprised gt 10 of totalcatchment area wetland DOM was the most significantdriver of stream DOM composition during periods of highhydrologic connectivity Although wetlands comprise an av-

erage of 37 of our study area they do not appear to be thesingle leading driver of variability in DOC concentration andDOM composition Other factors such as watershed slopethe depth of organic and mineral soil materials and the pres-ence of lakes also appear to be influence DOC concentrationand DOM composition The presence of cryptic wetlands(Creed et al 2003) and limitations of the wetland mappingmethod could also weaken the link between wetland extentDOC and DOM

In these watersheds soils with pronounced accumulationsof organic matter are not restricted to wetland ecosystemsPeat accumulation in wetland ecosystems results in the for-mation of organic soils (Hemists) where mobile fractions ofDOM accumulate under saturated soil conditions and limiteddrainage resulting in the enrichment of poorly biodegrad-able more stable humic acids (Stevenson 1994 Marschnerand Kalbitz 2003) Although Hemist soils comprise 278 of our study area Folic Histosols which form under morefreely drained conditions such as steeper slopes occur overan additional 257 of the area (Supplement S12) In freelydrained organic soils high rates of respiration can result infurther enrichment of aromatic and more complex moleculesand this material may be rapidly mobilized and exported tostreams (Glatzel et al 2003) This suggests the importanceof widely distributed alternative soil DOM source poolssuch as Folic Histosols and associated Podzols with thick for-est floors on hillslopes available to contribute high amountsof terrestrial carbon for export

Although lakes make up a relatively small proportion ofthe total landscape area their influence on DOM export ap-pears to be important The proportion of lake area can bea good predictor of organic carbon loss from a catchmentsince lakes often increase hydrologic residence times andthus increase opportunities for biogeochemical processing(Algesten et al 2004 Tranvik et al 2009) In our studywatersheds with a larger percentage of lake area exhibited aslower response following rain events (Fig S22) and lowerDOC yields and lake area was correlated with parametersthat represent greater autochthonous DOM production ormicrobial processing such as higher freshness index SRand fluorescence index and a higher proportions of compo-nent C6 In contrast watersheds with a high percentage ofwetlands contributed DOM that was more allochthonous incomposition Lakes are known to be important landscape pre-dictors of DOC as increased residence time enables removalvia respiration thus reducing downstream exports from lakeoutlets (Larson et al 2007) The proximity of wetlands andlakes to the watershed outlet can also play an important rolein the composition of DOM exports (Martin et al 2006)

5 Conclusions

Previous work has demonstrated that freshwater dischargeis substantial along the coastal margin of the North Pacific



temperate rainforest and plays an important role in pro-cesses such as ocean circulation (Royer 1982 Eaton andMoore 2010) Our finding that small catchments in this re-gion contribute high yields of terrestrial DOC to coastal wa-ters suggests that freshwater inputs may also influence oceanbiogeochemistry and food web processes through terrestrialorganic-matter subsidies Our findings also suggest that thisregion may be currently underrepresented in terms of its rolein global carbon cycling Currently there is no region-widecarbon flux model for the Pacific coastal temperate rainforestor the greater Gulf of Alaska which would quantify the im-portance of this region within the global carbon budget Ourestimates point to the importance of the hypermaritime outer-coast zone of the CTR where subdued terrain high rainfallocean-moderated temperatures and poor bedrock have gen-erated a distinctive bog-forest landscape mosaic within thegreater temperate rainforest (Banner et al 2005) Howevereven within our geographically limited study area we ob-served a range of DOC yields across watersheds To quan-tify regional-scale fluxes of rainforest carbon to the coastalocean further research will be needed to estimate DOCyields across complex spatial gradients of topography cli-mate hydrology soils and vegetation Long-term changesin DOC flux have been observed in many places (eg Wor-rall et al 2004 Borken et al 2011 Lepistouml et al 2014Tank et al 2016) and continued monitoring of this systemwill allow us to better understand the underlying drivers ofexport and evaluate future patterns in DOC yields Coupledwith current studies investigating the fate of terrestrial ma-terial in ocean food webs this work will improve our un-derstanding of coastal carbon patterns and increase capacityfor predictions regarding the ecological impacts of climatechange

Data availability Chemistry and DOC flux data are avail-able in Oliver et al (2017) (httpsdoiorg10219661321324)Ecosystem comparison plot data are available in Giesbrechtet al (2015) (httpsdoiorg1021966156481) Stage and dis-charge time series data are available in Floyd et al (2016)(httpsdoiorg10219661243102) Lidar-derived watersheds andassociated metrics data are available in Gonzalez Arriola etal (2015) (httpsdoiorg1021966115311)

The Supplement related to this article is available onlineat httpsdoiorg105194bg-14-3743-2017-supplement

Author contributions AAO prepared the paper with contributionsfrom all authors designed analysis protocols analyzed samples andperformed the modelling and analysis for dissolved organic carbonfluxes parallel factor analysis of dissolved organic-matter compo-sition and all remaining statistical analyses SET assisted with de-signing the study and overseeing laboratory analyses crafting the

scope of the paper and determining the analytical approach IGled the initial DOC sampling design helped coordinate the researchteam oversaw routine sampling and data management and led thewatershed characterization MCK developed the rating curves andconducted the statistical analysis of discharge measurement uncer-tainties and rating curve uncertainties WCF led the hydrology com-ponent of this project selected site locations installed and designedthe hydrometric stations and developed the rating curves and fi-nal discharge calculations CB and PS collected and analyzed soilfield data and prepared the digital soils map of the watersheds KPLconceived of and co-led the overall study of which this paper is acomponent helped assemble and guide the team of researchers whocarried out this work and provided input to each stage of the study

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This work was funded by the Tula Foundationand the Hakai Institute The authors would like to thank manyindividuals for their support including Skye McEwan Bryn FedjeLawren McNab Nelson Roberts Adam Turner Emma MyersDavid Norwell and Chris Coxson for sample collection and datamanagement Clive Dawson and North Road Analytical for sampleprocessing and data management Keith Holmes for creatingour maps Matt Foster for database development and supportShawn Hateley for sensor network maintenance Jason JacksonColby Owen James McPhail and the entire staff at Hakai EnergySolutions for installing and maintaining the sensors and telemetrynetwork and Stewart Butler and Will McInnes for field supportThanks to Santiago Gonzalez Arriola for generating the watershedsummaries and associated data products and Ray Brunsting foroverseeing the design and implementation of the sensor networkand the data management system at Hakai Additional thanks toLori Johnson and Amelia Galuska for soil mapping field assistanceand Francois Guillamette for PARAFAC consultation Thanksto Dave DrsquoAmore for inspiring the Hakai project to investigateaquatic fluxes at the coastal margin and for technical guidanceLastly thanks to Eric Peterson and Christina Munck who providedsignificant guidance throughout the process of designing andimplementing this study

Edited by Steven BouillonReviewed by three anonymous referees

References

Aringgren A Buffam I Jansson M and Laudon H Importance ofseasonality and small streams for the landscape regulation of dis-solved organic carbon export J Geophys Res-Biogeosci 112httpsdoiorg1010292006JG000381 2007

Aringgren A Buffam I Berggren M Bishop K Jansson M andLaudon H Dissolved organic carbon characteristics in borealstreams in a forest-wetland gradient during the transition be-tween winter and summer J Geophys Res-Biogeosci 113httpsdoiorg1010292007JG000674 2008



Akaike H Likelihood of a model and information crite-ria J Econometrics 16 3ndash14 httpsdoiorg1010160304-4076(81)90071-3 1981

Aitkenhead J A and McDowell W H Soil C N ra-tio as a predictor of annual riverine DOC flux at localand global scales Global Biogeochem Cy 14 127ndash138httpsdoiorg1010291999GB900083 2000

Alaback P B Biodiversity patterns in relation to climate Thecoastal temperate rainforests of North America Ecol Stud 116105ndash133 httpsdoiorg101007978-1-4612-3970-3_7 1996

Algesten G Sobek S Bergstroumlm A Aringgren A TranvikL and Jansson M Role of lakes for organic carbon cy-cling in the boreal zone Glob Change Biol 10 141ndash147httpsdoiorg101111j1365-2486200300721x 2004

Alvarez-Cobelas M Angeler D Saacutenchez-Carrillo S andAlmendros G A worldwide view of organic carbon ex-port from catchments Biogeochemistry 107 275ndash293httpsdoiorg101007s10533-010-9553-z 2012

Amon R M W and Benner R Bacterial utilization of differentsize classes of dissolved organic matter Limnol Oceanogr 4141ndash51 1996

Aufdenkampe A Mayorga E Raymond P Melack J DoneyS Alin S Aalto R and Yoo K Riverine coupling of biogeo-chemical cycles between land oceans and atmosphere FrontEcol Environ 9 53ndash60 httpsdoiorg101890100014 2011

Austnes K Evans C D Eliot-Laize C Naden P S andOld G H Effects of storm events on mobilisation andin-stream processing of dissolved organic matter (DOM) ina Welsh peatland catchment Biogeochemical 99 157ndash173httpsdoiorg101007s10533-009-9399-4 2010

Banner A LePage P Moran J and de Groot A (Eds) TheHyP3 Project pattern process and productivity in hypermar-itime forests of coastal British Columbia ndash a synthesis of 7-year results Special Report 10 Res Br British Columbia Min-istry Forests Victoria British Columbia 142 pp available athttpwwwforgovbccahfdpubsDocsSrsSrs10htm (last ac-cess 11 August 2017) 2005

Battin T J Kaplan L A Findlay S Hopkinson C S Marti EPackman A I Newbold D and Sabater F Biophysical con-trols on organic carbon fluxes in fluvial networks Nat Geosci 195ndash100 2008

Bauer J E Cai W J Raymond P A T S BianchiHopkinson C S and Regnier P A G The changingcarbon cycle of the coastal ocean Nature 504 61ndash70httpsdoiorg101038nature12857 2013

Berggren M Laudon H Haei M Stroumlm L and Jansson M Ef-ficient aquatic bacterial metabolism of dissolved low-molecular-weight compounds from terrestrial sources ISME J 4 408ndash416httpsdoiorg101038ismej2009120 2010

Boehme J and Coble P Characterization of ColoredDissolved Organic Matter Using High-Energy LaserFragmentation Environ Sci Technol 34 3283ndash3290httpsdoiorg101021es9911263 2000

Borken W Ahrens B Schultz C and Zimmermann L Site-to-site variability and temporal trends of DOC concentrations andfluxes in temperate forest soils Glob Change Biol 17 2428ndash2443 httpsdoiorg101111j1365-2486201102390x 2011

Borges A V Darchambeau F Teodoru C R Marwick T RTamooh F Geeraert N Omengo F O Gueacuterin F Lambert

T Morana C Okuku E and Bouillon S Globally signifi-cant greenhouse-gas emissions form African inland waters NatGeosci 8 637ndash642 httpsdoiorg101038ngeo2486 2015

Boyer E W Hornberger G M Bencala K E and McKnightD Overview of a simple model describing variation of dissolvedorganic carbon in an upland catchment Ecol Modell 86 183ndash188 1996

Burnham K P and Anderson D R Model selection and multi-model inference 2nd Edn Springer New York 2002

Carmack E Winsor P and William W The contiguous panarcticRiverine Coastal Domain A unifying concept Prog Oceanogr139 13ndash23 httpsdoiorg101016jpocean201507014 2015

Castillo M M Allan J D Sinsabaugh R L and Kling G WSeasonal and interannual variation of bacterial production in low-land rivers of the Orinoco basin Freshwater Biol 49 1400ndash1414 httpsdoiorg101111j1365-2427200401277x 2004

Clark J M Lane S N Chapman P J and Adamson J K Ex-port of dissolved organic carbon from an upland peatland dur-ing storm events Implications for flux estimates J Hydrol 347438ndash447 httpsdoiorg101016jjhydrol200709030 2007

Coble P Castillo C and Avril B Distribution and opticalproperties of CDOM in the Arabian Sea during the 1995Southwest Monsoon Deep-Sea Res Pt II 45 2195ndash2223httpsdoiorg101016S0967-0645(98)00068-X 1998

Cole J Prairie Y Caraco N McDowell W Tranvik L StrieglR Duarte C Kortelainen P Downing J Middelburg J andMelack J Plumbing the Global Carbon Cycle Integrating In-land Waters into the Terrestrial Carbon Budget Ecosystems 10172ndash185 httpsdoiorg101007s10021-006-9013-8 2007

Cory R M and McKnight D M Fluorescence spectroscopy re-veals ubiquitous presence of oxidized and reduced quinines indissolved organic matter Environ Sci Technol 39 8142ndash8149httpsdoiorg101021es0506962 2005

Creed I F Beall F D Clair T A Dillon P J andHesslein R H Predicting export of dissolved organiccarbon from forested catchments in glaciated landscapeswith shallow soils Glob Biogeochem Cy 22 GB4024httpsdoiorg1010292008GB003294 2008

Creed I F Sanford S E Beall F D Molot L A and DillonP J Cryptic wetlands integrating hidden wetlands in regressionmodels of the export of dissolved organic carbon from forestedlandscapes Hydrol Process 17 3629ndash3648 2003

DrsquoAmore D V Edwards R T and Biles F E Biophysicalcontrols on dissolved organic carbon concentrations of Alaskancoastal temperate rainforest streams Aquat Sci 2 381ndash393httpsdoiorg101007s00027-015-0441-4 2015a

DrsquoAmore D V Edwards R T Herendeen P A HoodE and Fellman J B Dissolved organic carbon fluxesfrom hydropedologic units in Alaskan coastal temperate rain-forest watersheds Soil Sci Soc Am J 79 378ndash388httpsdoiorg102136sssaj2014090380 2015b

DrsquoAmore D V Biles F E Nay M and Rupp T S Watershedcarbon budgets in the southeastern Alaskan coastal forest regionin Baseline and projected future carbon storage and greenhouse-gas fluxes in ecosystems of Alaska US Geological Survey Pro-fessional Paper 1826 196 pp 2016

Dai M Yin Z Meng F Liu Q and Cai WJ Spa-tial distribution of riverine DOC inputs to the ocean an



updated global synthesis Curr Opin Sust 4 170ndash178httpsdoiorg101016jcosust201203003 2012

Deirmendjian L Loustau D Augusto L Lafont S ChipeauxC Poirier D and Abril G Hydrological and ecological con-trols on dissolved carbon concentrations in groundwater and car-bon export to surface waters in a temperate pine forest watershedBiogeosciences Discuss httpsdoiorg105194bg-2017-90 inreview 2017

DellaSala D A Temperate and Boreal Rainforests of the WorldIsland Press Washington DC 2011

Emili L and Price J Biogeochemical processes in the soil-groundwater system of a forest-peatland complex north coastBritish Columbia Canada Northwest Sci 88 326ndash348httpsdoiorg1039550460870406 2013

Fasching C Behounek B Singer G and Battin T Microbialdegradation of terrigenous dissolved organic matter and potentialconsequences for carbon cycling in brown-water streams SciRep 4 4981 httpsdoiorg101038srep04981 2014

Fasching C Ulseth A Schelker J Steniczka G and Bat-tin T Hydrology controls dissolved organic matter export andcomposition in an Alpine stream and its hyporheic zone Lim-nol Oceanogr 61 558ndash571 httpsdoiorg101002lno102322016

Fellman J Hood E DrsquoAmore D Edwards R and White DSeasonal changes in the chemical quality and biodegradability ofdissolved organic matter exported from soils to streams in coastaltemperate rainforest watersheds Biogeochemistry 95 277ndash293httpsdoiorg101007s10533-009-9336-6 2009a

Fellman J Hood E Edwards R and DrsquoAmore D Changesin the concentration biodegradability and fluorescent prop-erties of dissolved organic matter during stormflows incoastal temperate watersheds J Geophys Res-Biogeo 114httpsdoiorg1010292008JG000790 2009b

Fellman J Hood E and Spencer R Fluorescence spectroscopyopens new windows into dissolved organic matter dynamics infreshwater ecosystems A review Limnol Oceanogr 55 2452ndash2462 httpsdoiorg104319lo20105562452 2010

Fellman J Nagorski S Pyare S Vermilyea A W Scott D andHood E Stream temperature response to variable glacier coverin coastal watersheds of Southeast Alaska Hydrol Process 282062ndash2073 httpsdoiorg101002hyp9742 2014

Finlay J C and Kendall C Stable isotope tracing of temporal andspatial variability in organic matter sources and variability in or-ganic matter sources to freshwater ecosytems in Stable Isotopesin Ecology and Environmental Science edited by Michener Rand Lajtha K Blackwell Publishing Ltd Oxford UK 2 283ndash324 2007

Fitzgerald D Price J and Gibson J Hillslope-swamp in-teractions and flow pathways in a hypermaritime rainfor-est British Columbia Hydrol Process 17 3005ndash3022httpsdoiorg101002hyp1279 2003

Gibson J J Price J S Aravena R Fitzger-ald D F and Maloney D Runoff generation ina hypermaritime bog-forest upland Hydrol Pro-cess 14 2711ndash2730 httpsdoiorg1010021099-1085(20001030)1415lt2711AID-HYP88gt30CO2-2 2000

Glatzel S Kalbitz K Dalva M and Moore T Dissolved or-ganic matter properties and their relationship to carbon dioxideefflux from restored peat bogs Geoderma 113 397ndash411 2003

Gonzalez Arriola S Frazer G W and Giesbrecht I LiDAR-derived watersheds and their metrics for Calvert Island HakaiInstitute httpsdoiorg1021966115311 2015

Gorham E Lehman C Dyke A Clymo D and JanssensJ Long-term carbon sequestration in North American peat-lands Quaternary Sci Rev 58 77ndash82 2012

Graeber D Gelbrecht J Pusch M Anlanger C andvon Schiller D Agriculture has changed the amount andcomposition of dissolved organic matter in Central Euro-pean headwater streams Sci Total Environ 438 435ndash446httpsdoiorg101016jscitotenv201208087 2012

Green R N Reconnaissance level terrestrial ecosystem mappingof priority landscape units of the coast EBM planning area Phase3 Prepared for British Columbia Ministry Forests Lands andNatural Resource Ops Blackwell and Associates VancouverCanada 2014

Guillemette F and Giorgio P Reconstructing the vari-ous facets of dissolved organic carbon bioavailability infreshwater ecosystems Limnol Oceanogr 56 734ndash748httpsdoiorg104319lo20115620734 2011

Hansen A M Kraus T E C Pellerin B A Fleck J ADowning B D and Bergamaschi B A Optical proper-ties of dissolved organic matter (DOM) Effects of biologicaland photolytic degradation Limnol Oceanogr 61 1015ndash1032httpsdoiorg101002lno10270 2016

Harrell F E and Dupont C Hmisc Harrell MiscellaneousR package version 40-2 httpsCRANR-projectorgpackage=Hmisc 2016

Harrison J Caraco N and Seitzinger S Global patterns andsources of dissolved organic matter export to the coastal zoneResults from a spatially explicit global model Global Bio-geochem Cy 19 httpsdoiorg1010292005gb002480 2005

Helms J Stubbins A Ritchie J Minor E Kieber D and Mop-per K Absorption spectral slopes and slope ratios as indica-tors of molecular weight source and photobleaching of chro-mophoric dissolved organic matter Limnol Oceanogr 53 955ndash969 httpsdoiorg104319lo20085330955 2008

Helton A Wright M Bernhardt E Poole G Cory Rand Stanford J Dissolved organic carbon lability increaseswith water residence time in the alluvial aquifer of a riverfloodplain ecosystem J Geophys Res-Biogeo 120 693ndash706httpsdoiorg1010022014JG002832 2015

Hoffman K M Gavin D G Lertzman K P Smith D J andStarzomski B M 13 000 years of fire history derived from soilcharcoal in a British Columbia coastal temperate rain forest Eco-sphere 7 e01415 httpsdoiorg101002ecs21415 2016

Hope D Billett M F and Cresser M S A review of the export ofcarbon in river water Fluxes and processes Environ Pollut 84301ndash324 httpsdoiorg1010160269-7491(94)90142-2 1994

Hopkinson C S Buffam I Hobbie J Vallino J and PerdueM Terrestrial inputs of organic matter to coastal ecosystems Anintercomparison of chemical characteristics and bioavailabilityBiogeochemistry 43 211ndash234 1998

Hudson N Baker A and Reynolds D Fluorescence anal-ysis of dissolved organic matter in natural waste andpolluted waters-a review River Res Appl 23 631ndash649httpsdoiorg101002rra1005 2007



Hurvich C M and Tsai C Regression and time seriesmodel selection in small samples Biometrika 76 297ndash307httpsdoiorg1023072336663 1989

IUSS Working Group WRB World Reference Base for Soil Re-sources International soil classification system for naming soilsand creating legends for soil maps World Soil Resources Re-ports No 106 Food and Agricultural Organization of the UnitedNations Rome Italy 2015

ISO Standard 9196 Liquid flow measurement in open channels ndashFlow measurements under ice conditions International Organi-zation for Standardization available online at wwwisoorg (lastaccess 1 November 2016) 1992

ISO Standard 748 Hydrometry ndash Measurement of liquid flow inopen channels using current-meters or floats International Or-ganization for Standardization available online at wwwisoorg(last access 1 November 2016) 2007

Johannessen S C Potentier G Wright C A Masson D andMacdonald R W Water column organic carbon in a Pacificmarginal sea (Strait of Georgia Canada) Mar Environ Res66 S49ndashS61 httpsdoiorg101016jmarenvres2008070082008

Johnson M Couto E Abdo M and Lehmann J Fluorescenceindex as an indicator of dissolved organic carbon quality in hy-drologic flowpaths of forested tropical watersheds Biogeochem-istry 105 149ndash157 httpsdoiorg101007s10533-011-9595-x2011

Johnson P C D Extension of Nakagawa andSchielzethrsquos R2

GLMM to random slopes models Methods EcolEvol 5 944ndash946 httpsdoiorg1011112041-210X122252014

Judd K Crump B and Kling G Variation in dissolved or-ganic matter controls bacterial production and community com-position Ecology 87 2068ndash2079 httpsdoiorg1018900012-9658(2006)87[2068VIDOMC]20CO2 2006

Kalbitz K Schmerwitz J Schwesig D and MatznerE Biodegradation of soil-derived dissolved organic mat-ter as related to its properties Geoderma 113 273ndash291httpsdoiorg101016S0016-7061(02)00365-8 2003

Kling G Kipphut G Miller M and OrsquoBrien W Integration oflakes and streams in a landscape perspective the importance ofmaterial processing on spatial patterns and temporal coherenceFreshwater Biol 43 477ndash497 httpsdoiorg101046j1365-2427200000515x 2000

Koehler A-K Murphy K Kiely G and Sottocornola M Sea-sonal variation of DOC concentration and annual loss of DOCfrom an Atlantic blanket bog in South Western Ireland Bio-geochemistry 95 231ndash242 httpsdoiorg101007s10533-009-9333-9 2009

Lakowicz J R Principles of Fluorescence Spectroscopy 2Kluwer Academic New York 1999

Larson J H Frost P C Zheng Z Johnston C A Bridgham SD Lodge D M and Lamberti G A Effects of upstream lakeson dissolved organic matter in streams Limnol Oceanogr 5260ndash69 httpsdoiorg104319lo20075210060 2007

Leighty W W Hamburg S P and Caouette J Effects of man-agement on carbon sequestration in forest biomass in SoutheastAlaska Ecosystems 9 1051 httpsdoiorg101007s10021-005-0028-3 2006

Lalonde K Middlestead P Geacutelinas Y Automation of 13C12Cratio measurement for freshwater and seawater DOC using hightemperature combustion Limnol Oceanogr-Meth 12 816ndash829 httpsdoiorg104319lom201412816 2014

Lambert T Bouillon S Darchambeau F Massicotte P andBorges A V Shift in the chemical composition of dissolvedorganic matter in the Congo River network Biogeosciences 135405ndash5420 httpsdoiorg105194bg-13-5405-2016 2016

Leach J Larsson A Wallin M Nilsson M and Laudon HTwelve year interannual and seasonal variability of stream car-bon export from a boreal peatland catchment J Geophys Res121 1851ndash1866 httpsdoiorg1010022016JG003357 2016

Legendre P and Durand S rdaTest Canonical redundancyanalysis R package version 111 available at httpadnbiolumontrealca~numericalecologyRcode (last access 1 January2017) 2014

Lepistouml A Futter M N and Kortelainen P Almost 50 years ofmonitoring shows that climate not forestry controls long-termorganic carbon fluxes in a large boreal watershed Glob ChangeBiol 20 1225ndash1237 httpsdoiorg101111gcb12491 2014

Liaw A and Wiener M Classification and Regression by random-Forest R News 2 18ndash22 2002

Lochmuller C H and Saavedra S S Conformational changes ina soil fulvic acid measured by time dependent fluorescence de-polarization Anal Chem 38 1978ndash1981 1986

Lorenz D Runkel R and De Cicco L rloadest River Load Esti-mation R package version 042 available at httpsgithubcomUSGS-Rrloadest 2015

Ludwig W Probst J and Kempe S Predicting the oceanic inputof organic carbon by continental erosion Global BiogeochemCy 10 23ndash41 httpsdoiorg10102995GB02925 1996

Mann P J Spencer R G M Dinga B J Poulsen J RHernes P J Fiske G Salter M E Wang Z A Hoer-ing K A Six J and Holmes R M The biogeochem-istry of carbon across a gradient of sreams and rivers withinthe Congo Basin J Geophys Res-Biogeo 119 687ndash702httpsdoiorg1010022013JG002442 2014

Marschner B and Kalbitz K Controls on bioavailability andbiodegradability of dissolved organic matter in soils Geoderma113 211ndash235 2003

Martin S L and Soranno P A Lake landscape po-sition Relationships to hydrologic connectivity andlandscape features Limnol Oceanogr 51 801ndash814httpsdoiorg104319lo20065120801 2006

Masiello C A and Druffel E R M Carbon isotope geochemistryof the Santa Clara River Global Biogeochem Cy 15 407ndash416httpsdoiorg1010292000GB001290 2001

Mayorga E Seitzinger S Harrison J Dumont E Beusen ABouwman A F Fekete B Kroeze C and Drecht G GlobalNutrient Export from WaterSheds 2 (NEWS 2) Model develop-ment and implementation Environ Model Softw 25 837ndash853httpsdoiorg101016jenvsoft201001007 2010

McClelland J Townsend-Small A Holmes R Pan F StieglitzM Khosh M and Peterson B River export of nutri-ents and organic matter from the North Slope of Alaskato the Beaufort Sea Water Resour Res 50 1823ndash1839httpsdoiorg1010022013WR014722 2014

McKnight D Boyer E Westerhoff P Doran P KulbeT and Andersen D Spectrofluorometric characterization



of dissolved organic matter for indication of precursor or-ganic material and aromaticity Limnol Oceanogr 46 38ndash48httpsdoiorg104319lo20014610038 2001

McLaren D Fedje D Hay M B Mackie Q Walker I JShugar D H Eamer J B R Lian O B and Neudorf C Apost-glacial sea level hinge on the central Pacific coast of CanadaQuaternary Sci Rev 97 148ndash169 2014

Meybeck M Carbon nitrogen and phosphorus transport byworld rivers Am J Sci 282 401ndash450 available from httpearthgeologyyaleedu~ajs198204198201Maybeckpdf (lastaccess 11 August 2017) 1982

Milliman J D and Syvitski J P M Geomorphic tectonic con-trol of sediment discharge to the ocean The importance of smallmountainous rivers J Geol 100 525ndash544 1992

Moore R D Introduction to salt dilution gauging for stream-flow measurement part III Slug injection using salt in solutionStreamline Watershed Management Bulletin 8 1ndash6 2005

Morrison J Foreman M G G and Masson D Amethod for estimating monthly freshwater discharge affect-ing British Columbia coastal waters Atmos-Ocean 50 1ndash8httpsdoiorg101080070559002011637667 2012

Mulholland P and Watts J Transport of organic carbon tothe oceans by rivers of North America a synthesis of exist-ing data Tellus 34 176ndash186 httpsdoiorg101111j2153-34901982tb01805x 1982

Murphy K Stedmon C Graeber D and Bro R Fluorescencespectroscopy and multi-way techniques PARAFAC Anal Meth-ods 5 6557ndash6566 httpsdoiorg101039C3AY41160E 2013

Murphy K Stedmon C Wenig P and Bro R Open-Fluor ndash A spectral database of auto-fluorescence by organiccompounds in the environment Anal Methods 6 658ndash661httpsdoiorg101039C3AY41935E 2014

Naiman R J Characteristics of sediment and organic carbon ex-port from pristine boreal forest watersheds Can J Fish AquatSci 39 1699ndash1718 httpsdoiorg101139f82-226 1982

Nakagawa S and Schielzeth H A general and simple methodfor obtaining R2 from generalized linear mixed-effects modelsMethods Ecol Evol 4 133ndash142 httpsdoiorg101111j2041-210x201200261x 2013

Olefeldt D Roulet N Giesler R and Persson A To-tal waterborne carbon export and DOC composition fromten nested subarctic peatland catchments-importance of peat-land cover groundwater influence and inter-annual variabil-ity of precipitation patterns Hydrol Process 27 2280ndash2294httpsdoiorg101002hyp9358 2013

Oliver A A Tank S E Giesbrecht I Korver M C Floyd WC Sanborn P Bulmer C and Lertzman K P Aquatic carbonflux data package httpsdoiorg10219661321324 2017

Pinheiro J Bates D DebRoy S Sarkar D and R Core Teamnlme Linear and Nonlinear Mixed Effects Models R packageversion 31-128 2016

Pojar J Klinka K and Demarchi D A Chapter 6 Coastal West-ern Hemlock Zone in Special Report Series 6 Ecosystems ofBritish Columbia edited by Meidiner D and Pojar J Ministryof Forests British Columbia Victoria 330 pp 1991

Poulin B Ryan J and Aiken G Effects of iron on optical prop-erties of dissolved organic matter Environ Sci Technol 4810098ndash106 httpsdoiorg101021es502670r 2014

R Core Team R A language and environment for statistical com-puting R Foundation for Statistical Computing Vienna AustriahttpwwwR-projectorg (last access 11 August 2017) 2013

Raymond P Saiers J and Sobczak W Hydrological andbiogeochemical controls on watershed dissolved organicmatter transport pulse-shunt concept Ecology 97 5ndash16httpsdoiorg10189014-16841 2016

Regnier P Friedlingstein P Ciais P Mackenzie F Gru-ber N Janssens I Laruelle G Lauerwald R LuyssaertS Andersson A Arndt S Arnosti C Borges A DaleA Gallego-Sala A Goddeacuteris Y Goossens N HartmannJ Heinze C Ilyina T Joos F LaRowe D Leifeld JMeysman F Munhoven G Raymond P Spahni R Sunthar-alingam P and Thullner M Anthropogenic perturbation ofthe carbon fluxes from land to ocean Nat Geosci 6 597ndash607httpsdoiorg101038ngeo1830 2013

Roddick J R Geology Rivers Inlet-Queens Sound BritishColumbia Open File 3278 Geological Survey of Canada Ot-tawa Canada 1996

Royer T C Coastal fresh water discharge in the northeast PacificJ Geophys Res 87 2017ndash2021 1982

Runkel R L Crawford C G and Cohn T A Load Estimator(LOADEST) A FORTRAN program for estimating constituentloads in streams and rivers US Geological Survey Techniquesand Methods Book 4 Chapter A5 65 pp 2004

Sanderman J Lohse K A Baldock J A and Amundson RLinking soils and streams Sources and chemistry of dissolvedorganic matter in a small coastal watershed Water Resourc Res45 W03418 httpsdoiorg1010292008WR006977 2009

Spencer R Butler K and Aiken G Dissolved organic car-bon and chromophoric dissolved organic matter properties ofrivers in the USA J Geophys Res-Biogeo 117 G03001httpsdoiorg1010292011JG001928 2012

Spencer R G Hernes P J Ruf R Baker A Dyda R Y Stub-bins A and Six J Temporal controls on dissolved organicmatter and lignin biogeochemistry in a pristine tropical riverDemocratic Republic of Congo J Geophys Res 115 G03013httpsdoiorg1010292009JG001180 2010

Stackpoole S M Butman D E Clow D W Verdin K LGaglioti B and Striegl R Carbon burial transport and emis-sion from inland aquatic ecosystems in Alaska in Baseline andprojected future carbon storage and greenhouse-gas fluxes inecosystems of Alaska edited by Zhiliang Z and David AUS Geological Survey Professional Paper 1826 196 pp 2016

Stackpoole S M Butman D E Clow D W Verdin K LGaglioti B V Genet H and Striegl R G Inland waters andtheir role in the carbon cycle of Alaska Ecol Appl 27 1403ndash1420 httpsdoiorg101002eap1552 2017

Stedmon C and Bro R Characterizing dissolved or-ganic matter fluorescence with parallel factor analy-sis a tutorial Limnol Oceanogr-Meth 6 572ndash579httpsdoiorg104319lom20086572b 2008

Stedmon C and Markager S Tracing the production and degra-dation of autochthonous fractions of dissolved organic matterby fluorescence analysis Limnol Oceanogr 50 1415ndash1426httpsdoiorg104319lo20055051415 2005

Stedmon C Markager S Bro R Stedmon C Markager Sand Bro R Tracing dissolved organic matter in aquatic envi-ronments using a new approach to fluorescence spectroscopy



Mar Chem 82 239ndash254 httpsdoiorg101016S0304-4203(03)00072-0 2003

Stevenson F J Humus Chemistry Genesis Composition Reac-tions 2 Jon Wiley and Sons Inc New York United States ofAmerica 1994

Symonds M R E and Moussalli A A brief guide to model selec-tion multimodel inference and model averaging in behaviouralecology using Akaikersquos information criterion Behav Ecol So-ciobiol 65 13ndash21 httpsdoiorg101007s00265-010-1037-62011

Tallis H Kelp and rivers subsidize rocky intertidal communities inthe Pacific Northwest (USA) Mar Ecol-Prog Ser 389 8596httpsdoiorg103354meps08138 2009

Tank S Raymond P Striegl R McClelland J Holmes RFiske G and Peterson B A land-to-ocean perspective on themagnitude source and implication of DIC flux from major Arc-tic rivers to the Arctic Ocean Global Biogeochem Cy 26GB4018 httpsdoiorg1010292011GB004192 2012

Tank S Striegl R G McClelland J W and Kokelij S V Multi-decadal increases in dissolved organic carbon and alkalinity fluxfrom the Mackenzie drainage basin to the Arctic Ocean EnvironRes Lett 11 httpsdoiorg1010881748-93261150540152016

Thompson S D Nelson T A Giesbrecht I Frazer Gand Saunders S C Data-driven regionalization of forestedand non-forested ecosystems in coastal British Columbia withLiDAR and RapidEye imagery Appl Geogr 69 35ndash50httpsdoiorg101016japgeog201602002 2016

Trant A J Niijland W Hoffman K M Mathews D LMcLaren D Nelson T A and Starzomski B M Inter-tidal resource use over millennia enhances forest productivityNat Commun 7 12491 httpsdoiorg101038ncomms124912016

van Hees P Jones D Finlay R Godbold D and Lund-stroumlm U The carbon we do not see-the impact of lowmolecular weight compounds on carbon dynamics and respi-ration in forest soils a review Soil Biol Biochem 37 1ndash13httpsdoiorg101016jsoilbio200406010 2005

Wallin M Weyhenmeyer G Bastviken D Chmiel H Pe-ter S Sobek S and Klemedtsson L Temporal controlon concentration character and export of dissolved organiccarbon in two hemiboreal headwater streams draining con-trasting catchments J Geophys Res-Biogeo 120 832ndash846httpsdoiorg1010022014jg002814 2015

Wang T Hamann A Spittlehouse D L and Murdock T QClimateWNA- High resolution spatial climate data for West-ern North America J Appl Meterol Climatol 51 16ndash29httpsdoiorg101175JAMC-D-11-0431 2012

Weishaar J L Aiken G R Bergamaschi B A Fram M SFujii R and Mopper K Evaluation of specific ultraviolet ab-sorbance as an indicator of the chemical composition and reactiv-ity of dissolved organic carbon Environ Sci Technol 37 4702ndash4708 httpsdoiorg101021es030360x 2003

Whitney F A Crawford W R and Harrison P J Physical pro-cesses that enhance nutrient transport and primary productivity inthe coastal and open ocean of the subarctic NE Pacific Deep-SeaRes Pt II 52 681ndash706 2005

Wickland K Neff J and Aiken G Dissolved Organic Carbon inAlaskan Boreal Forest Sources Chemical Characteristics andBiodegradability Ecosystems 10 1323ndash1340 2007

Wilson H F and Xenopoulos M A Effects of agricultural landuse on the composition of fluvial dissolved organic matter NatGeosci 2 37ndash41 httpsdoiorg101038ngeo391 2009

Wolf E C Mitchell A P and Schoonmaker P K The RainForests of Home An Atlas of People and Place Ecotrust PacificGIS Inforain and Conservation International Portland Oregon24 pp available at httpwwwinforainorgpdfsctrf_atlas_origpdf 1995

Worrall F Burt T and Adamson J Can climatechange explain increases in DOC flux from uplandpeat catchements Sci Total Environ 326 95ndash112httpsdoiorg101016jscitotenv200311022 2004

Xenopoulos M A Lodge D M Frentress J Kreps T ABridgham S D Grossman E and Jackson C J Regionalcomparisons of watershed determinants of dissolved organic car-bon in temperate lakes from the Upper Great Lakes region andselected regions globally Limnol Oceanogr 48 2321ndash23342003

Yamashita Y and Jaffeigrave R Characterizing the Interac-tions between Trace Metals and Dissolved Organic Mat-ter Using ExcitationndashEmission Matrix and Parallel Fac-tor Analysis Environ Sci Technol 42 7374ndash7379httpsdoiorg101021es801357h 2008

Yamashita Y Kloeppel B Knoepp J Zausen G and Jaffeacute REffects of Watershed History on Dissolved Organic Matter Char-acteristics in Headwater Streams Ecosystems 14 1110ndash1122httpsdoiorg101007s10021-011-9469-z 2011


Abstract

Introduction

Methods

Study sites

Soils and watershed characteristics

Sample collection and analysis

Hydrology precipitation and stream discharge

DOC flux

Optical characterization of DOM

Evaluating relationships in DOC concentration and DOM composition with stream discharge and temperature

Redundancy analysis relationships between DOC concentration DOM composition and watershed characteristics

Results

Hydrology

Temporal and spatial patterns in DOC concentration yield and flux

Temporal and spatial patterns in DOM composition

PARAFAC characterization of DOM

Relationships between watershed characteristics DOC concentrations and DOM composition

Discussion

DOC export from small catchments to the coastal ocean

DOM composition

DOC and DOM export sources and seasonal variability

Relationships between watershed attributes and exported DOM

Conclusions

Data availability

Author contributions

Competing interests

Acknowledgements

References


coastal ocean (Cole et al 2007) with typically gt 50 quan-tified as dissolved organic carbon (DOC) (Meybeck 1982Ludwig et al 1996 Alvarez-Cobelas et al 2012 Mayorgaet al 2010) Rivers draining coastal watersheds serve as con-duits of DOC from terrestrial and freshwater sources to ma-rine environments (Mulholland and Watts 1982 Bauer et al2013 McClelland et al 2014) and can have important im-plications for coastal carbon cycling biogeochemical inter-actions ecosystem productivity and food webs (Hopkinsonet al 1998 Tallis 2009 Tank et al 2012 Regnier et al2013) In addition because the transfer of water and organicmatter from watersheds to the coastal ocean represents an im-portant pathway for carbon cycling and ecological subsidiesbetween ecosystems better understanding of these linkagesis needed for constraining predictions of ecosystem produc-tivity in response to perturbations such as climate change Inregions where empirical data are currently scarce quantify-ing land-to-ocean DOC export is therefore a priority for im-proving the accuracy of watershed and coastal carbon models(Bauer et al 2013)

While quantifying DOC flux within and across systemsis required for understanding the magnitude of carbon ex-change the composition of DOC (as dissolved organic mat-ter or DOM) is also important for determining the ecologi-cal significance of carbon exported from coastal watershedsThe aquatic DOM pool is a complex mixture that reflectsboth source material and processing along the watershedterrestrialndashaquatic continuum and as a result it can show sig-nificant spatial and temporal variation (Hudson et al 2007Graeber et al 2012 Wallin et al 2015) Both DOC con-centration and DOM composition can serve as indicators ofwatershed characteristics (Koehler et al 2009) hydrologicflow paths (Johnson et al 2011 Helton et al 2015) and wa-tershed biogeochemical processes (Emili and Price 2013)DOM composition can also influence its role in downstreamprocessing and ecological function such as susceptibility tobiological (Judd et al 2006) and physiochemical interac-tions (Yamashita and Jaffeacute 2008)

The coastal temperate rainforests (CTRs) of Pacific NorthAmerica extend from the Gulf of Alaska through BritishColumbia to Northern California and span a wide rangeof precipitation and climate regimes Within this rainforestregion the ldquoperhumidrdquo zone has cool summers and sum-mer precipitation is common (gt 10 of annual precipita-tion) (Alaback 1996) (Fig 1) The perhumid CTR extendsfrom southeast Alaska through the outer coast of centralBritish Columbia and contains forests and soils that have ac-cumulated large amounts of organic carbon above and be-low ground (Leighty et al 2006 Gorham et al 2012) Dueto high amounts of precipitation and close proximity to thecoast this area represents a potential hotspot for the trans-port and metabolism of carbon across the land-to-ocean con-tinuum and quantifying these fluxes is pertinent for under-standing global carbon cycling

Within the large perhumid CTR there is substantialspatial variation in climate and landscape characteristicsthat create uncertainty about carbon cycling and patternIn Alaska for example riverine DOC concentrations varywith wetland cover (DrsquoAmore et al 2015a) and glacialcover (Fellman et al 2014) Previous studies have shownthat streams in southeast Alaska can contain high DOCconcentrations (Fellman et al 2009a DrsquoAmore et al2015a) and produce high DOC yields (DrsquoAmore et al2015a b Stackpoole et al 2016) but no known fieldestimates have been generated for the perhumid CTR ofBritish Columbia an area of approximately 97 824 km2

(adapted from Wolf et al 1995) Within the perhumid CTRof British Columbia terrestrial ecologists have defined alarge (29 935 km2) ldquohypermaritimerdquo subregion where rain-fall dominates over snow seasonality is moderated by theocean and wetlands are extensive (Pojar et al 1991 areaestimated using British Columbia Biogeoclimatic Ecosys-tem Classification SubzoneVariant mapping Version 1031 August 2016 available at httpscataloguedatagovbccadatasetf358a53b-ffde-4830-a325-a5a03ff672c3) Previ-ous work in the hypermaritime CTR showed that DOC con-centrations are high in small streams and tend to increaseduring rain events (Gibson et al 2000 Fitzgerald et al2003 Emili and Price 2013) Taken together these condi-tions should be expected to generate high yields and fluxes ofDOC 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 ex-tensive hypermaritime region of British Columbiarsquos perhu-mid CTR to describe the temporal and spatial dynamicsof exported DOC concentration and DOM composition andto identify relationships between DOC concentration DOMcomposition and watershed characteristics

2 Methods

21 Study sites

Study sites are located on northern Calvert Island and adja-cent Hecate Island on the central coast of British ColumbiaCanada (lat 51650 long minus128035 Fig 1) Average annualprecipitation and air temperature at sea level from 1981 to2010 was 3356 mm yrminus1 and 84 C (average annual min09 C average annual max 179 C) (available online athttpwwwclimatewnacom Wang et al 2012) with pre-cipitation dominated by rain and winter snowpack persist-ing only at higher elevations Sites are located within thehypermaritime region of the CTR on the outer coast ofBritish Columbia Soils overlying the granodiorite bedrock(Roddick 1996) are usually lt 1 m thick and have formedin sandy colluvium and patchy morainal deposits with lim-ited areas of coarse glacial outwash Chemical weatheringand organic-matter accumulation in the cool moist climate



Tabl

e1

Wat

ersh

edch

arac

teri

stic

sdi

scha

rge

DO

Cco

ncen

trat

ions

and

DO

Cyi

elds

for

the

seve

nst

udy

wat

ersh

eds

onC

alve

rtan

dH

ecat

eis

land

sA

dditi

onal

deta

ilson

the

met

hods

used

tode

term

ine

wat

ersh

edch

arac

teri

stic

sca

nbe

foun

din

the

Supp

lem

ent

Wat

ersh

edA

rea

Avg

L

akes

Wet

land

sA

vgd

epth

Avg

dep

thTo

tal

DO

C1

aQ

-wei

ghte

dD

OC

DO

CD

OC

slop

eor

gani

cm

iner

alQ

yiel

d1av

gD

OC

1an

nual

mon

thly

mon

thly

soils

soils

yiel

dbyi

eldb

yiel

db

WY

2015

1w

etse

ason

2dr

yse

ason

3

(km

2 )(

)(

Are

a)(

Are

a)(c

m)

(cm

)(m

m)

(mg

Lminus

1 )(m

gLminus

1 )(M

gC

kmminus

2 )(M

gC

kmminus

2 )(M

gC

kmminus

2 )

626

32

217

47

480

394plusmn

243

308plusmn

83

3673

110plusmn

35

153

377

359

062

(31

9ndash44

2)

(30

5ndash4

18)

(04

9ndash0

77)

1015

33

342

91

238

395plusmn

172

337plusmn

86

3052

112plusmn

16

129

247

256

027

(23

6ndash25

8)

(24

5ndash2

78)

(02

5ndash0

28)

819

48

301

03

502

379plusmn

191

298plusmn

57

3066

140plusmn

35

193

357

380

057

(31

7ndash40

2)

(33

7ndash5

10)

(04

8ndash0

67)

844

57

325

03

352

354plusmn

180

291plusmn

64

4129

131plusmn

36

159

436

424

054

(34

2ndash54

9)

(33

6ndash5

30)

(03

6ndash0

77)

708

78

285

75

463

362plusmn

197

299plusmn

60

3805

95plusmn

24

109

241

267

038

(22

2ndash26

0)

(24

6ndash4

07)

(03

4ndash0

43)

693

93

302

44

428

354plusmn

161

302plusmn

64

5866

77plusmn

25

84

297

319

041

(25

9ndash34

0)

(27

9ndash4

94)

(03

2ndash0

52)

703

128

403

19

243

373plusmn

165

358plusmn

134

6058

63plusmn

26

90

370

348

064

(32

5ndash42

0)

(30

7ndash4

02)

(05

2ndash0

77)

All

469

327

37

371

374plusmn

177

322plusmn

92

4730

104plusmn

38

111

333

335

050

(28

9ndash38

1)

(29

4ndash4

40)

(04

1ndash0

62)

1C

alcu

late

dfo

rwat

erye

ar20

15(W

Y20

151

Oct

ober

2014

ndash30

Sept

embe

r201

5)2

Wet

-per

iod

aver

age

mon

thly

yiel

dca

lcul

ated

from

Oct

ober

toA

pril

and

Sept

embe

rW

Y20

15a

ndO

ctob

erto

Apr

ilW

Y20

163

Dry

-per

iod

aver

age

mon

thly

yiel

dca

lcul

ated

from

May

toA

ugus

tW

Y20

15a

Mea

nplusmn

stan

dard

devi

atio

nb

Tota

lplusmn95

co

nfide

nce

inte

rval
















25 DOC flux















3 Results

31 Hydrology








= 057marginal R2





o










=




= 035marginal R2




= 059marginal R2








Table2Spectralcom

positionforthe

sixfluorescence

components

identifiedusing

PAR

AFA

Cincluding

excitation(E

x)andem

ission(E

m)


positionacross

allsamplesand

likelystructure

andcharacteristics

ofthefluorescentcom

ponentbasedon

previousstudies

Com

ponentE

x(nm)

Em

(nm)

composition

lowast


Previousstudies

with

comparable

results

C1

315436

341plusmn

22(311ndash393)

Hum


terrestrialhigh

molecularw

eightwidespread

buthighestin

wetland

andforestenvironm

ent

Garcia

etal(2015)(C1)G

raeberetal(2012)

(C1)W

alkeretal(2014)

(C1)

Yam

ashitaetal(2011)

(C1)C

oryand

McK

night(2005)(C1)

C2

270380484

202plusmn

19(161ndash256)

Hum

ic-likeresembles

fulvicacid

widespreadhigh

molecularw

eightterrestrial

Stedmon

andM

arkager(2005)

(C2)

Stedmon

etal

(2003)(C

3)C

oryand

McK

night(2005)(C5)

C3

270478

178plusmn

18(128ndash208)

Hum



asrefractory

Stedmon

andM

arkager(2005)

(C1)

Yam

ashitaetal(2010)(C

2)

C4

305435522

148plusmn

26(94ndash223)

Notcom

monly



fromsoils

Lochm

ullerandSaavedra

(1986)(E)

C5

325442

98plusmn

35(00ndash159)

Aquatic

humic-like

fromterrestrial


icrobialproducedm

aybe

photoproduced

Boehm

eand

Coble

(2000)(Peak

C)

Coble

etal(1998)

(PeakC

)Stedmon

etal(2003)(C3)

C6

285338

34plusmn

25(00ndash93)

Am


Freshlyadded


Rapidly

photodegradable

Murphy

etal

(2008)(C

7)Shutova

etal(2003)(C

4)Stedm

onetal(2007)

(C7)Y

amashita

etal(2003)(C5)

lowastM

eanplusmn

SD(m

inndashmax)from

allsamples



con

tribu

tion

to to

tal f

luor

esce

nce








4 Discussion





1

2

3

4

5

6

1

2

34

6

1

2 34

5

6

1

2

34

56

1

2

3

4

5

6

1

23

45

6

1 2

3

45

6

DOC

DO13C

Sr

FI

SUVA FRESHC1

C4

C6

Lakes

Slope

Wetlands

MinThick

OrgThick

-1

0

1

2

-2 -1 0 1 2RDA1

RD

A2

Watershed626

693

703

708

819

844

1015







42 DOM composition


















5 Conclusions












References

























































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References


Tabl

e1

Wat

ersh

edch

arac

teri

stic

sdi

scha

rge

DO

Cco

ncen

trat

ions

and

DO

Cyi

elds

for

the

seve

nst

udy

wat

ersh

eds

onC

alve

rtan

dH

ecat

eis

land

sA

dditi

onal

deta

ilson

the

met

hods

used

tode

term

ine

wat

ersh

edch

arac

teri

stic

sca

nbe

foun

din

the

Supp

lem

ent

Wat

ersh

edA

rea

Avg

L

akes

Wet

land

sA

vgd

epth

Avg

dep

thTo

tal

DO

C1

aQ

-wei

ghte

dD

OC

DO

CD

OC

slop

eor

gani

cm

iner

alQ

yiel

d1av

gD

OC

1an

nual

mon

thly

mon

thly

soils

soils

yiel

dbyi

eldb

yiel

db

WY

2015

1w

etse

ason

2dr

yse

ason

3

(km

2 )(

)(

Are

a)(

Are

a)(c

m)

(cm

)(m

m)

(mg

Lminus

1 )(m

gLminus

1 )(M

gC

kmminus

2 )(M

gC

kmminus

2 )(M

gC

kmminus

2 )

626

32

217

47

480

394plusmn

243

308plusmn

83

3673

110plusmn

35

153

377

359

062

(31

9ndash44

2)

(30

5ndash4

18)

(04

9ndash0

77)

1015

33

342

91

238

395plusmn

172

337plusmn

86

3052

112plusmn

16

129

247

256

027

(23

6ndash25

8)

(24

5ndash2

78)

(02

5ndash0

28)

819

48

301

03

502

379plusmn

191

298plusmn

57

3066

140plusmn

35

193

357

380

057

(31

7ndash40

2)

(33

7ndash5

10)

(04

8ndash0

67)

844

57

325

03

352

354plusmn

180

291plusmn

64

4129

131plusmn

36

159

436

424

054

(34

2ndash54

9)

(33

6ndash5

30)

(03

6ndash0

77)

708

78

285

75

463

362plusmn

197

299plusmn

60

3805

95plusmn

24

109

241

267

038

(22

2ndash26

0)

(24

6ndash4

07)

(03

4ndash0

43)

693

93

302

44

428

354plusmn

161

302plusmn

64

5866

77plusmn

25

84

297

319

041

(25

9ndash34

0)

(27

9ndash4

94)

(03

2ndash0

52)

703

128

403

19

243

373plusmn

165

358plusmn

134

6058

63plusmn

26

90

370

348

064

(32

5ndash42

0)

(30

7ndash4

02)

(05

2ndash0

77)

All

469

327

37

371

374plusmn

177

322plusmn

92

4730

104plusmn

38

111

333

335

050

(28

9ndash38

1)

(29

4ndash4

40)

(04

1ndash0

62)

1C

alcu

late

dfo

rwat

erye

ar20

15(W

Y20

151

Oct

ober

2014

ndash30

Sept

embe

r201

5)2

Wet

-per

iod

aver

age

mon

thly

yiel

dca

lcul

ated

from

Oct

ober

toA

pril

and

Sept

embe

rW

Y20

15a

ndO

ctob

erto

Apr

ilW

Y20

163

Dry

-per

iod

aver

age

mon

thly

yiel

dca

lcul

ated

from

May

toA

ugus

tW

Y20

15a

Mea

nplusmn

stan

dard

devi

atio

nb

Tota

lplusmn95

co

nfide

nce

inte

rval
















25 DOC flux















3 Results

31 Hydrology








= 057marginal R2





o










=




= 035marginal R2




= 059marginal R2








Table2Spectralcom

positionforthe

sixfluorescence

components

identifiedusing

PAR

AFA

Cincluding

excitation(E

x)andem

ission(E

m)


positionacross

allsamplesand

likelystructure

andcharacteristics

ofthefluorescentcom

ponentbasedon

previousstudies

Com

ponentE

x(nm)

Em

(nm)

composition

lowast


Previousstudies

with

comparable

results

C1

315436

341plusmn

22(311ndash393)

Hum


terrestrialhigh

molecularw

eightwidespread

buthighestin

wetland

andforestenvironm

ent

Garcia

etal(2015)(C1)G

raeberetal(2012)

(C1)W

alkeretal(2014)

(C1)

Yam

ashitaetal(2011)

(C1)C

oryand

McK

night(2005)(C1)

C2

270380484

202plusmn

19(161ndash256)

Hum

ic-likeresembles

fulvicacid

widespreadhigh

molecularw

eightterrestrial

Stedmon

andM

arkager(2005)

(C2)

Stedmon

etal

(2003)(C

3)C

oryand

McK

night(2005)(C5)

C3

270478

178plusmn

18(128ndash208)

Hum



asrefractory

Stedmon

andM

arkager(2005)

(C1)

Yam

ashitaetal(2010)(C

2)

C4

305435522

148plusmn

26(94ndash223)

Notcom

monly



fromsoils

Lochm

ullerandSaavedra

(1986)(E)

C5

325442

98plusmn

35(00ndash159)

Aquatic

humic-like

fromterrestrial


icrobialproducedm

aybe

photoproduced

Boehm

eand

Coble

(2000)(Peak

C)

Coble

etal(1998)

(PeakC

)Stedmon

etal(2003)(C3)

C6

285338

34plusmn

25(00ndash93)

Am


Freshlyadded


Rapidly

photodegradable

Murphy

etal

(2008)(C

7)Shutova

etal(2003)(C

4)Stedm

onetal(2007)

(C7)Y

amashita

etal(2003)(C5)

lowastM

eanplusmn

SD(m

inndashmax)from

allsamples



con

tribu

tion

to to

tal f

luor

esce

nce








4 Discussion





1

2

3

4

5

6

1

2

34

6

1

2 34

5

6

1

2

34

56

1

2

3

4

5

6

1

23

45

6

1 2

3

45

6

DOC

DO13C

Sr

FI

SUVA FRESHC1

C4

C6

Lakes

Slope

Wetlands

MinThick

OrgThick

-1

0

1

2

-2 -1 0 1 2RDA1

RD

A2

Watershed626

693

703

708

819

844

1015







42 DOM composition


















5 Conclusions












References

























































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References















25 DOC flux















3 Results

31 Hydrology








= 057marginal R2





o










=




= 035marginal R2




= 059marginal R2








Table2Spectralcom

positionforthe

sixfluorescence

components

identifiedusing

PAR

AFA

Cincluding

excitation(E

x)andem

ission(E

m)


positionacross

allsamplesand

likelystructure

andcharacteristics

ofthefluorescentcom

ponentbasedon

previousstudies

Com

ponentE

x(nm)

Em

(nm)

composition

lowast


Previousstudies

with

comparable

results

C1

315436

341plusmn

22(311ndash393)

Hum


terrestrialhigh

molecularw

eightwidespread

buthighestin

wetland

andforestenvironm

ent

Garcia

etal(2015)(C1)G

raeberetal(2012)

(C1)W

alkeretal(2014)

(C1)

Yam

ashitaetal(2011)

(C1)C

oryand

McK

night(2005)(C1)

C2

270380484

202plusmn

19(161ndash256)

Hum

ic-likeresembles

fulvicacid

widespreadhigh

molecularw

eightterrestrial

Stedmon

andM

arkager(2005)

(C2)

Stedmon

etal

(2003)(C

3)C

oryand

McK

night(2005)(C5)

C3

270478

178plusmn

18(128ndash208)

Hum



asrefractory

Stedmon

andM

arkager(2005)

(C1)

Yam

ashitaetal(2010)(C

2)

C4

305435522

148plusmn

26(94ndash223)

Notcom

monly



fromsoils

Lochm

ullerandSaavedra

(1986)(E)

C5

325442

98plusmn

35(00ndash159)

Aquatic

humic-like

fromterrestrial


icrobialproducedm

aybe

photoproduced

Boehm

eand

Coble

(2000)(Peak

C)

Coble

etal(1998)

(PeakC

)Stedmon

etal(2003)(C3)

C6

285338

34plusmn

25(00ndash93)

Am


Freshlyadded


Rapidly

photodegradable

Murphy

etal

(2008)(C

7)Shutova

etal(2003)(C

4)Stedm

onetal(2007)

(C7)Y

amashita

etal(2003)(C5)

lowastM

eanplusmn

SD(m

inndashmax)from

allsamples



con

tribu

tion

to to

tal f

luor

esce

nce








4 Discussion





1

2

3

4

5

6

1

2

34

6

1

2 34

5

6

1

2

34

56

1

2

3

4

5

6

1

23

45

6

1 2

3

45

6

DOC

DO13C

Sr

FI

SUVA FRESHC1

C4

C6

Lakes

Slope

Wetlands

MinThick

OrgThick

-1

0

1

2

-2 -1 0 1 2RDA1

RD

A2

Watershed626

693

703

708

819

844

1015







42 DOM composition


















5 Conclusions












References

























































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References











3 Results

31 Hydrology








= 057marginal R2





o










=




= 035marginal R2




= 059marginal R2








Table2Spectralcom

positionforthe

sixfluorescence

components

identifiedusing

PAR

AFA

Cincluding

excitation(E

x)andem

ission(E

m)


positionacross

allsamplesand

likelystructure

andcharacteristics

ofthefluorescentcom

ponentbasedon

previousstudies

Com

ponentE

x(nm)

Em

(nm)

composition

lowast


Previousstudies

with

comparable

results

C1

315436

341plusmn

22(311ndash393)

Hum


terrestrialhigh

molecularw

eightwidespread

buthighestin

wetland

andforestenvironm

ent

Garcia

etal(2015)(C1)G

raeberetal(2012)

(C1)W

alkeretal(2014)

(C1)

Yam

ashitaetal(2011)

(C1)C

oryand

McK

night(2005)(C1)

C2

270380484

202plusmn

19(161ndash256)

Hum

ic-likeresembles

fulvicacid

widespreadhigh

molecularw

eightterrestrial

Stedmon

andM

arkager(2005)

(C2)

Stedmon

etal

(2003)(C

3)C

oryand

McK

night(2005)(C5)

C3

270478

178plusmn

18(128ndash208)

Hum



asrefractory

Stedmon

andM

arkager(2005)

(C1)

Yam

ashitaetal(2010)(C

2)

C4

305435522

148plusmn

26(94ndash223)

Notcom

monly



fromsoils

Lochm

ullerandSaavedra

(1986)(E)

C5

325442

98plusmn

35(00ndash159)

Aquatic

humic-like

fromterrestrial


icrobialproducedm

aybe

photoproduced

Boehm

eand

Coble

(2000)(Peak

C)

Coble

etal(1998)

(PeakC

)Stedmon

etal(2003)(C3)

C6

285338

34plusmn

25(00ndash93)

Am


Freshlyadded


Rapidly

photodegradable

Murphy

etal

(2008)(C

7)Shutova

etal(2003)(C

4)Stedm

onetal(2007)

(C7)Y

amashita

etal(2003)(C5)

lowastM

eanplusmn

SD(m

inndashmax)from

allsamples



con

tribu

tion

to to

tal f

luor

esce

nce








4 Discussion





1

2

3

4

5

6

1

2

34

6

1

2 34

5

6

1

2

34

56

1

2

3

4

5

6

1

23

45

6

1 2

3

45

6

DOC

DO13C

Sr

FI

SUVA FRESHC1

C4

C6

Lakes

Slope

Wetlands

MinThick

OrgThick

-1

0

1

2

-2 -1 0 1 2RDA1

RD

A2

Watershed626

693

703

708

819

844

1015







42 DOM composition


















5 Conclusions












References

























































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References


o










=




= 035marginal R2




= 059marginal R2








Table2Spectralcom

positionforthe

sixfluorescence

components

identifiedusing

PAR

AFA

Cincluding

excitation(E

x)andem

ission(E

m)


positionacross

allsamplesand

likelystructure

andcharacteristics

ofthefluorescentcom

ponentbasedon

previousstudies

Com

ponentE

x(nm)

Em

(nm)

composition

lowast


Previousstudies

with

comparable

results

C1

315436

341plusmn

22(311ndash393)

Hum


terrestrialhigh

molecularw

eightwidespread

buthighestin

wetland

andforestenvironm

ent

Garcia

etal(2015)(C1)G

raeberetal(2012)

(C1)W

alkeretal(2014)

(C1)

Yam

ashitaetal(2011)

(C1)C

oryand

McK

night(2005)(C1)

C2

270380484

202plusmn

19(161ndash256)

Hum

ic-likeresembles

fulvicacid

widespreadhigh

molecularw

eightterrestrial

Stedmon

andM

arkager(2005)

(C2)

Stedmon

etal

(2003)(C

3)C

oryand

McK

night(2005)(C5)

C3

270478

178plusmn

18(128ndash208)

Hum



asrefractory

Stedmon

andM

arkager(2005)

(C1)

Yam

ashitaetal(2010)(C

2)

C4

305435522

148plusmn

26(94ndash223)

Notcom

monly



fromsoils

Lochm

ullerandSaavedra

(1986)(E)

C5

325442

98plusmn

35(00ndash159)

Aquatic

humic-like

fromterrestrial


icrobialproducedm

aybe

photoproduced

Boehm

eand

Coble

(2000)(Peak

C)

Coble

etal(1998)

(PeakC

)Stedmon

etal(2003)(C3)

C6

285338

34plusmn

25(00ndash93)

Am


Freshlyadded


Rapidly

photodegradable

Murphy

etal

(2008)(C

7)Shutova

etal(2003)(C

4)Stedm

onetal(2007)

(C7)Y

amashita

etal(2003)(C5)

lowastM

eanplusmn

SD(m

inndashmax)from

allsamples



con

tribu

tion

to to

tal f

luor

esce

nce








4 Discussion





1

2

3

4

5

6

1

2

34

6

1

2 34

5

6

1

2

34

56

1

2

3

4

5

6

1

23

45

6

1 2

3

45

6

DOC

DO13C

Sr

FI

SUVA FRESHC1

C4

C6

Lakes

Slope

Wetlands

MinThick

OrgThick

-1

0

1

2

-2 -1 0 1 2RDA1

RD

A2

Watershed626

693

703

708

819

844

1015







42 DOM composition


















5 Conclusions












References

























































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References



=




= 035marginal R2




= 059marginal R2








Table2Spectralcom

positionforthe

sixfluorescence

components

identifiedusing

PAR

AFA

Cincluding

excitation(E

x)andem

ission(E

m)


positionacross

allsamplesand

likelystructure

andcharacteristics

ofthefluorescentcom

ponentbasedon

previousstudies

Com

ponentE

x(nm)

Em

(nm)

composition

lowast


Previousstudies

with

comparable

results

C1

315436

341plusmn

22(311ndash393)

Hum


terrestrialhigh

molecularw

eightwidespread

buthighestin

wetland

andforestenvironm

ent

Garcia

etal(2015)(C1)G

raeberetal(2012)

(C1)W

alkeretal(2014)

(C1)

Yam

ashitaetal(2011)

(C1)C

oryand

McK

night(2005)(C1)

C2

270380484

202plusmn

19(161ndash256)

Hum

ic-likeresembles

fulvicacid

widespreadhigh

molecularw

eightterrestrial

Stedmon

andM

arkager(2005)

(C2)

Stedmon

etal

(2003)(C

3)C

oryand

McK

night(2005)(C5)

C3

270478

178plusmn

18(128ndash208)

Hum



asrefractory

Stedmon

andM

arkager(2005)

(C1)

Yam

ashitaetal(2010)(C

2)

C4

305435522

148plusmn

26(94ndash223)

Notcom

monly



fromsoils

Lochm

ullerandSaavedra

(1986)(E)

C5

325442

98plusmn

35(00ndash159)

Aquatic

humic-like

fromterrestrial


icrobialproducedm

aybe

photoproduced

Boehm

eand

Coble

(2000)(Peak

C)

Coble

etal(1998)

(PeakC

)Stedmon

etal(2003)(C3)

C6

285338

34plusmn

25(00ndash93)

Am


Freshlyadded


Rapidly

photodegradable

Murphy

etal

(2008)(C

7)Shutova

etal(2003)(C

4)Stedm

onetal(2007)

(C7)Y

amashita

etal(2003)(C5)

lowastM

eanplusmn

SD(m

inndashmax)from

allsamples



con

tribu

tion

to to

tal f

luor

esce

nce








4 Discussion





1

2

3

4

5

6

1

2

34

6

1

2 34

5

6

1

2

34

56

1

2

3

4

5

6

1

23

45

6

1 2

3

45

6

DOC

DO13C

Sr

FI

SUVA FRESHC1

C4

C6

Lakes

Slope

Wetlands

MinThick

OrgThick

-1

0

1

2

-2 -1 0 1 2RDA1

RD

A2

Watershed626

693

703

708

819

844

1015







42 DOM composition


















5 Conclusions












References

























































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References


Table2Spectralcom

positionforthe

sixfluorescence

components

identifiedusing

PAR

AFA

Cincluding

excitation(E

x)andem

ission(E

m)


positionacross

allsamplesand

likelystructure

andcharacteristics

ofthefluorescentcom

ponentbasedon

previousstudies

Com

ponentE

x(nm)

Em

(nm)

composition

lowast


Previousstudies

with

comparable

results

C1

315436

341plusmn

22(311ndash393)

Hum


terrestrialhigh

molecularw

eightwidespread

buthighestin

wetland

andforestenvironm

ent

Garcia

etal(2015)(C1)G

raeberetal(2012)

(C1)W

alkeretal(2014)

(C1)

Yam

ashitaetal(2011)

(C1)C

oryand

McK

night(2005)(C1)

C2

270380484

202plusmn

19(161ndash256)

Hum

ic-likeresembles

fulvicacid

widespreadhigh

molecularw

eightterrestrial

Stedmon

andM

arkager(2005)

(C2)

Stedmon

etal

(2003)(C

3)C

oryand

McK

night(2005)(C5)

C3

270478

178plusmn

18(128ndash208)

Hum



asrefractory

Stedmon

andM

arkager(2005)

(C1)

Yam

ashitaetal(2010)(C

2)

C4

305435522

148plusmn

26(94ndash223)

Notcom

monly



fromsoils

Lochm

ullerandSaavedra

(1986)(E)

C5

325442

98plusmn

35(00ndash159)

Aquatic

humic-like

fromterrestrial


icrobialproducedm

aybe

photoproduced

Boehm

eand

Coble

(2000)(Peak

C)

Coble

etal(1998)

(PeakC

)Stedmon

etal(2003)(C3)

C6

285338

34plusmn

25(00ndash93)

Am


Freshlyadded


Rapidly

photodegradable

Murphy

etal

(2008)(C

7)Shutova

etal(2003)(C

4)Stedm

onetal(2007)

(C7)Y

amashita

etal(2003)(C5)

lowastM

eanplusmn

SD(m

inndashmax)from

allsamples



con

tribu

tion

to to

tal f

luor

esce

nce








4 Discussion





1

2

3

4

5

6

1

2

34

6

1

2 34

5

6

1

2

34

56

1

2

3

4

5

6

1

23

45

6

1 2

3

45

6

DOC

DO13C

Sr

FI

SUVA FRESHC1

C4

C6

Lakes

Slope

Wetlands

MinThick

OrgThick

-1

0

1

2

-2 -1 0 1 2RDA1

RD

A2

Watershed626

693

703

708

819

844

1015







42 DOM composition


















5 Conclusions












References

























































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References


con

tribu

tion

to to

tal f

luor

esce

nce








4 Discussion





1

2

3

4

5

6

1

2

34

6

1

2 34

5

6

1

2

34

56

1

2

3

4

5

6

1

23

45

6

1 2

3

45

6

DOC

DO13C

Sr

FI

SUVA FRESHC1

C4

C6

Lakes

Slope

Wetlands

MinThick

OrgThick

-1

0

1

2

-2 -1 0 1 2RDA1

RD

A2

Watershed626

693

703

708

819

844

1015







42 DOM composition


















5 Conclusions












References

























































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References


1

2

3

4

5

6

1

2

34

6

1

2 34

5

6

1

2

34

56

1

2

3

4

5

6

1

23

45

6

1 2

3

45

6

DOC

DO13C

Sr

FI

SUVA FRESHC1

C4

C6

Lakes

Slope

Wetlands

MinThick

OrgThick

-1

0

1

2

-2 -1 0 1 2RDA1

RD

A2

Watershed626

693

703

708

819

844

1015







42 DOM composition


















5 Conclusions












References

























































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References


42 DOM composition


















5 Conclusions












References

























































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References









5 Conclusions












References

























































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References










References

























































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References






















































































































































Abstract

Introduction

Methods

Study sites




DOC flux




Results

Hydrology





Discussion


DOM composition



Conclusions

Data availability


Competing interests

Acknowledgements

References

a global hotspot for dissolved organic carbon in

Documents