AbstractIn the past decade, significant increases in surface water dissolved organic carbon (DOC) have been reported for large aquatic ecosystems of the Northern Hemisphere and have been attributed variously to global warming, altered hydrologic conditions, and atmospheric deposition, among other factors. We analyzed a 25-yr DOC record (1988–2012) available for a forested headwater stream in the United States and documented two distinct regimes of stream DOC trends. From 1988 to 2001, annual mean volume-weighted DOC concentration (DOCvw, mg L-1) and annual DOC flux (kg ha-1 yr-1) declined by 34 and 56%, respectively. During 1997 to 2012, the decline in DOCvw and DOC flux increased by 141 and 165%, respectively. Declining DOCvw from 1988 to 2001 corresponded to a decline in growing season runoff, which has the potential to influence mobilization of DOC from uplands to streams. Increasing DOCvw from 1997 to 2012 corresponded to increased precipitation early in the growing season and to an increase in the number and intensity of short-duration fall storms capable of mobilizing long-accrued DOC from forest litter and soils. In contrast, total annual runoff declined throughout the period. Rising air temperature, atmospheric acid deposition, and nitrogen depositions did not offer any plausible explanation for the observed bidirectional annual trends of stream DOCvw. Our study highlights the critical role of long-term datasets and analyses for understanding the impacts of climate change on carbon and water cycles and associated functions of aquatic and terrestrial ecosystems.

Hydro-Climatological Influences on Long-Term Dissolved Organic Carbon in a Mountain Stream of the Southeastern United States

Nitin K. Singh,* Wilmer M. Reyes, Emily S. Bernhardt, Ruchi Bhattacharya, Judy L. Meyer, Jennifer D. Knoepp, and Ryan E. Emanuel

Headwater streams constitute almost 70% of the total length of streams in the United States (Leopold et al., 1964). These streams can influence the global

carbon cycle by outgassing CO2 (Butman and Raymond, 2011) and by transporting large amounts of organic carbon to down-stream environments (e.g., Reichstein et al., 2013; West et al., 2011). Headwater streams are intimately linked with their sur-rounding terrestrial ecosystems and are highly sensitive to bio-geochemical processes within their contributing areas (Lowe and Likens, 2005). Due to interactions between headwater streams and the surrounding terrestrial ecosystems (Beckman and Wohl, 2014; Gomi et al., 2002), headwater streams receive most of their organic carbon in the form of dissolved organic carbon (DOC) (e.g., Meyer and Tate, 1983). In forested headwater streams, DOC influences energy supplies (e.g., Aitkenhead-Peterson et al., 2003), trophic resources (e.g., Hall and Meyer, 1998), and water quality (e.g., McKnight, 1981; Stewart and Wetzel, 1982). Furthermore, DOC not only dominates the carbon cycle of headwater streams (Dahm, 1981); it also mediates biogeochemi-cal cycles of nitrogen (Bernhardt and Likens, 2002), phosphorus (Ardón et al., 2006), and other elements. At high concentrations, DOC imparts color to the stream and protects microorganisms from ultraviolet radiation (Curtis and Schindler, 1997).

Table 1 summarizes most of the long-term (>10 yr) studies that have reported significant change in freshwater DOC con-centration and export in the Northern Hemisphere. Several of these studies have investigated the potential drivers of long-term DOC patterns, but no single driver has been implicated among all studies. Some of the major drivers include rising air tempera-ture (e.g., Freeman et al., 2001), hydrological change (e.g., Eimers et al., 2008a, 2008b), recovery from acidification (e.g., Monteith et al., 2007), nitrogen deposition (Findlay, 2005), and land use/land cover change (e.g., Aitkenhead-Peterson et al., 2009). However, many of these studies were conducted on large, fresh-water bodies such as lakes and rivers of mid to high latitudes,

Abbreviations: DOC, dissolved organic carbon; HTC, high-temperature combustion; PS, persulfate method; SS, Sen Slope; WY, water year.

N.K. Singh, W.M. Reyes, R. Bhattacharya, and R.E. Emanuel, Dep. of Forestry and Environmental Resources, North Carolina State Univ., Raleigh, NC 27695; E.S. Bernhardt, Dep. of Biology and Nicholas School of the Environment, Duke Univ., Durham, NC 27708; J.L. Meyer, Odum School of Ecology, Univ. of Georgia, Athens, GA 30602; J.D. Knoepp, Coweeta Hydrologic Laboratory, USDA Forest Service-Southern Research Station, Otto, NC 28763. Assigned to Associate Editor Marianne Bechmann.

Copyright © American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America. 5585 Guilford Rd., Madison, WI 53711 USA. All rights reserved. J. Environ. Qual. doi:10.2134/jeq2015.10.0537 Supplemental material is available online for this article.Received 23 Oct. 2015. Accepted 21 Mar. 2016. *Corresponding author (nksingh01@gmail.com).

Journal of Environmental QualityLAndscApe And WAtershed processes

technIcAL reports

core Ideas

• Long-term records show a steady decline in stream DOC fol-lowed by a steady increase.• Declining runoff and changes in storm timing and intensity point to hydro-climatological drivers.• Neither temperature nor atmospheric deposition explained bi-directionality of trends.

Published May 13, 2016

Journal of Environmental Quality

tabl

e 1.

sum

mar

y of

rece

nt s

tudi

es re

port

ing

long

-ter

m (>

10 y

r) c

hang

es in

sur

face

wat

er d

isso

lved

org

anic

car

bon.

Loca

tion

tim

e pe

riod

Leng

th o

f do

c†

reco

rdsa

mpl

ing

freq

uenc

yM

ean

do

cd

oc

tren

d re

port

ed/

site

‡ch

ange

repo

rted

refe

renc

es

yrm

g L-

1

Uni

ted

Stat

es19

88–2

012

25bi

wee

kly

1.45

(-)

0.04

mg

L-1 y

r–1th

is s

tudy

(+)

0.09

mg

L-1 y

r–1

Uni

ted

King

dom

1977

–200

226

wee

kly–

mon

thly

NA

§(+

) 216

site

s0.

06 m

g L-

1 yr–1

Wor

rall

and

Burt

, 200

7(-

) 55

site

s0.

075

mg

L-1 y

r–1

(0) 4

4 si

tes

Uni

ted

Stat

es19

88–2

008

15w

eekl

y0.

7(-

)0.

09 m

g L-

1 (19

89–1

997)

Lutz

et a

l., 2

012

(+)

0.10

66 m

g L-

1 (200

3–20

08)

Cana

da19

80–2

001

221.

3–1.

9 pe

r wee

k3.

4–10

.6(+

)0.

05–0

.15

mg

L-1 y

r–1Ei

mer

s et

al.,

200

8a

Uni

ted

King

dom

1993

–200

210

wee

kly

NA

(+)

0.14

mg

L-1 y

r–1 (s

easo

nal r

ate

of c

hang

e)Cl

ark

et a

l., 2

005

Uni

ted

King

dom

1961

–200

28–

42da

ily–y

early

NA

(+)/

153

site

s, (0

) 45

site

s0.

17 m

g L-

1 yr–1

Wor

rall

et a

l., 2

004b

Finl

and

1990

–200

314

NA

NA

(+)

0.19

–0.2

1 m

g L-

1 yr–1

Futt

er e

t al.,

200

8Sw

eden

1989

–200

811

–20

wee

kly–

mon

thly

7.2–

18.7

(+) 2

site

s, (0

) 2 s

ites

NA

Win

terd

ahl e

t al.,

201

1Cz

ech

Repu

blic

1993

–200

715

wee

kly

18.8

–20.

2(+

)0.

42–0

.43

mg

L-1 y

r–1H

rusk

a et

al.,

200

9U

nite

d Ki

ngdo

m19

88–2

002

15m

onth

ly–q

uart

erly

NA

(+)

0.06

–0.5

1 m

g L-

1 yr–1

Evan

s et

al.,

200

5G

erm

any

1987

–200

923

NA

NA

(+)

NA

Knor

r, 20

13Fi

nlan

d19

75–2

010

36m

onth

ly12

(0) 4

–61

Räik

e et

al.,

201

2(+

) 64–

679.

2% fo

r the

per

iod

Uni

ted

Stat

es19

92–2

001

10N

AN

A(+

) 4.

7 µm

ol y

r-1

Burn

s et

al.,

200

6(+

) 7.

6 µm

ol y

r-1

Uni

ted

King

dom

1988

–200

013

NA

NA

(+)

65%

for t

he p

erio

dFr

eem

an e

t al.,

200

1Ca

nada

1978

–199

821

5–24

sam

ples

per

yea

r1.

8–5.

1 (+

)ic

e-fr

ee P

AR¶

per

iod,

33.

6%H

udso

n et

al.,

200

3w

inte

r pre

cipi

tatio

n pe

riod,

16.

8%su

mm

er p

reci

pita

tion

perio

d, 9

.1%

Cana

da19

70–1

990

21w

eekl

yN

A(-

)in

flow

ing

stre

ams,

40%

Schi

ndle

r et a

l., 1

997

(-)

outfl

ow o

f lak

es: 1

5–25

% fo

r the

per

iod

Nor

th A

mer

ica,

Eur

ope

1990

–200

415

NA

NA

(+) 3

630.

02–0

.1 m

g L-

1 yr–1

Mon

teith

et a

l., 2

007

(-) 1

39>0

.15

mg

L-1 y

r–1

Swed

en19

93–2

010

18w

eekl

y–bi

wee

kly

15–3

1(+

) C2;

(0) C

4 an

d C7

NA

Oni

et a

l., 2

013

Uni

ted

King

dom

1987

–200

612

–16

wee

kly–

biw

eekl

y2.

19–1

1.59

(+) 4

site

s; (0

) 2 s

ites

0.15

–0.7

9 m

g L-

1 yr–1

Daw

son

et a

l., 2

008

Uni

ted

Stat

es19

96–2

010

2–14

mon

thly

2.0–

10(+

)0.

476

mg

L-1 o

f inc

reas

e/1°

C T

Tian

et a

l., 2

013

Uni

ted

Stat

es19

91–2

005

15w

eekl

y N

AN

AN

ASe

best

yen

et a

l., 2

009

Cana

da19

80–1

992

1255

sam

ples

per

yea

rN

AN

AN

AD

illon

and

Mol

ot, 1

997

Nor

way

1985

–200

319

wee

kly

4.9–

11.6

(+) 3

site

s0.

06–1

3 m

g L-

1 yr–1

de W

it et

al.,

200

7U

nite

d St

ates

1995

–200

915

1–3

sam

ples

per

wee

k1.

19(0

) 10

site

sN

ARü

egg

et a

l., 2

015

† D

isso

lved

org

anic

car

bon.

‡ (+

), (-

), an

d (0

) ind

icat

e in

crea

sing

, dec

reas

ing,

and

no

tren

d, re

spec

tivel

y.

§ N

ot a

vaila

ble.

¶ Ph

otos

ynth

etic

ally

act

ive

radi

atio

n.

Journal of Environmental Quality

many of which are heavily influenced by wetlands and snowmelt-dominated runoff. Fewer studies have investigated long-term (>10 yr) trends of stream DOC and associated potential driv-ers for humid mountainous watersheds (e.g., Lutz et al., 2012). Lutz et al. (2012) were limited in data and used a combination of observed and modeled DOC time series for the analysis.

Here, we analyzed a 25-yr DOC record along with ancillary hydro-climatic and atmospheric deposition data for a forested headwater stream in the Appalachian Mountains of the south-eastern United States. Long-term datasets provided a unique opportunity to understand the possible impacts of climate change on carbon and water cycles in streams that are more sensi-tive to anthropogenic forcing (e.g., Lowe and Likens, 2005). Our objectives were (i) to quantify and evaluate long-term trends of DOC concentration and flux in the stream over the period 1988 through 2012 and (ii) to investigate the potential drivers respon-sible for the observed DOC trends.

Materials and MethodsStudy Site

The study was conducted at Coweeta Hydrologic Laboratory (hereafter Coweeta), a US Forest Service Experimental Forest and a National Science Foundation Long-Term Ecological Research site located in the Southern Appalachian Mountains of North Carolina (35°03¢ N, 83°25¢ W) (Supplemental Fig. S1). The climate is classified as maritime-humid temperate with frequent, short, and small rainfall events distributed year-round (e.g., Swift et al., 1988). Since 1934, Coweeta has been monitor-ing long-term precipitation and temperature at a low-elevation (685 m) climate station (CS01) and has observed annual pre-cipitation of 1802 mm and annual temperature of 12.87°C for the 63-yr period of record (1950–2012).

The study watershed (WS27) has a surface area of 39 ha and elevations ranging from a minimum of 1061 m at the outlet to a maximum of 1455 m at the ridge. The watershed is characteris-tic of the region’s headwaters with a mean channel bed slope of 33%, a mean areal slope of 55%, and soil depths <2 m. Soils in WS27 are predominantly coarse-loamy and belong to chandler soil series. High-elevation streams of Coweeta (including those in WS27) have relatively cool waters (Webster and Golladay, 1984) that are depleted in nutrients and ionic strength (Swank and Waide, 1988). Mean annual precipitation within WS27 is 2363 mm, and mean annual runoff is 1658 mm. Here, we adopt the hydrological definition of runoff as discharge per unit water-shed area.

The watershed is considered a reference watershed for paired watershed studies at Coweeta and has not been harvested since 1929. Two vegetation communities are present within WS27; a northern hardwood community dominated by northern red oak (Quercus rubra L.), yellow birch (Betula alleghaniensis Britton), and sweet birch (Betula lenta L.), and a mixed oak community dominated by Quercus spp. and Carya spp. (Elliott et al., 1999). The bedrock underlying WS27 is metamorphic and comprises biotite, gneiss, metasandstone, and schists (Velbel, 1985). Unfractured bedrock beneath WS27 is highly impermeable, suggesting that runoff measured at the weir accurately repre-sents total water drainage from the study watershed (Swift et al., 1988).

MethodsDissolved organic carbon concentrations were measured in

water samples (n = ~1000) collected every 1 to 2 wk from WS27 between 1988 and 2012. Samples were not collected in 1998 and 2002. Water samples were collected from the stream immediately upstream of the weir pond, filtered using a Gelman A/E glass fiber filter (0.3 µm pore size), and refrigerated until analyzed. Samples were analyzed for DOC using two different methods: (i) an Oceanography International Organic Carbon analyzer using persulfate oxidation (1988–2001) and (ii) a Shimadzu TOC-5000A Total Carbon Analyzer (2001–2012) using high-temperature combustion (HTC). Dissolved organic carbon concentrations obtained from the persulfate method (PS) were converted to DOC concentrations from HTC method using the regression equation HTC = 0.87 PS - 0.07 (R2 = 0.6; p < 0.05), where the 95% confidence intervals ranged from 0.72 to 1.008 for slope and from -0.14 to -0.003 for the intercept. These ranges are comparable to other studies juxtaposing both methods (e.g., Benner and Hedges, 1993). This regression-based approach for converting DOC values from HTC to PS is consistent with other work at Coweeta summarized by Meyer et al. (2014). Detailed comparison of DOC values in samples analyzed using the two methods revealed a small bias, with PS overestimating HTC by an average of 0.13 mg L-1 (Supplementary Material). Overall, the methodological difference is believed to account for approximately 5% of total uncertainty in DOC estimates from freshwater systems (Kaplan, 1992). The difference in methods may affect the overall magnitude of DOC reported in this study but is unlikely to affect the direction of DOC trends at Coweeta.

For each month, the volume-weighted DOC concentration (DOCvw) was calculated as:

1vw

1

DOC

nj jj

njj

C Q

Q=

=

=åå

[1]

where Cj is the DOC concentration of a biweekly sample (mg L-1), and Qj is the instantaneous discharge (L s-1) at the time of sampling. We calculated the standard error (SE) for DOCvw on an annual basis as:

SEns

= [2]

where s is the standard deviation in DOCvw, and n is the sample size. Monthly DOC flux was calculated by multiplying monthly DOCvw by mean monthly discharge for WS27. We assessed DOC patterns on both monthly and water year (WR) bases. We used the WR definition of Meyer et al. (2014), which is 1 November of the base year to 31 October of the following year (e.g., WY 1989 is 1 Nov. 1989 through 31 Oct. 1990).

Discharge for the second-order perennial stream draining WS27 (Hard Luck Creek) is monitored by the US Forest Service using an instrumented 120° v-notch weir. A total of 7 mo of dis-charge data were missing during the 25-yr time series: August through October 2002, November 2003 through January 2004, and January 2013. We computed monthly averages of daily max-imum air temperature using records from Coweeta’s long-term climate station (CS01). We computed monthly precipitation

Journal of Environmental Quality

totals from daily precipitation recorded at a climate station within the watershed (CS77). Annual precipitation and runoff were summed from monthly totals for each water year. The annual runoff ratio was calculated as the ratio of annual runoff to annual precipitation. Before the establishment of CS77 in 1992, we calculated precipitation using a simple regression model: Y = 1.1057×X + 7.022 (R2 = 0.95), where Y is modeled monthly precipitation (mm) for the missing months of CS77, and X is observed monthly precipitation (mm) at CS17, a climate station located in a nearby watershed. Discharge data, air temperature data, and precipitation data were acquired from the Coweeta data portal (Coweeta-Long Term Ecological Research, 2013).

Trend analysis on all datasets was performed at annual and monthly time scales using the Mann–Kendall test (Kendall, 1975; Mann, 1945), and the rates of change (i.e., slopes) of sig-nificant trends were estimated using the Sen Slope (SS) method (Sen, 1968). The Mann–Kendall test is a robust statistical test due to its insensitivity to the shape of distribution and outliers of the time series. Where any gaps in time series variables occurred, we used linear least squares regression to estimate the trend slope. The Kendall correlation coefficient (t) was also used to quantify the relationships between potential drivers and DOC for the study period. We performed Wilcoxon rank sum tests (Wilcoxon, 1945) to evaluate differences between median values of two time series.

We performed a breakpoint analysis to identify the period of time in which trends in stream DOC changed direction. The breakpoint analysis involved computing correlation statistics (one-sided t and a corresponding p value) for the DOC time series both before and after a potential breakpoint year. We sys-tematically tested each year from 1989 to 2011 to determine whether it met the following criteria for a breakpoint in the time series. We considered a breakpoint in the time series to occur near the maximum sum of |t| values before and after the potential breakpoint year where both p values were significant. We allowed for uncertainty in the analysis by computing summed |t| to two significant digits only, allowing us to identify a range of consecu-tive years during which the trend in DOC changed direction. A Bonferroni adjustment for testing two simultaneous, one-sided correlations yielded a critical p value of 0.05 (0.10/2). Breakpoint analyses were also applied to annual and monthly time series of maximum air temperature, precipitation, and runoff.

ResultsTemporal Trends in Stream Dissolved Organic Carbon Concentrations and Fluxes

The breakpoint analysis identified a breakpoint in stream DOC trends between 1997 and 2001, with a declining trend in stream DOCvw (t = -0.66; p < 0.05) with a linear slope of -0.04 mg L-1 yr -1 from 1988 to 2001 followed by a rising trend in stream DOCvw (t = 0.72; p < 0.05) with a linear slope of 0.09 mg L-1 yr -1 from 1997 to 2012 (Fig. 1). Using the linear rates of decline and increase, we estimated a decline of 34% (1988–2001) followed by an increase of 141% (1997–2012) in annual mean DOCvw (Fig. 1). The lowest annual DOCvw was observed in 2000 (0.88 mg L-1) during the breakpoint-defined interval (1997–2001). We separated all data sets into “pre” (1988–2001) and “post” (1997–2012) intervals to investigate the drivers

influencing rising and declining trends in DOCvw during each of these intervals.

To further understand trends in DOCvw, we compared the temporal patterns of DOCvw at a monthly scale for the 1988–2001 and 1997–2012 intervals. For the later interval, significant increases in monthly mean DOCvw were observed for the grow-ing season months of April and June and the early fall months of September and October compared with the earlier interval (Wilcoxon, p < 0.05) (Fig. 2a–d). A bidirectional trend in mean monthly DOCvw was only observed for the month of September

Fig. 1. trends for annual mean volume-weighted dissolved organic carbon (docvw) for the study period (1988–2012) in the study water-shed (Ws27) at coweeta hydrologic Laboratory, otto, nc. solid red lines show linear trends, dashed line shows the breakpoint interval (1997–2001), and black error bars in the docvw represent se based on the number of samples collected during the study period.

Fig. 2. Monthly trends of volume-weighted dissolved organic carbon concentration (docvw) for the months of April (a), June (b), september (c), and october (d) during the study period (1988–2012). dashed lines and double arrows show the duration of lengths for which trends were significant (p < 0.05).

Journal of Environmental Quality

(Fig. 2c). For September, we observed a declining trend in DOCvw (1988–2001: t = -0.60; p < 0.05; linear slope = -0.09 mg L-1), followed by a rising trend in DOCvw (1997–2012: t = 0.67; p < 0.05; linear slope = 0.17 mg L-1) that was consis-tent in shape to the declining and rising behavior of annual mean DOCvw during the study period (Fig. 1).

Annual DOC flux declined from 1988 through 2001 (t = -0.60; p < 0.05) with a linear slope of -1.49 kg-1 ha-1 yr-1. The declining trend was followed by rising DOC from 1997 through 2012 (t = 0.75; p < 0.05) with a linear slope of 1.81 kg-1 ha-1 yr-1 (Fig. 3). Based on linear slopes, we computed a decline of 56% between 1988 and 2001, followed by an increase of 165% in annual DOC flux between 1997 and 2012 (Fig. 3). We also compared seasonal and monthly trends in DOC flux for the 1988–2001 and 1997–2012 intervals and did not detect any sig-nificant changes in mean monthly fluxes (Wilcoxon, p > 0.05) between both intervals.

Hydro-Climatological TrendsWe identified a significant increase in annual maximum air

temperatures of 1.25°C over the study period (t = 0.23; p < 0.05) with a SS of 0.05°C yr-1 (Supplemental Fig. S2). Summer months in particular became warmer, with increasing trends in monthly maximum air temperatures observed for June (t = 0.45; p < 0.05) and August (t = 0.34; p < 0.05). No sig-nificant breakpoints were identified for annual maximum air temperatures.

We observed a decline in runoff of 22 mm yr-1 (t = -0.28; p < 0.1) and runoff ratio (t = -0.32; p < 0.05), but no significant trends were detected for annual precipitation (p > 0.05) during the study period (Fig. 4a–c). The declining trend in annual runoff suggested a decline of 550 mm (SS, -22 mm yr-1) in annual runoff from 1988 to 2012, which was further supported by declining runoff ratios (Fig. 4c). No significant breakpoints were identified for annual runoff, annual precipitation, and annual runoff ratio.

We found no significant differences in monthly precipitation or monthly runoff between the two time intervals (Wilcoxon, p > 0.05). During the 1988–2001 interval, we identified significant declining trends in monthly runoff for the months of August (t = -0.39; p < 0.05), September (t = -0.45; p < 0.05), and October (t = -0.45; p < 0.05) (Fig. 5). During the 1997–2012 interval, April precipitation increased significantly (t = 0.43; p < 0.05) (Fig. 5a). We observed an increase in extreme precipitation

Fig. 3. trends for annual dissolved organic carbon (doc) flux for the study period (1988–2012) in Ws27 at coweeta hydrologic Laboratory, nc.

Fig. 4. trends in annual precipitation (a), annual runoff (b), and annual runoff ratio (c), for the study watershed (Ws27) from 1988 to 2012. solid red lines show the sen slope (ss) estimates, and red dashed lines represent 95% confidence interval for the ss estimates.

Fig. 5. trends for monthly precipitation and monthly runoff for the months of April (a, b), June (c, d), september (e, f), and october (g, h) during the study period. symbols represent type of data, precipita-tion (open circle), and runoff (filled circle). dashed lines and double arrows show the duration of lengths for which trends were significant (p < 0.05).

Journal of Environmental Quality

(storms exceeding 75th percentile; i.e., 16.5 mm) for the month of September (Fig. 5e), which was previously documented in an analysis by Laseter et al. (2012) at a lower-elevation climate sta-tion within the Coweeta basin (Fig. 6a, b). Laseter et al. (2012) found an increase in the number of high-intensity storms (75th and 85th percentile) from 2000 onward (Fig. 6b), which they attributed to tropical storm activity.

Stream Dissolved Organic Carbon and DischargeInstantaneous discharge and DOC concentrations at the

time of sampling covered a wide range of flows under base flow and storm flow conditions (Supplemental Fig. S3); how-ever, we found no significant correlation between sampled DOC concentration (not volume-weighted) and instan-taneous discharge for the study period. When we exam-ined samples during two intervals separately, we found that stream DOC concentrations during 1997 to 2012 were 35% higher on average than DOC concentrations during 1988 to 2001 despite the approximately same mean instantaneous discharge of 20 L s-1 (Supplemental Fig. S3a). During the month of September, we found that mean instantaneous discharge increased by 32% from the 1988–2001 interval to the 1997–2012 interval and that mean instantaneous stream DOC increased by 55% from the earlier interval to the later interval. Mean instantaneous DOC also increased during the months of April, June, and October from the earlier interval to the later interval, but the changes were less than that of September (Supplemental Fig. S3b–e).

We found a strong correlation between runoff during the growing season months and annual mean DOCvw (t = 0.46; p < 0.05) and annual DOC fluxes (t = 0.53; p < 0.05), respectively, during the 1988–2001 interval. For the same interval, a stronger correlation (t = 0.53; p < 0.05) was observed between the runoff from the month of September and annual mean DOCvw. The 1997–2012 interval exhibited no similar relationships.

DiscussionRising Air Temperature

Generally, higher air temperatures increase microbial activity, leading to increased DOC production in aquatic ecosystems (e.g., Evans et al., 2005; Freeman et al., 2001). In forested headwater streams, temperature exerts a strong control over the seasonal variability of DOC formation and consumption (e.g., Cronan and Aiken, 1985; Mulholland and Hill, 1997). High air temperature during the growing season increases microbial decomposition in forest soils, lead-ing to increased DOC formation (e.g., Liechty et al., 1995). At Coweeta, maximum annual air temperature increased by 1.25°C during the study period (Supplemental Fig. S2). These annual trends were accompanied by significant increases in mean and maximum air temperatures during the growing season month of June. Increasing June temperatures may have contributed to significantly higher DOCvw in June for the later interval (1997–2012).

Air temperature can also influence the transport of DOC from uplands to the stream indirectly by altering the water bal-ance of the watershed. Specifically, rising air temperatures can cause increased atmospheric demand for water vapor, which may

increase evapotranspiration when soil water supplies are suffi-cient to meet vegetation requirements (e.g., Emanuel et al., 2007, 2010; Wang and Dickinson, 2012). These influences of tempera-ture on DOC can elicit complex behavior by counteracting each other or by operating at different temporal scales.

Hydrological ChangesSeveral field studies highlight relationships between runoff

and DOC concentrations and fluxes in aquatic ecosystems (e.g., McDowell and Likens, 1988; McGlynn and McDonnell, 2003; Tate and Meyer, 1983). Significant declines in annual runoff (-22 mm yr-1) and runoff ratios at Coweeta (Fig. 4c) suggest drier conditions, potentially linked to rising air temperatures (Supplemental Fig. S2). In earlier work at Coweeta, Tate and Meyer (1983) revealed that the biological activities respon-sible for DOC formation and consumption were most active during the growing season due to favorable conditions for in-stream biological activity, litter decomposition, and leaching in soils. During the 1988–2001 interval, we observed a decline in monthly runoff for some of these growing season months (Fig. 5) and detected significant correlations between grow-ing season runoff and annual mean DOCvw and DOC fluxes. September exhibited much stronger correlation between the monthly runoff and annual mean DOCvw for the same dura-tion. These correlations were indicative of strong seasonal con-trol of runoff on annual patterns of mean DOCvw (e.g., Eimers et al., 2008b). Declining runoff during the growing season (Fig. 5) might be associated with reduced transport of more reactive fraction of DOC from uplands to the stream. Furthermore, drier conditions can result in a lack of hydrologic connectivity within the watershed, reducing the capacity of these landscapes to contribute new DOC to the stream (Pacific et al., 2010).

The role of runoff in mobilizing DOC has been widely recognized in various ecosystems (e.g., Eimers et al., 2008b;

Fig. 6. total number of storms observed in the month of september (a) and percentage (%) of storms exceeding the 75th percentile (16.5 mm) (b) during 1988 to 2008. data provided by Laseter et al. (2012). The gap represents no storms exceeding the 75th percentile for that particular year.

Journal of Environmental Quality

Lutz et al., 2012; Oni et al., 2013). Eimers et al. (2008b) reported that seasonal runoff changes had a direct impact on modulating the annual DOC dynamics for Canadian wetland–dominated watersheds. Oni et al. (2013) reported that runoff explained significant variability in DOC for both upland and wetland-dominated watersheds of Sweden. Our study joins these other studies to further support the direct linkages between runoff and DOCvw, which could potentially explain the declining annual patterns of DOCvw observed during the 1988–2001 interval.

During the 1997–2012 interval, monthly DOCvw and monthly precipitation increased for the months of April (Fig. 2a, 5a) and September (Fig. 2c), along with an increase in high-intensity storms for the month of September (Fig. 6b) (Laseter et al., 2012). Increasing precipitation, together with warmer conditions in early growing season months (e.g., April), may enhance biological productivity, leading to high stream DOC during that month. Earlier work at Coweeta suggests that grow-ing season storms can flush high concentrations of DOC from soils (Meyer and Tate, 1983). The timing of September storms was crucial because most of the earlier growing season months were relatively dry.

Leaf fall at Coweeta generally begins in late September or early October. Experimental studies in other Coweeta streams have demonstrated a direct link between DOC concentration and the amount of decomposing leaves in the stream (Meyer et al., 1998). September storms may have provided high antecedent conditions for October to mobilize in-stream decomposing fresh leaf litter to the catchment outlet. To summarize, during the later interval annual runoff declined through time but DOCvw continued to rise, likely due to changes in precipitation patterns during fall months.

A number of field studies have analyzed the relationship between DOC and storms, focusing on event-based DOC flush-ing (Dhillon and Inamdar, 2013; Hinton et al., 1997; McDowell and Likens, 1988). McDowell and Likens (1988) attributed the consistent flushing of high DOC concentrations during mul-tiple storms to freshly fallen leaves in a forested stream of New Hampshire. A few studies have also suggested that the strength of the flushing signal is even stronger after prolonged dry periods (Inamdar et al., 2008; Worrall et al., 2004a).

Other Potential DriversIn general, other drivers that may influence the long-term

patterns of stream DOC include land use/land cover changes (Aitkenhead-Peterson et al., 2009), recovery from acidification (Monteith et al., 2007), and nitrogen deposition (Findlay, 2005). We evaluated some of these drivers for WS27 but were unable to implicate any of them as controls on observed long-term patterns of stream DOC. In particular, WS27 is a reference watershed for the Coweeta basin, so no land use/land cover changes occurred at our study site during the study period.

Studies show an inverse relationship between atmospheric [SO4

2-] and stream DOC (e.g., Evans et al., 2005; Monteith et al., 2007). These studies suggest that increasing [SO4

2-] (i.e., acidity) suppresses the solubility of organic compounds in soil solution, restraining the formation and consumption of DOC in soils and streams (e.g., Clark et al., 2005; Hruska et al., 2009; Thurman, 1985). Annual mean volume-weighted

[SO42-] from the National Atmospheric Deposition Program

wet-deposition site at Coweeta (National Atmospheric Data Program, 2015) declined at an average rate of -0.02 mg L-1 yr-1 (t = -0.52; p < 0.001) during the study period (Supplemental Fig. S4a). These data suggest that the water-shed has been recovering from acidification since 1988 or earlier. Steadily declining [SO4

2-] might be expected to cause steadily rising stream DOC (Erlandsson et al., 2011). However, no such pattern was observed in stream DOC for WS27 (Fig. 1). Although a lag in recovery of DOC from acid-ification might be expected, such a long lag (>10 yr) seems highly unlikely given observations of recovery in other sites (e.g., Evans et al., 2005). Recovery from acidification may provide part of an explanation for long-term trends in DOC at Coweeta, but, given the bidirectional DOC trends, acidifi-cation recovery alone is unlikely to explain our observations.

Annual mean concentrations of [NO32-] of wet deposition

from the same National Atmospheric Deposition Program site declined gradually (t = -0.48; p < 0.001) (Supplemental Fig. S4b) for the study period. Furthermore, no significant trend was detected for annual mean concentrations of [NH4

+] at the site during the study year period (Supplemental Fig. S4c). These observations suggest a minor decline in inorganic nitro-gen deposition for the watershed during the study period. Laboratory and field experiments provide equivocal evidence of a direct (e.g., Pregitzer et al., 2004), an inverse (e.g., Park et al., 2002), or no relationship (e.g., Magill et al., 2000) between DOC and nitrogen deposition. However, a long-term sur-face water DOC export study in the Hudson River suggested that high rates of nitrogen deposition could accelerate micro-bial decomposition, leading to the release of humic substance and contributing to the formation of DOC (Findlay, 2005); however, given that nitrogen deposition at Coweeta declined during the study period, it is unlikely that this trend is linked to observed patterns in stream DOC.

Among the potential drivers of long-term stream DOC discussed in this paper, hydro-climatological factors emerge as the most likely source of the bidirectional trends observed in WS27. Other studies have reported bidirectional or periodic trends in long-term DOC. For example, Hejzlar et al. (2003) found bidirectional trends in dissolved organic matter in an upland stream, which was attributed to variability in runoff, air temperature, and atmospheric deposition. Recently, Lutz et al. (2012) attributed the observed rising and declining patterns in stream DOC to changes in runoff and runoff source areas for a forested watershed in Tennessee. Together with our results, these studies demonstrate the potential for hydro-climatolog-ical factors to elicit or contribute to nonmonotonic trends in long-term stream DOC.

Global Change Implications and Limitations of AnalysisOur results provide evidence of strong linkages between

long-term DOC trends and hydro-climatological variabil-ity. Annual maximum air temperatures are increasing at Coweeta (Supplemental Fig. S2) (Ford et al., 2011; Laseter et al., 2012), a trend that is consistent with global model pre-dictions for the southeastern United States (Stocker et al., 2013). Based on evidence of positive feedbacks between air

Journal of Environmental Quality

temperature and soil microbial activity (e.g., Davidson et al., 2000; Heimann and Reichstein, 2008), continued tempera-ture increases may be accompanied by increases in terrestrial DOC production, which may eventually contribute to rising concentrations of stream DOC. Rising temperatures can also alter hydrologic fluxes and associated hydrologic connectivity within a watershed, resulting in declining transport of DOC from uplands to streams. Given this potential for reduced connectivity, export of DOC could become more episodic in the future.

General circulation models predict that precipitation could become more extreme with increases in short-duration, high-intensity storms for much of the United States, including the Southeast (Stocker et al., 2013). Intense storms can signifi-cantly reduce the residence time of DOC within watersheds and flush a larger proportion of DOC from soils into stream channels (Dhillon and Inamdar, 2013; Sebestyen et al., 2009). Studies have predicted increasing DOC fluxes with increases in annual precipitation and annual runoff for future climate projections (Sebestyen et al., 2009); however, our results show that DOC export can increase even if annual precipitation magnitude remains unchanged or if annual runoff declines. This work suggests that over the long term, changes in storm characteristics (e.g., frequency, intensity, duration) may affect global DOC dynamics, carbon budgets, and overall ecosys-tem responses (e.g., Knapp et al., 2008) more than changes in total precipitation. In general, changing climatic patterns can severely affect the availability, formation, consumption, and transport of DOC in mountainous watersheds. Changing DOC patterns can have impacts on the closely coupled water-shed nitrogen cycle (Bernhardt and Likens, 2002) and overall stream chemistry. Due to the close coupling between head-water streams and their surrounding landscapes (Gomi et al., 2002), these changes may influence the healthy functioning of aquatic and terrestrial ecosystems.

Our study, along with other long-term datasets published in recent years (Argerich et al., 2013; Meyer et al., 2014), high-lights the potential complexity of long-term trends in stream biogeochemistry. In particular, these studies illustrate the abil-ity of long-term trends to vary in magnitude or even direction depending on the length of record or the period of record chosen for analysis. Collectively, these studies suggest caution while interpreting trends from a given time series and call for building and maintaining long-term datasets to understand more fully long-term ecological and hydrological responses of watersheds and their streams to environmental change. It stands to reason that such multi-decadal datasets will only increase in value as benchmarks for scientific research and for management and policy decisions.

AcknowledgmentsThe study was funded in part by the Department of Forestry and Environmental Resources at North Carolina State University. The US Department of Interior Southeast Climate Science Center supported N.K. Singh through a Global Change Fellowship. Additional support was provided by the USDA Forest Service Southern Research Station and by NSF awards DEB-0218001 and DEB-0823293 to the Coweeta Long-Term Ecological Research program.

ReferencesAitkenhead-Peterson, J.A., W. McDowell, and J.C. Neff. 2003. Sources, produc-

tion, and regulation of allochthonous dissolved organic matter inputs to surface waters. In: S. Findlay and R.L. Sinsabaugh, editors, Aquatic eco-systems: Interactivity of dissolved organic matter. Academic Press, Amster-dam, The Netherlands. p. 25–70.

Aitkenhead-Peterson, J.A., M.K. Steele, N. Nahar, and K. Santhy. 2009. Dis-solved organic carbon and nitrogen in urban and rural watersheds of south-central Texas: Land use and land management influences. Biogeo-chemistry 96:119–129. doi:10.1007/s10533-009-9348-2

Ardón, M., L.A. Stallcup, and C.M. Pringle. 2006. Does leaf quality mediate the stimulation of leaf breakdown by phosphorus in Neotropical streams? Freshwater Biol. 51:618–633. doi:10.1111/j.1365-2427.2006.01515.x

Argerich, A., S.L. Johnson, S.D. Sebestyen, C.C. Rhoades, E. Greathouse, J.D. Knoepp, M.B. Adams, G.E. Likens, J.L. Campbell, W.H. McDowell, and F.N. Scatena. 2013. Trends in stream nitrogen concentrations for forest-ed reference catchments across the USA. Environ. Res. Lett. 8:014039. doi:10.1088/1748-9326/8/1/014039

Beckman, N.D., and E. Wohl. 2014. Carbon storage in mountainous headwa-ter streams: The role of old-growth forest and logjams. Water Resour. Res. 50:2376–2393. doi:10.1002/2013WR014167

Bernhardt, E.S., and G.E. Likens. 2002. Dissolved organic carbon enrichment alters nitrogen dynamics in a forest stream. Ecology 83:1689–1700. doi:10.1890/0012-9658(2002)083[1689:DOCEAN]2.0.CO;2

Benner, R., and J.I. Hedges. 1993. A test of the accuracy of freshwa-ter DOC measurements by high-temperature catalytic oxidation and UV-promoted persulfate oxidation. Mar. Chem. 41:161–165. doi:10.1016/0304-4203(93)90114-4

Burns, D.A., M.R. McHale, C.T. Driscoll, and K.M. Roy. 2006. Response of sur-face water chemistry to reduced levels of acid precipitation: Comparison of trends in two regions of New York, USA. Hydrol. Processes 20:1611–1627. doi:10.1002/hyp.5961

Butman, D., and P. Raymond. 2011. Significant efflux of carbon dioxide from streams and rivers in the United States. Nat. Geosci. 4:839–842. doi:10.1038/ngeo1294

Clark, J.M., P.J. Chapman, J.K. Adamson, and S.J. Lane. 2005. Influ-ence of drought-induced acidification on the mobility of dissolved organic carbon in peat soils. Glob. Change Biol. 11:791–809. doi:10.1111/j.1365-2486.2005.00937.x

Coweeta-Long Term Ecological Research. 2013. Hydrologic and climatic datasets. https://coweeta.uga.edu/dbpublic/data_catalog.asp (accessed 1 Feb. 2013).

Cronan, C.S., and G.R. Aiken. 1985. Chemistry and transport of sol-uble humic substances in forested watersheds of the Adirondack Park, New York. Geochim. Cosmochim. Acta 49:1697–1705. doi:10.1016/0016-7037(85)90140-1

Curtis, P.J., and D.W. Schindler. 1997. Hydrologic control of dissolved organic matter concentration and quality in low-order Precambrian Shield lakes. Biogeochemistry 36:125–138. doi:10.1023/A:1005787913638

Dahm, C.N. 1981. Pathways and mechanisms for removal of dissoIved organic carbon from leaf leachate in streams. Can. J. Fish. Aquat. Sci. 38:68–76. doi:10.1139/f81-009

Davidson, E.A., S.E. Trumbore, and R. Amundson. 2000. Biogeochemis-try: Soil warming and organic carbon content. Nature 408:789–790. doi:10.1038/35048672

Dawson, J.J., C. Soulsby, D. Tetzlaff, M. Hrachowitz, S.M. Dunn, and I.A. Malcom. 2008. Influence of hydrology and seasonality on DOC exports from three contrasting upland catchments. Biogeochemistry 90:93–113. doi:10.1007/s10533-008-9234-3

de Wit, H.A., J. Mulder, A. Hindar, and L. Hole. 2007. Long-term increase in dissolved organic carbon in streamwaters in Norway is response to reduced acid deposition. Environ. Sci. Technol. 41:7706–7713. doi:10.1021/es070557f

Dhillon, G.S., and S. Inamdar. 2013. Extreme storms and changes in particulate and dissolved organic carbon in runoff: Entering uncharted waters? Geo-phys. Res. Lett. 40:1322–1327. doi:10.1002/grl.50306

Dillon, P.J., and L.A. Molot. 1997. Effect of landscape form on export of dis-solved organic carbon, iron, and phosphorus from forested stream catch-ments. Water Resour. Res. 33:2591–2600. doi:10.1029/97WR01921

Eimers, M.C., S.A. Watmough, and J.M. Buttle. 2008a. Long-term trends in dis-solved organic carbon concentration: A cautionary note. Biogeochemistry 87:71–81. doi:10.1007/s10533-007-9168-1

Eimers, M.C., J.M. Buttle, and S.A. Watmough. 2008b. Influence of seasonal changes in runoff and extreme events on dissolved organic carbon trends in wetland- and upland-draining streams. Can. J. Fish. Aquat. Sci. 65:796–808. doi:10.1139/f07-194

Journal of Environmental Quality

Elliott, K.J., J.M. Vose, W.T. Swank, and P.V. Bolstad. 1999. Long-term patterns in vegetation-site relationships in a southern Appalachian Forest. J. Torrey Bot. Soc. 126:320–334. doi:10.2307/2997316

Emanuel, R.E., P. D’Odorico, and H.E. Epstein. 2007. Evidence of optimal water use by vegetation across a range of North American ecosystems. Geophys. Res. Lett. 34:L07401.

Emanuel, R.E., H.E. Epstein, B.L. McGlynn, D.L. Welsch, D.J. Muth, and P. D’Odorico. 2010. Spatial and temporal controls on watershed ecohydrol-ogy in the northern Rocky Mountains. Water Resour. Res. 46:W11553.

Erlandsson, M., N. Cory, J. Fölster, S. Köhler, H. Laudon, G.A. Weyhenmeyer, and K. Bishop. 2011. Increasing dissolved organic carbon redefines the ex-tent of surface water acidification and helps resolve a classic controversy. Bioscience 61:614–618. doi:10.1525/bio.2011.61.8.7

Evans, C.D., D.T. Monteith, and D.M. Cooper. 2005. Long-term increases in surface water dissolved organic carbon: Observations, possible causes and environmental impacts. Environ. Pollut. 137:55–71. doi:10.1016/j.envpol.2004.12.031

Findlay, S.E.G. 2005. Increased carbon transport in the Hudson River: Unex-pected consequence of nitrogen deposition? Front. Ecol. Environ 3:133–137. doi:10.1890/1540-9295(2005)003[0133:ICTITH]2.0.CO;2

Ford, C.R., S.H. Laseter, W.T. Swank, and J.M. Vose. 2011. Can forest manage-ment be used to sustain water-based ecosystem services in the face of cli-mate change? Ecol. Appl. 21:2049–2067. doi:10.1890/10-2246.1

Freeman, C., C.D. Evans, D.T. Monteith, B. Reynolds, and N. Fenner. 2001. Export of organic carbon from peat soils. Nature 412:785. doi:10.1038/35090628

Futter, M.N., M. Starr, M. Forsius, and M. Holmberg. 2008. Modeling the effects of climate on long-term patterns of dissolved organic carbon concentra-tions in the surface waters of a boreal catchment. Hydrol. Earth Syst. Sci. 12:437–447. doi:10.5194/hess-12-437-2008

Gomi, T., R.C. Sidle, and J.S. Richardson. 2002. Understanding process-es and downstream linkages of headwater systems headwaters differ from downstream reaches by their close coupling to hillslope pro-cesses, more temporal and spatial variation, and their need for dif-ferent means of protection from land use. Bioscience 52:905–916. doi:10.1641/0006-3568(2002)052[0905:UPADLO]2.0.CO;2

Hall, R.O., Jr., and J.L. Meyer. 1998. The trophic significance of bacte-ria in a detritus-based stream food web. Ecology 79:1995–2012. doi:10.1890/0012-9658(1998)079[1995:TTSOBI]2.0.CO;2

Heimann, M., and M. Reichstein. 2008. Terrestrial ecosystem carbon dynamics and climate feedbacks. Nature 451:289–292. doi:10.1038/nature06591

Hejzlar, J., M. Dubrovsky, J. Buchtele, and M. Ruzicka. 2003. The apparent and potential effects of climate change on the inferred concentration of dis-solved organic matter in a temperate stream. Sci. Total Environ. 310:143–152. doi:10.1016/S0048-9697(02)00634-4

Hinton, M.J., S.L. Schiff, and M.C. English. 1997. The significance of storms for the concentration and export of dissolved organic carbon from two Precambrian Shield catchments. Biogeochemistry 36:67–88. doi:10.1023/A:1005779711821

Hruska, J., P. Kram, W.H. Mcdowell, and F. Oulehle. 2009. Increased dissolved organic carbon (DOC) in Central European streams is driven by reduc-tions in ionic strength rather than climate change or decreasing acidity. Environ. Sci. Technol. 43:4320–4326. doi:10.1021/es803645w

Hudson, J.J., P.J. Dillon, and K.M. Somers. 2003. Long-term patterns in dis-solved organic carbon in boreal lakes: The role of incident radiation, pre-cipitation, air temperature, southern oscillation and acid deposition. Hy-drol. Earth Syst. Sci. 7:390–398. doi:10.5194/hess-7-390-2003

Inamdar, S., J. Rupp, and M. Mitchell. 2008. Differences in dissolved organic car-bon and nitrogen responses to storm-event and ground-water conditions in a forested, glaciated watershed in western New York. J. Am. Water Re-sour. Assoc. 44:1458–1473. doi:10.1111/j.1752-1688.2008.00251.x

Kaplan, L.A. 1992. Comparison of high‐temperature and persulfate oxidation methods for determination of dissolved organic carbon in freshwaters. Limnol. Oceanogr. 37:1119–1125. doi:10.4319/lo.1992.37.5.1119

Kendall, M.G. 1975. Rank correlation methods, 4th ed. Charles Griffin, London.Knapp, A.K., C. Beier, D.D. Briske, A.T. Classen, Y. Luo, M. Reichstein, M.D.

Smith, S.D. Smith, J.E. Bell, P.A. Fay, and J.L. Heisler. 2008. Consequences of more extreme precipitation regimes for terrestrial ecosystems. Biosci-ence 58:811–821. doi:10.1641/B580908

Knorr, K.H. 2013. DOC-dynamics in a small headwater catchment as driven by redox fluctuations and hydrological flow paths: Are DOC exports me-diated by iron reduction/oxidation cycles? Biogeosciences 10:891–904. doi:10.5194/bg-10-891-2013

Laseter, S.H., C.R. Ford, J.V. Vose, and L.W. Swift. 2012. Long-term temperature and precipitation trends at Coweeta Hydrologic Laboratory, Otto, North Carolina, USA. Hydrol. Res. 43:890–901. doi:10.2166/nh.2012.067

Leopold, L.B., M.G. Wolman, and J.P. Miller. 1964. Fluvial processes in geomor-phology. Dover, New York.

Liechty, H.O., E. Kuuseoks, and G.D. Mroz. 1995. Dissolved organic car-bon in northern hardwood stands with differing acidic inputs and temperature regimes. J. Environ. Qual. 24:927–933. doi:10.2134/jeq1995.00472425002400050021x

Lowe, H.W., and G.E. Likens. 2005. Moving headwater streams to the head of the class. BioScience 55:196–197.

Lutz, B.D., P.J. Mulholland, and E.S. Bernhardt. 2012. Long-term data reveal pat-terns and controls on stream water chemistry in a forested stream: Walker Branch, Tennessee. Ecol. Monogr. 82:367–387. doi:10.1890/11-1129.1

Magill, A.H., J.D. Aber, G.M. Berntson, W.H. McDowell, K.J. Nadelhoffer, J.M. Melillo, and J. Steudler. 2000. Long-term nitrogen additions and nitrogen saturation in two temperate forests. Ecosystems 3:238–253. doi:10.1007/s100210000023

Mann, H.B. 1945. Non-parametric test against trend. Econometrica 13:245–259. doi:10.2307/1907187

McDowell, W.H., and G.E. Likens. 1988. Origin, composition, and flux of dissolved organic carbon in the Hubbard Brook Valley. Ecol. Monogr. 58:177–195. doi:10.2307/2937024

McGlynn, B.L., and J.J. McDonnell. 2003. Role of discrete landscape units in controlling catchment dissolved organic carbon dynamics. Water Resour. Res. 39:1090. doi:10.1029/2002WR001525

McKnight, D.M. 1981. Chemical and biological processes controlling the re-sponse of a freshwater ecosystem: A field study of the CuSO4 treatment of Mill Pond Reservoir, Burlington, MA. Limnol. Oceanogr. 26:518–531.

Meyer, J.L., and C.M. Tate. 1983. The effects of watershed disturbance on dissolved organic carbon dynamics of a stream. Ecology 64:33–44. doi:10.2307/1937326

Meyer, J.L., J.B. Wallace, and S.L. Eggert. 1998. Leaf litter as a source of dis-solved organic carbon in streams. Ecosystems 1:240–249. doi:10.1007/s100219900019

Meyer, J.L., J.R. Webster, J.D. Knoepp, and E.F. Benfield. 2014. Dynamics of dis-solved organic carbon in a stream during a quarter century of forest suc-cession. In: W.T. Swank, and J.R. Webster, editors, Long-term response of a forest watershed ecosystem: Clearcutting in the southern Appalachians. Oxford Univ. Press, New York. p. 102–117.

Monteith, D.T., J.L. Stoddard, C.D. Evans, H.A. de Wit, M. Forsius, T. Høgåsen, A. Wilander, B.L. Skjelkvåle, D.S. Jeffries, J. Vuorenmaa, B. Keller, J. Kopácek, and J. Vesely. 2007. Dissolved organic carbon trends resulting from changes in atmospheric deposition chemistry. Nature 450:537–540. doi:10.1038/nature06316

Mulholland, P.J., and W.R. Hill. 1997. Seasonal patterns in stream water nu-trient and dissolved organic carbon concentrations: Separating catch-ment flow path and in-stream effects. Water Resour. Res. 33:1297–1306. doi:10.1029/97WR00490

National Atmospheric Deposition Program. 2015. NADP maps and data. http://nadp.sws.uiuc.edu/data/ (accessed 1 May 2015).

Oni, S.K., M.N. Futter, K. Bishop, S.J. Köhler, M. Ottosson-Löfvenius, and H. Laudon. 2013. Long-term patterns in dissolved organic carbon, major ele-ments and trace metals in boreal headwater catchments: Trends, mecha-nisms and heterogeneity. Biogeosciences 10:2315–2330. doi:10.5194/bg-10-2315-2013

Pacific, V.J., K.G. Jencso, and B.L. McGlynn. 2010. Variable flushing mechanisms and landscape structure control stream DOC export during snowmelt in a set of nested catchments. Biogeochemistry 99:193–211. doi:10.1007/s10533-009-9401-1

Park, J.H., K. Kalbitz, and E. Matzner. 2002. Resource control on the production of dissolved organic carbon and nitrogen in a deciduous forest floor. Soil Biol. Biochem. 34:813–822. doi:10.1016/S0038-0717(02)00011-1

Pregitzer, K.S., D.R. Zak, A.J. Burton, J.A. Ashby, and N.W. MacDonald. 2004. Chronic nitrate additions dramatically increase the export of carbon and nitrogen from northern hardwood ecosystems. Biogeochemistry 68:179–197. doi:10.1023/B:BIOG.0000025737.29546.fd

Räike, A., P. Kortelainen, T. Mattsson, and D.N. Thomas. 2012. 36 year trends in dissolved organic carbon export from Finnish rivers to the Baltic Sea. Sci. Total Environ. 435-436:188–201. doi:10.1016/j.scitotenv.2012.06.111

Reichstein, M., M. Bahn, P. Ciais, D. Frank, M.D. Mahecha, S.I. Seneviratne, J. Zscheischler, C. Beer, N. Buchmann, D.C. Frank, D. Papale, A. Rammig, P. Smith, K. Thonicke, M.V.D. Velde, S. Vicca, A. Walz, and M. Watten-bach. 2013. Climate extremes and the carbon cycle. Nature 500:287–295. doi:10.1038/nature12350

Rüegg, J., J.J. Eichmiller, N. Mladenov, and W.K. Dodds. 2015. Dissolved or-ganic carbon concentration and flux in a grassland stream: Spatial and temporal patterns and processes from long-term data. Biogeochemistry 125:393–408. doi:10.1007/s10533-015-0134-z

Journal of Environmental Quality

Schindler, D., J. Curtis, S. Bayley, B. Parker, B. Ken, and M. Stainton. 1997. Cli-mate-induced changes in the dissolved organic carbon budgets of boreal lakes. Biogeochemistry 36:9–28. doi:10.1023/A:1005792014547

Sebestyen, S.D., E.W. Boyer, and J.B. Shanley. 2009. Responses of stream nitrate and DOC loadings to hydrological forcing and climate change in an upland forest of the northeastern United States. J. Geophys. Res. 114:G02002. doi:10.1029/2008JG000778

Sen, P.Q. 1968. Estimates of the regression coefficient based on Kendall’s tau. J. Am. Stat. Assoc. 63:1379–1389. doi:10.1080/01621459.1968.10480934

Stewart, A.J., and R.G. Wetzel. 1982. Influence of dissolved humic ma-terials on carbon assimilation and alkaline phosphatase activ-ity in natural algal-bacterial assemblages. Freshwater Biol. 12:369–380. doi:10.1111/j.1365-2427.1982.tb00630.x

Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex, and P.M. Midgley, editors. 2013. Climate change 2013: The physical science basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge Univ. Press, Cambridge, UK, and New York.

Swank, W.T., and J.B. Waide. 1988 Characterization of baseline precipitation and stream chemistry and nutrient budget of control watersheds. Ecol. Stud. 66:57–79.

Swift, L.W., G.B. Cunningham, and J.E. Douglass. 1988. Climate and hydrology. Ecol. Stud. 66:35–55.

Thurman, E.M. 1985. Organic geochemistry of natural waters. Kluwer Academ-ic, Bonn, Germany.

Tate, C.M., and J.L. Meyer. 1983. The influence of hydrologic conditions and successional state on dissolved organic carbon export from forested water-sheds. Ecology 64:25–32. doi:10.2307/1937325

Tian, Y.Q., Q. Yu, A.D. Feig, C. Ye, and A. Blunden. 2013. Effect of climate and land-surface processes on dissolved organic carbon export to major U.S. coastal rivers. Ecol. Eng. 54:192–201. doi:10.1016/j.ecoleng.2013.01.028

Velbel, M.A. 1985. Geochemical mass balance and weathering rates in for-ested watersheds of the southern Blue Ridge. Am. J. Sci. 285:904–930. doi:10.2475/ajs.285.10.904

Wang, C., and R.E. Dickinson. 2012. A review of global terrestrial evapotranspi-ration: Observation, modeling, climatology, and climatic variability. Rev. Geophys. 50:RG2005. doi:10.1029/2011RG000373

Webster, J.R., and S.W. Golladay. 1984. Seston transport in streams at Coweeta Hydrologic Laboratory, North Carolina, U.S.A. Verh. Int. Ver. Theor. An-gew. Limnol. 22:1911–1919.

West, A.J., C.-W. Lin, T.-C. Lin, R.G. Hilton, S.-H. Liu, C.-T. Chang, K.-C. Lin, A. Galy, R.B. Sparkes, and N. Hovius. 2011. Mobilization and transport of coarse woody debris to the oceans triggered by an extreme tropical storm. Limnol. Oceanogr. 56:77–85. doi:10.4319/lo.2011.56.1.0077

Wilcoxon, F. 1945. Individual comparisons by ranking methods. Biom. Bull. 1:80–83. doi:10.2307/3001968

Winterdahl, M., J. Temnerud, M. Futter, S. Löfgren, F. Moldan, and K. Bishop. 2011. Riparian zone influence on stream water dissolved organic carbon concentrations at the Swedish integrated monitoring sites. Ambio 40:920–930. doi:10.1007/s13280-011-0199-4

Worrall, F., T.P. Burt, and J. Adamson. 2004a. Can climate change explain in-creases in DOC flux from upland peat catchments? Sci. Total Environ. 326:95–112. doi:10.1016/j.scitotenv.2003.11.022

Worrall, F., R. Harriman, C.D. Evans, C.D. Watts, J. Adamson, C. Neal, E. Tip-ping, T. Burt, I. Grieve, D. Monteith, and P.S. Naden. 2004b. Trends in dissolved organic carbon in UK rivers and lakes. Biogeochemistry 70:369–402. doi:10.1007/s10533-004-8131-7

Worrall, F., and T.P. Burt. 2007. Trends in DOC concentration in Great Britain. J. Hydrol. 346:81–92. doi:10.1016/j.jhydrol.2007.08.021