Biogeochemistry (2011) 104:275–291

DOI 10.1007/s10533-010-9501-y

Controls on the origin and cycling of riverine dissolved

inorganic carbon in the Brazos River, Texas
Fan-Wei Zeng • Caroline A. Masiello •

William C. Hockaday

Received: 22 January 2010 / Accepted: 21 June 2010 / Published online: 23 July 2010
Ó Springer Science+Business Media B.V. 2010

Abstract Rivers draining watersheds that include carbonate in the lower Brazos, in addition to natural
carbonate bedrock or organic matter (OM)-rich sedi- marl and pedogenic carbonate. Additionally, the
mentary rocks frequently have 14C-depleted dissolved generally low pCO2 and high DIC concentration in
inorganic carbon (DIC) relative to rivers draining the Brazos River lead to a low CO2 outgassing:DIC
carbonate- and OM-free watersheds, due to dissolution export ratio, distinguishing the Brazos River from
of carbonate and/or decomposition of ancient OM. other rivers.
However, our results from a subtropical river, the
Brazos River in Texas, USA, show that in this Keywords Carbon cycle  Dissolved inorganic
watershed human activities appear to dominate basin carbon  Lithology  pCO2  River system
lithology in controlling the origin and metabolism of
DIC. The middle Brazos flows through limestone and
coal-bearing bedrock, but DIC isotope data suggest no Introduction
limestone dissolution or respiration of ancient OM, and
instead reflect efficient air–water CO2 exchange, Conservative estimates show that globally, rivers
degradation of relatively young OM and photosynthe- receive at least 1 Gt C year-1 (in both organic and
sis in the river as a result of river damming and urban inorganic forms) from land (Cole et al. 2007;
treated wastewater input. The lower Brazos drains only Meybeck 1993), equivalent to half of the net oceanic
small areas of carbonate and coal-bearing bedrock, but uptake of anthropogenic CO2 (Sarmiento and Sund-
DIC isotope data suggest the strong influence of quist 1992). Of the terrestrial carbon transported by
carbonate dissolution, with a potentially minor contri- river systems, at least 0.35 Gt year-1 is released to
bution from decomposition of old soil organic matter the atmosphere as CO2, and about 0.7 Gt year-1 is
(SOM). Oyster shells and crushed carbonate minerals exported to the ocean (Cole et al. 2007).
used in road construction are likely sources of One important control on the source and fate of
terrestrial carbon in river systems is basin lithology, as
reflected in the isotope values of riverine dissolved
F.-W. Zeng (&)  C. A. Masiello  W. C. Hockaday
Department of Earth Science, Rice University, inorganic carbon (DIC = dissolved CO2 ? HCO3- ?
P.O. Box 1892, Houston, TX 77251-1892, USA CO32-). Black shales rich in organic matter (OM) in
e-mail: fwzeng@rice.edu; fwzeng@gmail.com the northeast US provide rivers with very old organic
carbon, and decomposition of this ancient terrestrial
W. C. Hockaday
Geology Department, Baylor University, organic carbon may have contributed to the low D14C
One Bear Place #97354, Waco, TX 76798, USA values of the DIC (-69 to -6%) and to evasion of CO2

123
276 Biogeochemistry (2011) 104:275–291

from the northeast US rivers (Cole and Caraco 2001; longest river (2060 km) in Texas. It runs through
Raymond et al. 2004). In watersheds composed of large areas of carbonate bedrock in its middle reach,
carbonate rocks (e.g., limestone and dolomite), disso- which has minor outcrops of coal-bearing bedrock.
lution of carbonate is generally an important source of Several cities are developing along this middle reach
riverine DIC (Chakrapani and Veizer 2005; Helie et al. of the Brazos River, and the river main stem is
2002; Kanduc et al. 2007b; Karim and Veizer 2000; dammed in two of these cities. These characteristics
Lambs et al. 2009; Zhang et al. 2009). Even in the Tana make the Brazos River an excellent field site to study
River in Kenya, in which river particulate organic carbon the effects of basin lithology and human activities on
(POC) is significantly decomposed during transit, the the sources and carbon isotope values (d13C and
dissolution of carbonate and silicate in the watershed still D14C) of riverine DIC.
dominates OM respiration as the major source of river The second objective of this study was to determine
DIC (Bouillon et al. 2009). the fraction of the inorganic carbon in the Brazos River
Human activities, such as river damming and urban that is exported to the ocean. Cole et al. (2007) estimated
development, are another control on river carbon that about 0.23 Pg year-1 of terrestrial carbon is lost
cycling. Wachniew (2006) found that rapid decompo- from rivers to the atmosphere through CO2 outgassing,
sition of labile OM from effluents of two wastewater and about the same amount of terrestrially derived
plants lowered d13C values of DIC in the Vistula River inorganic carbon (0.26 Pg year-1) is exported to the
in Poland by about 1%. Enrichment of 13C in the DIC ocean, suggesting that the CO2 outgassing:DIC export
pool due to enhanced air–water CO2 exchange in ratio of global rivers is about 1:1. However, tropical
reservoirs also affects river carbon isotope values rivers tend to outgas substantially more CO2 to the
(Brunet et al. 2005; Wachniew 2006). Increased DIC atmosphere relative to the inorganic carbon they deliver
export from urban watersheds as a result of increased to the ocean. For example, in the Amazon River system,
chemical weathering, CO2 production, and wastewater the amount of terrestrial carbon returned to the atmo-
input has also been observed (Barnes and Raymond sphere each year is more than ten times the total carbon
2009). (organic and inorganic) the river exports to the ocean
Studies assessing the effects of lithology versus (CO2 outgassing:DIC export[10:1) (Richey et al. 2002).
human activities on river carbon cycling are few, and Similarly, the Nyong River in Cameroon releases
these studies have generally found that the influence CO2 to the atmosphere at a rate four times that of
of lithology is more important than that of human DIC export to the Gulf of Guinea (CO2 outgassing:DIC
activities. For example, although carbonates compose export = 4:1) (Brunet et al. 2009). If river CO2 evasion
only *5% of the bedrock geology of the watershed in and DIC export are equal (0.23–0.26 Pg year-1) on
the densely populated Lagan River basin in Northern a global scale, then some rivers must export more
Ireland, they were a more important source of river inorganic carbon to the ocean than they release as
DIC than wastewater discharged from industries and CO2 (i.e., CO2 outgassing:DIC export \1:1). One
urban–rural areas (Barth et al. 2003). In the Vistula example is the Ottawa River, of which CO2 evasion is
River in Poland, which receives high loads of pollu- about 30% of DIC export by the river (CO2 outgas-
tants from its watershed and drains a watershed mostly sing:DIC export = 0.3:1) (Telmer and Veizer 1999).
composed of carbonate bedrocks, d13C values of DIC Therefore, the second objective of this study was to
suggest a predominant influence of carbonate rock quantify the CO2 outgassing and DIC export fluxes,
dissolution (Wachniew 2006). Similarly, lithology and the CO2 outgassing:DIC export ratio of the Brazos
showed a dominant control on carbon geochemistry River.
in a mountainous tributary of the Zhujiang (Pearl We measured concentration and carbon isotopes
River) in southeast China, although impacts of agri- (d13C and D14C) of DIC as well as pCO2 to demonstrate
cultural, urban, and river damming activities were also that the Brazos River differs from previously studied
observed (Zhang et al. 2009). rivers in that (1) human activities appear to have
Our first objective in this study was to determine overprinted the natural isotope values of DIC in the
the relative influence of basin lithology and human Brazos River derived from basin lithology, and (2) the
activities on carbon cycling in the Brazos River, a Brazos River degasses much less CO2 to the atmo-
subtropical river in Texas. The Brazos River is the sphere than it delivers DIC to the ocean.

123
Biogeochemistry (2011) 104:275–291 277

Study area and methods Pleistocene and Holocene (11.7 Ka BP to present)

(Phillips 2007; Stricklin 1961). This alluvium consists
Study area of aluminosilicate gravel, sand, clay and silt (Shah et al.
2007). The presence of old carbon in the drainage
The Brazos River begins at the juncture of the Salt basin bedrock is very limited in the lower Brazos
Fork and the Double Mountain Fork, which merge at compared to the middle Brazos (Fig. 1a), but includes
33°160 N, 100°0.50 W, about 110 km upstream of (1) carbonate existing as Cretaceous Taylor marl
Seymour, TX. It then flows 1344 km across Texas (99.6–65.5 Ma BP) about 40 km downstream Waco,
in a southeasterly direction, and drains into the Gulf TX in a belt 160 km long and 56 km wide parallel to
of Mexico at 28°52.50 N, 95°22.50 W, 3 km south of the coast (Cronin et al. 1963); (2) lignite in the Tertiary
Freeport in Brazoria County, TX (Fig. 1a). Of the Wilcox and Jackson groups (Ruppert et al. 2002); (3)
116000 km2 Brazos River basin, 107520 km2 are in pedogenic carbonate in the Beaumont Formation in the
Texas. The relief of the Brazos River basin is gener- coastal region of Texas (Nordt et al. 2006); and (4)
ally low, with a channel gradient of no more than oyster shells and crushed carbonate minerals used in
0.67 m per kilometer (Stricklin 1961). Besides the road construction in the US Gulf coast (Doran 1965;
Salt and Double Mountain forks, the Brazos has five Titi et al. 2003). Although these small sources of old
other principal tributaries: the Clear Fork, the Bos- carbon are numerous, they are scattered and account
ques River, the Little River, Yegua Creek and the for only a small portion of regional lithology.
Navasota River (Fig. 1a). The predominant land use in the entire Brazos
The climate in the middle and lower Brazos varies watershed is agriculture (Phillips 2007) (Fig. 1b).
from temperate to subtropical. During the entire study Within the middle and lower Brazos watershed,
period (March 2007 to July 2009), the mean annual forests and pasture dominate the northern region from
precipitation was 990 mm in Granbury, TX (middle 80 km upstream of Graham to Waco, while grassland
Brazos, 32°26.50 N, 97°47.50 W) and 1050 mm in and row-crop agriculture dominate downstream of
Freeport, and the mean annual temperature was Waco. Both sides of the land adjacent to the lower
16°C in Granbury and 21°C in Freeport (NOAA Brazos from Waco to College Station, TX are
2009). Daily discharge variations were consistent dominated by cropland, with little or no riparian
along the Brazos River, with high river discharges in borders. As the Brazos flows toward the Gulf of
the spring and summer of 2007 and 2008, and spring Mexico, it runs through or near several cities
and fall of 2009 (data from USGS). (Fig. 1b). The total population in the entire Brazos
The lithology of the Brazos River basin varies along drainage basin is about 3.5 million (Phillips 2007).
the river (Fig. 1a). In the upper Brazos (upstream from The Brazos is dammed in several places, and all three
Graham, TX), the bedrock is primarily composed of major dams are located north of Waco.
highly erodible Pennsylvanian (upper Carboniferous, We collected samples at seven sites in the main stem
318–299 Ma BP) and Permian (299–251 Ma BP) of the middle and lower Brazos (Fig. 1a). The two sites
red beds, with a large fraction of sand (Stricklin (1 and 2) in the middle Brazos are located where the
1961). The middle Brazos (Graham to Waco, TX) is Brazos flows through urban areas, Granbury and Waco
underlain by resistant Pennsylvanian and Cretaceous (Table 1). The main stem of the Brazos is dammed
(145.5–65.5 Ma BP) limestone and red beds (Moore several kilometers downstream of these sites, forming
and Plummer 1922; Stricklin 1961). Bituminous coal is Lake Granbury in Granbury and a town-lake in Waco.
present in the Middle-Upper Pennsylvanian Strawn, The river is wide and flows slowly at both sites 1 and
Canyan, and Cisco groups (Hackley et al. 2009; Moore 2 (Table 1). The five sites in the lower Brazos are
and Plummer 1922). In the lower Brazos (downstream located in rural areas. From north to south, they are at
from Waco), the predominant bedrock is Eocene FM 60 about 10 km southwest of College Station (site
(56–34 Ma BP), Miocene (23–5.3 Ma BP) and Pleis- 3), at US 290 (site 4), at FM 1458 (site 5), about 20 km
tocene (2.6 Ma to 11.7 Ka BP) sand, silt and clay upstream from the mouth (site 6), and the mouth
deposits (Cronin et al. 1963; Stricklin 1961). The (site 7) (Fig. 1a, b; Table 1). We sampled sites 1–5 on
channel floor and walls of the lower Brazos are bridges over the Brazos River and sites 6 and 7 using a
alluvium (green color in Fig. 1a) formed in the boat.

123
 278

123
Fig. 1 a Surface geologic map of the Brazos drainage basin made according to digital maps by Stoeser et al. (2005) and Tewalt et al. (2008). Compositions of strata are: Permian age—
red beds, with a large fraction of sand; Pennsylvanian age—red beds and limestone; Cretaceous age—limestone (Stricklin 1961). The seven tributaries of the Brazos (letters in circles)
are Salt Fork (A), Double Mountain Fork (B), Clear Fork (C), Bosques River (D), Little River (E), Yegua Creek (F), and Navasota River (G). Cities are labeled in pink. There are seven
sampling sites numbered in the boxes. Sites 1 and 2 are dammed several kilometers downstream. Dams cannot be seen because they are masked by sampling site symbols (yellow stars).
b Land cover map of the Brazos drainage basin made according to the digital map by National Center for Earth Resources Observation and Science and U.S. Geological Survey (2005)
Biogeochemistry (2011) 104:275–291
 Table 1 Basin and hydrology information for the Brazos River basin
Site Site location Distance from the Catchment area Elevation City City area (km2) Population Distance upstream Width of the river (m) Annual mean
number site to the upstream of sites (m) adjacent (in 2007) of the nearest dam discharge (m3 s-1)
mouth (km) (km2) (km) (year dam Upstream Downstream (period)
completed) of dam of dam
Biogeochemistry (2011) 104:275–291

1 In Granbury 460 65,350 191.9 Granbury 15.9 8,029 12 (1969) 400 80 29.3 (1970–2008)
(32°26.50 N, 97°460 W)
2 In Waco 344 75,671 106.5 Waco 247.4 122,222 4 (1970) 100 50 62.2 (1971–2008)
(31°33.70 N, 97°7.70 W)
3 At FM 60* 230 99,965 57.7 – – – – 70 149.9 (1994–2008)
(30°33.50 N, 96°25.50 W)
4 At US 290* 160 112,333 32.9 – – – – 100 198.7 (1939–2008)
(30°7.70 N, 96°11.20 W)
5 At FM 1458* 125 113,903 20.7 – – – – 140 206.2 (1939–2008)
(29°48.50 N, 96°5.70 W)
6 Near the mouth 20 116,068 14.9 – – – – 120 240.4 (1968–2008)
(29°1.50 N, 95°27.70 W)
7 At the mouth 0 0.9 – – – – 200
(28°52.70 N, 95°22.80 W)

Sites in the middle Brazos: sites 1 and 2, in Granbury and Waco, respectively. Sites in the lower Brazos: sites 3–7
Sites 1–5 were sampled on bridges crossing the Brazos River. Site 6 and 7 were sampled using a boat
Discharge, catchment and elevation data were obtained from US Geological Survey. For site 1: average of data from stations 08090800 and 08091000; for site 2: station 08096500; for site 3: station
08108700 (about 15 km upstream of site 3); for site 4: station 08111500; for site 5: average of data from stations 08111500 and 08114000; for sites 6 and 7: station 08116650 (about 30 km upstream of
site 6)
* Numbers are the names of roads or highways
279

123
280 Biogeochemistry (2011) 104:275–291

Methods riverine particulate organic matter (POM) at site 3,

which is downstream from most of the lignite
This study was conducted between March 2007 and deposits (Fig. 1a). We made this measurement using
July 2009. At each site, we measured surface water cross-polarization and magic angle spinning (CP/
temperature and pH in situ using a YSI pH10 pH/ MAS) solid-state 13C NMR (nuclear magnetic reso-
Temperature meter. On July 23, 2009, we also nance) and a molecular mixing model. We collected
measured dissolved oxygen content (DO) of surface about 120 L of river water, which was stored in a
water in situ at sites 1 and 2 using an HI 9828 refrigerator overnight to let the particles settle. We
multimeter from Hanna Instruments. All the water then removed the top clear water and used a
pCO2 and DIC samples were taken directly from the centrifuge to concentrate sediment from the bottom
middle part of river cross-sections at depths between water. Finally we dried the sediment using a freeze
10 and 30 cm from bridges using a submersible pump dryer. We conducted NMR experiments on Rice
(SHURflo 9325-043-101). University’s 200 MHz Bruker Avance spectrometer,
For each site we collected three ambient air pCO2 equipped with a 4 mm magic angle spinning (MAS)
and three water pCO2 samples. We took ambient air probe. We aquired CP spectra with a 1 ms contact
CO2 samples on bridges, and collected river water time, at 5 kHz MAS frequency, and 2 s recycle delay.
pCO2 samples by headspace equilibration as For each experiment, we collected 40,000 acquisi-
described in Raymond et al. (1997). All pCO2 tions (scans). We then processed and deconvoluted
samples were transported to the laboratory and the NMR spectrum with the parameters and molec-
analyzed using an infrared CO2 analyzer (Li-Cor ular mixing model described by Baldock et al.
7000) on the same day they were collected, using a (2004). The molecular mixing model generates
set of 349, 1500, 3000, and 10100 latm CO2 gas an estimate of black carbon content, among other
standards (Scott Specialty Gases). properties.
We collected DIC samples using 250 mL pre- We applied Student’s t-test to examine if there is
cleaned (soapy water, deionized water and Milli-Q significant difference in river pCO2, DIC concentra-
water) and pre-combusted (500°C for 2 h) bottles. To tion, d13C and D14C of DIC between the middle and
minimize CO2 exchange between the sample and the lower Brazos. p values equal to or less than 0.05
air, we overflowed the bottle from the bottom three suggest that the difference is significant, and p values
times. DIC samples were stored at 0°C from time of less than 0.01 suggest that the difference is very
collection to poisoning with saturated HgCl2 solution significant.
and sealing with Apiezon-N grease in the lab. All
DIC samples were sent to the National Ocean
Sciences Accelerator Mass Spectrometry Facility Results
(NOSAMS) for analysis of DIC concentration and
isotopic composition (D14C and d13C). There, DIC Longitudinal variation in pCO2, concentration
samples were acidified using 85% H3PO4 and then and isotope values of DIC
sparged with high-purity N2 gas (99.99%) (McNichol
et al. 1994). The evolved CO2 was collected Measured concentration of atmospheric CO2 at all
cryogenically using liquid nitrogen traps (-190°C) study sites averaged 424 ± 43 latm (Table 2). Riv-
on a vacuum line (McNichol et al. 1994). Small erine pCO2 values in the Brazos were generally low.
aliquots (10%) of the CO2 gas were analyzed on an Average pCO2 was 760 ± 243 latm in the middle
isotope ratio mass spectrometer VG PRISM or VG Brazos and was 1174 ± 418 latm in the lower
OPTIMA for d13C. Larger aliquots (20%) of the CO2 Brazos (Table 2; Fig. 2a), showing no significant
gas were reduced to a graphite target, which was difference (p = 0.0564) between the middle and
finally analyzed by accelerator mass spectrometry lower Brazos. Riverine pCO2 values at sites 6 and 7
(AMS) to determine D14C (McNichol et al. 1994). were about three times the atmospheric pCO2,
To explore the possible contribution of old organic slightly lower than those at sites 4 and 5 (4–5 times
carbon from coal weathering to the Brazos River’s the atmospheric pCO2) probably due to mixing of
DIC pools, we analyzed the black carbon content of high-pCO2 river water with low-pCO2 ocean water in

123
Biogeochemistry (2011) 104:275–291 281

Table 2 Measured parameters for all sampling sites in the Brazos River
Site Date pH T (°C) DO pCO2,a pCO2,wa DICb D14C-DICc* d13C-DICd*
number sampled (mg L-1) (latm) (latm) (mM) (%) (%)

The middle Brazos

1 3/5/2009 8.08 15.6 – 361 439 3.0 ?10 -4.2
7/23/2009 8.55 27.3 3.36 415 966 2.3 -10 -3.5
Mean ± SD 8.32 ± 0.33 21.5 ± 8.3 – 388 ± 38 703 ± 373 2.6 ± 0.5 0 ± 14 -3.9 ± 0.5
2 1/6/2008 8.65 12.6 – 443 858 3.5 ?73 -7.5
7/3/2008 7.92 27.5 – 447 805 2.8 ?162 -5.9
10/30/2008 7.70 18.7 – 389 483 2.8 ?96 -5.7
7/23/2009 8.69 28.9 5.47 389 1007 2.6 ?179 -4.6
Mean ± SD 8.24 ± 0.50 21.9 ± 7.7 – 417 ± 32 788 ± 221 2.9 ± 0.4 ?128 ± 51 -5.9 ± 1.2
1–2 Mean ± SD 8.27 ± 0.42 21.8 ± 7.0 – 407 ± 34 760 ± 243 2.8 ± 0.4 ?85 ± 77 -5.2 ± 1.4
The lower Brazos
3 3/5/2009 8.01 21.1 – 363 435 4.2 -84 -6.8
4 3/25/2007 8.22 22.8 – 456 1351 2.4 -135 -9.4
6/13/2007 – – – 480 1770 4.1 -220 -6.6
Mean ± SD 8.22 22.8 – 468 ± 17 1561 ± 296 3.3 ± 1.2 -177 ± 60 -8.0 ± 1.9
5 3/25/2007 8.24 22.7 – 468 1451 2.5 -141 -9.2
6 11/28/2007 8.32 16.6 – 451 1120 2.3 -194 -9.9
11/19/2008 7.83 17.4 – 379 1029 2.0 -192 -10.6
Mean ± SD 8.08 ± 0.35 17.0 ± 0.60 – 415 ± 51 1075 ± 64 2.2 ± 0.3 -193 ± 1.5 -10.3 ± 0.6
7 11/28/2007 7.94 15.8 – 466 1060 2.2 -94 -9.9
3–7 Mean ± SD 8.09 ± 0.19 19.4 ± 3.2 – 438 ± 47 1174 ± 418 2.8 ± 0.9 -151 ± 52 -8.9 ± 1.6
Sites 3, 5, and 7 were measured only once during the study period. We applied Student’s t-test to examine if there is significant
difference between the middle Brazos (sites 1 and 2) and the lower Brazos (sites 3–7) in the parameters
* Means that the difference is significant
a b c d
p = 0.0564, p = 0.9732, p \ 0.0001, p = 0.0011

the Gulf of Mexico (Green et al. 2006; Lohrenz and Black carbon content of riverine POM at site 3
Cai 2006). in the lower Brazos
The concentration of DIC was consistent along the
Brazos (p = 0.9732), averaging 2.8 ± 0.4 mM in the Black carbon is detected within the NMR peak
middle Brazos and 2.8 ± 0.9 mM in the lower Brazos occurring between 110 and 145 ppm. This peak is
(Table 2; Fig. 2b). However, the D14C (p \ 0.0001) small relative to other peaks in the spectrum of riverine
and d13C (p = 0.0011) values of DIC differed signif- POM at site 3 (Fig. 3). Because aromatic carbon within
icantly between the middle and lower Brazos (Table 2; lignin is also detected between 110 and 145 ppm, it is
Fig. 2c–d). Average D14C values of DIC in the middle necessary to use a molecular mixing model to deter-
and lower Brazos were ?85 ± 77% and -151 ± mine the black carbon concentration. Using the
52%, respectively, and average d13C values of DIC in molecular mixing model of Baldock et al. (2004), we
the middle and lower Brazos were -5.2 ± 1.4% calculated black carbon to be 3.8% of the riverine
and -8.9 ± 1.6%, respectively (Table 2). Temporal POM. Since NMR-measured black carbon identifies
variations in these parameters were minor relative both coal-derived black carbon and soil-derived char-
to longitudinal variations (Table 2). The statisti- coal black carbon as one pool, our final measurement of
cally significant differences in carbon isotope values 3.8% is an upper estimate for coal input to river POM,
between the middle and lower Brazos suggest distinc- allowing us to conclude that coal contributed to no
tive, regional sources and metabolism patterns of DIC. more than 3.8% of POM in the Brazos.

123
282 Biogeochemistry (2011) 104:275–291

Fig. 2 Longitudinal variations in river water pCO2 (a), DIC sites. Numbers on the figures are study site numbers (Table 1).
concentration (b), D14C of DIC (c), and d13C of DIC (d) in the Vertical bars represent the standard deviation of all the data for
middle and lower Brazos. The dashed line in a is the measured the site
mean pCO2 value of the atmosphere (428 latm) at the study

invasion, OM respiration (young and old), and

dissolution of carbonate (Fig. 4). There are three
potential sources of carbonate in the Brazos
watershed: (1) natural carbonate rocks (Pennsylva-
nian and Cretaceous limestone in the middle Brazos
and Cretaceous Taylor marl in the lower Brazos), (2)
natural pedogenic carbonate in the lower Brazos, and
(3) anthropogenic carbonate minerals (oyster shells
400 350 300 250 200 150 100 50 0 -50 -100 -150
and crushed limestone used in road construction) in
Chemical Shift (ppm) the lower Brazos. Each potential DIC source has a
characteristic range of carbon isotope ratios (Fig. 4),
Fig. 3 CP/MAS 13C NMR spectrum of riverine POM at site 3 which we outline below.
in the lower Brazos
Atmospheric CO2 invasion is one potential source
Discussion of DIC. The D14C values of atmospheric CO2 were
?55 to ?66% in 2004 (Hsueh et al. 2007), and have
Isotope values of potential DIC sources been dropping at a rate of 6% per year (Trumbore
in the Brazos River basin et al. 2006), implying that for 2007–2008 DIC from
atmospheric CO2 invasion had D14C values of ?31 to
We investigated the origins of riverine DIC using a ?48% and d13C values of -5 to -3% (Raymond
plot of d13C versus D14C for our samples and et al. 2004).
potential DIC sources (Raymond et al. 2004; Zeng A second potential source of DIC is CO2 derived
and Masiello 2010) which include atmospheric CO2 from respiration of terrestrial OM. d13C values of

123
Biogeochemistry (2011) 104:275–291 283

Fig. 4 d13C versus D14C of riverine DIC in the middle (pink are four possible sources of carbonate in the study watershed (in
triangles) and lower (blue diamonds) Brazos. Also plotted are dashed boxes): natural sedimentary carbonate minerals and
d13C versus D14C of potential DIC sources (in solid boxes): crushed carbonate minerals (blue), natural pedogenic carbonate
atmospheric CO2 (orange), young OM respiration (red), old (black), and oyster shells (green). Carbon isotope values of each
OM respiration (brown), and carbonate dissolution (blue). There source are discussed in detail in the text

DIC from terrestrial OM respiration are typically d13C values of -14.5 to -9%. Pedogenic carbonate
in the range -23 to -14%, after correction for in Texas has D14C values of -1000 to -860%, and
fractionation (Barth et al. 2003; Mook et al. 1974; d13C values of -6 to ?1% (Nordt et al. 1998;
Raymond et al. 2004; Zhang et al. 1995), and D14C Rightmire 1967; Valastro et al. 1968). Therefore, the
values of DIC from OM respiration are ?50 to DIC sourced from pedogenic carbonate and soil CO2
?170% for young OM and -230 to -100% for old has D14C values of -490 to -355% and d13C values
OM (Raymond et al. 2004). A recent study (Griffith of -17.5 to -8.5%. The D14C and d13C values of
et al. 2009) found that OM in municipal wastewater shell carbonates have been measured in many studies
treatment plant effluent had a mean radiocarbon age (Douglas and Staines-Urias 2007; Gentry et al. 2008;
of about 1600 ± 500 years (D14C = -180 ± 50%). Keller et al. 2002; McConnaughey and Gillikin 2008;
Thus, DIC from the respiration of municipal waste Wisshak et al. 2009). Combining data from all these
falls within old OM respiration in Fig. 4. studies, we estimate shell carbonate D14C = -489 to
The final DIC source is carbonate dissolution. -240% and d13C = -7 to ?4%. The dissolution of
Isotope values of DIC produced by carbonate disso- shell carbonate by soil CO2 generates DIC with D14C
lution vary depending both on the source of carbonate values of -235 to -45% and d13C values of -18 to
and on the dissolving acid. Soil CO2 is generally the -7%.
dominant source of acidity that dissolves carbonate in The broad ranges of isotopic values for every
river systems (Garrels and Mackenzie 1971; Karim potential DIC source precluded identifying the
and Veizer 2000). Dissolution of different types of amount of DIC derived from each source. For
carbonate by soil CO2 produces DIC with isotope example, carbonate D14C values could range from
values intermediate between the carbonate and soil -490% to -45%; respirable OM D14C values could
CO2, which for soil CO2 are D14C = ?20 to ?150% range from -230% to ?170%, and atmospheric CO2
and d13C = -29 to -18% (Raymond et al. 2004; D14C values could range from ?31 to ?48%. There
Trumbore 2000; Trumbore et al. 2006). The dissolu- are also large ranges of values for the d13C of each
tion of carbonate rocks and crushed limestone source. The range of D14C values for carbonates
(D14C = -1000%, d13C = *0%) by soil CO2 pro- alone, holding all other potential contributors’ D14C
duces DIC with D14C values of -490 to -425% and and d13C values constant, is enough to cause the

123
284 Biogeochemistry (2011) 104:275–291

potential carbonate contribution to vary between 32 rivers. In China, river damming activities and increas-
and 94% of DIC in the lower Brazos. Isotopic data do ing eutrophy driven by urban growth have reduced
allow us to say conclusively what the major sources dissolved CO2 concentrations in the Changjiang
of DIC are likely to be, but we cannot assign specific River, the largest river in China (Wang et al. 2007).
fractions to each source. Iwata et al. (2007) also observed that streams draining
Assuming that basin lithology is the dominant more urbanized watersheds were less heterotrophic
control on the source and cycling of riverine DIC in (meaning with lower dissolved CO2 concentra-
the Brazos River as in other river systems (Helie et al. tions) than streams in intensively farmed watersheds.
2002; Kanduc et al. 2007b; Raymond et al. 2004), we Damming and urbanization control the geochemistry
expected to see a clear longitudinal trend in carbon of river water by: (1) enhancing exchange between
isotope values of riverine DIC. In the middle Brazos, riverine DIC and atmospheric CO2, (2) leading to
DIC should be highly depleted in 14C relative to the respiration of relatively young OM in river sediments
atmosphere resulting from dissolution of large areas above the dams; and (3) stimulating surface algal
of carbonate bedrock and perhaps respiration of production. The influence of each mechanism is dis-
ancient OM from bituminous coal erosion. In the cussed below.
lower Brazos, due to the limited sources of old
carbon, we expected DIC with a higher D14C value
compared to DIC in the middle Brazos. Our obser- Enhanced air–water CO2 exchange
vation that DIC in the middle Brazos was generally
enriched in 14C relative to the lower Brazos was Damming of rivers creates reservoirs that have large
opposite to predictions based upon lithology. This surface areas and relatively long hydraulic residence
suggests a dominant influence of human activities on times, both of which enhance the exchange of gas
carbon cycling in the Brazos River. between river water and the atmosphere. At sites 1 and
2 the river is 2–5 times its original width (Table 1).
Major controls on riverine DIC in the middle Extensive air–water exchange reduces the concentra-
Brazos: damming and urbanization tion gradient of dissolved gases between river water
and the atmosphere, and draws riverine pCO2 values
The D14C values of DIC in the middle Brazos closer to that of the atmospheric CO2 (for sites 1 and
(?85 ± 77%) suggest that the Pennsylvanian and 2, average water pCO2 = 760 ± 243 latm, average
Cretaceous limestone and bituminous coal contribute ambient air pCO2 = 407 ± 34 latm).
little or no carbon to riverine DIC (Fig. 4), although Extensive air–water exchange also resulted in
they outcrop in large areas in the middle Brazos. The relatively high D14C and d13C values of riverine
relatively low D14C-DIC values at site 1 (?10% and DIC. Atmospheric CO2 has D14C values of ?31 to
-10%) could be due to a very minor input of these ?48% in 2007–2008 (Hsueh et al. 2007; Trumbore
ancient carbonate rocks. They could also be due to et al. 2006) and a d13C value of -8% (Levin et al.
decomposition of old OM from treated urban waste- 1987). The invasion of atmospheric CO2, therefore,
water. However, either case suggests that carbonate shifted both D14C and d13C values of riverine DIC
dissolution is not the predominant source of DIC at to heavier values. Since dissolved CO2 is more
13
site 1. The average D14C-DIC value of ?128 ± 51% C-depleted relative to other DIC components (i.e.,
at site 2 rules out a carbonate input, and instead likely HCO3- and CO32-) (Alin et al. 2008; Mook et al.
reflects an 8- to 50-year residence time of carbon in 1974), removal of dissolved CO2 through outgassing
drainage basin ecosystems (Burchuladze et al. 1989; could increase d13C values of DIC in the water. Our
Hsueh et al. 2007; Trumbore et al. 2006). pCO2 and DIC isotope data in the middle Brazos are
The concurrently high D14C-DIC and d13C-DIC consistent with the results of previous studies (Brunet
values and low pCO2 values in the middle Brazos are et al. 2005; Finlay 2003; Zhang et al. 2009), in which
likely due to damming and urbanization of the Brazos high d13C values of DIC (up to -1.8%) and lower
River in Granbury and Waco. The low pCO2 values concentrations of dissolved CO2 were attributed
that we observed in the middle Brazos were consistent to efficient air–water exchange in large rivers and
with observations in dammed, urbanized Chinese impoundment due to dam construction.

123
Biogeochemistry (2011) 104:275–291 285

Respiration of young OM bacteria via methanogenesis, producing highly 13C-

depleted methane (CH4) and 13C-enriched CO2
The D14C-DIC values of ?73 to ?179% at site 2 can relative to the OM source (Gu et al. 2004; Stiller
only be explained by decomposition of relatively and Magaritz 1974; Stiller et al. 1985; Wachniew and
young OM, because neither atmospheric CO2 inva- Rozanski 1997). Lake Apopka in central Florida
sion nor carbonate dissolution could produce DIC (Gu et al. 2004) has d13C-DIC values as high as ?6 to
with values this high (Fig. 4). Northern hemisphere ?8% due to methanogenesis of sedimentary OM.
atmospheric 14C values spiked at *900% between Compared to the DO content of Lake Apopka (about
1962 and 1963, at the peak of atmospheric nuclear 7.5 mg L-1), measured DO contents in the surface
weapons testing (Burchuladze et al. 1989). Since the water at sites 1 and 2 (averaging 4.4 ± 1.5 mg L-1,
late 1960s, the bomb spike has mixed into the oceans Table 2) were lower, suggesting that methanogenesis
and terrestrial biosphere, leading to a drop in the of OM may have occurred and provided 13C-heavy
atmosphere’s 14C value (Burchuladze et al. 1989). CO2 to the water column. However, without CH4
Because of the spiked shape of the atmosphere’s 14C measurements in the water column, this hypothesis
value since 1950, environmental 14C values [0% do cannot be confirmed.
not correspond to unique carbon residence times.
Instead, each 14C value corresponds to two possible Algal growth
carbon residence times, one on each side of the bomb
spike. The D14C-DIC values (approx. ?170%) in CO2 uptake by algae can also lead to 13C enrichment
July 2008 and July 2009 suggest that the OM of DIC. In algal photosynthesis, dissolved CO2 is
decomposed was either *20 or *50 years old preferentially used because of lower energetic costs
(Burchuladze et al. 1989), while the D14C-DIC values of dissolved CO2 acquisition relative to HCO3-
(approx. ?85%) in January and October of 2008 (Burkhardt et al. 2001; Rotatore et al. 1995). Like
correspond to OM 8–10 or *50 years old (Hsueh removal of dissolved CO2 through outgassing, CO2
et al. 2007; Trumbore et al. 2006). The longer 14C- uptake by algal photosynthesis causes a reduction of
derived residence times in this ecosystem are dissolved CO2 concentrations, and increases d13C-
unlikely, given the region’s high temperature and DIC values. For example, in a study conducted in the
precipitation. The Amazon, draining a similarly Dead Sea (Oren et al. 1995), d13C-DIC values as high
warm and wet climate, has a carbon residence time as ?5.1% were driven by a Dunaliella bloom, while
of *5 years (Mayorga et al. 2005), making the 8–10 d13C-DIC values were 8.5% lighter before the
and 20 year residence time interpretations of our 14C bloom.
data most likely. Chlorophyll a concentration in Lake Granbury
If the ?170% D14C value reflects about a 20 year (site 1) can reach 90 lg L-1 in March (Roelke et al.
residence time, and the ?85% D14C value reflects an 2007), comparable to levels published for a eutrophic
8–10 year residence time at our site 2 (Waco), then lake in central Florida (96 lg L-1) (Gu et al. 2004),
the Brazos respires older OM in the summer and suggesting intensive algal photosynthetic activity in
younger OM in the winter. Respiration of older OM Lake Granbury. The growth of algae is favored by
in summer may be a result of more intensive nutrient enrichment. Total phosphorus was about
heterotrophic microbial activity or a shift in microbial 50 lg L-1 in Lake Granbury (BRA 2009), within the
community stimulated by warmer water (Table 2) range of mesotrophic conditions (moderately nutrient
(Boer et al. 2009). Alternatively, crops are typically enriched, total phosphorus ranging from 25 to
harvested and replanted in May and June in this part 75 lg L-1) (EPA 2001). Most of the lake nutrients
of the US. Increased soil mobility after crop harvest are from wastewater treatment plants and on-site
in summer could have led to loss of old soil organic sewage facilities near the Granbury site (Riebschlea-
matter (SOM) to the river. ger and Karthikeyan 2008).
The respiration of OM at site 2 may involve The river water passing through Waco is partially
methanogenesis, which can highly enrich 13C in the from Lake Waco, a reservoir constructed in the
DIC pool. It has been reported that OM in bottom Bosque River (Fig. 1) about ten river kilometers
sediments of lakes can be respired by anaerobic before the Bosque River merges with the Brazos

123
286 Biogeochemistry (2011) 104:275–291

River at a point only three river kilometers upstream and Hopkins 1917). This soft bedrock material is
of our sampling site in Waco. Lake Waco has been susceptible to erosion and could have added to the
classified to be mesotrophic or eutrophic in different riverine DIC pool.
time periods primarily due to the input of nutrients Pedogenic carbonate present in the Vertisols of the
from dairy waste and forage fields in the Lake Beaumont Formation in coastal Texas is another
Waco-Bosque River watershed (McFarland et al. source of carbonate (Nordt et al. 2006). Pedogenic
2001; McFarland and Hauck 2001). Enhanced algal carbonate is generally formed in arid and semi-arid
growth in Lake Waco is confirmed by increased areas. However, Vertisols in the Beaumont Formation
chlorophyll a concentration (Jones 2009; McFarland of the east Texas Gulf coast preserve pedogenic
et al. 2001). carbonate despite the high rainfall ([1000 mm per
year), due to their high clay content and shrink-swell
Natural and anthropogenic carbonate input properties (Nordt et al. 2006).
to the lower Brazos Oyster shells and crushed carbonate minerals are a
significant regional anthropogenic source of soil
DIC in the lower Brazos was strongly influenced by carbonate. Oyster shells were widely used in the US
old carbon (Fig. 4), although ancient carbon sources, Gulf coast as aggregate for the construction of roads
carbonate and lignite outcrops, are very limited in the and parking lots through the mid-twentieth century
lower Brazos. The black carbon content from NMR (Doran 1965). In the 1960s, shell roads could be
analysis was 3.8% of the riverine POM at site 3, of found in a belt about 100 km wide along the Texas
which a fraction is from the Mollisol soils drained in coast (Doran 1965), and their remnants are ubiquitous
this region. These soils have been shown to be rich in in coastal Texas urban environments. Crushed lime-
charcoal (Skjemstad et al. 2002), leading us to stone has also been used in Gulf coast roads (Titi
conclude that lignite is much less than 3.8% of the et al. 2003). In Buffalo Bayou, one of the rivers
riverine POM. Longworth et al. (2007) observed draining Houston about 50 km east of the lower
young dissolved organic matter (DOM) in watersheds Brazos, dissolution of shell and crushed limestone/
draining OM-rich shale, suggesting little or no dolomite imbedded in old city roads has contributed
contribution of shale organic carbon to riverine to the river’s old DIC (D14C = -117 ± 45%) (Zeng
DOM. Assuming that contributions of lignite black and Masiello 2010). Our isotope data show that shells
carbon to the DOM pool are also negligible in the and crushed carbonate minerals are also likely an
lower Brazos River, we conclude that lignite decom- important source of DIC in the lower Brazos, as
position has a trivial impact on 14C content of illustrated by Fig. 4.
riverine DIC. Soil erosion due to agricultural prac-
tices in the lower Brazos could have caused loss of Comparison of Brazos pCO2 and DIC
old SOM to the river as described in other studies with other rivers
(Krusche et al. 2002; Longworth et al. 2007;
Raymond et al. 2004) and decomposition of this old Riverine pCO2
OM may be a source of old riverine DIC. However,
this process is less important compared to carbonate Measured pCO2 values in the middle and lower
dissolution in driving the carbon isotope values of Brazos ranged from 435 to 1770 latm, with an
DIC, as indicated by the relatively high d13C values average (±SD) of 983 ± 397 latm during the study
of DIC (Fig. 4). period (March 2007 to July 2009). Compared to other
Possible carbonate sources in the lower Brazos subtropical rivers, pCO2 values in the Brazos were
include naturally occurring Cretaceous Taylor marl much lower than those of the Xijiang River in
and pedogenic carbonate, and oyster shells and southeast China (about 2600 latm, Yao et al. 2007),
crushed carbonate minerals used in road construction the Satilla and Altamaha rivers in the coastal region
(Doran 1965; Titi et al. 2003; Zeng and Masiello of Georgia (4000–8500 latm, Cai and Wang 1998),
2010). Taylor marl consists of highly calcareous clay and Buffalo Bayou and Spring Creek in Harris
or clay marl, which is a loose mixture of clay County, TX (3000–4200 latm, Zeng and Masiello
(65–35%) and calcium carbonate (35–65%) (Matson 2010).

123
Biogeochemistry (2011) 104:275–291 287

The low pCO2 values in the Brazos River are For the middle Brazos, the high DIC concentration
probably due to the combined effects of lower was not due to dissolution of carbonate bedrock, but
precipitation, an important control on river pCO2 instead due to urbanization. It has been observed that
(Yao et al. 2007; Zeng and Masiello 2010), larger wastewater treatment plant effluent contributes up
river size, and damming and urbanization as dis- to 22% of the total DIC in the Hockanum River,
cussed above. The mean annual temperature in the Connecticut (Barnes and Raymond 2009). Weath-
middle and lower Brazos was 16.5–21.3°C for the ering of bedrock material exposed to the surface due
study period, similar to that in the Xijiang River to urban soil disturbance, and dissolution of urban
(14–22°C) (Yao et al. 2007) and that in the Satilla concrete are two other mechanisms for the high DIC
and Altamaha river basins in 1995 (about 21°C, concentration in urbanized watersheds (Baker et al.
NOAA). However, the mean annual precipitation in 2008).
the study area for the study period, 898–1050 mm, is
much lower than that in the Xijiang River basin
CO2 outgassing and DIC export from the Brazos
(1451 mm, China Meteorological Data Sharing Ser-
vice System), the Satilla and Altamaha river basins
We used pCO2, DIC concentration and river dis-
(1580 mm in 1995, NOAA), and the Buffalo Bayou
charge data (obtained from the USGS) to calculate
and Spring Creek watersheds (1220–1350 mm,
the fluxes of CO2 outgassing and DIC export from the
NOAA). Also, large rivers in general are less
Brazos in a preliminary attempt to estimate the
supersaturated in CO2 compared to smaller streams
Brazos River basin as a C source to the atmosphere
under similar climate conditions (Finlay 2003).
and the ocean.
CO2 outgassing is calculated using the following
Riverine DIC concentration
equation (modified from Raymond et al. 1997):


The DIC concentration in the Brazos River, Outgas ¼ kCO2  pCO2;w  Kh  pCO2;a  Kh  A
2.8 ± 0.7 mM, far exceeded the average value of
world rivers (0.9 mM) (Livingstone 1963). It is an where Kh is Henry’s law constant at a given
order of magnitude higher than some northeast US temperature and salinity; pCO2,w and pCO2,a are the
rivers (Raymond et al. 2004), and higher than those in partial pressure of CO2 in surface water and the
some rivers in the Coast Range of northern California overlying air, respectively; kCO2 is the gas exchange
(Finlay 2003). coefficient for CO2 at a given temperature for a given
High DIC concentrations have also been observed type of river; and A is the area of water surface.
in the St. Lawrence River (0.5–5 mM) (Helie et al. Measured mean pCO2,w and pCO2,a values for the
2002), the Sava River (2.9 ± 1.0 mM) (Kanduc et al. Brazos were 983 and 424 latm, respectively. The
2007a), three major US rivers (Ohio, upper Missis- average temperature was 20.6°C, so Kh was 0.0383
sippi, and Missouri, 1–5 mM) (Raymond and Oh mol l-1 atm-1 (Stumm and Morgan 1996). We chose
2007), and Buffalo Bayou (2.4 ± 1.0 mM) (Zeng and a commonly used range of kCO2 values, 2.5–5 cm h-1
Masiello 2010). Most of these high DIC concentra- (Raymond et al. 1997; Richey et al. 2002; Zeng and
tions are associated with carbonate input to the river Masiello 2010), in our calculation. Using a water
systems, and sources of carbonate include bedrock, surface area (A) of 565.79 km2 for the Brazos River
oyster shells and crushed limestone used in road (including the river itself and the three main reser-
construction and as agricultural soil amendments. voirs along the main stem river channel) (BRA 2010),
Dissolution of limestone and dolomite accounts for we calculated CO2 outgassing rate from the Brazos
up to 26% of DIC in the Sava River (Kanduc et al. River to be 0.03–0.06 Tg C year-1.
2007a), and agricultural liming contributed 17% of DIC export from the Brazos River to the Gulf of
DIC in the Ohio River (Oh and Raymond 2006). Mexico is the product of average DIC concentration
Similarly, D14C and d13C values of DIC in the lower and average river discharge at the river mouth. Average
Brazos suggest that although lithologic sources of DIC concentration at the Brazos mouth was 2.8 ±
carbonate are minor, anthropogenic and possibly 0.7 mM. River discharge data were obtained from a US
pedogenic carbonates are a significant source of DIC. Geological Survey station (08116650) located about

123
288 Biogeochemistry (2011) 104:275–291

50 km upstream of the river mouth. Annual mean The pCO2 values in the Brazos River are generally
river discharge for 2007–2008 at this station was lower than other subtropical rivers, while the DIC
349 m3 s-1. Our estimated rate of river DIC export to concentration in the Brazos River is relatively high
the Gulf of Mexico is 0.37 ± 0.09 Tg C year-1. Since compared to many other rivers. As a result, the amount
the discharge at the river mouth should be higher than of DIC the Brazos River releases to the atmosphere as
at the 50 km upstream USGS station, the estimated CO2 gas is much smaller than the amount of DIC it
DIC export here is a conservative value. delivers to the ocean.
The CO2 outgassing : DIC export ratio of the Brazos
River is, therefore, about 0.1:1. This value is not only Acknowledgments CAM and FWZ acknowledge the support
of the Texas Water Resources Institute, the U.S. Geological
lower than those of tropical rivers, but also lower than Survey, and the National Institutes for Water Research. CAM
those of northern rivers such as the Ottawa River and FWZ also acknowledge the generous support of Hans O.
(Brunet et al. 2009; Richey et al. 2002; Telmer and and Suse Jahns. FWZ is grateful to T.J. Perez, S.R. Alin and D.
Veizer 1999). The low CO2 outgassing:DIC export Lockwood for their assistance with instrument setup and
sampling procedures, and John B. Anderson, Fred M. Weaver,
ratio implies that of the total DIC in the Brazos River,
Xinfeng Shi, LaQuanti Calligan, Kaijian Liu, Krystle Hodge,
about 90% is exported to Gulf of Mexico, while only Yongbo Zhai, and Morgan Gallagher for their assistance in
about 10% is released to the atmosphere. It is possible sample collection.
that this very low CO2 outgassing:DIC export ratio is
also a result of human activities, with pCO2 driven
down by damming and urban treated wastewater input References
and DIC driven up by carbonate dissolution.
Alin SR, Aalto R, Goni MA, Richey JE, Dietrich WE (2008)
Biogeochemical characterization of carbon sources in the
strickland and fly rivers, Papua New Guinea. J Geophys
Res Earth Surf 113:21
Conclusions Baker A, Cumberland S, Hudson N (2008) Dissolved and total
organic and inorganic carbon in some British rivers. Area
Our results show that human activities (i.e., damming 40:117–127
Baldock JA, Masiello CA, Gelinas Y, Hedges JI (2004)
and urbanization) dominate geology as a control on
Cycling and composition of organic matter in terrestrial
the origin and cycling of DIC in the Brazos River. and marine ecosystems. Mar Chem 92:39–64
Although limestone bedrock underlies a large area of Barnes RT, Raymond PA (2009) The contribution of agricul-
the middle Brazos River watershed, this limestone tural and urban activities to inorganic carbon fluxes within
temperate watersheds. Chem Geol 266:327–336
does not influence riverine DIC, as shown by the
Barth JAC, Cronin AA, Dunlop J, Kalin RM (2003) Influence
radiocarbon values (?85 ± 77%) of riverine DIC in of carbonates on the riverine carbon cycle in an anthro-
the middle Brazos. Neither does the coal-bearing pogenically dominated catchment basin: evidence from
bedrock present in the watershed. D14C (?85 ± major elements and stable carbon isotopes in the Lagan
River (N. Ireland). Chem Geol 200:203–216
77%) and d13C (-5.2 ± 1.4%) values of DIC in the
Boer SI, Arnosti C, van Beusekom JEE, Boetius A (2009)
middle Brazos are likely due to extensive air–water Temporal variations in microbial activities and carbon
CO2 exchange, respiration of decades old OM in river turnover in subtidal sandy sediments. Biogeosciences 6:
sediments, and algal growth, all of which are driven 1149–1165
Bouillon S, Abril G, Borges AV, Dehairs F, Govers G, Hughes
by a combination of damming and wastewater input
HJ, Merckx R, Meysman FJR, Nyunja J, Osburn C, Mid-
from urban areas. In contrast, D14C (-151 ± 52%) delburg JJ (2009) Distribution, origin and cycling of carbon
and d13C (-8.9 ± 1.6%) values of DIC in the lower in the Tana River (Kenya): a dry season basin-scale survey
Brazos suggest input of carbonate-sourced DIC to the from headwaters to the delta. Biogeosciences 6:2475–2493
BRA (2009) Brazos River Watershed Information Center, Bra-
lower Brazos, despite the absence of carbonate
zos River Authority. http://crpdata.brazos.org/. Accessed
bedrock in the lower Brazos watershed. Oyster shells 20 Aug 2009
and crushed limestone used as aggregate in road BRA (2010) Brazos River Authority, Van M. Walker, GIS
construction in the Gulf coast areas are two likely Coordinator, personal communication, 13 Jan 2010
Brunet F, Gaiero D, Probst JL, Depetris PJ, Lafaye FG, Stille P
sources of carbonate, in addition to naturally occur-
(2005) Delta C-13 tracing of dissolved inorganic carbon
ring marl and pedogenic carbonate in the lower sources in Patagonian rivers (Argentina). Hydrol Process
Brazos. 19:3321–3344

123
Biogeochemistry (2011) 104:275–291 289

Brunet F, Dubois K, Veizer J, Ndondo GRN, Ngoupayou JRN, Hackley PC, Guevara EH, Hentz TF, Hook RW (2009) Ther-
Boeglin JL, Probst JL (2009) Terrestrial and fluvial mal maturity and organic composition of Pennsylvanian
carbon fluxes in a tropical watershed: Nyong basin. coals and carbonaceous shales, north-central Texas:
Cameroon Chem Geol 265:563–572 implications for coalbed gas potential. Int J Coal Geol 77:
Burchuladze AA, Chudy M, Eristavi IV, Pagava SV, Povinec 294–309
P, Sivo A, Togonidze GI (1989) Anthropogenic 14C Helie JF, Hillaire-Marcel C, Rondeau B (2002) Seasonal
variations in atmospheric CO2 and wines. Radiocarbon 31: changes in the sources and fluxes of dissolved inorganic
771–776 carbon through the St. Lawrence River—isotopic and
Burkhardt S, Amoroso G, Riebesell U, Sultemeyer D (2001) chemical constraint. Chem Geol 186:117–138
CO2 and HCO3- uptake in marine diatoms acclimated Hsueh DY, Krakauer NY, Randerson JT, Xu XM, Trumbore
to different CO2 concentrations. Limnol Oceanogr SE, Southon JR (2007) Regional patterns of radiocarbon
46:1378–1391 and fossil fuel-derived CO2 in surface air across North
Cai WJ, Wang Y (1998) The chemistry, fluxes, and sources of America. Geophys Res Lett 34:6
carbon dioxide in the estuarine waters of the Satilla Iwata T, Takahashi T, Kazama F, Hiraga Y, Fukuda N, Honda
and Altamaha Rivers, Georgia. Limnol Oceanogr 43(4): M, Kimura Y, Kota K, Kubota D, Nakagawa S, Nakamura
657–668 T, Shimura M, Yanagida S, Xeu L, Fukasawa E, Hiratsuka
Chakrapani GJ, Veizer J (2005) Dissolved inorganic carbon Y, Ikebe T, Ikeno N, Kohno A, Kubota K, Kuwata K,
isotopic compositions in the Upstream Ganga river in the Misonou T, Osada Y, Sato Y, Shimizu R, Shindo K (2007)
Himalayas. Curr Sci 89:553–556 Metabolic balance of streams draining urban and agricul-
Cole JJ, Caraco NF (2001) Carbon in catchments: connecting tural watersheds in central Japan. Limnology 8:243–250
terrestrial carbon losses with aquatic metabolism. Mar Jones CA (2009) Texas A&M University—Lake Granbury and
Freshw Res 52:101–110 Bosque River assessment, final scientific/technical report.
Cole JJ, Prairie YT, Caraco NF, McDowell WH, Tranvik LJ, Texas Water Resources Institute Technical Report, TR-350
Striegl RG, Duarte CM, Kortelainen P, Downing JA, Kanduc T, Ogrinc N, Mrak T (2007a) Characteristics of sus-
Middelburg JJ, Melack J (2007) Plumbing the global pended matter in the River Sava watershed, Slovenia.
carbon cycle: Integrating inland waters into the terrestrial Isotopes Environ Health Stud 43:369–386
carbon budget. Ecosystems 10:171–184 Kanduc T, Szramek K, Ogrinc N, Walter LM (2007b) Origin
Cronin JG, Follett CR, Shafer GH, Rettman PL (1963) and cycling of riverine inorganic carbon in the Sava River
Reconnaissance investigation of the ground-water watershed (Slovenia) inferred from major solutes and
resources of the Brazos River basin, Texas. Texas Water stable carbon isotopes. Biogeochemistry 86:137–154
Commission Bulletin 6310 Karim A, Veizer J (2000) Weathering processes in the Indus
Doran E (1965) Shell roads in Texas. Geogr Rev 55:223–240 River Basin: implications from riverine carbon, sulfur,
Douglas R, Staines-Urias F (2007) Dimorphism, shell Mg/Ca oxygen, and strontium isotopes. Chem Geol 170:153–177
ratios and stable isotope content in species of Bolivina Keller N, Del-Piero D, Longinelli A (2002) Isotopic compo-
(benthic foraminifera) in the Gulf of California, Mexico. sition, growth rates and biological behaviour of Chamelea
J Foraminiferal Res 37:189–203 gallina and Callista chione from the Gulf of Trieste
EPA (2001) Ambient water quality criteria recommendations. (Italy). Mar Biol 140:9–15
Information supporing the development of state and tribal Krusche AV, Martinelli LA, Victoria RL, Bernardes M, de
nutrient criteria. Rivers and streams in nutrient ecoregion Camargo PB, Ballester MV, Trumbore SE (2002) Com-
X, p 20 position of particulate and dissolved organic matter in a
Finlay JC (2003) Controls of streamwater dissolved inorganic disturbed watershed of southeast Brazil (Piracicaba River
carbon dynamics in a forested watershed. Biogeochemis- basin). Water Res 36:2743–2752
try 62:231–252 Lambs L, Brunet F, Probst JL (2009) Isotopic characteristics of
Garrels RM, Mackenzie FT (1971) Evolution of sedimentary the Garonne River and its tributaries. Rapid Commun
rocks. W.W. Norton, New York Mass Spectrom 23:2543–2550
Gentry DK, Sosdian S, Grossman EL, Rosenthal Y, Hicks D, Levin I, Kromer B, Wagenback D, Minnich KO (1987) Carbon
Lear CH (2008) Stable isotope and Sr/Ca profiles from the isotope measurements of atmospheric CO2 at a coastal
marine gastropod Conus ermineus: testing a multiproxy station in Antarctica. Tellus 39B:89–95
approach for inferring paleotemperature and paleosalinity. Livingstone DA (1963) Chemical composition of rivers and
Palaios 23:195–209 lakes. U.S. Geol. Survey Prof Paper: 44-G
Green RE, Bianchi TS, Dagg MJ, Walker ND, Breed GA Lohrenz SE, Cai WJ (2006) Satellite ocean color assessment of
(2006) An organic carbon budget for the Mississippi air-sea fluxes of CO2 in a river-dominated coastal margin.
River turbidity plume and plume contributions to air-sea Geophys Res Lett 33:4
CO2 fluxes and bottom water hypoxia. Estuar Coasts 29: Longworth BE, Petsch ST, Raymond PA, Bauer JE (2007)
579–597 Linking lithology and land use to sources of dissolved and
Griffith DR, Barnes RT, Raymond PA (2009) Inputs of fossil particulate organic matter in headwaters of a temperate,
carbon from wastewater treatment plants to US rivers and passive-margin river system. Geochim Cosmochim Acta
oceans. Environ Sci Technol 43:5647–5651 71:4233–4250
Gu B, Schelske CL, Hodell DA (2004) Extreme 13C enrich- Matson GC, Hopkins OB (1917) The Corsicana oil and gas
ments in a shallow hypereutrophic lake: implications for field Texas. Contribution to economic geology, Part II.
carbon cycling. Limnol Oceanogr 49:1152–1159 Bulletin 661-F:211–252

123
290 Biogeochemistry (2011) 104:275–291

Mayorga E, Aufdenkampe AK, Masiello CA, Krusche AV, Raymond PA, Bauer JE, Caraco NF, Cole JJ, Longworth B,
Hedges JI, Quay PD, Richey JE, Brown TA (2005) Young Petsch ST (2004) Controls on the variability of organic
organic matter as a source of carbon dioxide outgassing matter and dissolved inorganic carbon ages in northeast
from Amazonian rivers. Nature 436:538–541 US rivers. Mar Chem 92:353–366
McConnaughey TA, Gillikin DP (2008) Carbon isotopes in Richey JE, Melack JM, Aufdenkampe AK, Ballester VM, Hess
mollusk shell carbonates. Geo Mar Lett 28:287–299 LL (2002) Outgassing from Amazonian rivers and wet-
McFarland AMS, Hauck LM (2001) Determining nutrient lands as a large tropical source of atmospheric CO2.
export coefficients and source loading uncertainty using Nature 416:617–620
in-stream monitoring data. J Am Water Resour Assoc 37: Riebschleager KJ, Karthikeyan R (2008) Bacterial impairment
223–236 assessment for Lake Granbury watershed. Report as of
McFarland A, Kiesling R, Pearson C (2001) Characterization FY2007 for 2007TX266B
of a Central Texas Reservoir with emphasis on factors Rightmire CT (1967) A radiocarbon study of the age and origin
influencing algal growth. Texas Institute for Applied of caliche deposits. M.A. thesis, University of Texas,
Environmental Research, Tarleton State University, Ste- Austin
phenville, Texas, TR0104 Roelke DL, Brooks BW, Grover JP (2007) Water quality
McNichol AP, Jones GA, Hutton DL, Gagnon AR, Key RM program for Lake Granbury Texas, 4th quarter report.
(1994) The rapid preparation of seawater sigma-CO2 for http://lakegranbury.tamu.edu/docs/projectupdate_2007-
radiocarbon analysis at the National Ocean Sciences AMS 10-30_ga.pdf. Accessed 10 Aug 2009
facility. Radiocarbon 36:237–246 Rotatore C, Colman B, Kuzma M (1995) The active uptake of
Meybeck M (1993) Riverine transport of atmospheric carbon— carbon dioxide by the marine diatoms Phaeodactylum
sources, global typology and budget. Water Air Soil Pollut tricornutum and Cyclotella sp. Plant Cell Environ 18:
70:443–463 913–918
Mook WG, Bommerso JC, Staverma Wh (1974) Carbon iso- Ruppert LF, Kirschbaum MA, Warwick PD, Flores RM, Af-
tope fractionation between dissolved bicarbonate and folter RH, Hatch JR (2002) The US Geological Survey’s
gaseous carbon-dioxide. Earth Planet Sci Lett 22:169–176 national coal resource assessment: the results. Int J Coal
Moore RC, Plummer FB (1922) Pennsylvanian stratigraphy of Geol 50:247–274
north central Texas. J Geol 30:18–42 Sarmiento JL, Sundquist ET (1992) Revised budget for the
National Center for Earth Resources Observation and Science oceanic uptake of anthropogenic carbon-dioxide. Nature
and U.S. Geological Survey (2005) Conterminous United 356:589–593
States Land Cover 1992—200-meter resolution. http:// Shah SD, Houston NA, Braun CL (2007) Hydrogeologic
nationalatlas.gov/atlasftp.html characterization of the Brazos River alluvium aquifer,
NOAA (2009) National climatic data center. http://www.ncdc. Bosque County to Fort Bend County, Texas. U.S. Geo-
noaa.gov. Accessed 20 Aug 2009 logical Survey
Nordt LC, Hallmark CT, Wilding LP, Boutton TW (1998) Skjemstad JO, Reicosky DC, Wilts AR, McGowan JA (2002)
Quantifying pedogenic carbonate accumulations using Charcoal carbon in US agricultural soils. Soil Sci Soc Am
stable carbon isotopes. Geoderma 82:115–136 J 66:1249–1255
Nordt L, Orosz M, Driese S, Tubbs J (2006) Vertisol carbonate Stiller M, Magaritz M (1974) Carbon-13 enriched carbonate in
properties in relation to mean annual precipitation: interstitial waters of Lake Kinneret sediments. Limnol
implications for paleoprecipitation estimates. J Geol 114: Oceanogr 19:849–853
501–510 Stiller M, Rounick JS, Shasha S (1985) Extreme carbon-
Oh NH, Raymond PA (2006) Contribution of agricultural isotope enrichments in evaporating brines. Nature 316:
liming to riverine bicarbonate export and CO2 sequestra- 434–443
tion in the Ohio River basin. Glob Biogeochem Cycles Stoeser DB, Shock N, Green GN, Dumonceaux GM, Heran
20:17 WD (2005) Geologic map database of Texas: U.S. Geo-
Oren A, Gurevich P, Anati DA, Barkan E, Luz B (1995) A logical Survey open data series. http://pubs.usgs.gov/
bloom of Dunaliella parva in the Dead Sea in 1992: ds/2005/170
biological and biogeochemical aspects. Hydrobiologia 297: Stricklin FL (1961) Degradational stream deposits of the
173–185 Brazos River, central Texas. Geol Soc Am Bull 72:19–36
Phillips JD (2007) Geomorphic context, constraints, and Stumm W, Morgan JJ (1996) Aquatic chemistry: chemical
change in the lower Brazos and Navasota Rivers, Texas. equilibria and rates in natural waters. Wiley, New York
Phase 2 of the project ‘‘Field data collection in support of Telmer K, Veizer J (1999) Carbon fluxes, pCO2 and substrate
geomorphic classification of the lower Brazos and Nava- weathering in a large northern river basin, Canada: carbon
sota Rivers’’, Texas Water Development Board report isotope perspectives. Chem Geol 159:61–86
Raymond PA, Oh NH (2007) An empirical study of climatic Tewalt SJ, Kinney SA, Merrill MD (2008) GIS representation
controls on riverine C export from three major U.S. of coal-bearing areas in North, Central, and South
watersheds. Glob Biogeochem Cycles 21:GB2022. doi: America: U.S. Geological Survey open-file report 2008-
10.1029/2006GB002783 1257. http://pubs.usgs.gov/of/2008/1257/
Raymond PA, Caraco NF, Cole JJ (1997) Carbon dioxide Titi H, Rasoulian M, Martinez M, Becnel B, Keel G (2003)
concentration and atmospheric flux in the Hudson River. Long-term performance of stone interlayer pavement.
Estuaries 20:381–390 J Transp Eng ASCE 129:118–126

123
Biogeochemistry (2011) 104:275–291 291

Trumbore S (2000) Age of soil organic matter and soil respi- composition, and stable isotope signature of the giant
ration: radiocarbon constraints on belowground C deep-sea oyster Neopycnodonte zibrowii sp n. from the
dynamics. Ecol Appl 10:399–411 NE Atlantic. Deep Sea Res I Oceanogr Res Pap 56:
Trumbore S, Da Costa ES, Nepstad DC, De Camargo PB, 374–407
Martinelli L, Ray D, Restom T, Silver W (2006) Yao GR, Gao QZ, Wang ZG, Huang XK, He T, Zhang YL,
Dynamics of fine root carbon in Amazonian tropical Jiao SL, Ding J (2007) Dynamics of CO2 partial pressure
ecosystems and the contribution of roots to soil respira- and CO2 outgassing in the lower reaches of the Xijiang
tion. Glob Change Biol 12:217–229 River, a subtropical monsoon river in China. Sci Total
Valastro S, Davis EM, Rightmir Ct (1968) University of Texas Environ 376:255–266
at Austin radiocarbon dates 6. Radiocarbon 10:384–401 Zeng F-W, Masiello CA (2010) Sources of CO2 evasion from
Wachniew P (2006) Isotopic composition of dissolved inor- two subtropical rivers in North America. Biogeochemis-
ganic carbon in a large polluted river: The Vistula. Pol try. doi:10.1007/s10533-010-9417-6
Chem Geol 233:293–308 Zhang J, Quay PD, Wilbur DO (1995) Carbon-isotope frac-
Wachniew P, Rozanski K (1997) Carbon budget of a mid- tionation during gas-water exchange and dissolution of
latitude, groundwater-controlled lake: isotopic evidence CO2. Geochim Cosmochim Acta 59:107–114
for the importance of dissolved inorganic carbon recy- Zhang SR, Lu XX, Sun HG, Han JT, Higgitt DL (2009) Major
cling. Geochim Cosmochim Acta 61:2453–2465 ion chemistry and dissolved inorganic carbon cycling in a
Wang FS, Wang YC, Zhang J, Xu H, Wei XG (2007) Human human-disturbed mountainous river (the Luodingjiang
impact on the historical change of CO2 degassing flux in River) of the Zhujiang (Pearl River), China. Sci Total
River Changjiang. Geochem Trans 8:7 Environ 407:2796–2807
Wisshak M, Correa ML, Gofas S, Salas C, Taviani M, Ja-
kobsen J, Freiwald A (2009) Shell architecture, element

123