Biogeochemistry of Carbon in the Amazonian Floodplains - Ainfo

30 Pages·2011·2.14 MB·English

EarthInteractions d Volume15(2011) d PaperNo.4 d Page1
Copyright(cid:2)2011,Paper15-004;11902words,10Figures,0Animations,6Tables.
http://EarthInteractions.org
Biogeochemistry of Carbon
in the Amazonian Floodplains
over a 2000-km Reach: Insights
from a Process-Based Model
Vincent Bustillo,*,1 Reynaldo Luiz Victoria,#
Jose Mauro Sousa de Moura,@ Daniel de Castro Victoria,&
Andre Marcondes Andrade Toledo,** and Erich Colicchio11
1Centro de Energia Nuclear na Agricultura, Laborato´rio de Geoprocessamento e Trata-
mento de Imagens, Piracicaba, Brazil, and Universite´ Franxcois Rabelais de Tours,
UMR CNRS/INSU 6113 Institut des Sciences de la Terre d’Orle´ans, Universite´
d’Orle´ans, Tours, France
#Centro de Energia Nuclear na Agricultura, Laborato´rio de Geoprocessamento e Trata-
mento de Imagens, and USP–ESALQ, NUPEGEL, Piracicaba, Brazil
@Universidade Federal do Oeste do Para, Santarem, Brazil
&Embrapa Monitoramento por Sate´lite, Campinas, Brazil
**Universidade Federal de Mato Grosso, Rondono´polis, Brazil
11Universidade Federal do Tocantins, AgroUnitins, Palmas, Brazil
Received 13 March 2010; accepted 16 September 2010
ABSTRACT: The influence of Amazonian floodplains on the hydrological,
sedimentary,andbiogeochemicalriverbudgetwasinvestigatedovera2000-km
reach. A process-based model relying on the closure of chemical fluxes and
isotopicsignalswasimplemented.Onaverageforthewholestudiedreach,the
overall fluxes of carbon associated with mineralization and aquatic photosyn-
thesiswereestimatedto35.7and15.3TgCyr21,respectively.Almost57%of
the carbon sequestrated by photosynthesis comes from aerial sources (flooded
* Corresponding author address: Vincent Bustillo, Universite´ Franxcois Rabelais de Tours,
ParcGrandmont,UFR Sciences et Techniques, Baˆtiment E,37200Tours, France.
E-mail address:[email protected]
DOI:10.1175/2010EI338.1
EarthInteractions d Volume15(2011) d PaperNo.4 d Page2
forest); the remaining 43% resulted from aquatic sources (va´rzea grasses and
phytoplankton).Theprocessratessubstantiallyfluctuateovertheannualcycle,
depending particularly on the extension of flooded area and on the river–
floodplain connectivity. As the river level declines, the drastic decrease of
turbidity and the lower supply of carbon substrates promote autotrophy to the
detriment of heterotrophy, leading to substantial changes of pH and gaseous
equilibria in the river water. The main consequences are (i) the side-chain
oxidation of dissolved organic matter leading to the concomitant rises of the
carbon to nitrogen atomic ratio and nitrate contents and (ii) the sorption of
hydrophobic humic acids, which fractionate 13C and thus lead to 13C-depleted
particulate organic matter (fine fraction) compared to remaining dissolved or-
ganic matter. As the river flow rises, the heterotrophy prevails over autotrophy
and this tends to attenuate the chemical signature imprinted by the latter. The
significantcontributionofaerialautochthonoussourcestothebudgetofcarbon
indicates that the fluxes of mineralization are sustained by the net primary
productionofrivercorridors.Thevariableextensionofsubmergedareasdefines
the proportions of CO exported by the river and released to the atmosphere.
2
The rate of CO outgassing on the studied reach (18.8 Tg C yr21) represents
2
about50%oftheincomingdissolvedinorganiccarbonflux.Therateofmethane
emissionisestimatedas2.2TgCyr21andthatofdenitrificationisestimatedas
0.87 Tg N yr21, representing 1.5 times the flux of dissolved inorganic nitrogen
(DIN)exportedbytheAmazonRiveratthestationofO´bidos(0.64TgNyr21).
KEYWORDS: Amazon River;Floodplains; Biogeochemical cycles; Carbon;
CO outgassing
2
1. Introduction
Thispaperfocusesonthequantificationofprocessesdrivingthebiogeochemical
budget of carbon within the Amazonian floodplains (FP) over a 2000-km reach
extending between the stations of Vargem Grande (VG; upstream boundary) and
O´bidos (O´bi; downstream boundary). The Amazonian forest is commonly con-
sidered to be a very significant sink of carbon because of the storage capacity of
soils (Bernoux et al. 2002) and biomass (Brown and Lugo 1992). Scientific in-
vestigations on the Amazon River demonstrate that flooded areas constitute very
important sources of carbonaceous greenhouse gases, involving not only CO
2
(Richeyetal.1988;Richeyetal.2002)butalsoCH (Devoletal.1988).Thestrong
4
seasonality of the Amazon River discharge promotes changes in stage height of as
much as 10 m (Richey et al. 1989; Mayorga and Aufdenkampe 2002), thus de-
termining verysignificant changes inthe extent of flooded areas within the central
lowlands: from 100 000 km2 in November to 350 000 km2 in May (Richey et al.
2002). According to these authors, the rivers of Central Amazonia (quadrant of
1.77 3 106 km2)return about 210 Tg Cyr21as carbon dioxide tothe atmosphere.
Mayorga et al. (Mayorga et al. 2005) showed that this outgassing flux of carbon
dioxide was driven by the respiration of terrestrial organic carbon that is less than
5 years old and mainly composed of autochthonous sources (va´rzea grasses), as
stated previously by Quay et al. (Quay et al. 1992).
Recently, the River Basin Organic Matter and Biogeochemistry Synthesis
(ROMBUS) model was implemented by Richey et al. (Richey et al. 2004) to
decipher the complex signals from dissolved organic matter (DOM) of variable
molecular size. Within ROMBUS, each of the organic and inorganic carbon pools
EarthInteractions d Volume15(2011) d PaperNo.4 d Page3
Figure 1. Map of the Amazon River basin and location of the main tributaries. The
sampling stations are 1) VG, 2) Santo Antonio do Icxa, 3) Xibeco (Xib), 4)
Tupe (Tup), 5) Jut, 6) Ita, 7) Anori (Ano), 8) Man, 9) Sa˜o Jose da Amatari
(SJA), 10)Pau, and 11) O´bi.
is represented by state variables that characterize the nitrogen-to-carbon ratio (for
the OM pools), d13C signature, and age (via D14C). A survey of the spatial and
temporalvariabilityinD14Csignaturesofdifferentcarbonfractionswasperformed
1) to investigate the biogeodynamics of carbon in Amazonian river of differing
types and sizes and 2) to constrain better age, turnover, and pathways of these
principalcarbonfractionsfortheimplementationandvalidationofanewROMBUS
model. In this context, the implementation of a process-based model, including all
carbon fractions, is potentially of great interest to quantify the magnitude of pro-
cesses driving the biogeodynamics of carbon within the river system and at its
interfaces with the atmosphere and with the sediment.
1.1. Preliminary work
Biogeochemical mass balances were established (Bustillo et al. 2010) by com-
paring incoming and outgoing mass balances at 10 monitoring stations (Figure 1)
located along the studied fluvial reach. Six end-member mixing models were im-
plemented to revisit the interpretations given to the seasonally structured excesses
EarthInteractions d Volume15(2011) d PaperNo.4 d Page4
and deficits with respect to inputs and outputs. By comparing the outputs of many
different models designed for different purposes, the nature and the magnitude of
processeslinkingwaterandbiogeochemicalbudgetsoftheAmazonianfloodplains
were clarified. Each one provides a specific insight on the soil–river system dy-
namics by coupling hydrological, sedimentary, and biogeochemical budgets. The
comparison of the model outputs, and the analysis of their reciprocal consistency,
enabledustodeciphermuchmoreassuredlythenatureandmagnitudeofprocesses
occurringinthefloodplains.Thesearemixingmodelsbasedon1)variableregional
sources with and without correction of inputs by small tributaries; 2) variable
hydrological sources with three end members to determine their individual com-
positional evolution, with contrasted response depending on hydrograph stage,
throughout their course in the floodplains; 3) variable hydrological sources with
threeendmembers,includingacorrectiononthebaseflowtoaccountforin-stream
biogenic transformations; and 4) mixed approaches combining the regional vari-
ability offivechemical signals(between river basins) andthevariability related to
thehydrologicalsource(betweencontributingrunoffsorendmembers),takinginto
considerationthedefaultsoffloodplainswaterbalance.Thesemodelsenabledusto
identify the main factors (hydrological source, water budget of the floodplains,
nature of hydrobiological pattern; i.e., photosynthesis versus mineralization, air–
water gaseous exchanges, etc.) controlling the biogeochemical and sedimentary
budgets of Amazonian floodplains.
This study revealed that most of the chemical baseline of the Amazon River
basin was acquired before the studied 2000-km Amazonian reach. However, the
tightconnectionbetweenthehydrographstageoftheriverandthechemicalsignals
provided insightful information on the dynamics of its floodplains. The chemical
expression of biotic and abiotic processes occurring in the Amazonian floodplains
can be more easily identified during falling waters. It appears delayed in time
compared to the maximum extension of submerged area because the alternating
water circulation direction (filling versus emptying) between the main channel
(MC) and the adjacent floodplains determines delayed emptying of floodplains
during falling waters. It results also in a longer residence time in the hydrograph
network, which strengthens the rate of transformation of transiting materials and
solutes. Biotic and biologically mediated processes tend to accentuate changes in
river water chemistry initiated upstream, in each subbasin, along river corridors,
indicatingthatprocessesoperatingdownstreamprolongthosefromupstream(e.g.,
floodplains of the large tributaries). Conversely, the flood wave propagation tends
to lessen the seasonal variability as a result of thewater storage in the floodplains,
which admixes waters of distinct origins (in time and space). The morphology of
floodplains,determiningthedepositionandthediagenesisofthesedimentsaswell
as the variable extension of submerged areas or the chronology of floodplains
storage/emptying,appearstobethemainfactorcontrollingthebiogeodynamicsof
the floodplains. By coupling classical end-member mixing models (providing in-
sight on hydrological source) with a variable regional contribution scheme, rele-
vant information on the biogeochemical budget of the Amazonian floodplains
could be captured. However, the identification, magnitude, and chronological
succession of major processes driving the biogeochemistry of carbon were not
completely elucidated. Therefore, this paper aims to bridge the knowledge gap by
testing a process-based model constrained by chemical and isotopic data.
EarthInteractions d Volume15(2011) d PaperNo.4 d Page5
1.2. Organization of the paper
Inlightoftheresultsofthesixmixingmodels,twomainissuesdealingwiththe
biogeochemistry and hydrology of floodplains are addressed:
1) nature, magnitude, and seasonal fluctuations of biotic processes (autotro-
phy versus heterotrophy) and
2) nature and magnitude of abiotic processes involving notably the sorption
of dissolved organic matter and the gaseous exchanges at the air–water
interface.
This paper aims to investigate more in detail these two topics, which are in-
trinsically related and determine most of the biogeochemical budget of the Ama-
zonianfloodplainswithrespecttocarbonandnitrogen.Toachievethisobjective,a
process-based model describing the balance of the chemical and isotopic budgets
ofC,N,andOwasimplemented.Thismodelreliesontheclassicalsurveyofriver
waterchemistrythroughtheCarbonintheAmazonRiverExperiment(CAMREX).
Specifically, the process-based model requires the assimilation of appropriate in-
fieldstudiesorremotesensingdatatoconstrainchemicalandisotopicbudgets.The
new modeling approach, implemented here, provides an original and pertinent
frameworktoappreciatequantitativelyandqualitativelythebiogeochemicalbudget
ofcarbonovera2000-kmreach.Overthelastthreedecades,aconsiderableamount
of field data were collected, but few studies using the whole dataset of the
CAMREX project are based on a modeling approach. We propose here to gather
theveryvaluableamountofinformationprovidedbytheCAMREXprojectandits
extensions [e.g., Large-scale Biosphere-Atmosphere Experiment in Amazonia
(LBA)] to implement a comprehensive process-based model of carbon biogeo-
chemistry at a very large scale.
The main originality of this study resides in the implementation of a modeling
approach, constrained by a very large amount of chemical data. This model is
expected to assimilate and synthesize this information, thus providing unique in-
sight and quantified issues on the major processes that drive carbon and nitrogen
biogeochemical budgets over a 2000-km reach.
2. Study area and available dataset
TheAmazonRiveristhelargestriverintheworldbydischargewithatotalriver
flow greater than the next eight largest rivers combined. Its drainage basin is the
largest in the world, covering about 6 915 000 km2 at its mouth. It gathers its
waters from 58N to 208S. The quantity of water released by the Amazon to the
Atlantic Ocean might exceed 300 000 m3 s21 during the rainy season (Mayorga
and Aufdenkampe 2002). The Amazon is responsible for about 20% of the total
volume of water entering the oceans worldwide. The confluence of Andean
tributaries in the Amazonian plain gathers a considerable amount of water and
generates huge discharges. The flatness of the Amazonian valley (2.1 cm km21
betweenVargemGrandeandO´bidos)promotesthelarge-scalefloodingofalluvial
plains. During the wet season, some parts of the Amazon might reach 100 km in
width. The area covered by the water of the Amazon River and its tributaries is
multiplied by 3 over the course of a year: on average, during the dry season,
EarthInteractions d Volume15(2011) d PaperNo.4 d Page6
100 000 km2 are flooded, whereas thewater-covered area rises to 350 000 km2 in
the wet season. Because of its vast dimensions, the Amazonian FP constitutes an
ideal open-sky laboratory for studying the dynamics of sediments, carbon species,
biogenic species, and solutes, whose concentrations and mass balance change
during their course throughout the floodplains. The study focuses on the 2000-km
fluvialreachextendingbetweenVargemGrande(upstreamarea5106 km2),close
´
to the frontier between Peru and Brazil, and Obidos (Amazon River, upstream
area 5 4.62 3 106 km2), the outlet of the studied reach located 700 km landward
from the mouth and where tidal influence remain low. Nine major tributaries join
the Amazon main stem over this section: Ixca, Japura´, and Negro on the northern
side; Juta´ı, Jurua´, Purus, Negro, and Madeira on the southern side; and Solimo˜es
from the Peruvian Andes. A view of the Amazon River basin upstream from
´
Obidos, including the hydrograph network and the location of CAMREX’s mon-
itoring station along the Amazon River main stem, is presented in Figure 1.
The dataset used in this study is from the CAMREX project (1982–91). The
objective of CAMREX project was to define by mass balances and direct mea-
surements those processes that control the distribution of bioactive elements
(C, N, P, and O) in the main stem of the Amazon River in Brazil. Representative
flux-weighted water samples for comprehensive chemical analysis and for taking
ratemeasurementswereobtainedover18differentsiteswithina2000-kmreachof
the Brazilian Amazon main stem, including major intervening tributaries. The
CAMREX dataset represents a time series unique in its length and detail for very
large river systems. Samples were collected on 13 different cruises (1982–91)
during contrasting hydrographic stages, consisting in depth-integrated, discharge-
weighted composite water samples (20–40 L). In this study, the samples from
cruises9–13werenotusedbecausethechemicalparametersrequiredformodeling
purposes were not all analyzed. The sampling and analytical procedures used
during CAMREX were presented by Richey et al. (Richey et al. 1986), Hedges
et al. (Hedges et al. 1986a), and Devol et al. (Devol et al. 1987). These will
therefore be described only briefly. Materials in water were first separated into
fractions by size. The three size classes (coarse, fine, and dissolved) display very
distinct transport dynamics, degradation patterns, and compositional characteris-
tics (Mayorga and Aufdenkampe 2002). Coarse particulate, fine particulate, and
dissolved fractions were thus operationally defined using a sieve pore size of
63 mm to separateparticulatefractions:coarsesuspendedsediments(CSS)versus
fine suspended sediments (FSS) and filter pore sizes of 0.45 mm to isolate dis-
solved constituents. FSS was isolated by continuous-flow centrifugation. FSS
consist of clays and silts, material between 0.45 and 63 mm in size as defined by
CAMREX (Mayorga and Aufdenkampe 2002). Elemental compositions, reported
as weight percentages of organic carbon and total nitrogen, were measured with a
Carbo Erba model 1106 carbon, hydrogen, and nitrogen (CHN) analyzer within
each size fraction. Related concentrations [particulate organic carbon (POC) and
particulate organic nitrogen (PON) within fine (POCF and PONF) and coarse
(POCC and PONC) size fractions] were then calculated as the product of weight
measurement and suspended concentrations. Remaining material (,0.45 mm)
constitutesthedissolvedfraction(DOM),whosecarboncontent[dissolvedorganic
carbon (DOC)] was measured using a potassium persulfate oxidation procedure.
The concentration of dissolved inorganic carbon (DIC; mM) was established from
EarthInteractions d Volume15(2011) d PaperNo.4 d Page7
pH and alkalinity titration measurements. The d13C values for DIC, POCF, and
POCCarereportedversusmarinecarbonate[PeeDeeBelemnite(PDB)standard].
The measurements of 13C:12C were madewith two mass spectrometers, a Nuclide
6–60 and Finnigan MAT 251.
Oneof thegreatest strengths ofthe CAMREX database resides inthevery wide
spectrum of analyzed parameters: major ions (Na1, K1, Ca21, Mg21, HCO2,
3
SO22,andCl2;concentrationsgiveninmM);organicspecies[DOC(mgL21;size,
4
0.45 mm), DIC (mM), and particulate organic carbon (POC; mg L21)]; sus-
pended sediments (mg L21); dissolved organic nitrogen (DON; mM); NO2 (mM);
3
NH1 (mM); dissolved phosphorus (PO4; mM); total phosphorus (Pt; mg L21);
4
particulate organic nitrogen (PON; mg L21); dissolved gases (O and CO ; mM);
2 2
pH,alkalinity,andisotopicdataforriverwater(d18O;‰SMOW);andcarbonspecies
fd13C (‰; PDB) for dissolved inorganic carbon [DIC] 5 [HCO2] 1 [CO ] 1
3 2
[CO22],POCF,andPOCCg.Notetheabsenceofdataford13C(DOC)andNO2,
3 2
which would have been ideally required to better constrain carbon and nitrogen
budgets. Available data correspond to cross-sectionally integrated samples. Ac-
cordingtoQuayetal.(Quayetal.1992)andbasedonreplicateanalyses(n53)of
thesamesample,theprecisionsofthemeasurements(61standarddeviation)were
61.5% for fine materials [fine particulate OM (FPOM) and FSS]; 63.0% for
coarse materials [coarse particulate OM (CPOM) and CSS]; 63.0% for dissolved
inorganic carbon; and 60.1‰ for d13C of FPOM, CPOM, and DIC.
Thedataset,extractedfrompre-LBACD-ROM(Richeyetal.2008),constitutes,
untilthatdate,thebasisofmorethan130CAMREXpublicationsthathavefocused
on understanding physical and biogeochemical dynamics throughout the basin
using a large variety of approaches.
3. Modeling strategy
The modeling strategy consists in assessing the biogeochemical balance of
carbon and nitrogen in the floodplains by constraining chemical and isotopic
budget of carbon and nitrogen species. The imbalances between incoming (IN*)
andoutgoing(OUT)fluxesareinterpretedastheresultofseveralprocessesarising
in the Amazon River. The composition of the incoming flux Ci,j is given by the
k,IN*
discharge-weighted mixture of the waters coming from the n major tributaries
located upstream from the jth sampling station (concentration Ci,j and total dis-
k,M1
charge upstream from the nth confluent Qj 5 Pj5nQj ) and those from un-
t,k j51 t,k
gauged areas (concentration Ci and total discharge Qj ),
k,UA k,UA
Ci,j 3Qj 1Ci 3Qj
Ci,j 5 k,M1 t,k k,UA k,UA. (1)
k,IN* Qj 1Qj
t,k k,UA
The composition of the flow supplied by ungauged areas (Ci ) is approximated
k,UA
by the composition of the Juta´ı River (76 000 km2), which was monitored along
the eight cruises of the CAMREX project. This choice is consistent because lith-
ological and hydroclimatological features of the Juta´ı River basin are representa-
tive of those of ungauged areas (Tardy et al. 2009) and because the chemical
composition and variability of the Juta´ı River are in good agreement with those of
black water rivers draining Central Amazonian area [e.g., see dataset from the
EarthInteractions d Volume15(2011) d PaperNo.4 d Page8
Figure2.Overview of major processes driving Cand Nbiogeodynamics in rivers.
HydrologyandGeochemistryoftheAmazonBasin(HiBAm)projectattheoutletof
theTrombetasRiver].ThedischargeoftheungaugedareaQj isestimatedbythe
k,UA
differencebetweentheoutgoingdischargedownstreamfromthenthmajortributary
(Qj ) and the sum of discharge of the n major tributaries (Pj5nQj ), leading to
k,OUT j51 t,k
j5n
X
Qj 5Qj (cid:2) Qj . (2)
k,UA k,OUT t,k
j51
The imbalances between IN* and OUT fluxes, noted DF(Cj ), are calculated as
i,k
follows:
DF(Cj )5[Ci,j (cid:2)Ci,j ]3Qj , (3)
i,k k,OUT k,IN* k,OUT
where Ci,j is the concentration of the ith chemical specie observed at the jth
k,OUT
sampling station for the kth cruise.
The process-based description of C and N biogeochemical cycles implies 14
riverprocesses(Figure2andTable1):nonaquatic(aerial)photosynthesisfollowed
by the incorporation of newly formed organic matter in the aquatic environment
(flux noted FV); aquatic photosynthesis removing DIC and dissolved inorganic
nitrogen(DIN)toformPOCCandPONC(fluxnotedFP );aquaticphotosynthesis
C
removing DIC and DIN to form POCF and PONF (flux noted FP ); physical size
F
reduction of coarse particulate organic matter CPOM through mechanical means
(flux noted FM )releasingPOCFandPONF without consuming O ;physical size
c 2
reduction of fine particulate organic matter FPOM (flux noted FM ) releasing
F
EarthInteractions d Volume15(2011) d PaperNo.4 d Page9
Table 1. Reactants and products of each process considered by the model and
stoichiometry of the reactions. The secondary reactants and products (e.g., H O)
2
are not shown. Stoichiometric coefficients are not specified for Ox (oxidative re-
actionsthatdonotreleaseCO )becausethecontributionofeachreactantonthe
2
finalbalanceofthisprocessremainsunknown.Otherwise,weassume,forinstance,
that1CPOMrepresentsonemoleof(carbon1nitrogen)intheCPOMfractionand
that 1 CPOC stands for 1 mol of C in the CPOM fraction. The indices (a) and (w)
attached to CO and O correspondto air and water, respectively.
2 2
Process Reactants Products
P CO (w) CPOM1O (w)
C 2 2
P CO (w) FPOM1O (w)
F 2 2
V CO (a) CPOC
2
M CPOM FPOM
C
M FPOM DOM
F
M DOM1O (w) CO (w)
D 2 2
N DON DIN
D DOM;NO2 5/4CO (w)
3 2
CH DOM CH (a)
4 4
R CO (w) CO (a)
2 2
O O (w) O (a)
2 2 2
Ox NH1;O (w); NO2;NO2;
4 2 3 2
DOM(HMW) DOM(LMW)
S DOC FPOC
S DON FPON
N
DOC andDON without consumingO ; DOM sorption removing DOC (flux noted
2
FS) and DON (flux noted FS ) to form POCF and PONF, respectively; aerobic
N
decay of DOC (flux noted FM ) consuming O and releasing DIC; oxidation of
D 2
DON (flux noted FN) consuming O and releasing DIN; methanogenic pathways
2
removing DOM species (flux noted FCH ); denitrification (flux noted FD) re-
4
moving DIN and forming DIC; and the oxidation of inorganic and organic sub-
strates (flux noted FOx) consuming O without releasing CO . Moreover, the
2 2
balance of gaseous exchanges between at the air–water interface is controlled by
CO outgassing (flux noted FR) and by the invasion of O from the atmosphere
2 2
(flux noted FO ). The deconvolution of processes controlling the chemical (C and
2
N species) and isotopic imbalances is performed by solving the system of linear
equations presented in Table 2. To achieve this purpose, several required variables
must be established and/or calibrated. The isotopic composition of respired CO ,
2
noted d13C(R), is determined taking into account 1) the chemical fractionation as-
sociated to the speciation between HCO2 and CO , d13C(CO ) 5 d13C(HCO2) 2
3 2 2 3
8.97 (Faure 1977) and 2) the physical fractionation associated to the outgassing,
d13C(R)5d13C(CO )23.4(Craig1954);3)thepH-dependentspeciationofDIC,
2
pH 5 6.367 1 log[HCO2] 2 log[CO ] for T 5 288C [mean temperature of the
3 2
Amazon River, inferred from the quality survey of the Brazilian Water Agency
(ANA)], with [HCO2] 1 [CO ] 5 [DIC], leading to
3 2
1
d13C(R)5[0.53d13C(DIC) 10.53d13C(DIC) ](cid:2)
IN* OUT 11106.367(cid:2)pH
38.97(cid:2)3.4. (4)
EarthInteractions d Volume15(2011) d PaperNo.4 d Page10
n (cid:3)
o
carb FPcFPFFSFMCFMF FMD FV FOxFCH4 FDFSNFN
½
o
t 3
e (cid:3)
v
ati 00000 0 0 00 1210
el 00000 0 0 00 0211
r
s 40000 0 0 00 100
e 5/ 2
x )
u H4
fl C
g 021000 0 0 0C( 000
n 3
oi 1d
g 2
ut 00000 0 0 210 000
o
vs. 00010 0 VC() 00 000
g 13
n d
mi )D M)D
ofinco 1210013dMC( 0 0 21132dC( 000
alances 012100 132dMC()F 0 013dMC()F 0MN/C()F2PN/C()F
mb )C M)C )C
ei 001210 MC( 3C( 00 00MC(
th 13d 12d N/
o
ssest 021100 MC()S 0 03MC()S 000
e 13 1d
c d 2
riverpro 21010132dPC()F 13PdC()F 0 10 2PN/C()F0PN/C()F
ating14 21001132dPC()C 0 13PdC()C 10 2PN/C()C00
el ½
r
s 5
n
o (cid:3)
ati R)
u C(
req FR000133d 0 0 FO20 000
a
nees. FR
lici ½
ofe 1
mp (cid:3)
s
e2.Systenitrogen F(DIC)F(DOC)FPOCF()F(POCC)F3DIC13dCDIC)FPOCF313dCPOCF)F3POCC13dCPOCC)F)(O2F3(DOC13dCDOC)FDIN()F(DON)F(PONF)
bld DDDDD( D( D( DD DDD
Taan ½