This article was downloaded by: [81.191.120.38]On: 14 June 2013, At: 13:58Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: MortimerHouse, 37-41 Mortimer Street, London W1T 3JH, UK

Marine Biology ResearchPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/smar20

On seasonal changes of the carbonate system in theBarents Sea: observations and modelingEvgeniy V. Yakushev a & Kai Sørensen aa Norwegian Institute of Water Research (NIVA) , Oslo , Norway

To cite this article: Evgeniy V. Yakushev & Kai Sørensen (2013): On seasonal changes of the carbonate system in theBarents Sea: observations and modeling, Marine Biology Research, 9:9, 822-830

To link to this article: http://dx.doi.org/10.1080/17451000.2013.775454

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form toanyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss, actions,claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly orindirectly in connection with or arising out of the use of this material.

ORIGINAL ARTICLE

On seasonal changes of the carbonate system in the Barents Sea:observations and modeling

EVGENIY V. YAKUSHEV* & KAI SØRENSEN

Norwegian Institute of Water Research (NIVA), Oslo, Norway

AbstractThe role of seasonality of organic matter (OM) production and decay in the seasonal changes of the carbonate system wasstudied on the basis of data received for a transect from Tromsø to Spitsbergen with a Ferrybox-equipped vessel. A 2Dsimplified vertical model was used for parameterization of the hydrophysical processes for a Coast�Open Arctic Transect.The biogeochemical processes were parameterized using OxyDep, a simplified biogeochemical model, that consideredinorganic nutrient (NUT), dissolved organic matter (DOM), particular organic matter (POM) and biota (BIO). Thecarbonate system equilibrium was considered as a fast process and calculated at each time step using an iteration procedurefor pH. According to the model estimates, OM production and decay play leading roles in the carbonate system seasonaldynamics. The modelled seasonal variations of pH (�0.2) are close to those observed, i.e. 7.94�7.99 in February and8.04�8.16 in August (total scale). The surface layer pCO2 varies from 280 ppm during the OM production period to about390 ppm in the centre of the sea and 430�460 ppm in the coastal regions in winter. The summer CO2 invasion is replacedby winter evasion. These estimates can be helpful for the planning of expedition studies and analysing the archived fielddata, as well as for elaborating the interannual and multidecade dynamics models.

Key words: Modelling, carbonate system, ocean acidification, Arctic

Introduction

Increasing partial pressure of CO2 in the atmosphere is

interconnected with the CO2 partial pressure in the

surface layer of the ocean. This leads to ocean

acidification and an increase in the acidity of the

seawater, expressed by a reduced pH (Caldeira &

Wickett 2003; Raven et al. 2005). An increased

concentration of dissolved CO2 in the seawater also

implies a reduced concentration of carbonate ions. This

has consequences for the calcium carbonate saturation

state of the seawater and leads to what is gradually

becoming more difficult for marine organisms � to

build carbonate shells. Corals, including those living

on cold-water reefs, and some pelagic organisms,

including potential key species of the phytoplankton

and zooplankton, are likely to be significantly and

negatively affected by the ongoing acidification (Orr

et al. 2005; Chierici & Fransson 2009).

The problem of estimating ocean acidification

using observations is that the interannual changes of

pH are superposed with large temporal (daily and

seasonal) variability and spatial variability (for exam-

ple, at the frontal zones). Also, the commonly applied

potentiometric technique has a very poor precision

and accuracy (worse than �0.02; Zeebe & Wolf-

Gladrow 2001) compared with the observed trends

and that makes it difficult to compare data from

different sources. This situation is even worse in the

Arctic region, where the available data are scarce,

especially for the winter season.

This article aims to study the role of seasonality of

the biogeochemical processes involved in organic

matter (OM) production and decay in the seasonal

changes of the carbonate system (pH, pCO2, arago-

nite saturation) using observations and a model. The

observations were performed during 4 cruises of a

cargo ship that participated in the Ships-Of-

Opportunity programme, equipped with a Ferrybox

system (http://www.ferrybox.org/eu_project_ferrybox/

index.html.en; Sørensen et al. 2008). This is the

only ship of the EuroGOOS programme that oper-

ates in the Arctic Ocean, covering the line between

*Correspondence: Evgeniy V. Jakushev, Norwegian Institute of Water Research (NIVA), Gaustadalleen 21, NO-0349 Oslo, Norway. E-mail:

eya@niva.no

Published in collaboration with the Institute of Marine Research, Norway

Marine Biology Research, 2013

Vol. 9, No. 9, 822�830, http://dx.doi.org/10.1080/17451000.2013.775454

(Accepted 15 July 2012; Published online 4 June 2013; Printed 14 June 2013)

# 2013 Taylor & Francis

Dow

nloa

ded

by [

81.1

91.1

20.3

8] a

t 13:

58 1

4 Ju

ne 2

013

Tromsø (Northern Norway) and Longyearbyen

(Spitsbergen). This article uses observations col-

lected in 2010. The model includes blocks describing

the biogeochemical processes of OM production and

decay, carbonate system processes and water trans-

port. A biogeochemical model, OxyDep (Yakushev

et al. 2013), was used for simulations of the seasonal

changes in the organic and inorganic nutrients, and

therefore to parameterize the changes in organic and

inorganic carbon. Inorganic carbon and alkalinity

were used for the carbonate system parameter calcu-

lations that were performed using the standard

approaches (i.e. recently described by Dickson

2010). A simplified two-dimensional vertical model

was used to parameterize the advective and turbulent

transport over a 400-km long transect. In this model

we aimed to reveal the reasons for the observed ranges

of the seasonal variations in the carbonate system

parameters and to check the possibility of the

proposed biogeochemical/carbonate system block

using three-dimensional circulation models.

Material and methods

Data

Data used for the comparison with the model results

were taken from a transect Tromsø�Longyearbyen

with a Ferrybox equipped vessel, MS ‘Norbjørn’

(Figures 1 and 2). The Norwegian Institute for Water

Research (NIVA) Ferrybox system allows for contin-

uous measurements of temperature, salinity, turbid-

ity, fluorescence, dissolved oxygen and chlorophyll-a.

The ship’s water intake is positioned at about 4 m

depth and the water is pumped through the sensor’s

measurement compartments. The system is also

equipped with a refrigerated 24�1 litre sampler

allowing automatic sampling in the chosen positions.

The results of measurements are available online

through the NIVA’s server (www.ferrybox.no).

During four cruises in February, May, August and

November 2010, onboard measurements of pH with

potentiometric and spectrophotometric techniques

were made (Dickson et al. 2007) and water sampled

for the determination of total titrated alkalinity and

total inorganic carbon (TIC) was taken. The posi-

tions of the sampling stations along the Tromsø�Longyearbyen transect are shown in Figure 1.

Onboard measurements were made by traditional

potentiometric pH techniques (pH-P) in parallel

with spectrophotometric techniques (pH-S) recom-

mended for the ocean acidification studies (Dickson

2010). pH-P and pH-S operate with different pH

scales: NBS(NIST, IUPAC) scale for pH-P and total

scale for pH-S. The total scale defines pH in terms of

the sum of the concentrations of free hydrogen ions

and HSO�4 (Dickson 2010).

pH-P was measured with a pH-meter, and the

electrode was calibrated before each measurement.

We performed pH-S measurements with a 5-cm cell

equipped HACH DR-2800 field spectrophotometer

that allowed for the measurement of the absorbance at

three wavelengths simultaneously according to Dick-

son et al. (2007). The m-creosol purple dye solution

remained stable during the 2-day cruises. Double the

amount of dye was added to each sample to allow for a

correction of the pH of the dye used. On the basis of

the differences in duplicate measurements of pH-S, it

was possible to estimate the repeatability and measure

the short-term standard deviation. The calculated

repeatability, sR�0.0043, was close to the Dickson

et al. (2007) estimate of 0.003.

The measurements of alkalinity and TIC were

performed at NIVA (Oslo) using the techniques

described in Dickson et al. (2007). The measured

alkalinity and TIC values were converted from mM

into mmol/kg. The seawater density was calculated

with the seawater_properties_v21.xls tool (http://

www.ecy.wa.gov/programs/eap/models.html), based

on the UNESCO equation of state. Carbonate system

parameters (value of pCO2, concentrations of bicar-

bonate and carbonate ions and the aragonite satura-

tion) were calculated with the co2sys_ver14.xls tool

(http://www.ecy.wa.gov/programs/eap/models.html).

Model description

The general equation that is used for the coupled

hydrodynamic-biogeochemical models is the following:

@C

@tþrC~V �rðKrCÞ ¼ RC � @

@zðwCCÞ (1)

Figure 1. Approximate location of the sampling stations along a

transect between Tromsø and Longyearbyen during the 2010

cruises (circles) and the position of the model’s transect (thick line).

Seasonal changes of carbonate system in Barents Sea 823

Dow

nloa

ded

by [

81.1

91.1

20.3

8] a

t 13:

58 1

4 Ju

ne 2

013

where the first term reflects the changes with time,

the second term reflects advective transport with

velocities (~V ) and the third term reflects turbulent

exchange with the turbulent coefficient (K). RCi is

the biogeochemical sink/source term for the consid-

ered variables, C is the concentration of a variable

and wC is the sinking rate of the particulate matter.

Biogeochemical block, OxyDep. The biogeochemical

sink/source terms RCi were parameterized using

the Oxygen Depletion model, OxyDep, which was

applied to the modelling of the oxygen regime of a fjord

system (Yakushev et al. 2013) and to the modelling of

the propagation of hazardous substances in the North

Sea (Green et al. 2011). The idea behind OxyDep is to

parameterize biogeochemical processes in the water

column and in the sediment/water boundary in case of

changeable redox conditions in the simplest possible

way. We aimed to use a simple tool capable of coupling

with hydrodynamic 3D models (i.e. GETM: Stips

et al. 2004; Burchard et al. 2006; HAMSOM: Schrum

et al. 2000) and to couple with additional model

blocks, i.e. for pollutant partitioning (Green et al.

2011) or carbonate system processes.

We presume that, generally, the choice of the

number of variables (Ci) and details of the parame-

terization of the fluxes between them should depend

on the process time scales and the scales of concen-

trations (Yakushev 2002). In the case of studying the

behaviour of large concentration scale parameters

(i.e. oxygen and nutrient, with concentrations of

10�100 mM) at relatively large time scales (seasonal),

it is possible to merge in one compartment the

biological variables (with concentrations less than

0.001�0.01 mM). The following five variables (Ci)

were considered in the model:

. BIO � all the living biota. BIO grows due to

photosynthesis, loses inorganic matter due to

respiration, and loses total (particulate and

dissolved) OM due to metabolism, mortality,

cannibalism.

. NUT � the oxydized form of nutrients (i.e. NO3

for N), that do not need additional oxygen for

nitrification.

. POM � all kinds of labile particulate organic

matter.

. DOM � all kinds of labile dissolved organic

matter and reduced forms of inorganic nutrients

(i.e. NH4 and urea for N).

. OXY � concentration of dissolved oxygen.

The flowchart of biogeochemical processes con-

sidered in the OxyDep is shown in Figure 3. These

processes were parameterized as follows.

The specific growth rate of BIO,

GrowthBio ¼ KNF ftðtÞ fiðiÞ fnðNUT Þ BIO; (2)

is a multiplicative function of temperature, light and

availability of nutrients with the maximum specific

growth rate KNF.

The following formula was used for dependence

on temperature:

ftðtÞ ¼0:2 þ 0:22ðexpð0:21tÞ � 1Þ=ð1 þ 0:28

expð0:21tÞÞ(3)

To describe the dependence on light in accordance

with:

fiðiÞ ¼ fuðuÞI0

Iopt

expð�khÞ exp

�1 � I0

Iopt

expð�khÞ�

(4)

the following parameters were used: incident light

(I0), optimal light (Iopt), extinction coefficient (k),

depth (h) and variation of light with latitude and

time:

fuðuÞ ¼ cos ðu � 23:5 sin ð2T=365:2Þ; (5)

where T is time (days) and 8 is latitude (degrees).

For NUT limitation description we used a satura-

tion curve dependence:

fnðNUTÞ ¼ ðNUT=BIOÞ2

ðNUT=BIOÞ2 þ KNUT

(6)

where KNUT is a constant.

The excretion rate of BIO with specific rate of

excretion KFD was described as:

ExcrBIO ¼ KFD BIO: (7)

The natural mortality rate of BIO with specific

rates of mortality KBP in oxic and KBPA in anoxic

conditions was described as:

MortBIO ¼ KBPBIO þ fs OXYð ÞKABPBIO

þ KCBP 0:5 1 � tanh BIOCan � KcanBIOð ÞBIOðð

(8)

The last term was added to parameterize an addi-

tional mortality due to ‘cannibalism’, that starts

when the BIO concentrations exceeds the threshold

value BIOCan.

We considered the formation of DOM from POM

(autolysis) with a constant specific rate as:

DissPOM ¼ KPDPOM: (9)

The DOM decay takes place due to oxic decay in

oxic conditions (the first term in the following

824 E. V. Yakushev and K. Sørensen

Dow

nloa

ded

by [

81.1

91.1

20.3

8] a

t 13:

58 1

4 Ju

ne 2

013

equation) and denitrification in suboxic conditions

(second term):

DecayDOM ¼KDOMf Dt tð Þfo OXYð ÞDOM

þ KSDOMf D

t tð Þfs OXYð Þf DN NUTð ÞDOM

(10)

where f Dt (t) and f D

t (NUT) are dependences of decay

on temperature and NUT.

The POM decay was parameterized as:

DecayPOM ¼KPOMf Dt tð Þfo OXYð ÞPOM

þ KSPOMf D

t tð Þfs OXYð Þf DN NUTð ÞPOM

(11)

The dependences of decay on temperature, f Dt (t),

were parameterized as:

f Dt ðtÞ ¼Bda

t2

t2 þ t2da

(12)

where Bda and tda are temperature control coeffi-

cients.

f Dt (NUT) is a dependence of decay on NUT

(checking for availability of NO3 and NO2 necessary

for denitrification).

f DN NUTð Þ ¼ 1 � tanh NUTDen � NUTð Þð Þ (13)

where NUTDen is a threshold value.

The changes between the processes occurring in

oxic and suboxic conditions were parameterized with

soft switches based on hyperbolic tangents functions.

fo OXYð Þ ¼ 1 � 0:5 1 þ tanh OXY � Obf2

� �� �(14)

and

fs OXYð Þ ¼ 0:5 1 þ tanh OXY � Obf2

� �� �(15)

where Obf2 is a constant that defines the oxygen

concentration, in which the changes are occurring.

Changes in OXY and total dissolved inorganic

carbon (DIC) were calculated with the Redfield

ratio.

Carbonate system calculations. The carbonate system

equilibration was considered as a fast process and

calculated at every time step using the iteration

procedure. The carbonate system modeling was

described on the basis of a standard approach

(Wanninkoff 1992; Roy et al. 1993; Lewis & Wallace

1998; Zeebe & Wolf-Gladrow 2001).

Hydrodynamic block. To parameterize the hydro-

physical processes of advection and turbulence we

proposed a vertical two-dimensional model that

reproduces the vertical and horizontal transport in

the 400-km long and 200-m deep section, positioned

between the Norwegian coast and Bear Island. The

position of this section is shown in Figure 1. The

hypothetical stream function was used to parameter-

ize the vertical and horizontal advective components

with an upwelling near the coastal boundary and

downwelling at the marine boundary and a lateral

transport between them with a maximum speed of

1 cm s�1 in the surface layer.

The sinking was parameterized as the last term of

Equation (1) and was considered for BIO and POM

with the same sinking velocity wC (Table I).

Boundary conditions. Upper boundary. In the case of

the upper boundary the surface fluxes of the

modelled chemical constituents were assumed to

be zero, except for OXY and CO2.

OXY exchange is given by the flux equation:

QO2¼ k660 ðSc=660Þ�0:5 ðOxsat � O2Þ (16)

where Oxsat is the concentration of the oxygen

saturation as a function of temperature and salinity,

according to UNESCO (1987); Sc is the Schmidt

number; k660 is the reference (Sc�660, CO2 at

208C) gas-exchange transfer velocity. To describe

k660 as a function of wind speed, the following

equation (Schneider et al. 2002) was used:

k660 ¼ 0:365 u2 þ 0:46 u (17)

Simulations were carried out based on a mean wind

speed of 2 m s�1.

Exchange of CO2 was described by the similar flux

equation:

QCO2¼ k660 ðSc=660Þ�0:5 ðCOatm

2 � CO2Þ (18)

where CO2 is the gaseous carbon dioxide concentra-

tion in the surface water and COatm2 ¼ 380 ppm is the

year-averaged concentration of carbon dioxide in

the air.

Lower boundary. The surface fluxes for dissolved

matter were assumed to be zero (a solid impene-

trable lid).

For particulate organic matter (BIO, POM) we

assumed the decrease of concentrations due to

burial (modified on the basis of an approach used

in Savchuk & Wulff 2009):

QCi¼ �BuHvertCi; (19)

where Bu is the burial coefficient, Hvert is the model’s

vertical resolution, and Ci is the concentration of

BIO or POM.

Seasonal changes of carbonate system in Barents Sea 825

Dow

nloa

ded

by [

81.1

91.1

20.3

8] a

t 13:

58 1

4 Ju

ne 2

013

Marine boundary. At the northern marine boundary

of the model transect we assumed this to be sinu-

soidal with time variability of temperature, salinity,

nitrate and alkalinity between the typical summer

and typical winter distributions shown in Figure 4.

These typical distributions were estimated on the

basis of the NODC data found for the 5�5 degree

region near Bear Island.

Computational aspects. The initial calculations em-

ployed a uniform distribution of the considered

variables. Numerical integration was conducted

with the Eurelian scheme with process splitting.

The time step was 0.01 day�1 for all the processes.

The vertical resolution was 10 m, horizontal resolu-

tion 100 km. A quasi-stationary solution with seaso-

nal forced oscillations was reached. There were no

changes in the year-averaged concentrations of the

variables for at least 50 model years.

Results

The results of observations completed along a

transect from Tromsø to Longyearbyen during the

cruises in February, May, August and November

2010 are shown in Figure 2. A minimum pH-S value

in the surface water (7.91, pH total scale) was

observed in winter near Tromsø, the highest (8.23)

was found in May near the Spitsbergen coast. The

observed seasonal variations of pH in the central

Barents Sea were about 0.2, i.e. 7.94�7.99 in

February and 8.04�8.16 in August. The calculations

on the basis of the observed pH-S and Alk surface

Table I. Notations, values, units and names of parameters used in the model (see detailed references in Yakushev et al. 2011).

Notation Value Units Parameter

GrowthBIO day�1 Specific growth rate

fi(i) � Photosynthesis dependence on irradiance

f8 (8) � Irradiance dependence on latitude

ft(t) � Photosynthesis dependence on temperature

fn(NUT) � Photosynthesis dependence on nutrient

KNB 4.0 day�1 Maximum specific growth rate

I0 80 W m�2 Optimal irradiance at the surface

k 0.10 m�1 Extinction coefficient

Iopt 25 W m�2 Optimal irradiance

bm 0.12 8C�1 Coefficient for uptake rate dependence on t

cm 1.4 � Coefficient for uptake rate dependence on t

KNUT 0.02 � Half-saturation constant for uptake of NUT by BIO

KBN 0.05 day�1 Specific respiration rate

KBP 0.01 day�1 Specific rate of mortality in oxic conditions

KBD 0.10 day�1 Specific rate of excretion

KA

BP0.5 day�1 Specific rate of mortality in anoxic conditions

KC

BP0.6 day�1 Specific rate of addtional mortality (cannibalism)

BIOCan 1 mM N Threshold BIO value for cannibalism

KCan 0.8 � Coeficient for the cannibalism description

KPD 0.10 day�1 Specific rate of POM decomposition (autolis)

DecayPOM day�1 Mineralization of POM

KPOM 0.003 day�1 Specific rate of POM oxic decay

KSPOM 0.001 day�1 Specific rate of POM denitrification

DecayDOM day�1 Mineralization of DOM

KDOM 0.05 day�1 Specific rate of DOM oxic decay

KSDOM 0.0005 day�1 Specific rate of DOM denitrification

tda 13 � Coefficient for dependence of decay on t

Bda 20 � Coefficient for dependence of decay on t

Bu 0.22 day�1 m�1 Burial coeficient for lower boundary

NUTDen 1 mM N Threshold NUT value for denitrification

COtoN �8.625 � O to N Redfield ratio (138/16)

fs(OXY) � Function parameterizing switches between oxic and suboxic processes

Obf2 20 mM O Oxygen threshold concentration, in which the changes between

suboxic and oxic processes occur

a0 31.25 mM O Coefficient for the oxygen saturation calculations

a1 14.603 Coefficient for the oxygen saturation calculations

a2 0.4025 8C�1 Coefficient for the oxygen saturation calculations

wC 0.5 m day�1 Sinking velocity

Bu 0.22 day�1 m�1 Burial coeficient for lower boundary

826 E. V. Yakushev and K. Sørensen

Dow

nloa

ded

by [

81.1

91.1

20.3

8] a

t 13:

58 1

4 Ju

ne 2

013

layer partial pressure of CO2 (pCO2) varied from

180�340 ppm during the spring (with a minimum

near Svalbard) to about 390 ppm in the centre of the

transect and 430�460 ppm near the Scandinavian

coast and Spitsbergen in winter. The aragonite

saturation values were in the range of 1.4�2.6 with

a minimum in February near the Spitsbergen coast

and a maximum in the same region in May.

In this work we used a simple hydrodynamic

model to simulate the characteristic features of the

Barents Sea water column seasonal changes. An

application of a two-dimensional model used here

has an advantage compared to a one-dimensional

model, because it allows us to force the model with

changes that happen not only at the surface and on

the bottom but also inside the water column. This is

a similar approach to so the called ‘1.5-dimensional’

model (Konovalov et al., 2000) that is practically

two-dimensional vertical models with 2 horizontal

columns. In this model, the seasonal variability in

the northern boundary of temperature, salinity,

nitrate and alkalinity (Figure 4) forced the distribu-

tions in the water column inside the integration area.

Other forcing factors were seawater fluxes of oxygen

and carbon dioxide. The calculation started with the

uniform distributions of all the considered para-

meters and continued until a quasi-stationary solu-

tion was reached.

The results of modelling of the seasonal variability

of the considered components in the water column

near the marine boundary of the model transect are

shown in Figure 5.

The model simulated the main features of the

seasonal variability of dissolved oxygen and organic

matter. The BIO model compartment allowed us to

parameterize the synthesis of organic matter during

the summer period (from May to August), that

resulted in the formation of large amounts of DOM

(up to 6 mM N or 42 mM C) and POM (up to 1.5

mM N or 10 mM C). These calculated concentrations

correspond to the labile forms of OM. These

modelled values for carbon correspond to the limits

of those observed in the Barents Sea in the July 2011

concentration of DOC (3�323 mM C) and POC

(9�30 mM C) (Chierici et al. 2012).

The sinking of POM and BIO in the summer

period led to the impoverishment of the surface layer

0100

200300

70

72

74

76

78

7.9

8

8.1

8.2

Day

Latitude

pH

-S,to

t

pH-S,tot7.9

7.9875

8.075

8.1625

8.25

0100

200300

70

72

74

76

78

200

250

300

350

400

450

Day

Latitude

pCO

2,pp

m

pCO2,ppm180

250

320

390

460

0100

200300

70

72

74

76

78

1.4

1.6

1.8

2

2.2

2.4

Day

Latitude

Ara

g.S

at.

Arag.Sat.1.4

1.7

2

2.3

2.6

Figure 2. Distributions of pH (total scale) (top), pCO2 (centre)

and aragonite saturation (bottom) along a transect between

Tromsø and Longyearbyen in February, May, August and

October 2010.

NUT

POM

DOM

BIO

OXY DIC

Alk

pH

HCO3-

CO3--

Figure 3. Flowchart of biogeochemical processes in the OxyDep

and carbonate system. The arrows represent the fluxes of matter

between the model compounds. The changes in dissolved oxygen

and carbon were calculated on the basis of the Redfield ratios.

Detailed explanations are in the text.

Seasonal changes of carbonate system in Barents Sea 827

Dow

nloa

ded

by [

81.1

91.1

20.3

8] a

t 13:

58 1

4 Ju

ne 2

013

with NUT and to an enrichment of the subsurface

layer (25�50 m) with OM and NUT. After the

termination of biota development in August, oxygen

in the subsurface layer was quickly decreased for the

mineralization of this OM.

Discussion

The calculated model seasonal variations of pH

(�0.2) are close to the observed ones. That means

that the model is capable of reproducing the features

of the seasonal variability of the carbonate system in

the Arctic. Modelled upper-layer water pCO2 varies

from about 390 ppm (equilibrated with the atmo-

sphere) in winter to 280 ppm during the OM pro-

duction period and this is also close to the observed

one in the central part of the transect (Figures 1

and 2). Both the model and observations show that

the summer invasion of CO2 is replaced by the

winter evasion, more pronounced (according to the

Figure 4. Seasonal changes of temperature (A), salinity (B), nitrate (C) and alkalinity (D) from the surface to a depth of 250 m. NODC

data for the 5�5 degree region near Bear Island are shown as straight crosses for summer and saltire-shaped crosses for winter. Parameters

for the model distributions are shown as a rhombus for winter and as inverted triangles for summer.

828 E. V. Yakushev and K. Sørensen

Dow

nloa

ded

by [

81.1

91.1

20.3

8] a

t 13:

58 1

4 Ju

ne 2

013

observations) near the coasts. These results do not

correspond to those recorded with the empirical

relationship estimates (Omar et al. 2007), that the

Atlantic Sector of the Barents Sea is an annual sink

of the CO2. That can probably be explained by the

small amount of winter data used for this empirical

relationship, first of all near the coasts.

The modelled aragonite saturation dynamics

reproduce the observed tendencies (Figures 2

and 5) with an increase of values from 1.25 in winter

to 2.0 in summer.

The model calculations took into account both

seasonal changes of organic matter production

and decay and temperature-dependent seawater�atmosphere flux variability. Meanwhile, calculations

performed without the influence of OM on the

carbonate system dynamics (not shown) demon-

strated the absence of significant changes in the

carbonate system. In this case the seasonal changes

of pH were less than 0.001, which is much lower

than those observed. This indicates to that the

summer formation of DOM and POM and their

further decomposition play a dominant role in the

carbonate system seasonal dynamics.

Conclusions

OM production and decay lead to seasonal transfor-

mations of carbon between inorganic (TIC) and

organic carbon in the form of biota, detritus and

dissolved carbon that affect pH.

The modelled seasonal variations of pH (�0.2)

are close to those observed along the Tromsø�

Longyearbyen transect, i.e. 7.94�7.99 in February

and 8.04�8.16 in August (pH(Tot), in situ). This

effect should be taken into consideration while

analysing archived data and estimating interannual

variability.

The elaborated model blocks of data for para-

meterization of biogeochemical processes (OxyDep)

and carbonate system processes were verified

with the observations. These blocks can be used

for coupling with three-dimensional hydrophysical

models of the Arctic for estimating the state of

Ocean Acidification in the different layers and

regions.

The results obtained can be helpful in planning

expedition studies and analysis of the archived

field data, as well as for elaborating the changes in

interannual and multidecadal models.

Acknowledgement

This research was supported by the FP7 ECO2

project under grant agreement No 265847 and

projects of KLIF, FRAM and NIVA.

References

Burchard H, Bolding K, Kuhn W, Meister A, Neumann T,

Umlauf L. 2006. Description of flexible and extendable

physical�biogeochemical model system for the water column.

Journal of Marine Systems 61:180�211.Caldeira K, Wickett ME. 2003. Anthropogenic carbon and ocean

pH. Nature 425:365�70.

BIO, uM N

100 200 300175

150

125

100

75

50

25

0

1

2

3

4

5

6

7

8POM, uM N

100 200 300175

150

125

100

75

50

25

0

0.2

0.4

0.6

0.8

1

1.2

1.4

TIC, uM

100 200 300175

150

125

100

75

50

25

0

2100

2120

2140

2160

2180

2200

2220pH, tot

100 200 300175

150

125

100

75

50

25

0

7.9

8

8.1

8.2

8.3

NUT, uM N

100 200 300175

150

125

100

75

50

25

0

5

10

15

20DOM, uM N

100 200 300175

150

125

100

75

50

25

0

1

2

3

4

5

6pCO2, ppm

100 200 300175

150

125

100

75

50

25

0

200

300

400

500

600Arag.Sat.

100 200 300175

150

125

100

75

50

25

0

0.8

1

1.2

1.4

1.6

1.8

2

O2, uM

100 200 300175

150

125

100

75

50

25

0

390

400

410

420

430

440CO2, uM N

100 200 300175

150

125

100

75

50

25

0

15

20

25

30

35 CO3, uM

100 200 300175

150

125

100

75

50

25

0

80

100

120

140

160

180HCO3, uM

100 200 300175

150

125

100

75

50

25

0

1900

1950

2000

2050

2100

Figure 5. Calculated seasonal variability of the model parameters. Vertical axis in metres, horizontal axis in days.

Seasonal changes of carbonate system in Barents Sea 829

Dow

nloa

ded

by [

81.1

91.1

20.3

8] a

t 13:

58 1

4 Ju

ne 2

013

Chierici M, Fransson A. 2010. CaCO3 saturation in the surface

water of the Arctic Ocean: Undersaturation in freshwater

influenced shelves. Biogeosciences 6:2421�32.Chierici M, Sørensen K, Johannessen T, Børsheim KY, Olsen A,

Yakushev E, et al. 2012. Tilførselsprogrammet 2011. Over-

vaking av forsuring av norske farvann. Klima- og forurens-

ningsdirektoratet (Klif), Rapport TA 2936/2012. NIVA

rapport 6365�2012. (in Norwegian)

Dickson AG. 2010. The carbon dioxide system in seawater:

equilibrium chemistry and measurements. In: Riebesell U,

Fabry VJ, Hansson L, Gattuso J-P, editors. Guide to Best

Practices for Ocean Acidification Research and Data Report-

ing. Luxembourg: Publications Office of the European Union,

p 17�40.

Dickson AG, Sabine CL, Christian JR. 2007. Guide to Best

Practices for the Ocean CO2 Measurements. PICES Special

Publication 3. 191 pages.

Green NW, Heldal HE, Mage A, Aas W, Gafvert T, Schrumm C,

et al. 2011. Tilførselsprogrammet 2010. Overvaking av

tilførsler og miljøtilstand i Nordsjøen. Klima- og forurens-

ningsdirektoratet (Klif), Rapport TA 2810/2011. NIVA rap-

port 6187-2011. (in Norwegian)

Konovalov S, Ivanov LI, Samodurov AS. 2000. Oxygen, nitrogen

and sulphide fluxes in the Black Sea. Mediterranean Marine

Science 1(2):41�59.Lewis E, Wallace DWR. 1998. Program Developed for CO2

System Calculations. ORNL/CDIAC-105. Carbon Dioxide

Information Analysis Center, Oak Ridge National Laboratory,

US Department of Energy, Oak Ridge, Tennessee. http://

cdiac.ornl.gov/oceans/co2rprt.html (accessed 15 June 2012).

Computer program.

Omar A, Johannessen T, Olsen A, Kaltin S, Rey F. 2007. Seasonal

and interannual variability of the air�sea CO2 flux in the

Atlantic sector of the Barents Sea. Marine Chemistry 104:

203�13.Orr JC, Fabry VJ, Aumont O, Bopp L, Doney SC, Feely RA, et al.

2005. Anthropogenic ocean acidification over the twenty-first

century and its impact on calcifying organisms. Nature

437:681�86.Raven J, Caldeira K, Elderfield H, Hoegh-Guldberg O, Liss PS,

Riebesell U, et al. 2005. Ocean Acidification Due to Increasing

Atmospheric Carbon Dioxide. Policy Document 12/05.

London: The Royal Society. 60 pages.

Roy RN, Roy LN, Vogel KM, Porter-Moore C, Pearson T, Good

CE, et al. 1993. The dissociation constants of carbonic acid in

seawater at salinities 5 to 45 and temperatures 0 to 458C.

Marine Chemistry 44:249�67.Savchuk OP, Wulff F. 2009. Long-term modelling of large-scale

nutrient cycles in the entire Baltic Sea. Hydrobiologia

629:209�24.Schrum C, Janssen F, Hubner U. 2000 Recent climate modelling

in North Sea and Baltic Sea. Part A: Model description

and validation. Berichte des Zentrums fur Meeres- und

Klimaforschung, Universitat Hamburg, 37. 60 pages.

Schneider B, Nausch G, Kubsch H, Peterson I. 2002. Accumula-

tion of total CO2 during stagnation in the Baltic deep water

and its relationship to nutrient and oxygen concentrations.

Marine Chemistry 77:277�91.Stips A, Bolding K, Pohlman T, Burchard H. 2004. Simulating

the temporal and spatial dynamics of the North Sea using the

new model GETM (General Estuarine Transport Model).

Ocean Dynamics 54:266�83.Sørensen K, Durand D, Folkestad A, Magnusson J, Wehde H.

2008. Operational high-frequency observations from Ferrybox.

Proceedings of 5th EuroGOOS Conference. Exeter, UK, May

2008.

UNESCO. 1987. Progress on oceanographic tables and standards

1983�1986: Work and recommendations of the UNESCO/

SCOR/ICES/IAPSO Joint Panel. UNESCO Technical Papers

in Marine Science, no. 50. 59 pages.

Wanninkhof RH. 1992. The relationship between wind speed and

gas exchange over the ocean. Journal of Geophysical Research

97:7373�82.Yakushev EV. 2002. On parameterization of biogeochemical

processes in modelling objects at different time scales. Electro-

nic Journal Studies in Russia (Issledovano v Rossii) 141:1587�94. (in Russian)

Yakushev EV, Debolskaya EI, Kuznetsov IS, Staalstrøm A. 2013.

Modelling of the meromictic fjord Hunnbunn (Norway) with

an Oxygen Depletion model (OxyDep). In: Yakushev EV,

editor. The Handbook of Environmental Chemistry. Chemical

Structure of Pelagic Redox Interfaces: Observation and Mod-

eling. Berlin, Heidelberg: Springer-Verlag, p 235�52.

Zeebe RE, Wolf-Gladrow D. 2001. CO2 in Seawater: Equili-

brium, Kinetics, Isotopes. Amsterdam: Elsevier Oceanography

Series, 65. 346 pages.

Editorial responsibility: Tom Fenchel

830 E. V. Yakushev and K. Sørensen

Dow

nloa

ded

by [

81.1

91.1

20.3

8] a

t 13:

58 1

4 Ju

ne 2

013