Manfredi Manizza1 M. J. Follows1, S. Dutkiewicz1, C. N. Hill1

Towards Modeling The Carbon Cycle of the Arctic Ocean

1EAPS-MIT, Cambridge, MA2Jet Propulsion Laboratory, Pasadena, CA

3U. Texas at Austin, Port Aransas, TX4Marine Biological Laboratory, Woods Hole, MA

D. Menemenlis 2, J. W. McClelland 3, B. J. Peterson4

TALK OUTLINE

1) The Arctic Ocean Carbon Cycle (AOCC)

2) A role for riverine DOC in the AOCC

3) Transport, Fate & Lifetime of riverine DOC inthe Arctic Ocean: Model-Data comparison.

4) Conclusions

Arctic Ocean Carbon Cycle

ATMOSPHERE

CO2 Fluxes

Primary Production

= 0.4 PgC yr-1 PhytoplanktonShel

f

Coas

t

LandBiospher

e

Surface Ocean

Deep Ocean

CO2

ORGANICCARBONEXPORT

RiversSea Ice

DOC

RiverineDOC flux = 0.25

PgC yr-1

Approach of “Data-Model Fusion”

1) Realistic riverine DOC flux from land to ocean.

2) Realistic representation of Large Scale Ocean Circulation in the Arctic basin.

3) Proper time scale regulating the rate of decay for riverine DOC while in the ocean .

1- Riverine Discharge of DOC in the Arctic

From Lammers et al., JGR 20013754 Hydro-stations

RunoffClimatology

DOCdata fromTownsend

McClellandPeterson

Pan-Arctic DOC Riverine Flux

Large seasonal Cycle of River RunOff peaking in June

0

3.5

Time (months)

Eurasian Load > North-American Load

1011molC month-1

Regional Arctic Set-up of MITgcm

1 - Regional version of MITgcm (cubed sphere) for theArctic domain with open boundary conditions.

2 - Eddy-resolving (~20 Km), using 1992-2001 NCEPAtmospheric Forcing - 3 repeated simulated decades.

3 - ON-LINE version of MITgcm for tracer transportand marine biogeochemical processes (carbon cycle).

4 - This study: ARCTIC OCEAN CIRCULATION and PASSIVE & DECAYING TRACER simulating transport and fate of riverine DOC.

2) - Large Scale Circulation of the Arctic Ocean

Surface flow (0-200 m)

(Figure from Rudels, 2006)

Slide 9

Regional Arctic MITgcm

Surface Salinity & Surface velocity

38 24

Riverine Fluxes & Ocean Dynamics

MacKenzie

Eastern Arctic (EA)

Greenland

[Passive Tracer] & Surface velocity

(Units : microM)

1) Atlantic Waters : Low [TRC]

2) EA Coastal Waters : High [TRC]

3) Pacific Waters : Low [TRC]

30th simulated year - annual mean

0

100

4) WA Coastal Waters : High [TRC] (I) Evident link between :Location of riverine tracer fluxes

Ocean circulation patterns

Western Arctic (WA)

(II) Strong haline gradients set the differences in both water masses and

[trc] distributions

Cross-Shelf Transport, Mixing, and Decay of DOC

DOC(decaying

tracer)

Salinity (passive tracer)

Mixing Line

Coastal Waters (Higher DOC & Lower Salinity)

Open-Ocean Waters(Lower DOC & Higher

Salinity )

can impact mixing line. Hansell et al. 2004

Salinity

Black Dots = Western ArcticWhite Dots = Eastern Arctic

= Decay time scale of Tracer

Data-Model Comparison (I)

500

250

0

500

0

250Western Arctic Eastern Arctic

Without any decayTracer values too high !!

With 5 years decayLower Tracer values

BUT values OFFSET = 5 years

= 0

Manizza et al., in prep

Data-Model Comparison (II)

DOCtotal = DOCriver+ DOCin-situ

DOCin-situ ~ 50 microM

(Anderson , 2002)Eastern Arctic Western Arctic

500

250

0

CONCLUSIONS

1 - Regional Ocean Model (MITgcm) captures main ocean circulation features and water masses distribution in the Arctic Ocean.

3 - This study will help to constrain the carbon transfer for DOC to DIC pool to correctly estimate air-sea CO2 fluxes at basin scale.

2 - Model-data fusion is a robust approach and it’s reliable as benchmark for testing data-based hypotheses.

Assessing Lifetime of Riverine DOC

Circulation Only

∂C/∂t = Adv. + Diff.

∂C/∂t = Adv. + Diff. - C

= 1/ = Decay Time Scale

= 1 - 10 years (5 years example)

C = Idealized Tracer

Circulation +

Decay

Passive Tracer

Decaying Tracer