spatial variability spatial variability

(1)

3 3 - - Dimensional coupled physical Dimensional coupled physical - - biological modelling

biological modelling the North Atlantic:

the North Atlantic:

impact of biogeochemical parameters impact of biogeochemical parameters

spatial variability spatial variability

Svetlana Losa, Oceanography Department, Dalhousie University

Alain F. Vezina, Bedford Institute of Oceanography Dan Wright, Bedford Institute of Oceanography Youyu Lu, Bedford Institute of Oceanography Keith Thompson, Oceanography Department,

Dalhousie University Svetlana

Svetlana LosaLosa, Oceanography Department, , Oceanography Department, Dalhousie University

Dalhousie University Alain F.

Alain F. VezinaVezina, Bedford Institute of Oceanography , Bedford Institute of Oceanography Dan Wright, Bedford Institute of Oceanography Dan Wright, Bedford Institute of Oceanography

Youyu

Youyu Lu, Bedford Institute of Oceanography Lu, Bedford Institute of Oceanography Keith Thompson, Oceanography Department, Keith Thompson, Oceanography Department,

Dalhousie University Dalhousie University

(2)

What is it all about ? What is it all about ?

The coupled model description

Physical pool Strategy for coupling

Biogeochemical model

Model validation

Impact of biological model parameters spatial variability

Parameter estimation problem

Conclusions

The coupled model description The coupled model description

Physical pool Physical pool Strategy for coupling

Strategy for coupling Biogeochemical model

Biogeochemical model

Model validation Model validation

Impact of biological model parameters Impact of biological model parameters spatial variability

spatial variability

Parameter estimation problem Parameter estimation problem

Conclusions Conclusions

(3)

The Physical Model The Physical Model

A North Atlantic circulation model, based on the Los Alamos Parallel Ocean Program (POP) (Smith et al. 1992) with an implicit treatment of the Coriolis term and vertical diffusion.

The K-profile parameterization (Large et al., 1994) is used for vertical mixing.

1 horizontal resolution (10S-80N, 99W-20E).

23 vertical levels (10, 20, 35, 55, 75, 100, 135, 185, 260, 360, 510, 710, 985,1335,

1750, 2200, 2700, 3200, 3700, 4200, 4700, 5200, 5700).

A North Atlantic circulation model, based on A North Atlantic circulation model, based on the Los Alamos Parallel Ocean Program (POP) the Los Alamos Parallel Ocean Program (POP) (Smith et al. 1992) with an implicit treatment (Smith et al. 1992) with an implicit treatment

of the

of the Coriolis Coriolis term and vertical diffusion. term and vertical diffusion.

The K The K - - profile parameterization (Large et al., profile parameterization (Large et al., 1994) is used for vertical mixing.

1994) is used for vertical mixing.

1 horizontal resolution (10S 1 horizontal resolution (10S - - 80N, 99W 80N, 99W - - 20E). 20E).

23 vertical levels (10, 20, 35, 55, 75, 100, 23 vertical levels (10, 20, 35, 55, 75, 100, 135, 185, 260, 360, 510, 710, 985,1335, 135, 185, 260, 360, 510, 710, 985,1335,

1750, 2200, 2700, 3200, 3700, 4200, 4700, 1750, 2200, 2700, 3200, 3700, 4200, 4700,

5200, 5700).

(4)

Forcing Forcing Forcing

Climatological monthly mean wind stress (da Silva et al., 1994).

Climatological monthly mean temperature and salinity (I. Yashayaev, Bedford Institute of Oceanography).

Northern and Southern boundaries are closed with sponge layers at which the water

temperature and salinity are relaxed to climatological values.

Climatological Climatological monthly mean wind stress monthly mean wind stress ( ( da da Silva et al., 1994). Silva et al., 1994).

Climatological Climatological monthly mean temperature monthly mean temperature and salinity (I.

and salinity (I. Yashayaev Yashayaev , Bedford Institute , Bedford Institute of Oceanography).

of Oceanography).

Northern and Southern boundaries are closed Northern and Southern boundaries are closed with sponge layers at which the water

with sponge layers at which the water temperature and salinity are relaxed to temperature and salinity are relaxed to

climatological

climatological values. values.

Initial and Boundary conditions

(5)

The Ecosystem Model The Ecosystem Model

P N

Z D

PP D_P

εD_Z

(1- β)G_P

βG_D G_P

D_D

wD

I

(6)

Biological boundary and initial condition

Biological boundary and Biological boundary and

initial condition initial condition

Climatological seasonal mean nitrate

estimates (World Ocean Database, 1998)

Climatological monthly mean chlorophyll estimates obtained by averaging SeaWiFS data over the period 1997-2003.

Z

_init

= 0.02 and decreases exponentially with the depth.

D

_init

= 0.1

Climatological Climatological seasonal mean nitrate seasonal mean nitrate

estimates (World Ocean Database, 1998) estimates (World Ocean Database, 1998)

Climatological Climatological monthly mean chlorophyll monthly mean chlorophyll estimates obtained by averaging

estimates obtained by averaging SeaWiFS SeaWiFS data over the period 1997

data over the period 1997 - - 2003. 2003.

Z Z

_init_init

= 0.02 and decreases exponentially with = 0.02 and decreases exponentially with the depth.

the depth.

D D

_init_init

= 0.1 = 0.1

(7)

(8)

(9)

The model has captured essential features of the ocean phytoplankton dynamics:

The model has captured essential features of The model has captured essential features of

the ocean phytoplankton dynamics:

a strong seasonal cycle in biological productivity in the mid- to high latitudes in the N. Atlantic;

phytoplankton biomass remains low and relatively invariant year-round in the subtropical to equatorial parts of the basin (except for regions of elevated biomass along west Africa and the equator);

subsurface chlorophyll maximum.

The coupled model had difficulty simulating the nitrate seasonal cycle.

a strong seasonal cycle in biological productivity in a strong seasonal cycle in biological productivity in the mid

the mid-- to high latitudes in the N. Atlantic;to high latitudes in the N. Atlantic;

phytoplankton biomass remains low and relatively phytoplankton biomass remains low and relatively invariant year

invariant year--round in the subtropical to equatorial round in the subtropical to equatorial parts of the basin (except for regions of elevated parts of the basin (except for regions of elevated biomass along west Africa and the equator);

biomass along west Africa and the equator);

subsurface chlorophyll maximum.subsurface chlorophyll maximum.

The coupled model had difficulty simulating the The coupled model had difficulty simulating the

nitrate seasonal cycle.

(10)

(11)

(12)

Horizontal distribution of optimized model parameters Horizontal distribution of optimized model parameters

((Losa, Losa, KivmanKivman and Ryabchenkoand Ryabchenko/Journal of Marine System, 2004)/Journal of Marine System, 2004) α, mg C^m² /(mg Chl W h)

Vp^* , mg C /(mgChl h)

Kyewalyanga et al., 1998

I II III IV V spring 0.073 +- 0.048 0.100 +- 0.045 0.075 +- 0.028 0.078 +-0.025 0.069 +-0.032 autumn 0.019 +- 0.007 0.040 +- 0.020 0.037 +- 0.006 0.049 +- 0.037 0.023 +-0.008

Kyewalyanga et al., 1998

I II III IV V spring 3.30 +- 2.63 6.01 +- 2.35 6.88 +- 3.30 8.23 +-2.62 6.64 +-3.37 autumn 2.24 +- 1.56 4.51 +- 2.09 4.58 +- 2.32 4.43 +-2.68 4.88 +-1.28

(13)

((Losa, Losa, KivmanKivman and Ryabchenkoand Ryabchenko/Journal of Marine System, 2004)/Journal of Marine System, 2004)

(14)

((Losa, Losa, KivmanKivman and Ryabchenkoand Ryabchenko/Journal of Marine System, 2004)/Journal of Marine System, 2004)

(15)

The spatial distribution of the parameters is, obviously, a result of a combined effect of several factors such as solar irradiance, temperature, etc., which may affect lati- tudinal changes of chemical conditions and plankton species composition. However, it is rather difficult to distinguish which of the physical and biological factors, and in which region, contributes more to the spatial vary- ability of the physiological model parameters.

The spatial distribution of the parameters is, The spatial distribution of the parameters is,

obviously, a result of a combined effect of obviously, a result of a combined effect of

several factors such as solar

several factors such as solar irradiance, irradiance, temperature, etc., which may affect

temperature, etc., which may affect latilati-- tudinal

tudinal changes of chemical conditions and changes of chemical conditions and plankton species composition. However, it is plankton species composition. However, it is rather difficult to distinguish which of the rather difficult to distinguish which of the physical and biological factors, and in which physical and biological factors, and in which

region, contributes more to the spatial vary region, contributes more to the spatial vary--

ability of the physiological model parameters.

(16)

(17)

August horizontal distribution of the surface chlorophyll

“a” concentration (mgChl m^-3) in the North Atlantic

a) the model solution obtained with constant biological parameters; b) the model solution obtained

with spatially variable biological parameters and c) SeaWiFS (http://seawifs.gsfc.nasa.gov/SEAWIFS.html) data averaged over 1997-2003.

(18)

Chlorophyll vertical profiles

(19)

Conclusions Conclusions

Using some of the biological parameters,

Using some of the biological parameters, -- previously previously considered as constants,

considered as constants, -- spatially variable allows to spatially variable allows to get a significant improvements in the model

get a significant improvements in the model--data data agreement

agreement

Relationships between physical and biological Relationships between physical and biological patterns appears to be different in physically patterns appears to be different in physically distinct regions.

distinct regions.

Parameterization of the different biological Parameterization of the different biological respond to the variability in physics, under respond to the variability in physics, under

different environmental conditions, still remains different environmental conditions, still remains of a real challenge.

of a real challenge.

Correct formulation of data assimilation problem Correct formulation of data assimilation problem for biology is a powerful tool for investigating for biology is a powerful tool for investigating mentioned above problem, as well as for

mentioned above problem, as well as for forecasting purposes.

forecasting purposes.