[PPT] - Development and Validation of Data Assimilative East Sea Regional PowerPoint Presentation

SLIDE 1

Development and Validation of Data Assimilative East Sea Regional Ocean Model

Kyung-Il Chang1, Young Ho Kim2, Gyun-Do Park1, Young-Gyu Kim3

1Research Institute of Oceanography/School of Earth and Environmental

Sciences, Seoul National University

2Coastal Engineering Research Department, Korea Ocean Research and

Development Institute

3Agency for Defense Development

SLIDE 2

Regional setting: East Sea

SLIDE 4

Regional setting: East Sea

SLIDE 5

Regional setting: Circulation

Miniature Ocean

Warm & cold water regions
Subpolar front
Deep water formation
Deep circulation
Double-gyre upper circulation
Mesoscale eddies

Courtesy of Dr. J.J. Park

SLIDE 6

Naganuma (1977) Senjyu et al. (2005)

Regional setting: Circulation

LCC: Liman Cold Current NKCC: North Korea Cold Current EKWC: East Korean Warm Current NB: Nearshore Branch

SLIDE 7

Tsushima Current Water North Korean Cold Water (Coastal mode of the East Sea Intermediate Water)

Regional setting: Water Masses

Deep water masses (< 1ºC)

SLIDE 8

Ulleung Warm Eddy

Regional setting: Eddies

SLIDE 9

A Miniature Ocean in Change

200 240 280 320

Dissolved Oxygen (μ M)

0.2 0.4 0.6 0.8

Potential Temperature (oC)

3500 3000 2500 2000 1500 1000 500

Depth (m)

1932 :Uda, 1934 1954 :USSR AOS, 1957 1969 :Sudo, 1986 1979 :Gamoand Horibe, 1983 1996 :Kim et al., 1999 2005:CREAMS

Global ocean temperature change: substantial warming in the upper 3000m, averaging about 0.037°C between 1955 and 1998

Levitus et al. (2005, GRL)

SLIDE 10

Before 1981 (1st workshop) Cooperative Study of the Kuroshio and Adjacent Regions (1965-1977) 1981-1992 Bilateral Collaboration (Korea/Tsushima Strait submarine cable voltage measurement) 1993-1997 CREAMS (Circulation Research of the East Asian Marginal Seas) Multi-national, multi-disciplinary collaboration 1998-2002 CREAMS II Japan/East Sea Program (USA/ONR) 2005 CREAMS/PICES Program under PICES (North Pacific Marine Science Organization) EAST-I Program (East Asian Seas Time-series: East/Japan Sea)

Brief History of International Programs

SLIDE 11

International collaborations

Joint surveys along meridional and zonal baselines; material flux measurements across the Korea Strait; joint workshops

Eulerian

time-series measurements

Volume transport monitoring; HF radar; coastal buoy and Super-Station; Volunteer

bserving ships; Moored observations
Lagrangian

measurements

Argo floats; Argos drifters; gliders

EAST( East Asian Seas Time-series) - I

SLIDE 12

Research Tasks (EAST-I)

Establishment of integrated ocean time-series system
Ecosystem structure and variability in response to physical forcing
Air-sea interaction, mixed layer dynamics and ecosystem response
Monitoring and understanding the thermohaline circulation
Carbon cycle and its response to climate change
Role of straits in climate and ecosystem
Physical-biological coupled modeling & future climate projection

SLIDE 13

125 130 135 140

1000 1500 2000 2500 3000 100 200 300 400 500 600 700 800 APEX CTD-911

SSH (T/P, ERS) SSW (ECMWF) SST (NOAA) T/S.. Profiler (APEX) Hydrographic Data (CREAMS, JODC, KODC) Real-time Monitoring Buoy Surface Current (Surface Drifter)

Input data Validation data

East Sea Regional Ocean Model

35 40 45 50

Inflow (Cable Voltage)

140

Temperature (PIES)

Observation Systems in the East/Japan Sea

SLIDE 14

Highly-resolved Observation in the UB

ONR JES Program: URI, KORDI, KU 16 current meters, 23 pressure-gauge- equipped inverted echo sounders Daily T & dynamic fields between June 1999 and June 2001

SLIDE 15

Mitchell et al. (2005); mean surface dynamic height

Regional setting: Circulation & Variability (UB)

SLIDE 16

130 135 140 35 40 45 50

500 5 500 500 5 1000 1000 1000 1 1000 1000 1000 1000 2000 2 2000 2000 2000 2000 3000 3000 3000

UB JB YR YB K S TS S S MS

Tuman

Horizontal Domain (127.5 ~ 142.5 °E, 33.0 ~ 52.0 °N) Horizontal resolution: 0.06~0.1º(zonal), 0.1º (meridional) Modelling periods: 1993~2002

East Sea Regional Ocean Model (ESROM)

Based on GFDL MOM3

Z-coordinate level model Parallel Processing (MPI) Hydrostatic and Boussinesq approximations

Open Boundary Conditions

Barotropic velocity of inflow and outflow – Estimated from the transport estimated by submarine cable Baroclinic structure of inflow – historical hydrography

Surface Boundary Conditions

Heatflux - Calculated from meteorological variables by Bulk Formula Saltflux - Restoring to observed SSS Windstress - ECMWF

Features

Explicit free surface Smagorinsky SGS for momentum Robert-Marshall Isoneutral SGS for tracers KPP Vertical SGS Parameterization Partial cell

ESROM

SLIDE 17

Heatflux – Bulk Formula

) (

lw lat sen sw net

Q Q Q Q Q + + − = ) (

1 10

θ ρ − =

a h a p a sen

T W C C Q ) (

1 10

q q W C L Q

a E e a lat

− = ρ

[ ]

( )⎪

⎭ ⎪ ⎬ ⎫ ⎪ ⎩ ⎪ ⎨ ⎧ − + − − =

a a a a SB lw

T T c F e T Q

1 3 5 . 4

4 ) ( ) ( 05 . 39 . θ εσ

Saltflux – Restoring to SSS

( )

τ τ

γ

surf

bs

surf

S S S − =

+1

Forced by monthly mean Surface Boundary Conditions and Open Boundary Conditions

Large, William G., et. al., 1997, Sensitivity to Surface Forcing and Boundary Layer Mixing in a Global Ocean Model : Annual-Mean Climatology, J. of Phys. Oceano., vol. 27, 2418-2447

(a) (b)

(WOA2001)

Winter Summer

Surface Boundary Condition ESROM

SLIDE 18

2.0

2

/cm dyne

3 8

/ 10 cm dyne

−

(a) (b)

Windstress (ECMWF) Winter Summer

Surface Boundary Condition ESROM

SLIDE 19

Open Boundary Conditions

Radiation condition for the tracers and barotropic velocity

= ∂ ∂ + ∂ ∂ + ∂ ∂ y C x C t

y x

φ φ φ

( ) ( )

2 2 2 2

/ / / y x x t Cx ∂ ∂ + ∂ ∂ ∂ ∂ ∂ ∂ = φ φ φ φ

( ) ( )

2 2 2 2

/ / / y x y t C y ∂ ∂ + ∂ ∂ ∂ ∂ ∂ ∂ = φ φ φ φ

An additional nudging term is added for the influxes ( )

ext y x

y C x C t φ φ τ φ φ φ − − = ∂ ∂ + ∂ ∂ + ∂ ∂ 1 > =

x

ut

C if τ τ < = = =

x y x in

C if C C and τ τ

Volume constraint

[ ]

∫ ∫∫ ∫∫∫

⋅ = ⋅ = =

b b

L S V

dL n u h dS n u dV dt d dt dV

Zero-gradient condition across the boundaries for the sea surface elevation

Marchesiello, P., McWilliams, J.C., and Shchepetkin, A. (2001) Open boundary conditions for long-term integration of regional oceanic models, ocean modeling, 3: 1-20.

ESROM

SLIDE 20

Open Boundary Conditions

Inflow (Cable Voltage) Jan-98 Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04

Date (month-year)

1 2 3 4 5

Transport (Sv) Volume transport through the Korea Strait by a submarine cable between Pusan and Hamada

ESROM

SLIDE 21

Theoretical Implementation

Weaver and Courtier (2001) Correlation modeling on the sphere using a generalized diffusion equation Size of background error covariance matrix : ~5 x 1011 (x 8 byte) = 4,000 Gbyte

neither estimated completely nor

even stored explicitly

Modeling B matrix as a sequence of operators.

A central task in the development of a statistical data assimilation Estimation of background error covariance

3-DVar

SLIDE 22

Variational assimilation system with atmospheric models Background error covariance - Correlation functions in terms of a spherical harmonic expansion It is not practical for the ocean due to lateral boundary Assimilation system with oceanic model Lorenc(1992, 1997) and Parrish et al.(1997) : Recursive grid-point filters (UKMO) Derber and Rosati (1989) : Iterative Laplacian grid-point filter (NCEP)

☺

Very efficient and flexible for geographical variations

_-

Limited flexibility in the shape of the correlation function difficult to make anisotropic Objectives : 3D univariate correlation models numerically efficient and sufficiently general; correlation functions with different shape (not just Gaussian), geographically variable length-scale, horizontal/vertical non-separability, and 3D anisotropy.

Theoretical Implementation

3-DVar

SLIDE 23

3D correlation model

3D covariance operator Vertical correlation model

v

M R r r v v r v R

D t I L } ) ( {

1

∑

=

− Δ − = κ

2 / 2 / 1 2 / 1 2 / 1 2 / 1 2 / 1 2 / 1 2 / 1 1 1 T v R T h P h h P v R T h P h h P T v R v v R h h P v v R

L L W L L L W L L W L W L W L

− − − − −

= =

2 / 2 / 1 2 / 2 / 1 2 / 1 2 / 1 2 / 1

,

T v R T h P h T h h P v R

L L W C W L L C

− −

= Λ =

α α

Sequence of operations for

(i) Multiply each element of the input vector by the inverse of the square root of its associated volume element (ii) Perform Mh/2 integration steps of the horizontal diffusion equation (iii) Perform Mv/2 integration steps of the vertical diffusion equation (iv) multiply each element of the filtered vector by it corresponding normalization factor

2 / 1 α

C Applied in reverse order for with adjoint code of the diffusion equation

2 / T

Cα , Diffusion equation , Correlation model

3-DVar

SLIDE 24

128E 129E 130E 131E 132E

Longitude

34 N 35 N 36 N 37 N 38 N 39 N

Latitude

KOREA JAPAN

Jumunjin Jukbyn Mukho Kampo

107 106 105 104 103 102 209 208

EAST SEA

207

7 7 7 7 7 5

100 200 300 400

Distance (km)

0.0 0.2 0.4 0.6 0.8 1.0

Correlation

fn(r) = Exp(-r/L) L : Decorrelation Length Scale

Characteristics of the hydrography taken by the NFRDI, Korea

200
150
100
50

50 100 150 200

X-Distance (km)

200
150
100
50

50 100 150 200

Y-Distance (km)

Correlation at 10 m

40 60 80 100 120 140 160

Decorrelation Length Scale (km)

400 350 300 250 200 150 100 50

Depth (m)

De-Correlation Length Scale

Horizontal Decorrelation Length Scale : 77 km Vertical Decorrelation Length Scale : 77 m

3-DVar

SLIDE 25

3D-Var Assim. Sys. For East Sea Regional Ocean Model

call save_an Do Loop { call setulb call update_dx call cost call grad_Jv } End Do Loop Loop minimizing cost function call grids call setocn call ioinit call topog Start call mom Ocean model save restart The end

Ocean Model Structure Ocean Model Structure

call grids call setocn call ioinit call topog Start call mom save restart The end call setocn_3dvar

#ifdef DA_3DVAR #endif #ifdef DA_3DVAR #endif

call init_assim

Ocean Model & Ocean Model & D.Assimilation D.Assimilation Structure Structure

Update T&S Observed T&S

MPI parallel Processing !!

3-DVar

SLIDE 26

Reanalysis with

1. SST Satellite image
2. Temperature of CREAMS(SNU)
3. Temp. of NFRDI
4. Temp. of JODC
5. Temp. taken by ARGO floats

Data distribution

3-DVar

SLIDE 27

Theoretical Implementation

Cooper and Haines (1996) Surface data assimilation problem – Requirement of a rearrangement of water parcels in space without modifying their T,S properties or their potential vorticity. Hydrostatic connection between

s

p Δ and subsurface pressure updates,

( )

s z

p z p g dz ρ Δ = Δ + Δ

∫

If we set

( )

p z H Δ = − = as a bottom constraint, this will ensure that the bottom pressure and current distribution (through geostrophy) are not altered. This bottom constraint gives the relationship

H s

g dz p ρ

−

Δ = Δ

∫

the change in weight of the entire water column should compensate for the change in surface pressure observed by the altimeter

3-DVar

SLIDE 28

Assimilating Sea Surface Height

Cooper and Haines (1996)

s s

p g ρ η Δ =

z

g dz ρ Δ

∫

( )

s H

p z H p g dz ρ

−

Δ = − = Δ + Δ

∫

No Change of the Bottom Pressure !!!

Sea Surface Isopycnic Surface

3-DVar

SLIDE 29

Using AVISO Product

Dok Is. Ulleung Is. Sea Level at Ulleung and Dok Islands Nearest point of the AVISO from Obs. Ulleung Island Dok Island

3-DVar

SLIDE 30

Surface current and Height (Model)

3-DVar

SLIDE 31

Ulleung Basin Temperature at 100 dbar (PIES)

129oE 30 130oE 30 ’ 131oE 30 ’ 132oE 30 ’ 133oE 30 ’ 36oN 30 ’ 37oN 30 ’ 38oN 30 ’

22 OCT 1999

Ulleung Basin Temperature at 100 dbar (PIES)

129oE 30 130oE 30 ’ 131oE 30 ’ 132oE 30 ’ 133oE 30 ’ 36oN 30 ’ 37oN 30 ’ 38oN 30 ’

21 NOV 1999

Ulleung Basin Temperature at 100 dbar (PIES)

129oE 30 130oE 30 ’ 131oE 30 ’ 132oE 30 ’ 133oE 30 ’ 36oN 30 ’ 37oN 30 ’ 38oN 30 ’

21 DEC 1999

Ulleung Basin Temperature at 100 dbar (PIES)

129oE 30 130oE 30 ’ 131oE 30 ’ 132oE 30 ’ 133oE 30 ’ 36oN 30 ’ 37oN 30 ’ 38oN 30 ’

20 JAN 2000

Ulleung Basin Temperature at 100 dbar (PIES)

129oE 30 130oE 30 ’ 131oE 30 ’ 132oE 30 ’ 133oE 30 ’ 36oN 30 ’ 37oN 30 ’ 38oN 30 ’

19 APR 2000

Ulleung Basin Temperature at 100 dbar (PIES)

129oE 30 130oE 30 ’ 131oE 30 ’ 132oE 30 ’ 133oE 30 ’ 36oN 30 ’ 37oN 30 ’ 38oN 30 ’

19 MAY 2000

Ulleung Basin Temperature at 100 dbar (PIES)

129oE 30 130oE 30 ’ 131oE 30 ’ 132oE 30 ’ 133oE 30 ’ 36oN 30 ’ 37oN 30 ’ 38oN 30 ’

18 JUN 2000

C

2 4 6 8 10 12 14 16 18 20

)