1
The NCEP GODAS Ocean Analysis of the Tropical Pacific Mixed Layer Heat Budget
on Seasonal to Interannual Time Scales
Yan Xue
Climate Prediction Center, NCEP
Acknowledgements:
Boyin Huang, David Behringer, Arun Kumar (CPC) Dongxiao Zhang, Michael J. McPhaden (PMEL)
The CTB Seminar, COLA, Maryland, November 19, 2009
2
Seasonal to Interannual Forecasting at NCEP
Global Ocean Data Assimilation
System (GODAS)
Climate Forecast System (CFS)
SST XBT
Moorings Altimeter
Argo
Reanalysis 2 Surface Fluxes
SST Anomaly Forecast
Forecasters
Official ENSO Forecast
Official Probabilistic Surface Temperature
& Rainfall Forecasts Seasonal Forecasts for North America
CCA, CA Markov
CCA, OCN MR, ENSO
Ocean Initial Conditions
IRI
IRI
3 Operational Global Ocean Data Assimilation System
(GODAS)
Model
GFDL’s Module Ocean Model v3
Quasi-global 74oS – 64oN, 1ox1o (1/3o in tropics), 40 vertical levels
Monthly/pentad outputs, 1979-present, 1 day delay
Forcing
Wind stress, heat fluxes, E-P from Reanalysis 2
SST relaxed to OI SST
SSS relaxed to Levitus climatology
Assimilation method
3D variational scheme, limited to upper 750m
Univariate in temperature and salinity
Error covariance varies geographically and temporally
Assimilation data
Temperature from XBTs, Argo, TAO/TRITON/PIRATA
Synthetic salinity constructed from temperature and local T-S climatology
Sea surface height from Jason-1 since March 14 2007
4
Global Ocean Monitoring Products Based on GODAS
(delivering climate information to society)
http://www.cpc.ncep.noaa.gov/products/GODAS/
• Synthesis of global ocean
observations by NCEP’s Global Ocean Data Assimilation
System (GODAS)
• Climatology and anomaly
plots for each month in 1979 – present
• Monthly Ocean Briefing Annual Ocean Review
• Intended for use by
operational climate prediction centers, researchers, fishery managers, industries, news media, program managers, teachers and students
Contact: Yan Xue, NOAA/CPC
Synthesis of Ocean Observations
5 The Ocean Component of the NCEP ENSO CFS
NOAA/OGP/CTB Project (Mar 2006 – Feb 2009)
PIs: Michael J. McPhaden, Dongxiao Zhang (NOAA/PMEL) Yan Xue, David Behringer (NOAA/NCEP)
• Diagnose the physical processes in the upper tropical Pacific that determine the ENSO development using observations, GODAS, CFS forecasts.
• Conduct tropical Pacific temperature, salinity and current analysis based on Argo, TAO, and XBT data.
• Construct and maintain a web site for real time
comparison between observations, GODAS, and CFS
forecasts.
6
Motivations for Heat Budget Analysis
• What physical processes control the evolution of SST anomaly in the tropical Pacific?
• Can GODAS provide a tool for monitoring and understanding the heat budget of the tropical Pacific mixed layer in real time?
• How well does GODAS simulate the climatological heat budget?
• How well does GODAS simulate the anomalous heat budget associated with ENSO?
Huang, B., Y. Xue, X. Zhang, A. Kumar, and M. J. McPhaden, 2009: The NCEP GODAS ocean analysis of the tropical Pacific mixed layer heat budget on seasonal to interannual time scales, Submitted to J. Clim..
Xue, Y., Alves, O., Balmaseda, M., Ferry, N., Good, S., Ishikawa, I., Lee, T., McPhaden, M., Peterson, D.,
& Rienecker, M., 2010: Ocean state estimation for global ocean monitoring: ENSO and beyond ENSO. In Proc. "OceanObs’09: Sustained Ocean Observations and Information for Society" Conference (Vol. 2), Venice, Italy, 21-25 September 2009, Hall, J., Harrison D.E. and Stammer, D., Eds., ESA Publication WPP-306.
7
Model Data Sets
• Pentad Temperature, salinity, U, V, W with data assimilation (GODAS)
• Pentad Temperature, salinity, U, V and W without data assimilation (CTL)
• Pentad R2 surface heat fluxes
• OI SST version 2
• Levitus temperature and salinity climatology
• TAO currents at 165
oE, 170
oW, 140
oW, 110
oW
• OSCAR surface currents
• OAFlux latent and sensible
• ISCCP short and long wave radiation
Validation Data Sets
8
GODAS OI
GODAS - OI
Sea Surface Temperature Biases 1
oS-1
oN
9 Mean Surface Heat Flux Biases
Mean Surface Heat Flux Biases (1984-2002) (W/m
(1984-2002) (W/m
22) ) Net Heat
- 60 W/m
2Short Wave
Net Heat Flux Latent Flux
10
Surface Heat Flux Annual Cycle (1oS-1oN) Net Surface Heat Flux Biases R2 - OAFlux
Surface Heat Flux Correction (Relaxation to OI SST)
GODAS – OI SST
11 Anomaly Correlation with OAFlux
Anomaly Correlation with OAFlux Net Heat
good correlation
Short Wave
Net Heat Flux Latent Flux
Net heat flux anomaly correlates well with OAFlux in the NINO3.4 region.
12
Mixed Layer Depth Annual Cycle (m)
GODAS 82-04
GODAS-WOD 2001
Levitu s
GODAS
GODAS - Levitus
1
oS-1
oN
13
Sea Surface Salinity and Mixed Layer Depth, 5oS-5oN Average
GODAS PMEL SSS
GODAS-PMEL SSS
MLD
SSS Diff.
MLD MLD Diff.
White contour shows the 29°C isotherm, indicating the edge of warm pool
14 OSCAR
GODAS
GODAS-GODAS
OSCAR
Zonal Current at 15m Depth (1
oS-1
oN, 1993-2007)
GODAS
GODAS - OSCAR
15 Surface Current Biases: GODAS vs TAO
U
V
165E 170W 140W 110W
165ºE 170ºW 140ºW 110ºW
OSCAR GODAS OSCAR GODAS OSCAR GODAS OSCAR GODAS
MBIAS -18 -24 13 23 -4 13 -2 -18
RMSE 19 26 14 26 8 15 5 18
ACC 0.93 0.76 0.94 0.80 0.95 0.96 0.98 0.98
ARMSE 5 10 6 11 7 7 5 5
16
Methodology for Mixed Layer Heat Budget
zz q
w v
u
t
Q Q Q Q Q
T
E Q
Q Q
Q Q
T t L u L v L w L q L zz
x u T
x u T
x u T
x u T Q u
1. Mixed layer temperature equation
2. Low and high (<75 days) frequency decomposition
3. Annual cycle and anomaly decomposition
Qu: Zonal advection;
Qv: Meridional advection;
Qw: Vertical entrainment Qzz: Vertical diffusion
Qq: (Qnet - Qpen + Qcorr)/ρcph; Qnet = SW + LW + LH +SH;
Qpen: SW penetration; Qcorr: Flux correction due to relaxation to OI SST
Stevenson and Niiler ,1983 Wang and McPhaden, 1999
Kessler et al.
1998
17
Climatological Heat Budget Closure: NINO3.4
CTL
ACC=0.99
GODAS
ACC=0.97
18
Anomalous Heat Budget Closure: NINO3.4
CTL
ACC=0.95
GODAS
ACC=0.70
19
Mean Heat Budget (1982-2004,
oC/month)
Qu:
cooling too strong
Qv:
cooling too strong
Eddy:
heating too weak
20 Annual Cycle Heat Budget (0.5
oN,
oC/month )
Eddy Tt
Qw+Qzz Qq
Qu Qv
21 Annual Cycle Heat Budget at TAO Sites
Qu Qv
Qw+Qzz Qq
Tt Tt
Qq Qw+Qzz
Qu Qv
Wang and McPhaden, 1999 Too weak
22
ENSO Composite
(1
oS-1
oN)
Composite members:
82-83, 91-92, 94-95 02-03, 04-05, 06-07 86-88, 97-98 excluded
Onset phase: Qu, Qv, Qw+Qzz Damping factors: Qq, Eddy
Warm biases Decay phase: Qu, Qq
23
ENSO Composite Heat Budget: NINO3.4
U’Tbar
Qu Qq
24 NINO3.4 Heat Budget: 09/10 El Nino
The model simulated
intraseasonal variability in Qu, but not in Qw+Qzz during Jul-Oct. This
insensitivity of Qw+Qzz to oceanic Kelvin waves
might be model biases that also lead to
imbalance of heat budget.
Qq and Qu contributed to the decay of the cold anomaly in the early spring 09.
Qu and Qw+Qzz contributed to the warming tendency in Mar-Jun.
The recent large
warming tendency since Oct is largely due to Qu.
25
26
27
28 NINO3.4 Heat Budget: 06/07 El Nino
Qu, Qv and Qw+Qzz all contributed to the onset of the warm anomaly in the summer 06.
Qu contributed to the sudden decrease of the warm anomaly in the early spring 07 due to upwelling oceanic Kelvin waves.
The model generally underestimated
intraseasonal variability in the temperature tendency.
29
30
31 09/10 El Nino
06/07 El Nino
32
09/10 El Nino 06/07 El Nino
33 NINO3.4 Heat Budget: 02/03 El Nino
Qw+Qzz, Qv and Qu all contributed to the onset of the warm anomaly in the early summer 02.
Qu and Qq contributed to the decay of the warm anomaly in the late winter 02/03 and spring 03.
34
35
36 NINO3.4 Heat Budget: 97/98 El Nino
Qw+Qzz, Qv and Qu all contributed to the onset of the warm anomaly in the spring 97.
Qq contributed to the decay of the warm anomaly in spring 98, while all the advection terms contributed to the sudden cooling in late May 98.
The model underestimated the cooling tendency in spring 98, probably due to the westward zonal current anomaly biases.
37
38
39 NINO3.4 Heat Budget: 91/92 El Nino
Qw+Qzz, and Qv contributed to the onset of the warm anomaly in the spring/summer 91.
Qu and Qq contributed to the decay of the warm anomaly in the spring 92.
40
41
Summary
• Short wave radiation from Reanalysis 2 has about 50 W/m2 negative biases, which results in about 50 W/m2 deficient fluxes into the GODAS ocean.
• GODAS simulates SST reasonably well due to relaxation to OI SST, which provides about 30 W/m2 warming to
compensate the 50W/m2 deficit in Reanalysis 2 net surface heat fluxes.
• GODAS simulates 20m too deep mixed layer depth west of the Dateline due to underestimation of the fresh pool and its interannual variability associated with ENSO.
• GODAS has large (40 cm/s) westward surface current biases in the far western and eastern Pacific, and
eastward biases in the east-central Pacific. However, zonal current anomaly is simulated reasonably well.
42
Summary
• GODAS simulates the mean and annual cycle of the heat budget in the tropical Pacific mixed layer with a moderate skill due to some biases in zonal and meridional advection and underestimation of eddy heating from TIW.
• The heat budget anomaly for the NINO3.4 region for each El Nino event since 1980 has been examined. GODAS
simulates interannual variability of the heat budget well, but underestimates the intraseasonal variability.
• The heat budget closure is satisfied reasonably well for moderate events (91/92, 94/95, 02/03, 06/07), but not well for strong events (82/83, 86/87, 97/98), pointing to
analysis and model errors.
• All advection terms (Qu, Qw+Qzz, Qv) contribute to the onset of El Nino, Qq and Qu to the decay phase, and Qu and Qw+Qzz, to the onset of La Nina.
