Ocean wave ensemble

1
Ocean wave ensemble

• Degui Cao, Hendrik Tolman, H.S. Chen

Marine Modeling and Analysis Branch
NOAA/NCEP/EMC, Camp
Springs, MD

Acknowledgement Zoltan Toth, Bo
Cui, Yuejian Zhu, Binbin Zhou, Mozheng Wei
May, 2008.
Laurel, Maryland, USA
2
Generation and evolution of ocean wave

• The generation and evolution of the wind
  waves are strongly dependent on external wind
  forcing. The wave nonlinear term plays the
  stabilization role on the wave perturbation in
  the wind-wave evolution. Wind sea is not
  sensitive to its initial conditions (spectrum,
  frequency ) unlike the chaotic systems. This
  makes the wave (wind sea) weak nonlinearity.
  Swell is sensitive to its initial conditions, and
  should include perturbation consistent with its
  generation.
• Wind sea actively generated waves
• Swell freely propagating waves with no
  local forcing

3
Status of ensemble systems

• Current operation system (old system)
• Wave ensemble has been operational since
  2005
• Running on old deterministic global model grid
  (1.251) as the control
• 10 wave members generated through GEFS using the
  breeding method.
• Same initial conditions for all members.
• 5 day (126 h) forecast.
• Wind data raw wind data
• Validation
• Nearly 50 buoys in the northern hemisphere.

4
Status of ensemble systems

• New ensemble system
• Wave ensemble has been parallel running
  since April 2008
• Running on 11 wave model grid as the control.
• 20 wave members generated through GEFS using ETR
  method
• Cycling initial conditions for individual members
  to introduce uncertainty in swell results.
• 10 day forecast using the GEFS bias corrected 10m
  wind (future operation)
• 5 day (126 hour) forecast using the raw wind data
  ( testing only)

5
Status of ensemble systems

• New ensemble system

• Validation
• 145 buoys in the northern hemisphere
• altimeter wave data (ENVISAT, GFO, Jason-1)

6
Validation of the old ensemble system

• Data
• Ensemble forecast data
• From Jun. 5, 2006 to Mar. 31, 2007
• Observation data
• NDBC hourly buoy data
• Climate data
• Hourly wave height and wind climate data
  generated using 10 years NDBC buoy datasets from
  Jan. 1997 to Dec. 2006.

7
Validation of the old ensemble system
Global buoy distribution used in the old ensemble
system verification
8
Validation of the old ensemble system
Skill scores against climatology slowly decrease
with forecast lead time. Better skill for waves
then for winds. Lack of data for higher parameter
values.
9
Validation of the old ensemble system
Skill scores against control run slowly increase
with forecast lead time. Better skill for waves
then for winds. Lack of data for higher parameter
values.
10
Validation of the old ensemble system
Hs gt 2m day 5
Hs gt 4m day 5
There is a clear lack of spread in the wave
heights, and a positive bias in particular for
higher wave heights.
11
Validation of the old ensemble system
U10 gt 10m/s day 5
U10 gt 14m/s day 5
There is a clear lack of spread in the wave
heights, and a positive bias in particular for
higher wave heights.
12

Validation of the old ensemble system
13
Comparison of the ensemble systems
old
cycle
Old ensemble setup, ensemble with cycling of
initial conditions and wind bias correction
(BC). Mean wave height (contours) and spread
(shading) 2008/03/28 t06z nowcast
cycle, BC
14
Comparison of the ensemble systems
old
cycle
Old ensemble setup, ensemble with cycling of
initial conditions and wind bias correction
(BC). Mean wave height (contours) and spread
(shading) 2008/03/28 t06z 48h forecast
cycle, BC
15
Comparison of the ensemble systems
old
cycle
Old ensemble setup, ensemble with cycling of
initial conditions and wind bias correction
(BC). Mean wave height (contours) and spread
(shading) 2008/03/28 t06z 120h forecast
cycle, BC
16
Contribution to the ensemble theory

• Perturbation of the wave initial
  conditions to generate the wave ensemble member
  has limited liability in the wave ensemble
  forecast. Its ensemble member is generated
  through
• the GEFS (forcing) ensemble system.
• The uncertainties come from
• 1). Observation analysis
• 2). Model physics and
  parameterization
• 3). External forcing, system
  interaction

17
Conclusions

• New wave ensemble shows 2-5 day memory
  of wind perturbations in wave system.
• Tentatively realistic swell perturbation
• Continuity in perturbations in GEFS between model
  cycles is critical.

18
Future plan

• Validate the new ensemble system with buoys and
  altimeter data
• Compare of the new and old ensemble systems
• Bias corrected wind from GEFS is expected to
  reduce (remove) wave bias in ensemble.
• Cycling of initial conditions will increase
  variability of wave heights (partial solution ?)
• Work towards a combined NCEP-FNMOC ensemble
• Analyze the role of swell played in the wave
  ensemble