Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
Slide 1
Workshop working groups:
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Modelling:
Coupling.
How to deal with multi scales.
Dangerous sea states.
Physics
Dissipation.
Non linear source term.
Wave effect on the oceans.
Wave/current interaction.
Wave under extreme winds.
New source terms.
Data
Analysis and re-analysis.
Future satellite missions.
Validations techniques.
Forecast products.
Slide 2
PRESENT STATUS of WAVE FORECASTING
AT E.C.M.W.F.
Jean-Raymond Bidlot
Marine Aspects Section
European Centre for Medium-range Weather Forecasts
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Slide 3
Symmetric Slope
2001 2011
Introduction: sustained improvement over the years
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
For example: global wave height forecast against buoy measurements:
analysis
1 day FC
3 day FC
5 day FC
7 day FC
In-situ wave observations
Scatter Index
2001 2011
Slide 4
Limited extend.
11 km grid spacing.
Stand alone.
Forced by 10m neutral wind fields.
Use surface currents from TOPAZ4.
Data assimilation of altimeter data.
2 daily forecasts extending to day 5.
Output every hour, including spectra
(*).
ECMWF Wave Model Configurations
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
1) Limited Area Wave model (LAW)
TOPAZ4
surface currents
Slide 5
Global.
Coupled to the atmospheric model.
Data assimilation of altimeter data.
Part of all forecasting components
(high resolution, ensemble,
monthly, seasonal, re-analyses)
ECMWF Wave Model Configurations
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Atm
osp
heri
c
mo
de
l
Wa
ve
mo
de
l
neutral wind
wind gustiness
air density
roughness
Stokes drift
2) Global models
Global from 81°S to 90°N
Slide 6
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
ECMWF Wave Model Configurations
Ensemble forecasts
55 km grid spacing.
30 25 frequencies *.
24 12 directions *.
Coupled to TL639 (32 km)
TL319 model *.
(50+1) (10+5) day forecasts from
0 and 12Z (monthly twice a
week).
High resolution
28 km grid spacing.
36 frequencies.
36 directions.
Coupled to the TL1279 model
(16km).
Analysis every 6 hrs and 10 day
forecasts from 0 and 12Z. * Change in resolutions after 10 days
NB: also in seasonal forecast at lower resolutions
Slide 7
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Ocean Wave Modelling: ECWAM
The ocean wave modelling at ECMWF is based on the wave
mode WAM cycle 4 (Komen et al. 1994), albeit with frequent
improvements.
Wave model page on the Centre‟s web site:
http://www.ecmwf.int/products/forecasts/wavecharts/index.html#forecasts
General documentation:
http://www.ecmwf.int/research/ifsdocs/CY36r1/index.html
Slide 8
Latest upgrade to operational system (CY38R1)
(19 June 2012):
Wave model main changes:
- Sinput + Sdiss + Sbottom
- Bug fix to wave stress table.
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
with bug corrected
Drag Coefficient versus wind speed (coupled runs):
Slide 9
Wave height (m)
Stdev
error
N.H.
Compared to model analysis:
CY38R1: wave scores
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Mean wave period (s)
Positive = improvement
old
new
Bias (m)
S.H.
Scatter Index
S.H.
Compared to altimeter data:
Slide 10
Comparison with 17 operational centres at a set of buoys as part of the
activities of JCOMM Expert Team on Waves and Storm surges (ETWS):
Wave height
Wind speed
Peak period
Global systems at all buoys, February-May 2012:
ECMWF old
Bias (model-obs) Scatter Index
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Slide 11
Comparison with 17 operational centres at a set of buoys as part of the
activities of JCOMM Expert Team on Waves and Storm surges (ETWS):
Wave height
Wind speed
Peak period
Global systems at all buoys, February-May 2012:
ECMWF new
Bias (model-obs) Scatter Index
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Slide 12
CY38R1: comparison with buoy spectra
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
CY37R3
CY38R1
Data from NDBC (US), ISDM (Canada), CDIP (US)
Slide 13
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Present status of ECWAM:
The 2-D spectrum F(f,θ) follows from the energy balance
equation (in its simplest form: deep water case):
where the group velocity Vg is derived from the dispersion
relationship which relates frequency and wave number.
dissnling SSSFt
F
V
Sin: wind input source term (generation).
Snl: non-linear 4-wave interaction (redistribution).
Sdiss: dissipation term due to whitecapping (dissipation).
Slide 14
Grid and Advection:
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Irregular lat-lon grid to keep the distance
between grid points roughly constant
Corner Transport Upstream scheme:
Unresolved bathymetry obstructions:
Model bathymetry:
Obstructions for low frequency from South:
Slide 15
Wind input Sin:
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
based on Janssen (1991):
))(cos(
ln*
2
zc
uc
gzZ o
where Z is the effective wave age:
βmax = 1.2 (as before)
oz
uU
10ln*
10
neutral 10m wind U10
*u
von Karman constant
oz
friction velocity
surface roughness
Fc
uZeS Z
w
ain 0 ),cos( max
2
*4
2
max
air density
zα : wave age tuning parameter:
zα = 0.008 (0.011 before CY38R1) New:
roughness oz
a
w
air density
water density
wind direction,
c phase speed,
),( fF
g
sea state dependent
Slide 16
Wind input Sin:
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
FSin
c
u*
0 ),cos( max
2
*
c
u
w
a
4
2
max ZeZ
Slide 17
Wind input Sin: gustiness parameterisation
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
wind gustiness FSin
**2
*
2
**
*
* )( 2
)( exp
2
1)( duu
uuu
)( )( 5.0)( ***** uuu
3/13
10
10
** 5.0
10 08.0 5.01
L
zb
C
U
U
u i
d
10
3 10)08.08.0( UCd
* : standard deviation of u*
10* UCu d
iz Inversion height
L Monin Obukhov length
from the atmospheric model :
0b
Slide 18
Wind input Sin: linear swell damping
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
FSin
)cos(20 ),cos( max
2
*
2
*
V
c
c
u
c
u
w
a
V : wind speed at height z = 1/k
Following Janssen (2004), the small effect of turbulent eddies
on the waves can be modelled as
Slide 19
Sdiss
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Sdiss following Bidlot, Janssen and Abdalla (BJA) 2007,
back to Komen et al. 1994 form:
2
2
0
2 )1( meanmean
meanmeandsdiss
k
k
k
kmkCS
33.1dsC 5.0
0
m
dfdFmean
0
m
dfdFkkmean
dfdFm 0
Slide 20
Snl:
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
deep ),(shallow nlnl TransfhkfTransf
The calculation of the non linear source term is still based on the Discrete
Interaction approximation (DIA).
For shallow water, the transfer coefficients are re-scaled:
Following Janssen and Onorato (2005), using the narrow band approximation, it was
shown that the scaling factor could be written as
)1()(4)]1([ 22222 TTkhTkhTk
vg
where
),tanh(khT ,))2sinh(
21(5.0
kh
khcgv ,
kc
,ghsc gkT
2
k
vT
Rhkf
g
8
2
),(
Slide 21
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Shallow water Snl
1
)2/2(1
8
910922
2
3
24
gs
g
vc
cv
khT
TTR
k
vT
Rhkf
g
8
2
),(
),tanh(khT
,))2sinh(
21(5.0
kh
khcgv
,k
c
,ghsc
gkT2
0 1 2 3 4
kD
0
2
4
6
8
R^2
Depth dependent correction factor Kinetic Equation
Eq. (39) (no wave-induced current)
Eq. (39) (with wave-induced current)
kh=1.363
2R
Slide 22
Sbottom:
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Sbottom (see Komen et al. 1994):
) 2 sinh( 2
hk
kCS botbottom
gCbot
038.0
Slide 23
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Bottom induced wave breaking:
Dissipation due to bottom induced wave breaking was added to the source
terms. Following Battjes, Janssen and Beji:
( , )dis BJ bS C Q f F f
hH maxH
H
s2
2max2
wavesbreaking offraction : bQ
6.0 rameterbreaker pa
1BJC encymean frequf
)1(exp bb QQ
Slide 24
Coupling to the waves: Warm skin layer model
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Following Takaya et al. (JGR 2010), a skin layer model is used to represent
the Daily SST Amplitude.
In this scheme, the temperature profile is controlled by the turbulent
diffusivity Kw(z):
)/(
)( )(
*
Lz
LfuzzK
h
aww
Following Grant and Belcher (JPO 2009),
for stable condition only and for f(La) >1:
3/2
1)(
a
aL
Lf
Stokes
wa
U
uL
*
Langmuir number
*
wu )/( Lzh: friction velocity in water : similarity function, L: Obukhov length
Stokes drift
Slide 25
Data assimilation:
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Currently only using Jason-2 wave heights.
Resuming use of Jason-1 under evaluation.
Still using the Optimum Interpolation scheme from Lionello et al.
(1992).
Some minor adaptations:
Slide 26
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Land recognition:
Hs analysis
increments
default Land detected
The model background error should recognize the presence of land
Slide 27
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Shallow water spectral update:
Egu
E )(2
4
* *
tgut )/*(*
But in shallow water
E* is limited by depth,
Hence use deep water
relation but only to
E*max
E*max
dgu
d )( *2
*
Egu
E )(2
4
* *
fugf *)/(*
Windsea update relies on deep water model growth curves in order to
determine an update to u* and mean frequency:
Slide 28
Data assimilation
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Impact still limited:
Jason-2 + ENVISAT
Jason-2
ENVISAT
No data
Slide 29
Products:
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
All “standard” wave parameters in the operational catalogue.
Including wave spectra.
A few „freak wave parameter‟, including Hmax.
New set of parameters to include wave effects on the ocean
fluxes:
Slide 30
New parameters:
A small portion of the stress is retained by the wave field to be
released later.
Hence, one can compute the stress that is actually acting on the
oceans.
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Monthly mean of the normalised stress into the ocean
as derived from ERA-Interim data
ua
2
*It is normalised by
Slide 31
New parameters
Similar consideration can be made for the energy fluxes passing first
into the waves.
It is then dissipated by the waves and transferred into the upper
oceans where it will contribute to the mixing of the top of the oceans.
Both quantities are connected to the wave model source terms.
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Monthly mean of the normalised energy
flux into the ocean
as derived from ERA-Interim data.
ua
3
*It is normalised by
Slide 32
Can we still improve ?
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Comparison to buoy spectra,
stratified by energy level:
Comparison to altimeter wave heights :
Bias
SI
Latest CY38R1 compared to observations, January to May 2012
Slide 33
Can we still improve ?
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Shallow water issues:
62042: UK CEFAS buoy, located on the north shore of East Anglia
in 18m of water
Slide 34
Ongoing developments:
Integration of the atmosphere-wave-ocean models.
Unstructured grid option.
Sea ice damping.
Sbottom.
…
Many more following this workshop…
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Slide 35
Future developments: unstructured grid
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Slide 36
Future developments: unstructured grid
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Slide 37
Future developments: unstructured grid
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
20 m/s
Slide 38
Future developments: sea ice damping:
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
D2
a ciα
From Kohout and Meylan (JGR 2008), Figure 6.
h The non-dimensional
attenuation
coefficient “a” was
found to depend only
on wave period and
sea ice thickness “h”
Slide 39
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Hs (m)
Tp (s)
Tm (s)
Raw buoy data
(further QC still needed!)
Model,
with wave propagation only
When sea ice cover > 30%
Model,
with sea ice damping
up to sea ice cover of 90%
(ciblock=0.9)
Slide 40
Questions/comments?
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
Slide 41
New decomposition:
Future developments: spectral partitioning
Workshop on Ocean Waves, ECMWF, June 25-27, 2012
windsea
Swell 1
Swell 2
Swell
windsea total
Operational:
Slide 42
Verification against buoy frequency spectra
f
E(
f )
area under spectrum = <2>
2
4H s
where A is the
area under curve for
a given frequency bin
Hs,
significant wave height. EWH,
Equivalent Wave Height
for a given frequency bin:
AEHW 4
EWH for waves with
periods between
17s and 19s.
EWH for waves with
periods between
5.5 s and 6.0 s.
Workshop on Ocean Waves, ECMWF, June 25-27, 2012