UNIVERSITY Kuroshio Extension Meanders: Model Data ...Instead, meanders alternately grew and decayed...
1
The SSH variance is decomposed into baroclinic and barotropic components. SSH tot =SSH BC + SSH BT = Where g is gravity, ρ b is bottom density, φ is the geopotential anomaly calculated relative to a deep pressure reference, p r (p r =5000dbar for HYCOM and 5200dbar for OFES). The variance of baroclinic SSH (termed ‘upper’) and barotropic SSH (termed ‘deep’) were calculated for 2004-2005 in 3 frequency bands: 60-30, 30-15 and 15-7 day. Kuroshio Extension System Study (KESS) consisted of 46 CPIES that provided synoptic measurements in the Kuroshio Extension (KE) between 143°E and 149°E with mesoscale resolution. The array collected data from June 2004 to September 2005. We want to quantify frequency and wavelength of KE meanders in three high-resolution ocean circulation models. Several recent studies provide quantitative metrics that can be used to test the models. This study is timely due to the availability of two data sets: satellite sea surface height (SSH) which provides a basin-wide view and KESS which resolves full-water column meanders. In particular, we are motivated by Tracey et al. which indicates that baroclinic meanders with the right phasing with westward propagating deep fluctuations jointly intensify. Here we examine the structure of KE meanders in three ocean general circulation models: •Hybrid Coordinate Ocean Model (HYCOM) •1/12° spatial resolution •Assimilated (Hindcast) •Free-running (Simulated) •Ocean General Circulation Model for the Earth Simulator (OFES) •1/10° horizontal grid 0 0.2 0.4 0.6 0.8 1 0 60 120 180 240 300 360 Simulated 140 o E 145 o E 150 o E 155 o E 30 o N 32 o N 34 o N 36 o N 38 o N Simulated 145 o E 150 o E 155 o E 145 o E 150 o E 155 o E Hindcast 140 o E 145 o E 150 o E 155 o E 30 o N 32 o N 34 o N 36 o N 38 o N Hindcast 145 o E 150 o E 155 o E 145 o E 150 o E 155 o E Deep Spatial Amplitude Mode 2, 30−15 day OFES 140 o E 145 o E 150 o E 155 o E 30 o N 32 o N 34 o N 36 o N 38 o N Deep Spatial Phase Mode 2, 30−15 day OFES 145 o E 150 o E 155 o E 145 o E 150 o E 155 o E 30 o N 32 o N 34 o N 36 o N 38 o N Upper Variance OFES Deep Variance OFES 30 o N 32 o N 34 o N 36 o N 38 o N Hindcast Hindcast 145 o E 150 o E 155 o E 30 o N 32 o N 34 o N 36 o N 38 o N Simulated 145 o E 150 o E 155 o E Simulated [cm 2 ] 0 0.5 1 1.5 [cm 2 ] 0 0.05 0.1 145 ° E 150 ° E OFES 60−30d 01/04 02/23 04/13 06/02 07/22 09/10 10/30 12/19 0 500 1000 145 ° E 150 ° E Distance Along KE Path [km] 30−15d c 01/04 02/23 04/13 06/02 07/22 09/10 10/30 12/19 0 500 1000 2004 145 ° E 150 ° E 15−7d 01/04 02/23 04/13 06/02 07/22 09/10 10/30 12/19 0 500 1000 145 ° E KESS 60−30d 07/18 09/06 10/26 0 500 1000 145 ° E Distance Along KE Path [km] 30−15d c 07/18 09/06 10/26 0 500 1000 2004 145 ° E 15−7d 07/18 09/06 10/26 0 500 1000 145 ° E 150 ° E Hindcast 60−30d 02/08 03/29 05/18 07/07 08/26 10/15 12/04 0 500 1000 145 ° E 150 ° E Distance Along KE Path [km] 30−15d c 02/08 03/29 05/18 07/07 08/26 10/15 12/04 0 500 1000 2004 145 ° E 150 ° E 15−7d 02/08 03/29 05/18 07/07 08/26 10/15 12/04 0 500 1000 145 ° E 150 ° E Simulated 60−30d 01/04 02/23 04/13 06/02 07/22 09/10 10/30 12/19 0 500 1000 145 ° E 150 ° E Distance Along KE Path [km] 30−15d c 01/04 02/23 04/13 06/02 07/22 09/10 10/30 12/19 0 500 1000 2004 145 ° E 150 ° E 15−7d 01/04 02/23 04/13 06/02 07/22 09/10 10/30 12/19 0 500 1000 Upper Variance: High along the KE (mean path for each model for 2004-2005 shown as black line). • Simulated has the lowest variance for all frequencies (bottom panels). Deep Variance: Hindcast shows a ridge of high variability extending eastward from Shatsky Rise in the 60-30 and 30-15 day bands. •In the 15-7 day band, similar spatial structure and amplitude apparent in all models. Phase speeds are fastest in high frequencies (30 km/day,11 day period, 300km wavelength) with decreasing speeds with longer periods (10 km/day, 45 day period, 500km wavelength). Simulated values are most comparable to KESS. Kuroshio Extension Meanders: Model Data-Intercomparison Stefanie E. Zamorski, Kathleen A. Donohue [email protected] www.po.gso.uri.edu/dynamics/KESS www.uskess.org Complex EOFs were calculated using both the upper and deep fields for 2004-2005 (upper not shown). •Mode 1: All the models are similar: broad-scale distribution of amplitude and phase indicates relatively fast speeds -- likely driven by atmospheric forcing. •Mode 2: Westward propagation from Shatsky Rise and into KE region in OFES and Hindcast. •Phase offset (not shown): OFES has the deep leading by 0.25 of a wavelength which is indicative of baroclinic instability. Hindcast phase is difficult to interpret, the deep field both leads and trails the upper field. B1417 •OFES represents joint intensification between upper meanders and external deep westward propagating features. • At times Hindcast shows this development, but not consistently. We note that since Hindcast assimilates mainly upper-ocean data, deep variability generated at Shatsky Rise may not have the same propagation characteristics as observed. •The relatively weak Simulated KE tends to intersect Shatsky Rise further south than the other models and may not generate strong deep variability at Shatsky Rise. 60-30 day 30-15 day 15-7 day 1. Introduction 3. SSH Variance: Upper and Deep 2. Large Scale Metrics 4. Meander Propagation and Upper-Deep Coupling 5. Deep Variability: 30-15 day band 6. Phase Speed 7. Conclusion Following Tracey et al, SSH anomalies along the mean KE path are contoured as a function of time to determine KE displacement and hence meander amplitude, wavelength and phase speed. The upper and deep have been band-passed filtered for three bands: 60-30 day, 30-15 day, and 15-7 day. Westward propagation slopes while eastward slopes The propagation in the upper is mainly eastward and phase speeds vary within each band, with the fastest in the highest frequency band. The anomalies between KESS and the models are similar but they do not correspond one-to-one. OFES: Most northern position. Strength and path length are comparable to Hindcast. Hindcast: Metrics are highly correlated with satellite (r > 0.9 for strength & position). Model captures the regime shift from stable to unstable. Simulated: The mean position is 1.26° south of Hindcast. Weakest strength. % Variance: Mode 1 & 2: 33.3 & 14.2 % Variance: 27.5 & 15.6 % Variance: 41.5 & 11.6 Deep Upper Positive Negative Contour Intervals [m] Deep: -0.15 - 0.15 Upper: 60-30d: -0.5 - 0.5 30-7d: -0.35 - 0.35 Phase [deg] Normalized Amplitude Normalized Amplitude Phase [deg] Satellite KESS is funded by National Science Foundation Grants OCE02-21008 and OCE08-51246. Masami Nonaka from Japan Agency for Marine-Earth Science and Technology generously provided the OFES output. OFES: The only model with westward deep propagation phasing consistent with baroclinic intensification. Hindcast: Deep westward propagation and growth, but the phasing is not consistent with baroclinic intensification. Simulated: Weak amplitude in the deep. Interaction between upper and deep is also weak. KESS results from Tracey et al showed “most meanders did not grow systematically downstream. Instead, meanders alternately grew and decayed as they interacted with remotely-generated deep eddies which propagated into the region from the northeast and east. Interactions have different outcomes depending on the phasing of the upper and deep anomalies”. Tracey, K. L., et al., 2012: Propagation of Kuroshio Extension meanders between 143°E and 149°E. J. Phys. Oceanogr., In press. See poster B1413 Zamorski, S. E., 2012: Kuroshio Extension meanders: Model data-intercomparison. MS Thesis. THE UNIVERSITY OF RHODE ISLAND GRADUATE SCHOOL OF OCEANOGRAPHY Deep leads Upper Deep leads Upper Note: Color bars are different in each band Deep leads Upper 30 o N 32 o N 34 o N 36 o N 38 o N Upper Variance OFES Deep Variance OFES 30 o N 32 o N 34 o N 36 o N 38 o N Hindcast Hindcast 145 o E 150 o E 155 o E 30 o N 32 o N 34 o N 36 o N 38 o N Simulated 145 o E 150 o E 155 o E Simulated [cm 2 ] 0 0.5 [cm 2 ] 0 0.05 30 o N 32 o N 34 o N 36 o N 38 o N Upper Variance OFES Deep Variance OFES 30 o N 32 o N 34 o N 36 o N 38 o N Hindcast Hindcast 145 o E 150 o E 155 o E 30 o N 32 o N 34 o N 36 o N 38 o N Simulated 145 o E 150 o E 155 o E Simulated [cm 2 ] 0 1 2 [cm 2 ] 0 0.1 0.2 0 0.2 0.4 0.6 0.8 1 0 60 120 180 240 300 360 Simulated 140 o E 145 o E 150 o E 155 o E 30 o N 32 o N 34 o N 36 o N 38 o N Simulated 145 o E 150 o E 155 o E Hindcast 140 o E 145 o E 150 o E 155 o E 30 o N 32 o N 34 o N 36 o N 38 o N Hindcast 145 o E 150 o E 155 o E 145 o E 150 o E 155 o E Deep Spatial Amplitude Mode 1, 30−15 day OFES 140 o E 145 o E 150 o E 155 o E 30 o N 32 o N 34 o N 36 o N 38 o N Deep Spatial Phase Mode 1, 30−15 day OFES 145 o E 150 o E 155 o E 145 o E 150 o E 155 o E 0 0.01 0.02 0.03 0 5 10 15 20 25 30 35 40 Wavenumber Phase speed [km/hr] Phase Speed 2004−2005 OFES Hindcast Simulated KESS 2004 2006 KE Position KE (140−155 ° E) Hindcast 32 o N 33 o N 34 o N 35 o N 36 o N 37 o N 2004 2006 Simulated 32 o N 33 o N 34 o N 35 o N 36 o N 37 o N 2004 2006 OFES 32 o N 33 o N 34 o N 35 o N 36 o N 37 o N 2004 2006 1000 1500 2000 2500 3000 3500 Path Length [km] KE Path Length KE (140−155 ° E) 2004 2006 1000 1500 2000 2500 3000 3500 Path Length [km] 2004 2006 1000 1500 2000 2500 3000 3500 Path Length [km] 2004 2006 0 0.5 1 1.5 Strength [m] Time KE Strength KE (140−155 ° E) 2004 2006 0 0.5 1 1.5 Strength [m] Time 2004 2006 0 0.5 1 1.5 Strength [m] Time 1.7 2 2.3 Izu Ridge Shatsky Rise Japan Japan Trench CPIES Profiler/ADCP/CM Profiling Float Deployments KEO Jason−1 135°E 140°E 145°E 150°E 155°E 160°E 28°N 30°N 32°N 34°N 36°N 38°N 40°N 42°N
Transcript of UNIVERSITY Kuroshio Extension Meanders: Model Data ...Instead, meanders alternately grew and decayed...
The SSH variance is decomposed into baroclinic and barotropic components.
SSHtot=SSHBC + SSHBT =
Where g is gravity, ρb is bottom density, φ is the geopotential anomaly calculated relative to a deep pressure reference, pr (pr=5000dbar for HYCOM and 5200dbar for OFES).
The variance of baroclinic SSH (termed ‘upper’) and barotropic SSH (termed ‘deep’) were calculated for 2004-2005 in 3 frequency bands: 60-30, 30-15 and 15-7 day. !
Kuroshio Extension System Study (KESS) consisted of 46 CPIES that provided synoptic measurements in the Kuroshio Extension (KE) between 143°E and 149°E with mesoscale resolution. The array collected data from June 2004 to September 2005. We want to quantify frequency and wavelength of KE meanders in three high-resolution ocean circulation models. Several recent studies provide quantitative metrics that can be used to test the models. This study is timely due to the availability of two data sets: satellite sea surface height (SSH) which provides a basin-wide view and KESS which resolves full-water column meanders. In particular, we are motivated by Tracey et al. which indicates that baroclinic meanders with the right phasing with westward propagating deep fluctuations jointly intensify. Here we examine the structure of KE meanders in three ocean general circulation models: • Hybrid Coordinate Ocean Model (HYCOM)
• 1/12° spatial resolution • Assimilated (Hindcast) • Free-running (Simulated)
• Ocean General Circulation Model for the Earth Simulator (OFES) • 1/10° horizontal grid
0 0.2 0.4 0.6 0.8 1 0 60 120 180 240 300 360
Simulated
140oE 145oE 150oE 155oE30oN
32oN
34oN
36oN
38oNSimulated
145oE 150oE 155oE145oE 150oE 155oE
Hindcast
140oE 145oE 150oE 155oE30oN
32oN
34oN
36oN
38oNHindcast
145oE 150oE 155oE145oE 150oE 155oE
Deep Spatial AmplitudeMode 2, 30−15 day
OFES
140oE 145oE 150oE 155oE30oN
32oN
34oN
36oN
38oN
Deep Spatial PhaseMode 2, 30−15 day
OFES
145oE 150oE 155oE145oE 150oE 155oE
30oN32oN34oN36oN38oN
Upper VarianceOFES
Deep VarianceOFES
30oN32oN34oN36oN38oN
Hindcast Hindcast
145oE 150oE 155oE30oN32oN34oN36oN38oN
Simulated
145oE 150oE 155oE
Simulated
[cm2]0 0.5 1 1.5
[cm2]0 0.05 0.1
145°E
150°E
OFES
60−30d01/04 02/23 04/13 06/02 07/22 09/10 10/30 12/19
0
500
1000
145°E
150°E
Dist
ance
Alo
ng K
E Pa
th [
km]
30−15d c01/04 02/23 04/13 06/02 07/22 09/10 10/30 12/19
0
500
1000
2004
145°E
150°E15−7d
01/04 02/23 04/13 06/02 07/22 09/10 10/30 12/19
0
500
1000
145°E
KESS
60−30d07/18 09/06 10/26
0
500
1000
145°E
Dis
tanc
e Al
ong
KE P
ath
[km
]
30−15d c07/18 09/06 10/26
0
500
1000
2004
145°E
15−7d07/18 09/06 10/26
0
500
1000
145°E
150°E
Hindcast
60−30d02/08 03/29 05/18 07/07 08/26 10/15 12/04
0
500
1000
145°E
150°E
Dis
tanc
e Al
ong
KE P
ath
[km
]
30−15d c02/08 03/29 05/18 07/07 08/26 10/15 12/04
0
500
1000
2004
145°E
150°E15−7d
02/08 03/29 05/18 07/07 08/26 10/15 12/04
0
500
1000
145°E
150°E
Simulated
60−30d01/04 02/23 04/13 06/02 07/22 09/10 10/30 12/19
0
500
1000
145°E
150°E
Dis
tanc
e Al
ong
KE P
ath
[km
]
30−15d c01/04 02/23 04/13 06/02 07/22 09/10 10/30 12/19
0
500
1000
2004
145°E
150°E15−7d01/04 02/23 04/13 06/02 07/22 09/10 10/30 12/19
0
500
1000
Upper Variance: High along the KE (mean path for each model for 2004-2005 shown as black line).
• Simulated has the lowest variance for all frequencies (bottom panels).
Deep Variance: Hindcast shows a ridge of high variability extending eastward from Shatsky Rise in the 60-30 and 30-15 day bands.
• In the 15-7 day band, similar spatial structure and amplitude apparent in all models.
Phase speeds are fastest in high frequencies (30 km/day,11 day period, 300km wavelength) with decreasing speeds with longer periods (10 km/day, 45 day period, 500km wavelength). Simulated values are most comparable to KESS.
Kuroshio Extension Meanders: Model Data-Intercomparison Stefanie E. Zamorski, Kathleen A. Donohue
[email protected] www.po.gso.uri.edu/dynamics/KESS www.uskess.org
Complex EOFs were calculated using both the upper and deep fields for 2004-2005 (upper not shown).
• Mode 1: All the models are similar: broad-scale distribution of amplitude and phase indicates relatively fast speeds -- likely driven by atmospheric forcing.
• Mode 2: Westward propagation from Shatsky Rise and into KE region in OFES and Hindcast.
• Phase offset (not shown): OFES has the deep leading by 0.25 of a wavelength which is indicative of baroclinic instability. Hindcast phase is difficult to interpret, the deep field both leads and trails the upper field.
B1417
• OFES represents joint intensification between upper meanders and external deep westward propagating features. • At times Hindcast shows this development, but not consistently. We note that since Hindcast assimilates mainly upper-ocean data, deep variability generated at Shatsky Rise may not have the same propagation characteristics as observed. • The relatively weak Simulated KE tends to intersect Shatsky Rise further south than the other models and may not generate strong deep variability at Shatsky Rise.
60-30 day 30-15 day 15-7 day
1. Introduction 3. SSH Variance: Upper and Deep
2. Large Scale Metrics 4. Meander Propagation and Upper-Deep Coupling
5. Deep Variability: 30-15 day band
6. Phase Speed
7. Conclusion
Following Tracey et al, SSH anomalies along the mean KE path are contoured as a function of time to determine KE displacement and hence meander amplitude, wavelength and phase speed.
The upper and deep have been band-passed filtered for three bands: 60-30 day, 30-15 day, and 15-7 day.
Westward propagation slopes while eastward slopes
The propagation in the upper is mainly eastward and phase speeds vary within each band, with the fastest in the highest frequency band. The anomalies between KESS and the models are similar but they do not correspond one-to-one.
OFES: Most northern position.
Strength and path length are comparable to Hindcast.
Hindcast: Metrics are highly correlated with satellite (r > 0.9 for strength & position).
Model captures the regime shift from stable to unstable.
Simulated: The mean position is 1.26° south of Hindcast.
Weakest strength.
% Variance: Mode 1 & 2: 33.3 & 14.2 % Variance: 27.5 & 15.6 % Variance: 41.5 & 11.6
Deep Upper
Positive Negative
Contour Intervals [m] Deep: -0.15 - 0.15 Upper: 60-30d: -0.5 - 0.5 30-7d: -0.35 - 0.35
Phase [deg] Normalized Amplitude
Normalized Amplitude
Phase [deg]
Satellite
KESS is funded by National Science Foundation Grants OCE02-21008 and OCE08-51246. Masami Nonaka from Japan Agency for Marine-Earth Science and Technology generously provided the OFES output.
OFES: The only model with westward deep propagation phasing consistent with baroclinic intensification.
Hindcast: Deep westward propagation and growth, but the phasing is not consistent with baroclinic intensification.
Simulated: Weak amplitude in the deep. Interaction between upper and deep is also weak.
KESS results from Tracey et al showed “most meanders did not grow systematically downstream. Instead, meanders alternately grew and decayed as they interacted with remotely-generated deep eddies which propagated into the region from the northeast and east. Interactions have different outcomes depending on the phasing of the upper and deep anomalies”.
Tracey, K. L., et al., 2012: Propagation of Kuroshio Extension meanders between 143°E and 149°E. J. Phys. Oceanogr., In press. See poster B1413 Zamorski, S. E., 2012: Kuroshio Extension meanders: Model data-intercomparison. MS Thesis.
THE UNIVERSITY OF RHODE ISLAND
GRADUATE SCHOOL OF OCEANOGRAPHY
Deep leads Upper Deep leads Upper
Note: Color bars are different in each band
Deep leads Upper
30oN32oN34oN36oN38oN
Upper VarianceOFES
Deep VarianceOFES
30oN32oN34oN36oN38oN
Hindcast Hindcast
145oE 150oE 155oE30oN32oN34oN36oN38oN
Simulated
145oE 150oE 155oE
Simulated
[cm2]0 0.5
[cm2]0 0.05
30oN32oN34oN36oN38oN
Upper VarianceOFES
Deep VarianceOFES
30oN32oN34oN36oN38oN
Hindcast Hindcast
145oE 150oE 155oE30oN32oN34oN36oN38oN
Simulated
145oE 150oE 155oE
Simulated
[cm2]0 1 2
[cm2]0 0.1 0.2 0 0.2 0.4 0.6 0.8 1 0 60 120 180 240 300 360
Simulated
140oE 145oE 150oE 155oE30oN
32oN
34oN
36oN
38oNSimulated
145oE 150oE 155oE
Hindcast
140oE 145oE 150oE 155oE30oN
32oN
34oN
36oN
38oNHindcast
145oE 150oE 155oE145oE 150oE 155oE
Deep Spatial AmplitudeMode 1, 30−15 day
OFES
140oE 145oE 150oE 155oE30oN
32oN
34oN
36oN
38oN
Deep Spatial PhaseMode 1, 30−15 day
OFES
145oE 150oE 155oE145oE 150oE 155oE
0 0.01 0.02 0.030
5
10
15
20
25
30
35
40
Wavenumber
Phas
e sp
eed
[km
/hr]
Phase Speed2004−2005
OFESHindcastSimulatedKESS
2004 2006
KE Position KE (140−155°E)Hindcast
32oN33oN34oN35oN36oN37oN
2004 2006
Simulated
32oN33oN34oN35oN36oN37oN
2004 2006
OFES
32oN33oN34oN35oN36oN37oN
2004 2006100015002000250030003500
Path
Len
gth
[km
]
KE Path Length KE (140−155°E)
2004 2006100015002000250030003500
Path
Len
gth
[km
]
2004 2006100015002000250030003500
Path
Len
gth
[km
]
2004 20060
0.5
1
1.5
Stre
ngth
[m]
Time
KE Strength KE (140−155°E)
2004 20060
0.5
1
1.5
Stre
ngth
[m]
Time2004 20060
0.5
1
1.5
Stre
ngth
[m]
Time
1.7
2
2.3
IzuRidge
ShatskyRise
Japan
Japa
n T
renc
h
CPIESProfiler/ADCP/CM Profiling Float Deployments
KEO Jason−1
135°E 140°E 145°E 150°E 155°E 160°E
28°N
30°N
32°N
34°N
36°N
38°N
40°N
42°N
![arXiv:1408.3858v3 [math.CO] 30 Nov 2015quasi-random graphs. This allows to represent the graph by a so-called cluster graph. Then, instead of solving the original problem, one can](https://static.fdocument.org/doc/165x107/60a93d90ba4f67251a5e1f65/arxiv14083858v3-mathco-30-nov-2015-quasi-random-graphs-this-allows-to-represent.jpg)
