You are here: Home / Lionel Gourdeau / My Filing Cabinet / Altimetry2

Altimetry2

by Webmaster Legos last modified Mar 27, 2012 06:25 PM

Works on altimetry in the Solomon Sea (start:29/06/2007)

Go back



Lien sur les sites d'Angélique


Lien sur les sites de Billy:
Plots from ERS winds
Plots from XBT track

Bathymetric maps from Maxsea, details on the Solomon Sea:
Woodlark Island.jpeg
SE_PNG.jpeg
Vitiaz.jpeg


A. Mean Dynamic Topography
B. Mapped Sea Level Anomalies
C. Processing of along track data
D. Some plots illustrating observations in-around the Solomon Sea
E. What to tell
F. EKE
G. EOF Analysis
H. Harmonic Analysis (Annual cycle)
I. Climatologic year
J. Transport
K. Low Frequency
L. Ideas for a discussion based on the Low frequency signature in the Solomon Sea
M. Sea Level Variability in the Solomon Sea
N. Idea for an Angelique's paper on altimetry
O. Paper on altimetry



A. Mean Dynamic Topography
1. "Bingham" solution (0.5°x0.5° grid)
a. unfiltered solution:  Pacific (30n-30s) ;   South Pacific
b. Filtered solution following Thierry's processing
1. same filtering than for Grace02 (Castruccio) (x=5°, y=1°)
Pacific (30n-30s) ;     South Pacific
2. fltx5y1  (x=2.5°,y=.5°):  Pacific (30s-30n)South Pacific

2. "Maximenko" solution
(0.5°x0.5° grid)
a. unfiltered solution:
i. MDT Pacific (30n-30s) South Pacific

ii. Geostrophic Current
a.  Zonal component (Pacific)
b.  Circulation in the South West Pacific


b. Filtered solution following Thierry's processing
1. same filtering than for Grace02 (Castruccio) (x=5°, y=1°)
i. MDT:   Pacific (30n-30s) South Pacific
ii. Geostrophic Current
a.  Zonal component (Pacific)
b.  Circulation in the South West Pacific

TOP



B.  Mapped Sea Level Anomalies
time sampling: 7 days

1. unfiltered data:  RMS South West Pacific RMS Salomon Sea
a. Seasonal cycle:   RMS Salomon Sea

2. Filtered data (SBX:5):   RMS Salomon SeaRMS residual
a. Seasonal cycle:   RMS Salomon Sea
b. Interannual:         RMS Salomon Sea

TOP


C. Processing of along track data

C.1 Traitment of altimetric data

Topex/Poseidon along track dataset has been treated by the CTOH team of Toulouse, with special algorithms whose aim is to recover good data in the open ocean that would have been juged erroneous, and to recover data near the coasts.
A post-processing treatment has been applied to flag erroneous left data. A 4 sigma filter is first applied. Then, data are filtered with a 2 sigma cycle to cycle difference filter. Afterward, data are "replaced" on a reference track, which is the barycentre of the track for every cycle.
Another filter is used : the difference of rms between the first point after or before land and his neighbour is not allowed to rise above 2 cm. This filter is applied several times.
Finally, a point must have a minimum of 200 valid cycles to be valid. Otherwise, this point can't be used.


Tracks of interest and bathymetry

Example of the number of valid cycles and sla variance for a track of the new product :   Track 251

C.2 Comparison of the new product with Aviso and the MSLA gridded data

-   Track 23 ;
zoom Salomon
a. RMSRMS Salomon
b. Hovmuller:  MSLADEGEOAviso


Track 73zoom Salomon
a. RMS RMS Salomon
b. Hovmuller:  MSLA DEGEO Aviso

Track 10zoom Salomon
a. RMS ;  RMS Salomon
b. Hovmuller:  MSLADEGEO

Track 112 zoom Salomon
a. RMS RMS Salomon
b. Hovmuller:  MSLA DEGEO

Track 149zoom Salomon
a. RMSRMS Salomon
b. Hovmuller:  MSLADEGEOAviso

Track 162 ; zoom Salomon
a. RMSRMS Salomon
b. Hovmuller:  MSLADEGEO

Track 199 zoom Salomon
a. RMS RMS Salomon
b. Hovmuller:  MSLA DEGEO

Track 238zoom Salomon
a. RMSRMS Salomon
b. Hovmuller:  MSLADEGEO

Track 86zoom Salomon
a. RMSRMS Salomon
b. Hovmuller:  MSLADEGEO

Track 99 zoom Salomon
a. RMS RMS Salomon
b. Hovmuller:  MSLA DEGEO

Track 251 zoom Salomon
a. RMSRMS Salomon
b. Hovmuller:  MSLADEGEO Aviso

Track 175 zoom Salomon
a. RMSRMS Salomon
b. Hovmuller:  MSLADEGEO

Track 188zoom Salomon
a. RMSRMS Salomon
b. Hovmuller:  MSLADEGEO

Track 225zoom Salomon
a. RMSRMS Salomon
b. Hovmuller:  MSLA DEGEO


Track 36
a. RMS
b. Hovmuller:  MSLADEGEO

Track 123
a. RMS
b. Hovmuller:  MSLA DEGEO

Track 47
a. RMS
b. Hovmuller:  MSLADEGEO

Track 60
a. RMS
b. Hovmuller:  MSLA DEGEO

C.3 Variance along the tracks
Sla rms map after filtering, using all TP tracks
Sla seasonal cycle rms map after filtering, using all TP tracks
Sla interannual variability rms map after filtering plus 3 months filtering, using all TP tracks

TOP



D. Some plots illustrating observations in-around the Solomon Sea
(Most from Billy's website)
NICU.png PNGzoom_vectors.pngvitiaz_sect.png

Boug_Kiri_map_with_adcp.gifBoug_Kiri_ug_1000m.gifwepocs2_solomon_sea_adcp_ony.gif

poi1map.pdf poi1sol.pdf

poi2map.pdfpoi2wsol.pdf

mw9304_adcp_vectors3.gif

Alex's page

TOP


E. What to tell

The Solomon sea is characterized by a complex bathymetry and ocean dynamic as illustrated on:
Fig.E1.

How altimetry can help to describe the variability of such dynamics?

- This area exhibits the strongest variability over the Pacific ocean between 10°N-19°S as shown from the gridded altimetric product:
Fig.E2: RMS.gif

- The gridded product, which merged TOPEX/POSEIDON and ERS, has a  1/3° horizontal resolution, and a 5 day temporal resolution. As shown on the figure above, the gridded data don't take account of the complex bathymetry of the Solomon Sea. Therefore, this product could be irrelevant for our study.

- We decide to use a new processing for alond track altimetric data. Compared to the classical product, data are gained near the coast, and more cycles are available. Details on the processing can be found on the Angelique'web site. Here are the tracks that have been used:
Fig. E3:  tracks of interest and bathymetry

Only 10 tracks span the Solomon Sea.

The benefit from this new processing can be illustrated by comparing the
Old (Fig.E4a) with the New (Fig.E4b) Sla along track 251.
This new Sla data have been validated against tide gauge data available in our region (look at "
Tracks of interest and bathymetry" for the location of the tide gauge). We present the time series of the tide gauge, and of the nearest altimetric point for the Lombrum, Madang, Rabaul, Townsville, Honiara sites (faire plots)

- Here is the rms variability along the tracks:
along track sla rms  (Fig.E5)
The highest variability (15 cm rms) is centered along 8°S and extends between 11°S-5°S in latitude, and between 150°E-170°E in longitude.

In the Solomon Sea, the eastern part of the basin, along the Solomon Islands, exhibits  higher variability than the western part, alonb the Papua New Guinea Coast.
Is this difference of variability representative of different dynamics between the west and the east into the Solomon Sea?

The western part of the basin is characterized by Western Boundary Currents, the NGCUC flowing northward below the NGCC; whereas  the eastern part is characterized by complex recirculation.

East of the Solomon Islands, the highest variability are not against the coast but a few degrees to the east. Qiu and Chen (2004) have studied the seasonal cycle in this area, and have found that at seasonal time scale, this high variability is explained from barotropic instability associated with the horizontal shear of the SECC-SEC system.

North (the Bismark Sea) and south of the Solomon Sea (11°S) the variability falls from 15 cm to 10 cm.

In conclusion, the highest variability are concentrated in the Solomon Sea, and just to the east of the Solomon Islands. The explanation of Qiu and Chen (2004) for the seasonal variability east of the Solomon Islands doesn't seem a good one for the Solomon Sea where neither the SECC nor the SEC can easily flow into the Solomon Sea

Some questions:
- From which temporal frequencies is this variability representative?
- How is different the variability inside and outside the Solomon Sea?
- Can the variability of the WBC be observed from altimetry?


Filtering of the data

- The SLA are filtered with a 1-month triangle filter:
Fig.E6a: rms of along track filtered sea level.gif
Fig.E6b: rms of the sla annual cycle.gif
Fig.E6c: rms of sea level once the seasonal cycle is removed.gif
Fig.E6d: rms of the interannual sla signal.gif

A 3 cm rms noise is filtered from the raw data. The description of the variability of the filtered sla is the same than above (check the track 188). The spatial distributions of the variability of the annual and interannual sla signals are similar, but east of the Solomon Islands, the highest annual variability is centered at 6°S whereas it is at 8°S  for interannual variability. The interannual sla signal exhibits higher variability than the annual cycle, 13 cm rms against 8 cm rms respectively. The comparison between the interannual variability and the sla variability free from the annual cycle shows that other time frequencies could exist (bi annual?? noise? see spectrum)

- Check of the track 188. All the data are plotted on:  track 188.gif (Fig. E7) ;  some data seems wrong. Sla greater than .45m are filtered.
The other tracks are also checked

Spectra
- Spectrum at some locations characteristic of different dynamics:
- North of the Vitiaz strait:                 Track 23, 4°S  (Fig. E8a); AvisoT023,4S
- South of the Vitiaz strait:                  Track 112, 7°S
(Fig. E8b); AvisoT112,7S
- Solomon Sea, east part:                     Track 10, 8°S
(Fig. E8c); AvisoT010,8S
- Milne Bay:                                      Track 188, 10°S
(Fig. E8d); AvisoT188,10S
- 10°S, middle west of Solomon Sea     Track 073, 10°S
(Fig. E8e); AvisoT073,10S
- 10°S, middle east of Solomon Sea:     Track 149, 10°S
(Fig. E8f); AvisoT149,10S
-  Makira:                                             Track 225, 10.5°S
(Fig. E8g); AvisoT225,10.5S
- East of Solomon Islands:                   Track 239, 8°S
(Fig. E8h); AvisoT238,8S
- South of Solomon Sea:                       Track 086, 14°S
(Fig. E8i); AvisoT086,14S
- West of the Coral Sea:                      Track099, 11°S
(Fig. E8j); AvisoT099,11S


- Around the Solomon Sea
South of the Solomon Sea, at 14°S, interannual and annual frequencies are dominant with a 5 cm magnitude. More to the west, at 11°S, there is no dominant variability. High interannual signal is present east of the Solomon Island at 8°S with magnitude up to 11 cm. In addition, three peaks between 1 and 1.5 years (4-6 cm) exist (??). North of the Vitiaz strait, the dominant frequency is dominant with a 9 cm magnitude. There is also a clear annual signal (4 cm), and there are also some energy at  4-6 months.
- Inside the Solomon Sea
We check the signal at 10°S from the east to the west between 152°E-162°E. The interannual signal is dominant (up to 10 cm) everywhere, except in the far west part of the section (down to 3 cm). The annual signal is always present but with different amplitude (>4 cm). The amplitude is maximum (up to 9 cm) on the central east part of the section. In the west part of the section, frequencies greater than the annual cycle are distinguishable, particularly a 60 days period. In the basin, there are more energy, at all frequencies, in the eastern part than in the western part. In the East, we retrieve both the 10 cm amplitude of the interannual signal, and the 9 cm amplitude of the annual signal. The 60 day peak is also high with a 3 cm amplitude. In the West, the annual cycle is dominant (6 cm), the interannual signal is only of 4 cm. The 60 day period is also present.
- In conclusion, a 60 day period exists in the Solomon Sea and not outside. The annual cycle is relatively high in the Solomon Sea compared to the surrounding areas. In the Solomon Sea, the interannual variability is particularly located on the central-east part of the basin. Outside, the interannual signal is high everywhere north of 10°S


Difference Between the east-West sides of the Solomon Sea

For each tracks, the points inside the Solomon Sea are selected. Another selection consists to distinguish the west and east parts of the Solomon Sea based on the location of the tracks inside the basin, and a criterion in variability (rms < or > 11 cm, respectively West and East). The west part is the area where is the NGCC. This separation between the East and the West is critical. Noted that the area of the West part is smaller than the East one.

Both time serie are highly correlated (0.87):  Fig.E9:  West/east correlation

And present energy at the same frequencies, mostly at annual and interannual period: Fig.E10: West/east spectra
Logically, there are more energy in the West part than in the East. The interannual signal is dominant in the East part (11 cm),  whereas the annual and interannual signals have same magnitude in the west part (4 cm). A peak at 60 days is visible in the West, and a semi annual signal is visible in the East. We retrieve the conclusion from the spectra at individual locations.

We look at the correlation between the West and the East function of latitude:  Fig.E11: West/East correlation
The correlation is high, around .9 for the annual and interannual signals from 6°S to 9°S, and decrease to .7 at 10.8°S. South of 10.8°S, we are out of the Solomon Sea, and there is a break in the correlation. For the interannual signal, the correlation increases to .99 whereas it decreases for the annual signal

In conclusion, the solomon sea seems to invole in phase between the West and the East, particularly north of 9°S (north of Milne bay) for period higher or equal to the annual signal. South of 9°S, the West part is representative of Milne bay and there are just few points available in this area, therefore may be correlations are less robust. The break between 11°S and 10.5°S seems to delineate the south boundary of the Solomon Sea.

Solomon Sea: East side

Is the SLA correlated between the north and the south of the domain. We  look at the correlation at 7°S function of latitude: Fig.E12: Correlation function of latitude.
The correlation is very close to 1 for the annual and interannual signals. It begins to decrease south of 10°S, and north of 6°S which are the limits of the domain. for the high frequency signal, the correlation falls down to .8 in a  distance of 100° km (Decorrelation scale, rayon de rossby??)

Hovmuller:  Fig. E13 a)full signal b)Annual c)Interannual
Sea level anomalies may each 30-35 cm. The amplitude of the annual signal is of 10-15 cm. It is interesting to see that the maximum (in march) and the minimum (in september) have a 2° extension in latitude, and they are not centered at the same place, 8°S and 9°S respectively. It deseappears at 10.5°S (the south boundary would be Guadalcalnal and not Makira. May be a significant flow exist between the two islands), and it reappears more south with a 4 months lag. The amplitude of the interannual signal is of 15-20 cm, the positive anomalies being higher than the negative anomalies. There are 3 negative events: the 1993, and the 1997-1998 are the most significant. The third one is present during the second half of 1994. There are 3 positive events, Each one having a marked signature during the first half of a year. The 2000- 2001 are the stongest events. There is another one in 1996.
The anomalies in the full signal are clearly a combination of annual and interannual signals.
.
Solomon Sea: West side
Is the SLA correlated between the north and the south of the domain. We  look at the correlation at 7°S function of latitude: Fig.E14: Correlation function of latitude.
The correlation decreases slowly down to .9 at 10°S. South of 10°S, the annual and interannual signals behave differently. For the annual signal the correlation falls down quickly wheras for interannual signal the correlation decreases significantly south of 11°S and reachs .5 at 12°S. For the high frequency, the correlation decreases firstly drasticaly in less than 1°, before to decrease continuously with the latitude. Is the observed decrease of correlation with latitude for the different signals due to some propagation??

Hovmuller: Fig. E15 a)full signal b)Annual c)Interannual
There is a break at 10°5°S in the SLA signal which means that the SLA signature of the Solomon Sea is different from the surroundings. The annual signal has a 5 cm amplitude extending between 9.5°S and 7.5°S. The anomalies are maximum in march and minimun in September. There is a clear 4 months lag with the signal at 11°S (maximum in December). The interannual signal is relatively small with a maximum amplitude between 5-10 cm.

Relation between the Solomon Sea (east side) and the variability at the east of the Solomon Islands

Sal-eastof_ft.gif

Sal-eastof_spe.gif

East of Solomon Islands:
Hovmuller:  full signal Annual Interannual HF
Correlation fonction of latitude


Relation between the Solomon Sea and the variability at the south of the Solomon Islands
Because the south west corner of the Solomon Sea is located more to the south than the south east corner, we look at the east side of the Solomon Sea with the south of the Solomon Sea between 10.5°S et 11.5°S, and we look at the west side of the Solomon Sea with the south of the Solomon Sea between 11.5°S and 13°S.

The two latitudinal bands don't have similar variability:  temporal series of SLA south of the Solomon Seacorresponding spectra

Relation between the east side of the Solomon sea and the [10.5°S-11.5°S] band: temporal seriesspectra

Relation between the west side of the Solomon sea and the [11.5°S-13°S] band: temporal seriesspectra

Relation between the west side of the Solomon sea and the south of PNG: temporal seriesspectra

North of New Britain
Relation with the west side of the Solomon Sea:
temporal series spectra
Relation with the ocean at the east of the Solomon Islands: temporal seriesspectra



First summary:

What is the role of the Solomon Sea to exchange anomalies between the subtropics and the western equatorial Pacific? What do we learn from altimetry??
The Solomon Sea is a pathway for the western boundary current and this boundary current connects both regions. Therefore the WBC may play a significant role in ENSO variability. It is a question to know the exact role of the WBC.
The solomon Sea is divided is two areas: a West side characteristic of the WBC (SalW) and a East Side (SalE).
The Solomon Sea is located inside an area of high sea level anomalies. It means that this high variability at the east of the Solomon Islands (EastOf) is at the latitude of the Solomon Sea, and could be due to barotropic instability (Qiu and Chen, 2004) at seasonnal time scale and also to propagating Rossby waves.
A crucial question is to know the story of these anomalies once they meet the Solomon Islands.
May be, there are some similitude with Hawaii.


Spectra of sla provide some insights:

- First, SalE and EastOf have similar spectra:
SalE-eastof_spe.gif
Ce qui veut dire que les îles Salomons ne sont pas un obstacle à la propagation du signal. Le signal qui arrive à la côte est des iles Salomons va se propager le long des îles et va pouvoir penetrer par les extremites nord (Solomon strait) et sud (Guadalcanal-Makira). En effet, la région au sud de Guadalcanal (
SudOf, 10.5°S-11°S) montre une variabilité interannuelle proche de celle observée sur EastOf: SalE-Sudof_spe.gif

- Time series SalE/EastOf:   Climatology
; low frequency; residual
Pour les 3 signaux (Clim, interannual, haute fréquence), SalE et Eastof ont la même variabilité. Si l'on regarde le signal haute fréquence, il est en retard (1mois) dans SalE par rapport à EastOf

- Time series SalE/SudOf :  Climatologylow frequency; residual
Si la variabilité climatologique est nettement differente entre SalE et SudOf, la variabilité haute et basse fréquence est relativement concordante.

- Second, the variability north of New Britain (NorthOf) is more related to EastOf than to SalW:
NorthOf-EastOf_spectra.gif;    NorthOf-SalW_spectra.gif
Ce qui voudrait dire que la variabilité au nord ouest de la mer des Salomons est directement associée à celle qui de trouve à l'est des iles Salomons et cette relation ne se fait pas forcemment par la mer des Salomons. Le signal arrivant de l'est va donc contourner la Nouvelle Irlande et se propager vers l'Ouest.

- Time series NorthOf/EastOf: Climatologylow frequency residual
Pour les 3 signaux (Clim, interannual, haute fréquence), SalE et Eastof ont la même variabilité.

- Time series NorthOf/SalW: Climatologylow frequency residual
NorthOf a clairement un signal semi annuel visible aussi dans EastOf mais qui n'existe pas dans SalW. La variabilité interannuelle est moins prononcée dans SalW que NorthOf mais les deux signaux sont en phase.

- Third, SalW exibits a lower variability than SalE that seems to be a combination of Signals from SalE, SudOf (11°S-13°S), and from the WBC south of PNG (SouthPNG)

SalW/SalE spectra.gif
Time series SalW/SalE:  Climatology low frequency residual
La différence des spectres suggèrent que la partie ouest et est de la mer des Salomons ont des régimes dynamiques différents.  La variabilité basse fréquence est en phase entre les deux séries.

SalW/SouthPNG_spectra.gif
Time series SalW/SalE:  Climatology low frequencyresidual
SouthPNG montre une faible variabilité interannuelle ce qui suggère que la variabilité interannuelle observée en mer des Salomon (SalW) ne provient pas de façon majoritaire de cette région. La ressemblance des spectres pour les fréquences à 60 jours et annuelle indique bien que le WBC au sud de la PNG contourne l'extremité sud est de la png et longe la cote est de la png. La variabilité "climatologique" est similaire entre les deux séries avec un retard de 2 mois pour SalW

SalW/SouthOf_spectra.gif
Time series SalW/SalE:  Climatology low frequency residual
Les spectres sont similaires surtout pour les fréquences annuelles et interannuelles. Les series climatologiques SouthPNG, SouthOf, SalW sont trés ressemblantes avec des déphasages. Il n'est pas facile de voir si SouthOf est capable d'influencer SalW. On peut penser que si c'est le cas l'effet de SouthOf se fait au détriment de SouthPNG

Il est clair que SalW est associé à SouthPNG et à SalE, le role de Southof est moins évident. Cela voudrait dire que SouthOf se propage vers l'ouest avant de rejoindre SouthPNG.

TOP



F. EKE
Use of gridded SLA from AVISO
Qiu and Chen (2004) have focussed on the Seasonal cycle of the EKE in the SECC box (150°E-170°W; 15°S-5°S).
First, we do similar plots than in their paper.
We retrieve the results from the figures 4a and 5a of Chen and Qiu (2004)
- F.1:  EKE averaged over the SECC box.gif
- F.2: EKE as a function of calendar month in the SECC box.gif

But when looking at the spatial distribution of the mean EKE, we see strong heterogeneous areas. High EKE is cencentrated in the Solomon Sea:
- Fig.3: Spatial distribution of mean eke.gif

The SECC box is divided in two parts: 150°E-160°E and 160°E-170°W
- Fig. 4: EKE averaged over the differents boxes.gif
Mean EKE in the box included the Solomon Sea: 341 cm2/s2 against 174 cm2/s2 for the other part. Results from Qiu and Chen (2004) are relative to the area at the east of the Solomon Sea.

EKE in the Solomon Sea:
- Fig. 5: EKE averaged in the Solomon Sea.gif; 150°E-155°E: 9°S-5°S; Mean EKE: 679 cm2/s2
- Fig.6:  EKE as a function of calendar month in the Solomon box compared to the east part.gif

Contribution of the U component (y derivative):
Fig.7:
EKE averaged in the Solomon Sea.gif
Fig.8:
EKE as a function of calendar month in the Solomon box
Spatial Distribution of mean EKE: U2; V2

- EKE has a dominant interannual signal
- Fig.9: Low frequency
- Fig. 10: High frequency

TOP



G. EOF Analysis
Use of the gridded SLA from AVISO: 1992-2004

EOF analysis are performed both on a climatological series and on a series where the climatological signal has been removed.

First, we consider the domaine 142°E-170°E; 13°S-12°S

- Climatological series: Mode 1Mode 2Mode 3 Mode 4Mode 5
The mode 1 explains 74% of the variance; and the first 3 modes 92%

- "Inter-intraannual" series: Mode 1 Mode 2 Mode 3 Mode 4 Mode 5Mode 6
The mode 1 explains 71% of the variance

EOF performed on the Solomon Sea only:
- raw data:
Mode 1 The mode 1 explains 81% of the variance (Variance=9.8E-3 m2)

- Climatological series: Mode 1Mode 2Mode 3
The mode 1 explains 88% of the variance

- "Inter-intraannual" series: Mode 1Mode 2Mode 3
The mode 1 explains 81% of the variance

TOP



H. Harmonic Analysis (Annual cycle)
- Amplitude and Phase
- Velocity anomalies in March and September

- Amplitude and phase curl tau

TOP



I. Climatologic year

- Hovmuller at different latitute: msla_hov

TOP



J. Transport
L'approche de Ridgway (1993) est utilisée ici pour estimer la variabilité des transports entrant et sortant de la mer des Salomon à partir de l'altimétrie. La SLA est censée représenter la variabilité des 150 premières mètres sus la surface. On utilise les sorties du modèle ORCA05 pour confirmer cette approche.

1. Rms de SLA sur le domaine: Modèle Altimétrie

Le transport est  estimé soit par différence de sla aux extrémités de la section en utilsant l'altimétrie ou la SSH du modèle, soit directement avec le courant méridien du modèle.


2. Flux entrant en mer des Salomons (entre la pointe sud est de la PNG et Makira)
a. Courant méridien du modèle le long de la section: Mean RMS
b. Spectres des transports
c. transports fonction du temps: Non filtrés basse fréquence climatologiques

3. Flux sortant par Vitiaz (trace 99)
a. Courant méridien du modèle le long de la section:
Mean;   RMS
b. Spectres des transports
c. transports fonction du temps: Non filtrés basse fréquence climatologiques

4. Flux sortant par Solomon strait
a. Courant méridien du modèle le long de la section:
Mean; RMS
b. Spectres des transports
c. transports fonction du temps: Non filtrés (0-150) (0-bottom)basse fréquenceclimatologiques

5. Flux sortant à travers la section allant de Makira  l'ouest du détroit des Salomon
a. Spectres des transports
b. transports fonction du temps: Non filtrés (0-150) (HF filtrée) basse fréquenceclimatologiques


6. Flux sortant à travers la section allant du sud est de la PNG à l'est de Vitiaz
a. Spectres des transports
b. transports fonction du temps: Non filtrés (0-150),  basse fréquence climatologiques

7. Comparaison du flux entrant avec la somme des flux sortants
AltimetrieModèle

TOP



K. Low Frequency

Les données sont filtrées à 7 mois (spz41, dt=10 jours)
Les sea level sont soit des données along track, soit des données AVISO grillées
Le signal EKE est issu de AVISO grillé


1. Along track sea level anomalies averaged over the Solomon box
a. Sla/SOI.gif
b: Sla/EKE

2. SLA Aviso averaged over the Solomon box:

a. SLA calculé partir d'un masque sur la mer des Solomon: Slab/SOI.gif

3. L'activité turbulente (EKE) sur la mer des Salomon est liée à la variabilité interannuelle de type ENSO (représentée par la SOI):
a. EKE/SOI.gif
b. EKE calculé sur la mer des Salomon: EKEb/SOI.gif

4.Relation entre l'activité turbulente (EKE) et la variabilité des transports mesurée dans les détroits.
Inflow: ce qui rentre par le sud et l'est
Outflow: ce qui sort par Vitiaz et Solomon straits

A. transports calculés à partir des données along track.
1. Correspondance entre Inflow et Outflow: In/out flow.gif
2. Correspondance entre EKE et Outflow (variables normalisées): EKE/outflow.gif
3. transports at Vitiaz and Solomon strait: Vitiaz/Solomon.gif

B. transports calculés à partir de Aviso (résultats similaires avec le produit along track):

1. Correspondance entre Inflow et Outflow: In/out flow Aviso.gif
2. Correspondance entre EKE et Outflow (variables normalisées): EKE/outflow.gif

5. Relation avec le Warm Water Volume de Meinen (ouest)
1. en terme de niveau de la mer sur les Salomon: SLA/wwva.gif
2. en terme de transport à travers Vitiaz et Solomon strait: Outflow/d(wwv)/dt.gif

6. EOFs
1. SLA Pacifique (14s-14n)/SOI:   Mode1/WWV West.gif;   Mode2/WWV Pac.gif
2. SLA Pacifique Sud (14s-5s)/SOI:    Mode1/WWV West.gif;   Mode2/WWV Pac.gif
3. SLA Salomon/SOI/WWVaWest:  Mode 1.gif; Mode 2.gif

7. Wind Curl from ERS
1. Pacifique (14s-14n)/SOI:  Mode1/WWV Pac.gif Mode2/WWV West.gif
2. Pacifique Sud Ouest/SLA_Sal/SOI: Evolution temporelle.gif
3. Pacifique sud Ouest/Sv: Temporal Evolution.gif

8. TAO
1. 5S156E, Heat Content/SLA_Sal/SOI: Temporal evolution.gif
2. 5S156E/Sv: Temporal Evolution.gif

9. SST
1. Anomalies over the Solomon Sea/SLA_Sal/SOI: Temporal Evolution.gif

10. Relation between SLA and transports in the Solomon Sea: Temporal Evolution,.gif

TOP



L. Ideas for a discussion on Low frequency variability in the Solomon Sea

On the specific role of the Solomon sea to fill/empty the equatorial band in the western Pacific


The importance of the South West Pacific Ocean, and of the Solomon Sea, to connect the subtropical region to the Western equatotrial region is largely discussed in the SPICE programm.
This study is based on altimetric data and focusses on the interannual variability in the Solomon Sea. The Sla are analyzed with regard to the WWV anomalies provided by C. Meinen. Results are supported by model analysis. This study emphasizes that most of the water necessary to balance the discharge/recharge of the equatorial Pacific transit through the Solomon Sea. They point out the importance of the South hemisphere and of the western boundary current in ENSO dynamics.

The results presented below is a synthesis of the work
Data are filtered with a triangle filter and a 7-months cutoff

1.  Maxima of high variability are located in the western Pacific ocean of each hemisphere around 8°. The western end of the ocean is characterized by the recirculation of zonal currents crossing the Pacific basin into western boundary currents. In the South hemisphere,
the high SLA variability between 5S-10S and 140E-170 encompasses the Solomon Sea. We will try to characterize the specific role of the Solomon Sea at ENSO time scale.
FigA:  RMS LF 92-07.gif

2. An EOF analysis over the tropical Pacific (14S-14N) shows two dominant modes explaining
respectively 54% et 23% of the variance. These two modes are discussed a lot in the literature. The mode 1 represents a tilting mode and the mode 2 a discharge and recharge of WWV. Mode 1 is in phase with the SOI whereas mode 2 is negatively correlated with the SOI. (Fig.B)
Fig.B: a) Mode1/WWV West.gif;  b) Mode2/WWV Pac.gif


3. altimetry2plots is a good way to track changes in WWV (already shown by Meinen, 2005). It is worth noting that the WWVa for the western region lags the mode 1 in SLA (and the SOI) by 5-7 months, and the WWVa for the equatorial band is in phase with the mode 2 in SLA. (Fig. B).
The Mode 1 is correlated with the western WWV by about 0.58, and leads the western WWV by about 5 months with a 0.83 correlation.

4. A zoom in the Solomon shows that the first EOF mode explains up to 95% of the interannual variance.
In a statistical point of view, the Solomon Sea exhibits a relatively simple variability. This mode is highly correlated to the SOI and leads the western WWVa by few months (Fig.C). The temporal function of this mode is very closed to that one of the mode 1 over the tropical Pacific (Fig. Cb), and their spatial structure over the Solomon Sea are similar with maxima of varialility along the western coast of the Solomon Islands. The correlation between the two EOF temporal function is high 0.82. The Solomon Mode is correlated with the western WWV by about 0.87, and 0.89 considering a two months lag.
Fig.Ca: Mode_Sal 1.gif
Fig.Cb: Mode1: Sal/Pac.gif

5. For the spatial EOF 1 mode, looking at zonal sections in the latitudinal range spanning the Solomon sea from 5S to 10S, we clearly see that inside the Solomon Sea, the slope of SLA variability is inverse of the slope outside the Solomon Sea (Fig.D). The crosses in Fig.Db delineate the location of the Solomon Islands. It means that the corresponding meridional flow is in phase opposition inside and outside the Solomon Sea. When the interior ocean discharges (recharges), it recharges (discharges) through the Solomon Sea. Looking at a similar figure for the northern hemisphere, we don't see a so strong signature. This result argues for the predominance of the Solomon Sea to fill or deplete the Western Equatorial Pacific.
Fig.D: a) EOF1 South.gifb) EOF1 South West.gif (the Solomon Islands are located by crosses);  c) EOF1 North.gif

6. To illustrate the discussion above, the velocity field corresponding to the spatial EOF1 mode is plotted (Fig.E). When there is a meridional divergence in the equatorial band, a westward surface geostrophic flow
south of 10°S enters the Solomon Sea, and goes north. It bifurcates at the New Britain coast before to escape throuh Vitiaz and Solomon straits. Just a smal part of this westward flow continues  to the Australian coast decreasing the NQC. It seems that a part of the interannual variability of the Australian WBCs is in fact controlled by the SEC inflow. Similar conclusion have been done By Kessler and Gourdeau (2007) when looking at the annual cycle
Fig.E:  EOF1_vector.gif

7. It is an uneasy task to physically interpret the statistical EOF modes. Because there is only one dominant mode in the Solomon Sea, there is no particular interest any more to use EOFs to analyse interannual varaibility inside the Solomon Sea. We retrieve the same relations between the western WWVa, the SOI and the EOF mode than using directly the high-pass filtered SLA time series. Compare Fig.F with Fig.C
Fig.F: Solomon Sea: relation between SLA, SOI and western WWVA.gif

8. We use a numerical simulation: the one used in Kessler and Gourdeau (2007) to check the interannual variability of the model with regard to the results above. We performed similar EOF analysis (Fig.Ga), and in the Solomon Sea we look at the ssh variability (Fig.Gb) and the relations between the SSH signal averaged in the Salomon Sea, the SOI and the western WWV (Fig.Gc). Model analysis are very close to the altimetric analysis giving confidence both in the model and in the data analysis.
ORCA05:
Fig.Ga: EOF1.gif;   EOF2.gif
Fig.Gb: RMS.gif
Fig.Gc: Relation between SSHa averaged over the Solomon Sea, the SOI and the western WWVa.gif

9. Once the relations between the SLA in the Solomon Sea and the SOI and the Western WWVA are established, we try to provide an estimation of the transports crossing the Solomon Sea. We use the same idea already developped in Ridgway et al (1993) with tide gauges. The geostrophic mass transport is estimated from the expression gdHD/f where dH is the sea level difference between each side of the straits, and D=150 is representaive of the depth of the upper thermocline. It is a crude estimation of the transport because D is fixed and transports at deeper levels may exist, particularly in the WBC. These estimations are assessed with the use of the numerical simulation.

10.  Transports are estimated at the south entrance of the Solomon Sea between the south east extremity of the PNG coast and the southern islands of the Solomon Islands (y=10.5°S) for the inflow and at the Vitiaz  and Solomon straits for the outflow (y=-6.41°S, and x=151°E is the longitude which divide th eoutflow into the straits). We need only two SLA measurements (one at each extremity of the sections) to estimate the transports.
We check that the inflow balances the outflow (Fig.H). The inflow is defined by the addition of the flow at the souterhn entrance of the Solomon Sea and of the flow crossing the Solomon Islands. except some differences existing at the peaks of transports anomalies, both curves are greatly similar.  Transports anomalies may reach 10 Sv during the 1997-1998 ENSO event
Fig.H: Inflow/outflow transports from Aviso.gif

11. The model is used to verify the estimation  from
altimetrie (old plots). We remember that the bathymetry of the model is not very good in this area, very poor representation of the Solomon Islands. transports from the model are estimated in differents ways: by using the same method used for altimetry (using modeled SSH), ans directly by using the velocity fields (0/150m and 0/bottom transports are estimated) (Fig.I)
- Inflow at the south entrance: Fig.Ia Comparison Model/altimetry ; Spectra

- Outflow at Vitiaz straits: Fig.Ib Comparison Model/altimetry ; Spectra
- Outflow at Solomon straits: Fig.Ic comparison Model/altimetry ; Spectra
altimetry2plots and model provide similar estimation with similar amplitude and a good phasing. Therefore the crude astimation of transports from altimetry is not so bad. Discrepancies exist for the Solomon strait. It is not surprizing because of an unrealistic bathymetry in the model.

12. The model has the advantage to provide the vertical structure of the currents (Fig.J)
- Inflow at the south entrance: Fig.Ja Mean;   RMS
- Outflow at Vitiaz straits: Fig.Jb Mean;   RMS
- Outflow at Solomon straits: Fig.Jc   Mean;   RMS

13. Relation between the flux in the Solomon Sea (through Vitiaz and Solomon straits) and the flux from the recharge/discharge of the Western WWVa (dWWVa/dt) (Fig.K).
Both time series are out of phase and their magnitude are similar.
Fig.K      Outflow/d(wwv)/dt.gif It is the most important plot!

14. We can separate the contribution of Vitiaz and Soloon straits (Fig. L.)
Fig.L:  Vitiaz/Solomon straits.gif


Reste à developper tout ceci avec la biblio qui va bien, le message étant 1. que la mer des Salomon est un endroit majeur pour "équilibrer" les recharges/décharges qui ont lieu en plein océan. 2. que l'altimetrie avec juste quelques points de part et d'autres des détroits donne des informations assez fiables sur la variabilité basse fréquence" de la circulaiton à travers la mer des Salomon.


References:

Alory G. and T. Delcroix, 2002
Bosc C. and T. Delcroix, Observed equatorial Rossby waves ans ENSO related warm water volume changes in the equatorial Pacific Ocean, J. Geophys. Res., in press,2008
Butt J., and E. Lindstrom, Currents off the east coast of New Ireland, Papua New Guinea, and their relevance to regional undercurrents in the Western equatoriall Pacific, 1994.
Clarke A.J., S.V. Gorder and G. Colantuono, Wind stress curl and ENSO discharge/recharge in the equatorial Pacific, J. Phys. Oceanogr., 37, 1077-1091, 2007
Fine R.A., R. Lukas, F. Bingham, M.J. Warner and R.H. Gammon, The western equatorila Pacific: A water mass crossroads, J.Geophys. res, 99, C12, 25,063-25,080, 1994.
Johnson G.C. and M. McPhaden, Interior pycnocline flow from the subtropical to the equatorila Pacifci ocean, J. Phys. Oceanogr., 29, 3073-3089, 1999.
Johnson T.M.S. and M.A. Merrifield, Interannual geostrophic current anomalies in the near equatorial Western Pacific, J. Phys. Oceanogr., 30, 3-14, 2000.
Kug J.S andI.S. Kang, Symmeric and antisymmetric mass exchanges between the equatorial abnd off-equatorial Pacific associated with ENSO, J. Geophys. Res., 108,C8, 3284, doi:10.1029/2002JC001671, 2003
Holland C.L. and G. Mitchum, Interannual volume variability in the tropical Pacific, J. Geophys. Res., 108, C11, 3369, doi:10.1029/2003JC001835, 2003
Holland C.L. and G. Mitchum, Interannual temperature variability in thetropical Pacific and Lagrangian heat transport pathways, J. Geophys. Res., 110, C03017, doi:10.10029/2004JC002466, 2005
Lee T., and I. Fukomori, Interannual to decadal variations of tropical-subtropical exchange in the Pacific ocean: Boundary versus interior transports, J. Clim., 16, 4022-4042, 2003
Fukomori I., T. Lee, B. Cheng and D. Menemenlis, 2004
Meinen C.S, and M.J. McPhaden, Observations of warm water volume changes in teh equatorial Pacific and their relationship to El Nino and La Nina, J. Climate, 3551-3559, 2000.
Meinen C.S, and M.J. McPhaden, Interannual variability in warm water volume transports in the equatorial Pacific during 1993-1999, J. Phy. Oceanogr., 31, 1324-1345, 2001
Meinen C.S., Meridional extent and interannual variability of the Pacific ocean tropical-subtropical warm water exchange, J. Phus. Oceanogr, March, 323-335, 2005.
Qiu B. and S. Chen, 2004
Ridgway K.R., J.S. Godfrey, G. Meyers, and R. Bailey, Sea level response to the 1986-1987 El Nino-Southern Oscillation event in the western Pacific in the vicinity of Papua New Guinea, J. Geophys. Res., 98 C9, 16,387-16,395, 1993.
Schott F.A, J.P. McCreary, G.C. Johnson, Shallow overturning circulations of the tropical-subtropical oceans, Geophysical Monograph, 147, doi:10.1029/147GM15, 261-304, 2004.
Sloyan B.M., G.C. Johnson and W.S. Kessler, The pacific cold tongue: a pathway for interhemispheric exchange, J. Phys. Oceanogr., 33, 1027-1043, 2003.
Tsuchiya M., et al., Source waters of the PAcific equatorial undercurrent, Prog. Oceanogr., 23(2), 101-147, 1989.
Ueki I., Y. Kashino and Y. Kuroda, 2003

TOP





M. Sea level variability in the Solomon Sea


I. introduction

- La région de la mer des Salomon présente une variabilité de  sla la plus forte sur tout le Pacifique tropical Sud


Fig.A: rms Pacifique Sud.gif (rms AVISO 1992-2007)

- La mer des Salomon est un endroit clé de la circulation océanique car c'est l'endroit où transite le NGCC, le WBC qui connecte les régions subtropicales à l'équateur. Cette connection est partie intégrante des STCs. La variablité de cette circulation méridienne participe activement à la variabilité climatique de type ENSO et notamment à sa modulation basse fréquence.
La mer des Salomon se caractérise par une bathymétrie complexe. C'est une mer semi fermée connectée aux régions équatoriales par 3 détroits: Vitiaz, St Georges, et Solomon strait.

Fig.B: Circulation in the Solomon Sea

- Information apportée par le papier.
- une description exhaustive de la variabilité de sla en mer des Salomon qui se traduit par:
- une différentiation est-ouest
- la dominance des périodes annuelle et interannuelle
- un signal EKE important au sud de la N.B.

- la relation entre la SLA en mer des Salomon vis à vis des alentours

- une interprétation dynamique des signaux observés

- la mise en évidence des WBC à partir d'un traitement spécifique de l'altimétrie


II.  Data

III. Description générale

La sla moyennée sur la mer des Salomons présente des variations de +/- 20 cm. On observe une forte anomalies négative associée à l'ENSO 1997-1998 et de fortes variations aux échelles saisonnière

Fig C: SLA_ft.gif

Sur la Fig.A, il est clair que la variabilité de SLA est forte de part et d'autre des îles Salomons et semble relativement faible dans la partie ouest de la mer des Solomon caractérisée par le NGCC qui s'écoulent le long des côtes de la PNG. Il semble donc que l'on puisse distinguer 2 régions distinctes de variabilité en mer des Salomons. On va donc distinguer les parties ouest et est

Fig D: masque Solomon.gif

Si les SLA pour ces 2 régions sont trés fortement corrélées (0.91), les spectres montrent des niveaux d'énergie trés différent. Les 2 fréquences prépondérantes correspondent aux signaux saisonnier et interannuelle. La variabilité interannuelle est la plus importante.
Pour la partie est l'amplitude du signal interannuel est de 9 cm contre 4 cm pour la partie est. En ce qui concerne le signal saisonnier, l'amplitude est de 5 cm pour la partie ouest contre 3 cm pour la partie est

Fig. E: spectre.gif

La partie est de la mer des Salomon est caractérisée par 2 patchs de rms, un situé au nord (152°E-155°E; 8°S-5°S) et un autre situé au sud (155°E-160°E, 10°S-8°S).

Le patch au nord est principallement associé à de la variabilité interannuelle alors que celui au sud est aussi associé à un signal saisonnier même si la variabilité interanuelle reste dominante.

Fig. E1: a) low frequency.gif ;  b) climatology.gif

Le calcul de l'EKE montre un fort signal en mer des Salomon situé au sud de la Nouvelle Bretagne. Le long de la Nouvelle Bretagne coule vers l'est le NBCU qui correspond à la partie du NGCU qui va s'échapper par le détroit des Salomon. Il existe également des entrées d'eau en mer des Salomon par ce même détroit. L'hypothèse est que ce signal d'EKE serait généré par des instabilités (barotropes, baroclines??) associées au cisaillement entre le NBCU et le flux entrant par le détroit des Salomon. Ce signal d'EKE est cohérent avec la forte rms de SLA au même endroit sur la figure A. qui a une signature interannuelle importante.

Fig. E2: eke.gif

La fonction temporelle de ce maximum d'EKE (8°S-5°S; 152°E-155°E) montre de trés grandes variations avec des EKE allant de 200 (cm/s)**2 à 2800 (cm/s)**2 dans une gamme de fréquence étendue.

Fig. E3: EKE_ft.gif

Le spectre met en évidence des pics annuels et semi annuels mais les amplitudes les plus fortes sont aux échelles interannuelles (entre 2.5 et 7 ans)

Fig.E4: spectre EKE.gif




IV. Signal saisonnier

Une analyse harmonique sur la période Aviso 1993-2006 permet d'avoir le cycle annuel de SLA en mer des Salomon

Fig. F: harmo.gif

Le maximim de variabilité saisonnière est situé de part et d'autre des îles Salomon, à 5°S-158°E coté large et 9°S-158°E dans la mer des Salomon. Donc si le maximun de variabilité saisonnière en sla est localisé dans la partie est de la mer des Salomon, celle-ci augmente vers le sud. L'entrée de la mer des salomon à 11°S est marquée par l'absence de signal saisonnier. Le cycle saisonnier est en phase dans la mer des Salomon, le signal est maximum au mois de mars. Au sud de 11°S, le signal est en opposition de phase.

Fig. G: annualcurrent.gif

Au mois de mars, alors que la partie équatoriale du SEC (0-4°S) est intensifiée  on a une vidange par le sud de la mer des Salomon  qui s'échappe vers l'est et participe à l'intensification du SECC. L'anomalie du NQC, maximum à cette période ne pénétre donc pas en mer des Salomon. La situation est inverse en septembre avec une entrée d'eau par le sud via le SEC (JNV, 10°-11°S) associée à une baisse d'intensité du NQC. En sortie  de la mer des Salomon vers le nord, la circulation vers l'est est associée à un affaiblissement du SEC équatorial. En terme de courant, la circulation en mer des Salomon est concentrée dans la partie ouest du bassin, au niveau du NGCC.

Quelles explications pour ce cycle saisonnier en mer des Salomon?
A l'est des Salomons, l'inclinaison des lignes de phase est caractéristique de la propagation d'ondes de Rossby annuelles dont la vitesse augmente vers l'équateur. Ce sont des ondes de Rossby raisonnantes qui se propagent avec les anomalies de vent. L'arrivée de ces ondes sur la côte des iles Salomon est bien mise en évidence sur les Hovmuller. Le signal en mer des salomon est en phase avec l'onde de Rossby qui arrive au niveau du détroit de salomon. Le signal de downwelling en avril se traduit par une  augmentation de sla en avril et le signal d'upwelling se traduite par une baisse de sla en aout.

Fig. H: msla_hov

L'analyse harmonique du rotationnel de vent issu de ERS (1993-2000) et de Quickscat (2000-2007) montre des différences importantes entre les deux champs de vent dans la région des Salomon. Notamment le signal ERS est en phase dans la mer des Salomons, ce qui n'est pas le cas pour Quickscat. La région de changement de phase (qui correspond à la disparition du cycle annuel) visible dans ERS à 10°S au niveau de l'entrée sud de la mer des Salomon s'étends vers le nord en mer des Salomon dans Quickscat.
Il y a beaucoup de trous dans les données Quickscat dans cette région: est ce que ces données sont fiables?
La correspondance entre la SLA (Fig.F) et ERS est notable (on retrouve la même région à 10°S où le signal saisonnier disparait) ce qui pousse à avoir davantage confiance en ERS. Le vent est en avance de 2 mois par rapport à la SLA. Pour une réponse locale de la sla au vent on s'attendrait à un déphasage de 3 mois.

Fig. I: a. curlERS.gif, b. curlquickscat

La région est caractérisée par un rotationnel négatif donc le forçage saisonnier est minimum en janvier et maximum en août.

Fig.I1: ERS_clim.gif

contribution de l'effet stérique sur le signal de SLA (déphasage de 20-30 jours)??

Il semble que la SLA en mer des Salomon soit bien associé au signal remote des ondes plutôt qu'au forçage local du vent. Par quelle processus le signal de Rossby se propage par le détroit des Salomon et envahit la mer des Salomon se traduisant par une variabilité sur la partie est de la mer des Salomon?
Il semble aussi que la signature particulière du curl de vent à 10°S joue un rôle important sur la sla en annihilant le signal saisonnier et en forçant les structures de courant associées à s'orienter est-ouest.


Le cycle d'EKE ne présente pas un signal vraiment régulier, c'est pourquoi pour extraire le signal saisonnier correspondant on commence par faire une demodulation complexe autour de la période de 1 an.

Fig.J: complex demodulation (1 year) of the EKE signal.gif

Le cycle annuel du maximum d'EKE montre un pic d'énergie en avril et un minimum en octobre en accord avec l'intensité du NBCU. L'amplitude du cycle saisonnier (+/- 160 cm2/s2) est relativement faible devant la valeur moyenne ( 745 cm2/s2)

Fig.K: EKE annual cycle from complex demodulation.gif

La circulation en mer des salomon est caractérisée par des courants fort le long des côtes de PNG et de Nouvelle Bretagne: NGCC, NBCU. Ces courants permettent d'évacuer le flux entrant à la frontière sud de la mer des salomon. Ici nous tentons d'évaluer la variabilité saisonnière de ces courants à partir de l'altimétrie et de la relier à le description faite au dessus. Ces courants ont des extensions trés fines et le produit AVISO ne permet d'acceder à leur variabilité associée. Nous allons donc utiliser les données along track avec un traitement spécifique "côtier" pour accéder à l'information sur ces courants. Au vue de la taille du bassin trés peu de traces T/P sont disponibles. Les traces choisies sont les traces:
251; 188; 175; 112; 99; 225, 149

Fig.L: traces TP.gif

La variabilité du courant géostrophique cross track est calculée le long de la trace. La dérivé est calculée sur 3 points (+/- 20 km) alors que le rayon de Rossby est de l'ordre de 150 km. L'explication est que l'on a une situation anisotrope avec des courants étroits et intenses dans une direction. Le nombre de Rossby correspond à une situation isotrope, dans le cas anisotrope on se trouve avec des echelles de Rossby trés différentes selon les directions.

La trace 251échantillonne le courant entrant dans Milne bay et au nord des Woodlark. Au nord de Woodlark, la variabilité du NGCU semble  s'étendre entre Woodlark (9°S) et 8.2°S, et présente une variabilité annuelle et semi annuelle.

Fig. M1: a) t251.gif; b) North Woodlark.gif; c) NW_cl.gif; d) Mil Bay cl.gif

La trace 188 échantillone les même structures que la trace 251 plus le NBCU

Fig. M2: a) t188.gif; b) NW cl.gif; c) Milne Bay.gif;  d)Mil Bay cl.gif

Fig. M3: a) t175.gif; b)North Woodlark.gif; c) NW NB cl.gif

Fig.M4:  a) t112.gif;  b) North Woodlark.gif;  c) NW NB cl.gif

Fig. M5: a)t099.gif;   b) North Woodlark.gif;  c) NW NB cl.gif

Fig. M6: a) t149.gif

TOP


N.  Ideas for an Angelique's paper on altimetry


Sur l'idée que la disymetrie de la variabilité en mer des salomons entre l'est et l'ouest qui caractérise une difference de régime entre les 2 régions; a l'ouest les courants de bord ouest, à l'est  une activité tourbillonnaire; on peut valoriser le travail d'Angélique au niveau de l'altimétrie en:
1. utilisant le travail fait sur le traitement côtier pour mettre en avant la variabilité des courants de bord ouest (on s'interesse à la partie ouest)
2. en analysant le signal EKE associé aux eddies de la région est.


I. introduction

- La région de la mer des Salomon présente une variabilité de  sla la plus forte sur tout le Pacifique tropical Sud

Fig.A: rms Pacifique Sud.gif (rms AVISO 1992-2007)

- La mer des Salomon est un endroit clé de la circulation océanique car c'est l'endroit où transite le NGCC, le WBC qui connecte les régions subtropicales à l'équateur. Cette connection est partie intégrante des STCs. La variablité de cette circulation méridienne participe activement à la variabilité climatique de type ENSO et notamment à sa modulation basse fréquence.
La mer des Salomon se caractérise par une bathymétrie complexe. C'est une mer semi fermée connectée aux régions équatoriales par 3 détroits: Vitiaz, St Georges, et Solomon strait.

Fig.B: Circulation in the Solomon Sea + RMS SLA
- Information apportée par le papier.
- une description exhaustive de la variabilité de sla en mer des Salomon qui se traduit par:
- une information sur la variabilité des courants de bord ouest
- une analyse du signal EKE qui caractérise la partie est.

II. DATA
a. Altimetrie
1. AVISO
2. Along track
SLA
Ug

b. Modèle Salomon

III. Description générale
- le signal de SLA en mer des Salomon est un signal grande echelle. Il varie en phase avec les signaux à l'est et au nord de la mer des Salomon. La corrélation est moins bonne avec le signal de sla au sud de 10°S

Fig.C1: SLA averaged over boxes.gif;  C2: East-North.gif; C3: South.gif

- L'analyse spectrale montre que l'on 2 fréquences prépondérantes: Annuelles et interannuel.

Fig. D:
spectre.gif


- En mer des Salomon la partie est se distingue de la partie ouest par une rms plus ellevée. Cela est tres bien montré sur l'analyse spectrale des deux regions. Si les fréquences sont les mêmes l'amplitude des signaux est nettement différente. (Fig. B,D)

- La partie ouest de la mer des Salomons correspond  à une région de forts courants:
Le NGCC longe la côte ouest de la PNG (Fig.B). Dans la section IV, nous allons étudier la variabilité associée au NGCC

- La partie est est caractérisé par une zone de recirculation (Fig.B).
Elle se caractérise par 2 patchs de forte rms
: au sud de Nouvelle Bretagne et auprés de la côte sud ouest des iles Salomon.

- Ce patch au nord est associé à un fort signal EKE et semble avoir une composante interannuele forte comparée au patch sud:

Fig. E : eke.gif; Fig. F: low frequency.gif

- Le patch au sud n'est pas associé à un signal EKE et sa composante annuelle est plus importante que celle du patch nord.

Fig. G:
climatology.gif

L'étude de ces 2 patch de variabilité est présentée en section V


IV. Variabilité des WBCs
- Annuelle
L'analyse harmonique du signal annuel des données AVISO est effectuée. On en déduit la circulation associée. On retrouve une variabilité des courants similaires à celle du papier d'Angélique avec une vidange maximum par le sud de la mer des Salomon en mars et un augmentation des courants entrant en septembre.

Fig. H: annualcurrent.gif

La Fig. H montre des courants qui envahissent la mer des Salomon. Le signal Aviso est à trop grande échelle spatiale pour accéder à la variabilité intrinsèque du NGCC aux échelles montrées par la Fig. B.

Nous allons donc utiliser les données along track avec un traitement spécifique "côtier" pour accéder à l'information sur ces courants. Au vue de la taille du bassin trés peu de traces T/P sont disponibles. Les traces choisies sont les traces:
251; 188; 175; 112; 99; 225, 149

Fig.I: traces TP.gif

La variabilité du courant géostrophique cross track est calculée le long de la trace. La dérivé est calculée sur 3 points (+/- 20 km) alors que le rayon de Rossby est de l'ordre de 150 km. L'explication est que l'on a une situation anisotrope avec des courants étroits et intenses dans une direction. Le nombre de Rossby correspond à une situation isotrope, dans le cas anisotrope on se trouve avec des echelles de Rossby trés différentes selon les directions.

La trace 251échantillonne le courant entrant dans Milne bay et au nord des Woodlark. Au nord de Woodlark, la variabilité du NGCU semble  s'étendre entre Woodlark (9°S) et 8.2°S, et présente une variabilité annuelle et semi annuelle.

Fig. K1: a) t251.gif; b) North Woodlark.gif; c) NW_cl.gif; d) Mil Bay cl.gif

La trace 188 échantillone les même structures que la trace 251 plus le NBCU

Fig. K2: a) t188.gif; b) NW cl.gif; c) Milne Bay.gif;  d)Mil Bay cl.gif

Fig. K3: a) t175.gif; b)North Woodlark.gif; c) NW NB cl.gif

Fig.K4:  a) t112.gif;  b) North Woodlark.gif; c) NW NB cl.gif

Fig. K5: a) t099.gif;   b) North Woodlark.gif;  c) NW NB cl.gif

On est à même de définir pour chaque trace l'extension du courant et de montrer une variabilité cohérente avec celle d'écrite par Angelique en ce qui concerne le NGCC, NGCC_VS et NGCC_NB. C'est vrai uniquement pour la phase des signaux. L'altimétrie permet d'estimer l'amplitude de la variabilité des courants.

- Interannuelle

Fig. K6: a) t251.gif; b) Hovmuller.gif; c) WBCs.gif
Fig. K7: a) t188.gif; b) Hovmuller.gif;  c) WBCs.gif
Fig. K8: a) t175.gif; b) Hovmuller.gif; c) WBCs.gif
Fig.K9:  a) t112.gif;  b) Hovmuller.gif;  c) WBCs.gif
Fig. K10: a)t099.gif;   b) Hovmuller.gif;  c) WBCs.gif

V. Variabilité de la partie est
a. Patch nord
Fig. K11: sla.gif

Ce patch caractérisé par une forte rms en SLA est associé à un tourbillon qui a une forte signature en EKE. Ici on ne regarde que les anomalies.

- Variabilité annuelle
Relation SLA/EKE
:  Fig.L: EKE + SLA annual cycle from complex demodulation.gif (EKE:noir)
Minimum de EKE en octobre, EKE Max en avril. Le cycle annuel de la SLA n'est pas identique à celui de l'EKE. En effet la SLA est minimum en August et max en avril.
En fait le signal de SLA moyenné sur le patch correspond au signal moyenné sur la zone Salomon ce qui signifie que tout bouge en phase dans cette zone (déja decrit dans la partie III). En fait le signal de sla en mer des Salomon correspond à l'arrivée des ondes de Rossby:
Fig. M: msla_hov

puisque ces ondes de Rossby sont des ondes raisonnantes, cad elles se déplacent en phase avec le forçage, le signal de sla est egalement en phase avec le signal de curl de vent.

Le signal d'EKE correspond a des anomalies de courant cyclonique (minimum d'EKE) /anticyclonique (maximum d'EKE)

Fig. N: geostrophic current: April - October.gif; August.gif

De façon saisonnière la mer des Salomon est soumis à un flux nord/sud. On a un flux de sud max en avril associé à un max de SLA et de EKE et un flux vers le nord max en aout correspond à un min de SLA. La sortie vers le nord se fait par le detroit de Vitiaz et celui des Salomon. Les 2 détroits n'évoluent pas en phase (sans doute ils n'obéissent pas à la même dynamique). Vitiaz est en avance de 2 mois sur le detroit des salomon (aout versus octobre). Vitiaz joue le role principal ce qui explique pourquoi la sla est en phase avec Vitiaz.Le NGCC s'écoule par Vitiaz. NGCC et NGCU sont en phase. Le NGCC répond directement au signal de mousson avec des anomalies de sud est max en juillet-aout qui correspondent à un max du NGCC et des anomalies de nord ouest en mars, période de NGCC min.

Fig. N1: ERS wind annual cycle.gif; March-July ERS wind.gif
Fig.N2: ERScurl_clim.gif La région est caractérisée par un rotationnel négatif donc le forçage saisonnier est minimum en janvier et maximum en août.

Relation avec le papier d'Angélique:
Le signal d'EKE est phasé avec celle du NBCU. le maximum du NBCU correspond à un minimum d'EKE. Ca s'explique car le NBCU est un courant profond et le signal altimétrique correspond au courant de surface. Lorsque les anomalies de courant de surface sont dans la meme direction que le NBCU alors l'EKE est minimum. A l'inverse lorsque les anomaliesde surface sont de direction opposée au NBCU alors on doit avoir declenchement d'instabilité barocline et création d'un fort signal EKE.

Pourquoi a t'on création d'un tourbillon??



- Variabilité interannuelle
La variabilité interannuelle est la variabilié dominante.
Il y a une tres forte relation entre SOI, SLA et EKE

Fig.O1: lf_tseries.gif
;   Fig.O2: lf_tseries_shf.gif (vert: EKE; noir: SLA; rouge: SOI)

La SLA est en retard de 2 mois vis à vis de la SOI.
L'EKE est en retard de 7 mois vis à vis de la SLA et est en opposition de phase. Par exemple l'ENSO 97-98 s'est traduit par une baisse globale de la SLA en mer des Salomons qui traduit la vidange du reservoir d'eau chaude en décembre 1997.  Fin 1998 est associée une augmentation significative de l'EKE.
Explication?
Le signal d'EKE est associé à la vitesse à laquelle la mer des Salomon si vide ou se remplit. Il faut donc regarder la dérive temporelle de la SLA.

Fig.P: eke_sladt.gif

Les deux signaux sont bien en phase. Donc l'activité EKE est associé à l'intensité de la circulation pour les echelles interannuelles. En fait l'anomalie interannuelle marquante correspond à l'ENSO 97-98

Fig. P1: time series EKE patch nord.gif


Fig. P2: Situation au pic d'EKE: Current1298.gif

Fig. P3: Situation Nino: averagedcurrent0497-0198.gif

Fig. P4: Situation Nina: averagedcurrent0399-1200.gif

Explication:
En période Nino, le réservoir d'eau chaude  migre le long de l'équateur. Cela est associé à des anomalies de vitesse vers l'est. Cette vidange de la partie equatorial ouest (SLA <0) est compensé par un flux de nord dans la mer des Salomon. Le flux entrant par le sud ressort par les detroits de Vitiaz et Salomon, aspiré par les conditions équatoriales. L'energie EKE est alors minimum car les anomalies sont dans le sens de la circulation moyenne.

En période Nina: toute la bande tropicale 0-8S est soumise à des anomalies de courants vers l'ouest. On  a alors remplissage de la région équatorial ouest (sla >0). La mer des salomon draine vers le sud le flux entrant par les detroits de Vitiaz et des Salomon. on observe alors des anomalies de courant vers l'est du SEC au sud de 10°S. En condition Nina developpé, on doit avoir disparition du NBCU donc on a plus de cisaillement de courant dans cette région ce qui exlique le minimum d'EKE.

En période de transition: La bande équatoriale est sous l'influence de la Nina alors que la mer des Salomon n'est pas encore en équilibre avec les conditions Nina: on a deux flux entrant par le sud et par Vitiaz  et un flux sortant par le detroit des Salomon. Mais cette fois les conditions equatoriales de favorisent pas la sortie et on observe une dissipation d'energie (max d'EKE) et le flux sortant alimente des anomalies vers l'est du SEC entre 6 et 8°S

b. Patch sud
Fig. Q: sla.gif
Fig. Q1: eke.gif

En terme de sla les deux patchs evoluent de façon similaire. Ils ont le même signal interannuel, le signal annual est un peu plus fort pour le patch sud.
le patch sud n'est pas associé à un signal EKE.
Explication: C'est une  zone où la circulation moyenne est relativement faible, notamment au niveau de la thermocline. Cela limite le developpement d'instabilité barocline.

TOP
O. Paper on altimetry



Variability of the Solomon Sea from altimetry sea level data

A. Melet, L. Gourdeau, J. Verron, and F. Birol



Abstract


1. Introduction

The intense boundary currents which take place on the western side of the oceans are often characterized by high sea level variability associated with intense eddy activities. The Gulf Stream, the Kuroshio, and the East Australian Current are well known examples of such energetic areas.(ref?). In the same way, the tropical South West Pacific ocean exhibits the highest sea level variability of the entire South Tropical Pacific (Fig.A1), and its Low Latitude boundary currents are part of the SubTropical Cells (STC) system bringing back subducted waters from the subtropical gyre to the Equator (Ref?).

The South West Pacific ocean contains most of the numerous islands and archipelago that exist in the all Pacific. Between 10°S and 5°S, the western boundary of the South Pacific is formed by the Solomon Sea which is delineated by Papua New Guinea (PNG), New Britain (NB) and the Solomon islands (SI) (Fig.A2). This semi-enclosed sea has a complex bathymetry, and is connected to the Equator region through two main straits: Vitiaz, and Solomon straits. The Solomon Sea is the final transit area for the tropical waters flowing westward into the South Equatorial Current to returning northward back to the Equator, in particular through the New Guinea Coastal Undercurrent (NGCU). This WBC, flowing along the PNG coast, is the primary source for the Equatorial UnderCurrent  (EUC) (Tsuchiya et al 1989). Its variability in term of advected water masses can modulate the equatorial thermocline and Pacific cold tongue and ultimately the atmospheric circulation playing a major role in the low modulation of ENSO (Luo et al., 2005, Yeager and Large 2004, Schneider 2004). Therefore the Solomon Sea is of particular interest in a climatic context and one of the focuses of the South Pacific Circulation and Climate Experiment (SPICE) program (http://www.clivar.org/ organization/pacific/pacific_SPICE.php).

To our knowledge, little work has been specifically dedicated so far to discussing the Solomon sea circulation and its variability. Ridgway et al. (1993) analysed the sea level response to the 1986-1987 ENSO event through tide gauges, and XBT data. Recently, Melet et al. (2009) proposed a scheme for the thermocline circulation in this sea, as a result of a high-resolution model analysis.  They detailed a double system of boundary currents. The main system flowing along the PNG coast is formed both by the NGCU that permanently flows equatorward, and the New Guinea Coastal Current (NGCC) in the surface layers that may reverse as a consequence of a marked seasonal cycle in relation to the monsoonal winds (Ref?).

Little observations are available in the Solomon Sea. There are some XBT measurements along commercial ship lines from the VOS program, and only few oceanographic campaigns have been realised (Gorgone, WEPOCS 1986, 1987) in this remote region of the world ocean. Some ARGO floats made their route within the Solomon Sea but specific deployment of ARGO floats are uneasy in this region of very uneven bathymetry. Thanks to SPICE, a recent observational effort is being undertaken. Lately, the FLUSEC1 cruise (?ref?) allowed to sample the South inflow section of the Solomon Sea. Also gliders are regularly deployed since 2007 to monitor the entering flow in the Solomon Sea.

The approach of the present study is to use altimeter data to better describe and possibly understand the variability features of the Solomon Sea where high Sea Level Anomalies (SLA) are observed. The most readily available products provided by AVISO, are the gridded product which merged satellite data from different altimetric missions like Topex/Poseidon (T/P) and ERS. Such gridded data do not take account of the complex bathymetry of the Solomon Sea. Therefore, this product might be partly irrelevant for our study. This is the reason why a complementary analysis of altimetric data have been performed based on the original along track T/P data on which a new specific processing have been performed.  

The objective of this paper is to explore further the Solomon Sea annual and interannual variability as revealed by altimetry, to develop specific analysis of altimetric data in this respect and possibly to understand the origins of the various variability signals, in time and space that will emerge from the analysis. altimetry2plots give direct access to the SLA signal but sea level gradients also give access to current structures and variability. Therefore the discussion will focus on the variability of the current system of the Solomon Sea, and especially of the WBCs.

The altimeter datasets that have been used, from the AVISO database and the retreated T/P data, are first introduced in Section 2. An overall discussion of the sea-level variability in the Solomon Sea is presented in Section 3.  Then, Section 4 detailed the variability related the WBCs. Section 5 discusses the physical mechanisms at the origin of locally intense SLA variability. Summary and conclusions are presented in Section 6.




FigA1.gif;   FigA2.gif (a modifier)

2. Data and model

2.1. Gridded AVISO data

2.2. Along-track DEGEO data

FigB.gif

3. Overall variability

a. Spatial description

Sea Level variability is untimely linked to the mean circulation but existent climatology provide a too crude description of the mean state because of the relatively sparse data available and the difficulty to take account the complex bathymetry in the spatial interpolation. This is also true for the most relevant CSIRO Atlas of Regional Seas (CARS, http:). Since observational diagnoses are limited by a severe lack of data, Melet et al. (2009) have developed a high resolution (1/12°) regional model able to handle the complex bathymetry of the Solomon Sea. These authors provide a detailed description of the mean and annual circulation in the Solomon Sea at the thermocline level. They describe boundary currents which have their maximum in subsurface, but these currents extends to the surface, and most of their variability is concentrated in the first 150 m depth. Therefore, in our analysis of the surface variability, the same terminology as in Melet et al. (2009) is used but the fact that these currents are undercurrent are not put forward (by example, here the NGCC defines both the NGCU and the NGCC).

To briefly summarize their results, the NGCC is fed both by the North Queensland Current flowing along the Australian coast and the south coast of PNG, and directly by the North Vanuatu Jet that is formed by the interaction of the SEC with the northern extremity of the Vanuatu archipelago. When entering the Solomon Sea, the NGCC may flow into Milne Bay (this branch is named NGCC_MB in the following) or may go around the Woodlarks (this branch is named NGCC_NW). The two branches join when flowing to Vitiaz strait. Here the NGCC is divided in two branches: one branch crosses Vitiaz, and another branch, the New Britain Coastal Current (NBCC), flows along the south coast of New Britain, and exits at Solomon strait (Fig. A2). Different processes as equatorial dynamics, remote off-equatorial Rossby waves north of 10°S, and the spin up and down of the subtropical gyre south of 10°S explain the rather complex annual cycle at the thermocline level. This description is model dependant but as shown in the following, these model results support the analysis of the altimetric sea level data.

The Solomon Sea is embedded in a large region of high sea level variability (Fig.A1). This large region has already been analyzed by Qiu and Chen (2004) from altimetry. This region defined between 15°S-5°S, 150°E-170°W encompasses the South Equatorial Counter Current (SECC) region. The SECC is a relatively narrow jet, confined to the surface 150-m layer extending eastward from the Solomon Islands at around 8°S (Gouriou and Toole, 1993). Qiu and Chen (2004) showed that the seasonal modulation of the SECC’s EKE field, explained by barotropic instability associated with the horizontal shear of the SECC-SEC system, is at the origin of the observed variability. Their explanation doesn’t hold for the Solomon Sea where the current system is rather different. The Solomon Sea covers a relatively small area, compared to the one diagnosed by Qiu and Chen (2004), and spans approximately 8° in longitude, and 5° in latitude. The variability ranges between 6 and 16 cm rms, and presents a well-marked spatial distribution between the Western and Eastern basin (Fig.C1). The Western basin is the place of the main WBC system with the NGCC, and is characterized by relatively low SLA amplitude, particularly at the coast. On the contrary the Eastern Solomon Sea is characterized by high SLA amplitude, and could be the signature of a complex current recirculation area as shown by Melet et al. (2009). The highest variability, up to 16 cm rms, is a relatively local signal, centered at 7°S-153°E South of New Britain. This signal is referred to as the NB patch. The eddy kinetic energy computed from altimetry gives a time mean EKE level of 679 (cm/s)2 for the Solomon Sea. In fact, the EKE field shows that it is principally located in the North East part of the basin (Fig. C2) related to the NB patch, and extends to the south in the centre of the basin. The highest EKE energy level is equivalent to an rms velocity anomaly of 30 cm/s. This area, roughly the eastern side of the basin, is dominated by a high eddy activity as suggested by Melet et al (2009). At the opposite, the smallest EKE energy level bounds the coast, except at Vitiaz strait and in the south east extremity of New Britain. Therefore the mean flow of the WBC that can reach 60 cm/s dominates the western side of the basin.



FigC12.gif


b. Temporal sla description

The Solomon Sea is small enough for the sea level to evolve in phase everywhere in the basin. Spatial correlations give correlation coefficients superior to .8 for the raw data, and are as high as .9 once the high frequencies are filtered. Correlation between the Solomon Sea and the surrounding ocean falls down drastically south of 11°S but stays high in the Bismark Sea, and east of the Solomon Islands (not shown). The time evolution of SLA averaged over the Solomon Sea shows a +/- 20 cm range of variability (Fig.D). There is a clear annual signal that is strongly modulated at interannual timescale. By example, the annual signal is no visible during the 1997-1998 ENSO event. Both components explains X% of the SLA variance.


FigD.gif

Annual time scale

The annual signal is dominant in the eastern side of the Solomon Sea, and particularly in the south east (Fig.E). Spectral analysis of the time series averages over the Solomon Sea gives a 5 cm magnitude for the annual component. The maximum SLA is in March/April, and the minimum SLA is in August/September (Fig. F). The annual SLA variability induces a maximum southeastward surface geostrophic velocity anomaly during March-April that traduces a decrease of the flow entering the Solomon Sea by the southern boundary. In the same time, south of the Solomon Sea, the NVJ decreases whereas the NQC in the Gulf of Papua increases (Fig.E). The opposite situation exists in September-October. This description of surface geostrophic velocity is very close to the one in Melet (2009) on thermocline transport.


FigE.gif; FigF.gif

Since the WEPOCS cruises (Lindstrom et al., 1987), the most relevant works on the NGCC are the two papers by Kuroda (2000) and Ueki et al. (2003) who have analysed data from moorings deployed especially to observe the temporal variability of the NGCC and the NGCU. The most southern position of the moorings is at 2.5°S, north of the Solomon Sea, in a dynamical regime highly influenced by equatorial conditions. The westward NGCUC was observed to persist year around, and it is intensified in boreal summer, whereas seasonal reversal of the surface intensified NGCC was clearly observed. In the boreal summer characterized by the south-easterly monsoon, the direction was northwestward. In the boreal winter the direction was southeastward in response to the northwesterly monsoonal winds. In fact, the wind forcing is strongly related to variation of the NGCC, and the highest correlation coefficient between both were obtained when the surface current were shifted by 2.5 day ahead (Kuroda, 2000). Ueki et al. (2003) explain that northwesterly alongshore winds drive offshore surface Ekman flows, thereby inducing upwelling near the coast of New Guinea and subsurface onshore flows. Consistent with the upwelling, a southeastward coastal surface jet was generated associated with the shoaled pycnocline near the coast. In the case of the southeasterly alongshore winds, the opposite results occurred. The strong surface southeastward current in winter is considered as this coastal surface jet forced by monsoonal wind, of which generation mechanism was studied on the current off the west coast of Somalia by Mc Creary and Kundu (1985).  The situation could be different in the Solomon Sea which is bounded by the PNG to the west and the Solomon Islands to the east. When monsoonal winds induced an upwelling in the west coast, it also induces a downwelling at the east coast. Both contribute to enhance a southeastward surface jet.We can ask why variability are not maximum along the coast and minimum at mid basin?? Five moorings with current meters were deployed across the Vitiaz strait sill from February 1992-April 1993 (Murray et al., 1995). In august 1992, during the southeasterly wind monsoon, the NGCC is in phase with the northwestward flowing NGCUC, and speeds reach 110 cm/s at mid strait at the surface. In march 1992, there is a low flow period during the north west monsoon, and surface layer seeds just reach 80 cm/s. So no reversal of the NGCC was observed. The subsurface maximum speed characteristic of the NGCUC, clearly seen in the 1985-1986 ADCP data of Lindstom (1987) is notably absent from the current meter data fields. It appears that increased energy in the surface layer (the NGCC) is able to mask the subsurface speed maximum of the undercurrent.


In the Solomon Sea,  Melet el al. (2009) find that the annual cycle of the thermocline circulation is mostly explained by three regimes of the basin-wide wind forcing. First, the annual march of the ITCZ produces equatorial waves that control EUC variability, and therefore its drawing of off-equatorial water during its austral spring-summer surge. Second, the same forcing generates resonant Rossby waves in the off-equatorial central Pacific. North of 10°S, this remote forcing modulates the SEC inflow to the western boundary. Third, the annual march of the tradewinds generates stationary, in-phase Rossby waves south of 10°S. This is responsible for the spin up and down of the subtropical gyre, which modulates the NVJ and the NQC at its equatorward edge (Kessler and Gourdeau 2007). The annual cycle of the western boundary currents of the Solomon Sea responds to those dynamics driven by the basin-wide wind forcing which modulates incoming and outgoing transports of the EUC, the SEC, and the NVJ. For instance, the NGCU is maximum in October in Milne Bay, in phase with the NVJ, whereas the eastern part of the NGCU owing in the central basin is advanced in phase, with a maximum transport in August-September. North of the Woodlark Islands, the NGCU is maximum in September. In the northern Solomon Sea, the two branches of the NGCU have different timing. The NGCU is maximum in August along the northern coast of PNG, whereas the NBCU is maximum in October. North of 10°S, in the open ocean, the seasonal variability is dominated by a remotely-forced Rossby wave regime. As the seasonal wind forcing and the ocean response propagate westward at similar speeds, the Rossby waves are resonant (Chen and Qiu 2004). The Rossby waves are maximum with a 6 cm amplitude between 9°S-5°S and modulate the SEC (Ref). These Rossby waves impinge the Solomon Islands with a phase lag reflecting the decrease of the Rossby wave speed with latitude. Downwelling Rossby waves arrive at 6°S in March-April whereas they arrive at 11°S (south of San Cristobal) in August-September. The phase lag with latitude existing east of the Solomon Islands doesn’t exist in the Solomon Sea. It argues for a significant contribution of the Rossby waves arriving at 6°S, the latitude of the Solomon strait, to explain the seasonal cycle observed in the Solomon Sea (Fig. D).



Interannual time scale

The interannual signal is the dominant signal in the Solomon Sea with a 7 cm rms magnitude as given by a spectral analysis of the time series averaged over the Solomon Sea. Most of the interannual variability is located in the eastern Solomon Sea (Fig. G), with the highest variability at the NB patch. Sea level is directly linked to ENSO, as represented here by the SOI index (Fig.D). Therefore, the Solomon Sea variability is related to the large scale ENSO variability which is abundantly discussed in the literature. Most of the recent papers focus on the theoritical “recharge oscillator” paradigm of Jin (1997). The oscillation proposed by Jin (1997) results from phase lags between changes in the zonally averaged main thermocline depth and changes in the eastern Pacific SST, and zonal thermocline tilt, the latter two being in phase with one another. The zonally averaged main thermocline depth integrated over the entire equatorial Pacific defines the Warm Water Volume (WWV; Meinen and McPhaden, 2000). WWV changes, affected by the meridional transports resulting from the anomalous eastward (westward) zonal winds near the equator, are representative of the discharge (recharge) of the warm water to (from) higher latitudes. It is the incompletely balanced meridional interior transport by western boundary current transports on interannual time scale, which leads to low frequency changes in the WWV near the equator associated with the ENSO cycle. In another way, changes in the zonal thermocline tilt affect WWV defined only over the western equatorial Pacific (WWV-w). Meinen and McPhaden (2000) confirm that variations of WWV changes were consistent with Jin’s hypothesized oscillator. Correlation analysis indicated that the peak correlation occurs with SST lagging WWV by seven months.


FigG.gif

The net meridional transport at the origin of WWV changes is the sum of both interior Sverdrup transport and WBC transport. A model study by Lee and Fukomori (2003) shows that the variations of the boundary transport is smaller than that of the interior geostrophic transports, and they are found to be generally anticorrelated to each other. The variability in off-equatorial wind stress curl in the western Pacific changes the strength of horizontal circulation and results in a variation of boundary flow that is opposite in direction but comparable in magnitude to that of the interior geostrophic flow. The variability of near equatorial zonal wind stress primarily affects the strength of the shallow meridional overturning circulation with net geostrophic flow opposing the surface Ekman flow. The covariability of these two forcings leads to an enhancement of interior transport. In fact, these transports are highly influenced by meridional asymmetry between the northern and southern hemisphere. First The PV front associated to the ITCZ deflects water and keeps it from flowing northward. As a consequence the WWV discharged from the equatorial band reached much farther to the south than it did to the north and the WWV flow toward the equator from the south in the western Pacific did not occur until the WWV discharge from the equatorial band was already underway (Meinen, 2005). As a consequence, the refilling of the western subtropical Pacific during the 1997-1998 El Nino event, occurs roughly 6-8 months earlier in the northern hemisphere than in the southern hemisphere (Meinen, 2005). A model study by Kug et al. (2003) shows that the net ENSO-related meridional mass transport in the northern hemisphere is larger than that in the southern hemisphere. The antisymmetric characteristics are mainly due to a southward shift of the maximum zonal wind stress anomaly during the ENSO mature phase, and the north-south antisymmetric geometry at the western boundary that generates a relatively stronger WBC in the SH than in the NH. The southward shift of the zonal wind stress is at the origin of strong wind stress curl in the south western Pacific that generated large negative SLA. Accordingly, the zonal slope of the sea level is steeper in the SH so that the geostrophic current is stronger in the SH. Recently, Ishida et al. (2008) using the MOM2 model confirm that the interior transport in the South Hemisphere is larger than in the NH; however the WWV variability is determined by the net transport variability in the NH associated with the lags of the interior transport and WBC transport. Negative wind stress curl in the north equatorial Pacific after the mature phase of El Nino is key to generating the lagged northward WBC transport. In the south hemisphere most of the interior transport is compensated with the WBC transport in the south hemisphere.


Two main EOFs modes are representative of the interannual sea level variability in the tropical Pacific (Alory; Meinen), explaining repectively X% of the SLA variability. The first mode is the well known ENSO mode characterized by a zonal tilt (Fig. G), and the second mode, characterized by a meridional seasaw at 5°N, is seen as a recharge/discharge mode. The first mode, in phase with ENSO, traduces the zonal migration of the warm pool, and by consequence is related to the WWV changes in the western equatorial Pacific (depletion/repletion of the warm pool) as shown by Meinen (2005, see their figure 4). Whereas the second mode is related to the WWV changes over the all equatorial Pacific, and leads ENSO by 7 months (Meinen, 2005). It is notable that 96% of the variance of the low frequency signal in the Solomon Sea is explained by the first EOF mode described above meaning that a strong relation ship exist between the exchange of warm water between the eastern and western Pacific, the WWV changes in the western Pacific, the changes in strength of the South equatorial Currents and the sea level variability in the Salomon Sea (Fig. D).
The surface geostrophic velocity anomaly field deduced from the spatial EOF1 mode in SLA is plotted for the Solomon Sea area at the date of the mature phase of El Nino (Fig.G).  In the equatorial band the south east velocity anomaly shows the depletion of the WWV-w during El Nino.  South of 10°S a westward surface geostrophic flow anomaly enters the Solomon Sea, and goes north. It bifurcates at the New Britain coast before to escape through Vitiaz and Solomon straits. Just a small part of this westward flow anomaly extends to the Australian coast decreasing the NQC. Therefore the interannual variability of the Australian WBCs could be mostly controlled by the SEC inflow around 10°S. Similar conclusion has been done By Kessler and Gourdeau (2007) when looking at the annual cycle. The circulation in the Solomon Sea counter balance the interior flow, therefore the WBCs are in phase opposition with the interior circulation. It means when there is a divergence of the interior flow, in the same time, there is equatorward anomalies of the WBC to feed the western equatorial Pacific.
To summarize, SLA inside and surroundings the Solomon Sea evolves in the same way at seasonal and interannual time scales. The marked dissymetry of the variability features between East and West of the Solomon Sea is at the origin of the variability of the surface geostrophique circulation. Higher variability are concentrated in the eastern Salomon Sea, particularly in the NB patch associated to a high EKE signal. The circulation in the Solomon Sea is principally related to the WBCs. During the austral spring (autumn) SLA is maximum (minimum), and surface geostrophic velocity shows an exiting (entering) flow at the southern boundary. At interannual time scale, the WBC counter balance the interior flow in term of variability. In the two next sections, we detailed the variability of the WBC, and the dynamical processes at the origin of the NB patch


4. Variability of the Solomon Sea western boundary current

The circulation in the western Salomon Sea is characterized by a surface NGCC and a thermocline level NGCU.  These currents are supposed to be thin coastal currents. So, the gridded SLA data cannot be used to track them. Along track data have been carefully processed in coastal area (see section 2). Given the relatively small size of the basin, the available tracks are not so numerous since only 10 T/P tracks span the Solomon Sea. We considered only the T/P tracks crossing the western boundary current.  Three descending tracks (251,175, 99) are considered that are sufficiently orthogonal to the WBC to favourably determine it. Track 251 crosses Milne Bay and the NGCC at the eastern extremity of the Woodark Islands. More to the West, the T/P175 track samples NGCC along the northern shelf break of the Woodlark Islands, and gives an additional insight of the NBCC flowing along the southern coast of New Britain. Noted that this track is not sufficiently oriented in the cross-NBCC direction to really interpret the derived geostrophic along shore current. The T/P 099 track is useful to provide information on the current when arriving at Vitiaz strait. Our dataset permits monitoring the variability of the boundary current to within about few km of the shore defined as the shelf break intersecting the 1000 m isobath. Data gaps are small enough to adequately resolved periods greater than 2 months.

In the considered latitudinal band, the Rossby radius is of order of 100 km (ref.), larger than the suspected signature of the WBC along the altimetric tracks crossing these currents. The Rossby radius, as computed by X, is relative to the open ocean, and relatively isotropic flow. Here, the WBCs present an anisotropic structure. They are elongated on shore with a small width in the off shore direction. Therefore, Rossby radius is different following the considered direction, and is around X km in the offshore direction of interest. The altimeter data is first smoothed because the velocity filed is expected to adjust to the density field (and therefore to sea level) only at scales larger than the Rossby radius of deformation prior to derive the cross-track surface geostrophic current anomaly.

Because the WBCs are supposed to have a marked seasonal cycle, the location of the WBCs along the track is repaired thanks to the annual current anomaly. The WBCs are characterized by coherent cross track current anomaly extending off the coast, and away from the coast the circulation is supposed to be more chaotic. We first look at the spatial coherency of the signal, and we reparired the point of maximal annual variability corresponding to the WBC. These points are labelled in Fig. H. The width of the WBC is determined by the distance where the correlation with points of maximum variability began to be smaller than 0.85. Table 1 gives the precise information about the WBC as repaired along the track. The WBC anomalies are defined as the current anomalies averaged over the width as defined in Table 1.

Fig H.gif

Table1:
A1         A2        B1        B2        C1
Longitude (°E)    153.5        153    -8            -6            -7
Latitude (°N)        -8.6        -9.8    150.9        151.6    148.4
Width (km)                85        68    77            52            119

Annual cycle (Fig. Ia)

The NGCC in Milne Bay (A2) is minimum in March and maximum in September with a 15 cm/s amplitude. The NGCC at the eastern extremity of north Woodlark (A1) is minimum in January-February and maximum in July-August with a 10 cm/s amplitude. More to the West along north of Woodlark, the NGCC (B1) is minimum in March and maximum in September with a 18 cm/s amplitude. In front of Vitiaz, the signal (C1) is wide (119 km) and the amplitude is high (25 cm/s) it is in phase with the signal in A1 (minimum in February and maximum in July-August) Along New Britain the current (B2) is in phase with the current along the Woodlarks (B1) with a 15 cm/s magnitude.

There are some observational evidences of the seasonal cycle of the NGCC. Kuroda (2000), Ueki et al. (2003) analysed time series of current data from ADCP moorings at 2°S, 142°S and 2.5°S,142°E. They found a seasonal reverse of the NGCC associated to the wind forcing monsoon. The NGCC is maximum in boreal summer characterized by the southeasterly monsoon, and minimum in boreal winter characterized by the northwesterly monsoon. The NGCU has a similar annual cycle even if always flow equatorward. Measurements at Vitiaz strait (Murray et al., 1995) show similar results except that the reverse of the NGCC was not observed in 1992.

altimetry2plots provides similar results for the surface current in front of Vitiaz (C1) but as shown some lag exist in the Solomon Sea between the different locations where the WBC are analyzed which cannot be explained by local wind variability. In fact, the phase of the surface current looks like very muck to the phase of the NGCU as described by the model study of Angelique et al. (2009) (see their figure 4). The NGCU flowing into Milne bay and crossing the Woodlark is maximum in September-October in phase with the NBCU (as the variability of point A2, B1 and B2 shown here), and lags the NGCU at Laughlan (A1), and at Vitiaz (C1) by 1-2 months.

The direct forcing by the monsoon doesn’t hold for the NGCU but we may suspect that similar forcing hold both for the NGCU and the NGCC in addition to the locol monsson forcing. In fact the two months current fields at thermocline level reveal that in July-August the signal is phase locked with the Rossby wave at 8°S entering the Solomon Sea through Indispensable strait, and there is a direct pathway from Indispensable strait, laughlan and Vitiaz. The maximum of the NGCC in A2 and B2 in September seems related to the annual cycle of the NVJ. Part of it enters the Solomon Sea by Milne bay crossing the Woodlaks by the channels.

In summary, this analysis using the DEGEO data reveals the fine structure of the western boundary currents. Interestingly it confirms the scheme provided by the numerical model by Melet et al. (2009). In particular, it confirms some important feature that has been devised on the branching of the NGCC/NGCU at Vitiaz strait: one branch of the NGCU continuing North and one branch continuing eastward and denoted as NBCU by Melet et al. (2009). The surface annual variability of the NGCC is relatively similar than the NGCU one. It could be surprising because the NGCU isa  relatively deep current but most of its variability is in fact conditioned by the NVJ and the Rossby wave north of 10 ° S which have  a strong signature at the surface.

Fig I.gif

Interannual

At interannual scales, the WBCs are conditioned by the large scale ENSO forcing. Anomalies up to 30 cm/s are observed (not shown). altimetry2plots could help to study the interannual response of the southwestern Pacific boundary to ENSO event. The inflow entering the Solomon Sea by the south is divided in two parts, one flowing through Milne bay, the other one contouring the Woodlarks. These two branches might correspond to A1 and A2 respectively. Recent information from gliders shows an e-folding of 200 m for the surface current anomalies. The spatial integration of the surface current summed up for the two branches gives an estimation of transport anomalies at ENSO time scale (Fig.Ib). A 8 Sv transport anomalies is associated to the peak of the 1997 El Nino. Thanks to in situ tide gauges and XBT data, Ridgway et al. (1993) have already studied the sea level response of the Solomon Sea to the 1986-1987 ENSO, and they provide estimation of  anomalous geostrophic transport anomalies in the Solomon Sea trough the simple expression gdHD/f where dH is the sea level difference from tide gauges between Port Moresby and Honiara, and D is a e-folding depth of 150 m for the anomalous current approximated from XBT data. The same method can be applied with the altimetric product. The same D than Ridgway et al. (1993) is used. The sea level difference between PNG and Solomon islands can be done using alongtrack data or Aviso gridded data. Both data set give similar result, therefore we choose the longest time series that is the Aviso gridded data. Both estimations are in good accordance (Fig.Ib) and the complete series is shown on Fig.J. The strongest are related to the 1997-1998 ENSO event. As already mentioned in section 3., SLA in the Solomon Sea is well correlated with WWV-w. In the same way, temporal variation of the WWV-w is relatively anti correlated with the anomalous of the WBC transport. When there is a discharge of the WWV-w (dWWV-w/dt < 0), the WBCs are higher than normal bringing more water in the western equatorial Pacific to compensate the discharge. In fact the WBC leads the variation of WWV-w by 4 months and the correlation between both values is 0.7.

Fig J.gif

Same estimation of transport anomaly was done for Vitiaz and Solomon strait choosing points on each side of the straits. The sum of the transport at Vitiaz and Solomon balance the transport at the south boundary, therefore, we have confidence in the estimation for each strait (Fig Jb). The highest anomalies are observed at Solomon strait. They reach 6 Sv whereas at Vitiaz strait  anomalies are less than 4 sv even during the 1997-1998 ENSO. Transport at Solomon strait evolves relatively in phase with dWWV-w/dt whereas transport at Vitiaz strait leads dWWV-w/dt by 7 months.
These calculations are supported by model analysis.



5. The New Britain patch

It is characterized by high rms (Fig.1), high EKE (Fig. 6) and traduces an eddy activity (Fig.5).


a. The time evolution of the EKE signal corresponding to the patch shows both an annual and inerannual signal (There is also semi annual signal): Fig11.gif

The mean EKE signal is high (670 cm2/s2). The annual variability is relatively low compared to the mean EKE, Whereas interannual variability has large amplitude. In fact, during the 1993-2006 period. The interannual EKE signal is mainly  associated to a peak in 1998-1999 that must be related to the 1997-1998 ENSO event.

b. At annual time scale, the EKE is max in March when anticyclonic eddy is well developped. The EKE is  minimum in October when cyclonic anomaly exist. Firstly, I think that the annual modulation of the vertical shear between the NBCC and the NBCU could be at the origin of baroclinic instability but we have a situation where the EKE is max when the shear is minimum. Vertical section of the current at the different time shows that barotropic instability could work. The NBCC is highly variable and reverse during the year. The horizontal shear could be between the NGCC and the SEC. The EKE variation are related the the SEC anomaly associated to the Rossby waves.

Fig12.gif

c. At interannual time scale, the EKE evolution (red) is intimely related to the time variability of the SLA signal (black) that is the speed of depletion/repletion of the Solomon Sea. The max of EKE at the end of 1998 is phased with the transition period between EL Nino and la Nina.

Fig13.gif

during El nino, there is northward flux anomaly  entering the Solomon Sea. During La Nina, there is southward flux anomaly exiting the Solomon. The phase transition is marked by La Nina condition in the equatorial band but always by El Nino condition more to the south (10°S). Therefore, from the north southward anomalies enter by the straits, and from the south northward anomaly enter the Solomon Sea by the south. So in this case there is some difficulty for the flow to exit and we have energy disspation by a high EKE level.

Fig14.gif

TOP













Document Actions

logo cnes logo IRD Logo université de Toulouse Logo université Paul Sabatier Logo CNRS
Logo bibliothèque OBS Logo Observatoire Midi Pyrénées