For the upper and lower thermocline the penetration of salty water from SPSW is controled by the mixing induced by internal tides in Halmahera and Seram Sea Due to its strategic geographic position, the Indonesian ThroughFlow (ITF), the only low-latitude passage between major ocean basins, has always been suspected to play an important role in the ocean circulation and regulation. The main route consist of a part of the Mindanao current advecting North Pacific Water. The second route divides South pacific waters consecutivly through the Halmahera, Seram and Banda sea (Gordon and Fine, 1996). Its upper part represents only 10% with no penetration of the saltier SPSW into the upper thermocline of Banda (150 m) but the signature of this water mass is visible in the lower thermocline (Gordon 1996, 2005). The Pacific salinity maximum of the thermocline water disappears as it reaches Banda Sea to form a unique characterized water mass with homogeneous salinity bellow 20°C (hautala & reid 1996, …) (34.58 psu, box 8, fig 1). It is commonly accepted that Banda plays a key role in the mixing. (Gordon 2005) In fact the transformation are so intense that diffusion advection model calculated an averaged vertical diffusivity of 1 cm²/s to reproduce the observed water mass modification from Pacific Ocean to the central Banda Sea (Ffield and Gordon 1992 and Hautala and Reid 1996). But this value is a measure of vertical mixing integrated along the flow path, therefore these models do not answer the question whether the mixing occurs at sills and boundaries, or in basin interiors. Many studies suggest that internal tides are responsible for this transformation (Schiller (2004) and Simmons et al., 2004, hatayama 2004, Robertson 2005). Indeed, the energy transferred from barotropics to baroclinics tides that generate internal tidal is highly concentrated in this region (fig1) (15 % the total global transfer). The Indonesian archipelago forms a unique place in the world that gather a strong internal tides generation and no possibility to radiate them away (semi closed basins). The issue of where and how the transformations of the water mass happened is still unanswered. In this study, we aim at investigating water mass transformation. To that end we use an OGCM with a specific parameterization to mimic the internal tides effect in this particular region. Contact : Influence of internal tidal mixing on the water mass transformation in the Indonesian Throughflow Ariane Koch-Larrouy (1), Gurvan Madec (1), Pascale Bouruet aubertot (1), Theo Gerkema (2), Laurent Bessière (3), Agus Atmadipoera (1), Robert Molcard (1) (1) Laboratoire d’Océanographie : Expérimentation et Analyse Numérique (LOCEAN), (2) NIOZ, Laboratoire des Ecoulement Géophysique et de l’Océanographie Spatiale (LEGOS) 1 - Introduction 2 - Model We use the global NEMO/OPA ocean model [Madec et al.1998] with 0.25° horizontal resolution (Barnier et al. 2006). The domain extends from 95 E to 145 E over 25 S to 25 N, with open boundaries conditions from a global climatological simulation. The model is forced by a daily climatology derived from weekly ERS 10-year ( ) wind stress. Surface heat fluxes and evaporation are computed with climatologies from NCEP/NCAR and observations using bulk formulas. Surface flux used an additional relaxation to the surface salinity of Levitus et al.[1998]. Alford, M. H., M. C. Gregg, and M. Ilyas, 1999, Diapycnal mixing in the Banda Sea: Results of the first microstructure measurements in the Indonesian Throughflow, Geophys. Res. Lett., 26(17), 2741–2744. Blanke, B., and S. Raynaud, 1997: Kinematics of the Pacific Equatorial Undercurrent: a Eulerian and Lagrangian approach from GCM results. J. Phys. Oceanogr., 27, Bessières L, Madec G, Lyard F, Le Provost C (2006) Improved tidally driven mixing in a numerical model of the ocean general circulation. Ocean Modell, submitted for publication. Egbert GB, Ray RD (2001) Estimates of M 2 tidal energy dissipation from TOPEX/POSEIDON altimeter data. J Geophys Res 106: Ffield, A., and R. Robertson (2005). Indonesian Seas finestructure variability, Oceanography, vol 18, December, Ffield, A. and A. L. Gordon, 1992 : Vertical mixing in the Indonesian thermocline. J. of Phys. Oceanogr., 22 (2), Ffield, A. and A. L. Gordon,1996, Tidal mixing signatures in the Indonesian Seas, J. Phys. Oceanogr., 26, Gordon, A.L.,(2005), Oceanography of the Indonesian Seas and Their Throughflow. Oceanography 18(4): December Gerkema T., Lam F.–P. A. and Maas, L.R.M. Internal tides in the Bay of Biscay: conversion rates and seasonal effects, Deep Sea Research Part II: Topical Studies in Oceanography, Volume 51, Issues 25-26, December 2004, Pages Hautala, S., J. L. Reid, and N. A. Bray, 1996: The distribution and mixing of Pacific water masses in the Indonesian Seas. J. Geophys. Res., 101 (C5), 12,375-12,390. Hatayama, T., 2004: Transformation of the Indonesian throughflow water by vertical mixing and its relation to tidally generated internal waves, J. of Oceanog., 60, Jayne SR, S t Laurent LC (2001) Parameterazing tidal dissipation over rough topography. Geophys Res Lett 28: Le Provost C, Genco ML, Lyard F, Vincent P, Canceil P (1994) Spectroscopy of the world ocean tides from a finite element hydrodynamic model.. J Geophys Res 99: Madec G., P. Delecluse, M. Imbard, and C. Lévy, 1998: OPA 8.1 Ocean General Circulation Model reference manual. Note du Pôle de Modélisation, Institut Pierre-Simon Laplace, N°11, 91 pp. ( Schiller A., 2004, Effects of explicit tidal forcing in an OGCM on the water-mass structure and circulation in the Indonesian throughflow region, Ocean Modelling, Volume 6, Issue 1, Pages Simmons H. L., Jayne S. R., St. Laurent L.C. and Weaver A. J., 2004, Tidally driven mixing in a numerical model of the ocean general circulation, Ocean Modelling, Volume 6, Issues 3-4, Pages Influence of tidal mixing NP surfNPSWNPCWNPIWSP surfSPIW Transport (Sv) Resident time (year) Mean depth (m) Mean temperature °C Mean salinity (psu) Mean density Banda Sea seasonal horizontal blender Fig 5 : Salinity at 325 m for December (left panel) and August (right panel). Black arrow represent the seasonal circulation During Northwest Monsoon : - Salt penetration in Banda due to a southward flux from november to february (not shown, represented by black arrows in left pannel fig3) - Water from Makassar stay south of 6°S and goes through Ombai Sait and Timor Passage without recirculating in Banda Sea During Southeast Monsoon : - water from makassar goes to Malukku sea and recirculate in the north western part of Banda Sea before exiting Fig 6 : stream function of lagrangian calculation (ARIANE, B. Blanke) for western route (left pannel) and eastern route (right panel) for TIDES exp : no Penetration of SPSW in the upper thermocline and weak signature in the lower thermocline in Banda sea as observed (Gordon 2005). for NOTIDES exp : In the upper thermocline salty water from SPSW invade the entire Banda Sea. In the lower the salinity that penetrate is too salty. how can the internal tides prevent a salt invasion in the upper thermocline ? -> the salinity in Seram sea is so reduced that the southward flow in the winter time (see fig4) export freshened water from SPSW. Fig 3 : Central 2D maps of energy transfer Logarithme from barotropic to baroclinic tides from Le Provost & Lyard Only red patterns can generates a significant vertical diffusivity. In the average boxes from 1 to 9 T,S diagram are calculated for the observations (World Ocean Data Base 2001 and Levitus 98, in black) and for the model experiments (NOTIDES in blue, TIDES in red). In these boxes are also calculated the averaged vertical diffusity given in cm2/s. Vertical scale is the depth from 0 to 2000 m. Fig 4 : Salinity in the upper and lower thermocline for the TIDES and NOTIDES run Banda Sea may have a seasonal role of an horizontal blender that mix water from North Pacific to South Pacific The parameterization improved the model reproducing the highly localized transformations all along the routes Vertical mixing du to internal tides of the SPSW occurs before entering Banda Sea Vertical diffusivity in Banda is not important compared to Flores, Seram and Timor 6 - Influence of tidal mixing in the thermocline T. Gerkema 2004 Tidal model Internal tidal model 2D Maximum of energy in the thermocline E(x,y) F(z) Maximum of energy in Molucca Halmahera Sea Jayne, St Laurent 2001 Where on the horizontal ?Where on the vertical ? Ref Le Provost & Lyard Parameterization of vertical diffusivity Because the Indonesian archipelago is the unique place in the world that gather strong internal tides generation and no possibility to radiate them away (semi closed basins), a specific parameterization is necessary to reproduce the physic of the vertical mixing. On the horizontal the distribution of energy is given by the tidal model of Le Provost et al. (1994). Whereas the vertical distribution is inferred from a 2D linear model (Gerkema et al. 2004). The Jayne & St Laurent parameterization is used. It allows the system to evolve with the stratification. Two 10 years experiments : TIDES with parameterizatio, NOTIDES without 4 - Routes Tab 1 : transport, resident time, etc..calulated with lagrangian method, for the surface water (NPsurf), subtropical water (NPSW), central water (NPCW) and intermediate water (NPIW, SPIW). In blue values on pacific sections (north or south depending of the water mass) in red values in the Indian Ocean section Salinity maximum of the NPSW is erased and replace by fresher water NPCW and NPIW become warmer and lighter Mean resident time is 6 months except for deep eastern route of the SPIW at 1000 m that stays in the ITF 10 years at minimum The total transport in the model is 16.6 Sv (1 Sv 10 6 m 3 /s) in good agreement with the Island Rule calculation (Godfrey1999). The paths within the Indonesian archipelago are well represented by the model, even if the total transport is overestimated. Indeed, the major route is going through Makassar with a North Pacific origin. This flow is divided into the 3 exits in a good proportion compared with observations. Part of the flow recirculates in west of Banda Sea before exits through Timor passage. On the eastern route, a deep flow enters the Indonesian seas from the south pacific via Maluku and Lifamatola Strait. Furthermore, in the density range of the salinity maximum of the SPSW in the model ( sigma0, that correspond to m) the net flow is northward (with a NP origin). This brief validation gives us confidence in studying the water mass transformation in the model. Fig2 : barotropic stream function of TIDES exp, quasi identical to the one of NOTIDES exp. Contours are in Sverdrup 8 a) - Perspectives : quantify route properties Next ? describe the evolution of properties along the route. The observed SPSW salinity maximum is strongly eroded from its entrance in Halmahera sea (box 5) to locally vanish around Seram Sea (box 6-7). Therefore contrary to the common acceptance that the mixing occurs in Banda Sea these observational results clearly show that this mixing already happened before entering Banda. In the NOTIDES experiment, the salinity maximum is gradually eroded but its signature is still visible in the Banda sea in contrast to what observed. In the TIDES run, the vertical mixing induced by internal tides improved the T,S properties in all the different sub basins. In particular, for the eastern route, the salinity maximum is attenuated in all the sub basins as the observations, suggesting that the transformation happened in the right place all along the route. In particular, the disappearance of the SPSW signature at the exit of Seram Sea is well reproduced by the model. The comparison between the two experiments suggest that vertical mixing induced by internal tides in Halmahera and Seram Seas is responsible for the transformation of the SPSW before entering Banda Sea. Indeed, the parameterization of internal tides can generates vertical diffusivity as high as 4 cm2/s in the thermocline in average for the Seram and Halmahera seas (boxes 5-6), with local maximum of 20 cm2/s. Furthermore the average vertical diffusivity in the Banda sea is about 0.1cm2/s with local maximum that does not exceed 0.5 cm2/s. Fig1 a : Energy transfer logarithme from barotropic to baroclinic tides Fig1 b : Energy of internal tides generated in Timor Passage(left panel), energy profil in Searm and Flores Sea (right panel) 8 b) - Perspectives : Comparison with INSTANT data