the only tropical inter-ocean exchange site (~15 Sv) transports heat and freshwater from Pacific into Indian Ocean pressure gradient between Pacific

Transcription

1 Upper-Ocean Processes and Air-Sea Interaction in the Indonesian Seas Janet Sprintall, Scripps Institution of Oceanography, USA Arnold L. Gordon, Asmi M. Napitu, LDEO, USA Ariane Koch-Larrouy,, LEGOS, France Susan E. Wijffels,, CSIRO, Australia

2 the Indonesian Seas and Throughflow (ITF) the only tropical inter-ocean exchange site (~15 Sv) transports heat and freshwater from Pacific into Indian Ocean pressure gradient between Pacific (high) and Indian Ocean (low) (Wyrtki, 1987) ascending branch of Walker Circulation closely coupled to the Australasian Monsoon system, MJO, ENSO and IOD extends across Indo-Pacific warm pool many many islands, deep basins, wide and shallow marginal seas mixing from strong tides and enhanced air-sea heat and freshwater fluxes

3 Strong Mixing Regional T-S Plots (color coded by depth) Koch-Larrouy et al. (2007) Sprintall et al., (2014) Signatures of the Smax in the thermocline and Smin in the Intermediate layer disappear quickly in the Indonesian seas through vigorous mixing from tides, air-sea interaction and complex bathymetry to form cool and fresh Indonesian Seas Water masses

4 Processes that Drive SST Variability Annual SST: Monsoon driven upwelling not the whole story! February August NW Monsoon: Also large P -> high R/O warms SST by limiting latent heat release & mixing. Freshwater caps trap heat in near surface SE Monsoon: Wind driven upwelling and Arafura shelfbreak cools SST Kida and Richards, JGR, 2009 Kida and Wijffels, JGR, 2012

5 Seasonal SST: Role of the ITF in Summer Standard Deviation Note weak seasonality in Flores Sea. What other processes are important to SST here? Flores Sea: Control NO-ITF Heat balance shows in austral summer, lateral advection warms SST while vertical mixing and Q cools SST, i.e the ITF impacts SST by advection of warm water Kida and Wijffels, JGR, 2012

6 Resolving Mixing is Important Annual SST and precipitation differences in coupled simulations with/without tides OLR (Rainfall) Models that include tides can reproduce the rainfall (Δ20%), SST (Δ2 o C) and heat flux (Δ20 Wm -2 ) patterns observed in the Indonesian Seas than those without Koch-Larrouy et al (2010) Kida and Wijffels (2012) Jochem and Potemra (2008) Sprintall et al. (2014)

7 Intraseasonal SST Contribution of intraseasonal variability to total SST variance Largest intraseasonal SST variance (55-60%) found in Banda Sea, Timor Sea, and in the Sulawesi Sea. Napitu, Gordon and Pujiana, submitted, 2014

8 Intraseasonal SST Coherence of SST and OLR at intraseasonal timescales. Highest amplitude in Banda Sea and Timor Sea OLR leads SST by 1-2 weeks Napitu, Gordon and Pujiana, submitted, 2014

9 Mixing Impact on Intraseasonal OLR Std dev OLR bandpass filtered (30-1d) Coupled model with tidal parameterization reduced intraseasonal variability in good agreement with observations OBS CTL TIDES Koch-Larrouy et al (in prep)

10 Oceanic Processes Influence SST Too! Napitu, Gordon and Pujiana, submitted, 2014

11 Indo-Pacific MJO Surface Forcing OLR U-WND STRESS Q o MJO Composites (Nov-Apr) based on Wheeler and Hendon (2004) Index Active MJO phase convective cells (OLR<0) lead strong westerly wind anomalies in Indian Ocean (esp. IAB in phases 5-7) but are more aligned in Pacific, and so have different impacts on the SST and mixed layer. Drushka, Sprintall, Gille and Wijffels, J. Climate, 2012 Variations in Q (Q LH ) more closely follow wind stress magnitude

12 MLD Response to MJO Forcing MLD MLD-T Argo data MLD amplitudes ~ 10 m Largest signal in IAB ~0.6C No profiles available in Indonesian seas Spatial patterns resemble t and Q Active MJO: diabatic cooling & wind stirring cause MLD deepening and cooling Suppressed MJO: Surface warming and light winds lead to MLD shoaling and warming MLD-T lags Q by one phase, i.e., consistent with model that Q drives MLD-T Drushka, Sprintall, Gille and Wijffels, J. Climate, 2012

13 Barrier Layer (BL) Influence on MJO Heat Flux Precip Wind Stress Zonal V Temperature THIN BL THICK Thin BL:- 15% stronger heat flux and wind stress; 10% stronger P; higher MLD-T; deeper isothermal layer Thick BL:- entrainment cooling during MJO reduced so MJO drives weaker SST anomaly Modulation of SST by BL thickness can have significant consequences for response of ocean to MJO and in turn, the feedbacks of ocean to atmosphere on MJO time scales Drushka, Sprintall and Gille, JGR, 2014

14 Summary SST and upper ocean characteristics in the Indonesian Seas are the product of several competing processes (both atmospheric and oceanic driven) over many time scales Response may be to local (winds, tides, inertial waves, air-sea interaction) and remote (Kelvin waves) forcing Mixing is important! The efficiency of mixing processes depends on stratification, bathymetry and the background oceanic and atmospheric large-scale conditions that vary across the Indonesian seas.

15 Few Direct Observations of Mixing in Indonesian Seas shear strain dissipation rate Near-inertial phase lines evident in shear and strain, leading to pulses of mixing every 4.4 days, the wave period Turbulence mixing occurs when low Ri=N 2 /S 2 (base of MLD) At time of observations, local winds were light -> down-ward propagating internal wave likely generated 3 weeks prior and ~200 km away. diapycnal diffusivity Alford and Gregg, JGR, 2001

16 Direct measure of vertical mixing : INDOMIX VMP direct Dissipation Measurement CTD Indirect Overturn Thorpe method Banda Sea Far field Near field Above Straits Koch-Larrouy, Atmadipoera, van Beek, Madec, Aucan, Lyard, Grelet, and Souhart, submitted DSR, 2015

17 Mixing Processes: Discussion Issues Where does mixing occur? At what depths? Seasonal preference? Impact of the ITF advection stream on SST and air-sea interaction if ITF absent or significantly reduced, may enhance the zonal SST gradient Shallow versus Deep Basins? Precipitation and presence of barrier layer? Large P -> high R/O warms SST by limiting latent heat release & mixing. Also freshwater caps can trap heat in near surface Presence of a barrier layer may significantly affect SST anomalies. SST gradients in response to MJO forcing zonal SST gradient in response to seasonal migration across Indonesia and its atmospheric convective activity (convective limit at SST ~27.5C) Diurnal variability in SST and surface layer? Strong vs. Weak Wind scenarios? Regional variability within Indonesian Seas? EEZ Issues? Use of ROV? What is the impact on biology? Ecosystem/fisheries interest and also might feedback via solar absorption to SST Need in situ SST and MLD observations within Indonesian Seas to validate remotely sensed data and model output

18

19 MJO Driven Deep Vertical Motions Nov-Apr Mean T along equator MJO composite dt/dz at thermocline depth shows eastward propagation of 0.5C anomaly along the equator at speed of ~2 m/s consistent with Kelvin wave speed (and slower than 5 m/s propagation speed of MJO) T Anomaly at Thermocline Depth Drushka, Sprintall and Gille, JGR, 2014

20 ITF: Influence on Circulation Temperature on sigma-0 = 25.0, upper thermocline cool ITF ITF cools and freshens through mixing in the Indonesian Seas and also the Indian Ocean thermocline [on isopycnals] relative to the ambient thermocline. Salinity on sigma-0 = 25.0, upper thermocline Low salinity freshwater input Indonesian Mix-Master Song, Gordon and Visbeck, JGR, 2004

21 Seasonal SST: The role of Mixing SST(CTRL-NO_TIDE) Along shelf-break - enhanced bottom mixing lifts the thermocline and enables intrusion of thermocline water onshore. Shallow seas - water column well mixed so vertical mixing does not result in enhanced cooling. Annual February SST(NOWind-NOWindTide) February North Nusa Tenggara: well mixed thermocline water from Makassar intrudes below and upwells to cool SST in NW monsoon. NE Ekman flow advects this to Banda Sea. Kida and Wijffels, JGR, 2012

22 MJO Driven MLD Heat Budgets dt/dt Q Hor adv Vert ent Mixed layer temperature variations driven by anomalous Q Central I.O. rapid deepening of MLD during onset of active MJO induces entrainment cooling No entrainment cooling in WPAC despite MLD deepening is related to barrier layer (water below MLD is same T and so no net cooling). Using a fixed MLD overestimates heat budget by 40% during active MJO (i.e. using a slab MLD will overestimate the T changes forced by MJO -> important consideration for models simulating MJO temperature and air-sea interaction associated with MJO Large error bars! Event to event variation! Drushka, Sprintall, Gille and Wijffels, J. Climate, 2012

23 Barrier Layers Influence SST THIN BL THICK BL MJO composite surface temperature anomalies are stronger when BL is thin Modulation of SST by BLT can have significant consequences for response of ocean to MJO and feedbacks of ocean to atmosphere on MJO time scales Difference Drushka, Sprintall and Gille, JGR, 2014