Interannual Variability of the Gulf of Mexico Loop Current

Transcription

1 Interannual Variability of the Gulf of Mexico Loop Current Dmitry Dukhovskoy (COAPS FSU) Eric Chassignet (COAPS FSU) Robert Leben (UC) Acknowledgements: O. M. Smedstad (Planning System Inc.) J. Metzger and A. Wallcraft (NRL SSC) C. Hall (UC) R. Nedbor-Gross (COAPS FSU) Funded by HYCOM Consortium

2 The Loop Current and Loop Current Eddies North Atlantic LATEX Shelf Ocean Gulf of Mexico Yucatan Peninsula Yucatan Channel Cuba Caribbean Sea from NOAA NGDC

3 Loop Current Cycle LC Northward Intrusion LCE Shedding and Rapid Retreat LCE Shedding Period (Lugo-Fernandez, 2007) LC in the Retracted Phase

4 What we still don t know about the LC? 1. What controls the eddy shedding? Flow variability in the Yucatan Channel (Ezaer et al., 2003) Potential vorticity flux to the GoM (Candela et al., 2002; Oey et al., 2003) Flow variability in the Florida Straits (Sturges et al., 2009) Frontal cyclones (Cochrane, 1972; Zavalo-Hidalgo et al., 2003; Shmitz, 2005) Caribbean eddies (Oey et al., 2003; Athie et al., 2012) Wind-induced transport through Great Antilles Passages (Oey et al., 2003) Effect of westward winds (Chang & Oey, 2010) Dynamical instability (Cherubin et al., 2006) Deep flows in the Yucatan Channel (Bunge et al., 2002) Deep ocean dynamics and deep eddies (Lee & Mellor, 2003; Oey, 2008) 2. Is LC eddy shedding stochastic? Can it be predicted? Eddy-shedding is dominated by natural variability O(12mo) (Hurlbert & Thompson, 1980; Pichevin & Nof, 1997) Eddy-shedding is stochastic (Nowlin et al., 2000; Zavala-Hidalgo et al., 2006) LC behavior is not chaotic and can be approximated as a nonlinear driven and dampen oscillator with a very short memory (Lugo-Fernandez, 2007)

5 LC Variability from Observations LCE Shedding Period from Altimetry Lugo-Fernandez (2007) Infrared images Altimetry Near-surface T in August, GOMl0.04 Concerns: 1. Only ~20 years of reliable satellite observations of the LC 2. Spatial and temporal coverage of satellite observations may not be sufficient for accurate tracking of the LC and LCEs 3. Errors in assimilation and interpolation techniques employed for production of the SSH gridded products 4. Robustness of methods for LC / LCE identification and tracking T/P and Jason 1 ground tracks in the Gulf of Mexico Yu et al., 2009

6 Multi-Decadal Simulation of the Gulf of Mexico Model Initialization: Spin-up: 54-year run: HYCOM Gulf of Mexico 1/25 (GOMl0.04), 20 vertical layers Atmospheric forcing: CFSR 1hr fields, m air temperature specific humidity radiative fluxes (net sh/wave net longwave) penetrating shortwave wind stress wind speed Lateral OB forcing: monthly mean fields from the near-real time nowcast/forecast 1/12 North Atlantic HYCOM, from Levitus climatology 2 cycles of with CFSR forcing 3 cycles with CFSR (Fields in December 2009 are ramped to January 1, 1992) Cycle 1 Cycle 2 Cycle 3 54 years

7 Percentage of time the LC front is present in 0.1 x 0.1 bins from the 54-year HYCOM GOMl0.04 >25 NOAA/AVHRR SST composite image, 01/31/2002 Vukovich, 2007

8 Comparison of LC eddy shedding in the 54-year HYCOM simulation and altimetry Altimetry separations (Leben) 18 years 27 total mean separation period: 7.9 months HYCOM separations 54 years 69 total mean separation period: 9.4 months 9.4 mo if infrared observations are included (Leben, 2005) Satellite (includes infrared obs.) separations (Vukovich, 2007) 32 years 35 total mean separation period: 11 months

9 Separation Period Binning Altimetry (18 yrs) HYCOM (54 yrs) % of separations Separation period (months) Separation period (months) Non-shedding state

10 18-year Block Mean Separation Period Years of integration Block 2: 3: mean T T Block 1: mean T Satellite Satellites observations 32 yrs (Vukovich, (Vukovich, 2007) 2007) HYCOM HYCOM GOMl0.04 Altimetry Altimetry 18 yrs (Leben, 2012) Courtesy of R. Nedbor-Gross

11 Period-Retreat Regression Altimetry HYCOM Retreat Latitude ( ) Separation Period (days) Separation Period (days)

12 Identification and Tracking of LCEs in SSH Fields

13 Trajectories of LCEs from the GOMl0.04 Southern Path (14%) Vukovich, 2007

14 Decay of the normalized diameter of LCEs at the surface over time Statistics of LCE from the GOMl0.04 Decay of the normalized area of LCEs over time

15 Eddy Shedding and PV Flux through the Yucatan Channel LC length (km)

16 48 months of Non-Shedding State Mass leakage through WFS jet Model Recovered from the non-shedding state (Hetland et al., 1999) Blocking cyclone Zavala-Hidalgo et al., 2006

17 Closing Remark HYCOM GOMl0.04 multi-decadal simulation of the Gulf of Mexico and satellite altimetry observations can address several problems related to the variability of the Loop Current: Accuracy of SSH gridded products derived from the satellite observations Robustness of methods of LC and LCE identification and tracking Reliability of existing estimates of the LC/LCE statistics Stochastic and deterministic components in the LC behavior Testing hypothesis of the LC variability Non-shedding behavior in the models

18

19 Volume Flux through the Yucatan Channel Seasonal LATEX Transports through the Yucatan Channel (black) and the Florida Straits Yucatan Transport from GOMl0.04 North Atlantic Ocean Gulf of Mexico Sv Yucatan Channel Cuba Yucatan Peninsula /- 5 Sv Rousset and Beal, JPO, 2010 Caribbean Sea 23 ± 3 Sv Candela et al., Sv Ochoa et al., Sv Johns et al., 2002 from NOAA NGDC

20 Cycle 1 SSH Fields from GOMl0.04 on the Same Forcing Date (06/14/1995) but Different 18-year Cycles Cycle 2 Cycle 3

21 What else is known? Planetary β-effect is essential in the LC variability such as northward/westward penetration and LCE shedding. The f-plane solution is a steady source-sink flow (Reid, 1972; Hurlbert & Thompson, 1980; Pichevin & Nof, 1997) Nonlinearity is necessary for shedding, the linear solution is a steady source-sink flow; Shedding can occur at constant inflow at the Yucatan Channel; Barotropic inflow stops eddy shedding (Hurlbert & Thompson, 1980) The LC tends to grow (penetrate into the Gulf) and shed eddies in response to the momentum imbalance of a zonal boundary current (Pichevin & Nof, 1997; Nof & Pichevin, 2001) The LCEs have horizontal scales O( km), propagate westward driven by β-effect (Schmitz, 2005; Cherubin et al., 2006; Sutyrin, 2001); have lifetimes of several months to a year (Leben, 2005)

22 Histograms of Separation Periods (months) Altimetry HYCOM GOMl0.04, 18-yr Blocks 10-90% Zavala-Hidalgo et al., Courtesy of R. Nedbor-Gross

23 Maximum LC North Latitude Altimeter Data (CCAR UC) Maximum Latitude HYCOM GOMl