Mixing of the Faroe Bank Channel Overflow Ilker

Mixing of the Faroe Bank Channel Overflow Ilker Fer Geophysical Inst. , University of Bergen, Norway Ilker. Fer@gfi. uib. no

Acknowledgements Thanks to n Bert Rudels, Knut S. Seim, Gunnar Voet, Katrin Latarius, Kjersti L. Daae for their help during the cruise n Knut S. Seim for providing movies from a regional simulation n Elin Darelius for preparation of the mooring data n Detlef Quadfasel, Bogi Hansen and Svein Østerhus for contributions to the moorings n Peter Rhines and Charlie Eriksen for their efforts to coordinate Seaglider deployments with my cruise Work is supported through the IPY project BIAC (Bipolar Atlantic Thermohaline Circulation) by the Research Council of Norway

FBC Overflow Olsen et al. (2008) • Overflows across the GSR (~ 6. 4 Sv) contribute significantly to AMOC. • FBC overflow is 1/3 of the total overflow • Overflows sink, spread and merge to form a deep boundary current (13 Sv near Cape Farvel) 70% of the total overturning! Hansen et al. (2001) See: 10 years of FBC overflow observations at the sill: Hansen and Østerhus, 2007, Pi. O. Downstream mixing of the FBC overflow: Mauritzen et al. 2005, DSR-I.

Experiment: 29 May – 8 June 2008 Squares : Microstructure / CTD/ LADCP stations Circles : Time series (Microstructure / CTD / LADCP) Triangles : Moorings (F line is U. Hamburg, rest U. Bergen) 2 months of mooring data (14. 05 – 17. 07. 2008) 63 CTD-LADCP Profiles / 92 deep Microstructure Profiles

VMP 2000 during deployment LADCP & CTD VMP

VMP-C sensor failed. Density is inferred from T within an rms error of 0. 01. Figure shows the 3 rd deg. poly. fit to all SBE-CTD data.

Plume Interface, Ambient and Water Masses Dr

LADCP detided using Egbert et al. (1994) for the European Shelf Model (res: 1/30°)

Survey Mean Profiles

Section-averaged Properties Following Baringer and Price (1997) and Girton and Sanford (2003), we use density-anomaly and transport-weighted averages of various properties. E. g. , anomaly weighted bottom depth:

Sill 60 80 100 120 Downstream Distance (km) 30

A 2 Ric

A 5 Ric

Mean Profiles at Time Series Stations

Volume Transport: T-classes

Volume Transport: Below the interface Mauritzen et. al (2005) : 2. 4 Sv Hansen, Østerhus (2007) : 2. 2 Sv Duncan et. al (2003) : 1. 9 Sv Mauritzen et. al (2005) Duncan et. al (2003) Sill – 0 km : 2. 0 Sv : 1. 9 Sv 30 km Mauritzen et. al (2005) Duncan et. al (2003) : 2. 0 Sv : 1. 9 Sv 60 km

Overall Entrainment where W 0. 5 is the plume half-width which comprises 50% of the plume mass anomaly (Girton and Sanford, 2003). Compare entrainment with Arneborg et al. 2007: Using Up = 0. 6 m/s; CD = 3 x 10 -3; Fr = 0. 7 and Ek = 0. 15 w. E = 2. 2 x 10 -4 m/s

Stress Entrainment from dissipation profiles (following Arneborg et al. 2007):

tb = 1. 9 Pa compares with: 2 Pa for a similar survey (Mauritzen et al, 2005) and 3. 5 ± 1. 3 Pa at the sill (Johnson and Sanford, 1992) Residence time of plume water in the channel system is about 45 days (Mauritzen et al. , 2005) Spindown time = rhp. Up /2 tb is about 10 hours using survey-averaged hp = 225 m and Up = 0. 6 m/s.

Around 90 km: - Mauritzen et al. (2005) infer the largest entrainment rates - Pratt et al. (2007) identify a transverse hydraulic jump Rate of descent is not entirely due to friction

Low Frequency Oscillations Current-meter time series show 3. 5 – 4 day period oscillations in the channel, more energetic than inertial and M 2 period. Regional simulations using the Bergen Ocean Model (BOM), sigma-coordinate, domain -bathymetry and forcing identical to Riemenschneider and Legg (2007). 32 vertical layers, 2 km horizontal resolution. Movie by Knut S. Seim

Low Frequency Oscillations Movie by Knut S. Seim

Secondary Circulation in a V-shaped canyon Courtesy of E. Darelius

Umlauf and Arneborg, 2009, JPO in press.

UX UD UG

Concluding remarks n First microstructure measurements of the FBC overflow were conducted n CTD&LADCP system resolves the plume, gradient Ri and the fine-structure n n Plume is characterized by a bottom-mixed layer (up to 100 m thick) with enhanced dissipation, a quiescent core near the velocity maximum and highly dissipative interface of 100 -200 m thickness. Dissipation values near the bottom and at the interface are very large (comparable to the most dissipative tidal fjords). Consistently, 4 -m gradient Ri is well below unity, frequently less the critical value of 0. 25. n There is change in the descent rate at about 90 km downstream from the sill n Descent rate cannot be explained by bottom stress alone n n Evidence for cross-circulation. It enhances mixing, but detailed analysis is needed to quantify. The nature of the 3. 5 -4 day period oscillations and their affect on the mixing need further work. Entrainment and interfacial stress may be important on the overflow dynamics, but the bottom stress dominate. Entrainment rate agree with the empirical formulation of Arneborg et al. (2007).