1. Updates on 2D model results McGillicuddy, D.J., Lavelle, J.W., Thurnherr, A.M., Kosnyrev, V.K. and L.S. Mullineaux. Larval dispersion along an axially symmetric mid-ocean ridge. Submitted, Deep-Sea Research I. 2. Toward a conceptual model of the secondary circulations on an axisymmetric ridge? 3. Initial 3-D particle tracking simulations

2-D model currents: May-Nov 1999 Zonal Meridional Vertical Mean Standard Deviation

Flank currents: comparison with observations LADDER Bottom-mounted ADCP (Andreas Thurnherr) Time-mean: Nov 2006 – Nov 2007 Simulated currents: May-Nov 1999

Time-series of simulated currents z=2500m, ±20km of crest Zonal Meridional Vertical

Time-dependent particle release Sept-Oct animations 10 mab: plan view 10 mab: x-z 225 mab: plan view 225 mab: x-z

Particle Positions After 30 Days: Probability Density Functions

Trends in larval retention (1 km bin on ridge crest)

Time-mean PDFs as a function of depth and precompetency interval

Comparison with Marsh et al. (2001) Dec 1999 – April 2000 Current 175mab Example trajectory Dispersal distance Retention, dispersal vs. time “In the eight-month current profile from April 1999 to December 1999, the longest dispersal distance SSE along-axis was only 54 km.” Assumptions Spatially uniform currents 100% mortality of larvae transported beyond 25 km of ridge crest Considered all larvae within 25 km of ridge crest at the end of their larval period to be survivors

Comparison with Marsh et al Comparison with Marsh et al. (2001) Particles within ±25km after 26 days Marsh et al.: Dec 99- Apr 00 This study: May-Nov 99 It is important to note that the Marsh et al. study used a current meter record from a different time period (December 1999 to April 2000) than that used herein (May to November 1999). In comparing the two records, Marsh et al. found that the later time series yielded longer dispersal distances than the earlier record used in the present study. Therefore, were the comparison between these two approaches to be carried out using the same current meter record, the differences in dispersal would likely be even larger.

Conclusions Retention is sensitive to vertical position 3-fold variation in time-mean retention range: 1% (10mab) to 3% (225mab) Retention is time-dependent up to 10-fold variation in monthly mean retention Dispersal distance is sensitive to vertical position 10 mab: ±200km along-ridge 225 mab: ±100km along-ridge Flank currents play a major role in larval dispersal

Hypotheses All else being equal… -- Species with passive larvae would be the first to re-colonize a new area (longer dispersal distances) -- Species with a balloonist larval stage can achieve a self-sustaining population with lower fecundity than those with a passive larval stage (higher retention)

1. Updates on 2D model results 2. Toward a conceptual model of the secondary circulations on an axisymmetric ridge? 3. Initial 3-D particle tracking simulations

2-D model currents Zonal Meridional Vertical Mean Standard Deviation

Currents in the EPR system Oscillatory flow enhances vertical mixing. Vertical mixing reduces stratification: isopycnals dome updward above the ridge and plunge downward below the ridge. Displacement of the density surfaces leads to strong flank currents in approximate geostrophic balance. There is a secondary circulation: mean downwelling is fed by horizontal convergence at the top of the cold dome, and divergence at the crest. What drives this secondary circulation? What balances the cross-isopycnal buoyancy fluxes? Oscillatory flow

Animations of hourly variations in the temperature field http://science.whoi.edu/users/valery/outgoing/2dmodel/ 2 May - 20 June 20 June - 9 August 9 August - 28 September 28 September - 21 November C.f. Andreas’ density cross-sections from LADDER 1-2-3

1. Updates on 2D model results 2. Toward a conceptual model of the secondary circulations on an axisymmetric ridge? 3. Initial 3-D particle tracking simulations

Particle Tracking with the 3-D Model 10mab 75mab 125mab 175mab 225mab http://science.whoi.edu/users/valery/ladder/epr_3d.html

Particle depth distributions 10 mab 75 mab 125 mab 175 mab 225 mab

Active Vent Sites 1 Biovent 2 Bio1 Tica EastWall M MusselBed Perseverance Q Riftia RiftiaField RobinsRoost RustyRiftia 3 AlvinellidBlanket AlvinellidPillar Bio9 Bio9+ Bio9++ BM119 BM141 BM82 BM89 Io P Ty 4 Choo-Choo TubeWarmPillar Y 5 TevniaHole 6 V 7 A J T 8 Hot8 L 9 B 10 D 11 K DyeRelease 12 PBR600 Depth (m)

Population connectivity: 10 mab, Tp=5 d Source “effectiveness” Group 3 Group 2 Group 1 Destination “attractiveness” Relatively higher connectivity within the northernmost 2 groups of vents, which is expected given their proximity. Structure within/between groups: 1 reseeds itself but little impact on others 2-5 interconnected within group 1, but do not seed group 2 or beyond group 2 interconnected within itself and group 1; asymmetry: group 2 feeds group 1 but not vice-versa. More isolated vents to the south: can feed northern vents, but self-sustenance is the most likely. Source effectiveness: sum the rows to compute the fraction of particles release from each site that settle in a suitable environment; in this scenario, the two vents at the midpoint between the northern 2 groups are the most effective in feeding settlement. Destination attractiveness: sum the columns to compute the fraction of the total number of particles (from all sites) that settle on each site. The N-S gradient in attractiveness is presumably related to mean flow being mostly northward during this particular time period. Inspired by Mitarai et al. (2009)

Depth of particle release Precompetency period Tp

Future Directions Generate more robust dispersal statistics for the LADDER era [requires yearlong 3-D simulation] Implement behavior more explicitly Make more direct comparison with larval observations Dynamics of secondary circulation in 2-D model?