Vertical Eddy Iron Transport in the Southern Ocean

Presentation given at OSM20.20

Transcript

1. Vertical Eddy Iron Transport in the Southern Ocean Ryan Abernathey (LDEO) Takaya Uchida (LDEO ➜ now Grenoble) Dhruv Balwada (NYU ➜ now UW) Shafer Smith (NYU) Galen McKinley (LDEO) Marina Levy (LOCEAN-IPSL)
2. • Uchida, T., D. Balwada, R. Abernathey, G. McKinley, S. Smith & M. Lévy. Vertical Eddy Iron Fluxes Support Primary Production in the Open Southern Ocean. Nature Communications (accepted), 2020. • Uchida, T., D. Balwada, R. Abernathey, G. McKinley, S. Smith & M. Lévy. The contribution of submesoscale over mesoscale eddy iron transport in the open Southern Ocean. JAMES, 2019. • Uchida, T., D. Balwada, R. Abernathey, P. Channing, E. Boss & S. Gille. Southern Ocean Phytoplankton Blooms Observed by Biogeochemical Floats. JGR: Oceans, 2019. Takaya Uchida https://roxyboy.github.io/
3. Vertical Fluxes Matter! (especially in SO) • Global impact of transient eddies is most apparent in the vertical.
 Horizontal fluxes often cancel when averaging over large scales, standing meanders, etc. (e.g. Griffies et al., 2015) • Especially true in Southern Ocean
 (e.g. Zika et al., 2013; Abernathey & Cessi (2013)) ted by a dagger, i.e. a0 = a a , a† = hai a . e terms are associated with transient eddy fluctuations (turbule dagger terms are associated with stationary eddies, a.k.a. steady C. Both types of motion can transport buoyancy meridionally and v al average is reinterpreted as a “streamwise” average around a cir hich follows the ACC path, the dagger terms will vanish. In pra budget is generally dominated by the vertical flux terms, i.e. hwihbi ' hw0b0i . hwihbi ' v @ @z hbi Munk’s classic “abyssal recipe” Southern Ocean Deep ocean buoyancy budget
4. • Macronutrient consumption by SO phytoplankton regulates global productivity • Phytoplankton growth is limited mainly by light and iron availability • Bloom has strong seasonal cycle due to… • Light • Mixed layer depth • Mesoscale / submesoscale turbulence? Journal of Geophysical Research: Oceans 10.1029/2019JC0153 Figure 2. Time series of despiked and interpolated Cp masked out based on the Chl cut off (a–c) plotted against pressure for the floats shown in Figure 1 full time series of each float is given in Supporting Information S1. The black solid (dashed) lines show hML (hPAR ). (d–f) Time series of the vertically int carbon ⟨Cp⟩ in green solid, surface carbon concentration multiplied by hML in green dashed, and accumulation rates (rp ) in blue solid lines after a 30-day running mean is applied. The thin green lines show ⟨Cp⟩ before the running mean. Southern Ocean Phytoplankton Blooms Observed by Biogeochemical Floats Takaya Uchida1 , Dhruv Balwada2 , Ryan Abernathey1,3 , Channing J. Prend4 , Emmanuel Boss5 , and Sarah T. Gille4 1Department of Earth and Environmental Sciences, Columbia University, New York, NY, USA, 2Center for Atmosphere H ARTICLE C015355 of phytoplankton asonal cycle are optical backscatter ical Argo floats Southern Ocean Phytoplankton
5. Southern Ocean Phytoplankton Tagliabue et al. (2014) Entrainment vs. diapycnal diffusion ARTICLES NATURE GEOSCIENCE DOI: 10.1038/NGEO2101 Small phytoplankton Large phytoplankton Zooplankton Bacteria Fe supply Low High Low High dFe stock Mixed-layer depth Ferricline depth Winter Spring Summer Autumn High Low Mixed-layer dFe stock Recycling Winter entrainment Detrainment losses Biological uptake and abiotic losses Weak diapycnal inputs Dominance of recycling and shift to smaller cells Dominant physical processes High f e-ratio Moderate f e-ratio Low f e-ratio Reduction in dFe Stock dFe profile Entrainment Detrainment Diapycnal di￿usion High Low
6. Southern Ocean Phytoplankton Tagliabue et al. (2014) Entrainment vs. diapycnal diffusion ARTICLES NATURE GEOSCIENCE DOI: 10.1038/NGEO2101 Small phytoplankton Large phytoplankton Zooplankton Bacteria Fe supply Low High Low High dFe stock Mixed-layer depth Ferricline depth Winter Spring Summer Autumn High Low Mixed-layer dFe stock Recycling Winter entrainment Detrainment losses Biological uptake and abiotic losses Weak diapycnal inputs Dominance of recycling and shift to smaller cells Dominant physical processes High f e-ratio Moderate f e-ratio Low f e-ratio Reduction in dFe Stock dFe profile Entrainment Detrainment Diapycnal di￿usion High Low Our Hypothesis: Eddies transport iron vertically into the Euphotic zone
7. MITgcm Simulations Physics MITgcm @ 20km, 5km, 2km resolultion Seasonal cycle in heat flux, wind stress KPP Mixed Layer 2000 km Biology Based on Darwin (Dutkiewicz et al., 2009) 2 phytoplankton sp. + 1 zooplankton Phosphate + nitrate Dissolved + ligand-bound iron 5 km run
8. MITgcm Simulations Physics MITgcm @ 20km, 5km, 2km resolultion Seasonal cycle in heat flux, wind stress KPP Mixed Layer 2000 km Biology Based on Darwin (Dutkiewicz et al., 2009) 2 phytoplankton sp. + 1 zooplankton Phosphate + nitrate Dissolved + ligand-bound iron 5 km run
9. MITgcm Simulations Physics MITgcm @ 20km, 5km, 2km resolultion Seasonal cycle in heat flux, wind stress KPP Mixed Layer 2000 km Biology Based on Darwin (Dutkiewicz et al., 2009) 2 phytoplankton sp. + 1 zooplankton Phosphate + nitrate Dissolved + ligand-bound iron 5 km run
10. Seasonal Cycle in Biology a f e d c b 2km Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical diffusive fl Vertical eddy flux Iron conc. Vertical diffusive+GM+Redi flux a f e d c b Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical diffusive flux Vertical eddy flux Iron conc. Vertical diffusive+GM+Redi flux Flux Figure 3: Time-depth plots of the daily and spatial median of phytoplankton biomass; a. b,c The spatial mean of vertical eddy transport using 15-daily snapshot outputs and diffusive iron transport using daily-averaged outputs. Panels a-c, and f (daily-averaged iron concentration) are from the 2 km run. d,e Daily-averaged phytoplankton biomass and the sum of vertical diffusive, GM and Redi iron flux from the 100 km GM+R run. The dotted lines in all panels show the mixing (mixed) layer depth for the 2 km (100 km GM+R) run. The mixing-layer depth (MLD) was too sensitive to the winds in the 100 km GM+R run, likely due to the GM tapering interacting with KPP (42). In all of our other runs, the mixed-layer depth defined as the depth at which the potential temperature decreased by 0.2 C from the surface (43) (not shown), proved to be very similar to the MLD so we used the mixed-layer depth for the 100 km GM+R run. f e 2km hytoplankton conc. Iron conc. Vertical diffusive+GM+Redi flux Flux -depth plots of the daily and spatial median of phytoplankton biomass; a. b,c The vertical eddy transport using 15-daily snapshot outputs and diffusive iron transport using utputs. Panels a-c, and f (daily-averaged iron concentration) are from the 2 km run. d,e phytoplankton biomass and the sum of vertical diffusive, GM and Redi iron flux from the a f e d c b Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical diffusive flux Vertical eddy flux Iron conc. Vertical diffusive+GM+Redi flux Flux Figure 3: Time-depth plots of the daily and spatial median of phytoplankton biomass; a. b,c The spatial mean of vertical eddy transport using 15-daily snapshot outputs and diffusive iron transport using daily-averaged outputs. Panels a-c, and f (daily-averaged iron concentration) are from the 2 km run. d,e Daily-averaged phytoplankton biomass and the sum of vertical diffusive, GM and Redi iron flux from the 100 km GM+R run. The dotted lines in all panels show the mixing (mixed) layer depth for the 2 km (100 km GM+R) run. The mixing-layer depth (MLD) was too sensitive to the winds in the 100 km GM+R run, likely due to the GM tapering interacting with KPP (42). In all of our other runs, the mixed-layer depth defined as the depth at which the potential temperature decreased by 0.2 C from the surface (43) (not shown), proved to be very similar to the MLD so we used the mixed-layer depth for the 100 km GM+R run. Climatology over central part of simulation
11. Seasonal Cycle in Biology a f e d c b 2km Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical diffusive fl Vertical eddy flux Iron conc. Vertical diffusive+GM+Redi flux a f e d c b Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical diffusive flux Vertical eddy flux Iron conc. Vertical diffusive+GM+Redi flux Flux Figure 3: Time-depth plots of the daily and spatial median of phytoplankton biomass; a. b,c The spatial mean of vertical eddy transport using 15-daily snapshot outputs and diffusive iron transport using daily-averaged outputs. Panels a-c, and f (daily-averaged iron concentration) are from the 2 km run. d,e Daily-averaged phytoplankton biomass and the sum of vertical diffusive, GM and Redi iron flux from the 100 km GM+R run. The dotted lines in all panels show the mixing (mixed) layer depth for the 2 km (100 km GM+R) run. The mixing-layer depth (MLD) was too sensitive to the winds in the 100 km GM+R run, likely due to the GM tapering interacting with KPP (42). In all of our other runs, the mixed-layer depth defined as the depth at which the potential temperature decreased by 0.2 C from the surface (43) (not shown), proved to be very similar to the MLD so we used the mixed-layer depth for the 100 km GM+R run. f e 2km hytoplankton conc. Iron conc. Vertical diffusive+GM+Redi flux Flux -depth plots of the daily and spatial median of phytoplankton biomass; a. b,c The vertical eddy transport using 15-daily snapshot outputs and diffusive iron transport using utputs. Panels a-c, and f (daily-averaged iron concentration) are from the 2 km run. d,e phytoplankton biomass and the sum of vertical diffusive, GM and Redi iron flux from the a f e d c b Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical diffusive flux Vertical eddy flux Iron conc. Vertical diffusive+GM+Redi flux Flux Figure 3: Time-depth plots of the daily and spatial median of phytoplankton biomass; a. b,c The spatial mean of vertical eddy transport using 15-daily snapshot outputs and diffusive iron transport using daily-averaged outputs. Panels a-c, and f (daily-averaged iron concentration) are from the 2 km run. d,e Daily-averaged phytoplankton biomass and the sum of vertical diffusive, GM and Redi iron flux from the 100 km GM+R run. The dotted lines in all panels show the mixing (mixed) layer depth for the 2 km (100 km GM+R) run. The mixing-layer depth (MLD) was too sensitive to the winds in the 100 km GM+R run, likely due to the GM tapering interacting with KPP (42). In all of our other runs, the mixed-layer depth defined as the depth at which the potential temperature decreased by 0.2 C from the surface (43) (not shown), proved to be very similar to the MLD so we used the mixed-layer depth for the 100 km GM+R run. Climatology over central part of simulation c b 2km ytoplankton conc. Vertical diffusive flux Vertical eddy flux c b 2km Phytoplankton conc. Vertical diffusive flux Vertical eddy flux
12. Seasonal Cycle in Biology a f e d c b 2km Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical diffusive fl Vertical eddy flux Iron conc. Vertical diffusive+GM+Redi flux a f e d c b Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical diffusive flux Vertical eddy flux Iron conc. Vertical diffusive+GM+Redi flux Flux Figure 3: Time-depth plots of the daily and spatial median of phytoplankton biomass; a. b,c The spatial mean of vertical eddy transport using 15-daily snapshot outputs and diffusive iron transport using daily-averaged outputs. Panels a-c, and f (daily-averaged iron concentration) are from the 2 km run. d,e Daily-averaged phytoplankton biomass and the sum of vertical diffusive, GM and Redi iron flux from the 100 km GM+R run. The dotted lines in all panels show the mixing (mixed) layer depth for the 2 km (100 km GM+R) run. The mixing-layer depth (MLD) was too sensitive to the winds in the 100 km GM+R run, likely due to the GM tapering interacting with KPP (42). In all of our other runs, the mixed-layer depth defined as the depth at which the potential temperature decreased by 0.2 C from the surface (43) (not shown), proved to be very similar to the MLD so we used the mixed-layer depth for the 100 km GM+R run. f e 2km hytoplankton conc. Iron conc. Vertical diffusive+GM+Redi flux Flux -depth plots of the daily and spatial median of phytoplankton biomass; a. b,c The vertical eddy transport using 15-daily snapshot outputs and diffusive iron transport using utputs. Panels a-c, and f (daily-averaged iron concentration) are from the 2 km run. d,e phytoplankton biomass and the sum of vertical diffusive, GM and Redi iron flux from the a f e d c b Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical diffusive flux Vertical eddy flux Iron conc. Vertical diffusive+GM+Redi flux Flux Figure 3: Time-depth plots of the daily and spatial median of phytoplankton biomass; a. b,c The spatial mean of vertical eddy transport using 15-daily snapshot outputs and diffusive iron transport using daily-averaged outputs. Panels a-c, and f (daily-averaged iron concentration) are from the 2 km run. d,e Daily-averaged phytoplankton biomass and the sum of vertical diffusive, GM and Redi iron flux from the 100 km GM+R run. The dotted lines in all panels show the mixing (mixed) layer depth for the 2 km (100 km GM+R) run. The mixing-layer depth (MLD) was too sensitive to the winds in the 100 km GM+R run, likely due to the GM tapering interacting with KPP (42). In all of our other runs, the mixed-layer depth defined as the depth at which the potential temperature decreased by 0.2 C from the surface (43) (not shown), proved to be very similar to the MLD so we used the mixed-layer depth for the 100 km GM+R run. Climatology over central part of simulation c b 2km ytoplankton conc. Vertical diffusive flux Vertical eddy flux c b 2km Phytoplankton conc. Vertical diffusive flux Vertical eddy flux Continuous supply of iron during summer. Less need to invoke iron recycling
13. Iron Transport by Eddies Figure S5. Time-depth plots of the vertical eddy iron flux in [ ] for runs without any eddy parametrization are shown for each resolution a-c. Figure S6. Time-depth plots of the vertical total (KPP+Redi+MLI+”resolved” eddy) flux a is shown along with the parametrized MLI contribution b and resolved eddy μmol Fe m−2 yr−1 2 km 5 km 20 km a c b Month
14. Bulk Primary Production Response • Primary production and total biomass go up with resolution • …Even though ML shoals • This is due to strong dependence in eddy vertical iron flux on resolution 100km GM+R 2km 5km 20km 20km MLI+R 20km MLI A scatter plot showing the resolution and parametrization dependence on annual median kton biomass (hCpi) plotted against the annual mean of total vertical iron transport at the r 100 m whichever is deeper (Fz Fe ). The runs without any eddy parametrizations are shown in Avg eddy iron transport below ML base Total phytoplankton biomass
15. Seasonal Cycle in Turbulence • 2 km resolution: prominent asymmetry in vorticity and strong seasonality
 (submesoscale permitting) Sep. 15 Feb. 15 b -100 -300 b a manuscript submitted to Journal of Advances in Modeling Earth Systems (JAMES) Sep. 15 Feb. 15 b -100 -300 -300 -100 -100 -300 c b d a -300 -100 -100 -300 c d Figure 1. Relative vorticity normalized by the local Coriolis parameter (Ro = ⇣ f ) for the 2 km run on September 15 a and February 15 b in the top 300 m. c The zonal-annual mean stratification from the 2 km run plotted against depth and meridional distance. d Seasonal prob- Relative Vorticity Relative Vorticity Histogram Winter Summer
16. Mesoscale or Submesoscale? <{ ˆ w⇤ˆ c}(k, z) e f d Vertical Buoyancy Flux inv. wavenumber [cpkm] a b k [cp Vertical Iron Flux inv. wavenumber [cpkm] Vertical Flux Cross Spectrum: Variance Preserving Plots
17. Omega Equation Giordani & Planton (2000)
18. Omega Equation d Giordani & Planton (2000)
19. Can the flux be parameterized? 100km GM+R 2km 5km 20km 20km MLI+R 20km MLI scatter plot showing the resolution and parametrization dependence on annual median n biomass (hCpi) plotted against the annual mean of total vertical iron transport at the 00 m whichever is deeper (Fz Fe ). The runs without any eddy parametrizations are shown in a f e d c b 2km Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical eddy flux Vertical diffusive+GM+Redi flux a f e d c b 100km GM+R Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical eddy flux Vertical diffusive+GM+Redi flux Flux
20. Can the flux be parameterized? 100km GM+R 2km 5km 20km 20km MLI+R 20km MLI scatter plot showing the resolution and parametrization dependence on annual median n biomass (hCpi) plotted against the annual mean of total vertical iron transport at the 00 m whichever is deeper (Fz Fe ). The runs without any eddy parametrizations are shown in a f e d c b 2km Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical eddy flux Vertical diffusive+GM+Redi flux a f e d c b 100km GM+R Depth [m] Depth [m] Phytoplankton conc. Phytoplankton conc. Vertical eddy flux Vertical diffusive+GM+Redi flux Flux • YES! But it takes lots of tuning • MLI parameterization does NOT help
 (only active within ML, no deeper flux) • Gent-McWilliams + Redi can do it • Need spatial variability (Visbeck scheme) • Need vertical structure: Parameterization of eddy fluxes near oceanic boundaries (Ferrari et al. 2008) • Redi (isopycnal mixing is 50% of the transport)
21. Conclusions 2000 km 5 km run Parameterization of eddy fluxes near oceanic boundaries • Eddies can make a first-order contribution to the open-ocean ML iron budget by moving iron across the ML base • Dominant contribution to flux is from balanced mesoscale isopycnal stirring
22. Conclusions 2000 km 5 km run Parameterization of eddy fluxes near oceanic boundaries • Eddies can make a first-order contribution to the open-ocean ML iron budget by moving iron across the ML base • Dominant contribution to flux is from balanced mesoscale isopycnal stirring