Progress

WP1 Intermediate Progress Report

2018-09-03T05:37:27+00:00

WP1 Biogeochemical model calibration and evaluation

Task 1.2 Revisit of Iron sources in PISCES model (Icebergs and sea-ice)

Iron is a critical limiting micronutrient in the Southern Ocean, but its chemistry is complex; moreover, it has multiples sources, not all of which have been accounted for in ocean biogeochemical models. To better evaluate how “new” sources of iron to the Southern Ocean, previously unaccounted for in ocean models, in terms of how they affect ocean biogeochemistry, two model configurations were set up, one to assess effects of iron delivery from icebergs and the other to assess that from sea ice. Both relied on the eORCA1 (nominal 1° grid) configuration of the NEMO-LIM-PISCES ocean model. To assess effects of iron delivery from icebergs, we prescribed that the water flux from Southern Ocean icebergs (Merino et al., 2016, a monthly climatology) had a dissolved iron concentration of 4.5 nmol L^-1, as estimated from Raiswell et al. (2008). This iceberg iron source modified the simulated surface distribution of dissolved iron (Fig. 1) resulting in substantial changes in the spatiotemporal evolution of surface chlorophyll concentration (Fig. 2).

Figure 1. Difference in spatial distribution of the surface dissolved iron concentration in summer between the “iceberg” experiment (ICB, with iceberg iron source) and the control experiment (CTL, , no iceberg iron source).

Figure 2. Surface chlorophyll concentration in (left) the CTL experiment (no iceberg iron delivery) and in (right) the ICB experiment (with iceberg iron delivery) in November, December, and January in the Southern Ocean, south of 55°S.

A second configuration was set up to better represent the sea-ice iron source to the Southern Ocean via the sea-ice model (LIM) to the PISCES model. That is, in LIM we added a new variable, the iron content in sea-ice for each ice category. This variable is computed as a function of the iron availability in surface seawater in PISCES based on 2 different seawater-to-sea-ice flux formulations: (1) maintaining a constant maximal iron concentration of 10 nmol L^-1 in sea-ice, and (2) a constant iron flux from surface seawater to sea ice of 1x10^-6 nmol s^-1. The latter is expected to better represent of what it is observed in the ocean (Lannuzel et al., 2016). Figure 3 shows a comparison in winter and summer of the spatial distribution of surface iron concentration in the PISCES model in the experiment with no iron in sea-ice (Fe CTL) and the experiment with a maximal iron concentration in sea-ice (Fe CST) and their difference.

Figure 3. The iron concentration in surface seawater in the Southern Ocean, south of 55°S for (a, b) the control experiment (Fe CTL, no iron in sea-ice), (c,d) the Fe CST experiment, and (e,f) difference (Fe CST minus FeCTL) in austral winter (June, July, August) and summer (December, January, February).

Concurrently during SOBUMS, related work on improving simulated ocean biogeochemistry in the context of the iron-limited Southern Ocean has led to an article that was published in J. Geophys. Res. Oceans in 2018. This study focused on the role of iron in diatoms and how that affects the biological pump and the seasonal cycle of pCO₂ (Person, Aumont, & Lévy, 2018)). To cope with iron limitation in the Southern Ocean diatoms, a dominant phytoplankton group in that region, have been shown in cultures to have developed an original adaptation strategy to maintain efficient growth rates despite very low cellular iron quotas, even in low light conditions. With the NEMO-PISCES model, we explored the consequences of this physiological adaptation for the biological pump and the seasonal variability of both surface chlorophyll and surface pCO₂. In the PISCES model, we implemented a low intracellular Fe:C requirement in Southern Ocean diatoms. Relative to the control experiment (CTL), that new formulation (PHYSIO) exhibits a 10% increase in the relative contribution of diatoms to total Southern Ocean primary production, which strengthens the biological pump and enhances the biological contribution to the seasonal evolution of pCO₂ relative to the thermodynamic component. When compared to the observations, the simulated seasonal cycle of both surface chlorophyll and surface pCO₂ in the Polar Zone (PZ) are markedly improved in PHYSIO vs. CTL . This improvement is noteworthy given that virtually all ocean models, whether forced or coupled, have had major difficulties to simulate the seasonal cycle of pCO₂ in the Southern Ocean, the region that takes up most of the ocean's anthropogenic carbon and heat.

Figure 4. Seasonal cycles of the anomalies of ∆pCO2 (pCO₂^oc – pCO₂^atm) from observations, (Landschützer [dashed line, dark grey] and Takahashi data set [dashed line, light grey]), the CTL experiment (light blue), and the PHYSIO experiment (dark blue) averaged over (a) the Sub-Antarctic Zone (SAZ), (b) the Polar Zone (PZ), (c) the Antarctic Zone (AZ), and (d) the entire Southern Ocean (SO) south of the subtropical front.

Reference:

Lannuzel, D., Vancoppenolle, M., van der Merwe, P., de Jong, J., Meiners, K. M., Grotti, M., … Schoemann, V. (2016). Iron in sea ice: Review and new insights. Elementa: Science of the Anthropocene, 4, 000130. https://doi.org/10.12952/journal.elementa.000130

Merino, N., Le Sommer, J., Durand, G., Jourdain, N. C., Madec, G., Mathiot, P., & Tournadre, J. (2016). Antarctic icebergs melt over the Southern Ocean: Climatology and impact on sea ice. Ocean Modelling, 104, 99–110. https://doi.org/10.1016/j.ocemod.2016.05.001

Person, R., Aumont, O. & Lévy, M. (2018). The Biological Pump and Seasonal Variability of pCO2 in the Southern Ocean: Exploring the Role of Diatom Adaptation to Low Iron. J. Geophys. Res. Oceans. https://doi.org/10.1029/2018JC013775

Raiswell, R., Benning, L. G., Tranter, M., & Tulaczyk, S. (2008). Bioavailable iron in the Southern Ocean: the significance of the iceberg conveyor belt. Geochemical Transactions, 9(1), 7. https://doi.org/10.1186/1467-4866-9-7

WP2 Intermediate Progress Report

2018-09-03T08:51:34+00:00

WP2 : Global model integration and run production

Task 2.1 Preparation of the physical configuration

During SOBUMS, development of the circulation component of NEMO (OPA), i.e., without biogeochemistry (PISCES), was initiated in 2016, a collaboration between IPSL in Paris and the LGGE MEOM in Grenoble. The initial strategy was to start from an ORCA025 configuration developed within the DRAKKAR consortium, to increase the vertical resolution near the surface, to extend the grid to include under-ice regions in certain shelf seas, and to test new model parameterizations.

Thus during this upgrade to the eORCA025 configuration, the NEMO ocean model has been updated to its most recent version, which includes the newly tested tide-induced mixing parameterization and freshwater distribution within the water column. In addition, much effort has been devoted to adjust model parameters in order to reduce the small-scale effects and improve ventilation in the deep ocean (Fig. 1).

Fig. 1: North Atlantic meridional stream function. Sensitivity to tidal-induced mixing : (left) standard parameterization and (right) new parameterization (from De Lavergne).

Task 2.2 Integration of the biogeochemical component eORCA025-NEMO-PISCES

After developing and validating the physical component of NEMO, we adjusted the marine biogeochemistry component PISCES to optimize physical-biogeochemical coupling on the eORCA025 grid. Adjustments included optimizing the biogeochemical models parameters and boundary conditions and making technical modifications to optimize computational resources.

With that new NEMO-PISCES configuration on the eORCA025 grid, we then made our first reference simulations, forcing them with interannually varying fluxes from an atmospheric reanalysis (CORE II) over 1948-2009. Preliminary results suggest that generally the model satisfactorily simulates large-scale patterns of key biogeochemical tracers and fluxes, e.g. surface chlorophyll and sea-to-air CO₂ fluxes (Figure 2).

Fig. 2 : Annual mean of observed (left) and simulated (right) fields for surface chlorophyll (top) and the sea-to-air CO₂ flux (bottom).

Task 2.3 Production of reference and sensitivity experiments to changing stressor

For our first productions simulations, before engaging the immense computational resources needed to make simulations with the high-resolution eORCA025 configuration, with have made production simulations with the lower resolution configuration ORCA2 of the NEMO-PISCES model.

To test the effect of the atmospheric forcing, we made identical NEMOPISCES simulations over the same 20-year period (1990-2009), which differed only in terms of the atmospheric reanalysis used for the forcing. One simulation was made with the CORE2 forcing and another was made with the DFS5.2 forcing. Figure 3 reveals that relative to CORE2, the DFS5.2 forcing generally leads to smaller residuals between simulated and observed climatological annual-mean fields for dissolved oxygen and nutrients (WOA09).

In addition, one simulation was made to begin to assess potential future changes. In that 'future' simulations, the ORCA2 configuration of the NEMO-PISCES model was forced with the DFS5.2 atmospheric reanalysis, repeating its 1958-2012 loop until year 2100, while adding in atmospheric forcing anomalies from a former CMIP5 simulation of the IPSL-CM5 earth system model under the RCP8.5 scenario (DeltaClim, need reference here). A preliminary analysis indicates that during the next 100 years under the RCP8.5 forcing, there are large simulated changes in biogeochemical variables in the Southern Ocean and the North Pacific.

Fig. 3 : Annual means of surface nitrate, total dissolved inorganic phosphorus, total dissolved silicon, and dissolved oxygen: observations (WOA09) (column 1), CORE2-WOA09 and DFS5.2-WOA09 as averages over 1990-2009 (columns 2-3), and RCP8.5-DFS5.2 over 2012-2100.

WP4 Intermediate Progress Report

2018-09-03T05:29:28+00:00

WP4 : Quantification of the sources of uncertainty due to scale interactions.

Task 4.1 Impact of model resolution on tracer transport at basin scale

PIs: M. Lévy, X. Capet

The purpose of this task is to quantify how the scales that are not resolved in eORCA025 may affect tracer transport in the Southern Ocean. This is done through: 1) review of existing literature; 2) an idealized model configuration at high-resolution; 3) LES modeling

Task 4.1.1

Two review papers have been submitted, one on the role of small-scales on marine productivity (Lévy et al., 2018), the other on export production (Boyd et al., 2018).

Task 4.1.2

Set-up of an idealized model configuration to study to role of processes in the scale range R 1/24°<R<1/4°.

The open ocean conditions of the Southern Ocean are represented with a periodic reentry channel (Figure 1). In this configuration, the flow of the ACC is unobstructed by landmass allowing it to circumnavigate and reconnect back on itself. The configuration was initialized using a hierarchical approach of increasing horizontal resolution (dx), starting from dx = 1°, which was initialized from rest and adjusted for 200 years. This was used to initialize a run at dx = 1/6° which, initialized the dx = 1/24°. Our analysis will be based on the dx = 1/24° run, referred here on as B24S, which is mesoscale resolving and sub-mesoscale permitting run with seasonal and intraseasonal (storms) forcing.

Figure 1: Three-dimensional representation of the model setup. Top panel:Sea-surface Temperature (SST) overlaid with a snapshot of wind stress vectors in gray. Below is a model snapshot of surface chlorophyll. On the sides are sections of upper ocean dissolved Iron (DFe), with the purple line denoting the mixed-layer depth.

This model configuration is being used to evaluate the role of small-scale processes, combined with storms: A) on primary production, B) export and C) anthropogenic carbon penetration.

A) Primary production (S. Nicholson)

The experiments have been performed and analyzed. One publication is in preparation (Nicholson et al. in preparation)

B) Export (L. Resplandy + S. Nicholson + post-doc)

One paper has been submitted on the method (Resplandy et al., 2018)

The application of this method to Southern Ocean conditions will start in 2019 (post-doc)

C) Anthropogenic carbon (Thesis N. Aldea-Ramos)

The study is based on synthetic tracer experiments intended to mimic anthropogenic carbon penetration, which are included in the above model configuration. The high-simulation experiment has been performed, and the subduction rate of the tracer evaluated online. Subsequently, we will focus on how the subduction rate is affected by data resolution, and we will perform offline diagnostics based on the same fields projected onto coarser resolution model grids.

Task 4.1.3

The impact of resolution on small-scale vertical fluxes of nutrients was assessed with a model of even higher resolution (resolving submesoscales and some turbulent scales) using an LES model in a frontal zone configuration of 2 km x 2km horizontal size, i.e. nearly the size of one pixel in the previous configuration. One paper has been submitted (Whitt et al., 2018)

Task 4.2 Study of chaotic eddy-driven variability of Southern Ocean biogeochemistry!

PIs : M. Gehlen, T. Penduff,

The purpose of this task is to quantify how chaotic eddy-driven, low frequency variability of physical variables affect Southern Ocean biogeochemistry. This study has been carried out on the basis of an ensemble of three ¼° global simulations . Using the metrics developed by OCCIPUT project, the focus here is on primary production and air sea CO2 fluxes The outcome is the first quantification of the magnitude of chaotic eddy-driven low frequency variability of Southern Ocean biogeochemistry. A manuscript describing this study is in preparation (Gehlen et al.)