Introducing particle associated copepods in MEDUSA

The ocean’s biological carbon pump (BCP) is the ensemble of all the marine biological processes involved in removing inorganic carbon from the ocean surface and exporting it to the deep ocean and sediments. It is driven by living organisms from the photic zone (~0-200 metres) producing organic matter that eventually sinks down the water column as detrital particles of different sizes. These organic particles are deteriorated and remineralized as they sink to depth. In general, the deeper the particles go, the longer the carbon they contain will be stored away from the atmosphere, in the ocean’s interior, but most is remineralized above 1000m (Martin et al., 1987; Giering et al., 2014; Buesseler et al., 2020).

The efficiency of the ocean’s BCP can be estimated through the particle organic carbon (POC – the organic carbon component of sinking particles) flux attenuation with depth. This is usually expressed as an empirical power law known as the Martin curve (Martin et al., 1987):

where F_z is the POC flux at depth z; F₁₀₀ is the POC flux at reference depth 100m; and b is the sinking flux attenuation rate. The Martin curve is a simple and effective way of capturing the general pattern of the POC attenuation. However, in the real ocean, lots of different processes affect this attenuation.

Reflecting this, marine biogeochemistry components of Earth system models (ESMs) include a range of different formulations that vary in their levels of process complexity. For instance, the biological processes and interactions represented in models can include particle turnover rates that are fixed or depend on temperature or oxygen, increasing sinking speed of detritus with depth, or mineral ballasting (Dunne et al., 2010; Yool et al., 2013; Aumont et al., 2015; Kriest et al., 2015;  Schwinger et al., 2016). The uncertainties associated to these formulations are large, especially when considering climate change scenarios.

Observations of sinking particles using ARGO floats with optical backscattering sensors have highlighted the role of POC fragmentation as the principal factor controlling the POC degradation process (Briggs et al., 2020). Mayor et al. (2020) proposed that ubiquitous, particle associated copepods (PAC) are responsible for much of this fragmentation. These are small crustaceans that break-up sinking organic particles as they feed on the microorganisms living on and within them. To examine this, Mayor et al. (2020) explicitly included PAC and their effect on the POC sinking flux in a 1D water column version of the biogeochemical model MEDUSA, the marine biogeochemistry component of UKESM1 (Yool et al., 2013; Sellar et al., 2019; Yool et al., 2021). Figure 1 shows a schematic of the ecosystem of the resulting model. The main changes in the model are the addition of PAC itself; partitioning of the fast POC pool into 2 pools, one fast (sinking at 30m/d), and one very fast (120m/d); the introduction of PAC within MEDUSA’s plankton food-web; and finally the interactions between PAC and particles. The PAC are preferentially attracted by the fastest sinking particles, targeting these and fragmenting them into smaller particles, and eating a small fraction in the process.

The results of this work provide the first quantitative demonstration that PAC may be largely responsible for attenuating particle flux in the upper part of the ocean’s mesopelagic zone (200-1000 metres).

Figure 1: Diagram of MEDUSA’s ecosystem components including the PAC model. New PAC components are shown in purple.

To understand whether PAC are important on a global scale, here we translate the 1D water column work of Mayor et al. (2020) to a global, 3D version of MEDUSA (Yool et al., 2013). For this, we use an ocean-only configuration, and run the resulting model across the historical period (1850-2015) under surface forcing from the full UKESM1 model.

Adding a new plankton group in an ecosystem model is an extremely sensitive thing to do. The new component can easily unbalance the model’s trophic interactions by annihilating and replacing other organisms, or being completely consumed and not changing anything in the model. To examine the impact of the change, we look at the model’s net primary production (NPP), the total organic matter produced by autotrophs, as this gives a good first order idea of the model’s behaviour. Figure 2 shows that MEDUSA-PAC’s NPP is realistic, and with the same patterns as in the original MEDUSA model. The main differences with MEDUSA are slightly higher surface nutrient concentrations due to PAC increasing the slow detritus concentration when fragmenting the fast detritus – which has the consequence of recycling nutrients more quickly. Interestingly, the pattern of Southern Ocean NPP is better with the PAC model, but the mechanisms behind this improvement are still under investigation. Overall, PAC has not distorted MEDUSA’s surface trophic interactions too much, and as the model is yet to be tuned, this is a very good and encouraging result.

The worldwide distribution of PAC (Figure 3) is consistent with observations: namely that they are ubiquitous in the ocean, and occur at relatively high concentrations down to 500 m unlike other zooplankton in MEDUSA that are confined to the upper 200m. Interestingly, the surface concentrations of PAC are higher in oligotrophic areas where their predation by mesozooplankton is lower. There, the PAC population stays within the upper 200m due to the relative lack of fast particles lower down.

Figure 2: Total primary production for satellite-based estimates (left), MEDUSA (middle), MEDUSA-PAC (right).

Figure 3: PAC surface (up-left), and vertically integrated concentration (up-right). Down in Antarctic to Arctic section of PAC through the middle of the Atlantic Ocean down to 500m depth.

Figure 4:

Comparison of MEDUSA (left) and MEDUSA-PAC (right) models’ export of POC at 1000m depth (up), and transfer efficiency (right). The transfer efficiency is the ratio between POC export 1000 and 100m depth, expressed as a percentage. It represents how much of the POC flux at 100m reaches down to 1000m.

The impacts of PAC on the overall POC flux in the ocean interior are quite large (Figure 4). In the original MEDUSA model, the geographical pattern of POC flux at 1000m is very similar to that of NPP, whereas with the MEDUSA-PAC model the pattern is much more homogeneous. As can be seen from the transfer efficiency (TE; the fraction of the 100m flux that reaches 1000m), this is explained by PAC being much more active in the productive particle-rich regions where they reduce TE to 5-10%. In oligotrophic areas, there is not enough food to sustain an active PAC community and consequently, a large part of the POC is exported beyond 1000m depth, giving TE >40%. This result shows the potential importance of including the mechanistic effects of PAC in MEDUSA.

Conclusion

The MEDUSA-PAC model is very promising. The underlying theory was able to reproduce observed POC flux data, and its inclusion in PAC-MEDUSA does not significantly compromise the rest of the model’s performance. The next steps are to tune the MEDUSA-PAC model to fix the problem of high surface nutrients, and also to fully evaluate the model. Including PAC in MEDUSA adds an extra calculation cost of few percent, and that should also be considered if we want to include PAC in UKESM2. And, intriguingly, with PAC’s POC export being so different to MEDUSA’s, it would be interesting to estimate the impact of climate change, to more fully appreciate the importance of these animals on the BCP.

References:

Aumont, O., Ethé, C., Tagliabue, A., Bopp, L., & Gehlen, M. (2015). PISCES-v2: An ocean biogeochemical model for carbon and ecosystem studies. Geosci Mod. Dev., 8, 2465–2513.

Briggs, N., Dall’Olmo, G., & Claustre, H. (2020). Major role of particle fragmentation in regulating biological sequestration of CO2 by the oceans. Science, 367, 791–793.

Buesseler, K. O., Boyd, P. W., Black, E. E., & Siegel, D. A. (2020). Metrics that matter for assessing the ocean biological carbon pump. Proc. Natl. Acad. Sci. USA, 117, 9679–9687.

Dunne, J. P., Gnanadesikan, A., Sarmiento, J. L., & Slater, R. D. (2010). Technical description of the prototype version (v0) of Tracers Of Phytoplankton with Allometric Zooplankton (TOPAZ) ocean biogeochemical model as used in the Princeton IFMIP model. Biogeosci, 7: 3593.

Giering, S. L. C., Sanders, R., Lampitt, R. S., Anderson, T. R., Tamburini, C., Boutrif, M., […] Mayor, D. J. (2014). Reconciliation of the carbon budget in the ocean’s twilight zone. Nature, 507, 480–483.

Kriest, I., & Oschlies, A. (2015). MOPS-1.0: Towards a model for the regulation of the global oceanic nitrogen budget by marine biogeochemical processes. Geosci Mod. Dev., 8, 2929–2957.

Martin, J., Knauer, G., Karl, D., & Broenkow, W. (1987). VERTEX: Carbon cycling in the Northeast Pacific. Deep-Sea Res., 34, 267–285.

Mayor, D. J., Gentleman, W. C., & Anderson, T. R. (2020). Ocean carbon sequestration: Particle fragmentation by copepods as a significant unrecognized factor? BioEssays, 42, e2000149.

Schwinger, J., Goris, N., Tjiputra, J. F., Kriest, I., Bentsen, M., Bethke, I., […] Heinze, C. (2016). Evaluation of NorESM-OC (versions 1 and 1.2), the ocean carbon-cycle stand-alone configuration of the Norwegian Earth System Model (NorESM1). Geosci Mod. Dev., 9, 2589–2622.

Sellar, A. A., Jones, C. G., Mulcahy, J. P., Tang, Y., Yool, A., Wiltshire, A., […] Palmiéri, J., et al. (2019). UKESM1: Description and evaluation of the U.K. Earth System Model. Journal  of  Advances  in  Modeling  Earth  Systems,  11,  4513– 4558.

Yool, A., Popova, E. E., & Anderson, T. R (2013). MEDUSA-2.0: An intermediate complexity biogeochemical model of the marine carbon cycle for climate change and ocean acidification studies. Geosci Mod. Dev., 6, 1767–1811.

Yool, A., Palmiéri, J., Jones, C. G., de Mora, L., Kuhlbrodt, T., Popova, E. E., Nurser, A. J. G., Hirschi, J., Blaker, A. T., Coward, A. C., Blockley, E. W., and Sellar, A. A. (2021). Evaluating the physical and biogeochemical state of the global ocean component of UKESM1 in CMIP6 historical simulations, Geosci. Model Dev., 14, 3437–3472