In “Factors controlling the competition between Phaeocystis and diatoms in the Southern Ocean and implications for carbon export fluxes,” authors Nissen and Vogt use a coupled physical-biogeochemical model to assess the relative ecological importance of two major Southern Ocean phytoplankton functional groups: diatoms and Phaeocystis. The authors use an extensive CHEMTAX and HPLC ecological dataset to validate their model, and conduct several sensitivity experiments for model parameterization. Despite the uncertainties associated with many of the model parameters, the authors conclude that Phaeocystis makes a larger-than-expected contribution to NPP and POC export in the Southern Ocean.
My main concern with the paper is that given that the outcome is highly sensitive to the model parameterization, the authors should place a very high priority on reporting uncertainties throughout. In its current form uncertainties are seldom reported. I also recommend that the authors conduct more sensitivity tests that examine how the model output changes when more than one parameter is perturbed. The model framework constructed by the authors would be ideal for a thorough exploration of how NPP and export vary with changing parameters, and presents an excellent opportunity to isolate the environmental conditions where model uncertainty is greatest. Additionally, cellular Fe:C ratios should be included in these sensitivity experiments. As Fe:C will make a very large difference to the POC calculated, and is central to the paper title, it is necessary that the authors investigate how Fe:C, which is sensitive to both light and iron, can change the model outcome.
I believe that this paper will be much more useful to biogeochemists if error bars can be placed on the model estimates during major revisions, with particular consideration of parameter sensitivity to light and iron limitation.
Detailed comments:
Line 5: “improved” instead of “extended”
Line 7: This implies solitary Phaeocystis are not biogeochemically relevant, when commonly used to understand succession in the Ross sea
Line 7: Please report uncertainties
Line 9: Saying “temporal variability” here implies spatial variability is not a considerable factor, when in the following line you say that there is a difference at the coast
Line 11: Remove “Still,” from this sentence, it’s unclear how this sentence follows from the previous
Line 24-25: There is no previous statement that there is observed succession, just that the other groups contribute to biomass. Please include such a statement and provide references.
Line 32: Since there are few and not zero estimates, please briefly reference them here.
Line 48: This is unclear. Do you mean that grazing would not be as significant of a loss as aggregation?
Line 52: “expensive” instead of “difficult”
Line 66: Reference Yager et al., 2016
Line 67: Zooplankton grazing rates on Phaeocystis are low (Yang et al., 2016)
Line 68: There are also arguments that the hydrography results in the resurfacing of any sunken Phaeocystis-associated POC (Lee et al., 2017)
Line 79: State how the referenced models parameterize Phaeocystis differently, and the possible consequences on the model outcome
Line 88: The introduction is missing a description of observed succession in SO sectors outside of the Ross Sea
Line 93: Why did you not include solitary Phaeocystis when it’s been used in other successful models (such as those published by Kaufman)?
Line 95: Please give a brief statement here on how the model was validated.
Line 102: does the addition of the new PFG affect the validation metrics done in Nissen et al. 2018?
Line 107: isn’t solitary Phaeocystis also important in the SO?
Line 110: Wouldn’t you need to assume a minimum cell concentration for this to be valid?
Line 114: Please give your rationale for using this function instead of Eppley
Table 1: Why did you choose a slightly different grazing rate for diatoms and Phaeocystis?
Line 149: shouldn’t alpha be sensitive to the iron concentration? (Strzepek et al., 2019)
Line 162: this is a very small difference, what is the net sensitivity of the output to this parameter?
Line 190: what day of year are you initializing with?
Line 194: what is the sensitivity of the model outcome to your initial community composition?
Line 212: diatom and phaeo cellular Fe:C ratios should also be informed by light and iron limitation (Strzepek et al., 2011)
Line 240: Top 50 m is not deep enough for analyzing export.
Line 246: References to supplemental figures are not in order
Line 250: What does the outcome look like with an attempt at a quantitative comparison? This would at least be useful to see in the supplemental material.
Line 263: This doesn’t seem like an insurmountable problem; satellite chlorophyll data should be employed for model validation. Binning and temporal averaging are potential workarounds for the issues presented.
Line 263: if you are using timing from the literature, why not use bloom initiation from the literature as well?
Line 265: I recommend you also validate using bloom initiation as well.
Line 314: Could the overestimate have to do with modeled Fe:Chl ratios?
Line 315: Please state the reason why Chl was overestimated in Nissen et al., 2018
Figure 1: The model is not capturing spatial variability in chlorophyll concentration- if it’s truly due to a latitudinal bias in the ocean color product, please validate against data from shipboard CTDS
Line 320: How does modeled Chl compare to measurements from CTDs, gliders, BGC Argo floats, etc. from the region? Then you can determine if it’s a satellite underestimate issue.
Line 324: “distinct” instead of “distinctly different”
Table 3: Please include confidence intervals on ROMS-BEC values
Table 3: Why do you use a 100 m depth horizon for export? The 0.1% light depth horizon is more biogeochemically relevant (Buesseler et al., 2020)
Line 338: It would increase model confidence to include whether coccolithophore biomass corresponds to positioning of the great calcite belt.
Line 342: Please include standard deviations on these percentages.
Figure 2: Please use box-and-whisker plots instead of pie charts to display categorical data and the associated error.
Line 347: Again, please include error on your reported percentages.
Line 361: “agrees” instead of “suggests”
Line 370: Please also include a comparison to bloom initiation.
Line 374: This doesn’t seem right. Phaeocystis are often associated with bloom peaks in coastal polynyas.
Figure 3: Why don’t Phaeocystis concentrations reach values much higher than 3 ug/L? Much higher concentrations have been observed.
Line 382: How specifically does the diatom parameterization drive this bias?
Line 405: Wouldn’t the lower cell density associated with deeper MLD bias this assessment? It is known that Phaeocystis thrive under lower light conditions than diatoms, so this doesn’t seem right.
Line 408: It is important to note here that much of the SO is light limited, in particular the canyons near the WAP (Carvalho et al., 2016).
Line 414: If this analysis is not useful for the scientific questions proposed, it should be either removed or moved to the supplemental information.
Line 422: If this is the case, why do Phaeocystis never dominate at the bloom peak in the model?
Figure 4: The temperature range covered here is very large- where are water column temperatures getting so hot in the model? This may be killing off the Phaeocystis.
Line 430: It would be useful to see how photoacclimation effects on the Fe:C ratio would affect the outcome.
Figure 5: Iron is green here, not blue.
Line 442: It would also be good to see Pine Island Polynya and the Amundsen Sea Polynya to compare with the Ross Sea.
Line 445: In the model they should be considered, but what is your confidence in the model when it cannot reproduce a Chl max in Phaeocystis?
Line 457: It is necessary to know how sensitive the model is to the range of these parameters to determine how likely this result is to mirror reality.
Line 467: It would be great to include some necessary information in this manuscript about the way iron is cycled in ROMS-BEC. How are organic ligands parameterized? How about scavenging processes? Is the relief from iron limitation in the Ross Sea driven by wind-driven sediment resuspension (McGillicuddy et al., 2015)?
Line 475: Please include error bars on these estimates. Propagate the error using the sensitivity analyses, and the standard deviation across the domain.
Figure 6: Please report uncertainty on the numbers in this figure.
Line 492: It would be helpful to see a breakdown of these fluxes in comparison to other modeled and observed export in the SO.
Figure 7: Since there is a wide potential range of aggregation rates, how would this figure change when testing using that range?
Line 497: Why is the peak in Phaeocystis so late? This is much later than observations.
Line 525: I find this result hard to believe when the model is not taking into account changing cellular iron quotas as iron limitation shifts.
Line 539: More than Kfe needs to be constrained with environmental data. Alpha, Fe:C, Pmax are all also sensitive to light/iron limitation conditions.
Line 546: There is agreement in some conditions, but elsewhere there seems to be large discrepancies. Additionally, surface chlorophyll variability is not represented well by the model.
Line 555: It would be useful to conduct more sensitivity experiments while varying the aggregation parameter alongside other model parameters, in light of this finding.
Line 565: This paragraph seems somewhat redundant with the above paragraph. Please restructure.
Line 579: Please provide uncertainty on this percentage.
Line 587: Please propagate this uncertainty into your NPP, POC, and export estimates.
Line 590: Considering the discrepancies between the observations and the model, please include confidence intervals on your estimates.
Line 608: Why can’t you assess these effects? Assessing horizontal fluxes of model tracers should be relatively straightforward.
Line 616: How are you accounting for variability in mixed layer depth in the DMSP calculation?
Line 640: These topics should be discussed in your methods. The rationale for excluding solitary Phaeocystis needs to be justified.
Line 669: These non-Redfieldian dyanamics should be straightforward to implement in the model, and would be an interesting sensitivity study.
Line 689: It may be worth testing a range of parameters to get a general sense of the sensitivity of export. I’d imagine this would be useful for many people to see.
Buesseler, K. O., Boyd, P. W., Black, E. E., & Siegel, D. A. (2020). Metrics that matter for assessing the ocean biological carbon pump. Proceedings of the National Academy of Sciences of the United States of America, 117(18), 9679–9687. https://doi.org/10.1073/pnas.1918114117
Carvalho, F., Kohut, J., Oliver, M. J., Sherrell, R. M., & Schofield, O. (2016). Mixing and phytoplankton dynamics in a submarine canyon in the West Antarctic Peninsula. Journal of Geophysical Research: Oceans, 121(7), 5069–5083. https://doi.org/10.1002/2016JC011650
Lee, S. H., Hwang, J., Ducklow, H. W., Hahm, D., Lee, S. H., Kim, D., et al. (2017). Evidence of minimal carbon sequestration in the productive Amundsen Sea polynya. Geophysical Research Letters, 44(15), 7892–7899. https://doi.org/10.1002/2017GL074646
McGillicuddy, D. J., Sedwick, P. N., Dinniman, M. S., Arrigo, K. R., Bibby, T. S., Greenan, B. J. W., et al. (2015). Iron supply and demand in an Antarctic shelf ecosystem. Geophysical Research Letters, 42(19), 8088–8097. https://doi.org/10.1002/2015GL065727
Strzepek, R. F., Maldonado, M. T., Hunter, K. A., Frew, R. D., & Boyd, P. W. (2011). Adaptive strategies by Southern Ocean phytoplankton to lessen iron limitation: Uptake of organically complexed iron and reduced cellular iron requirements. Limnology and Oceanography, 56(6), 1983–2002. https://doi.org/10.4319/lo.2011.56.6.1983
Strzepek, R. F., Boyd, P. W., & Sunda, W. G. (2019). Photosynthetic adaptation to low iron, light, and temperature in Southern Ocean phytoplankton. Proceedings of the National Academy of Sciences, 116(10), 4388–4393. https://doi.org/10.1073/pnas.1810886116
Yager, P. L., Sherrell, R. M., Stammerjohn, S. E., Ducklow, H. W., Schofield, O. M. E., Ingall, E. D., et al. (2016). A carbon budget for the Amundsen Sea Polynya, Antarctica: Estimating net community production and export in a highly productive polar ecosystem. Elementa: Science of the Anthropocene, 4, 000140. https://doi.org/10.12952/journal.elementa.000140
Yang, E. J., Jiang, Y., & Lee, S. H. (2016). Microzooplankton herbivory and community structure in the Amundsen Sea, Antarctica. Deep Sea Research Part II: Topical Studies in Oceanography, 123, 58–68. https://doi.org/https://doi.org/10.1016/j.dsr2.2015.06.001 |