Modeling <i>p</i>CO<sub>2</sub> variability in the Gulf of Mexico

Xue, Zuo; He, Ruoying; Fennel, Katja; Cai, Wei-Jun; Lohrenz, Steven; Huang, Wei-Jen; Tian, Hanqin; Ren, Wei; Zang, Zhengchen

doi:https://doi.org/10.5194/bg-13-4359-2016

Articles | Volume 13, issue 15

https://doi.org/10.5194/bg-13-4359-2016

© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/bg-13-4359-2016

© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 13, issue 15

Research article

|

08 Aug 2016

Research article |

| 08 Aug 2016

Modeling pCO₂ variability in the Gulf of Mexico

Zuo Xue, Ruoying He, Katja Fennel, Wei-Jun Cai, Steven Lohrenz, Wei-Jen Huang, Hanqin Tian, Wei Ren, and Zhengchen Zang

Download

Final revised paper (published on 08 Aug 2016)
Preprint (discussion started on 27 Aug 2014)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC C5679: 'comments', Anonymous Referee #1, 04 Oct 2014
- AC C7596: 'Reply to comments of referee #1', Zuo Xue, 22 Dec 2014
RC C5831: 'Comments on “Modeling pCO2 variability in the Gulf of Mexico"', Anonymous Referee #2, 11 Oct 2014
- AC C7599: 'Reply to comments of referee #2', Zuo Xue, 22 Dec 2014

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Z. George Xue on behalf of the Authors (09 Dec 2015) Author's response Manuscript

ED: Referee Nomination & Report Request started (20 Dec 2015) by Minhan Dai

RR by Anonymous Referee #3 (02 Jan 2016)

Suggestions for revision or reasons for rejection

Modeling pco2 variability in the Gulf of Mexico

By Xue et al.

This manuscript presents the surface pco2 in the Gulf of Mexico from a coupled physical-biogeochemical model. The focus is on the annual mean and the seasonal cycle of surface pco2 and air-sea CO2 flux for the northern (Louisiana-Texas shelf) and the deep part (water depth >200m) of the gulf. Some discussion is also made for other areas such as the western Florida shelf and Mexican shelf.

The over-all results seem reasonable in that the modeled surface pco2 and air-sea CO2 fluxes agree reasonably well with available data. The open ocean flux, however, is about 2-3 times of the estimate by Robbins et al. (2014). In addition, it seems the model is unable to reproduce the large inter-annual variability of surface pCO2 (hence likely air-sea flux as well) as seen in the pco2 data, also reported by Huang et al. (2013, 2015a,b).

The biggest issues appear to be 1) a lack of key information, and 2) that some statements/conclusions need better justification. For example,

1) Page 5-8, description of the model setup is not detailed enough and at times, confusing. For example, there is no description of how carbon is being cycled through the food web and in the water column (uptake, sinking, remineralization etc.). Also is organic carbon input included in the river discharges or along the open boundaries? And if yes, what are the data sources and processing details?
2) Page 10-13, it will be helpful to include some information regarding the spatial distribution and seasonal cycle of new productivity (or net community productivity) since it is concluded that biological removal is the dominant driver for air-sea CO2 flux in the gulf (see, e.g. abstract). In order to evaluate the impacts of biological removal, a map showing the differences of pco2 and air-sea flux between exp1 and exp2 (similar to Figure 8) will be useful as well.
3) Page 10, section 4.1. In addition to temporal comparison, a spatial comparison is needed to verify the spatial pco2 pattern, e.g., by binning the data into 0.5x0.5 boxes and then making a 1 to 1 correlation analysis.
4) Page 18, for the carbon budget, some calculation of production, respiration (it was asserted respiration is the dominant process within the Mississippi river plume), and export (vertical and horizontal) for NGoM or entire gulf will be useful as well
5) Page 14-17, section 5.1. This is a rather long section but the discussion is rather meandering and sometime contradicting. For example (line 359-360), why the high pco2 in the plume and on the shelf is a result of respiration? Just two paragraphs above (lines 333-334), the authors also state that there is rich DIC input from the river, which would presumably bring high pco2 along with the plume. The bottom line is that there is a lack of quantitative information regarding the roles of biological production/removal and respiration. Also, why is the river plume “light-limited”?
6) Page 17, lines 368-370. This statement appears to conflict with Figure 6 (bottom panel), which indicates the biological removal strongly affects surface pco2 and air-sea flux throughout the year.
7) Page 18, last paragraph, what exactly are model uncertainties?

The writing seems confusing or prone to grammatical errors. Here are some examples,

1) page 5, line 95, pco2 -> CO2
2) page 6, line 123, we found -> we found that
3) page 6, line 124, were -> was
4) page 6, line 126, limited in -> limited to
5) page 8, line 155, Talk+50 is assigned as river DIC. This sounds like a strong assumption that needs to have a better justification.
6) Page 8, line 160-167. Grammar error.
7) Page 8, line 165-168. I assume the authors simply mapped the 1904-1910 conditions to 2004-2010 with a day to day match (e.g. Jan 20, 1904 will be corresponding to Jan 20, 2004). But, yes or no, this needs to be explicitly stated. Also, why 1904-1910? Any special reason to choose this period?
8) Page 9, line 178-182. How exactly is the data being processed? I assume the author use a 10-day binning for temporal processing. How about spatial means? Were the data within one domain (say NGoM) being summed up and divided by the total number of valid data points?
9) Page 9, line 188, fell in -> fell within
10) Page 10, line 217-218. I think this statement is not accurate. The pco2 maxima in the NGoM appear to be about 50 ppm higher. But this does not seem to be the case for low end pco2.
11) Page 13, line 271, wrong numbers.
12) Page 13, line 287, were simulated -> were
13) page 15, line 333-334, why this is true? What is the correlation relationship between Talk-sal (henceforth DIC-sal), Talk-DIN, and DIN-sal in the MS river?
14) Page 16, line 354-355. What does the bifurcation exactly imply?

Hide

RR by Chris Hunt (02 Feb 2016)

Suggestions for revision or reasons for rejection

Review of Xue et al, "Modeling pCO2 Variability in the Gulf of Mexico"

This revised manuscript addresses an important issue, namely, the status of the Gulf of Mexico as a regional source or sink of atmospheric carbon dioxide. The authors use a modeling approach to derive sea surface pCO2 from total alkalinity and dissolved inorganic carbon, which is a reasonable approach given the lack of direct pCO2 observations in many areas of the Gulf. The authors have done a good job to address the points raised by reviewers of the original manuscript, particularly regarding the atmospheric pCO2 data source. As discussed below, I think more discussion of the relationships used to derive total alkalinity and dissolved inorganic carbon, especially the strengths and weaknesses of these relationships in the Gulf of Mexico, is warranted.

Major Comments:
-I was surprised by the data presented in Figure 4, specifically the Mississippi/Atchafalaya River DIC input for the 1904-1910 and modern time periods. The river DIC flux appears largely similar, on average, between the two time periods. It was my understanding from the paper by Raymond et al. (Nature 2008) that DIC flux from the Mississippi increased substantially over the preceding century- an increase that the DLEM model used in this work does not seem to capture. This contrast is worth noting and discussing, although the discussion of results from the 1904-1910 simulation was very limited.
-The validation data for the WGoM and MX regions, and to a lesser extent the WF regions, are extremely sparse. Also, there is a very large region of the south/western Open Ocean region with little to no validation data. I acknowledge that the utility of a model lies in the opportunity to estimate conditions in unsampled times/regions, but all those empty spaces leave a lot of uncertainty. The WGoM, MX and WF regions are not specifically addressed at all in the Discussion section. I imagine that the overall Gulf of Mexico results/fluxes would not change substantially if these three regions were excluded from the totals, and just the Open Ocean and NGoM results were used. This comparison (Open Ocean+NGoM vs. the entire GoM) would be useful to present.
-The model presented in this manuscript relies on the empirical relationships of Lee et al. (2006) to determine total alkalinity and Lee et al. (2000) to determine total inorganic carbon boundary conditions. However, these relationships specifically exclude all of the Gulf of Mexico for total alkalinity, and about half of the Gulf of Mexico for inorganic carbon. The authors should explicitly describe which relationship/region from each Lee et al. paper was used in this work, and justify the choice to use these relationships. The authors should also explicitly describe that total alkalinity was parameterized from salinity and surface temperature (Lee et al. 2006), while inorganic carbon was presumably parameterized from just surface temperature (and not temperature and nitrate, Lee et al. 2000). This helps explain the larger apparent variability in DIC when compared to total alkalinity in Figure 9.
-Section 4.1 discusses Mississippi River discharge and pCO2 results, yet pCO2 is not shown in Fig. 4 where it is referenced. Instead NO3 is shown. This needs to be fixed.
Minor Comments
-The manuscript could use another proof-read. Some instances: Abstract L32 change to “Two model sensitivity”; P3L44 “In the face”; sP6L124 change were to was; P7L133 change to ‘regardless of which’; P7L138 change to calculations; P7L149 change to relatively; P8L173 change composed to compiled; P10L204 change to ‘model results in shallow’; P12L259 change to “uniform”; P12L268 change was to were; P13L271 change to ‘rest of the seasons’; P13L291 change to pCO2; P14L293 change to pCO2; P15L325 change to times; P18L390 change were to was; P19L425 change to relatively
-I echo the sentiment of the Anonymous Reviewer #2 that details of the biogeochemical component of the model are needed.
-P3L63: the authors may want to elaborate on how the loading of carbon causes eutrophication. I generally associate the term eutrophication with nutrient loading; the contributions of carbon (presumable organic) are more indirect that nutrient additions.
-P4L65-70 and throughout: I believe the authors are generally using the term ‘carbon’ in this paper to refer to inorganic carbon. If so, this should be made explicit, and care should be taken when describing total or organic carbon, as I presume this section does. Another example is P13L284: presumably the authors are talking about a CO2 sink, instead of an inorganic or total carbon sink.
-P4L71-76: Please show the Loop Current in Figure 1, for those like myself who are less familiar with Gulf of Mexico circulation.
-P5L95: This is a nitpick, but I believe one does not quantify pCO2 fluxes, as pCO2 refers to the partial pressure. Instead it is the CO2 flux which was quantified.
-P5L107-108: Please explain exactly what the terms “realistic atmospheric forcing” and “open boundary conditions” refer to.
-P7L154: Please list the major rivers whose observations were taken from USGS data.
-P8L167: Should be Exp3.
-P9L187: Specifically name the model simulation as the control (Exp1).
-P9L187-197: Why not show the model pCO2 results from all sub-regions in a table, as the fluxes are shown in Table 1?
-P10L222-223 and throughout: Please add uncertainties to the average fluxes listed here, as variability is quite important in understanding these results.
-P15L329-336: This is a very nice section of the discussion, and would benefit even more from additional description of the biogeochemical component of the model, as mentioned earlier in my review.
-Figure 2: Panel A is labeled TIC (total inorganic carbon), while the manuscript discussed DIC.
-Figure 3: Is there a way to show the 1904-1910 atmospheric pCO2 used in the model? Perhaps as an inset?
-Figure 6: The shading showing the model uncertainty is very light, and may not print well. Can it be darkened?
-Figure 10: I’m not quite sure what this figure is showing. What does each data point represent: one grid cell in each sub-region? The caption describes seasonal means from 2005-2010, but that would produce 24 data points for each sub-region, and there are many more data points shown here. Are these data points averaged over some sort of longitude range?
-P41L700: Isn’t salinity technically unit-less?
-P42L713: I believe Alk should be in units of micro-equivalents per liter, or milli-equivalents per cubic meter)

Referee Report: PDF

Hide

ED: Reconsider after major revisions (20 Feb 2016) by Minhan Dai

AR by Z. George Xue on behalf of the Authors (10 May 2016) Manuscript

ED: Publish subject to minor revisions (Editor review) (17 Jun 2016) by Minhan Dai

AR by Z. George Xue on behalf of the Authors (01 Jul 2016) Author's response Manuscript

ED: Publish as is (14 Jul 2016) by Minhan Dai

AR by Z. George Xue on behalf of the Authors (18 Jul 2016)

Short summary

In this study we used a state-of-the-science coupled physical–biogeochemical model to simulate and examine temporal and spatial variability of sea surface CO₂ concentration in the Gulf of Mexico. Our model revealed the Gulf was a net CO₂ sink with a flux of 1.11 ± 0.84 × 1012 mol C yr⁻¹. We also found that biological uptake was the primary driver making the Gulf an overall CO₂ sink and that the carbon flux in the northern Gulf was very susceptible to changes in river inputs.

Modeling pCO2 variability in the Gulf of Mexico

Download

Interactive discussion

Peer-review completion

Modeling pCO₂ variability in the Gulf of Mexico