the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Spatial reconstruction of long-term (2003–2020) sea surface pCO2 in the South China Sea using a machine learning based regression method aided by empirical orthogonal function analysis
Zhixuan Wang
Guizhi Wang
Xianghui Guo
Yan Bai
Yi Xu
Abstract. The South China Sea (SCS), the largest marginal sea of the North Pacific Ocean, is one of the world’s most studied model ocean margins in terms of its carbon cycle, where intensive field observations including sea-surface carbon dioxide partial pressure (pCO2) have been conducted over the last two decades. However, the datasets of cruise-based sea surface pCO2 are still temporally and spatially incomplete. Using a machine learning-based method facilitated by empirical orthogonal function (EOF) analysis capable of constraining the spatiality, this study provides a reconstructed dataset of the monthly sea surface pCO2 in the SCS with a reasonably high spatial resolution (0.05º×0.05º) and temporal coverage between 2003 and 2020. We validate our reconstruction with three independent testing datasets where, TEST.1 includes 10 % of our observed data, TEST.2 includes four independent underway datasets corresponding to four seasons, and TEST.3 includes a continuous observed dataset from 2003–2019 at the South East Asia Time-Series (SEATs) station located in the northern basin of the SCS. Our TEST.1validation demonstrated that the reconstructed pCO2 field successfully simulated the spatial and temporal patterns of sea surface pCO2. The root-mean-square error (RMSE) between our reconstructions and observed data in TEST.1 averaged to ~10 μatm, which is much smaller (by ~50 %) than that between the remote sensing (RS) and observed data. TEST.2 verified the accuracy of our reconstruction model in data months lacking observations, showing a near-zero bias (RMSE: ~8 μatm). TEST.3 tested the accuracy of the reconstructed long-term trend, showing that at the SEATs Station, the difference between the reconstructed pCO2 and observations ranged from -10 to 4 μatm (-2.5 to 1 %). In addition to the typical machine learning performance metrics, we present a new method to assess the uncertainty that includes the bias from the reconstruction and its sensitivity to the features, and successfully quantifies the spatial distribution patterns of uncertainty. These validations and uncertainty analysis strongly suggest that our reconstruction is effectively captures the main features of both the spatial and temporal patterns of sea surface pCO2 in the SCS. Using the reconstructed dataset, we show the long-term trends of sea surface pCO2 in 5 sub-regions of the SCS with differing physico-biogeochemical characteristics. We show that mesoscale processes such as the Pearl River plume and China Coastal Currents significantly impact sea surface pCO2 in the SCS during different seasons. While the SCS is overall a weak source of atmospheric CO2, the northern SCS acts as a sink, showing a trend of increasing strength over the past two decades.
Zhixuan Wang et al.
Status: closed
-
RC1: 'Comment on essd-2022-322', Anonymous Referee #1, 17 Nov 2022
The comment was uploaded in the form of a supplement: https://essd.copernicus.org/preprints/essd-2022-322/essd-2022-322-RC1-supplement.pdf
- AC1: 'Reply on RC1', Minhan Dai, 23 Dec 2022
-
RC2: 'Comment on essd-2022-322', Anonymous Referee #2, 24 Nov 2022
the presented study aimed to produce monthly sea surface pCO2 maps for the South China Sea (SCS). Given SCS is a typical temperate/subtropical maringal sea, the pCO2 sea surface maps for this waters is necessary for understanding the CO2 flux in temperate marginal sea and even global CO2 flux. From this perspective, the study and the data it present is very meaningful. Howeve, the mansucript still have some majors flaws which do not advise me to give a yes to publishing it in its current status.
Major comments
1. The manuscript was about a dataset generation, but from the abstract and the last section, what kind of data was used as input for the method
was missing.2. As I understand EOF was an important part of the method used for pCO2 maps generation, but in the entire section of methods, no paragraph or sentence was about EOF
3. The language of the manuscript still need some efforts. The current version contains too many redundent phrases and sentences without clear meaning and very difficult to read through and get the logical flow. Readers expect concise and precise expresssion in an acedemica paper.
and there are some gammar mistake and fuzzy expression.4. the range of legend in nearly all the map figures were to large and cannot show the spatial gradient of pCO2 distribution, e.g, figure 6, 8, 11, 12,13.
5. what is the intention of including figure 4, if it is the quality of the remote sensing based pCO2 maps included for further pCO2 maps derivation, should the authors just need to include the information from the data distributor?
6. the study site section(2.1) should just serve the question "why mapping pCO2 in SCS is important?", no other information is needed here.
7. be consitent with the teminology, sometimes it is "in-situ", but "observational data" and "observed data" were present many times.
8. in the abstract (line 12-14,), the importance of mapping pCO2 in SCS should be addressed before presenting the method, generated data and its quality.
9 part of the input pCO2 data of the presented study is from unpublished study (line 158), meaning not peer-reviewed.
10, line 308: Figure 7, validating the model output with the mdoel training data gives no useful information, suggest removing this part
Minor comments
line 15-17""Using a machine learning-based method facilitated by empirical orthogonal function (EOF).... between 2003 and 2020" should specifically mention what kind of data was used for the methods input.
linse 17- 20 "We validate our reconstruction with three independent testing datasets where,.... northern basin of the SCS." how independent are the three data set?
line 22 "our reconstructions and observed data" grammar mistake.
Line 27-28 "we present a new method to assess the uncertainty that includes the bias from the reconstruction and its sensitivity to the features,... quantifies the spatial distribution patterns of uncertainty." then the assessment method should be concisely introduced here. in addition, given this is a data presentation paper, the newly developed method should not in the highlight, unless it is a method presentation paper.
line 19 "that our reconstruction is effectively captures the main features of both the" ,check the grammar.
line 38,, "22–26%", I assume it should be 22%–26%.
line 54-55: ":The former typically use statistical interpolations and regression methods" does not fit with the neighouring sentence, rewrite it or delete it.
line 61- 63 ,"However, because of the complex and dynamic nature of biogeochemical and physical processes in coastal areas, characterization of sea surface pCO2 and subsequently the
air-sea CO2 fluxes both in time and space in marginal seas remains challenging", this sentence is too strong and undermines the motivation of presented study, rewrite it,line 67: "clear need", what kind of need is clear need? a need can be strong, urgent, but not clear, need itsself is a clear expression,
line 73: "(sea surface temperature, SST; chlorophyll a, Chl a),", pay attention to journal requirements on abbrevation
line 74: "underway "pay attention to the usage of underway, it is ambiguous in the manuscript.
line 82, "the whole China Sea", where is the China Sea? do you mean all the seas in China's territory?
line 84: "(reported in Wang et al., 2021).", pay attention to the format of the reference citation
line 84: "Bai et al. (unpublished) subsequently", if the work is not publised, then it should not be cited or discussed, as it is not peer-reviewed.
line 94-96: include the input data here.
line 137-138 : there is no asterisk in the table and the meaning of the asterisk led note is not clear.
line 144 "Figure 3 shows the spatial and temporal distributions of surface water pCO2.", the spatial distribution of in-situ measurements or data from other source?
line 157: how the remote sensing-derived pCO2 data were derived?which methods, what is the quality? and output from unpublished study should not be used.
line 184-187: "Wang et al. (in preparation) found a relatively high differential between the....observed data", meaning of this super long sentence is not clear.
line 198 "pCO2 filling method of", should explain the filling method here!
line 201: "pCO2 reconstruction model" pCO2 reconstruction was used many times in the manuscript, but sea surface pCO2 is not something one can reconstruct, it is a properties or variable of of the sea water, one can measure it ,describe it, retrieve its distribution, but not reconstruct pco2 itself. So, please pay attention to the verb usage.
Citation: https://doi.org/10.5194/essd-2022-322-RC2 - AC2: 'Reply on RC2', Minhan Dai, 23 Dec 2022
Status: closed
-
RC1: 'Comment on essd-2022-322', Anonymous Referee #1, 17 Nov 2022
The comment was uploaded in the form of a supplement: https://essd.copernicus.org/preprints/essd-2022-322/essd-2022-322-RC1-supplement.pdf
- AC1: 'Reply on RC1', Minhan Dai, 23 Dec 2022
-
RC2: 'Comment on essd-2022-322', Anonymous Referee #2, 24 Nov 2022
the presented study aimed to produce monthly sea surface pCO2 maps for the South China Sea (SCS). Given SCS is a typical temperate/subtropical maringal sea, the pCO2 sea surface maps for this waters is necessary for understanding the CO2 flux in temperate marginal sea and even global CO2 flux. From this perspective, the study and the data it present is very meaningful. Howeve, the mansucript still have some majors flaws which do not advise me to give a yes to publishing it in its current status.
Major comments
1. The manuscript was about a dataset generation, but from the abstract and the last section, what kind of data was used as input for the method
was missing.2. As I understand EOF was an important part of the method used for pCO2 maps generation, but in the entire section of methods, no paragraph or sentence was about EOF
3. The language of the manuscript still need some efforts. The current version contains too many redundent phrases and sentences without clear meaning and very difficult to read through and get the logical flow. Readers expect concise and precise expresssion in an acedemica paper.
and there are some gammar mistake and fuzzy expression.4. the range of legend in nearly all the map figures were to large and cannot show the spatial gradient of pCO2 distribution, e.g, figure 6, 8, 11, 12,13.
5. what is the intention of including figure 4, if it is the quality of the remote sensing based pCO2 maps included for further pCO2 maps derivation, should the authors just need to include the information from the data distributor?
6. the study site section(2.1) should just serve the question "why mapping pCO2 in SCS is important?", no other information is needed here.
7. be consitent with the teminology, sometimes it is "in-situ", but "observational data" and "observed data" were present many times.
8. in the abstract (line 12-14,), the importance of mapping pCO2 in SCS should be addressed before presenting the method, generated data and its quality.
9 part of the input pCO2 data of the presented study is from unpublished study (line 158), meaning not peer-reviewed.
10, line 308: Figure 7, validating the model output with the mdoel training data gives no useful information, suggest removing this part
Minor comments
line 15-17""Using a machine learning-based method facilitated by empirical orthogonal function (EOF).... between 2003 and 2020" should specifically mention what kind of data was used for the methods input.
linse 17- 20 "We validate our reconstruction with three independent testing datasets where,.... northern basin of the SCS." how independent are the three data set?
line 22 "our reconstructions and observed data" grammar mistake.
Line 27-28 "we present a new method to assess the uncertainty that includes the bias from the reconstruction and its sensitivity to the features,... quantifies the spatial distribution patterns of uncertainty." then the assessment method should be concisely introduced here. in addition, given this is a data presentation paper, the newly developed method should not in the highlight, unless it is a method presentation paper.
line 19 "that our reconstruction is effectively captures the main features of both the" ,check the grammar.
line 38,, "22–26%", I assume it should be 22%–26%.
line 54-55: ":The former typically use statistical interpolations and regression methods" does not fit with the neighouring sentence, rewrite it or delete it.
line 61- 63 ,"However, because of the complex and dynamic nature of biogeochemical and physical processes in coastal areas, characterization of sea surface pCO2 and subsequently the
air-sea CO2 fluxes both in time and space in marginal seas remains challenging", this sentence is too strong and undermines the motivation of presented study, rewrite it,line 67: "clear need", what kind of need is clear need? a need can be strong, urgent, but not clear, need itsself is a clear expression,
line 73: "(sea surface temperature, SST; chlorophyll a, Chl a),", pay attention to journal requirements on abbrevation
line 74: "underway "pay attention to the usage of underway, it is ambiguous in the manuscript.
line 82, "the whole China Sea", where is the China Sea? do you mean all the seas in China's territory?
line 84: "(reported in Wang et al., 2021).", pay attention to the format of the reference citation
line 84: "Bai et al. (unpublished) subsequently", if the work is not publised, then it should not be cited or discussed, as it is not peer-reviewed.
line 94-96: include the input data here.
line 137-138 : there is no asterisk in the table and the meaning of the asterisk led note is not clear.
line 144 "Figure 3 shows the spatial and temporal distributions of surface water pCO2.", the spatial distribution of in-situ measurements or data from other source?
line 157: how the remote sensing-derived pCO2 data were derived?which methods, what is the quality? and output from unpublished study should not be used.
line 184-187: "Wang et al. (in preparation) found a relatively high differential between the....observed data", meaning of this super long sentence is not clear.
line 198 "pCO2 filling method of", should explain the filling method here!
line 201: "pCO2 reconstruction model" pCO2 reconstruction was used many times in the manuscript, but sea surface pCO2 is not something one can reconstruct, it is a properties or variable of of the sea water, one can measure it ,describe it, retrieve its distribution, but not reconstruct pco2 itself. So, please pay attention to the verb usage.
Citation: https://doi.org/10.5194/essd-2022-322-RC2 - AC2: 'Reply on RC2', Minhan Dai, 23 Dec 2022
Zhixuan Wang et al.
Data sets
Reconstructed Data, Observed Data, and CO2rs Zhixuan Wang, Minhan Dai https://github.com/Elricriven/co2data
Zhixuan Wang et al.
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
440 | 149 | 20 | 609 | 10 | 12 |
- HTML: 440
- PDF: 149
- XML: 20
- Total: 609
- BibTeX: 10
- EndNote: 12
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1