The sensitivity of <i>p</i>CO<sub>2</sub> reconstructions to sampling scales across a Southern Ocean sub-domain: a semi-idealized ocean sampling simulation approach
<p>The Southern Ocean is a complex system yet is sparsely sampled in both space and time. These factors raise questions about the confidence in present sampling strategies and associated machine learning (ML) reconstructions. Previous studies have not yielded a clear understanding of the origi...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2022-09-01
|
| Series: | Biogeosciences |
| Online Access: | https://bg.copernicus.org/articles/19/4171/2022/bg-19-4171-2022.pdf |
| _version_ | 1828749208820121600 |
|---|---|
| author | L. M. Djeutchouang L. M. Djeutchouang N. Chang L. Gregor M. Vichi P. M. S. Monteiro |
| author_facet | L. M. Djeutchouang L. M. Djeutchouang N. Chang L. Gregor M. Vichi P. M. S. Monteiro |
| author_sort | L. M. Djeutchouang |
| collection | DOAJ |
| description | <p>The Southern Ocean is a complex system yet is sparsely
sampled in both space and time. These factors raise questions about the
confidence in present sampling strategies and associated machine learning
(ML) reconstructions. Previous studies have not yielded a clear
understanding of the origin of uncertainties and biases for the
reconstructions of the partial pressure of carbon dioxide
(<span class="inline-formula"><i>p</i></span>CO<span class="inline-formula"><sub>2</sub></span>) at the surface ocean
(<span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M7" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="384f359723e12c1d22fe40db07699d29"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00001.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00001.png"/></svg:svg></span></span>). We examine these questions through
a series of semi-idealized observing system simulation experiments (OSSEs)
using a high-resolution (<span class="inline-formula">±</span> 10 km) coupled physical and biogeochemical
model (NEMO-PISCES, Nucleus for European Modelling of the Ocean, Pelagic Interactions Scheme for Carbon and Ecosystem Studies). Here we choose 1 year of the model sub-domain of 10<span class="inline-formula"><sup>∘</sup></span> of latitude (40–50<span class="inline-formula"><sup>∘</sup></span> S) by 20<span class="inline-formula"><sup>∘</sup></span> of longitude (10<span class="inline-formula"><sup>∘</sup></span> W–10<span class="inline-formula"><sup>∘</sup></span> E). This
domain is crossed by the sub-Antarctic front and thus includes both the
sub-Antarctic zone and the polar frontal zone in the south-east Atlantic Ocean,
which are the two most sampled sub-regions of the Southern Ocean. We show
that while this sub-domain is small relative to the Southern Ocean scales,
it is representative of the scales of variability we aim to examine. The
OSSEs simulated the observational scales of
<span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M14" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="03654b13fda64c41bd29e3a2fae0ca33"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00002.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00002.png"/></svg:svg></span></span> in ways that are comparable to
existing ocean CO<span class="inline-formula"><sub>2</sub></span> observing platforms (ships, Wave Gliders,
carbon floats, Saildrones) in terms of their temporal sampling scales and
not necessarily their spatial ones. The <span class="inline-formula"><i>p</i></span>CO<span class="inline-formula"><sub>2</sub></span> reconstructions
were carried out using a two-member ensemble approach that consisted of two machine
learning (ML) methods, (1) the feed-forward neural network and (2) the
gradient boosting machines. The baseline data were from the ship-based
simulations mimicking ship-based observations from the Surface Ocean
CO<span class="inline-formula"><sub>2</sub></span> Atlas (SOCAT). For each of the sampling-scale scenarios, we applied
the two-member ensemble method to reconstruct the full sub-domain
<span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M19" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="05185283c850a6110de151686835d32b"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00003.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00003.png"/></svg:svg></span></span>. The reconstruction skill was then
assessed through a statistical comparison of reconstructed
<span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M20" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="2d2db79ec52577d5689a18b01aacfaef"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00004.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00004.png"/></svg:svg></span></span> and the model domain mean. The analysis
shows that uncertainties and biases for
<span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M21" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="a7732e51d91c5c5b155e15a8a70f9088"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00005.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00005.png"/></svg:svg></span></span> reconstructions are very sensitive to
both the spatial and the temporal scales of <span class="inline-formula"><i>p</i></span>CO<span class="inline-formula"><sub>2</sub></span> sampling in the
model domain. The four key findings from our investigation are as follows: (1) improving ML-based <span class="inline-formula"><i>p</i></span>CO<span class="inline-formula"><sub>2</sub></span> reconstructions in the Southern
Ocean requires simultaneous high-resolution observations (<span class="inline-formula"><i><</i>3</span> d)
of the seasonal cycle of the meridional gradients of
<span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M27" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="8ce241c9be9d6b7a0dd2d295e17dcfd6"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00006.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00006.png"/></svg:svg></span></span>; (2) Saildrones stand out as the
optimal platforms to simultaneously address these requirements; (3) Wave Gliders with hourly/daily resolution in pseudo-mooring mode improve on
carbon floats (10 d period), which suggests that sampling aliases from the
10 d sampling period might have a greater negative impact on their
uncertainties, biases, and reconstruction means; and (4) the present
seasonal sampling biases (towards summer) in SOCAT data in the Southern
Ocean may be behind a significant winter bias in the reconstructed seasonal
cycle of <span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M28" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="7a393e68c77e236d6b2e544506c216b1"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00007.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00007.png"/></svg:svg></span></span>.</p> |
| first_indexed | 2024-12-10T11:48:29Z |
| format | Article |
| id | doaj.art-2c99341eaafe4e46b0bac103b125b0a3 |
| institution | Directory Open Access Journal |
| issn | 1726-4170 1726-4189 |
| language | English |
| last_indexed | 2024-12-10T11:48:29Z |
| publishDate | 2022-09-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | Biogeosciences |
| spelling | doaj.art-2c99341eaafe4e46b0bac103b125b0a32022-12-22T01:50:00ZengCopernicus PublicationsBiogeosciences1726-41701726-41892022-09-01194171419510.5194/bg-19-4171-2022The sensitivity of <i>p</i>CO<sub>2</sub> reconstructions to sampling scales across a Southern Ocean sub-domain: a semi-idealized ocean sampling simulation approachL. M. Djeutchouang0L. M. Djeutchouang1N. Chang2L. Gregor3M. Vichi4P. M. S. Monteiro5SOCCO, CSIR, Rosebank, Cape Town, 7700, South AfricaMARIS, Department of Oceanography, University of Cape Town, Cape Town, 7701, South AfricaSOCCO, CSIR, Rosebank, Cape Town, 7700, South AfricaEnvironmental Physics, Institute of Biogeochemistry and Pollutant Dynamics, ETH Zürich, 8092 Zurich, SwitzerlandMARIS, Department of Oceanography, University of Cape Town, Cape Town, 7701, South AfricaSOCCO, CSIR, Rosebank, Cape Town, 7700, South Africa<p>The Southern Ocean is a complex system yet is sparsely sampled in both space and time. These factors raise questions about the confidence in present sampling strategies and associated machine learning (ML) reconstructions. Previous studies have not yielded a clear understanding of the origin of uncertainties and biases for the reconstructions of the partial pressure of carbon dioxide (<span class="inline-formula"><i>p</i></span>CO<span class="inline-formula"><sub>2</sub></span>) at the surface ocean (<span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M7" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="384f359723e12c1d22fe40db07699d29"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00001.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00001.png"/></svg:svg></span></span>). We examine these questions through a series of semi-idealized observing system simulation experiments (OSSEs) using a high-resolution (<span class="inline-formula">±</span> 10 km) coupled physical and biogeochemical model (NEMO-PISCES, Nucleus for European Modelling of the Ocean, Pelagic Interactions Scheme for Carbon and Ecosystem Studies). Here we choose 1 year of the model sub-domain of 10<span class="inline-formula"><sup>∘</sup></span> of latitude (40–50<span class="inline-formula"><sup>∘</sup></span> S) by 20<span class="inline-formula"><sup>∘</sup></span> of longitude (10<span class="inline-formula"><sup>∘</sup></span> W–10<span class="inline-formula"><sup>∘</sup></span> E). This domain is crossed by the sub-Antarctic front and thus includes both the sub-Antarctic zone and the polar frontal zone in the south-east Atlantic Ocean, which are the two most sampled sub-regions of the Southern Ocean. We show that while this sub-domain is small relative to the Southern Ocean scales, it is representative of the scales of variability we aim to examine. The OSSEs simulated the observational scales of <span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M14" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="03654b13fda64c41bd29e3a2fae0ca33"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00002.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00002.png"/></svg:svg></span></span> in ways that are comparable to existing ocean CO<span class="inline-formula"><sub>2</sub></span> observing platforms (ships, Wave Gliders, carbon floats, Saildrones) in terms of their temporal sampling scales and not necessarily their spatial ones. The <span class="inline-formula"><i>p</i></span>CO<span class="inline-formula"><sub>2</sub></span> reconstructions were carried out using a two-member ensemble approach that consisted of two machine learning (ML) methods, (1) the feed-forward neural network and (2) the gradient boosting machines. The baseline data were from the ship-based simulations mimicking ship-based observations from the Surface Ocean CO<span class="inline-formula"><sub>2</sub></span> Atlas (SOCAT). For each of the sampling-scale scenarios, we applied the two-member ensemble method to reconstruct the full sub-domain <span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M19" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="05185283c850a6110de151686835d32b"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00003.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00003.png"/></svg:svg></span></span>. The reconstruction skill was then assessed through a statistical comparison of reconstructed <span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M20" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="2d2db79ec52577d5689a18b01aacfaef"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00004.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00004.png"/></svg:svg></span></span> and the model domain mean. The analysis shows that uncertainties and biases for <span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M21" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="a7732e51d91c5c5b155e15a8a70f9088"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00005.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00005.png"/></svg:svg></span></span> reconstructions are very sensitive to both the spatial and the temporal scales of <span class="inline-formula"><i>p</i></span>CO<span class="inline-formula"><sub>2</sub></span> sampling in the model domain. The four key findings from our investigation are as follows: (1) improving ML-based <span class="inline-formula"><i>p</i></span>CO<span class="inline-formula"><sub>2</sub></span> reconstructions in the Southern Ocean requires simultaneous high-resolution observations (<span class="inline-formula"><i><</i>3</span> d) of the seasonal cycle of the meridional gradients of <span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M27" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="8ce241c9be9d6b7a0dd2d295e17dcfd6"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00006.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00006.png"/></svg:svg></span></span>; (2) Saildrones stand out as the optimal platforms to simultaneously address these requirements; (3) Wave Gliders with hourly/daily resolution in pseudo-mooring mode improve on carbon floats (10 d period), which suggests that sampling aliases from the 10 d sampling period might have a greater negative impact on their uncertainties, biases, and reconstruction means; and (4) the present seasonal sampling biases (towards summer) in SOCAT data in the Southern Ocean may be behind a significant winter bias in the reconstructed seasonal cycle of <span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M28" display="inline" overflow="scroll" dspmath="mathml"><mrow><mi>p</mi><msubsup><mi mathvariant="normal">CO</mi><mn mathvariant="normal">2</mn><mi mathvariant="normal">ocean</mi></msubsup></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="47pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="7a393e68c77e236d6b2e544506c216b1"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="bg-19-4171-2022-ie00007.svg" width="47pt" height="15pt" src="bg-19-4171-2022-ie00007.png"/></svg:svg></span></span>.</p>https://bg.copernicus.org/articles/19/4171/2022/bg-19-4171-2022.pdf |
| spellingShingle | L. M. Djeutchouang L. M. Djeutchouang N. Chang L. Gregor M. Vichi P. M. S. Monteiro The sensitivity of <i>p</i>CO<sub>2</sub> reconstructions to sampling scales across a Southern Ocean sub-domain: a semi-idealized ocean sampling simulation approach Biogeosciences |
| title | The sensitivity of <i>p</i>CO<sub>2</sub> reconstructions to sampling scales across a Southern Ocean sub-domain: a semi-idealized ocean sampling simulation approach |
| title_full | The sensitivity of <i>p</i>CO<sub>2</sub> reconstructions to sampling scales across a Southern Ocean sub-domain: a semi-idealized ocean sampling simulation approach |
| title_fullStr | The sensitivity of <i>p</i>CO<sub>2</sub> reconstructions to sampling scales across a Southern Ocean sub-domain: a semi-idealized ocean sampling simulation approach |
| title_full_unstemmed | The sensitivity of <i>p</i>CO<sub>2</sub> reconstructions to sampling scales across a Southern Ocean sub-domain: a semi-idealized ocean sampling simulation approach |
| title_short | The sensitivity of <i>p</i>CO<sub>2</sub> reconstructions to sampling scales across a Southern Ocean sub-domain: a semi-idealized ocean sampling simulation approach |
| title_sort | sensitivity of i p i co sub 2 sub reconstructions to sampling scales across a southern ocean sub domain a semi idealized ocean sampling simulation approach |
| url | https://bg.copernicus.org/articles/19/4171/2022/bg-19-4171-2022.pdf |
| work_keys_str_mv | AT lmdjeutchouang thesensitivityofipicosub2subreconstructionstosamplingscalesacrossasouthernoceansubdomainasemiidealizedoceansamplingsimulationapproach AT lmdjeutchouang thesensitivityofipicosub2subreconstructionstosamplingscalesacrossasouthernoceansubdomainasemiidealizedoceansamplingsimulationapproach AT nchang thesensitivityofipicosub2subreconstructionstosamplingscalesacrossasouthernoceansubdomainasemiidealizedoceansamplingsimulationapproach AT lgregor thesensitivityofipicosub2subreconstructionstosamplingscalesacrossasouthernoceansubdomainasemiidealizedoceansamplingsimulationapproach AT mvichi thesensitivityofipicosub2subreconstructionstosamplingscalesacrossasouthernoceansubdomainasemiidealizedoceansamplingsimulationapproach AT pmsmonteiro thesensitivityofipicosub2subreconstructionstosamplingscalesacrossasouthernoceansubdomainasemiidealizedoceansamplingsimulationapproach AT lmdjeutchouang sensitivityofipicosub2subreconstructionstosamplingscalesacrossasouthernoceansubdomainasemiidealizedoceansamplingsimulationapproach AT lmdjeutchouang sensitivityofipicosub2subreconstructionstosamplingscalesacrossasouthernoceansubdomainasemiidealizedoceansamplingsimulationapproach AT nchang sensitivityofipicosub2subreconstructionstosamplingscalesacrossasouthernoceansubdomainasemiidealizedoceansamplingsimulationapproach AT lgregor sensitivityofipicosub2subreconstructionstosamplingscalesacrossasouthernoceansubdomainasemiidealizedoceansamplingsimulationapproach AT mvichi sensitivityofipicosub2subreconstructionstosamplingscalesacrossasouthernoceansubdomainasemiidealizedoceansamplingsimulationapproach AT pmsmonteiro sensitivityofipicosub2subreconstructionstosamplingscalesacrossasouthernoceansubdomainasemiidealizedoceansamplingsimulationapproach |