RetroSearch Browse

Home - News ( United States | United Kingdom | Italy | Germany ) - Football scores

Showing content from https://pubmed.ncbi.nlm.nih.gov/32025087 below:

Reassessing Southern Ocean Air-Sea CO2 Flux Estimates With the Addition of Biogeochemical Float Observations

. 2019 Nov;33(11):1370-1388. doi: 10.1029/2019GB006176. Epub 2019 Nov 16. Reassessing Southern Ocean Air-Sea CO₂ Flux Estimates With the Addition of Biogeochemical Float Observations

Affiliations

¹ Program in Atmospheric and Oceanic Sciences Princeton University Princeton NJ USA.
² Now at Department of Oceanography University of Hawai'i at Mānoa Honolulu HI USA.
³ Max Planck Institute for Meteorology Hamburg Germany.
⁴ Max Planck Institute for Biogeochemistry Jena Germany.
⁵ School of Oceanography University of Washington Seattle WA USA.
⁶ Cooperative Institute for Research in the Atmosphere Colorado State University Fort Collins CO USA.
⁷ Scripps Institution of Oceanography University of California, San Diego La Jolla CA USA.
⁸ Department of Geosciences and Princeton Environmental Institute Princeton University Princeton NJ USA.
⁹ Monterey Bay Aquarium Research Institute Moss Landing CA USA.

PMID: 32025087
PMCID: PMC6988491
DOI: 10.1029/2019GB006176

Item in Clipboard

Reassessing Southern Ocean Air-Sea CO₂ Flux Estimates With the Addition of Biogeochemical Float Observations

Seth M Bushinsky et al. Global Biogeochem Cycles. 2019 Nov.

. 2019 Nov;33(11):1370-1388. doi: 10.1029/2019GB006176. Epub 2019 Nov 16. Affiliations

¹ Program in Atmospheric and Oceanic Sciences Princeton University Princeton NJ USA.
² Now at Department of Oceanography University of Hawai'i at Mānoa Honolulu HI USA.
³ Max Planck Institute for Meteorology Hamburg Germany.
⁴ Max Planck Institute for Biogeochemistry Jena Germany.
⁵ School of Oceanography University of Washington Seattle WA USA.
⁶ Cooperative Institute for Research in the Atmosphere Colorado State University Fort Collins CO USA.
⁷ Scripps Institution of Oceanography University of California, San Diego La Jolla CA USA.
⁸ Department of Geosciences and Princeton Environmental Institute Princeton University Princeton NJ USA.
⁹ Monterey Bay Aquarium Research Institute Moss Landing CA USA.

PMID: 32025087
PMCID: PMC6988491
DOI: 10.1029/2019GB006176

Item in Clipboard

Abstract

New estimates of pCO₂ from profiling floats deployed by the Southern Ocean Carbon and Climate Observations and Modeling (SOCCOM) project have demonstrated the importance of wintertime outgassing south of the Polar Front, challenging the accepted magnitude of Southern Ocean carbon uptake (Gray et al., 2018, https://doi:10.1029/2018GL078013). Here, we put 3.5 years of SOCCOM observations into broader context with the global surface carbon dioxide database (Surface Ocean CO₂ Atlas, SOCAT) by using the two interpolation methods currently used to assess the ocean models in the Global Carbon Budget (Le Quéré et al., 2018, https://doi:10.5194/essd-10-2141-2018) to create a ship-only, a float-weighted, and a combined estimate of Southern Ocean carbon fluxes (<35°S). In our ship-only estimate, we calculate a mean uptake of -1.14 ± 0.19 Pg C/yr for 2015-2017, consistent with prior studies. The float-weighted estimate yields a significantly lower Southern Ocean uptake of -0.35 ± 0.19 Pg C/yr. Subsampling of high-resolution ocean biogeochemical process models indicates that some of the differences between float and ship-only estimates of the Southern Ocean carbon flux can be explained by spatial and temporal sampling differences. The combined ship and float estimate minimizes the root-mean-square pCO₂ difference between the mapped product and both data sets, giving a new Southern Ocean uptake of -0.75 ± 0.22 Pg C/yr, though with uncertainties that overlap the ship-only estimate. An atmospheric inversion reveals that a shift of this magnitude in the contemporary Southern Ocean carbon flux must be compensated for by ocean or land sinks within the Southern Hemisphere.

Keywords: SOCCOM; Southern Ocean; biogeochemical profiling floats; global carbon cycle.

PubMed Disclaimer

Figures
Figure 1

Shipboard and SOCCOM float sampling…

Figure 1

Shipboard and SOCCOM float sampling distributions. Temporal and spatial sampling density for (a)…

Figure 1

Shipboard and SOCCOM float sampling distributions. Temporal and spatial sampling density for (a) SOCAT: 1982–2017, (b) SOCAT: 2014–2017, and (c) SOCCOM. Colors represent the number of calendar months (maximum 12) with any data present in 3° latitude × 4° longitude grid cells. Grid cell size was chosen to approximate the decorrelation length scales of surface pCO₂ in the Southern Ocean (Jones et al., 2012). The 1982–2017 SOCAT period represents the entire range of data used as inputs to the interpolation methods used in this study. Example summer (February, d) and winter (August, e) months for 2014–2017 illustrate the bias in shipboard observations toward warmer months as well as remaining gaps in SOCCOM float coverage. SOCCOM = Southern Ocean Carbon and Climate Observations and Modeling; SOCAT = Surface Ocean CO₂ Atlas v6.

Figure 2

Seasonal and spatial sampling density…

Figure 2

Seasonal and spatial sampling density of SOCCOM and SOCAT observations in the Southern…

Figure 2

Seasonal and spatial sampling density of SOCCOM and SOCAT observations in the Southern Ocean. Area sampled per month is determined by counting the percent of 3° latitude × 4° longitude boxes with at least one observation in a given month. SOCAT observations (red lines) in the more southern regions are biased toward summertime sampling (year labels are on 1 January) and only seasonally does their coverage exceed 5% of a region's area in a given month. SOCCOM (blue line) float deployments began in April of 2014 and have rapidly increased sampling density in all seasons, surpassing area sampled by SOCAT in all regions except the STZ. SOCCOM = Southern Ocean Carbon and Climate Observations and Modeling; SOCAT = Surface Ocean CO₂ Atlas v6; STZ = Subtropical Zone; SAZ = Subantarctic Zone; PFZ = Polar Frontal Zone; ASZ = Antarctic‐Southern Zone; SIZ = Seasonal Ice Zone.

Figure 3

Neural network and Jena CarboScope‐derived…

Figure 3

Neural network and Jena CarboScope‐derived Southern Ocean (south of 35°S) air‐sea carbon dioxide…

Figure 3

Neural network and Jena CarboScope‐derived Southern Ocean (south of 35°S) air‐sea carbon dioxide fluxes. Addition of SOCCOM floats to the mapping product‐derived CO₂ fluxes reduces the mean 2015–2017 Southern Ocean carbon sink from −1.14 ± 0.19 Pg C/yr (orange circle) to −0.75 ± 0.22 Pg C/yr (black triangle), and removal of shipboard observations reduces the sink by an additional 0.4 Pg C/yr to −0.35 ± 0.19 Pg C/yr (purple square). Symbols and error bars represent the average of both products and the combined standard deviation of interannual variability, differences between the means of each product, and a method uncertainty of ±0.15 Pg C. The nonmapped, float‐only estimate of Gray et al. (2018) is included for reference (gray “x”). SOCCOM = Southern Ocean Carbon and Climate Observations and Modeling; SOCAT = Surface Ocean CO₂ Atlas v6.

Figure 4

Mean 2015–2017 summer, winter, and…

Figure 4

Mean 2015–2017 summer, winter, and annual Southern Ocean fluxes from all three p…

Figure 4

Mean 2015–2017 summer, winter, and annual Southern Ocean fluxes from all three pCO₂ products. Fronts and boundaries shown as black lines are (from north to south) as follows: the subtropical front, subantarctic front, polar front, and seasonal ice extent. Neural network‐derived fluxes show strong summer (November to April) uptake in the SOCAT‐only product combines with moderate winter (May to October) outgassing around the Polar Front to give an annual flux with little outgassing evident. Addition of the SOCCOM floats produces stronger outgassing in the winter SOCAT+SOCCOM map, and has a moderate impact on the mean annual flux. The SOCCOM‐weighted product displays outgassing around the Polar Front even in the summer and a very strong winter signal. Equivalent fluxes for Jena interpolation scheme in Figure S1 in the supporting information. SOCCOM = Southern Ocean Carbon and Climate Observations and Modeling; SOCAT = Surface Ocean CO₂ Atlas v6.

Figure 5

Impact of adding SOCCOM floats…

Figure 5

Impact of adding SOCCOM floats on neural network CO ₂ fluxes. SOCAT+SOCCOM monthly…

Figure 5

Impact of adding SOCCOM floats on neural network CO₂ fluxes. SOCAT+SOCCOM monthly mean air‐sea CO₂ fluxes (2015–2017) minus SOCAT‐only monthly mean fluxes. Red values indicate regions where addition of float‐estimated pCO₂ decreased the Southern Ocean sink. Equivalent figure for the Jena interpolation scheme in supporting information (Figure S2). SOCCOM = Southern Ocean Carbon and Climate Observations and Modeling; SOCAT = Surface Ocean CO₂ Atlas v6.

Figure 6

Impact of removing SOCAT shipboard…

Figure 6

Impact of removing SOCAT shipboard data on neural network CO ₂ fluxes. SOCCOM‐weighted…

Figure 6

Impact of removing SOCAT shipboard data on neural network CO₂ fluxes. SOCCOM‐weighted monthly mean air‐sea CO₂ fluxes (2015–2017) were subtracted from SOCAT+SOCCOM monthly mean fluxes. Red values indicate regions where removal of ship‐measured pCO₂ decreased the Southern Ocean sink or increased outgassing. Equivalent figure for the Jena interpolation scheme in supporting information (Figure S3). SOCCOM = Southern Ocean Carbon and Climate Observations and Modeling; SOCAT = Surface Ocean CO₂ Atlas v6.

Figure 7

Mean annual fluxes for Southern…

Figure 7

Mean annual fluxes for Southern Ocean regions from 2015–2017 for two data‐based products…

Figure 7

Mean annual fluxes for Southern Ocean regions from 2015–2017 for two data‐based products and a float‐only average. Neural network fluxes (grays), Jena mixed layer scheme (blues), and float‐only fluxes (colors) all indicate a reduction in the Southern Ocean carbon sink as the influence of the float observations is increased. The largest impacts are in the Subantarctic Zone (SAZ), Polar‐Frontal Zone (PFZ), and Antarctic‐Southern Zone (ASZ), which combine with smaller differences in the fluxes of the other regions to yield a large overall impact in the total Southern Ocean. Error bars represent ±1 s.d. SOCCOM = Southern Ocean Carbon and Climate Observations and Modeling; SOCAT = Surface Ocean CO₂ Atlas v6.

Figure 8

True model air‐sea flux, air…

Figure 8

True model air‐sea flux, air sea fluxes based on neural network derived from…

Figure 8

True model air‐sea flux, air sea fluxes based on neural network derived from SOCAT‐only and SOCAT+SOCCOM sample locations, and flux differences from model. Air‐sea fluxes were calculated in the same way for true model pCO₂ and neural network mapped pCO₂. The region with the most consistent differences between the neural network derived fluxes and model fluxes is the Southern Ocean. Addition of the float locations to the neural network decreases some of the large biases near Antarctica and reduces the root‐mean‐square differences between model and mapped pCO₂ (Figure 9). SOCCOM = Southern Ocean Carbon and Climate Observations and Modeling; SOCAT = Surface Ocean CO₂ Atlas v6.

Figure 9

Model subsampling experiment with neural…

Figure 9

Model subsampling experiment with neural network results. (a) The difference between p CO…

Figure 9

Model subsampling experiment with neural network results. (a) The difference between pCO₂ RMSD for the SOCAT‐only and SOCAT+SOCCOM mapping of CM2.6 and SOSE (SOCAT‐only RMSD minus SOCAT+SOCCOM; positive values indicate an improvement in re‐creation of model output). (b) The percent improvement of neural network flux RMSD. For all cases, solid bars represent annual means, light bars represent summer (December–February), and outlined bars with no fill represent winter (June–August). The neural network approach for CM2.6 was run both with a climatological pCO₂ input to the SOM based on model output and without any pCO₂ climatology to demonstrate the sensitivity of the method to prior knowledge of the true seasonal cycle of pCO₂. SOCCOM = Southern Ocean Carbon and Climate Observations and Modeling; SOCAT = Surface Ocean CO₂ Atlas v6; STZ = Subtropical Zone; SAZ = Subantarctic Zone; PFZ = Polar Frontal Zone; ASZ = Antarctic‐Southern Zone; SIZ = Seasonal Ice Zone; RMSD = root‐mean‐square differences; RMSE = root‐mean‐square error.

Figure 10

Example of p CO2 observations…

Figure 10

Example of p CO2 observations from a mooring off the coast of Tasmania…

Figure 10

Example of pCO2 observations from a mooring off the coast of Tasmania (a) and the impact on the mean annual flux from subsampling pCO₂ observations (b). (a) Observations of pCO₂ made every 2 hr were subsampled at increasing intervals. Each subsampled record was then interpolated back to a daily time step, and air‐sea CO₂ fluxes were calculated from daily atmospheric reanalysis data using the same method as for SOCCOM floats. The difference between mean annual fluxes from daily and subsampled records are plotted for 5 years of data (b), with the shaded areas representing the range of 12 analysis iterations (i.e., starting the subsampling on a different 2‐hr time step). Float observations are made once every 10 days, which should bias the mean annual flux by less than ±0.1 mol·m⁻²·yr⁻¹ (<0.1 Pg C/yr over the entire Southern Ocean). SOCCOM = Southern Ocean Carbon and Climate Observations and Modeling.

Figure 11

Mean atmospheric inversion derived land…

Figure 11

Mean atmospheric inversion derived land carbon fluxes for 2015–2017. Global ocean CO _2…

Figure 11

Mean atmospheric inversion derived land carbon fluxes for 2015–2017. Global ocean CO₂ flux for SOCAT only and SOCAT+SOCCOM products based on the Jena CarboScope interpolation output (a) and land CO₂ fluxes (b) calculated from the Jena CarboScope atmospheric inversion (positive to the atmosphere). The atmospheric inversion uses the oceanic fluxes as a fixed boundary condition, so this is primarily a test of the needed change in land CO₂ fluxes to balance a change in the Southern Ocean flux. The difference indicates that compensating fluxes in the land or ocean would have a northern limit based on the atmospheric transfer model of ~5°S. SOCCOM = Southern Ocean Carbon and Climate Observations and Modeling; SOCAT = Surface Ocean CO₂ Atlas v6.