Storage and redistribution of anthropogenic CO2 in the western North Pacific: The role of subtropical mode water transportation

Highlights • Anthropogenic CO2 (CANT) storage was investigated in the western North Pacific.• CANT inventories in water column and in specific water masses were estimated.• Subtropical mode water dominates CANT dynamics in the Kuroshio Recirculation region.


Reference Simulation of the North Pacific Subtropical Mode Water
To examine the possible mechanism of subducted C ANT , the characteristics of the North Pacific Subtropical Mode Water (STMW) were simulated using the MASNUM three-dimensional, primitive equation, numerical ocean circulation model.This model adopts most of the numerical schemes of the Princeton Ocean Model (POM) but uses a two-time-level algorithm for time stepping.This modification helps remove the computational mode inherent in the three-time-level scheme such as the leapfrog method adopted by POM and many other ocean circulation models [1].In addition to inheriting other aspects of the numerical features of POM, MASNUM exhibited better performance in terms of numerical stability [2][3][4].
In this case, a MASNUM global simulation with horizontal resolution of 0.25° × 0.25° and 31 sigma levels was run under the GEBCO 08 topography (www.gebco.net).
The simulation started from rest with initial temperature and salinity distributions obtained from WOA09 [5][6] and forced by surface fluxes of the NCEP-DOE AMIP-II Reanalysis [7].The model was integrated until the total kinetic energy of the global ocean reached a quasi-equilibrium state [2].The upper ocean where the STMW exists was fully spun up to depict the seasonal evolution of this water mass.

Sampling and analytical analyses for nutrients
Descriptions of the methods used for sampling and determination of discrete water samples for nutrients (nitrate, nitrite, phosphate, and silicate) in May 2018 are given below.In the surveys, duplicate water samples were unfiltered and stored in 60 mL high-density polyethylene bottles.The samples were frozen at −20 °C and then analyzed later in the laboratory for nitrate, nitrite, dissolved inorganic phosphate (DIP), and dissolved silicate (DSi).Nitrate (NO 3 − ) and nitrite (NO 2 − ) were measured by reducing nitrate to nitrite with a cadmium column, followed by determination of nitrite using the pink azo dye spectrophotometric method [8].DIP and DSi were measured based on standard phosphomolybdenum blue (for DIP) and silicon molybdenum blue (for DSi) spectrophotometric procedures [9].Nitrate, nitrite, DIP, and DSi were determined using a reorganized Syslyzer Ⅲ analyzer (Systea S.P.A. Co., Italy).Detection limits were 0.1 μmol L −1 for nitrate, 0.04 μmol L −1 for nitrite, 0.08 μmol L −1 for DIP, and 0.08 μmol L −1 for DSi [10].Samples with high nutrient concentrations were properly diluted before determination.For data quality assurance, Reference Materials for Nutrients in Seawater from The General Environmental Technos Co., Ltd.(http://www.kanso.co.jp/) were used during the analyses, and a precision level of ± 0.1 μmol L −1 was achieved for most nutrient species other than ammonium.

Comparison of anthropogenic CO 2 estimations from different methods
To evaluate the validation and precision of the TrOCA method in relation to the North Pacific, we compared it with the traditional C* method used for calculating C ANT [11][12].Briefly, the C* method can be expressed as follows: where 170 is the O 2 /P ratio and  is potential temperature.
The C ANT values calculated using the TrOCA method (TrOCA_C ANT ) and the C* method (C*_C ANT ) were comparable with each other (Fig. S3).The result showed that the C ANT values obtained using the two methods were consistent with each other at deviation levels of ± 3 µmol kg 1 (n = 1112) between 100 and 2000 m, within the uncertainty of the TrOCA method (± 6 μmol kg −1 ), which provided confidence in the validation and precision of the TrOCA method for the western North Pacific.
Introduction to Simulation of the North Pacific Subtropical Mode Water Introduction to Sampling and analytical analyses for nutrients Comparison of anthropogenic CO 2 estimations from different methods Table S1.Characteristics of deep waters (~1000 m) in the western North Pacific (21-35N, 123-147E) in May and August 2018 and May 2005.Note that "(Aug 2018  May 2005)" indicates the differences in parameters between August 2018 and May 2005.

Fig. S1 .
Fig. S1.Different criteria used to define the Subtropical Mode Water along the 147°E transect in the North Pacific.Fig. S2.Comparison of anthropogenic CO 2 values calculated using different methods.Fig. S3.Monthly mean sea surface height (m) in March, April, May, and August 2018.
S1) C* = C m − C 280 − C bio , (S2) where the quasi-conservative tracer (C*) is defined as the difference between the measured DIC (C m ) corrected for biology (C bio ) and the air-equilibrated DIC with a preindustrial atmosphere level of CO 2 of 280 µatm at the water surface (C 280 ); C diseq is the airsea CO 2 difference expressed in terms of DIC, and the mean value of −6 ± 3 µmol kg 1 for the North Pacific [11] was used to calculate the C ANT ; C bio is the DIC change attributable to remineralization of organic matter and the dissolution of calcium carbonate particles and denitrification: C bio = 117/170  AOU + 0.5 (TA − TA + 16/170  AOU) − 106/104  (N* − N* mean ), (S3) where 117/170 and 16/170 are the C/O 2 and N/O 2 ratios, respectively [13].The denitrification term was 106/104  (N* − N* mean ) [14], where N* = (N − 16P + 2.90) and the mean N* (N* mean ) value for our dataset was 0 µmol kg 1 .In this study, TA was estimated for the North Pacific using a multiple linear regression of the surface alkalinity values to conservative tracers: TA = 148.7 + 61.36  Salinity + 16/170  (O 2 + 170  P) − 0.582   , (S4)

Fig. S1 .
Fig. S1.Different criteria used to define the Subtropical Mode Water along the 147°E transect in the North Pacific.Color fills denote the vertical derivative of the potential temperature, dT/dz.Pink, purple, and green lines represent contour lines of potential density, temperature, and potential vorticity (PV = 2  10 10 m 1 s 1 ), respectively.White lines denote the dT/dz criterion[15].The data were produced by a global run of the MASNUM ocean circulation model[2].

Fig. S3 .
Fig. S3.Monthly mean sea surface height (m) in March, April, May, and August 2018 obtained from https://marine.copernicus.eu(red rectangles indicate the region where a mesoscale eddy was observed in May 2018).

Table S1 .
Characteristics of deep waters (~1000 m) in the western North Pacific (123-147E, 21-35N) in May and August 2018 and May 2005.Note that "(Aug 2018  May 2005)" indicates the differences in parameters between August 2018 and May 2005.