Interpolated surface ocean carbon dioxide partial pressure for the South Atlantic Ocean (2002-2018) using different biological parameters

• Main Authors: Ford, Daniel J, Tilstone, Gavin H, Shutler, Jamie D, Kitidis, Vassilis
• Format: Dataset
• Language: English
• Published: PANGAEA - Data Publisher for Earth & Environmental Science 2021
• Subjects: chlorophyll-a; Net community production; net primary production; pCO2; South Atlantic Ocean; File content; Binary Object File Size; Binary Object; interpolated
• Online Access: https://dx.doi.org/10.1594/pangaea.935936; https://doi.pangaea.de/10.1594/PANGAEA.935936

The dataset contains interpolated fields of surface ocean partial pressure of carbon dioxide (pCO2sw) for the South Atlantic Ocean (10N – 60S; 25E-70W) on a monthly time and 1 degree latitude and longitude grid, between July 2002 and December 2018. The pCO2sw is interpolated using SOCATv2020 observational data, that have been corrected to the surface sub skin temperature, using a neural network interpolation scheme described in Ford et al. (2022).Five separate pCO2sw estimates are provided using different biological parameters as input to the interpolation scheme. These consist of net community production (NCP), net primary production, chlorophyll-a, and two variants with no biological parameters as input. Full details are given in the article published in Biogeosciences. This article highlights that the NCP variant is the most accurate and therefore is the recommended version. : The data submission contains five netCDF files. The files contain the pCO2sw estimates for the South Atlantic Ocean from five neural network interpolation schemes, using different biological parameters as input, as described in Ford et al. (2022). These biological parameters are net community production (FULL_PCO2_SAFNN_NCP_072002_122018.nc), net primary production (FULL_PCO2_SAFNN_NPP_072002_122018.nc) and chlorophyll-a (FULL_PCO2_SAFNN_CHL_072002_122018.nc). Two variants of the neural network without biological parameters are also presented: one trained with SST (FULL_PCO2_SAFNN_NO_BIO_1_072002_122018.nc) and the second with SST, salinity, and mixed layer depth (FULL_PCO2_SAFNN_NO_BIO_2_072002_122018.nc). Inputs parameters for all the variants are shown in Table 2 of the accompanying manuscript (Ford et al., 2022).Within each netCDF, five variables are present:latitude – The latitude grid in degrees north (10N to -60N).longitude – The longitude grid in degrees east (25E to -75E).time – MATLAB date number for the 1st day of each month. This is 198 months spanning July 2002 to December 2018pCO2sw – The interpolated partial pressure of CO2 in seawater corrected to the surface sub skin in μatm.pCO2sw_unc – The propagated uncertainty on the partial pressure of CO2 in seawater in μatm.