NorESM-bcoi-v1 reanalysis data for ocean biogeochemistry at >40° N, 1980-2020, and basis compilations of quality-controlled observations (NORBGCv1).

Datasets associated with Wallhead et al. "NorESM-bcoi-v1: A bias-corrected reanalysis of ocean biogeochemistry at >40° N, 1980-2020, based on a global ocean model hindcast" manuscript submitted to ESSD. Reanalysis data in files <variable name>_NorESM-bcoi-v1_40N_1980-2020.nc are 4D gridded biogeochemical datasets covering latitudes >40 degrees N and years 1980-2020 inclusive. Spatial resolution is on the original NorESM-OCv1.2 grid (27–82 km in the horizontal, with higher resolution further north, 5–1000 m in the vertical, higher near the surface). Temporal resolution is monthly. NorESM-bcoi-v1 reanalysis data are provided in NetCDF format for: no3no2 = nitrate plus nitrite (NO3+NO2 in manuscript), po4 = phosphate, si = silicate, o2 = dissolved oxygen, dissic = dissolved inorganic carbon (DIC in manuscript), talk = total alkalinity (TA in manuscript), dissicn = normalized dissolved inorganic carbon (DICn in manuscript), talkn = normalized total alkalinity (TAn in manuscript). Each of these variables is provided in a separate NetCDF file. All units are mol/m3 in the corresponding in situ conditions of (temperature, salinity, pressure). Each NetCDF file contains, in addition to one of the above reanalysis variables, the following variables: depth: mid-layer depths of the renalysis data (m). This is a 1D variable describing uniform depth levels ranging from 1.25 m to 7312.5 m, with irregular spacing, same as the original NorESM hindcast data. time: nominal timestamps of the reanalysis data (days since 1970-01-01 00:00:00). This is a 1D variable describing the nominal timestamps of the monthly reanalysis data, positioned on the 15th day of each month and covering 1980-01-15 through 2020-12-15. depth_bnds: depths of upper and lower layer surfaces (m). This is 2D variable describing the depths of upper/lower depth layer surfaces, where the second dimension 'bounds' specifies upper/lower, and the values range over 0-8000 m, same as the original NorESM hindcast data. lat: latitude of the horizontal grid locations (degrees north). This is a 2D variable describing the latitude of the each horizontal grid location defined by the dimension variables (x, y). Values range over 40.4 to 89.8 degN and comprise a subset of the original NorESM hindcast values. NOTE: The NorESM grid is irregular over (latitude, longitude), hence the latitudes and longitudes are described by 2D rather than 1D variables. lon: longitude of the horizontal grid locations (degrees east). This is a 2D variable describing the longitude of the each horizontal grid location defined by the dimension variables (x, y). Values range over -180.0 to 180 degE. The reanalysis variables themselves are dimensioned over (time, depth, y, x). This does mean that care is needed to extract subsets defined by latitude-longitude ranges. One must first establish subselection indices over the dimension variables (x, y) using the 2D variables (lat, lon). The irregular grid spacing over latitude-longitude should also be noted when interpolating the reanalysis output to new locations. We also provide the reanalysis data for (temperature, salinity) from SODA3.4.2, interpolated onto the NorESM grid, that were used to back-transform the normalized reanalysis data for (DICn, TAn), to obtain the provided reanalysis data for (DIC, TA). Users with higher resolution or more accurate regional reanalysis data for (temperature, salinity) should consider using these in combination with the (DICn, TAn) reanalysis data and the normalization relations described in the manuscript (also in the DIC, TA, variable attributes). We also provide the quality-controlled in situ observation compilations (NORBGCv1) used to develop the reanalysis data. These consist of 3 NetCDF files: NORBGCv1_NUT.nc, containing all (NO3+NO2, PO4, Si) data with coincident (T, S, O2) data where available; NORBGCv1_DO.nc, containing all O2 data with coincident (T, S) data where available; NORBGCv1_CO2.nc, containing all (DIC, TA) data with coincident (T, S, NO3+NO2, PO4, Si, O2) data where available. These files also contain indicator variables that will allow users to replicate the training and test subsets used in the manuscript for skill assessment (including the "final" test subsets, accounting for constraints on distance to shoreline and availability of interpolative data from NorESM, WOA, and GLODAPv2gridded products). A set of figures NORBGCv1_figs.zip is provided to illustrate the results of the observational quality control. In these plots, observations with suspect QC flag values (values >2) are coloured magenta. Data with acceptable QC flag values (0,1,2), that are used in the reanalysis, are coloured black (or white in the horizontal map subplots). These plots exclude data within 40 km of the coast, where QC is relaxed to allow for coastal effects, and all data in the Baltic Sea (to improve clarity in profile analyses). Matlab code used to develop and analyse the NorESM-bcoi-v1 reanalysis data, and to develop the NORBGCv1 compilations, is provided in code_NorESMbcoiv1_NORBGCv1.zip. All functions used by the relevant scripts are provided in subfolders. As intermediate products, we also provide .mat files for the observational data (obs_mat.zip), for training/test subsets of the observational data (obs_data_subsets_mat.zip), and model-observation matchup values (where the various model products are interpolated over time and space to the observational coordinates, see model-obs_matchups_mat.zip).