Dissolved iron (dFe) plays an important role in regulating marine biological productivity. In high-nutrient, low-chlorophyll regions (> 33% of the global ocean) iron is the primary growth-limiting nutrient, and elsewhere iron can regulate nitrogen fixation and growth by diazotrophs. The link between iron availability and carbon export is strongly dependent on the phytoplankton iron quotas, or cellular Fe:C ratios. This ratio can vary by more than an order of magnitude in the open ocean and is positively correlated with ambient dFe concentrations in sparse field observations. The Community Earth System Model ocean component was modified to simulate dynamic, group-specific, phytoplankton iron quotas (Fe:C) that vary as a function of ambient iron concentration. The simulated Fe:C ratios match the spatial trends in the observed  Fe:C ratios. Acclimation of phytoplankton Fe:C ratios dampens the biogeochemical response to varying atmospheric deposition of soluble iron, compared to a fixed Fe:C ratio. However, varying atmospheric soluble iron supply still has first-order impacts on global carbon and nitrogen fluxes, and on nutrient limitation spatial patterns. Our results suggest pyrogenic Fe is a significant dFe source that rivals mineral dust inputs in some regions. Changes in dust flux and iron combustion sources (anthropogenic and wildfires) will modify atmospheric Fe inputs in the future. Accounting for dynamic phytoplankton iron quotas is critical for understanding ocean biogeochemistry and projecting its response to future variations in atmospheric deposition.

This dataset contains NetCDF output from the CESM v1.98 model, ocean only, Biogeochemical Elemental Cycling. The files presented are a 20 year average for the final 20 years of a 300 year simulation (Years 281-300). Files are provided for every simulation conducted for the results of the paper. This includes the optimal variable stoichiometry model and  fixed Fe:C versions with values of 3, 7, 10 for the four iron deposition simulations (onlyBC is Pyro Only, noBC is Dust Only, CESM2opt is Pyro + Dust, BCx2 is 2xPyro + Dust). 

Supplemental runs include a limited variable run (with parameterization from Moore et al., 2004) and two fixed runs where Fe:C is set at 5 and 6, all three with the CESM2opt (Pyro + Dust) forcing.

The fec_full_diaz.csv file contains the observational data of phytoplankton Fe:C including the following fields: num (unique location number), UniqueStn (unique station name), CellType3 (aflag (autotrophic flagellate), diatom, diaz(diazotroph), or NA), CellTypeNum (1 = diatom, 2 = aflag, 3 = diaz, NA), FeC (phytoplankton Fe:C ratio, umol/mol), lowSe_FeC (lower bound standard error, umol/mol), highSE_FeC (upper bound standard error, umol/mol), dFe (dissolve iron, nM), lat_n (latitude, in degrees N), lon_E (longitude in degrees E), method (observational method, SXRF, GFAAS, NA) and citation for previously published data.

The observational data has also been read into a .mat file titled obs_full.mat via the script 'create_struct_obs_full_diaz.m'.

The .m script 'read_FeC_annual_anydeg.m' reads in the CESM NetCDF output and saves the data as a .mat file. Minimal processing is done other than converting units as noted. The phytoplankton Fe:C for each group is also calculated, and extraneous data overwritten for the maximum and minimum values. Diazotroph data is also removed where biomass is negligible (less than 0.01 mmol/m3/s).

The .m script 'create_surface_FeC_annual_3deg_diaz.m' extracts data from the previous .mat file at the observational locations described below for surface data only. Position within the CESM grid is determined by optimizing least squares.

The .m script 'create_mat_WOA18_NC.m' reads a .nc file of WOA18 data on the CESM gx3v7 grid and creates a .mat file with no other processing.

The .m script 'read_Fe_txt.m' reads a .txt file of dissolved iron observations from GEOTRACES and supplemental cruises and creates a .mat file, 'master_Fe_gx3v7.mat'.

The .m script 'HNLC_index.m' calculates the HNLC index for the CESM input data, which is defined as the ratio between the model and WOA18 of the fraction of the region between 25n and 25S extending from 150E to the American coast within the Pacific Ocean where nitrate is less than 0.3nM. This script also creates the file 'HNLC_grid.mat' which contains a mask where the value of 1 indicates the HNLC region defined above and 0 elsewhere.

The .m script 'hnlc_stats.m' calculates the total net primary productivity (NPP) and particulate organic carbon export at 100m (POC) within the HNLC region from 'HNLC_grid.mat'. 

The .m script 'statistics_fe_po4_no3.m' calculates CESM performance statistics for dissolved iron, nitrate and phosphate where the iron observations are compiled from  GEOTRACES and supplemental cruises, and nitrate and phosphate are from World Ocean Atlas 2018. The script also calculates CESM performance statistics for the Fe:C data. For all observational comparisons, the Spearman's rank correlation coefficient and RMSE are calculated.

The unprocessed .nc, global read .mat, and surface observation .mat files for each CESM run are included, as well as the Fe:C observations in .csv and .mat files. WOA18 on the gx3v7 grid is included as .nc and .mat files. Dissolved iron observations from GEOTRACES and supplemental cruises are provided in .txt and .mat files.

These .nc files can be opened via Panoply, Python, or any other program that can import NetCDFs. The excel file can be opened using Microsoft Excel, Google Sheets or any other program that can import CSVs. The .m and .mat files can be viewed and executed with MATLAB, or viewed with VSCode or Python using the Scipy or PyTables libraries.