We estimated Prochlorococcus, Synechococcus and picoeukaryotic phytoplankton global cell abundance (cell) using quantitative niche models (Flombaum et al 2013, 2020). Inputs for the niche models were temperature and nitrate from Earth System Models, and light from Ocean Color (oceancolor.gsfc.nasa.gov) which identical across simulations. We estimated Prochlorococcus and Synechococcus using two methodologies, parametric niche model, and artificial neural networks (ANN). For picoeukaryotic phytoplankton we used only the ANN. The parametric model used only temperature and photosynthetic active radiation as inputs, and the ANN also included nitrate concentration. Standard deviation values were estimated for the parametric model from a 10.000 iteration bootstrap on a yearly basis, and for the ANN as 100 iterations for each unique geographic and depth location.
\nData columns containing standard deviation from the mean global cell abundance are labeled after lineage, niche model, scenario, and Earth System Model.
\nUnits: cells.
\nLineages: Prochlorococcus, Synechococcus, and picoeukaryotic phytoplankton.
\nNiche model: Parametric; ANN.
\nScenarios: Historic, RCP2.6, RCP4.5, RCP8.5.
\nEarth System Models: GFDL-ESM2G, HadGEM2-ES, IPSL-CM5A-MR, MPI-ESM-LR, NorESM1-ME (Taylor et al 2012).
Prochlorococcus, Synechococcus and picoeukaryotic phytoplankton yearly standard deviation from mean global abundance estimated for four CMIP5 climate scenarios using an ensemble of five circulation models from 1900 to 2100.
BCO-DMO Processing:
\n- renamed columns to conform with BCO-DMO naming convetions;
\n- replaced NaN with nd (\"no data\");
\n- 2021-03-09: replaced v1 of dataset with v2.