10.48364/ISIMIP.920945 Xiao Liu Xiao Liu 0000-0002-4019-196X Charles Stock Charles Stock 0000-0001-9549-8013 John Dunne John Dunne 0000-0002-8794-0489 Minjin Lee Minjin Lee 0000-0002-8831-9387 Elena Shevliakova Elena Shevliakova 0000-0003-4910-2166 Sergey Malyshev Sergey Malyshev 0000-0001-6259-1043 Paul C.D. Milly Paul C.D. Milly 0000-0003-4389-3139 Matthias Büchner Matthias Büchner 0000-0002-1382-7424 ISIMIP Repository 2022 EARTH SCIENCE > OCEANS Julia Blanchard Julia Blanchard 0000-0003-0532-4824 Colleen Petrik Colleen Petrik 0000-0003-3253-0455 Cheryl S. Harrison Cheryl S. Harrison 0000-0003-4544-947X Matthias Büchner Matthias Büchner 0000-0002-1382-7424 Jochen Klar Jochen Klar 0000-0002-5883-4273 Iliusi Vega del Valle Iliusi Vega del Valle 0000-0001-6902-2257 Jan Volkholz Jan Volkholz 0000-0002-2533-3739 Jacob Schewe Jacob Schewe 0000-0001-9455-4159 Franziska Piontek Franziska Piontek 0000-0003-4305-7552 Christopher Reyer Christopher Reyer 0000-0003-1067-1492 Matthias Mengel Matthias Mengel 0000-0001-6724-9685 Christian Otto Christian Otto 0000-0001-5500-6774 Martin Park Martin Park 0000-0002-2467-3256 Katja Frieler Katja Frieler 0000-0003-4869-3013 Potsdam Institute for Climate Impact Research (PIK) 03e8s1d88 2022-11-17 en Input dataset https://protocol.isimip.org https://www.isimip.org https://doi.org/10.1029/2019ms001726 https://doi.org/10.1038/s41467-019-09468-4 https://doi.org/10.1029/2021GL094367 https://doi.org/10.1016/j.ocemod.2018.07.002 https://doi.org/10.1029/2019MS002043 1.0 Creative Commons Attribution 4.0 International (CC BY 4.0) The GFDL-MOM6-COBALT2 dataset is based on a 50-year ocean simulation using an enhanced resolution, global ocean physical-biogeochemical model incorporated with temporally changing river freshwater and nutrient fluxes (Liu et al., 2021). The ocean physical model used in this study was configured from the 6th generation of the Modular Ocean Model (MOM6) and its accompanying 2nd generation Sea Ice Simulator (SIS2) both developed at the NOAA Geophysical Fluid Dynamics Laboratory [Adcroft et al., 2019]. The model grid used in our simulations has a nominal horizontal resolution of 0.25° or approximately 25 km. MOM6 is integrated with the GFDL’s Carbon, Ocean Biogeochemistry and Lower Trophics (COBALTv2) ocean biogeochemical model [Stock et al., 2020]. The 33-tracer COBALT model simulates global-scale dynamics of carbon, nitrogen, phosphorus, iron, and oxygen, along with three explicit phytoplankton (small, large, and N2-fixing) groups and three explicit zooplankton groups by size. The temporally changing river freshwater and nutrient fluxes used in MOM6-COBALTv2 are derived from long-term simulation of GFDL’s land-watershed model LM3-TAN (Land Model version 3 with Terrestrial and Aquatic Nitrogen) [Lee et al., 2019]. The coupled MOM6-COBALTv2 ocean model was initialized at rest using long-term climatologies of observed temperature from the World Ocean Atlas (WOA 2013v2), and preindustrial dissolved inorganic carbon and total alkalinity from the GLobal Ocean Data Analysis Project (GLODAPv2). The simulation was spun up for a total of 208 years with eight repeating cycles of 26 years between 1959 and 1984 forced by the Japanese 55-year Reanalysis (JRA-55) products [Tsujino et al., 2018], from which two retrospective experiments (dynRN and climRN, see below) were initialized and ran for 52 years from 1959 to 2010. During the last spin-up cycle, climatological values of river freshwater runoff and nutrient concentrations averaged over the period just prior to initialization (1951-1958) were used. The dynRN experiment was forced with both dynamic atmospheric states (6-hourly JRA-55 reanalysis) and dynamic (monthly changing) river freshwater and nutrient fluxes from LM3-TAN (described in section 2.2), while climRN was forced with the same dynamic atmospheric states but river fluxes were held at the 1951-1958 climatological values consistent with the spin-up simulation. More details for the GFDL-MOM6-COBALT2 dataset can be found in Liu et al. [2021].<br>The remapping to a 0.25° and 1° spatial resolution was performed using CDO's bilinear remapping function. The vector-based zonal and meridional water velocities were previously rotated from the irregular tripolar grid into a direct relationship with the true Earth poles. Vertical resolution is preserved. <br>Data was converted where needed to match the unit definitions of the CMIP6 simulations available from ESGF in ISIMIP3b (FishMIP Phase 1 2020). This means in detail: (1) apply the N-C and atomic mass conversion ratio, (2) apply standard salt water density of 1035 kg m-3 when converting from mass to volumetric units, e.g. µg kg-1 -> kg m-3, and (3) convert H+ ion concentration to ph via ph=-1*log10(htotal).<br>For the use within ISIMIP, dynRN and climRN were translated to "obsclim" and "ctrlclim", respectively. global -179.5 179.5 -89.5 89.5