A global perspective on river alkalinity: drivers and implications for coastal ocean carbonate chemistry

Title: A global perspective on river alkalinity: drivers and implications for coastal ocean carbonate chemistry

Contact: Fei Da (fd6996@princeton.edu or feida6996@gmail.com)

This dataset contains supporting materials for the study analyzing global riverine total alkalinity (TA) and dissolved inorganic carbon (DIC) and their impacts on coastal ocean CO₂ dynamics using multiple linear regression (MLR) and GFDL's MOM6-COBALTv3 model. The files include processed river chemistry data, catchment properties, model grid information, and corresponding model outputs.

Content

• catchment_property_for_regression.mat
  Contains 18 catchment-scale environmental predictors (e.g., carbonate rock extent, land cover, and soil characteristics) for 159 river catchments compiled from global datasets. These catchment properties were used as explanatory variables in the MLR analyses.

• CO2_system_data_for_regression.mat
  Includes processed river chemistry data (from GLORICH and ArcticGRO) for 159 river monitoring stations, representing climatological means of riverine TA, DIC, pH, and pCO₂. These data, together with the catchment properties above, were used to generate figures and establish the statistical relationships presented in the paper (Figures 1–3 and Table 3).

• global_TA_DIC_estimates_from_MLR.mat
  Provides global riverine TA and DIC:TA ratio estimates for 380 major catchments (each with mean discharge > 500 m³ s⁻¹ according to GFDL’s LM3 land model; Figure 4). The estimates are derived from the MLR relationships (Table 3) applied to the corresponding catchment-averaged predictors.
  Note: Values for Antarctica and Greenland are updated based on glacial runoff data from Cantoni et al. (2020), as described in Section 3.2 of the paper.

• Model_grid_info_OM4_quarter_degree.mat
  Contains spatial grid information for the MOM6-COBALTv3 ocean model at ¼° resolution, including longitude, latitude, cell area, and the coastal mask defined in this study for computing coastal air-sea CO₂ fluxes (Figures 5-7).

• Model_IJ_location_near_river_mouth.mat
  Provides the indices and grid cell locations corresponding to major river mouths within the MOM6-COBALTv3 model grid (Figure 7).

• MOM6-COBALTv3_CO2_system.*_2008-2017.nc
  A set of NetCDF files containing model outputs from four river carbon experiments:

  • varRC: River TA values are spatially variable and derived from the MLR analysis; river DIC concentrations are also spatially variable, with DIC:TA ratios derived from MLR relationships (reference run).

  • eqratioRC: River TA values are spatially variable (from MLR), while river DIC concentrations are set equal to TA (i.e., DIC:TA = 1).

  • homogRC: River TA and DIC values are spatially homogeneous across all river inputs.

  • zeroRC: River TA and DIC are both set to zero (control simulation).

Each file represents five-year mean fields for two periods (2008–2012 and 2013–2017), with dimensions x = 1440, y = 1080, z = 2. These outputs were used to analyze and visualize coastal air–sea CO₂ flux differences (Figures 5–7).