Biogeochemical river inputs for global ocean models (RivR2O)

1. General Description

The global biogeochemical riverine export dataset (RivR2O) uploaded here is a synthesis product for yearly means of preindustrial C, N and P exports to the ocean and their historical evolutions, which are ready-to-use for global ocean models. They will serve as biogeochemical river inputs in the River-2-Ocean Model Intercomparison Study (R2OMIP). The files cover >10000 global catchments which can be read as lists with coordinates, or as gridded netcdf files (0.25°X0.25°). They cover the compounds DIC, DOC, POC, DIP and DIN. The assumed pre-industrial era is assumed to be pre-1900, whereas historical data will cover 1901-2020. 

Please site the dataset as: 

Lacroix, F., Liu, M., Ma, M., Resplandy, L., Beusen, A., Hauck, J., Lennartz, S., Li, Y., Tian, H., & Regnier, P. (2024). Biogeochemical river inputs for global ocean models (RivR2O) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.13799103

1.1. Preindustrial inputs and their transformations

The files are for preindustrial river inputs can be downloaded as netcdf (r2o_riverinputs_preindustrial.nc), or as catchment lists (DIC,DOC,POC,DIN: riverexports_list_CN.csv , DIP: riverexports_list_P.csv) with given coordinates. They quantify yearly means for every catchment without a significant anthropogenic perturbation. They were constructed in the following ways:

DIC, DOC, POC

Preindustrial DIC, DOC and POC were obtained by subtracting the estimated anthropogenic perturbations for every catchment, which were determined for the 1901-2020 time period by Tian et al. (2023), from the synthesis of present-day exports by Liu et al. (2024). We further accounted for a net DOC source in the tropics (+0.07 Pg C yr-1), and a source in the Southern Hemisphere (+0.01 Pg C yr-1) from estuaries and coastal vegetated ecosystems (including submerged) based on Regnier et al. (2022). Note that in the study, Northern Hemisphere lateral transfers of DOC due to estuaries and coastal vegetation are estimated to approximately zero. A fraction of POC was also removed from the dataset due to models misrepresenting burial on shelf and the remaining fraction (recycled POC) should be added to the semi-refractory DOC pool (see protocol). DIC inputs from groundwater discharge (0.016 Pg C yr-1) were distributed globally homogeneously at every river mouth. Globally, this then amounts to a total of 0.51 Pg C yr-1 of DIC, 0.35 Pg C yr-1 of DOC and 0.095 Pg C yr-1 of POC of available C export to the ocean over the preindustrial time period. 

DIN 

The DIN product averages over three river N exports models (ORCHIDEE-NLAT: Ma et al., in review; DLEM: Yang et al., 2015; Tian, pers. Com., IMAGE-GNM: Beusen et al., 2015, 2016) for every catchment. The resulting preindustrial DIN load to the ocean is 11 Tg N yr-1. In addition,  labile DON is accounted here as DIN (9 Tg N yr-1) based on the ratio C:N of 2583:103 from labile DOC given above (See R2O MIP protocol). This in total amounts to 20 Tg DIN yr-1 inputs to the ocean in the dataset.

DIP

The DIP product averages catchment estimates from IMAGE-GNM (Beusen et al., 2016) and Lacroix et al. (2020). The resulting preindustrial DIP load to the ocean is 2.28 Tg P yr-1. In addition, we account for labile DOP as DIP here (0.19 Tg P yr-1) based on the C:P ratio of 2583:1 (See R2O MIP protocol). This in total amounts to 2.47 Tg DIP yr-1 inputs to the ocean in the dataset.

1.2. Anthropogenic Perturbation (1901-2024)

The river input files from 1901 can be downloaded as a zip file (r2o_river_inputs_1901_2024.zip), which contains a netcdf files for every year of the time series (1901-2024) as rivr2o_riverinputs_{year}.nc. E.g. for 1901 -> rivr2o_riverinputs_{year}.nc 

DIC, DOC, POC

Preindustrial DIC, DOC and POC were obtained by interpolating linearly the estimated anthropogenic perturbations for every catchment, which were determined for the 1901-2024 time period by Tian et al. (2023), to the present-day exports by Liu et al. (2024). Based on Regnier et al. (2022), we assumed no lateral transfers of DOC due to estuaries and coastal vegetation for the present day. The same fraction of POC was also removed from the dataset due to models misrepresenting burial on shelf and the remaining fraction (recycled POC) should be added to the semi-refractory DOC pool (see protocol). DIC inputs from groundwater discharge (0.016 Pg C yr-1) were distributed globally homogeneously at every river mouth. Globally, this then amounts to a total of 0.53 Pg C yr-1 of DIC, 0.30 Pg C yr-1 of DOC and 0.12 Pg C yr-1 of POC of available C export to the ocean over the 2011-2020 period.

DIN 

The DIN product averages over three river N exports models (ORCHIDEE-NLAT: Ma et al., in review; DLEM: Yang et al., 2015; Tian, pers. Com., IMAGE-GNM: Beusen et al., 2015, 2016) for every catchment. The total amounts to 30.03 Tg DIN yr-1 inputs to the ocean in the dataset for the 2011-2020 average (including inputs from labile DON).

DIP

The DIP product averages catchment estimates from IMAGE-GNM (Beusen et al., 2016) and Lacroix et al. (2020).  This in total amounts to 4.92 Tg DIP yr-1 inputs to the ocean in the dataset.

2. Use for modelers within the R2O MIP 

We only briefly describe most important information on how to apply the river input data here and refer to the official R2O MIP protocol for more detail on our general simulation guidelines.

We firstly recommend the addition of a terrestrial dissolved organic carbon pools in the ocean models: tDOC semi-labile (DOC_sl).  Their only source should be that of the terrestrial inputs given here, it should be degraded with a first order constant of k_sl = 1 / 1.5yr (based on Hansell et al., 2012). The other tDOC compound given in the dataset, tDOC labile (tdoc_l), is assumed to be rapidly degraded and should therefore be added to the ocean model DIC pool.

The inputs should be added to the closest ocean model grid points where the ocean model has freshwater inputs. Note that the inputs are given as 10^6 C/N/P per year, and this should be taken into account in the addition of the inputs at the model timestep. We recommend scaling the inputs to the seasonality of the freshwater inputs.

The inputs from the riverine files should be added to the corresponding pool based on the following table:

River Input

(as named in rivr2o_riverinputs_preindustrial.nc)

Global Load (preindustrial)

Global Load 

(2011-2020 Mean)

Ocean Model Pool

DIC ->

0.51 Pg C yr-1

0.53 Pg C yr-1

DIC & Alkalinity (see protocol)

DOC_l ->

0.19 Pg C yr-1

0.21 Pg C yr-1

DIC

DOC_sl ->

0.16 Pg C yr-1

0.09 Pg C yr-1

DOC_sl (new ocean model pool) and associated DON and DOP

POC ->

0.095 Pg C yr-1

0.12 Pg C yr-1

marine DOC and associated nutrients (DON, DOP, see protocol)

DIP ->

2.47 Tg P yr-1

4.92 Tg P yr-1

DIP / Phosphate

DIN ->

20 Tg N yr-1

30.03 Tg N yr-1

DIN / Nitrate

3. References

Beusen, A. H. W., L. P. H. Van Beek, A. F. Bouwman, J. M. Mogollón, and J. J. Middelburg. Coupling Global Models for Hydrology and Nutrient Loading to Simulate Nitrogen and Phosphorus Retention in Surface Water-description of IMAGE–GNM and Analysis of Performance. Geoscientific Model Development, 8, no. 12 (2015): 4045–67. https://doi.org/10.5194/gmd-8-4045-2015. 

Beusen, A. H. W., Bouwman, A. F., Van Beek, L. P. H., Mogollón, J. M., and Middelburg, J. J.: Global riverine N and P transport to ocean increased during the 20th century despite increased retention along the aquatic continuum, Biogeosciences, 13, 2441–2451, https://doi.org/10.5194/bg-13-2441-2016, 2016.

Hansell, D. A., C. A. Carlson, and R. Schlitzer (2012), Net removal of major marine dissolved organic carbon fractions in the subsurface ocean, Global Biogeochem. Cycles, 26, GB1016, doi:10.1029/2011GB004069.

Lacroix, F., Ilyina, T., and Hartmann, J.: Oceanic CO2 outgassing and biological production hotspots induced by pre-industrial river loads of nutrients and carbon in a global modeling approach, Biogeosciences, 17, 55–88, https://doi.org/10.5194/bg-17-55-2020, 2020.

Liu et al. (2024). Global riverine land-to-ocean carbon export constrained by observations and multi-model assessment, Nature Geoscience, https://www.nature.com/articles/s41561-024-01524-z

Ma, M., Zhang, H., Lauerwald, R., Ciais, P., and Regnier, P.: Estimating lateral nitrogen transfer through the global river network using a land surface model, Earth Syst. Dynam. Discuss. [preprint], https://doi.org/10.5194/esd-2024-29, in review, 2024.

Regnier, P., Resplandy, L., Najjar, R.G. et al. The land-to-ocean loops of the global carbon cycle. Nature 603, 401–410 (2022). https://doi.org/10.1038/s41586-021-04339-9

Tian, H., Yao, Y., Li, Y., Shi, H., Pan, S., Najjar, R. G., et al. (2023). Increased terrestrial carbon export and CO2 evasion from global inland waters since the preindustrial era. Global Biogeochemical Cycles, 37, e2023GB007776. https://doi.org/10.1029/2023GB007776

Yang, Qichun, Hanqin Tian, Marjorie A. M. Friedrichs, Charles S. Hopkinson, Chaoqun Lu, and Raymond G. Najjar.: Increased Nitrogen Export from Eastern North America to the Atlantic Ocean Due to Climatic and Anthropogenic Changes during 1901–2008. Biogeosciences,120, no. 6 (2015): 1046–68. https://doi.org/10.1002/2014JG002763. 

4. Version Log

v1 -> pre-industrial river inputs with coastal vegetation and burial transformations

v2 -> Groundwater DIC discharge was added.

v3-> Bugfixes for groundwater discharge and blue carbon inputs.

v4 -> Corrected index with list riverexports_list_CN.csv for DIN inputs

v5 -> corrected tDOC splits according to R2O-MIP protocol

v8 -> Added submerged coastal vegetation fluxes to tDOC_semilabile

v9 -> labile DON and labile DOP are added to the DIP and DON pools (based on C:N:P ratio of 2583:106:1)

v10 -> slight correction in the labile DOM C:N:P ratio (C:N:P = 2583:103:1)

v11 -> correction of labile DOM C:N:P ratio in list files

v12 -> Addition of anthropogenic time series for 1901-2024

v13 -> Corrected unit mistake in historical timeseries for DIN (10^3 magnitude too large)