New estimates of silicate weathering rates and their uncertainties in global rivers

This study estimated the catchment- and global-scale weathering rates of silicate rocks from global rivers using global compilation datasets from the GEMS/Water and HYBAM. These datasets include both time-series of chemical concentrations of major elements and synchronous discharge. Using these datasets, we first examined the sources of uncertainties in catchment and global silicate weathering rates. Then, we proposed future sampling strategies and geochemical analyses to estimate accurate silicate weathering rates in global rivers and to reduce uncertainties in their estimates. For catchment silicate weathering rates, we considered uncertainties due to sampling frequency and variability in river discharge, concentration, and attribution of weathering to different chemical sources. Our results showed that uncertainties in catchment-scale silicate weathering rates were due mostly to the variations in discharge and cation fractions from silicate substrates. To calculate unbiased silicate weathering rates accounting for the variations from discharge and concentrations, we suggest that at least 10 and preferably ∼40 temporal chemical data points with synchronous discharge from each river are necessary. For the global silicate weathering rate, we examined uncertainties from infrequent sampling within an individual river, the extrapolation from limited rivers to a global flux, and the inverse model selections for source differentiation. For this weathering rate, we found that the main uncertainty came from the extrapolation to the global flux and the model configurations of source differentiation methods. This suggests that to reduce the uncertainties in the global silicate weathering rates, coverage of synchronous datasets of river chemistry and discharge to rivers from tectonically active regions and volcanic provinces must be extended, and catchment-specific silicate end-members for those rivers must be characterized. With current available synchronous datasets, we suggest that the global silicate weathering rate (Ca + Mg silicate weathering flux) was ∼2.17 (2 σ range of 1.59–2.75) × 1012 mol/yr, and the global CO2 consumption rates from silicate weathering was ∼7.85 (5.78–9.93) × 1012 mol/yr. Since current synchronous datasets are not available for some rivers in tectonically active regions and highly active volcanic rocks, our estimate should be considered as lower limit of estimates. Including the estimates from volcanic provinces from Dessert et al. (2003), the global CO2 consumption rates from silicates can be estimated as 11.93 (9.86–14.01) × 1012 mol/yr, which is similar to the previous estimates of Gaillardet et al. (1999).