A mechanistic model for gas transfer velocity in the Damodar river.

Abstract

Gas transfer across river interfaces remains a poorly quantified process in the Earth's carbon cycle. Standard frameworks overlook how turbulence, biofilm impedance, and suspended sediment interact. We propose a mechanistic k model integrating (i) turbulence-driven transfer via the O'Connor & Dobbins formulation, (ii) biofilm suppression as a power-law function of thickness (δ), (iii) sediment suppression as an exponential function of suspended sediment concentration (C_sed), and (iv) diurnal inhibition of k(Δk_photo). We apply this model to the Damodar River, India-a regulated, eutrophic tropical system. Field measurements (SF₆ tracer injections, floating chambers, micro-profiling, ADCP surveys) reveal that biofilms suppress k by 18-37% in dry-season low-flow reaches (δ = 1.2-3.0 mm), while suspended sediments dominate monsoon fluxes, reducing k by 22-29% at estimated C_sed of 300-850 mg L^-1. The model captures observed spatial and seasonal patterns with RMSE = 1.82 cm h^-1 and NSE = 0.64, outperforming conventional formulations (O'Connor and Dobbins, NSE = 0.42; Raymond et al., NSE = 0.45). Because sediment concentrations were estimated from published post-dam baselines, we report results as a sensitivity analysis across plausible C_sed ranges. These findings highlight the need to incorporate biofilm-sediment interactions into gas flux models, particularly in anthropogenically altered rivers.