Nitrous oxide, N2O, is the leading cause of stratospheric ozone depletion and one of the most potent greenhouse gases (GHG). Its concentration in the atmosphere has been rapidly increasing since the green revolution in the 1950s and 1960s. Riverine systems have been suggested to be an important source of N2O, although their quantitative contribution has been estimated with poor precision, ranging between 32.2 and 2100 GgN2O - N/yr. Here, we quantify reach scale N2O emissions by integrating a data-driven machine learning model with a physically-based upscaling model. The application of this hybrid modeling approach reveals that small streams (those with widths less than 10 m) are the primary sources of riverine N2O emissions to the atmosphere. They contribute nearly 36 GgN2O - N/yr; almost 50% of the entire N2O emissions from riverine systems (72.8 Gg2O - N/yr), although they account for only 13% of the total riverine surface area worldwide. Large rivers (widths wider than 175 m), such as the main stems of the Amazon River (~ 6 GgN2O - N/yr), the Mississippi River (~ 2 GgN2O - N/yr), the Congo River (~ 1 GgN2O - N/yr) and the Yang Tze River (~ 0.7 GgN2O - N/yr), only contribute 26% of global N2O emissions, which primarily originate from their water column. This study identifies, for the first time, near-global N2O emission and NO3 removal hot spots within watersheds and thus can aid the development of local- to global-scale management and mitigation strategies for riverine systems with respect to N2O emissions. The presented framework can be extended to quantified biogeochemical, besides N2O emissions, processes at the global scale.
Keywords: Global scale; Hybrid model; Nitrous oxide emissions; River networks.
Copyright © 2021 Elsevier B.V. All rights reserved.