Abstract:Estimates of global riverine nitrous oxide (N2 O) emissions contain great uncertainty. We conducted a meta-analysis incorporating 169 observations from published literature to estimate global riverine N2 O emission rates and emission factors. Riverine N2 O flux was significantly correlated with NH4 , NO3 and DIN (NH4 + NO3 ) concentrations, loads and yields. The emission factors EF(a) (i.e., the ratio of N2 O emission rate and DIN load) and EF(b) (i.e., the ratio of N2 O and DIN concentrations) values were comparable and showed negative correlations with nitrogen concentration, load and yield and water discharge, but positive correlations with the dissolved organic carbon : DIN ratio. After individually evaluating 82 potential regression models based on EF(a) or EF(b) for global, temperate zone and subtropical zone datasets, a power function of DIN yield multiplied by watershed area was determined to provide the best fit between modeled and observed riverine N2 O emission rates (EF(a): R2 = 0.92 for both global and climatic zone models, n = 70; EF(b): R2 = 0.91 for global model and R2 = 0.90 for climatic zone models, n = 70). Using recent estimates of DIN loads for 6400 rivers, models estimated global riverine N2 O emission rates of 29.6-35.3 (mean = 32.2) Gg N2 O-N yr-1 and emission factors of 0.16-0.19% (mean = 0.17%). Global riverine N2 O emission rates are forecasted to increase by 35%, 25%, 18% and 3% in 2050 compared to the 2000s under the Millennium Ecosystem Assessment's Global Orchestration, Order from Strength, Technogarden, and Adapting Mosaic scenarios, respectively. Previous studies may overestimate global riverine N2 O emission rates (300-2100 Gg N2 O-N yr-1 ) because they ignore declining emission factor values with increasing nitrogen levels and channel size, as well as neglect differences in emission factors corresponding to different nitrogen forms. Riverine N2 O emission estimates will be further enhanced through refining emission factor estimates, extending measurements longitudinally along entire river networks and improving estimates of global riverine nitrogen loads.

A scalable hybrid model to predict riverine nitrous oxide emissions from the reach to the global scale

Modeling Riverine N2O Sources, Fates, and Emission Factors in a Typical River Network of Eastern China.

Modeling nitrous oxide emission from rivers: a global assessment

Indirect nitrous oxide emission factors of fluvial networks can be predicted by dissolved organic carbon and nitrate from local to global scales

Modeling and forecasting riverine dissolved inorganic nitrogen export using anthropogenic nitrogen inputs, hydroclimate, and land-use change

[Estimation of Nitrous Oxide Emission from River System Based on Water Discharge and Dissolved Nitrous Oxide Concentration].

Modeling Riverine N 2 O Sources, Fates, and Emission Factors in a Typical River Network of Eastern China

Estimating lateral nitrogen transfer through the global river network using a land surface model

Predicting Dynamic Riverine Nitrogen Export in Unmonitored Watersheds: Leveraging Insights of AI from Data-Rich Regions

Developing a hybrid data-driven and informed model for prediction and mitigation of agricultural nitrous oxide flux hotspots

Increased global nitrous oxide emissions from streams and rivers in the Anthropocene

Modeling indirect N2O emissions along the N cascade from cropland soils to rivers

Machine learning for modeling N 2 O emissions from wastewater treatment plants: Aligning model performance, complexity, and interpretability

Hydrologic Connectivity Regulates Riverine N 2 O Sources and Dynamics

Modeling greenhouse gas emissions from riverine systems: A review

A hybrid deep learning approach to predict hourly riverine nitrate concentrations using routine monitored data

Modeling N2O emissions with remotely sensed variables using machine learning

Evaluating the suitability of large-scale datasets to estimate nitrogen loads and yields across different spatial scales

A deep learning-based model for estimating pollution fluxes from rivers into the sea and its optimization

Application of a hybrid mechanistic/machine learning model for prediction of nitrous oxide (N2O) production in a nitrifying sequencing batch reactor

Spatiotemporal Patterns of Methane and Nitrous Oxide Emissions in China's Inland Waters Identified by Machine Learning Technique