Factors influencing export of dissolved inorganic nitrogen by major rivers: A new, seasonal, spatially explicit, global model

Abstract

Substantial effort has focused on understanding spatial variation in dissolved inorganic nitrogen (DIN) export to the coastal zone and specific basins have been studied in depth. Much less is known, however, about seasonal patterns and controls of coastal DIN delivery across large spatial scales. Understanding seasonal patterns of DIN export is critical to efforts to predict impacts of coastal eutrophication, such as algal blooms and hypoxic areas, which are often seasonal phenomena. Here we describe, test, and apply a global model that predicts seasonal DIN export to coastal regions for >6000 rivers using the Nutrient Export from Watersheds (NEWS2) model. NEWS2-DIN-S used spatially explicit, seasonal N inputs and was calibrated with measured DIN yield (kg N km À2 season À1) for 77 rivers, distributed globally. Of the characteristics considered, DIN-transport efficiency was positively related to runoff and negatively related to temperature (r 2 = 0.34-0.60, depending on season p < 0.0001), likely due to flushing effects and increased retention by plants and soils, respectively. NEWS2-DIN-S incorporated these insights and performed well in predicting DIN yield (Nash-Sutcliffe Efficiency = 0.54-0.65, depending on season). Catchments were effective in retaining DIN and average export rates were lower during the growing season (3-5% of total nitrogen inputs) compared to other seasons (6-10%) for major latitude bands. Model output was insensitive to changes in the magnitude of N inputs, suggesting that refinement of seasonal N input budgets will not substantially improve model performance. Rather, better representation of land-to-river N transfers could improve future models because of strong landscape N attenuation.

Modeling nitrous oxide emission from rivers: a global assessment

“…Using model Eqns and DIN loads for 6400 global rivers recently calculated by the NEWS2‐DIN‐S model (McCrackin et al ., ), we estimated global riverine N 2 O emissions. The estimated 18.84 Tg N yr −1 of global riverine DIN load yielded a global riverine N 2 O emission rate of 32.2 (95% CI: 12.4–66.9) Gg N 2 O–N yr −1 , which is the average of four modeling approaches (Table ): 29.4 (95% CI: 12.0–58.1) Gg N 2 O–N yr −1 using Eqn , 30.4 (95% CI: 14.2–53.9) Gg N 2 O–N yr −1 using Eqn , 35.3 (95% CI: 10.8–85.0) Gg N 2 O–N yr −1 using Eqns and and 33.8 (95% CI: 12.8–70.7) Gg N 2 O–N yr −1 using Eqns and .…”

“…Increasing river water discharge decreases the water-bed contact area per unit water volume and decreases the efficiency of nitrate diffusion across the river water-sediment interface leading to lower denitrification and/or nitrification per unit N loading (Alexander et al, 2000;Peterson et al, 2001;Mulholland et al, 2008). In addition, increasing river discharge decreases the water residence time, decreasing the opportunity and duration for nitrification and denitrification (Chen et al, 2011;McCrackin et al, 2014). The gas transfer velocity rate that determines the evasion efficiency across the water-atmosphere interface with supersaturated N 2 O decreases with increasing river order due to declining water turbulence with increasing river order (increasing water volume buffering capacity) (Garnier et al, 2009;Raymond et al, 2012;Turner et al, 2015).…”

Reducing bias and quantifying uncertainty in watershed flux estimates: the R package loadflex

“…Quantifying solute and sediment fluxes from watersheds can yield insights into watershed processes, in-stream processes, and nutrient cycles at regional to global scales. Flux estimates are an essential component of whole-watershed manipulation experiments (Likens et al 1970), measurements of nutrient retention within catchments (Groffman et al 2004) and within stream reaches (Bowes and House 2001), calibration of global nutrient export models (McCrackin et al 2014), and assessments of export trends across time and space (Hruška et al 2009, Prokushkin et al 2011. A wide variety of quantification methods have been proposed, including simple period-weighted means, ratio estimators, binning methods, and both linear and nonlinear statistical models (Preston et al 1989, Letcher et al 1999, Asselman 2000, Cox et al 2008, Birgand et al 2010, Raymond and Saiers 2010, Worrall et al 2013).…”

River export of nutrients to the coastal waters of China: the MARINA model to assess sources, effects and solutions

“…Global NEWS-2 has been widely used for the regional analyses of nutrient export by rivers draining into the coastal waters of China (Li et al 2011b;Qu & Kroeze 2010;Qu & Kroeze 2012;Yan et al 2010), Indonesia (Suwarno et al 2014a;Suwarno et al 2014b), the Bay of Bengal Zinia & Kroeze 2015), Europe (Thieu et al 2010), Africa (Yasin et al 2010) and North America (McCrackin et al 2013;McCrackin et al 2014). Global NEWS-2 was calibrated and validated for large rivers worldwide, including large Chinese rivers (the Yellow, Yangtze and Pearl) ).…”

“…The model has been used for global analyses of nutrient export by rivers (Maavara et al 2015;Mayorga et al 2010;Seitzinger et al 2010;Van Cappellen & Maavara 2016). Many studies applied the 2010 version of the Global NEWS-2 model to analyze water pollution in Indonesia (Suwarno et al 2013;Suwarno et al 2014a;Suwarno et al 2014b), the Bay of Bengal Zinia & Kroeze 2015), Europe (Thieu et al 2010), Africa (Yasin et al 2010), North America (McCrackin et al 2013McCrackin et al 2014), the Black Sea ) and in China (Li et al 2011b;Qu & Kroeze 2010;Qu & Kroeze 2012;Yan et al 2010). I discuss these studies in Section 7.3.2.…”