Quantifying Virtual Nitrogen Transfers in the US Food Trade and Mapping Associated Uncertainties

Abstract:

Food production and consumption can increase the loss of reactive nitrogen (Nr; all nitrogen species except N₂) to the environment, which may cause detrimental impacts on both people and ecosystems. This study utilizes embedded resource accounting methods such as nitrogen footprint tool quantifying the release of anthropogenic Nr to the environment to understand trade-offs between environmental, economic, and social outcomes of food systems. To fulfill the gaps within the current nitrogen (N) footprint literature, this research creates US sub-national virtual N networks embedded in three classes of food commodities which are the cereal grain, fruit/vegetable, and meat/seafood products. This work also examines the strength and characteristics of these virtual N networks to understand the trading relationships and to compare properties to other similar networks for embedded resources, such as the virtual water networks. We utilize a Monte Carlo simulation approach to estimate the uncertainty in the footprint values due to sampling variability in the commodity trade. Finally, this work builds and implements a model to forecast virtual N flows by using machine learning approaches. Our preliminary results indicate that most of the US states exhibit a high nitrogen footprint for meat/seafood products, which also indicates high density and larger connectivity properties. However, sampling uncertainty in the N footprint values exhibits larger variability for the cereal grain and fruit/vegetable products.