Interacting drivers and their tradeoffs for predicting denitrification potential across a strong urban to rural gradient within heterogeneous landscapes. (15th September 2021)

• Record Type: Journal Article
• Authors: Stephan, Emily; Groffman, Peter; Vidon, Philippe; Stella, John C.; Endreny, Theodore
• Is Part Of: Journal of environmental management. Volume 294(2021)
• Publication Date: 2021-09-15
• Subjects: Denitrification -- Land use management -- Urbanization -- Nutrient pollution; Environmental policy -- Periodicals; Environmental management -- Periodicals; Environment -- Periodicals; Ecology -- Periodicals; 363.705
• DOI: 10.1016/j.jenvman.2021.113021
• Languages: English
• ISSNs: 0301-4797

Abstract

Abstract: Denitrification is a significant regulator of nitrogen pollution in diverse landscapes but is difficult to quantify. We examined relationships between denitrification potential and soil and landscape properties to develop a model that predicts denitrification potential at a landscape level. Denitrification potential, ancillary soil variables, and physical landscape attributes were measured at study sites within urban, suburban, and forested environments in the Gwynns Falls watershed in Baltimore, Maryland in a series of studies between 1998 and 2014. Data from these studies were used to develop a statistical model for denitrification potential using a subset of the samples (N = 188). The remaining measurements (N = 150) were used to validate the model. Soil moisture, soil respiration, and total soil nitrogen were the best predictors of denitrification potential (R 2 adj = 0.35), and the model was validated by regressing observed vs. predicted values. Our results suggest that soil denitrification potential can be modeled successfully using these three parameters, and that this model performs well across a variety of natural and developed land uses. This model provides a framework for predicting nitrogen dynamics in varying land use contexts. We also outline approaches to develop appropriate landscape-scale proxies for the key model inputs, including soil moisture, respiration, and soil nitrogen. Graphical abstract: Image 1 Highlights: Differences in denitrification across land uses are not statistically significant. Soil moisture, respiration, and total nitrogen are best denitrification predictors. Denitrification drivers differ across urban, suburban, and rural landscapes. Proxies for predictors are necessary to model spatial denitrification patterns.