Nitrogen nutrients in a subtropical river: Temporal variation and analysis at different spatial scales

Herein we provide a framework for evaluating the spatiotemporal variation of nitrate and ammonium and their relationships with environmental and anthropogenic variables at different scales. Bimonthly samples were collected from 28 sites in the Duero River, Mexico, from May to December 2013. The river flow changed and lost connectivity during the dry season because more water was diverted for irrigation. Four sites had the highest nitrate and ammonium values (7.6 and 22.1mgL-1 respectively), which were related to direct waste water discharge. The remaining 24 sites were analysed using a multivariate approach. Using hierarchical cluster analysis, seasonal trends were identified with two groupings during July and five in December. Sites were grouped into those with constant discharge and nitrates affected by springs and those with poor water quality (high concentrations of ammonium). Non-metric multidimensional scaling simultaneously revealed the variation in time and space, organised sites into an environmental ammonium gradient and differentiated between seasons according to nitrate levels. Regression tree analysis established a relationship between nutrients and independent variables. At the landscape level, the agricultural area affected nitrate (75%) and urban area affected ammonium (45%); at the basin level, road density influenced both parameters (10.68km per 25km2).