The Murderkill Estuary (Delaware, USA) receives NO3− principally from its upland watershed and from a wastewater treatment facility. Due to disparate
NO3− sources, one‐dimensional salinity‐based mixing models were inadequate for describing distributions of
NO3−, δ15
NNO3−, and δ18
ONO3−. Distance‐based mixing models with multiple, spatially‐specified inputs were, therefore, applied to describe conservative mixing of these constituents and determine the extent to which biogeochemical reactions lead to non‐conservative behavior of
NO3−. These models closely matched Si observations in both winter and summer, consistent with high wastewater silicate loads and light limitation, and serve to validate modeling parameters for both seasons. A close fit of distance‐based models to estuarine
NO3− observations suggested a lack of uptake and fractionation in early winter. In the summer, modeled predictions of
NO3−, δ15
NNO3−, and δ18
ONO3− diverged from estuarine observations, particularly in the oligohaline and polyhaline regions, consistent with in situ nitrogen cycling or additional sources and sinks. Effluent from an adjacent marsh in the lower estuary contained
NO3− with low δ15
NNO3− and δ18ONO3, low DO and high
NH4+ concentrations in late summer. This data and previous studies of adjacent Delaware Bay suggest that reactions in marshes and Bay waters likely drove the non‐conservative behavior of
NO3− and its stable isotopes. Potential uncertainty in watershed discharge, however, limited explicit quantitation of
NO3− loss in the estuary. Nonetheless, distance‐based models are useful tools for the study of
NO3−, δ15
NNO3− and δ18
ONO3− distributions and cycling patterns in complex marsh‐lined estuaries with multiple
NO3− inputs.