Research

Motivation: The presence of a stream on the state 303(d) list of impaired streams for elevated nutrient concentrations is often based on limited sampling protocols. This leaves uncertainty as to the scope of the nutrient pollution problem and the best management practices (BMPs) to improve upon it. Additionally, there is a lack of studies investigated seasonal trends of nitrogen dynamics in watersheds relevant to the southeastern region of the U.S. This study attempts to characterize catchment-scale nutrient dynamics in a heterogeneous urban watershed using isotope source tracking techniques in tandem with water quality and quantity modeling.

Methods: This study employs stable isotope analysis of nitrate and sulfate in tandem with a hydrologic monitoring plan. Three hydrologic monitoring sites and 21 baseflow sampling sites were established throughout a 1,664 acre urban catchment with headwaters in a recreational forest preserve. Baseflow and storm flow samples were collected from four seasons and analyzed for nitrate concentration as well as N, O, and S isotope compositions of nitrate and sulfate. 

Outcome: The nitrate concentration observed in the study catchment does not appear to be anthropogenic in origin. Nitrate concentrations during stormflow were on average ___ % lower than during baseflow, indicating a dilution rather than a contribution of nitrate during storm event. The seasonal response of the distribution of nitrate concentration reflected the land use characteristics of the watershed area. Isotope analysis of dissolved nitrate in the stream during both stormflow and baseflow indicated that the primary sources of nitrate to the stream were soil nitrogen processes. These results indicate that ecological relevance of dissolved nutrient presence in urban streams must be considered when designing site-specific mitigation practices. Future work focusing on a framwork for characterizing watersheds, their likely nutreint dynamics, and most effective BMPs based on spatial orientation of patchwork land use and an approachable sampling plan would be a valuable addition to our nutrient management strategy.

Motivation: Urban stormwater runoff drives transport of pollutants to receiving streams (Muller et al., 2019). Re-defined drainage conveyances and patchy land covers due to impervious surfaces obfuscates direct pathways of pollutants. To mitigate risk to urban ecology and downstream bodies, it is critical to accurately assess the timing and volumetric proportion of runoff from different surfaces within the urban watershed. This will allow for improved water quality modeling sensitive to site-specific nuance. Advancements in the field of microbiology have illuminated microbial source tracking using ribosomal DNA as a promising way to approximate source contributions by the mixing of their microbial communities (Knight et al., 2009; McCarthy et al., 2016). These techniques have been used in multiple contexts, but never to partition hydrographs by surface runoff source in urban watersheds. It is known that surface water microbial communities are shaped by the environment they encounter, but little is known as to the unique fingerprints of urban land cover microbial communities (McLellan et al., 2015). 

Methods: This study uses a Bayesian mixing model analysis to approximate microbial community contributions of grass runoff, road runoff, urban baseflow, and forest baseflow to a receiving urban stream with forested headwaters. Microbial communities are quantified using 16S rRNA metagenomic sequencing. Thes study provides evidence that MST may be an effective way to approximate the influence of runoff from different landscapes to the receiving stream’s microbial community. 

Outcome: The MST results were consistent with hydrologic theory, showing 118% increase in estimated contribution of runoff from grassy landcover between storm events of 0.52 and 1.9 in of rainfall. MST results predicted 9.34% of the water column microbial community during storm events was contributed to by the forested land cover microbial signature, while curve number method predicts ~5% of total runoff comes from forested landcover. More research is needed to develop sampling methodology that results in distinct microbial communities between land covers, because road runoff and grass runoff community diversity largely overlaps. Still, the results show there is promise in the technique to be sensitive to hydrologic timing of runoff arrival especially if a more sophisticated hydrologic model is applied.