Baltimore Ecosystem Study Institute of Ecosystem Studies

2015 BES Annual Meeting Presentation and Poster Abstracts

Use of water isotope analysis as a screening tool in contributing to understanding groundwater-surface water interactions in urban watersheds
Kemper, John
Co-Authors: John Kemper, Claire Welty, Nicholas Rogers, Lee Blaney, Peter Groffman

Abstract: During extended dry periods we have conducted longitudinal synoptics of base flow in Dead Run to advance our understanding of groundwater dynamics in this complex highly urbanized system. We have conducted three sampling synoptics to date (9-19-14, 5-6-15, and 9-1-15), where we collected water quality samples every 50 m over a 6 km distance in a period of about 5 hours after a 10-day period of no rain. d18O was used as an inexpensive screening tool to evaluate overall system behavior. In general, d18O should increase from a more negative, groundwater-dominated state in stream headwaters to a less negative, more evaporative state downstream. However, in urban systems where there are perturbations from leaking sewage pipes and potable water pipes, the signatures from these sources (typically evaporative) can confound expected patterns. Apparent perturbations in the 9-19-14 synoptic prompted further analysis using ion chromatography for chloride, sulfate, nitrate, and fluoride, as well as fluorescence EEM (excitation-emission matrices) analysis. These analyses revealed reaches of groundwater contamination entering the stream, where elevated fluoride levels were indicative of leaking potable water pipes and EEM data suggested sewage. Elevated and variable chloride levels tracked with sulfate concentrations and appear to be entering the stream system at multiple locations. Spatially, nitrate patterns were highly heterogeneous and hypothesized to be attributable to spatially variable channel incision. However, results from the May spatial synoptic differed dramatically from the 9-19-14 synoptic, highlighting the need for repeated sampling throughout the year at variable base flow conditions. We have also conducted temporal synoptics, where we have sampled stream base flow weekly at six nested Dead Run subwatersheds since March 2013, and all precipitation events since January 2013. Analysis of d18O content of the stream and precipitation time series confirmed expected seasonal and storm-to-storm variability. We evaluated the time series using spectral analysis methods to calculate mean watershed transit times. We found mean transit times to be 1.9 years for Dead Run Franklintown, 1.3 years for DR4, 1.2 years for DR3, and 0.88 years for DR1; these values scaled with watershed area. Future work includes comparison of calculated transit times with predictions from ongoing groundwater modeling efforts.