Groundwater Transit Times and the Fate of Aquifer Nitrate: Observations from Sampling in Stream Channels and Well Nests in an Agricultural Watershed, North Carolina, USA

Four groundwater sampling approaches were used to study the transit times and fate of agricultural nitrate in a surficial aquifer. Three stream-based measurement approaches were applied in a 2.5 km reach of West Bear Creek: point measurements beneath the streambed, seepage blankets (novel seepage meter design), and reach mass-balance. Groundwater was also sampled from two nearby well nests. Stream-based methods gave similar mean groundwater seepage rates into the stream (0.3-0.6 m/day) over the course of two separate 3-4 day field campaigns even though stream discharge differed by a factor of ~10 between the two campaigns. At low flow, data on the flow-weighted mean nitrate concentration in groundwater discharge, [NO3--]FWM, and nitrate flux from groundwater to the stream suggested lower values with increasing opportunity for channel-streambed influence as the measurement scale increased, e.g., [NO3--]FWM of 654, 561, and 451 muM for points, blankets, and reach mass-balance, respectively. At high flow the trend was reversed, likely because reach mass-balance captured greater inputs from shallow transient high-nitrate flow paths. Streambed point sampling may be better suited to estimating aquifer discharge of nitrate, and reach mass-balance better suited to estimating effective nitrate input into streams (which at high flow may be more than aquifer discharge due to transient flowpaths and at low flow may be less, due to channel retention). Modeling dissolved N2 from point and blanket samples and from samples collected from nearby well nests suggested about 50 -- 78% of groundwater nitrate was denitrified in the aquifer. Streambed sampling suggested both extent of denitrification and initial nitrate concentration in groundwater (700-1300 muM) were related to land use, and that these forms of streambed sampling for groundwater can reveal watershed spatial relations relevant to nitrate fate in the aquifer. Based on age-dating tracer data from streambed point sampling, the mean transit time (MTT) of groundwater (flowweighted mean apparent groundwater age) in the surficial aquifer is about 30 years, similar to the MTT from well data (27 years). Blanket data, after corrections for a fraction of stream water in blanket samples, suggested similar groundwater apparent ages compared to point samples collected adjacent to the blankets. Transit time distributions (TTD) for groundwater discharging to the stream were best fit by a gamma distribution model and were very consistent, despite different hydrologic conditions and sampling arrangements (closelyspaced transects in 58 m reach vs widely spaced transects in 2.5 km reach). This consistency suggests that coupled measurements of groundwater seepage rate and groundwater age may be a robust approach for determining aquifer TTD as well as MTT. An exponential-piston flow model also gave a reasonable fit to the TTD from point measurements, and to the age distributions observed from well nest data, which suggests significant spatial variation in recharge rates in the watershed. This has potentially important implications for future fluxes of NO3-- to the stream, because the TTD for aquifer discharge to the stream is dominated by a large fraction of older groundwater (20-40 years). As a result, NO 3-- fluxes will initially respond slowly to changes in nitrogen use in the watershed before responding more rapidly during a period 20-40 years after the changes are made. Historic patterns of [NO3 --] in groundwater suggest that peak [NO3-- ] may have occurred 10-20 years ago, which means that fluxes of NO 3-- from the aquifer could potentially peak in the next 1-2 decades before returning to current levels as high-nitrate groundwater is flushed from the aquifer. Collectively, the results of this study highlight the importance of directly sampling aquifer discharge to better understand historic and spatial patterns of contamination and the likely groundwater system response to nutrient management in agricultural watersheds.