Controls on decadal, annual, and seasonal concentration‐discharge relationships in the Sleepers River Research Watershed, Vermont, northeastern United States

Abstract

Past studies on concentration‐discharge (C‐Q) relationships have focused on short‐term or low‐temporal resolution data. While advancing understanding of catchment processes, these studies provided limited insight on catchment response over time or to climate change. Using 15 solutes from 1992 to 2015 at Sleepers River Research Watershed, Vermont, we compared C‐Q relationships over decades, years, and seasons to elucidate controls on stream chemical variation. We applied end‐member mixing analysis (EMMA) to identify solute sources and flow path routing. EMMA identified three end‐members: near‐surface runoff (NSR), riparian groundwater, and hillslope hollow groundwater. Shifting mixing proportions of these end‐members accounted for the temporal variability of conservative (no chemical reaction en route from source to stream) solutes in streamflow. For example, an increase in NSR fraction, typical of high flow, caused flushing (increased concentrations) of NO3−, DOC, Al, and Fe, which were greatest in NSR, dilution of specific conductance and base cation, SO42−, Si, Sr, Ba, and Mn concentrations, which were greatest in the two groundwater end‐members. This behaviour is reflected in the b‐coefficient of the C‐Q relation (C = aQb), which indicates the strength of dilution (b < −0.1) and flushing (b > 0.1) effects. For conservative solutes, the b‐coefficient decreased significantly (p < 0.01) with an increase in the groundwater to NSR concentration ratio. Solutes that are conservative and have relatively constant concentrations in end‐members over time showed consistent annual C‐Q patterns over years and decades. Furthermore, the strength of dilution or flushing was stronger during the snowmelt period, when the NSR fraction peaked, than during the dormant and growing seasons. With shorter snowmelt periods and snow to rain shifts, the flushing or dilution power of snowmelt runoff will weaken and alter catchment response to climate change. These insights provide more tools for the interpretation of catchment processes and responses to climate change.

Hydrology and biogeochemistry datasets from Sleepers River Research Watershed, Danville, Vermont, USA

“…Kendall et al (1999) used solute chemistry of hillslope groundwater to show that hillslopes are not always well‐connected to the stream, and Hjerdt (2002) extended this work to show that hillslope contributions were focused in convergent hollows. The extensive subsurface chemical database combined with stream event sampling have supported findings that subsurface, chemically evolved ‘old’ water dominates the high‐flow hydrograph (Porter et al, this issue; Kendall et al, 1999; McGlynn et al, 1999; Saiers et al, 2021; Sebestyen et al, 2008; Stewart et al, 2021). Shanley et al (2002) used water isotopes to show that this old water dominated even in the headwater realm, and did not increase with basin scale as hypothesized.…”

Catchment concentration–discharge relationships across temporal scales: A review

“…This divergence across time scales was attributed to the direct link among event precipitation, the associated deposition flux, and storm flows, a relationship that becomes relatively less direct at longer time scales (Knapp et al, 2020). Porter et al (2022) also compared C-Q slopes across multiple time scales, coupled with end-member mixing analysis (EMMA) to quantify solute contributions from hydrological and biogeochemical sources over time.…”

“…1 Long-term (10-30 years) monitoring data (Basu et al, 2010(Basu et al, , 2011Diamond & Cohen, 2018;Godsey et al, 2019;Knapp et al, 2022;Musolff et al, 2015Musolff et al, , 2017Porter et al, 2022;Thompson et al, 2011;Zarnetske et al, 2018) Discrete Hysteresis and flushing indices 1 >1 year of subhourly frequency sensor data (Cain et al, 2022;Kincaid et al, 2020;Lakoba et al, 2021;Speir et al, 2021;Vaughan et al, 2017) Discrete 2 Hourly stormflow samples from 44 events across >10 years (L. A. Rose et al, 2018) Machine-learning classification of hysteresis patterns 1 High-frequency suspended sediment-discharge patterns during events (Hamshaw et al, 2018) Discrete CV C /CV Q ratio 1 >1 year of 5-15 min frequency sensor data (Marinos et al, 2020;Wymore et al, 2021) Discrete 2 Long-term (10-30 years) monitoring data (Basu et al, 2010(Basu et al, , 2011Diamond & Cohen, 2018;Knapp et al, 2022;Musolff et al, 2015Musolff et al, , 2017L.…”

BioRT-HBV 1.0: a Biogeochemical Reactive Transport Model at the Watershed Scale

“…These reactions all follow the simplified rate law 𝑟 = 𝑘𝐴𝑓(𝑇)𝑓(𝑆 𝑤 )𝑓(𝑍 𝑤 ) (Equation 5), without explicitly considering dependence on other solutes. (Porter et al, 2022;Stewart et al, 2022). This highlights the need for denitrification processes in subsurface to capture NO 3 dynamics.…”