INTRODUCTION
Data documenting the temporal variation of solute concentrations in streams have advanced the knowledge about how catchments produce stream flow and impact streamwater quality. Lab analysis of discrete water samples has underlain this knowledge such as how piston displacement of stored water is a dominant stream flow generation mechanism in many catchments (Sklash and Farvolden 1979), how mineral weathering and soil development are impacted by human activities (Bailey 2020), and how hydrology and soil geochemistry interact to regulate streamwater solutes (e.g., Bennetin et al., 2015), just to name a few examples. However, weekly or less frequent sampling can inadequately capture temporal patterns in the variation of streamwater chemistry (Kirchner 2004). Advancements in technology have allowed us to supplement the information gained from water sampling with high-frequency data from electronic sensors (e.g., Pellerin et al. 2010; von Freyberg et al. 2017).
Stream solute data allow estimation of the fractional contribution of new precipitation versus older, stored water to streamflow (Hooper and Shoemaker 1986; Pellerin et al. 2008). This chemical hydrograph separation can provide insights into how catchments transport rainfall and how catchment characteristics or human modifications alter this hydrologic transport (Pellerin et al. 2008; Inserillo et al. 2107). Specific electrical conductivity (SC) is a common solute indicator used for ”new water” estimates (Sklash and Farvolden 1979; Pellerin et al. 2008). Sensors measuring SC are robust, relatively affordable, and have shown that stream chemistry is highly variable on short time scales (Walling 1975; Vogt et al. 2010). Analysis of temporal SC patterns remains an opportunity for advancing catchment hydrology and biogeochemistry. Improved methods for analysis of these high-frequency time series can help expand our understanding of water movement and solute transport through catchments.
During and immediately following storm events, the SC of streams exhibits patterns that provide insights into catchment hydrology and biogechemistry. During a storm, streamwater SC is diluted by rainwater with lower SC than the groundwater that is sourcing streamwater prior to the event (Sklash and Farvolden 1979). Following this, barring further precipitation events, the rainwater continues to drain from the catchment, is evapotranspired, or is integrated into stores, causing SC to recover towards its pre-storm conditions (Walling 1975). Within certain catchments, a solute flush occurs before dilution (Inamdar et al. 2009; Inserillo et al., 2017). Specific conductivity drastically increases on the rising limb of some streams’ hydrographs, indicating a flush of ions entering the stream. This flush may represent a distinct source of water entering the stream beyond that of mean stored groundwater and rainwater (Walling 1975, Creed and Band 1998, Inserillo et al. 2017). Pairing meteorologic conditions with sensed, high-frequency time series of stream solutes will aid understanding how precipitation, antecedent hydrologic conditions, and seasonal changes influence the variability of those solutes (e.g., Robson 1992, Biron et al. 1999, Kirchner 2004, Fovet et al. 2018).
The goal of this study was to analyze the temporal behavior of streamwater SC in response to a variety of different precipitation events. To this end, high frequency data from an in situ sensor deployment were used to analyze the magnitude of solute flushing (FSC), the magnitude of SC dilution (DSC), and the rate at which SC recovers after a dilution occurs (RSC). We developed an algorithm to objectively analyze these chemograph patterns and produce functional relationships between storm characteristics and stream SC behavior.