Strategies for karst groundwater flow characterization in remote, mountainous, snowmelt-dominated catchments

The Front Ranges of the Canadian Rockies are home to extensive carbonate assemblages that host karst aquifers. New caves and karst springs have recently been discovered in the mountains from Banff, Alberta extending all the way to the United States border, although little research has been conducted on them due to the challenging terrain. In this study, we focus on the Watridge Karst Spring, which is located on a forested hillside in a mountain range that reaches an elevation of 3400 m. This perennial spring can discharge up to 3000 L/s. Karst catchments in these snowmelt-dominated, glacierized areas have sparse vegetation, heavy snowfall, and high hydraulic gradients, leading to efficient groundwater infiltration. As a result, the hydrochemistry of these springs often exhibits strong and rapid fluctuations. The effects of rapid conduit flow are expressed at the Watridge Karst Spring by an increase in discharge followed by a lagged decrease in electrical conductivity (EC), occurring over a diurnal-scale and a longer-scale (e.g., episodic snowmelt or heavy storm). This research aims to use hydrologically relevant metrics to understand the recharge, flow paths, and storage capacity of the aquifer. 
Particularly, we used signal processing of the fluctuations in discharge, EC and air temperature to estimate groundwater response time, defined here as the lag time between a hydrologic event and a resulting change in hydrochemistry. Response time can be used to approximate celerity in the case of discharge, and velocity in the case of EC. Additionally, automatic water sampling allowed for the observation of rapid changes in major ion chemistry.
The results yielded an estimated groundwater conduit velocity on the order of 0.1 m/s that steadily decreases with diminishing flow. It was also found that a distinct shift in the EC signal phase and an associated change in mineral dissolution marks the drainage of an overflow conduit path. This is supported by dye tracer experiments of up to 14 km distance where a maximum velocity of 0.14 m/s has been recorded. Our results show that continuous hydrochemical monitoring of discharge and meteorological conditions at a high-temporal resolution can be used as a first step in characterizing conduit system response. For alpine karst springs with strong hydrochemical fluctuations, this strategy may limit the need to conduct tracer tests involving laborious field work in remote, mountainous locations.