DOUBLE- INJECTION DYE TRACING AND CONCENTRATION SAMPLING: A METHOD FOR DETECTING FLOW VARIATIONS IN KARST BEDROCK

Dye traces are important to understanding groundwater flow, especially in karst bedrock where flow can be highly anisotropic between groundwater inputs (sinkholes) and outputs (springs). Initially, simple dye traces establish point-to-point flow paths by injecting dye into a sinkhole and placing carbon packets at nearby springs to capture the dye. Improvements to the method include the use of automated water samplers to develop dye-breakthrough curves. More recently, we have piloted the use of constant-rate dye injection pumps and automated water samplers to provide enhanced dye breakthrough data which includes dye-dilution discharge at the receiving points.

In the Ward Branch watershed in Springfield, Missouri, quantitative dye traces have provided better insight into groundwater flow in the karst bedrock. Rhodamine WT dye was pumped at a constant rate into a stream sink with a known groundwater connection to two nearby springs to provide base discharge data. Lissamine Flavine FF dye was injected as a slug with 200 gallons of water into a newly forming sinkhole. Two 24-bottle autosamplers, along with carbon packets, were placed at two springs where previous dye traces have confirmed a groundwater connection between the stream sink and other nearby sinkholes. The autosamplers collected one 500mL water sample every hour from before dye injection began to 72 hours after injection. A fluorescent spectrophotometer was used to analyze dye concentrations. Detection of Lissamine Flavine FF with carbon packets indicates the newly forming sinkhole drains to both springs. Analysis of Rhodamine WT dye concentrations shows that although the dye travels from the stream sink to both springs it emerges at different concentrations and therefore indicates addition of a significant amount of groundwater mixing with the primary flow before re-emerging at the second spring. Results also suggest a sizable portion of this increased flow does not surface at the second spring, but appears to continue downstream as underflow beneath the watershed. By analyzing dye at both locations it shows that groundwater in the region has good lateral communication but there is a split in groundwater source between the stream sink and the newly forming sinkhole, a distinction which may have been missed using a less quantitative, single-dye trace.