Investigation of Nitrate Transfer in Karst Basins During Storm Events Using Submersible Underwater Nutrient Analyzers

Knowledge of nitrate sources and pathways in karst basins remains incomplete and hinders management of nutrients that cause algae blooms and degrade municipal water supply. However, the increased availability of optical, ultraviolet nitrate sensors allows advancement of nitrate source and transfer for water managers. A concept model is hypothesized for nitrate transfer during hydrologic events and baseflow that considers the multiple porosity of karst. To test the concept model, 15-minute nitrate sensor data is collected with submersible underwater nutrient analyzers over a multi-year period from two locations in a karst basin in the inner bluegrass region of central Kentucky, USA. Data results carried through quality assurance and quality control methods suggests fluctuations in nitrate and provides evidence of karst pathways with varying porosity. The inner bluegrass nitrate sensor results are compared with data from other karst basins in the United States, including the carbonate karst and their aquifers in Kentucky, Arkansas, Virginia and Maryland. Meta-analysis results from hydrographs, chemographs and hysteresis show evidence of nitrate transfers in karst including: a piston effect at the onset of an event as water stored in fractures and conduits is displaced by new water; a quick response during or just after storm event peaks that quickly dilutes nitrate levels in water; a concentrating effect after an event as high nitrate levels stored in soil water dominate transport; and a nitrate recession curve as nitrate stored in rock matrix become a larger contributor to nitrate flux. Mass balance un-mixing simulation was carried out to quantify nitrate sources for the inner bluegrass basin. The amount of water and nitrate load associated with the nitrate sources was quantified for four seasons and revealed the importance of the intermediate flow/fracture network pathway in transporting the majority of nitrate load. The piston flush occurring at the beginning of storm events was also noteworthy generating greater than 10% of the nitrate load. A reservoir model was formulated to represent the nitrate transfer processes for prediction. The reservoir model showed more insight including the impact of seasonality and sinkhole concentration on the distribution of water while at the same time showed the efficacy of the approach. New information from the reservoir modelling included the volume of water and nitrate stored in the karst aquifer, and these estimates will be useful for concerns of algal bloom proliferation at different times of the year