QUANTIFYING FACTORS THAT INFLUENCE DEICER EXPORT AND RETENTION IN A MULTI-LANDUSE WATERSHED IN UPSTATE NY

Chloride contamination of streams and groundwater has become a prevalent issue throughout urbanizing areas in the last half century, particularly in northern latitudes where deicing salts are applied to roadways. This study determined how deicer impacted runoff disperses through sub-urban and urban areas on seasonal and multi-year scales. Chloride concentration changes were then modelled under varying pollutant loading scenarios through an integrated catchment model (INCA-Cl).

Six in-stream conductivity/stage/temperature sondes, recording at 15 minute intervals, were installed within a small (~9.6 km²) multi-landuse watershed in Broome County NY and monitored over a 1 year period. Weekly grab samples were taken at each sonde site and analyzed for dissolved cations and anions to help interpret the sensor results. Data from these sensors and local weather stations were used as inputs to the INCA-Cl model. Results from weekly Fall and Spring stream and groundwater grab samples from 2006-2016 were used in conjunction with the model results to interpret long term trends.

Stream response to storm events was found to be dependent on season as well as amount of impervious surface. In contrast to the urban locations, sub-urban sites did not have an initial spike in TDS before dilution during summer and fall runoff events, and had overall smaller TDS spikes from winter and spring de-icer flushing events, as well as slower discharge response times. TDS of stream water within the watershed showed increasing concentrations over the 10 year period that cannot be solely accounted for by an increase in impervious surface, thus suggesting an accumulation of deicers in groundwater as well. These observations are consistent with seasonal cation and anion data which suggests baseflow composition retains elevated de-icer levels year-round in parts of the watershed. INCA-Cl was able to model seasonal discharge and chloride trends within Fuller Hollow Creek under variable loading conditions throughout the study period. However, chloride spikes from individual deicer flushing events could not be replicated with the model. By quantifying and understanding the effects of road salting practices on variable land use areas, better estimates of chloride export and retention can be developed in order to protect salt sensitive environments.