David W. Ostendorf

Publication Date



The goal of this research was to evaluate the use of an automated sampler in profiling concentrations of deicing agents in stonnmnoff at a research site in Plymouth, MA. The Massachusetts Highway Department applies a combination of deicing agents in the form of salt, premix and calcium magnesium acetate (CMA) to an 830 m long stretch of State Route 25 during the months of November to April when driving conditions are hazardous. Portions of the applied deicing agents applied are left behind in a depression storage layer which was found to be 2.67 mm. The depression storage layer govems the dissolution kinetics of deicing agents into runoff over the year. Precipitation, runoff and specific conductivity data for stonns over a 4 year period (2001-2004) were used in developing a sampling strategy for an ISCO 6712 automated sampler with 24 available samples. The best sampling rule for a yearly sampling strategy was found to be split into two parts. The first with five samples to capture the first flush at time steps of 15 minutes, while the second part of the mle has 19 samples following completion of the first part of the rule which were taken at 65 minute intervals. The error between the flux for the stonn and the flux captured by the sampling rule was 37%.· Dividing the sampling strategy into summer and winter seasons provided better results with more samples needed to profile first flush in the winter [11 samples - time sp~cing (Ll.t) of 20 min] compared to' fewer samples needed for the summer (5 samples - t.t 10 min). The second part of the rule for both summer and winter had time between samples taken of greater than an hour; errors decreased fi'om the yearly sampling. Differences between the flux were-34% for winter sampling and 38% for summer sampling. The results of this rule were used in successive sample collection using the automated sampler. Samples collected by the automated sampler were evaluated for cations and anions studied (HCO]-, Acetate ,cr ,S04-2 ,Na+ ,K+ ,Ca+2 ,Mg+2 ). Amounts of deicing agents were predicted from the observed concentrations of ions showing that maximum concentrations of deicing agents occurred in winter runoff. Sodium and chloride ions dominate the ionic. species in the runoff throughout the year. This is shown in the source strength predicted by the loadograph model (average - 1.26 x 10-7 kg/m2-s) and shown by the graphs of storm events. CMA and premix concentrations persist in significantly less quantities in stolm runoff throughout the year. Models of runoff and cr flux were developed using a one parameter Nash instantaneous unit hydro graph (IUH). Runoff was calibrated for storm events from 2004 using a Fibonacci search to find the mnoff decay constant for the study site. Optimum parameters of 1 linear reservoir and a mnoff decay constant of 7.79 x 10-4 S-I were found.. The chloride loadograph model. was calibrated using a two parameter Fibonacci search for the decay constant and the source term. The runoff decay constant for the loadograph was 9.51 x 10-4 S-I comparable tothat found for the hydrograph. The chloride source strength decreased with time varying between 2.17 x 10'9 in March to 7.42 X 10-7 kg/m2_s in early December with the highest cr source term obselved during deicing agent application. An initial high ion loading rate was observed during the first flush· followed by a slower progressive linear trend for the remaining portion of the event that was sampled. These trends decreased with successive events over the year.