Sediment Monitoring Bias by Automatic Sampler in Comparison with Large Volume Sampling for Parking Lot Runoff

Robert M. Roseen, University of New Hampshire
Thomas P. Ballestero, University of New Hampshire
George D. Fowler, University of New Hampshire
Qizhong Guo, Rutgers University
James J. Houle, University of New Hampshire


A field study was conducted to assess biases of suspended sediment concentration (SSC) analyses (ASTM Standard D3977-97)performed on discrete samples obtained by automatic sampler in comparison with actual sediment concentrations from large volume sam-pling. Research results indicate that the biases attributed to the monitoring of sediment event mean concentration (EMC) and median particle size in parking lot runoff by automated samplers (nonisokinetic) were minimal. Large volume samples (∼15;000 L) of the first-flush event runoff were taken from a storm-water sewer system for eighteen storm events over two years. The intent was to obtain a complete portion of a storm to accurately determine EMCs and particle size distributions (PSDs). Concurrently, flow-weighted discrete samples were obtained by automatic samplers for the same portion of the events. Thus, characteristics of sediments from a whole-storm sample were compared with those of subsamples obtained by an automatic sampler using nonisokinetic sampling. SSCs and PSDs were compared for the two respective field sampling methods. The two methods showed a strong correlation for median sediment EMCs (R2¼0:98,n¼18). Biases to particle size distributions were found to be both for the large particles (>75–150μm) and smaller fines (<25μm). Specific sediment size fractions captured by the large volume sampling and automatic sampler were not significantly different (α¼0:05) forD50, which = 58 and 46μm,respectively.