Abstract of Meeting Paper

Society for Risk Analysis 2001 Annual Meeting

Using Rolling Averages to Determine Exposure Point Concentrations for Time Series Groundwater Data. H. S. Strauss, H. Strauss Associates, Inc.

Trichloroethylene (TCE) and other chlorinated solvents were released to the groundwater, probably via wastewater disposal in cesspools, from a manufacturing facility in Eastern Massachusetts. The plume was discovered in 1985, and over the next few years was traced down the hill at the rear of the facility, underneath a portion of an elementary school, and into the adjacent wetlands. In 1997, TCE from the plume was detected within a few hundred feet of a municipal water supply well; 4500' from the source.

The extensive groundwater monitoring program at this Site now includes over 220 wells. For several wells, there are 10-15 years of monitoring data with up to 30 samples per well. Methods for both sampling and analysis have been largely the same over the entire time period, thus providing a consistent dataset.

As part of the human health risk, it was important to determine average exposure point concentrations under current and future conditions. This raised the question of whether TCE concentrations were increasing or decreasing at key monitoring wells (no groundwater model was available), and what time average to use (most recent sample, one year average, all data, etc). For individual wells, a plot of concentration vs. sample date was extremely noisy and difficult to interpret. However, use of rolling averages to calculate concentrations (one, two, or three year trailing averages) provided sufficient smoothing of the data to allow observation of trends for individual monitoring wells and predictions of the rate of change. Three year averages provided the smoothest curves, but the averaging period was too long to use for newer wells, despite quarterly sampling data. Use of rolling averages is widely used to evaluate air data, however it appears to be rarely applied to groundwater data. This poster demonstrates the utility of applying rolling averages to groundwater data where individual wells have been sampled over a period of time.


Go to . . .

2001 SRA Annual Meeting Table of Contents
2001 SRA Annual Meeting Author Index
Main Abstracts Menu Page
RiskWorld Home Page