Methods for Minimization and Management of Variability in Long-Term Groundwater Monitoring Results

Dr. Thomas McHugh | GSI Environmental, Inc.


Objectives of the Demonstration

The National Research Council (NRC) has estimated that annual monitoring costs more than $100 million at Department of Defense (DoD) facilities across the country. This cost includes ongoing monitoring of roughly 40,000 groundwater monitoring wells. A primary purpose of this monitoring is to determine the long-term reduction in contaminant concentrations due to natural attenuation or active remediation. However, short-term variability in contaminant concentrations limits the ability to accurately quantify contaminant attenuation rates, increasing monitoring costs and limiting the ability to make appropriate site management decisions.

The objectives of this project were to: (1) validate the use of alternative field sampling procedures for the collection of groundwater samples in order to minimize variability in groundwater monitoring results; (2) develop and validate an improved method to optimize monitoring frequency by evaluating the site-specific short-term variability and long-term attenuation rate; and (3) develop and validate an improved method to identify long-term concentration trends that better account for the potentially confounding effects of short-term variability. All three of these objectives were successfully met.

Back to Top

Technology Description

To validate sample collection procedures that minimize variability in monitoring results, the project provided a direct comparison of the short-term variability associated with commonly used sampling methods for volatile organic compounds (VOCs), including (1) three variations of low-flow purge; (2) SNAP Sampler (passive no purge); and (3) HydraSleeve (active no purge). The field demonstration was conducted at eight monitoring wells at each of the two demonstration sites, in Texas and in California. Each sampling method was used six times with a total of 96 samples per method and a total of 480 groundwater samples for the demonstration from both sites. Four VOCs were consistently detected in the samples from the Texas site, while 10 were consistently detected in the wells from the California site. The resulting 3,262 data points were used to evaluate the effect of sample method on short-term variability in the monitoring results and statistical bias (i.e., difference in concentration between methods). 

ER-201209 Graphic

Sampling methods investigated.

Back to Top

Demonstration Results

The demonstration results indicated that the sample method (except active no purge) has only a modest impact on monitoring variability and concentration, suggesting that sampling methods should be selected based on factors such as cost, ease of implementation, and sample volume requirements rather than concerns regarding data quality. At both sites, low flow standard (purging to parameter stability) and low flow alternative (small volume) showed the lowest variability. The results were consistent between the two sites except for the active no purge (HydraSleeve) method, which was more variable at the California site than the Texas site.

Although low flow alternative (large volume) and passive no purge (SNAP samplers) yielded slightly more variable groundwater monitoring results than low flow standard, this increase in variability would have little impact on the number of events needed to characterize the long-term concentration trend. However, the increased variability with the active no purge method would increase the number of sampling events required to characterize long-term concentration trends in the well. Low flow small volume purge and passive no-purge (SNAP sampler) were the two best sampling methods based on the combined goals of minimizing monitoring cost and minimizing variability in monitoring results.

Although statistically significant differences in concentration were observed between methods, the average bias was small for all methods. This finding is consistent with a number of previous studies on the effect of sample method on contaminant concentration, although some prior studies have suggested a low bias for the active no purge method.

The development and demonstration of improved data analysis methods (Objectives 2 and 3) is covered in detail in Appendix E of the project’s Final Report. 

ER-201209 Graphic 2

Semi-quantitative analysis of sampling methods. Gray dots indicate range of costs for shallow and deep wells.

Back to Top

Implementation Issues

All tested methods are mature technologies, have few end-user concerns, are straightforward to master, and can be easily applied without substantial implementation issues at most sites. Extensive peer-reviewed literature and guidance exists for all methods. Both the no purge sample methods and the alternative (i.e., fixed volume) low flow purge methods were found to be more cost effective than the standard method of low flow purge to parameter stability, with little clear benefits in data quality. The no purge methods result in little to no generation of purge waste and, therefore, may be more strongly favored at sites where management of purge waste is a logistical challenge or is expensive. Sample volume constraints for the no purge methods are the principal implementation concern where certain analyte suites require large water volumes. For those sites, the low flow alternative methods may be more applicable.

Back to Top