Over the last decade it has been recognized that releases of energetic constituents into the environment as a result of military training occur in an extremely heterogeneous pattern. Conventional soil sampling and sample preparation methodologies are inadequate to address the level of contaminant heterogeneity observed. Recently, there have been questions regarding whether the issues observed for the deposition of energetic constituents also substantively apply to other constituents such as metals, semi-volatile organic compounds, and polychlorinated biphenyls.

The objective of this project was to develop a sampling and laboratory analysis method for metals in surface soils on military training ranges that demonstrably produces higher quality data at lower costs than conventional sampling and analysis methods.

To reduce the influence of compositional and distributional error when estimating the mean concentration of an analyte within a decision unit, U.S. Environmental Protection Agency (USEPA) Method 8330B recommends collecting 30 or more evenly spaced increments to build a sample with a total sample mass of more than 1 kg. The objective of this sampling technique is to obtain a representative amount of every particle size, composition (e.g., Pb, As, Cu, Cr, Sb, Cd, and Zn), and configuration (e.g., spheres or elongated particles) and to not over-sample or miss any portion of the decision unit. To estimate the total uncertainty for estimating mean concentrations of munitions constituents (MCs), replicate multi-increment samples must be collected. If this step is not included in a sampling plan, the total characterization error cannot be determined. To obtain representative subsamples from a field sample, the processing protocol must also address the compositional and distribution heterogeneity. The U.S. Geological Survey recommends that the entire field sample be dried, passed through a #10 (2-mm) sieve, then mechanically pulverized to reduce the particle size to less than 0.15 mm. This step is necessary because, within the less than 2-mm soil size class, particles of MCs are present in a variety of sizes, densities, shapes, and compositions. The #10 sieve size encompasses those particles that dissolve more readily and is consistent with the classification of soil and sometimes is used in risk models for human exposure. In this project, multi-increment sampling and adequate sample processing was demonstrated for the characterization of metals in soils that have been introduced as a consequence of military training activities. Comparisons were made with existing protocols to evaluate data quality improvements.

Field and laboratory protocol development for the Incremental Sampling Methodology (ISM) was conducted at an active small arms range at Camp Ethan Allen, Vermont as reported in the Technical Report. In addition, a demonstration was conducted at three additional small arms ranges as reported in the Final Report. The inactive ranges assessed included the 1000-inch Rifle Range at Fort Eustis, Virginia and the Northern Area 3 of the Kimama Training Site (TS), Idaho. Both of these ranges are Military Munitions Response Program sites. A demonstration was also conducted at the active Range 16 Record Range located within the Small Arms Complex at Fort Wainwright, Alaska.

The demonstrations included collection of 63 ISM surface soil samples and 50 conventional grab/discrete samples at Fort Wainwright; 18 ISM and 30 grab samples from Kimama TS; and 27 ISM and 33 grab samples at Fort Eustis. ISM involves changes to the field sampling approach as well as laboratory sample preparation procedures. Each incremental sample was prepared (in the field) by combining a set of multiple increments (of roughly equal soil mass) that were collected over the same Sampling Unit using systematic random sampling.

The performance criteria used to determine whether ISM provided technically defensible data were: (1) reproducible results for surface soil samples containing metal particles, (2) improved performance of ISM as compared with conventional grab sampling techniques, and (3) ease of ISM implementability. Comparisons of ISM with the conventional grab sampling methodology demonstrated, in general, that ISM provided results more representative and reproducible for all three demonstration sites, consistent with the initial study using the results from Camp Ethan Allen.

Distributional heterogeneity was addressed by collecting at least 30-100 increments over the entire decision unit. However, multi-increment field sampling is insufficient in and of itself to overcome the distributional and compositional heterogeneity in the soil samples. Modifications to laboratory sample preparation procedures using USEPA Method 3050B are also necessary to reduce variability owing to sample heterogeneity and a proposed protocol is outlined in the Guidance Document. The proposed changes for metals adopted many of the recommendations for energetics outlined in USEPA Method 8330B such as air drying, milling, larger acid volumes to soil digestion ratios, larger digestion masses, and subsampling to build the digestate sample. Two types of milling equipment (e.g., Ball Mill and Puck Mill) yielded satisfactory results. As the digestion procedures in USEPA Method 3050B resulted in poor antimony and tungsten recoveries, alternative digestion methods were also developed for these metals.

In general, the demonstration results met the targeted performance criteria using ISM. However, there were instances where the performance criteria were not met, e.g., copper. In these situations, the results indicate the extreme contaminant heterogeneity was not adequately dealt with by the ISM approach used. Consequently, in some situations an iterative approach may be necessary, whereby the ISM process is modified to meet the performance objectives, e.g., increasing the milling interval, increasing the number of increments collected, increasing the digestion mass, etc.

In addition to the published Incremental Sampling Methodology (ISM) for Metallic Residues protocol, the investigators are currently working with the USEPA to modify Method 3050B by incorporating the recommended changes identified from this project into a proposed Method 3050C. These changes include modifications to the sample preparation methods as well as the incorporation of an Appendix outlining the multi-increment field sampling approach, similar to what was done for USEPA Method 8330B. In the interim, Technical and Regulatory Guidance: Incremental Sampling Methodology, ISM-1 (February 2012, Interstate Technology and Regulatory Council) is a good reference for understanding and implementing ISM. Protocols specific to ISM sampling of sites with metallic residues can be found in the Guidance Document.

There are no known limitations to the application of ISM, as the equipment used for ISM is the same as that for conventional grab sampling. Implementation costs for ISM are lower than conventional grab sampling because fewer samples are collected, prepared, and analyzed (e.g., less supplies are consumed, less time is required to survey and select sample locations, and there are less samples to label and ship to the laboratory for preparation and analysis). As multiple increments need to be collected to prepare a single sample in the field, the collection time for a single incremental sample is greater than that for a single grab sample. However, as relatively large numbers of grab samples are typically needed to provide data of comparable quality as a few incremental samples, ISM does not typically increase the total sample collection time. Once the Sampling Unit corners are surveyed, there is no need to survey individual sample increment locations. In contrast, each grab sample requires surveying each location. ISM samples typically have a larger mass than conventional grab samples, resulting in greater per sample shipping costs and sample preparation fees. Fewer ISM samples are analyzed than grab samples. The greatest cost saving is incurred at the laboratory preparation step owing to fewer samples requiring preparation and analysis. Again, some additional costs are incurred with the addition of the milling and subsampling step. However, the increased costs are more than offset by the fewer number of ISM samples. Although per-sample (unit) costs are higher for ISM, the total cost of soil sampling and analysis for ISM will generally be less than that for conventional grab sampling.

Cost savings are difficult to quantify because there is no standard procedure for determining the number of soil samples needed to characterize a study area. Conventional grab sampling designs are frequently judgmental in nature, entailing subjective criteria for selecting sampling locations and number of soil samples. However, based on a review of current practices, case studies, and the results of these demonstrations, ISM can result in a cost savings of 30-60% relative to conventional grab sampling methods.