Difference between revisions of "User:Debra Tabron/sandbox"

From Enviro Wiki
Jump to: navigation, search
(Matrix Diffusion Modeling Options)
 
(75 intermediate revisions by the same user not shown)
Line 1: Line 1:
REMChlor-MD is a free toolkit available for download and is capable of simulating matrix diffusion in groundwater contaminant plumes. It is a significant upgrade from REMChlor 1.0<ref>Falta, R.W., 2008. Methodology for comparing source and plume remediation alternatives. Groundwater, 46(2), pp.272-285. [https://doi.org/10.1111/j.1745-6584.2007.00416.x doi: 10.1111/j.1745-6584.2007.00416.x]</ref>, which does not include matrix diffusion in the plume.  REMChlor-MD is useful for planning-level approximations of contamination extent and duration at sites where matrix diffusion is important. REMChlor-MD employes a semi-analytical method for simulating mass transfer between high and low permeability zones that provides computationally accurate predictions of concentration distributions and mass discharge in the higher permeability portions of an aquifer.  Model run times for REMChlor-MD are much shorter than traditional fine grid numerical models.  
+
The heterogeneous distribution of munitions constituents, released as particles from munitions firing and detonations on military training ranges, presents challenges for representative soil sample collection and for defensible decision making. Military range characterization studies and the development of the incremental sampling methodology (ISM) have enabled the development of recommended methods for soil sampling that produce representative and reproducible concentration data for munitions constituents. This article provides a broad overview of recommended soil sampling and processing practices for analysis of munitions constituents on military ranges.  
 
 
 
<div style="float:right;margin:0 0 2em 2em;">__TOC__</div>
 
<div style="float:right;margin:0 0 2em 2em;">__TOC__</div>
  
 
'''Related Article(s)''':  
 
'''Related Article(s)''':  
*[[Source Zone Modeling]]
 
*[[Plume Response Modeling]]
 
*[[Matrix Diffusion]]
 
  
'''CONTRIBUTOR(S):''' [[Dr. Ron Falta]] and [[Kien Pham]]
+
 
 +
'''CONTRIBUTOR(S):''' [[Dr. Samuel Beal]]
  
  
 
'''Key Resource(s)''':  
 
'''Key Resource(s)''':  
*[https://www.serdp-estcp.org/Program-Areas/Environmental-Restoration/Contaminated-Groundwater/Persistent-Contamination/ER-201426 REMChlor-MD Toolkit package]
+
*[[media:Taylor-2011 ERDC-CRREL TR-11-15.pdf| Guidance for Soil Sampling of Energetics and Metals]]<ref name= "Taylor2011">Taylor, S., Jenkins, T.F., Bigl, S., Hewitt, A.D., Walsh, M.E. and Walsh, M.R., 2011. Guidance for Soil Sampling for Energetics and Metals (No. ERDC/CRREL-TR-11-15). [[media:Taylor-2011 ERDC-CRREL TR-11-15.pdf| Report.pdf]]</ref>
 +
*[[Media:Hewitt-2009 ERDC-CRREL TR-09-6.pdf| Report.pdf | Validation of Sampling Protocol and the Promulgation of Method Modifications for the Characterization of Energetic Residues on Military Testing and Training Ranges]]<ref name= "Hewitt2009">Hewitt, A.D., Jenkins, T.F., Walsh, M.E., Bigl, S.R. and Brochu, S., 2009. Validation of sampling protocol and the promulgation of method modifications for the characterization of energetic residues on military testing and training ranges (No. ERDC/CRREL-TR-09-6). Engineer Research and Development Center / Cold Regions Research and Engineering Lab (ERDC/CRREL) TR-09-6, Hanover, NH, USA. [[Media:Hewitt-2009 ERDC-CRREL TR-09-6.pdf | Report.pdf]]</ref>
 +
*[[media:Epa-2006-method-8330b.pdf| U.S. EPA SW-846 Method 8330B: Nitroaromatics, Nitramines, and Nitrate Esters by High Performance Liquid Chromatography (HPLC)]]<ref name= "USEPA2006M">U.S. Environmental Protection Agency (USEPA), 2006. Method 8330B (SW-846): Nitroaromatics, Nitramines, and Nitrate Esters by High Performance Liquid Chromatography (HPLC), Rev. 2. Washington, D.C. [[media:Epa-2006-method-8330b.pdf | Report.pdf]]</ref>
 +
*[[media:Epa-2007-method-8095.pdf | U.S. EPA SW-846 Method 8095: Explosives by Gas Chromatography.]]<ref name= "USEPA2007M">U.S. Environmental Protection Agency (US EPA), 2007. Method 8095 (SW-846): Explosives by Gas Chromatography. Washington, D.C. [[media:Epa-2007-method-8095.pdf| Report.pdf]]</ref>
  
 +
==Introduction==
 +
[[File:Beal1w2 Fig1.png|thumb|200 px|left|Figure 1: Downrange distance of visible propellant plume on snow from the firing of different munitions. Note deposition behind firing line for the 84-mm rocket. Data from: Walsh et al.<ref>Walsh, M.R., Walsh, M.E., Ampleman, G., Thiboutot, S., Brochu, S. and Jenkins, T.F., 2012. Munitions propellants residue deposition rates on military training ranges. Propellants, Explosives, Pyrotechnics, 37(4), pp.393-406. [http://dx.doi.org/10.1002/prep.201100105 doi: 10.1002/prep.201100105]</ref><ref>Walsh, M.R., Walsh, M.E., Hewitt, A.D., Collins, C.M., Bigl, S.R., Gagnon, K., Ampleman, G., Thiboutot, S., Poulin, I. and Brochu, S., 2010. Characterization and Fate of Gun and Rocket Propellant Residues on Testing and Training Ranges: Interim Report 2. (ERDC/CRREL TR-10-13.  Also: ESTCP Project ER-1481)  [[media:Walsh-2010 ERDC-CRREL TR-11-15 ESTCP ER-1481.pdf| Report]]</ref>]]
 +
[[File:Beal1w2 Fig2.png|thumb|left|200 px|Figure 2: A low-order detonation mortar round (top) with surrounding discrete soil samples produced concentrations spanning six orders of magnitude within a 10m by 10m area (bottom). (Photo and data: A.D. Hewitt)]]
  
==Matrix Diffusion of Dissolved Contaminants==
+
Munitions constituents are released on military testing and training ranges through several common mechanisms. Some are locally dispersed as solid particles from incomplete combustion during firing and detonation. Also, small residual particles containing propellant compounds (e.g., [[Wikipedia: Nitroglycerin | nitroglycerin [NG]]] and [[Wikipedia: 2,4-Dinitrotoluene | 2,4-dinitrotoluene [2,4-DNT]]]) are distributed in front of and surrounding target practice firing lines (Figure 1). At impact areas and demolition areas, high order detonations typically yield very small amounts (<1 to 10 mg/round) of residual high explosive compounds (e.g., [[Wikipedia: TNT | TNT ]], [[Wikipedia: RDX | RDX ]] and [[Wikipedia: HMX | HMX ]]) that are distributed up to and sometimes greater than) 24 m from the site of detonation<ref name= "Walsh2017">Walsh, M.R., Temple, T., Bigl, M.F., Tshabalala, S.F., Mai, N. and Ladyman, M., 2017. Investigation of Energetic Particle Distribution from High‐Order Detonations of Munitions. Propellants, Explosives, Pyrotechnics, 42(8), pp.932-941. [https://doi.org/10.1002/prep.201700089 doi: 10.1002/prep.201700089] [[media: Walsh-2017-High-Order-Detonation-Residues-Particle-Distribution-PEP.pdf| Report.pdf]]</ref>.
Dissolved contaminants, such as chlorinated volatile organic compounds (CVOCs), are found in groundwater at many current and former industrial sites globally. Plumes of CVOCs dissolved in groundwater often originate when the chemicals enter the subsurface as dense non-aqueous phase liquids (DNAPL). Being heavier than water and immiscible, DNAPLs are particularly problematic because they can spread vertically into the aquifer, generating extensive contaminated groundwater plumes. In Figure 1, CVOCs in DNAPL form are shown in bold red, and the resulting dissolved groundwater plume is represented by the pink halo emanating from the DNAPLs. One of the most significant lessons learned about chlorinated solvents is that their groundwater plumes can persist even when the source DNAPLs have been depleted. Plume persistence in the absence of source materials has been attributed to a phenomenon called matrix diffusion.   Matrix diffusion can affect any type of dissolved contaminant, particularly when dissolved concentrations are high relative to regulatory standards. In addition to CVOCs, matrix diffusion can occur with other dissolved contaminants including radionuclides, per- and polyfluoroalkyl substances (PFAS), dissolved petroleum hydrocarbons such as benzene, toluene, ethylbenzene, and xylene (BTEX), and with solvent stabilizers such as 1,4-dioxane.
 
[[File:Falta1w2 Fig1.png|thumb|Figure 1. Spread of CVOCs in the Subsurface]]
 
  
The life cycle of a matrix diffusion site is separated into two periods. In the loading period, the low-permeability (low K) zones act as storage areas for the contaminant mass. Following clean-up actions in the transmissive (high K) zones, the concentration gradient is reversed. The unloading period begins when the contaminant mass previously stored in the low-permeability zones diffuses back out into the transmissive zones. In Figure 1, the blue arrows represent the forward and backward diffusion gradients into the surrounding materials. Because the pace of back diffusion can be significantly slower than the loading of the contaminant, the plume resulting from the back diffusing mass can last for decades to centuries<ref>Chapman, S.W. and Parker, B.L., 2005. Plume persistence due to aquitard back diffusion following dense nonaqueous phase liquid source removal or isolation. Water Resources Research, 41(12). [https://doi.org/10.1029/2005WR004224 doi: 10.1029/2005WR004224]</ref><ref>Parker, B.L., Chapman, S.W. and Guilbeault, M.A., 2008. Plume persistence caused by back diffusion from thin clay layers in a sand aquifer following TCE source-zone hydraulic isolation. Journal of Contaminant Hydrology, 102(1-2), pp.86-104. [https://doi.org/10.1016/j.jconhyd.2008.07.003 doi: 10.1016/j.jconhyd.2008.07.003]</ref>.
+
Low-order detonations and duds are thought to be the primary source of munitions constituents on ranges<ref>Hewitt, A.D., Jenkins, T.F., Walsh, M.E., Walsh, M.R. and Taylor, S., 2005. RDX and TNT residues from live-fire and blow-in-place detonations. Chemosphere, 61(6), pp.888-894. [https://doi.org/10.1016/j.chemosphere.2005.04.058 doi: 10.1016/j.chemosphere.2005.04.058]</ref><ref>Walsh, M.R., Walsh, M.E., Poulin, I., Taylor, S. and Douglas, T.A., 2011. Energetic residues from the detonation of common US ordnance. International Journal of Energetic Materials and Chemical Propulsion, 10(2). [https://doi.org/10.1615/intjenergeticmaterialschemprop.2012004956 doi: 10.1615/IntJEnergeticMaterialsChemProp.2012004956] [[media:Walsh-2011-Energetic-Residues-Common-US-Ordnance.pdf| Report.pdf]]</ref>. Duds are initially intact but may become perforated or fragmented into micrometer to centimeter;o0i0k-sized particles by nearby detonations<ref>Walsh, M.R., Thiboutot, S., Walsh, M.E., Ampleman, G., Martel, R., Poulin, I. and Taylor, S., 2011. Characterization and fate of gun and rocket propellant residues on testing and training ranges (No. ERDC/CRREL-TR-11-13). Engineer Research and Development Center / Cold Regions Research and Engineering Lab (ERDC/CRREL) TR-11-13, Hanover, NH, USA. [[media:Epa-2006-method-8330b.pdf| Report.pdf]]</ref>. Low-order detonations can scatter micrometer to centimeter-sized particles up to 20 m from the site of detonation<ref name= "Taylor2004">Taylor, S., Hewitt, A., Lever, J., Hayes, C., Perovich, L., Thorne, P. and Daghlian, C., 2004. TNT particle size distributions from detonated 155-mm howitzer rounds. Chemosphere, 55(3), pp.357-367.[[media:Taylor-2004 TNT PSDs.pdf| Report.pdf]]</ref>
  
Given the prevalence and persistence of back diffusion plumes, modeling tools are needed that can accurately capture the dynamic progression of the contamination in both loading and unloading periods.
+
The particulate nature of munitions constituents in the environment presents a distinct challenge to representative soil sampling. Figure 2 shows an array of discrete soil samples collected around the site of a low-order detonation – resultant soil concentrations vary by orders of magnitude within centimeters of each other. The inadequacy of discrete sampling is apparent in characterization studies from actual ranges which show wide-ranging concentrations and poor precision (Table 1).
  
==Matrix Diffusion Modeling Options==
+
In comparison to discrete sampling, incremental sampling tends to yield reproducible concentrations (low relative standard deviation [RSD]) that statistically better represent an area of interest<ref name= "Hewitt2009"/>.
There have been limited options available for simulating matrix diffusion effects at groundwater contamination sites. Available analytical models require simplifications that may overlook the unique complexities at actual remediation sites.  Numerical models are often used to simulate solute transport, but require fine discretization to capture matrix diffusion concentration gradients, which occur at millimeters to centimeters scale.  This very fine discretization (fine model grid) greatly increases the number of model cells required to simulate transport at field sites, greatly increasing computer model run times. Figure 2 shows a typical fine grid model containing almost 3 million gridblocks (cells).
 
[[File:Falta1w2 Fig2.PNG|thumb|Figure 2.  A Typical Fine Grid Model Capable of Simulating Matrix Diffusion]]
 
  
Another type of numerical model of matrix diffusion is the dual-porosity model which is more computationally efficient because it uses a first order approximation to describe the mass transfer between the high K and low K zones. Dual-porosity models can be calibrated to match the effects of matrix diffusion on contaminant concentrations over short time periods.  However, the first order mass transfer coefficient is time-dependent<ref>Guan, J., Molz, F.J., Zhou, Q., Liu, H.H. and Zheng, C., 2008. Behavior of the mass transfer coefficient during the MADE‐2 experiment: New insights. Water Resources Research, 44(2). [https://doi.org/10.1029/2007WR006120 doi: 10.1029/2007WR006120]</ref>, so models calibrated to early time data will not accurately simulate concentrations at a later time.
+
{| class="wikitable" style="float: right; text-align: center; margin-left: auto; margin-right: auto;"
 
+
|+ Table 1. Soil Sample Concentrations and Precision from Military Ranges Using Discrete and Incremental Sampling. (Data from Taylor et al. <ref name= "Taylor2011"/> and references therein.)
The REMChlor-MD model was developed under ESTCP project [https://www.serdp-estcp.org/Program-Areas/Environmental-Restoration/Contaminated-Groundwater/Persistent-Contamination/ER-201426  ER-201426] to provide a practical alternative to previous approaches for modeling matrix diffusion.
+
|-
 
+
! Military Range Type !! Analyte !! Range<br/>(mg/kg) !! Median<br/>(mg/kg) !! RSD<br/>(%)
==REMChlor-MD’s Modeling Approach==
+
|-
REMChlor-MD employs a semi-analytical method adapted from a heat diffusivity estimation strategy used in petroleum engineering and applied here to the problem of chemical diffusion with first order decay kinetics<ref>Vinsome, P.K.W. and Westerveld, J., 1980. A simple method for predicting cap and base rock heat losses in'thermal reservoir simulators. Journal of Canadian Petroleum Technology, 19(03).  Pp 87-90 [https://doi.org/10.2118/80-03-04 doi: 10.2118/80-03-04]</ref><ref name= "Falta2017">Falta, R.W. and Wang, W., 2017. A semi-analytical method for simulating matrix diffusion in numerical transport models. Journal of contaminant hydrology, 197, pp.39-49. [https://doi.org/10.1016/j.jconhyd.2016.12.007 doi: 10.1016/j.jconhyd.2016.12.007]</ref>. This semi-analytical method utilizes a fitting function to describe the concentration of the chemical in the low K zones in, or adjacent to, each gridblock:
+
| colspan="5" style="text-align: left;" | '''Discrete Samples'''
 
+
|-
::'''Equation 1:'''&nbsp;&nbsp;&nbsp;&nbsp;<big>''C<sub>l</sub>(z<sub>l</sub>,t) = (C</big><sup>t + &Delta;t</sup><big> + pz<sub>l</sub> + qz<sub>l</sub><sup>2</sup>) e</big><sup>-z<sub>l</sub> / d''
+
| Artillery FP || 2,4-DNT || <0.04 – 6.4 || 0.65 || 110
{|
+
|-
 +
| Antitank Rocket || HMX || 5.8 – 1,200 || 200 || 99
 +
|-
 +
| Bombing || TNT || 0.15 – 780 || 6.4 || 274
 +
|-
 +
| Mortar || RDX || <0.04 – 2,400 || 1.7 || 441
 
|-
 
|-
| where:
+
| Artillery || RDX || <0.04 – 170 || <0.04 || 454
 
|-
 
|-
| ''C<sub>l</sub>'' || represents the concentration in the low K material;
+
| colspan="5" style="text-align: left;" | '''Incremental Samples*'''
 
|-
 
|-
| ''C<sup>t + &Delta;t</sup>'' || is the concentration in the high K zone during the next numerical time-step;
+
| Artillery FP || 2,4-DNT || 0.60 – 1.4 || 0.92 || 26
 
|-
 
|-
| ''z<sub>l</sub>'' || is the distance into the low K zone from the high K / low K interface;
+
| Bombing || TNT || 13 – 17 || 14 || 17
 
|-
 
|-
| ''p'' and ''q''&nbsp;&nbsp;&nbsp;&nbsp; || are fitting parameters that can be solved for by enforcing the diffusion differential equation at the interface and mass conservation in the matrix gridblock; and
+
| Artillery/Bombing || RDX || 3.9 – 9.4 || 4.8 || 38
 +
|-
 +
| Thermal Treatment || HMX || 3.96 – 4.26 || 4.16 || 4
 
|-
 
|-
| ''d'' || is the concentration penetration depth, which is defined in the REMChlor-MD User’s Manual in Appendix 1.
+
| colspan="5" style="text-align: left; background-color: white;" | * For incremental samples, 30-100 increments and 3-10 replicate samples were collected.
 
|}
 
|}
  
Using the trial function, the diffusive mass flux can be computed analytically, which is the key time-saving component of the method. The only variable that requires numerical simulation is the concentration in the high-permeability parts of the aquifer.
+
==Incremental Sampling Approach==
 +
ISM is a requisite for representative and reproducible sampling of training ranges, but it is an involved process that is detailed thoroughly elsewhere<ref name= "Hewitt2009"/><ref name= "Taylor2011"/><ref name= "USEPA2006M"/>. In short, ISM involves the collection of many (30 to >100) increments in a systematic pattern within a decision unit (DU). The DU may cover an area where releases are thought to have occurred or may represent an area relevant to ecological receptors (e.g., sensitive species). Figure 3 shows the ISM sampling pattern in a simplified (5x5 square) DU. Increments are collected at a random starting point with systematic distances between increments. Replicate samples can be collected by starting at a different random starting point, often at a different corner of the DU. Practically, this grid pattern can often be followed with flagging or lathe marking DU boundaries and/or sampling lanes and with individual pacing keeping systematic distances between increments. As an example, an artillery firing point might include a 100x100 m DU with 81 increments.
 +
[[File:Beal1w2 Fig3.png|thumb|200 px|left|Figure 3. Example ISM sampling pattern on a square decision unit. Replicates are collected in a systematic pattern from a random starting point at a corner of the DU. Typically more than the 25 increments shown are collected]]
  
There are three key geometrical model inputs to the semi-analytical method that can be used to model the diffusion process:
+
DUs can vary in shape (Figure 4), size, number of increments, and number of replicates according to a project’s data quality objectives.
*''V<sub>f</sub>'' is the volume fraction of the high-permeability materials in a gridblock;
 
*''A<sub>md</sub>'' is the interfacial area between the high-K and low-K materials in a gridblock; and
 
*''L'' is the characteristic maximum diffusion length.
 
  
Because the volume of low K material in the gridblock should equal the product of the interfacial area and the characteristic maximum diffusion length, the third parameter (L) can be calculated from the first two using the volume balance equation for a gridblock of volume V:
+
[[File:Beal1w2 Fig4.png|thumb|right|250 px|Figure 4: Incremental sampling of a circular DU on snow shows sampling lanes with a two-person team in process of collecting the second replicate in a perpendicular path to the first replicate. (Photo: Matthew Bigl)]]
  
::'''Equation 2:'''&nbsp;&nbsp;&nbsp;&nbsp;<big>''A<sub>md</sub> L = (1 – V<sub>f</sub>) V''</big>
+
==Sampling Tools==
 +
In many cases, energetic compounds are expected to reside within the soil surface. Figure 5 shows soil depth profiles on some studied impact areas and firing points. Overall, the energetic compound concentrations below 5-cm soil depth are negligible relative to overlying soil concentrations. For conventional munitions, this is to be expected as the energetic particles are relatively insoluble, and any dissolved compounds readily adsorb to most soils<ref>Pennington, J.C., Jenkins, T.F., Ampleman, G., Thiboutot, S., Brannon, J.M., Hewitt, A.D., Lewis, J., Brochu, S., 2006. Distribution and fate of energetics on DoD test and training ranges: Final Report. ERDC TR-06-13, Vicksburg, MS, USA. Also: SERDP/ESTCP Project ER-1155. [[media:Pennington-2006_ERDC-TR-06-13_ESTCP-ER-1155-FR.pdf| Report.pdf]]</ref>. Physical disturbance, as on hand grenade ranges, may require deeper sampling either with a soil profile or a corer/auger.
  
This semi-analytical method has undergone thorough testing and proved to be accurate in numerous scenarios<ref name= "Falta2017"/><ref name= "Muskus2018">Muskus, N. and Falta, R.W., 2018. Semi-analytical method for matrix diffusion in heterogeneous and fractured systems with parent-daughter reactions. Journal of Contaminant Hydrology, 218, pp.94-109. [https://doi.org/10.1016/j.jconhyd.2018.10.002 doi: 10.1016/j.jconhyd.2018.10.002]</ref>.  The following section shows common test scenarios where the method can be used.
+
[[File:Beal1w2 Fig5.png|thumb|left|200 px|Figure 5. Depth profiles of high explosive compounds at impact areas (bottom) and of propellant compounds at firing points (top). Data from: Hewitt et al. <ref>Hewitt, A.D., Jenkins, T.F., Ramsey, C.A., Bjella, K.L., Ranney, T.A. and Perron, N.M., 2005. Estimating energetic residue loading on military artillery ranges: Large decision units (No. ERDC/CRREL-TR-05-7). [[media:Hewitt-2005 ERDC-CRREL TR-05-7.pdf| Report.pdf]]</ref> and Jenkins et al. <ref>Jenkins, T.F., Ampleman, G., Thiboutot, S., Bigl, S.R., Taylor, S., Walsh, M.R., Faucher, D., Mantel, R., Poulin, I., Dontsova, K.M. and Walsh, M.E., 2008. Characterization and fate of gun and rocket propellant residues on testing and training ranges (No. ERDC-TR-08-1). [[media:Jenkins-2008 ERDC TR-08-1.pdf| Report.pdf]]</ref>]]
  
[[File:Falta1w2 Fig3.png|thumb|left|Figure 3.  Concentration profiles in fractures at 1, 49, 51 and 100 years.]]
+
Soil sampling with the Cold Regions Research and Engineering Laboratory (CRREL) Multi-Increment Sampling Tool (CMIST) or similar device is an easy way to collect ISM samples rapidly and reproducibly. This tool has an adjustable diameter size corer and adjustable depth to collect surface soil plugs (Figure 6). The CMIST can be used at almost a walking pace (Figure 7) using a two-person sampling team, with one person operating the CMIST and the other carrying the sample container and recording the number of increments collected. The CMIST with a small diameter tip works best in soils with low cohesion, otherwise conventional scoops may be used. Maintaining consistent soil increment dimensions is critical.
[[File:Falta1w2 Fig4.png|thumb|left|Figure 4: Simulated mass discharge at downstream edge of fine-grid MT3DMS and REMChlor-MD models.]]
 
[[File:Falta1w2 Fig5.png|thumb|right|Figure 5: Top view (''xy'') of concentration contours computed with the fine-grid numerical model (top) and with REMChlor-MD (bottom) at 30 years.]]
 
[[File:Falta1w2 Fig6.png|thumb|right|Figure 6: Top view (''xy'') of concentration contours computed with the fine-grid numerical model (top) and with REMChlor-MD (bottom) at 130 years.]]
 
  
==Modeling Examples==
+
The sampling tool should be cleaned between replicates and between DUs to minimize potential for cross-contamination<ref>Walsh, M.R., 2009. User’s manual for the CRREL Multi-Increment Sampling Tool. Engineer Research and Development Center / Cold Regions Research and Engineering Lab (ERDC/CRREL) SR-09-1, Hanover, NH, USA.  [[media:Walsh-2009 ERDC-CRREL SR-09-1.pdf | Report.pdf]]</ref>.
  
===Matrix Diffusion in Fractured Media===
+
==Sample Processing==
Contaminants in fractured rock media are often the most challenging to remediate because of the small volume of individual fractures and the large surface area of the matrix. Networks of fractures can store a large mass of contaminants and potentially become a secondary source over time.
+
While only 10 g of soil is typically used for chemical analysis, incremental sampling generates a sample weighing on the order of 1 kg. Splitting of a sample, either in the field or laboratory, seems like an easy way to reduce sample mass; however this approach has been found to produce high uncertainty for explosives and propellants, with a median RSD of 43.1%<ref name= "Hewitt2009"/>. Even greater error is associated with removing a discrete sub-sample from an unground sample. Appendix A in [https://www.epa.gov/sites/production/files/2015-07/documents/epa-8330b.pdf U.S. EPA Method 8330B]<ref name= "USEPA2006M"/> provides details on recommended ISM sample processing procedures.
  
The semi-analytical method was tested in a parallel fractured system, and the results were compared against an exact analytical solution<ref>Sudicky, E.A. and Frind, E.O., 1982. Contaminant transport in fractured porous media: Analytical solutions for a system of parallel fractures. Water Resources Research, 18(6), pp.1634-1642. [https://doi.org/10.1029/WR018i006p01634 doi: 10.1029/WR018i006p01634]</ref>. In this example, the fracture spacing was set at 2 m, and the fracture aperture was 100 ''µm''. A known concentration of trichloroethene (TCE) was introduced into the fractures at the upstream end of the model, where its concentration was held constant for 50 years. Then the contaminant source was completely removed, and clean water was flushed through the fractures for another 50 years. Figure 3 shows the concentration profiles in the fractures calculated using the semi-analytical method over 100 years of simulation time compared to the exact analytical solution.
+
Incremental soil samples are typically air dried over the course of a few days. Oven drying thermally degrades some energetic compounds and should be avoided<ref>Cragin, J.H., Leggett, D.C., Foley, B.T., and Schumacher, P.W., 1985. TNT, RDX and HMX explosives in soils and sediments: Analysis techniques and drying losses. (CRREL Report 85-15) Hanover, NH, USA. [[media:Cragin-1985 CRREL 85-15.pdf| Report.pdf]]</ref>. Once dry, the samples are sieved with a 2-mm screen, with only the less than 2-mm fraction processed further. This size fraction represents the USDA definition of soil. Aggregate soil particles should be broken up and vegetation shredded to pass through the sieve. Samples from impact or demolition areas may contain explosive particles from low order detonations that are greater than 2 mm and should be identified, given appropriate caution, and potentially weighed.
  
In the 1-year profile, TCE concentrations in the fractures decreased rapidly with distance, since the steep concentration gradient results in rapid mass transfer from the fracture into the low K matrix. Toward the end of the loading period at 49 years, TCE concentration declines more gradually with distance, since TCE has now diffused into the adjoining low K matrix, reducing the concentration gradient and associated mass flux. At 51 years, one year after source removal, there was a decrease in TCE concentration near the inlet, but almost no effect downstream in the fractures due to back diffusion of TCE from the matrix to the fractures. After 50 years of clean water flushing, back diffusing TCE was still acting as a secondary source, reaching a peak concentration of approximately 16 mg/L. The profiles show a typical fractured system prior to and after source removal with matrix diffusion effects preventing the system from reaching clean-up level.
+
The <2-mm soil fraction is typically still ≥1 kg and impractical to extract in full for analysis. However, subsampling at this stage is not possible due to compositional heterogeneity, with the energetic compounds generally present as <0.5 mm particles<ref name= "Walsh2017"/><ref name= "Taylor2004"/>. Particle size reduction is required to achieve a representative and precise measure of the sample concentration. Grinding in a puck mill to a soil particle size <75 µm has been found to be required for representative/reproducible sub-sampling (Figure 8). For samples thought to contain propellant particles, a prolonged milling time is required to break down these polymerized particles and achieve acceptable precision (Figure 9). Due to the multi-use nature of some ranges, a 5-minute puck milling period can be used for all soils. Cooling periods between 1-minute milling intervals are recommended to avoid thermal degradation. Similar to field sampling, sub-sampling is done incrementally by spreading the sample out to a thin layer and collecting systematic random increments of consistent volume to a total mass for extraction of 10 g (Figure 10).
  
The match between the analytical solution and the semi-analytical method was good, particularly during the loading period. In the unloading period, the semi-analytical method overestimated the peak concentration somewhat, but it still captured the long tailing of concentration almost exactly.
+
<li style="display: inline-block;">[[File:Beal1w2 Fig6.png|thumb|200 px|Figure 6: CMIST soil sampling tool (top) and with ejected increment core using a large diameter tip (bottom).]]</li>
 +
<li style="display: inline-block;">[[File:Beal1w2 Fig7.png|thumb|200 px|Figure 7: Two person sampling team using CMIST, bag-lined bucket, and increment counter. (Photos: Matthew Bigl)]]</li>
 +
<li style="display: inline-block;">[[File:Beal1w2 Fig8.png|thumb|200 px|Figure 8: Effect of machine grinding on RDX and TNT concentration and precision in soil from a hand grenade range. Data from Walsh et al.<ref>Walsh, M.E., Ramsey, C.A. and Jenkins, T.F., 2002. The effect of particle size reduction by grinding on subsampling variance for explosives residues in soil. Chemosphere, 49(10), pp.1267-1273. [https://doi.org/10.1016/S0045-6535(02)00528-3 doi: 10.1016/S0045-6535(02)00528-3]</ref> ]]</li>
 +
<li style="display: inline-block;">[[File:Beal1w2 Fig9.png|thumb|200 px|Figure 9: Effect of puck milling time on 2,4-DNT concentration and precision in soil from a firing point. Data from Walsh et al.<ref>Walsh, M.E., Ramsey, C.A., Collins, C.M., Hewitt, A.D., Walsh, M.R., Bjella, K.L., Lambert, D.J. and Perron, N.M., 2005. Collection methods and laboratory processing of samples from Donnelly Training Area Firing Points, Alaska, 2003 (No. ERDC/CRREL-TR-05-6). [[media:Walsh-2005 ERDC-CRREL TR-05-6.pdf| Report.pdf]]</ref>.]]</li>
 +
<li style="display: inline-block;">[[File:Beal1w2 Fig10.png|thumb|200 px|center|Figure 10: Incremental sub-sampling of a milled soil sample spread out on aluminum foil.]]</li>
  
[[File:Falta1w2 Fig7.mp4|thumb|left|Figure 7: Video tutorial demonstrating use of REMChlor-MD]]
+
==Analysis==
 
+
Soil sub-samples are extracted and analyzed following [[Media: epa-2006-method-8330b.pdf | EPA Method 8330B]]<ref name= "USEPA2006M"/> and [[Media:epa-2007-method-8095.pdf | Method 8095]]<ref name= "USEPA2007M"/> using [[Wikipedia: High-performance liquid chromatography | High Performance Liquid Chromatography (HPLC)]] and [[Wikipedia: Gas chromatography | Gas Chromatography (GC)]], respectively. Common estimated reporting limits for these analysis methods are listed in Table 2.
===Matrix Diffusion in Heterogeneous Media===
 
One of the motivations for REMChlor-MD was to develop an alternative to a computationally expensive fine grid model for simulating matrix diffusion.  To evaluate the ability of the semi-analytical method incorporated into REMChlor-MD to accurately simulate back-diffusion, a heterogeneous fine grid model was developed<ref name= "Muskus2018"/> using MT3DMS. A geostatistical model generated multiple realizations of the subsurface’s heterogeneity using synthetic borehole data.  Using the heterogeneous material distribution of sand and clay, a fine grid flow and transport model, consisting of almost 3 million gridblocks, was created. A TCE source was present near the upgradient end of the model for 30 years and then it was completely removed. Following source removal, clean water was flushed through the system for another 200 years.
 
 
 
For comparison, a REMChlor-MD model was developed with a much coarser grid (about 14,000 gridblocks) where flow and transport parameters were comparable to the fine grid model. The mass discharge at the outlet of the models was used as a comparison criterion between the two models.  Figure 4 shows a comparison of the computed mass discharge versus time at the downstream edge of the model.  The mass discharge predicted by the 3-million gridblock MT2DMS simulation is shown as the yellow dots.  The green line shows the REMChlor-MD prediction based on the geostatistical properties of the high and low K zones (characteristic maximum diffusion length, L= 1.85m).  The REMChlor-MD match to the fine-grid MT3DMS model was slightly improved by reducing the characteristic maximum diffusion length (L= 1.5m).  Both of these results are similar to the 3-million gridblock MT3DMS simulation result.
 
 
 
Plan view concentrations from the fine grid model and the semi-analytical model were extracted to generate contours of the TCE plume in Figures 5 and 6 after 30 and 130 years, respectively.  The top panels of these figures were generated with the fine grid MT3DMS model, whereas the bottom panels were generated with the REMChlor-MD model. The fine grid contour has jagged edges because small changes in concentration were captured at a higher spatial resolution. In contrast, the semi-analytical contour was smoother because the concentration changes were interpolated over larger gridblocks. However, the overall shape and contours from the two models are similar.  On a high performance 2018 workstation, the run time for the fine-grid MT3DMS model was many hours, while the REMChlor-MD model completed its run in a little over one minute.
 
 
 
==REMChlor-MD User Interface==
 
The semi-analytical method was programmed in FORTRAN, and a graphical user interface (GUI) was built using Visual Basic in Excel® to facilitate data inputs and output processing<ref>Farhat, S. K., Newell, C. J., Falta, R. W., and Lynch, K. (2018). REMChlor-MD toolkit user’s manual. ER-201426 [[media:2018-Falta-REMChlor_Modeling_Matrix_Diffusion_Effects.pdf | Report.pdf]]</ref>.  The Visual Basic REMChlor-MD interface calls the FORTRAN executable using a dynamic link library, and then processes the FORTRAN output files to produce various graphs.  A short video tutorial demonstrating data entry and some of the capabilities of REMChlor-MD is shown in Figure 7.  Additional detail is provided in an online [https://www.serdp-estcp.org/Tools-and-Training/Webinar-Series/02-07-2019 webinar].  The toolkit package, including a user’s manual is available free of charge and can be downloaded here ([https://www.serdp-estcp.org/content/download/48566/462193/file/REMChlorMD_64bit.zip 64bit version]) and ([https://www.serdp-estcp.org/content/download/48565/462183/file/REMChlorMD_32bit.zip 32 bit version]). Readers are referred to the [https://www.serdp-estcp.org/content/download/48433/460814/file/ER-201426%20REMChlor-MD%20User's%20Manual.pdf user manual] for a complete explanation of the program, its input parameters, outputs, and examples.
 
  
 +
{| class="wikitable" style="float: center; text-align: center; margin-left: auto; margin-right: auto;"
 +
|+ Table 2. Typical Method Reporting Limits for Energetic Compounds in Soil. (Data from Hewitt et al.<ref>Hewitt, A., Bigl, S., Walsh, M., Brochu, S., Bjella, K. and Lambert, D., 2007. Processing of training range soils for the analysis of energetic compounds (No. ERDC/CRREL-TR-07-15). Hanover, NH, USA. [[media:Hewitt-2007 ERDC-CRREL TR-07-15.pdf| Report.pdf]]</ref>)
 +
|-
 +
! rowspan="2" | Compound
 +
! colspan="2" | Soil Reporting Limit (mg/kg)
 +
|-
 +
! HPLC (8330)
 +
! GC (8095)
 +
|-
 +
| HMX || 0.04 || 0.01
 +
|-
 +
| RDX || 0.04 || 0.006
 +
|-
 +
| [[Wikipedia: 1,3,5-Trinitrobenzene | TNB]] || 0.04 || 0.003
 +
|-
 +
| TNT || 0.04 || 0.002
 +
|-
 +
| [[Wikipedia: 2,6-Dinitrotoluene | 2,6-DNT]] || 0.08 || 0.002
 +
|-
 +
| 2,4-DNT || 0.04 || 0.002
 +
|-
 +
| 2-ADNT || 0.08 || 0.002
 +
|-
 +
| 4-ADNT || 0.08 || 0.002
 +
|-
 +
| NG || 0.1 || 0.01
 +
|-
 +
| [[Wikipedia: Dinitrobenzene | DNB ]] || 0.04 || 0.002
 +
|-
 +
| [[Wikipedia: Tetryl | Tetryl ]]  || 0.04 || 0.01
 +
|-
 +
| [[Wikipedia: Pentaerythritol tetranitrate | PETN ]] || 0.2 || 0.016
 +
|}
  
 
==References==
 
==References==
 
+
<references/>
<references />
 
  
 
==See Also==
 
==See Also==
 +
*[https://itrcweb.org/ Interstate Technology and Regulatory Council]
 +
*[http://www.hawaiidoh.org/tgm.aspx Hawaii Department of Health]
 +
*[http://envirostat.org/ Envirostat]

Latest revision as of 18:58, 29 April 2020

The heterogeneous distribution of munitions constituents, released as particles from munitions firing and detonations on military training ranges, presents challenges for representative soil sample collection and for defensible decision making. Military range characterization studies and the development of the incremental sampling methodology (ISM) have enabled the development of recommended methods for soil sampling that produce representative and reproducible concentration data for munitions constituents. This article provides a broad overview of recommended soil sampling and processing practices for analysis of munitions constituents on military ranges.

Related Article(s):


CONTRIBUTOR(S): Dr. Samuel Beal


Key Resource(s):

Introduction

Figure 1: Downrange distance of visible propellant plume on snow from the firing of different munitions. Note deposition behind firing line for the 84-mm rocket. Data from: Walsh et al.[5][6]
Figure 2: A low-order detonation mortar round (top) with surrounding discrete soil samples produced concentrations spanning six orders of magnitude within a 10m by 10m area (bottom). (Photo and data: A.D. Hewitt)

Munitions constituents are released on military testing and training ranges through several common mechanisms. Some are locally dispersed as solid particles from incomplete combustion during firing and detonation. Also, small residual particles containing propellant compounds (e.g., nitroglycerin [NG] and 2,4-dinitrotoluene [2,4-DNT]) are distributed in front of and surrounding target practice firing lines (Figure 1). At impact areas and demolition areas, high order detonations typically yield very small amounts (<1 to 10 mg/round) of residual high explosive compounds (e.g., TNT , RDX and HMX ) that are distributed up to and sometimes greater than) 24 m from the site of detonation[7].

Low-order detonations and duds are thought to be the primary source of munitions constituents on ranges[8][9]. Duds are initially intact but may become perforated or fragmented into micrometer to centimeter;o0i0k-sized particles by nearby detonations[10]. Low-order detonations can scatter micrometer to centimeter-sized particles up to 20 m from the site of detonation[11]

The particulate nature of munitions constituents in the environment presents a distinct challenge to representative soil sampling. Figure 2 shows an array of discrete soil samples collected around the site of a low-order detonation – resultant soil concentrations vary by orders of magnitude within centimeters of each other. The inadequacy of discrete sampling is apparent in characterization studies from actual ranges which show wide-ranging concentrations and poor precision (Table 1).

In comparison to discrete sampling, incremental sampling tends to yield reproducible concentrations (low relative standard deviation [RSD]) that statistically better represent an area of interest[2].

Table 1. Soil Sample Concentrations and Precision from Military Ranges Using Discrete and Incremental Sampling. (Data from Taylor et al. [1] and references therein.)
Military Range Type Analyte Range
(mg/kg)
Median
(mg/kg)
RSD
(%)
Discrete Samples
Artillery FP 2,4-DNT <0.04 – 6.4 0.65 110
Antitank Rocket HMX 5.8 – 1,200 200 99
Bombing TNT 0.15 – 780 6.4 274
Mortar RDX <0.04 – 2,400 1.7 441
Artillery RDX <0.04 – 170 <0.04 454
Incremental Samples*
Artillery FP 2,4-DNT 0.60 – 1.4 0.92 26
Bombing TNT 13 – 17 14 17
Artillery/Bombing RDX 3.9 – 9.4 4.8 38
Thermal Treatment HMX 3.96 – 4.26 4.16 4
* For incremental samples, 30-100 increments and 3-10 replicate samples were collected.

Incremental Sampling Approach

ISM is a requisite for representative and reproducible sampling of training ranges, but it is an involved process that is detailed thoroughly elsewhere[2][1][3]. In short, ISM involves the collection of many (30 to >100) increments in a systematic pattern within a decision unit (DU). The DU may cover an area where releases are thought to have occurred or may represent an area relevant to ecological receptors (e.g., sensitive species). Figure 3 shows the ISM sampling pattern in a simplified (5x5 square) DU. Increments are collected at a random starting point with systematic distances between increments. Replicate samples can be collected by starting at a different random starting point, often at a different corner of the DU. Practically, this grid pattern can often be followed with flagging or lathe marking DU boundaries and/or sampling lanes and with individual pacing keeping systematic distances between increments. As an example, an artillery firing point might include a 100x100 m DU with 81 increments.

Figure 3. Example ISM sampling pattern on a square decision unit. Replicates are collected in a systematic pattern from a random starting point at a corner of the DU. Typically more than the 25 increments shown are collected

DUs can vary in shape (Figure 4), size, number of increments, and number of replicates according to a project’s data quality objectives.

Figure 4: Incremental sampling of a circular DU on snow shows sampling lanes with a two-person team in process of collecting the second replicate in a perpendicular path to the first replicate. (Photo: Matthew Bigl)

Sampling Tools

In many cases, energetic compounds are expected to reside within the soil surface. Figure 5 shows soil depth profiles on some studied impact areas and firing points. Overall, the energetic compound concentrations below 5-cm soil depth are negligible relative to overlying soil concentrations. For conventional munitions, this is to be expected as the energetic particles are relatively insoluble, and any dissolved compounds readily adsorb to most soils[12]. Physical disturbance, as on hand grenade ranges, may require deeper sampling either with a soil profile or a corer/auger.

Figure 5. Depth profiles of high explosive compounds at impact areas (bottom) and of propellant compounds at firing points (top). Data from: Hewitt et al. [13] and Jenkins et al. [14]

Soil sampling with the Cold Regions Research and Engineering Laboratory (CRREL) Multi-Increment Sampling Tool (CMIST) or similar device is an easy way to collect ISM samples rapidly and reproducibly. This tool has an adjustable diameter size corer and adjustable depth to collect surface soil plugs (Figure 6). The CMIST can be used at almost a walking pace (Figure 7) using a two-person sampling team, with one person operating the CMIST and the other carrying the sample container and recording the number of increments collected. The CMIST with a small diameter tip works best in soils with low cohesion, otherwise conventional scoops may be used. Maintaining consistent soil increment dimensions is critical.

The sampling tool should be cleaned between replicates and between DUs to minimize potential for cross-contamination[15].

Sample Processing

While only 10 g of soil is typically used for chemical analysis, incremental sampling generates a sample weighing on the order of 1 kg. Splitting of a sample, either in the field or laboratory, seems like an easy way to reduce sample mass; however this approach has been found to produce high uncertainty for explosives and propellants, with a median RSD of 43.1%[2]. Even greater error is associated with removing a discrete sub-sample from an unground sample. Appendix A in U.S. EPA Method 8330B[3] provides details on recommended ISM sample processing procedures.

Incremental soil samples are typically air dried over the course of a few days. Oven drying thermally degrades some energetic compounds and should be avoided[16]. Once dry, the samples are sieved with a 2-mm screen, with only the less than 2-mm fraction processed further. This size fraction represents the USDA definition of soil. Aggregate soil particles should be broken up and vegetation shredded to pass through the sieve. Samples from impact or demolition areas may contain explosive particles from low order detonations that are greater than 2 mm and should be identified, given appropriate caution, and potentially weighed.

The <2-mm soil fraction is typically still ≥1 kg and impractical to extract in full for analysis. However, subsampling at this stage is not possible due to compositional heterogeneity, with the energetic compounds generally present as <0.5 mm particles[7][11]. Particle size reduction is required to achieve a representative and precise measure of the sample concentration. Grinding in a puck mill to a soil particle size <75 µm has been found to be required for representative/reproducible sub-sampling (Figure 8). For samples thought to contain propellant particles, a prolonged milling time is required to break down these polymerized particles and achieve acceptable precision (Figure 9). Due to the multi-use nature of some ranges, a 5-minute puck milling period can be used for all soils. Cooling periods between 1-minute milling intervals are recommended to avoid thermal degradation. Similar to field sampling, sub-sampling is done incrementally by spreading the sample out to a thin layer and collecting systematic random increments of consistent volume to a total mass for extraction of 10 g (Figure 10).

  • Figure 6: CMIST soil sampling tool (top) and with ejected increment core using a large diameter tip (bottom).
  • Figure 7: Two person sampling team using CMIST, bag-lined bucket, and increment counter. (Photos: Matthew Bigl)
  • Figure 8: Effect of machine grinding on RDX and TNT concentration and precision in soil from a hand grenade range. Data from Walsh et al.[17]
  • Figure 9: Effect of puck milling time on 2,4-DNT concentration and precision in soil from a firing point. Data from Walsh et al.[18].
  • Figure 10: Incremental sub-sampling of a milled soil sample spread out on aluminum foil.
  • Analysis

    Soil sub-samples are extracted and analyzed following EPA Method 8330B[3] and Method 8095[4] using High Performance Liquid Chromatography (HPLC) and Gas Chromatography (GC), respectively. Common estimated reporting limits for these analysis methods are listed in Table 2.

    Table 2. Typical Method Reporting Limits for Energetic Compounds in Soil. (Data from Hewitt et al.[19])
    Compound Soil Reporting Limit (mg/kg)
    HPLC (8330) GC (8095)
    HMX 0.04 0.01
    RDX 0.04 0.006
    TNB 0.04 0.003
    TNT 0.04 0.002
    2,6-DNT 0.08 0.002
    2,4-DNT 0.04 0.002
    2-ADNT 0.08 0.002
    4-ADNT 0.08 0.002
    NG 0.1 0.01
    DNB 0.04 0.002
    Tetryl 0.04 0.01
    PETN 0.2 0.016

    References

    1. ^ 1.0 1.1 1.2 Taylor, S., Jenkins, T.F., Bigl, S., Hewitt, A.D., Walsh, M.E. and Walsh, M.R., 2011. Guidance for Soil Sampling for Energetics and Metals (No. ERDC/CRREL-TR-11-15). Report.pdf
    2. ^ 2.0 2.1 2.2 2.3 Hewitt, A.D., Jenkins, T.F., Walsh, M.E., Bigl, S.R. and Brochu, S., 2009. Validation of sampling protocol and the promulgation of method modifications for the characterization of energetic residues on military testing and training ranges (No. ERDC/CRREL-TR-09-6). Engineer Research and Development Center / Cold Regions Research and Engineering Lab (ERDC/CRREL) TR-09-6, Hanover, NH, USA. Report.pdf
    3. ^ 3.0 3.1 3.2 3.3 U.S. Environmental Protection Agency (USEPA), 2006. Method 8330B (SW-846): Nitroaromatics, Nitramines, and Nitrate Esters by High Performance Liquid Chromatography (HPLC), Rev. 2. Washington, D.C. Report.pdf
    4. ^ 4.0 4.1 U.S. Environmental Protection Agency (US EPA), 2007. Method 8095 (SW-846): Explosives by Gas Chromatography. Washington, D.C. Report.pdf
    5. ^ Walsh, M.R., Walsh, M.E., Ampleman, G., Thiboutot, S., Brochu, S. and Jenkins, T.F., 2012. Munitions propellants residue deposition rates on military training ranges. Propellants, Explosives, Pyrotechnics, 37(4), pp.393-406. doi: 10.1002/prep.201100105
    6. ^ Walsh, M.R., Walsh, M.E., Hewitt, A.D., Collins, C.M., Bigl, S.R., Gagnon, K., Ampleman, G., Thiboutot, S., Poulin, I. and Brochu, S., 2010. Characterization and Fate of Gun and Rocket Propellant Residues on Testing and Training Ranges: Interim Report 2. (ERDC/CRREL TR-10-13. Also: ESTCP Project ER-1481) Report
    7. ^ 7.0 7.1 Walsh, M.R., Temple, T., Bigl, M.F., Tshabalala, S.F., Mai, N. and Ladyman, M., 2017. Investigation of Energetic Particle Distribution from High‐Order Detonations of Munitions. Propellants, Explosives, Pyrotechnics, 42(8), pp.932-941. doi: 10.1002/prep.201700089 Report.pdf
    8. ^ Hewitt, A.D., Jenkins, T.F., Walsh, M.E., Walsh, M.R. and Taylor, S., 2005. RDX and TNT residues from live-fire and blow-in-place detonations. Chemosphere, 61(6), pp.888-894. doi: 10.1016/j.chemosphere.2005.04.058
    9. ^ Walsh, M.R., Walsh, M.E., Poulin, I., Taylor, S. and Douglas, T.A., 2011. Energetic residues from the detonation of common US ordnance. International Journal of Energetic Materials and Chemical Propulsion, 10(2). doi: 10.1615/IntJEnergeticMaterialsChemProp.2012004956 Report.pdf
    10. ^ Walsh, M.R., Thiboutot, S., Walsh, M.E., Ampleman, G., Martel, R., Poulin, I. and Taylor, S., 2011. Characterization and fate of gun and rocket propellant residues on testing and training ranges (No. ERDC/CRREL-TR-11-13). Engineer Research and Development Center / Cold Regions Research and Engineering Lab (ERDC/CRREL) TR-11-13, Hanover, NH, USA. Report.pdf
    11. ^ 11.0 11.1 Taylor, S., Hewitt, A., Lever, J., Hayes, C., Perovich, L., Thorne, P. and Daghlian, C., 2004. TNT particle size distributions from detonated 155-mm howitzer rounds. Chemosphere, 55(3), pp.357-367. Report.pdf
    12. ^ Pennington, J.C., Jenkins, T.F., Ampleman, G., Thiboutot, S., Brannon, J.M., Hewitt, A.D., Lewis, J., Brochu, S., 2006. Distribution and fate of energetics on DoD test and training ranges: Final Report. ERDC TR-06-13, Vicksburg, MS, USA. Also: SERDP/ESTCP Project ER-1155. Report.pdf
    13. ^ Hewitt, A.D., Jenkins, T.F., Ramsey, C.A., Bjella, K.L., Ranney, T.A. and Perron, N.M., 2005. Estimating energetic residue loading on military artillery ranges: Large decision units (No. ERDC/CRREL-TR-05-7). Report.pdf
    14. ^ Jenkins, T.F., Ampleman, G., Thiboutot, S., Bigl, S.R., Taylor, S., Walsh, M.R., Faucher, D., Mantel, R., Poulin, I., Dontsova, K.M. and Walsh, M.E., 2008. Characterization and fate of gun and rocket propellant residues on testing and training ranges (No. ERDC-TR-08-1). Report.pdf
    15. ^ Walsh, M.R., 2009. User’s manual for the CRREL Multi-Increment Sampling Tool. Engineer Research and Development Center / Cold Regions Research and Engineering Lab (ERDC/CRREL) SR-09-1, Hanover, NH, USA. Report.pdf
    16. ^ Cragin, J.H., Leggett, D.C., Foley, B.T., and Schumacher, P.W., 1985. TNT, RDX and HMX explosives in soils and sediments: Analysis techniques and drying losses. (CRREL Report 85-15) Hanover, NH, USA. Report.pdf
    17. ^ Walsh, M.E., Ramsey, C.A. and Jenkins, T.F., 2002. The effect of particle size reduction by grinding on subsampling variance for explosives residues in soil. Chemosphere, 49(10), pp.1267-1273. doi: 10.1016/S0045-6535(02)00528-3
    18. ^ Walsh, M.E., Ramsey, C.A., Collins, C.M., Hewitt, A.D., Walsh, M.R., Bjella, K.L., Lambert, D.J. and Perron, N.M., 2005. Collection methods and laboratory processing of samples from Donnelly Training Area Firing Points, Alaska, 2003 (No. ERDC/CRREL-TR-05-6). Report.pdf
    19. ^ Hewitt, A., Bigl, S., Walsh, M., Brochu, S., Bjella, K. and Lambert, D., 2007. Processing of training range soils for the analysis of energetic compounds (No. ERDC/CRREL-TR-07-15). Hanover, NH, USA. Report.pdf

    See Also