RepresentativeRepresentative SamplingSampling forfor EnergeticEnergetic CompoundsCompounds atat MilitaryMilitary TrainingTraining RangesRanges

Sampling near tank target Sampling near low-order 500 lb YakimaTraining Center bomb, Camp Guernsey Red crater at CFB Gagetown

Alan D. Hewitt, Thomas F. Jenkins, Marianne E. Walsh, Michael R. Walsh, Susan Taylor, Charles Ramsey, and Judith C. Pennington MajorMajor ClassesClasses ofof EnergeticEnergetic ChemicalsChemicals UsedUsed byby DoDDoD

NO 2 CH3 N O 2 N NO2

Nitroaromatics N N O N (TNT) 2 NO 2 NO2 Nitramines (RDX)

H C-O-NO 2 2 Nitrate Esters HC-O-NO (NG) 2

H 22 C-O-NO EnergeticEnergetic ChemicalsChemicals inin MilitaryMilitary ExplosivesExplosives

• Composition B (artillery/mortar) 60% Military grade RDX (Contains about 10% HMX) 39% Military grade TNT (Contains about 1% other TNT isomers and DNTs) • Composition C4 (demolition explosive) 91% Military Grade RDX • Tritonal (Air Force Bombs) Military grade TNT, aluminum • Composition A4 (40-mm grenades) Military grade RDX • TNT (artillery) Military grade TNT • Composition H-6 (Air Force bombs) Military grade RDX and TNT, aluminum • Octol (Antitank rockets) Military grade HMX and TNT EnergeticEnergetic ChemicalsChemicals inin MilitaryMilitary GunGun PropellantsPropellants

• Nitrocellulose (NC) Polymer used in all gun propellants • Nitroglycerin (NG) Component of double and triple base propellants • Nitroguanidine (NQ) Component of triple base propellants • 2,4-Dinitrotoluene (2,4-DNT) Energetic plasticizer in some single base propellants PhysicalPhysical andand ChemicalChemical PropertiesProperties ofof EnergeticEnergetic ChemicalsChemicals

• Most are solids at environmental temperatures • Sources often particulate and deposited at soil surface • Low aqueous solubilities • Surface contamination of “chunks” can persist for long periods (50-100 years) - particularly in arid areas • RDX does not degrade rapidly under aerobic conditions • Once dissolved, RDX can migrate rapidly through vadose zone to ground water for sites with shallow groundwater, permeable soils, and moderate rainfall • TNT readily biotransforms and migration much slower than RDX • Compounds are relatively non-volatile (10-4 to 10-14 torr) - samples can be homogenized (or composited) without losses • Thermally labile - some can decompose upon heating Field Experiments

Cold Lake Fort Wainwright, AK Petawawa, Ontario Air Weapons Range, Alberta Training Ranges Hand Grenade Range Air-Surface Range CFB Valcartier, Quebec Fort Greeley, AK CFB-Wainwright, Alberta Training Ranges (Donnelly Training Area) Training Ranges Test Ranges CFB Gagetown, New Brunswick CFB-Shilo, Manitoba Training Ranges Battlerun Fort Richardson, AK Training Ranges Snow Experiments

Fort Lewis, WA Training Ranges Camp Bonneville, WA Hand Grenade Range Camp Ethan Allen, VT Snow Experiments Yakima Training Center, WA Training Ranges Massachusetts Military Fort Ord, CA Reservation, MA Anti-Tank Range Impact Ranges Hill AFB, UT Fort Drum, NY Demolition Range Snow and WP Experiments 29 Palms, CA Blossom Point, MD Marine Corps Training Ranges Detonation Test Range Fort Bliss, NM Camp Guernsey, WY Training Ranges Jefferson Artillery Range Proving Ground, IN Holloman AFB, NM Artillery Range Bombing Range Schofield Barracks, HI Fort Hood, TX Redstone Arsenal, AL Training Ranges Fort Carson, CO Camp Shelby, MS Training Ranges Detonation Test Range Artillery Range Training Ranges Pohakuloa Military Reservation, HI Fort Leonard Wood, MO Fort Polk, LA Eglin Air Force Base, FL Training Ranges Hand Grenade Range Training Ranges Air-Surface Range ResearchResearch onon DepositionDeposition ofof EnergeticEnergetic ResiduesResidues • Firing points Residues deposited as fibers and particles of burnt and unburnt propellant

• Impact areas Residues deposited as small particles of explosive from low-order (partial) detonations PropellantPropellant FiberFiber DepositedDeposited fromfrom 105105--mmmm HowitzerHowitzer

Fiber contained 24 µg 2,4-DNT

1 mm ResiduesResidues fromfrom 105105--mmmm PropellantPropellant atat FiringFiring PointsPoints

Residues are deposited as burned and unburned fibers of nitrocellulose (6% of unburned fiber weight is 2,4-DNT) • Average measured fiber length: – 3 mm • Average mass of 2,4-DNT: – 12 μg / fiber • If soil concentration is 1 mg/kg, then you have only about 80 fibers / kilogram of soil CompositionComposition BB particlesparticles fromfrom lowlow--orderorder 8181--mmmm mortarmortar

1mm ResiduesResidues ofof CompositionComposition BB DepositedDeposited atat ImpactImpact AreasAreas

• Particles (chunks) of Composition B deposited range in size from micrometer to centimeter • Soil sized particles are defined as < 2 mm • One 1-mm spherical particle of Composition B: - weighs about 0.9 mg - contains ~ 0.50 mg of RDX 0.35 mg of TNT 0.05 mg of HMX • If the soil concentration is 1 mg/kg of RDX, a kilogram sample contains only 2 of these 1-mm particles WhatWhat isis thethe MultiMulti--incrementincrement SamplingSampling Approach?Approach?

• Pooling of several individual increments from withinwithin thethe decisiondecision unitunit • Typically used to obtain a more reliable estimate of the mean concentration • Reduces analytical costs compared with multiple discrete samples • Normalizes data SamplingSampling DepthDepth

• At firing points, residues are associated with NC polymeric propellants and remain at the surface

• At impact areas, residues are particulates from low- order detonations and ruptured rounds and the bulk also remains at the surface

• For most ranges, we recommend a sampling depth of 2.5 to 5 cm

• Including deeper soil just dilutes the concentrations NormalizedNormalized DepthDepth ConcentrationConcentration ProfilesProfiles forfor SoilsSoils atat VariousVarious RangesRanges

0

1

2 Depth (cm) 3 TNT RDX 4 00.20.40.60.81 Normalized Concentration (depth value/top value) ToolsTools forfor MultiMulti--IncrementIncrement SamplingSampling SamplingSampling ExperimentExperiment forfor 105105--mmmm HowitzerHowitzer FiringFiring Point,Point, DonnellyDonnelly TrainingTraining Area,Area, AKAK

10-m X 10-m Grid SamplingSampling Objective:Objective: WhatWhat isis thethe meanmean concentrationconcentration ofof 2,42,4--DNTDNT inin aa 1010--mm XX 1010--mm areaarea atat aa 105105--mmmm HowitzerHowitzer FiringFiring Point?Point?

Comparison: Discrete samples collected in 1-m x 1-m cells within a 10-m X 10-m grid. 30-increment samples collected randomly throughout entire 10-m X 10-m grid.

Results: Discrete Multi-increment Number of Samples: 100 10 Minimum: 0.0007 mg/kg 0.60 mg/kg Maximum 6.4 mg/kg 1.35 mg/kg Mean: 1.1 mg/kg 0.94 mg/kg Standard Deviation: 1.2 mg/kg 0.24 mg/kg Median: 0.65 mg/kg 0.92 mg/kg Distribution: Skewed Gaussian NormalNormal ProbabilityProbability Plot:Plot: 100100 DiscreteDiscrete SamplesSamples

7 100 discrete samples from 10-m X 10-m area Maximum 6.4 mg/kg 6

5

4

3 2,4-DNT Conc. (µg/g) 2,4-DNT Conc. (mg/kg) 2

Mean 1.1 mg/kg 1 Minimum 0.0007 mg/kg Median 0.65 mg/kg 0 .01 .1 1 5 10 20 30 50 70 80 90 95 99 99.9 99.99

Percent NormalNormal ProbabilityProbability Plot:Plot: Ten 30-Increment Composite Samples 1.4

1.3

1.2

1.1

1

0.9 2,4-DNT Conc. (µg/g) 2,4-DNT Conc. (mg/kg)

0.8

0.7

0.6 .01 .1 1 5 10 20 30 50 70 80 90 95 99 99.9 99.99 Percent CentralCentral LimitLimit TheoremTheorem AppliedApplied toto MultiMulti--incrementincrement (MI)(MI) SamplingSampling

• As the number of individual increments in each sample gets “large enough,” the distribution of replicate samples can be approximated by a normal distribution, regardless of the shape of the distribution of individual increments.

• “Large“Large enough”enough” - For most populations, the number of increments must be at least 30. FiringFiring PointPoint -- 2,42,4--DNTDNT mg/kgmg/kg withinwithin 1010 xx 1010--mm areaarea atat 105105--mmmm howitzerhowitzer trainingtraining rangerange

100100 discretesdiscretes (1 (1 xx 1-m1-m subsub grids)grids) ProPro UCLUCL -- Discrete Discrete samplessamples N = 7* –– Mean Mean 1.11.1 N = 7* –Median–Median 0.650.65 1.6,1.6, 3.6,3.6, 5.4,5.4, 0.72**0.72**,, 1.11.1 –– Min Min 0.00070.0007 –– Max Max 6.46.4 N = 15* –– Pro Pro UCLUCL 1.31.3 N = 15* 2.6,2.6, 2.8,2.8, 2.5,2.5, 2.4,2.4, 2.42.4 Multi-incrementMulti-increment SamplesSamples –– 10, 10, 30-increment30-increment (random)(random) NN == 30*30* •• Pro Pro UCLUCL 1.11.1 1.6,1.6, 1.4,1.4, 1.1,1.1, 1.5,1.5, 1.51.5

–– 4, 4, 30-increment30-increment (random)(random) ** SetSet ofof randomrandom discretediscrete samplesample •Pro•Pro UCLUCL valuesvalues 1.3,1.3, 1.1,1.1, 1.1,1.1, 1.4,1.4, 1.11.1 **** falsefalse negativenegative FiringFiring PointPoint -- NGNG mg/kgmg/kg withinwithin 1010 xx 1010--mm areaarea atat LawLaw RocketRocket trainingtraining rangerange

100100 discretesdiscretes (1 (1 xx 1-m1-m subsub grids)grids) ProPro UCLUCL -- Discrete Discrete samplessamples NN == 7*7* –– Mean Mean 0.6450.645 0.82,0.82, 0.57,0.57, 0.320.32,, 1.2,1.2, 0.740.74 –– Median Median 0.403 0.403 –– Min Min 0.023 0.023 NN == 15*15* –– Max Max 3.373.37 0.74,0.74, 0.76,0.76, 0.69,0.69, 0.86,0.86, 0.680.68 –– Pro Pro UCLUCL 0.5620.562

NN == 30*30* Multi-incrementMulti-increment SamplesSamples 0.74,0.74, 0.65,0.65, 0.56,0.56, 0.540.54,, 0.930.93 –– 1, 1, 30-increment30-increment (random)(random) •0.80 •0.80 ** SetSet ofof randomrandom discretediscrete samplesample valuesvalues **** falsefalse negativenegative SamplingSampling ExperimentExperiment atat anan Artillery/MortarArtillery/Mortar RangeRange ImpactImpact AreaArea 1010--mm xx 1010--mm chunkchunk residueresidue “Hot“Hot Spot”Spot”

10 x 10 m

Cobble-size pieces of Composition B (60/40 RDX/TNT) RDXRDX ConcentrationsConcentrations (mg/kg)(mg/kg) inin aa 1010-- xx 1010--mm chunkchunk residueresidue “Hot“Hot spot”spot” areaarea onon artillery/mortarartillery/mortar impactimpact rangerange

17.1 1.27 0.829 0.908 10.9 4.44 0.437 0.354 1.52 0.067

0.805 24.1 7.73 0.539 0.260 0.233 0.366 1.93 0.731 0.138 • Hotspots two to three

30.8 1.40 12.5 0.342 0.074 1.11 0.18 0.076 7.11 0.187 meters in size

12.7 138 53.7 3.85 4.94 1.22 4.63 0.470 2.41 1.06

331 9.70 3.96 1.44 3.67 0.243 3.21 0.254 1.03 0.073 • Mean 71 mg/kg

7.52 5.65 1.97 0.571 4.84 19.9 0.825 0.122 1.46 0.070 • Loading: 300 g RDX 1.65 1.56 8.51 10.6 2.24 25.2 7.15 0.248 0.175 0.037 100 g chunks 48.3 13.3 3.36 6.93 889 21.8 3.75 0.618 0.193 0.081 200 g top 2.5 cm soil

1.18 1.03 64.3 557 1790 2390 11.3 1.65 0.335 0.263

8.86 3.50 5.02 42.7 385 24.9 3.64 0.96 0.526 0.161 10-m PatternPattern ofof multimulti--incrementincrement samplesample collectioncollection usingusing systematicsystematic--randomrandom samplingsampling designdesign

Path of travel Sample collection points SamplingSampling Objective:Objective: WhatWhat isis thethe meanmean concentrationconcentration ofof RDXRDX inin aa 1010--mm XX 1010--mm areaarea atat anan artillery/mortarartillery/mortar impactimpact area?area?

Comparison: Discrete samples collected in 1-m x 1-m cells within a 10-m X 10-m grid, 25-increment samples collected randomly within entire 10-m X 10-m grid. Results: Discrete Random-MI Sys.-Random-MI Simulated Number: 100 10 4 Minimum: 0.037mg/kg 4.6 mg/kg 33 mg/kg Maximum: 2390 mg/kg 290 mg/kg 100 mg/kg Mean: 71 mg/kg 54 mg/kg 71 mg/kg Std. Dev. 315 mg/kg 86 mg/kg 43 mg/kg Median: 1.8 mg/kg 25 mg/kg 75 mg/kg Distribution: Skewed Skewed ?? MultiMulti--IncrementIncrement SamplingSampling ExperimentExperiment atat anan Artillery/Mortar/BombingArtillery/Mortar/Bombing ImpactImpact RangeRange SamplingSampling Objective:Objective: WhatWhat isis thethe meanmean concentrationconcentration ofof RDXRDX inin aa 1010--mm XX 1010--mm areaarea atat anan artillery/mortar/bombingartillery/mortar/bombing impactimpact area?area?

Comparison: Three totally random 25-increment samples versus three 25-increment samples collected using a systematic-random design from within 10-m X 10-m area.

Results: Totally Random-MI Systematic-Random-MI Number: 3 3 Minimum: 0.24 mg/kg 0.31 mg/kg Max:0.78 mg/kg 0.37 mg/kg Mean: 0.49 mg/kg 0.34 mg/kg Std. Dev.: 0.27 mg/kg 0.033 mg/kg Median: 0.45 mg/kg 0.34 mg/kg ImpactImpact RangeRange -- RDXRDX mg/kgmg/kg withinwithin 1010 xx 1010--mm areaarea aroundaround lowlow--orderedordered 8181 mmmm mortarmortar

100100 discretesdiscretes (1(1 xx 1-m1-m subsub grids)grids) ProPro UCLUCL NN == 7*7* 4.6**, 2.7, 9.0, 1300, 12 –– Mean Mean 70.970.9 4.6**, 2.7, 9.0, 1300, 12 –– Median Median 1.791.79 NN == 15*15* – Min 0.0007 – Min 0.0007 320,320, 240,240, 1300,1300,4.0 4.0,, 13001300 –– Max Max 23902390 –– Pro Pro UCLUCL 96.496.4 NN == 30*30* 1100,1100, 49, 49, 310,310, 120,120, 410410 Multi-incrementMulti-increment SamplesSamples N = 50* –– 10, 10, 25-increment25-increment (random)(random) N = 50* 550,550, 240,240, 90,90, 130,130, 170170 •• Pro Pro UCLUCL 120120 –– 4, 4, 25-increment25-increment (systematic-(systematic- * Set of random discrete values random)random) * Set of random discrete values •• Pro Pro UCLUCL 110110 **** FalseFalse negativenegative RDXRDX ConcentrationsConcentrations (mg/kg)(mg/kg) In a 30- x 30-m area at an artillery/mortar impact range Fort Richardson, AK

0.0 0.05 0.02 0.0 0.0 0.0 0.16 0.44 0.86 6.89 100 Discrete samples Mean = 5.5 0.0 0.0 0.02 0.05 0.03 0.0 0.0 119 172 1.51 Median =

0.0 0.0 0.0 0.0 0.0 0.03 0.04 40.2 4.47 0.55 Max = 172

0.0 0.0 0.0 0.0 0.04 0.15 0.03 0.7 0.37 0.21

0.0 0.07 0.0 0.0 0.0 0.0 0.0 62.5 0.05 0.02 100-increment samples (n=4) 2.9, 5.6, 11.8, 8.5 0.0 0.03 0.0 0.0 0.0 0.0 0.0 0.25 0.03 0.04

0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.17 0.0 0.13 Concentration (mg/kg) 0.0 0.0 0.0 0.11 0.0 0.06 0.0 4.34 121 2.90 < d 0.0 0.0 0.0 0.03 0.0 0.0 0.0 0.21 0.07 0.46 0.01 – 1.0 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.07 0.14 6.10 1 – 10 30-m >10 ImpactImpact RangeRange -- RDXRDX mg/kgmg/kg withinwithin 3030 xx 3030--mm areaarea aroundaround lowlow--orderedordered 120120 mmmm mortarmortar

100100 discretesdiscretes (3 (3 xx 3-m3-m subsub grids)grids) ProPro UCLUCL NN == 7*7* –– Mean Mean 5.465.46 0.33**,0.33**, 0.11,0.11, 0.31,0.31, 0.64,0.64,55 55 –– Median Median 0.00.0 NN == 15*15* –Min 0.0 –Min 0.0 61,61, 43,43, 2.2, 2.2,58, 58, 2.8 2.8 –– Max Max 172172 –– Pro Pro UCLUCL 16.316.3 NN == 30*30* 41,41, 22,22,1.5, 1.5,28, 28, 1313 Multi-incrementMulti-increment SamplesSamples NN == 50*50* –– 4, 4, 100-increment100-increment 16,16, 27,27, 32,32, 1.21.2,, 1414 (systematic-random)(systematic-random) •• Pro Pro UCLUCL 11.911.9 ** SetSet ofof randomrandom discretediscrete valuesvalues **** FalseFalse negativenegative DemolitionDemolition ofof MissileMissile MotorsMotors DetonationDetonation (one(one minuteminute later)later) DetonationDetonation CraterCrater HMXHMX ConcentrationsConcentrations (mg/kg)(mg/kg) forfor 100100--mm xx 100100--mm gridsgrids

0.132

0.328 0.550 1.11

0.618 2.20 2.62 0.392 0.604

0.522 3.18 4.13 3.54 1.54

2.90 3.90 2.48 2.66 4.38

Red numbers indicate HMX 1.26 0.770 0.384 concentrations (ppm) based on 100 incremental samples/grid FieldField ReplicatesReplicates forfor MissileMissile DemolitionDemolition RangeRange (100(100--mm xx 100100--mm grids)grids)

Triplicate 100-increment samples using a systematic- random design for four 100-m X 100-m grids

HMX Concentration (mg/kg) Grid 3 Grid 6 Grid 18 Grid 31 0.146 <0.04 4.26 <0.04 0.126 <0.04 3.96 <0.04 0.124 <0.04 4.16 <0.04 Mean=0.132 Mean=4.13 RSD = 9.2% RSD = 3.7%

For this example, the MI sampling was done by a contractor. ReplicateReplicate MultiMulti--incrementincrement ResultsResults

O G D E N A I R L O G I S T I C S C E N T E R

HMX (ppm) HMX (ppm) STL CRREL STL CRREL 0.13 0.15 <0.1 <0.1 <0.1 0.13 <0.1 <0.1 <0.1 0.12 <0.1 <0.1

HMX (ppm) STL CRREL 3.9 4.3 3.4 4.0 3.7 4.2 DiscreteDiscrete SamplesSamples vsvs MultiMulti--IncrementIncrement SamplesSamples

• Large range in data for discrete samples – much smaller range in multi-increment samples • Distribution of discrete data always non normal – distribution of data from multi-increment samples often normal • Median from discrete sample almost always less than mean – thus most discrete samples underestimate mean • Median and mean concentrations generally similar for multi-increment samples • Use of multi-increment samples often eliminates problem of dealing with “less than” data • Recommendation – collect replicate systematic multi-increment samples to represent decision unit MajorMajor SamplingSampling IssuesIssues

• How many increments are necessary to overcome spatial heterogeneity within the decision unit? – Recommend 30 to 100 • How should increments be collected (totally random or with some systematic component)? – Recommend systematic-random sampling design • What should the sample mass be to overcome compositional heterogeneity? – Recommend a minimum of 1 kg • How large of an area can be adequately characterized with a multi-increment sample? – Experiments have been successful for areas as large as 100 m X 100 m (1 hectare) • What should the sampling depth be? – Experiments indicate most residues are present in top 5 cm OverallOverall RecommendationsRecommendations forfor SamplingSampling ofof RangesRanges

• Stratify ranges into functional areas • Divide functional areas into decision units of a hectare or less • Collect multi-increment samples from within each decision unit using a systematic-random pattern (n from 30 to 100) • Collect field replicate samples to provide estimate of total uncertainty • Do the job once and do it right!