Introduction to Site
Characterization
Dr. Bob Johnson
Argonne National Laboratory
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Life Cycle Site Decision-Making is Based on Data
 Are contaminants present in environmental media at
levels above background at a site?
 Do those contaminants pose unacceptable dose or
risk concerns?
 Which portions of a site require remediation?
 Are remedial actions performing as expected?
 Does the site present ongoing and immediate health
and safety issues?
 When can remediation stop, and are we confident that
residual risks/doses are at acceptable levels?
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
For Every Step of the Process, Data Inputs are Key
CERCLA (Comprehensive
Environmental Response,
Compensation and Liability Act)
• Discovery; Preliminary
Assessment (PA)
RCRA (Resource Conservation
and Recovery Act)
Result s
Samples
• Site Investigation (SI)
Result s
• Extended Site
Investigation (ESI)
Samples
• Remedial
Investigation/Feasibility
Study (RI/FS)
Samples
Result s
s
Sample
• RCRA Facility
Assessment (RFA)
s
Sample
• RCRA Facility
Investigation (RFI)
Results
Results
s
Sample
Result s
Samples
• Discovery
Results
Results
s
Sample
• Corrective Measures
Study (CMS)
• Remedial Action
• Corrective Measures
Implementation (CMI)
• Closure
• Closure
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Bad Data Lead to Bad Consequences
 Missing site-specific dose or health risks that should be
addressed
 Spending resources on remedial actions that are not
truly necessary from a risk or dose perspective
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Agenda for Day 2
 Background to hazardous site characterization
 MARSSIM - U.S. approach to demonstrating compliance with
dose standards for closure purposes
 U.S. EPA’s Triad approach to site characterization
 The role of systematic planning and conceptual site models in
life cycle site management
 Overview of measuring radionuclides in the environment
 Dynamic data collection strategies for site characterization
 Data management, integration, visualization, and
communication
 Field demonstration of direct measurement technologies for
radionuclides
June 2010 | Argonne National Laboratory, USA
Environmental Data
Collection Realities
Dr. Bob Johnson
Argonne National Laboratory
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Environmental Media Typically of Concern
 Soils – main focus of remaining presentations
 Sediments
 Groundwater
 Surface water
 Biota
 Air
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Typical Soil Characterization Goals
 Looking for the presence of contamination above some
threshold
– Checking against background levels
– Comparison to never-to-exceed cleanup criteria
– Searching for elevated areas of concern
 Estimating average concentrations over an area of
interest
– Developing exposure point concentrations (EPC) for
risk/dose assessments
– Comparison to area-averaged cleanup requirements
 For radionuclides, parameter of interest is usually activity
concentrations of radionuclides of concern
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
When More than One Radionculide of Concern is
Present, Sum of Ratios Becomes Important
 Individual radionuclides alone may not be at levels of dose
concern, but taken together, cumulative dose may be an
issue
 Captured by sum of ratios (SOR) calculation:
– Concen1/Acceptable Concen1 + Concen2/Acceptable Concen2 +…
– Needs to be less than one, otherwise dose concerns
 Careful attention needs to be paid to whether the
acceptable concentration derived for a contaminant of
concern accounted for daughters in secular equilibrium or
not (e.g., Ra-226 decay chain)
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Soil Activity Concentration Estimates Can Come
From Several Sources
 Discrete soil sample and
subsequent analysis of
sample
 Composite soil sample and
subsequent analysis of
sample
 Direct in situ
measurements (mobile or
stationary)
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Fundamental Concepts for Sampling Design
and Statistics
 Decision Unit (DU): Area, volume, or set of objects
treated as single unit for decision-making
– Such as 2,000 m2 area, bin of soil, set of drums
– Examples: exposure units, survey units, remediation units…
 Sample: A portion of a population/decision unit
collected to characterize a population/decision unit
parameter of interest
 Sample Support: Physical dimensions and
characteristics of a (sub)sample
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
More Concepts
 Representativeness:
– Degree to which a sample reflects original population
in context of decision
– Ability to confidently extrapolate concentration results
from a tiny sample to represent the concentration of
the much larger volume of soil (area of inference)
from whence it came
– Example:
• MARSSIM FSS unit – more than 300 metric tonnes of soil
• Typical discrete sample – 400 grams of soil
• Typical alpha spec sub-sample – a few grams of soil
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Heterogeneity/Variability
 Heterogeneity: Variations throughout an area
or volume as observed in sample results
 Variability: Variations in measured
concentrations observed in (sub)sample results
– Within-sample heterogeneity
– Short-scale between-sample
heterogeneity (can affect agreement
between co-located samples)
– Long-scale between-sample
heterogeneity (on scale of
*
conventional distances
between samples)
*
June 2010 | Argonne National Laboratory, USA
7
6
1
5
*
*
*
*
*
*
*
*
*
*
2
3
4
ENVIRONET Environmental Remediation Training Course
Sample Support Determines the Range of
Variability Observed within a Decision Unit
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Decision Quality Only as Good as the Weakest
Link in the Data Quality Chain
Sampling
Sampling
Design
Sample
Support
SubSampling
Sample
Preservation
Analysis
Extract Cleanup
Method
Sample Prep
Method
Interpretation
Result
Reporting
Determinative
Method
Relationship between
Measurement Parameter
& Decision Parameter
Each link represents a variable contributing toward the
quality of the analytical result. All links in the data quality
chain must be intact for data to be of decision-making quality!
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Basic Statistical Terms
 Mean: Average concentration for a given decision
unit
 Median: Concentration at which half of a decision
unit would be below and half above
 Range: Concentration interval defined by the
minimum and maximum concentration values
 Variance: A measure of the “spread” of
concentration values for a set of samples or
measurements
 Standard Deviation: Square root of the variance
(continued)
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
What Contributes to Observed Variability in
Sample/Measurement Results?
 Natural heterogeneity
 Sample preparation/homogenization OR
geometry/environmental variations for in situ
measurements
 Measurement error
 All three of these are a function of concentration (i.e.,
all three are at a minimum when concentrations are
at background levels, but grow as concentration
levels in an area increase)
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Soil Heterogeneity at the “Within-sample” or
“Micro-scale”
Firing Range Soil Grain Size
(Std Sieve Mesh Size)
Pb Concentration in fraction by
AA (mg/kg)
Greater than 3/8” (0.375”)
10
Between 3/8” and 4-mesh
50
Between 4- and 10-mesh
108
Between 10- and 50-mesh
165
Between 50- and 200-mesh
836
Less than 200-mesh
1,970
Bulk Total
927
(wt-averaged)
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Short-Scale Heterogeneity Can Be Significant:
Arsenic in Samples from 3 Residential Yards
1 ft apart over 4 ft
As 129 221 61
Linear even spread
over 6 ft
As
37 290 625
94
Spread evenly over 7 ft
As 17
41
367 351 268
June 2010 | Argonne National Laboratory, USA
39 14
Same yard, 8 ft away from
group to left & spread over 6 ft
As
27
29
45
34
Same yard, 15 ft away from
group to left & spread over 4 ft
As 29
24
79
120
ENVIRONET Environmental Remediation Training Course
Short-Scale Variability Can Be Significant:
Uranium in Soils
Loc 13-1-10
Vertical Total
U Distributions
1-ft2
Uranium over
surface area
0
2
6
10
0
49 ppm
113 ppm
2
6
10
Depth (in)
0
496 ppm
2
Background conditions
6
10
0
2
6
10
0
2
6
30 ppm
116 ppm
10
0
200
400
600
Total U (ppm)
June 2010 | Argonne National Laboratory, USA
800
1000
1200
ENVIRONET Environmental Remediation Training Course
Short-Scale Variability Can Be Significant: Explosives in
Range Soils
416 ppm
2
286 ppm
7
41,400 ppm
1,220 ppm 6
3 136 ppm
1
2 ft
5
27,700 ppm
June 2010 | Argonne National Laboratory, USA
4
Figure adapted from
Jenkins (CRREL), 1996
42,800 ppm
ENVIRONET Environmental Remediation Training Course
Heterogeneity Overwhelms Variability from Different
Analytical Techniques
331 On-site
286 Lab
1,280 On-site
1,220 Lab
500 On-site
416 Lab
2
7
95% of data variability
due to sample location
over a 4 ft diameter
39,800 On-site
41,400 Lab
6
24,400 On-site
27,700 Lab
June 2010 | Argonne National Laboratory, USA
1
3
2 ft
5
4
164 On-site
136 Lab
27,800 On-site
42,800 Lab
ENVIRONET Environmental Remediation Training Course
The Biggest Cause of Bad Decisions
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
The Relationship between Analytical & Sampling
Uncertainties
Uncertainties add according to (a2 + b2 = c2)
Analytical Uncertainty (AU)
Total Uncertainty (TU)
Sampling Uncertainty (SU)
Examples:
• AU = 10 ppm, SU = 80 ppm: TU = 81 ppm
• AU = 5 ppm, SU = 80 ppm: TU = 80 ppm
• AU = 10 ppm, SU = 40 ppm: TU = 41 ppm
• AU = 20 ppm, SU = 40 ppm: TU = 45 ppm
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Historically, Focus Has Been Analytical Quality
 Emphasis on fixed-base laboratory analyses of all
samples following well-defined protocols
 Analytical costs driven to a large degree by QA/QC
requirements
 Sampling costs driven by analytical costs
 Result:
– Analytical error typically on order of +/-30% or less for replicate
analyses
– Analytical costs constrain the number of samples collected
– Traditional laboratory data treated as “definitive” – but definitive
(definite) about what?
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
How Do We Reduce Data Uncertainty?
 For analytical errors:
– Modify current technique or choose a different
analytical technique
– Improve QC on existing techniques
 For sample prep and handling errors:
– Improve sample preparation
 For sampling errors:
– Collect samples from more locations!
– Composite sampling is one cost-effective way to do
this
June 2010 | Argonne National Laboratory, USA
MARSSIM and Site
Closure
Dr. Bob Johnson
Argonne National Laboratory
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Closure Data Collection Demonstrates Dose/Risk
Standards Have Been Met
 In the United States, closure data collection protocols
described in MARSSIM
 MARSSIM: Multi-Agency Radiation Survey & Site
Investigation Manual
 Multi-agency consensus on closure process for
radioactively contaminated sites
 A flexible, overall framework for addressing radioactively
contaminated sites
 Technically defensible techniques for demonstrating
compliance
 Performance-based approaches for demonstrating
compliance
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Within MARSSIM’s Scope
MARSSIM’s closure scope includes sites:
 That have radionuclide contaminants
 With cleanup requirements that are originally risk- or dosebased that have been translated into activity concentration
guidelines
 That are applied to the surfaces of soils or structures
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Beyond MARSSIM’s Scope






Subsurface soils and sediments
Chemical contamination
Ground or surface waters
Building materials other than surfaces
Waste acceptance/disposal criteria
Translating dose or risk into concentration-based
guidelines
 Remedial alternative recommendations or
evaluations
 Stakeholder involvement
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
MARSSIM References
 U.S. Department of Energy:
www.etec.energy.gov/Cleanup/Documents/Radiation_Cle
anup_Standards/MARSSIM.pdf
 U.S. Nuclear Regulatory Commission:
http://www.nrc.gov/reading-rm/doccollections/nuregs/staff/sr1575/
 U.S. Environmental Protection Agency:
http://www.epa.gov/rpdweb00/marssim/
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
MARSSIM Life Cycle
 Site Identification
 Historical Site Assessment
 Scoping Survey
 Characterization Survey
 Remedial Action Support Survey
 Final Status Survey (FSS)
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Derived Concentration Guideline Levels
 Derived concentration guideline levels (DCGLs)
refer to the cleanup criteria that a site must meet
 Called derived because they are derived out of
basic dose or risk goals for the site
 Usually incremental to background
 Typically posed as activity concentrations (e.g.,
picoCuries per gram, or pCi/g) that apply to an area
of specified size
 MARSSIM assumes there will be two: a DCGLw and
a DCGLemc
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
DCGL Derivation
Site-Specific Risk or
Dose-Based
Requirements
Plant
Foods
Milk
Meat
Dust,
Radon
Soil
Ingestion
Infiltration
External
Radioactively Contaminated Material in Soil
Leaching
Groundwater
June 2010 | Argonne National Laboratory, USA
Drinking
Water
Fish
Surface
Water
ENVIRONET Environmental Remediation Training Course
DCGLw
 The DCGLw represents a wide-area average
goal. Applied to areas that are the size of
survey units.
 Average goal is an important concept. This
means that if one were to sample an area that
complied with a DCGLw, we could tolerate some
samples above the DCGLw as long as the
average was below.
 For soils, MARSSIM assumes one will use
discrete samples and statistics to show that
survey units comply with the DCGLw.
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
DCGLemc
 The DCGLemc represents an elevated small area
average.
 More commonly known as a “hot spot”.
 By definition, the DCGLemc > DCGLw, while the area
associated with the DCGLemc is smaller than a survey
unit.
 MARSSIM assumes scans will be used to identify the
presence or absence of elevated areas, with discrete
sampling a fall-back option.
 MARSSIM assumes these are primarily concerns in
Class 1 final status survey units.
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Reference Areas
 “Background” area where radionuclide
concentrations are believed to be at natural levels.
 Used as a point of comparison by MARSSIM
statistical techniques for situations in which the
DCGLw is close to background and radionuclides
are naturally occurring.
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Graded Approach to Data Collection
 Target the level of data collection to match the
likelihood of compliance problems
Low Contamination
Potential
High Contamination
Potential
Less Data Collection
More Data Collection
 MARSSIM captures this through the definition of
survey units
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Final Status Survey (FSS) Units
 Site is broken into FSS units
 Three types of survey units: Class 1 units, Class 2
units and Class 3 units
 Breakdown of an area into these classes based on
contamination potential
 Class 1 units have a greater density of data
collection than Class 2 units, and Class 2 units
have a greater density of data collection than
Class 3 units
 FSS decision-making is done on a unit-by-unit basis
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Example FSS Unit Layout
Estimated Excavation Footprint
based on ROD Criteria
Preliminary Class I
Unit (2,000 m2)
June 2010 | Argonne National Laboratory, USA
Preliminary Outer Boundary
of Class II Units
Preliminary Outer Boundary
of Class III Units
ROD = Record of Decision
ENVIRONET Environmental Remediation Training Course
MARSSIM Statistical Tests
 Used for evaluating DCGLw compliance
 Parametric versus non-parametric tests
 Wilcoxon Rank Sum Test
– When DCGLw is close to background levels
– Compares sample results from a survey unit to
results from a reference unit
 Sign Test
– When radionuclide is not in background, or when
DCGLw is much higher than background
– Looks at the number of samples that exceed the
DCGLw
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Common MARSSIM Implementation Issues
 Subsurface soil/sediment contamination
 Radionuclides mixed with other chemical
contamination
 Long lists of suspect radionuclides
 Determining when radionuclides are in secular
equilibrium
 Clearing soil piles for re-use
 Working with promulgated standards that don’t fit
MARSSIM’s DCGLw and DCGLemc concepts
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Typical MARSSIM FSS Process
Establish DCGL
Requirement
Remediate as
Necessary
Lay Out
FSS Units
Scan Each Unit
for DCGLemc Issues
Sample Each Unit
for DCGLw Issues
June 2010 | Argonne National Laboratory, USA
Done
ENVIRONET Environmental Remediation Training Course
Simple MARSSIM Example:
Michigan NORM Site
 Approx. 3-acre site used for
pipe storage
 Contaminated with NORM
 Remediated in the early ’90s;
 Covered under State of
Michigan rad guidance
(radium-226, 5 pCi/g over
100 square meters)
 Owner wants to sell land and
so needs closure
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Relevant DCGL Requirements
 DCGLemc
– 5 pCi/g averaged over 100 m2 area
– Evaluated using scans
 DCGLw
– 5 pCi/g averaged over a final status survey unit
– Evaluated using direct measurements and
statistics
– The practical effect of using statistics is that the
final average residual activity concentration has
to be significantly less than 5 pCi/g
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Layout of FSS Units
 Class 1 units cover area
known to have been
contaminated
 Rest of property is a Class
2 unit
 No Class 3 unit(s) for this
site because adjacent
properties are not
accessible
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Implementing the Final Status Survey Plan
• After an initial walkover of
the site, some remaining
DCGLemc concerns were
identified that were
addressed with spot removal
• Pre-FSS sampling, the site
still exhibited some evidence
of individual locations above
5 pCi/g, but it was believed
to be compliant with DCGL
specifications
June 2010 | Argonne National Laboratory, USA
ENVIRONET Environmental Remediation Training Course
Sampling/Measuring for DCGLw
Requirements
 DCGLw requirements were
established using direct gamma
spec measurements
 Nine locations were allocated to
each survey unit using a regular
grid
 The results from these readings
were used to calculate averages
for each FSS unit and to perform
statistics
June 2010 | Argonne National Laboratory, USA