c/EPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/4-90/017
July 1990
Direct/Delayed Response
Project: Laboratory
Operations and
Quality Assurance
Report for Preparation of
Soils from the
Mid-Appalachian
Region of the
United States
-------
EPA/600/4-90/017
July 1990
Direct/Delayed Response Project:
Laboratory Operations and Quality Assurance
Report for Preparation of Soils from the
Mid-Appalachian Region of the United States
by
M.L. Papp and R.D. Van Remortel
A Contribution to the
National Acid Precipitation Assessment Program
OILS-- KwJroni»ntal Protection Agency
^S^on 5, Library (cPL-16)
5-SQ' C_ Scat-barn Street, Room
1^" 6Q6G4
U.S. Environmental Protection Agency
Ofiice of Research and Development
Washington, DC 20460
Environmental Monitoring Systems Laboratory, Las Vegas, Nevada 89193
Environmental Research Laboratory, Corvallis, Oregon 97333
-------
Notice
The information in this document has been funded wholly or in part by the U.S. Environmental
Protection Agency under Contract Number 68-03-3249 to Lockheed Engineering & Sciences Company.
Additional cooperation has been provided under Contract Number 68-03-3246 to NSI Technology
Services Corporation. It has been subject to the Agency's peer and administrative review and it has
been approved for publication as an EPA document.
Mention of corporation names, trade names, or commercial pioducts does not constitute
endorsement or recommendation for use.
This document is one volume of a set which fully describes the Direct/Delayed Response
Project, Mid-Appalachian Soil Survey. The complete document set includes the major data report,
quality assurance plan, laboratory analysis handbook, field and preparation laboratory operations
and quality assurance reports, and analytical laboratory quality assurance report. Similar sets are
being produced for each Aquatic Effects Research Program component project. Colored covers,
artwork, and the use of the project name in the document title serve to identify each companion
document set.
The authors of this document are affiliated with the Lockheed Engineering & Sciences
Company, Las Vegas, Nevada. The correct citation of the document is:
Papp, M. L., and R. D. Van Remortel. 1990. Direct/Delayed Response Project: Laboratory Operations
and Quality Assurance Report for Preparation of Soils from the Mid-Appalachian Region of the
United States. EPA/600/x-xx/xxx U.S. Environmental Protection Agency, Las Vegas, Nevada.
183 pp.
-------
Abstract
The Direct/Delayed Response Project Soil Survey includes the mapping, characterization,
sampling, preparation, and analysis of soils in order to assess watershed response to acidic
deposition within various regions of the United States. Soil samples from the Mid-Appalachian
region were transported to a preparation laboratory for processing before delivery to analytical
laboratories. This document summarizes procedural, operational and data quality compliance with
the protocols used at the preparation laboratory. Deviations from the protocols and difficulties
encountered are identified and discussed. Estimates of uncertainty in the laboratory data are
made, and the relation of measurement uncertainty to overall uncertainty in the routine data is
assessed. Recommendations are made for program improvement.
A review of the soils data suggests that the integrity of the soil samples was maintained
during the preparation activities. In most cases, laboratory personnel adhered to the protocols.
Generally, measurement uncertainty is a very small part of the overall data uncertainty.
This report was submitted in partial fulfillment of contract number 68-03-3249 by Lockheed
Engineering & Sciences Company under the sponsorship of the U.S. Environmental Protection
Agency. The report covers a period from September, 1988 to April, 1989.
in
-------
Contents
Section
Notice [[[ ii
Abstract ...................... . ........................................ iii
Figures ............................................... . ........ . . , ..... vi
Tables ............................. , ............ , .......... . ........... vi
Acknowledgments ................... . .................................... vii
Abbreviations and Acronymns .............................................. viii
1 Introduction .................................................. . ...... 1
Mid-Appalachian Soil Survey ............................................ 2
Laboratory Objectives ................................................. 4
Organization of the Report ............................................. 4
2. Preparation Laboratory Operations ........................................ 5
Laboratory Design [[[ 5
Personnel . ................. . ....................................... 5
Equipment Disbursement ................. . ............... . ...... ...... 5
Sample Receipt [[[ 7
Sample Storage [[[ 9
Sample Drying .................................... , ................ . 9
Determination of Field-Moist pH (PH_MP) ................................ . . 9
Determination of Organic Matter by Loss on Ignition (OM_LOI) ..... ..... . ....... 10
Determination of Air-Dry Moisture (MOIST_P) .................. . ....... . ..... 10
Sample Disaggregation and Sieving ...................................... 10
Determination of Rock Fragment Content (RFJMG and RF_FG) ...... . ............ 11
Sample Homogenization and Subsampling ................................. 11
Determination of Bulk Density .................... . .................... . . 11
Ciod Method (BD_CLD) .............................................. 11
Known Volume Methods (BD_KV) ................. . ..................... 12
Sample Shipment ........................................... . ........ 12
Sample Archiving [[[ 13
Record Keeping [[[ 13
Data Entry [[[ 13
Sample Tracking [[[ 13
Progress Reporting [[[ 14
-------
(continued)
Sect/on Page
Special Studies , 16
Data Quality Assessment , 17
Quality Attributes . 17
Measurement Quality Objectives 17
Uncertainty Estimation for Laboratory Analyses . , 18
Measurer-tent Quality Samples 21
Technical Systems Audits 23
Data Verification , . 24
4. Results arid Discussion 29
Evaluation of Quality Attributes 29
Precision 29
Accuracy 32
Representativesos 32
Completeness 32
Comparability . , 33
Assessment of Data Uncertainty Components 33
Additive Components of Uncertainty 33
Effect of Measurement System on Data Uncertainty 34
5. Conclusions and Recommendations 36
Laboratory Operations 36
Laboratory Design 36
Equipment and Supplies 36
Sample Receipt 36
Sample Drying , , . , 36
Field-Moist pH , 37
Organic Matter by Loss on Ignition 37
Disaggregation and Sieving 37
Bulk Density 37
Sample Batching 37
Sample Archiving , 37
Quality Assurance Program , 38
Training 38
Measurement Quality Objectives 38
Technical Systems Audits 38
Data Verification . 38
References 39
Appendices
A. Preparation Laboratory Standard Operating Procedures 41
B. Preparation Laboi-itojy Data Forms 99
C. Preparation Laboratory Computer Entry/Tracking/Verincation Procedures ........ 110
D. Comparison of Soil Sample Hornogenization Techniques 118
E. Determination of Dust Constituents at the Preparation Laboratory Facility 137
F. Kraft Paper and Nylon Mesh Analysis 143
G. Technical Systems Audit Reports 152
H. Quality Assurance Reanalysis Templates 173
I. Laboratory Quality Control Charts 177
-------
Figures
Number Page
1 Layout of the Mid-Appalachian Soil Survey Preparation Laboratory 6
Tables
Number Page
1-1 Analytical Parameters Measured at the Preparation Laboratory 8
2-1 Contents of Initial Packages of Equipment and Consumable
Supplies Sent to Sampling Crews 7
2-2 List of Equipment and Consumable Supplies Used at the
Preparation Laboratory 8
2-3 Particle Density Values for Selected Parent Materials 12
3-1 Preparation Laboratory Within-Run Measurement Quality Objectives 19
3-2 Primary Horizon Types for Sampling Class/Horizon Groups 21
3-3 Samples Placed in Mineral and Organic Soil Batches 22
3-4 Summary of Achievements for Acceptance Criteria of PH_MP 26
3-5 Summary of Achievements for Acceptance Criteria of OM_LOI 27
3-6 Summary of Achievements for Acceptance Criteria of MOIST_P 28
4-1 Imprecision estimates for PH_MP 29
4-2 Imprecision estimates for OMJ.OI 30
4-3 Imprecision estimates for MOIST_P 30
4-4 Imprecision estimates for RF_FG and RF_MG 30
4-5 Imprecision estimates for BD_CLD and BD_KV 31
4-6 Multiple Survey Comparison of Imprecision for Rock Fragments
and Clod Bulk Density 33
4-7 Comparison of Measurement Uncertainty Components 34
4-8 Delta Values and Ratios for Assessment of Data Uncertainty Components 34
VI
-------
Ackno wledgments
External peer reviews by the following individuals are gratefully acknowledged: L K.
Fenstermaker, Environmental Research Center, University of Nevada, Las Vegas, Nevada; and
K. A. Dwire, NSI Technology Services Corporation, Corvallis, Oregon.
The authors wish to acknowledge the following individuals for their technical assistance
during the operations and data assessment phases of the preparation laboratory activities: J. J.
Lee, U.S. Environmental Protection Agency, Corvallis, Oregon; D. G. Easterly and S. C. Black, U.S.
Environmental Protection Agency, Las Vegas, Nevada; D. S. Coffey and J. S. Kern, NSI Technology
Services Corporation, Corvallis, Oregon; M. H. Bartling, Martin Marietta Corporation, Oak Ridge,
Tennessee; J. V. Burton, Woodward & Clyde Consultants, Wayne, New Jersey; and G. E. Byers,
J. E. Teberg, R. L. Tidwell, K. C. Kohorst, G. D. Merritt, G. E. Satterwhite, D. A. Jadhon, J. R. Parolini,
D. L Peres, M. E. Silverstein, S. O. Offemson, A. D. Tansey, B. A. Schumacher, and M. J. Miah,
Lockheed Engineering & Sciences Company, Las Vegas, Nevada. We appreciate contributions to
the project overview of this report by the DDRP Technical Director, M. R. Church, U.S. Environmental
Protection Agency, Corvallis, Oregon.
The following individuals provided editorial and graphics support and are gratefully
acknowledged: B. N. Cordova, A. M. Tippett, K. M. Peres, and L. A. Stanley, Lockheed Engineering
& Sciences Company, Las Vegas, Nevada.
Finally, we appreciate the support of our technical monitors, Louis J. Blume and Daniel T.
Heggem, during the course of this survey.
VII
-------
Acronymns and Abbreviations
AERP
BD_CLD
BD_KV
DDRP
EMSL-LV
EPA
ERL-C
FAL
FAO
FAP
kg
LAL
LAO
LAP
MASS
MOIST_P
MQO
NAPAP
NIST
OM_LOI
PH_MP
PLSOP
PVC
QA
QART
QC
QCAS
QCCS
QE
RF_FG
RF_MG
RSD
SCS
SD
set ID
S/H
USDA
Aquatic Effects Research Program
bulk density using clod samples
bulk density using known volume samples
Direct/Delayed Response Project
Environmental Monitoring Systems Laboratory - Las Vegas, Nevada
U.S. Environmental Protection Agency
Environmental Research Laboratory - Corvallis, Oregon
low-range field audit sample
organic field audit sample
mineral field audit sample
kilogram
low-range laboratory audit sample
organic laboratory audit sample
mineral laboratory audit sample
Mid-Appalachian Soil Survey
air-dry moisture
measurement quality objective
National Acid Precipitation Assessment Program
National Institute of Standards and Technology
organic matter by loss on ignition
field-moist pH in deionized water
preparation laboratory standard operating procedures
polyvinyl chloride
quality assurance
quality assurance reanalysis template
quality control
quality control audit sample
quality control check sample
quality evaluation
fine gravel rock fragments
medium gravel rock fragments
relative standard deviation
Soil Conservation Service
standard deviation
set identification code
sampling class/horizon group
United States Department of Agriculture
VIII
-------
Section 1
Introduction
The Direct/Delayed Response Project
(DDRP) is an integral part of the Aquatic
Effects Research Program (AERP) of the U.S.
Environmental Protection Agency (EPA). The
DDRP is administered under the federally-
mandated National Acid Precipitation Assess-
ment Program (NAPAP) and addresses con-
cerns relating to potential acidification of
surface waters by atmospheric deposition
within the eastern United States. Given that
acidification of some surface waters has
occurred, critical scientific and policy questions
focus on whether acidification is continuing in
the regions of concern, whether it is just
beginning in other regions, how extensive the
effects might become, and over what time
scales effects might occur. The DDRP was
designed as the terrestrial complement to the
aquatic resources program and includes soil
surveys encompassing the Northeastern re-
gion, the Southern Blue Ridge Province, and
the Mid-Appalachian region of the United
States (Church et al., 1989). The project draws
its name from the consideration of whether
acidification might be immediate, i.e., "direct",
or would lag in time, i.e., "delayed", because of
soil buffering characteristics.
The overall purpose of the DDRP is to
characterize three geographic regions of the
United States by predicting the long-term
response of watersheds and surface waters to
acidic deposition. The four specific goals of
the DDRP are to: (1) characterize the variability
of soil and watershed attributes across these
regions, (2) determine which soil and water-
shed characteristics are most strongly related
to surface-water chemistry, (3) estimate the
relative importance of key watershed proces-
ses in controlling surface water chemistry
across the regions of concern, and (4) classify
the sample of watersheds with regard to their
response characteristics and extrapolate the
results from the sample of watersheds to the
regions of concern.
The DDRP watersheds were selected as
a subset of lake and stream systems surveyed
in the Eastern Lakes Survey and National
Stream Survey (Church, 1989) and were char-
acterized as probability samples to ensure that
results could be extrapolated to a specified
target population. A series of maps of soils,
vegetation, land use, and depth to bedrock
were prepared for each DDRP watershed by
the United States Department of Agriculture
(USDA) Soil Conservation Service (SCS). Soil
sampling classes were defined for each DDRP
region, and soils selected from these classes
were sampled and analyzed for physical and
chemical characteristics (Lee et al., 1989).
Soils were aggregated within sampling classes
to develop class means and variances used in
representing, e.g., by mass or area weighting,
characteristics of the watersheds of interest.
A variety of complementary data ana-
lyses are being performed as part of the
DDRP, including the statistical evaluation of
interrelationships among atmospheric deposi-
tion, mapped watershed characteristics, soil
chemistry, and current surface water chem-
istry. The principal goal of these analyses is
to verify that the processes and relationships
incorporated in the subsequent modeling
analyses reasonably represent the systems
under study.
Recent trend analyses have indicated
that the rate of sulfur deposition is slowly
declining in the Northeastern United States but
is increasing in the Southeastern United States
(Church et al., 1989). If a "direct" response
exists between sulfur deposition and surface
water alkalinity, the extent of current effects
on surface water probably would not change
-------
much at current levels of deposition, and
conditions would improve as the levels of
deposition decline. If surface water chemistry
changes in a "delayed" manner, e.g., due to
slow changes in watershed chemistry, future
changes (even with current or declining rates
of deposition) become difficult to predict. This
range of potential effects has implications for
public policy decisions on sulfur emissions
control strategies.
The three DDRP surveys focused on
regions of the United States that were iden-
tified as potentially sensitive to surface water
acidification. The Northeastern Soil Survey
was conducted in 1985 in the states of Maine,
New Hampshire, Vermont, Massachusetts,
Connecticut, Rhode Island, New York, and
portions of Pennsylvania. The Southern Blue
Ridge Province Soil Survey was conducted in
1986 in portions of Virginia, Tennessee, North
Carolina, South Carolina, and Georgia. The
Mid-Appalachian Soil Survey (MASS) was
conducted in 1988 in portions of Pennsylvania,
Virginia, and West Virginia.
Mid-Appalachian Soil Survey
The MASS focused on watersheds within
the states of Pennsylvania, Virginia, and West
Virginia. The streams in this region were
sampled as part of the National Stream Sur-
vey, Phase I, which is an EPA program de-
signed to evaluate stream chemistry in regions
of the United States believed to be susceptible
to the effects of acidic deposition. A sampling
design was applied to allow for unbiased
characterization of Mid-Appalachian regional
populations, which resulted in the selection of
36 watersheds for detailed study.
Soil types that were identified in the Mid-
Appalachian sampling region were combined
into sampling classes which are either known
to have or are believed to have similar chemi-
cal and physical characteristics. Each of the
sampling classes was then sampled across a
number of different watersheds in which the
classes occur. Using this approach, a given
soil sample does not just represent the spec-
ific watershed from which it came. Rather,
each sample contributes to a set of samples
that collectively represent a specific sampling
class for all DDRP watersheds within the
sampling region (Lee et al., 1989).
The DDRP technical director at ERL-C
had overall responsibility for the MASS. There
were five sampling crews, each consisting of
three to four soil scientists, involved in the
sampling phase. In each of the 150 soil
pedons selected for sampling, all soil horizons
were described on a coded field data form. A
total of 844 soil horizons from these pedons
were sampled in accordance with the sampling
protocols. In addition to collecting 5-kilogram
routine soil samples, each sampling crew
collected one duplicate sample for QA pur-
poses from every third pedon sampled. The
sampling protocols and a detailed account of
the soil mapping and sampling activities are
contained in a separate field operations and
QA report (Kern and Lee, in press).
As part of the MASS, a preparation
laboratory was established to process soil
samples collected from the Mid-Appalachian
region and to perform preliminary analyses on
these samples. The laboratory was located at
the Lockheed Engineering & Sciences Company
warehouse facility in close proximity to the
EPA Environmental Monitoring Systems Labor-
atory in Las Vegas, Nevada (EMSL-LV) to
ensure the rapid turnaround of samples and to
promote real-time data quality evaluation. The
preparation laboratory was designed to be the
link between the sampling crews and the
analytical laboratories, and served as a dis-
tribution point for equipment and supplies
used by the sampling crews.
All participants were required to comply
with specified protocols documented in the
Preparation Laboratory Standard Operating
Procedures (PLSOP) manual (see Appendix A).
The laboratory was supervised by a laboratory
manager, who was an experienced soil scien-
tist with an advanced university degree in
soils. The manager was responsible for
ensuring that sample integrity was maintained
after the bulk field samples were received by
the preparation laboratory. The laboratory
analysts had previous laboratory experience
and most held a university degree in the biolo-
gical or earth sciences. The laboratory staff
was responsible for tracking soil samples
throughout processing, performing soil
analyses, preparing samples for shipment to
analytical laboratories, and documenting all
preparation activities.
-------
The four types of soil samples sent to
the preparation laboratory during the field
sampling phase of the MASS were:
• Bulk samples, those soil samples col-
lected from specific soil horizons for
each sampling class. Both mineral and
organic samples were collected; organic
soils were defined as having 20 percent
or more organic matter as determined by
loss on ignition. A bulk sample con-
tained approximately four liters of soil
(eight liters for organic soils) and
generally weighed about 5 kilograms.
The bulk samples were collected in large
plastic bags and placed within protective
outer canvas bags. Each bulk sample
arrived at the laboratory with a com-
pleted DDRP Label A (see Appendix B)
affixed to the inner plastic bag.
• Field audit samples were measurement
quality samples sent by the EMSL-LV QA
staff to the sampling crews to include
with the soils sampled from every third
pedon by each crew. The samples were
sieved, packaged, labeled, and shipped
in the same manner as their associated
routine bulk samples.
• Clod samples were fist-sized, structurally
intact bulk density samples collected in
triplicate, where possible, from desig-
nated horizons. Each clod was wrapped
in a hairnet and dipped briefly in a 1:5
saran:acetone solution to help maintain
the clod structure and reduce moisture
loss from the clod during transport and
storage. The clod was labeled and then
covered by a small plastic bag and
placed in a special clod box with dividers
and cushioned packing material.
• Known volume bulk density samples
were collected in duplicate, where pos-
sible, from horizons where clod samples
were unobtainable. There were two
types of known volume samples col-
lected in the MASS: volume-replacement
and volume-fill samples. These bulk
density samples were packaged in small
pre-labeled plastic bags in much the
same manner as the bulk soil samples.
The laboratory staff performed several
soil analyses on the samples, including deter-
minations of field-moist pH in deionized water,
bulk density by clod and known volume meth-
ods, and gravimetric measurements of percent
air-dry moisture, percent rock fragments of fine
and medium gravel size, and percent organic
matter by loss on ignition. After the samples
were processed and analyzed, homogenization
and subsampling occurred and the resulting
analytical samples were assembled into batch-
es and shipped via overnight courier to desig-
nated analytical laboratories for further
analysis.
An important function of batch formation
was to place measurement quality samples in
the batches without revealing their identity to
the analytical laboratories. The measurement
quality samples were intended to provide a
way to both control and assess data quality
within each batch and between batches. Both
mineral and organic soil batches were
assembled, each containing a field duplicate
sample, two preparation duplicate samples,
seven natural soil audit samples, and
approximately thirty routine samples. The
analytical samples were organized according
to their respective pedon/set codes and then
randomized within each batch by the labora-
tory manager. The samples were labeled only
with batch and sample numbers, and the
batches were subsequently distributed among
three analytical laboratories contracted by EPA
to perform physical and chemical analyses
(Byers et al., in press).
All information pertaining to sample
receipt, tracking, preparation, and analysis
activities was placed on raw data forms and
entered into a computer entry and verification
program. A personal computer was used to
track samples, evaluate laboratory progress,
calculate final data values, and identify dis-
crepancies in the data.
An intensive quality assurance (QA)
program was used at the preparation labora-
tory to maintain consistency in the application
of soil preparation protocols and to ensure
that the soil sample analyses would yield
results of known quality. Laboratory personnel
received hands-on training in the protocols and
analytical methods. In addition, representa-
tives of the QA staffs at EMSL-LV and the EPA
Environmental Research Laboratory at
Corvallis, Oregon (ERL-C) conducted on-site
technical systems audits of the preparation
laboratory. Daily communication between the
QA staff and laboratory personnel was
established to identify, discuss, and resolve
issues. Control charts were also developed by
-------
the QA staff to track the quality of data gen-
erated at a given point in time.
Laboratory Objectives
The primary objectives of the preparation
laboratory were: (1) to prepare homogeneous,
anonymous, air-dry subsamples from pro-
cessed bulk samples received from the sam-
pling crews, (2) to organize batches of sub-
samples, including quality evaluation (QE) and
quality control (QC) sampler, suitable for
shipment to analytical laboratories, and (3) to
perform preliminary anaiysas of the bulk roil
samples to supplement the data generated by
the analytical laboratories.
This report describes the operations and
data quality facets of the MASS preparation
laboratory program. A series of recommenda-
tions for improving both the quality and
efficiency of preparation laboratories in future
soil surveys is also included. Specific informa-
tion concerning the protocols used for the
MASS sample preparation can be found in
Appendix A.
Organization of the Report
This document has been organized into
five main sections. Section 1 provides an
overview of the survsy objectives and prepara-
tion laboratory analytical parameters. Section
2 systematically describes the MASS prep-
aration laboratory operations. Section 3 de-
scribes tue implementation of the overall QA
program, its relation to data quality assess-
ment, and the use of measurement quality
samples during the various stages of data
collection and verification. Section 4 provides
the data quality results and discussion for the
analytical parameters measured at the prepar-
ation laooratory. Section 5 provides the con-
clusions and recommendations that have been
generated from these findings, particularly in
ragard to issues of concern for sample prepar-
ation, improvement in QA design, and sugges-
tions for QA efforts in future surveys.
Throughout the report, data quality is
discussed in terms of precision, accuracy,
representativeness, completeness, and com-
parability. Data quality achievements in rela-
tion to objectives established at the beginning
of the survey tsre evaluated.
The chemical and physical soil parame-
ters analyzed at the preparation laboratory are
referenced either by a data-variable or descrip-
tive parameter name. A list of parameters and
their corresponding descriptions is given in
Table 1-1. The order of the parameters is
consistent with their order of presentation in
this report.
Table 1-1. Analytical Parameters Measured at ths Preparation Laboratory
Parameter
Description of Parameter
PH_MP Field-moist pH is determined in a deionized water extract using a 1:1 mineral soil to solution ratio or 1:4
organic soil to solution ratio, measured with a pH meter and combination eSectrodt,.
OM_LOI Percent organic matter is determined by loss on ignition and is used to distinguish organic soils from
mineral soils for analytical purposes.
MOIST_P Air-dry soil moisture measured prior to disaggregation/sieving at the preparation laboratory is expressed
as a percentage on an oven-dry weight basis; minaral soils are dried at 105°C, organic soils at 60°C.
RF_FG Fine gravel is the portion of the rock fragments in the soil with particle diameter between 2 mm and 4.75
mm, and is measured gravimetrically.
RF_MG Medium gravel is the portion of the rock fragments in the soil with particle diameter between 4.75 mm and
20 mm, and is measured gravimetrically.
BD_CLD Bulk density is measured from replicate intact clod samples collected in the field, and is measured by a
gravimetric water submersion procedure.
BD_KV Bulk density is gravimetrically measured from replicate known volume samples collected bv cithsr a volume-
fill or volume-replacement procedure.
-------
Section 2
Preparation Laboratory Operations
Brief explanations of the specified meth-
ods and procedures for the soil preparation
tasks are given below in conjunction with
descriptions of the operational components.
The procedures are discussed in sequential
order of performance by the preparation labor-
atory. The detailed methods and analytical
procedures used in soil preparation for the
MASS are provided in the PLSOP (see Appen-
dix A). This section describes preparation
laboratory operations, including difficulties with
the procedures and deviations from the proto-
cols.
Laboratory Design
The MASS soil processing and analyses
were performed in a secure, access-controlled
warehouse located in Las Vegas, Nevada.
Sample integrity was ensured by the use of
detailed sample labels, by documenting the
status of each sample during the processing
and analyses, and by avoiding physical or
chemical contamination during each processing
step. The warehouse contained a group of
mobile laboratory units that were used for
specific preparation activities, as shown in
Figure 1. Fume hoods and exhaust fans were
provided in each unit and were operable during
stages of sample processing as required by
the PLSOP. Ventilated sample drying and
sieving rooms were constructed to aid in
preserving sample integrity. The rooms were
constructed of wood frames and sheets of
clear plastic, and were equipped with electric
fans and filtered inlet louvres to allow air
exchange. Deionized water and a source of
filtered compressed air was available for
routine cleaning of equipment.
Personnel
The preparation laboratory manager was
required to be knowledgeable in laboratory
methods and procedures, and have demonstra-
ted ability to track large numbers of samples
and supervise laboratory personnel. Adequate
staffing was provided to ensure a fast and
efficient turnaround of samples from the field
to the analytical laboratories. All personnel
were thoroughly trained in the protocols and
procedures by the laboratory manager before
sample preparation began. The laboratory
was able to maintain a steady rate of proces-
sing with a regular staff of four full-time
analysts and two part-time technicians.
With few exceptions, the distribution of
tasks followed closely the outline of activities
provided in the PLSOP. The laboratory mana-
ger was responsible for assigning duties
according to the specific project needs at any
given time during the operation of the labora-
tory.
Equipment Disbursement
A laboratory technician, under the direc-
tion of the laboratory manager and the QA
logistics coordinator, was responsible for
distributing and tracking equipment and sup-
plies to the sampling crews and preparation
laboratory personnel. The equipment cache
was stored in an area of the warehouse
having restricted access and a series of sign-
out sheets was used.
-------
SAMPLE DRYING
NON-DORP
OlSAQOnEOATlON
SIEVING
ROCK FRAGMENT DETERMINATION
NW DOOR
01-00
BULK
DENSITY
LOSS
ON
IGNITION
SAMPLE
nECEIPT
NON-DDRP
AFP
WC
we
I
PARKING AREA
HOMOOENIWTION
SUB-SAMTL1NO
ROLL
DOOR
NW DOOR
fi
FRONT OFFICE
c
Figure 1. Layout of the Mid-Appalachian Soil Survey Preparation Laboratory.
-------
An initial package consisting of all essen-
tial sampling hardware and a two-week supply
of the necessary consumables was prepared
and distributed to each of the original four
sampling crews at the beginning of the survey.
A second shipment, containing a major portion
of the remaining consumables needed during
the survey, was sent to the appropriate United
States Department of Agriculture (USDA) Soil
Conservation Service (SCS) field office or
another convenient delivery point approximately
one week after the first shipment. Each initial
field equipment package included the items
presented in Table 2-1.
Table 2-1. Contents of Initial Packages of Equipment
and Consumable Supplies Sent to
the Sampling Crews
Equipment/Consumables
Quantity
Hand pump 1
35-mm camera w/built-in flash 1
Clod boxes 30
Brass sieve, square-holed, 19-mm 1
Stapler, large 1
Stapler, small 1
Gel-pacs 80
Coolers 10
Hole punch 1
Straight-edged trowel 1
Cylinder, clear plastic, 250-mL 1
Graduated cylinder, clear plastic, 2000-mL 1
Macroballoon beads, 3.18-9.53 mm diameter 4 L
Plug for packing beads in cylinder 1
Field pH kit 1
Random number table 2
SCS-SOI-232 field data form 50
Orange flagging rolls 10
Yellow marker flags 80
Indelible marker - thick 10
Indelible marker - thin 10
Golf tees (horizon delineation) 20
Photogray cards - 8" x 10" 20
Horizon depth tape 1
Metal paint can, teflon-lined, w/lid, 3.8-L 2
Hairnets 144
Plastic inner sampling bags 350
Canvas soil bags 350
DDRP Label A 200
Staples, box 2
Small plastic clod bags 600
Twist ties 300
Clod tags 300
Clod box labels 500
Log books 5
Saran™ powder, 540-g packets 10
Acetone, 19-L can 1
A fifth sampling crew was brought into
service later in the sampling period. This did
not cause any serious difficulties with the
distribution of equipment and supplies, al-
though two of the other four crews had to
share some of their supplies and field audit
samples with the newest crew. The sampling
crews usually obtained supplies by calling the
laboratory manager at least one week before
an anticipated shortage or by making these
needs known on the weekly conference call.
On occasion, a crew member would request
additional supplies when a particular consu-
mable was already depleted or could not be
found. All shipments to the crews were made
by an overnight carrier service.
Supply sheets were provided for the
sampling crews, who were requested to log all
consumables taken from or returned to the
SCS offices, which allowed better estimates to
be made of any additional consumable items
needed. Log entries identified the sampling
crew, the equipment taken, and the date.
Preparation laboratory personnel were
originally tasked with providing the saran-
acetone mixture used for coating bulk density
clod samples in the field. Because of leakage
and safety problems in the containment and
transport of the pre-mix solution, it was deci-
ded that the laboratory would instead send
pre-weighed packets of saran powder and
factory-sealed cans of acetone. The sampling
crews prepared their own mixtures of approxi-
mately 1:5 by weight of saran:acetone solution.
After soil sampling was completed, all
hardware and remaining consumables were
returned to the preparation laboratory. The
hardware and supply items listed in Table 2-2
were required to perform the prescribed soil
processing and analysis tasks.
Sample Receipt
To facilitate the receipt of sample ship-
ments, the sampling crews were asked to
make every effort to inform the preparation
laboratory when samples were being shipped
from the field. Because this was often logisti-
cally difficult, the preparation laboratory in-
itiated an alternating work week that enabled
laboratory personnel to be available to receive
shipments during all carrier delivery hours, six
days per week. As a result, no carrier ship-
ments had to be turned away because of the
laboratory being closed.
-------
Table 2-2. List of Equipment and Consumable
Supplies Used at the Preparation Laboratory
Items
Log books
Raw data forms
Munsell color book
Pens, indelible black ink
Kraft paper, 1-meter (36") wide rolls
Plastic gloves (unpowdered)
Dust masks
Organic vapor masks, cartridge-type
Metal paint cans, teflon-lined, w/ lid, 1-gallon
Aluminum weighing tins
Nalgene bottles: 2-L, 500-mL, 125-mL
Carboys, 13-gallon capacity
Crucibles, 450"C tolerance
Evaporating dishes, 450°C tolerance
Beakers, 1000-mL
Tongs
Ring stand
Thermometers, 0 to 100° C range
Sample drying tables, PVC / nylon mesh
Rolling pins, wooden
Rubber stoppers, No. 10 size
Brass sieves, square-holed: 2-mm and 4.75-mm
Riffle splitter, closed-bin, with 1.25-cm baffles
Top-loading balance, 0.01-g accuracy, 500-g capacity
Top-loading balance, 0.1-g accuracy, 5-kg capacity
pH meter, with proper electrodes
Desiccators
Desiccant
Hot plate
Convection oven
Muffle furnace
Shipping boxes
Packing material
Strapping tape
Saran™ powder
Acetone
Plastic containers, 25-mL and 50-mL
Bulk soil samples, known volume bulk
density samples, and field audit samples were
placed with two frozen gel-pacs in box-
enclosed styrofoam coolers. Sample shipment
forms and the original SCS-SOI-232 field data
forms (Kern et al., in press) corresponding to
the samples were sealed in plastic bags and
taped to the underside of the cooler covers.
The outer boxes enveloping the coolers were
firmly enclosed with strapping tape and the
coolers were then shipped directly to the
preparation laboratory via overnight carrier.
Because of their fragile nature, the soil clod
samples were securely packed with vermiculite
or crumpled newspaper in coolers separate
from the bulk soil samples. Each sample was
labeled in accordance with the protocols
outlined in Appendix A.
The field data forms were completed by
the sampling crews while at the sampling
sites. Each form listed all available site infor-
mation for a particular pedon, including des-
criptions of landform, vegetation, soil climate,
and a detailed pedon description. The sample
labels were checked for coding accuracy
against the field data form and sample ship-
ment form. A sample receipt raw data form
(see Appendix B) was filled out completely for
all samples received and logged in at cold
storage.
All preparation laboratory personnel
became familiar with the sample labeling
system, as it was the primary information
used in receiving and tracking samples. After
receiving a sample shipment, the preparation
laboratory staff checked the condition of
incoming samples and logged the sample
labeling information on a sample receipt form.
The number of samples collected and the
horizon, watershed, and sampling class desig-
nations were compared from the field data
form to the sample labels to ensure that all of
the samples collected had actually arrived with
the shipment and that the labeling was cor-
rect. The field data forms and sample ship-
ment forms were manually scanned for consis-
tency and any discrepancies were noted and
brought to the attention of the QA manager,
although a copy of each form was kept at the
preparation laboratory. Raw data from the
sample receipt form were entered into a data
entry and verification program on a personal
computer (described later in this section).
Discrepancies noted on the verification printout
were evaluated and discussed with the QA
manager and the ERL-C sampling QA staff.
While the laboratory manager validated
the accuracy of the sample labeling, the sam-
ples were grouped according to their respec-
tive pedon and set identification (set ID) codes
and were placed in cold storage. The QA
manager was notified of any labeling discrep-
ancies and immediate consultations with the
ERL-C sampling QA staff or the appropriate
sampling crews helped to resolve the inconsis-
tencies. Subsequently, the samples were
allowed to undergo the next stages of sample
preparation.
-------
Sample Storage
A walk-in cooler was constructed within
the preparation laboratory in order to contain
both the incoming field samples and the
samples being stored in the intermediate
stages of processing. A separate off-site cold
storage facility was used to store archived
subsamples from those bulk samples that had
already been processed and batched.
Upon receipt, all field samples were
placed in cold storage as soon as possible
and remained there until sufficient space was
available in the drying area for a "run" of
samples (ranging from 20 to 40 samples,
grouped according to set ID) to be air dried as
described below. After air drying, the samples
were returned to cold storage until further
processing. Samples were organized in cold
storage so that a particular sample within a
given set could be easily located. This was
best accomplished by grouping the samples
according to their component set and pedon
codes which were maintained during storage,
processing, and shipping.
The cold storage area was maintained at
a constant temperature of 4°C and was moni-
tored by a continuous rotation sensor and by
daily thermometer readings. A substantial
deviation in temperature (± 2°C or more)
occurred only once and was recorded in the
cold storage log book. Specifically, the labora-
tory had a temporary refrigeration failure on
April 24, 1989 that was reported to the QA
staff immediately. The unit was repaired
within approximately four hours. Sample
integrity was not threatened because the
failure occurred after sample processing had
been completed on the bulk samples. At the
time, only clod samples were in the facility and
they were not considered to be temperature-
sensitive.
Sample Drying
When there was sufficient space in
the enclosed drying area, the field-moist
bulk samples were taken from cold storage,
grouped by pedon code, and were spread out
on large sheets of kraft paper placed on
drying tables to air dry. Where necessary,
several layers of paper were placed under-
neath the samples to absorb excess moisture
and this paper was replaced periodically for
faster drying. A sheet of paper was placed
loosely over each sample to minimize the
possibility of cross contamination among
samples without restricting air circulation.
The use of nylon mesh drying tables and
a multiple-fan air exchange system enhanced
the air circulation around the samples, thereby
increasing the rate of sample drying. Labora-
tory personnel occasionally stirred the samples
to encourage uniform drying. Possible sources
of sample contamination, such as chemicals,
food, drinks, and smoking, were prohibited in
the drying area. A pair of gloves dedicated for
each sample was always worn when handling
the samples. The drying area was damp-
mopped at least once per week to control
dust, and was done after air-dry samples were
returned to cold storage and before field-moist
samples were laid out to dry.
Organic soils in advanced stages of
decomposition and mineral soils high in clay
had a tendency to harden nearly irreversibly if
allowed to dry without concerted efforts to
disaggregate clumps and structural peds
common in these types of samples. With this
in mind, efforts were made during the air-
drying period to disaggregate these clumps
and peds while the samples were still some-
what moist or friable before reaching an air-
dry state.
Determination of Field-Moist pH
(PH_MP)
Immediately after a field-moist bulk
sample was spread to air-dry, a composite
subsample was collected from across the
drying sample and placed in a 125-mL Nal-
gene™ bottle. The pH of the sample in deio-
nized water was determined in a mobile labor-
atory unit set up specifically for this measure-
ment. A digital pH/mV meter, capable of
measuring pH to 0.01 pH unit using a combina-
tion pH and reference electrode, was calibra-
ted immediately prior to each run of analysis
using standard stock buffer solutions of pH
7.0 and pH 4.0. Approximately 1:1 and 1:4 soil
to solution ratios were used for mineral and
organic soil analysis, respectively. The elec-
trode was rinsed with deionized water bet-
ween each sample in order to prevent solution
carryover.
-------
Early in the survey, the laboratory ana-
lysts experienced difficulties in the calibration
of the meter. The problems were traced to the
group of available electrodes that appeared to
be faulty because of fluctuating readings.
While additional electrodes were being requisi-
tioned, a borrowed electrode was used and
provided sufficient precision and accuracy
results.
All calibration readings and adjustments
were recorded in a soil pH logbook and ini-
tialed by the analyst. Data from each day's
analysis were recorded on a pH raw data form
and subsequently entered into a data entry
and verification program on a personal com-
puter.
Determination of Organic Matter
by Loss on Ignition (OM_LOI)
A loss-on-ignition method was used to
determine percent organic matter in the sam-
ples and to classify each sample as a mineral
or organic soil. Because organic soil con-
stituents are oxidized at high temperatures,
percent organic matter could be calculated on
a weight-loss basis. From this, the organic
carbon content can be estimated. A modified
version of the method described in MacDonald
(1977) was used. This method was added to
the DDRP preparation laboratory activities for
the MASS in order to identify organic samples,
containing 20 percent or more organic matter,
that would be grouped into separate batches
for analysis.
A composite subsample of a bulk soil
sample was placed into a crucible and dried
overnight in an oven to determine its air-dry
weight. Samples that were suspected to be
mineral soils were dried at 105°C and organic
samples were dried at 60°C. After weighing,
the crucibles were placed in a 450 °C muffle
furnace activated in an operable fume hood.
Determination of Air-Dry
Moisture (MOIST_P)
Each bulk sample was allowed to dry
until it appeared to be at or below the speci-
fied air-dry moisture content, which was 2.5
percent for mineral soils and 6.0 percent for
organic soils. Air drying generally took about
two or three days for most samples, although
certain saturated organic samples took a week
or more to visibly dry. At this time, a com-
posite subsample was selected for the gravi-
metric determination of air-dry moisture. If the
moisture criteria were not satisfied for a
sample, it was allowed to continue drying until
a later determination could be made. In a few
cases, certain samples would not dry suffi-
ciently to satisfy the criteria but were con-
sidered to be air dry for practical purposes.
Differences in soil texture and mineralogy are
thought to have influenced this variation in
moisture content.
Sample Disaggregation and
Sieving
After a bulk soil sample was determined
to be air dry, it was ready to be disaggregated
and sieved in order to remove rock fragments
and to prepare the sample for homogenization
and subsampling. The disaggregation and
sieving tables were covered with layers of
kraft paper that were replaced after each
sample had been sieved. Dust masks and
protective clothing were worn at all times, and
ventilation and heating was provided for the
sieving area. All equipment used during this
operation was cleaned with compressed air
and fresh tissue paper after each sample had
been sieved.
After recording the weight of the air-dry
bulk sample, the soil was disaggregated to
allow the less than 4.75-mm soil fraction to
pass a No. 4 sieve and the less than 2-mm
soil fraction to pass a No. 10 sieve. Depend-
ing on the consistence of rock fragments in
each sample, a wooden rolling pin or large
rubber stopper was used for disaggregation.
The rolling pin was successfully used when
there were no fragments in the sample or the
rock fragments were extremely firm. Because
of the highly weathered state of many of the
shaly fragments in some samples, various
types and sizes of rubber stoppers were used
to selectively disaggregate the individual soil
peds. This procedure allowed the passage of
soil through the sieves without breaking the
fragments.
10
-------
Determination of Rock Fragment
Content (RF_MG and RF_FG)
Rock fragments remaining from the
sieving procedure were retained on two sieves;
the No. 4 sieve retained the 4.75- to 20-mm
medium gravel fraction and the No. 10 sieve
retained the 2- to 4.75-mm fine gravel fraction.
Fragments larger than 20 mm were sieved out
of the samples by the field sampling crews.
Gravimetric analysis was used to determine
the percentage by weight of each fraction in
the bulk soil samples.
There were no deviations from the speci-
fied protocols for rock fragment determination.
The laboratory bagged and labeled the frag-
ments and awaited further instructions for
dispensation.
Sample Homogenization and
Subsampling
In order to obtain representative quan-
tities of soil suitable for physical and chemical
analysis at the analytical laboratories, the
protocols specified the use of a closed-bin
riffle splitter to prepare homogenous subsam-
ples from the less than 2-mm soil fraction.
The riffle splitter was cleaned with com-
pressed air after each sample was split. The
compressed air was passed through a desic-
cant filter trap. If the operator suspected that
the riffle splitter was still not thoroughly clean,
the riffle splitter was washed with deionized
water and allowed to air dry completely. The
work area was vented with an operating fume
hood and the surfaces were cleaned with a
hand brush after each sample was processed.
The preparation laboratory prepared two
different subsamples because separate con-
tracts were awarded to analytical laboratories
for the general analysis of 47 physical and
chemical parameters and for the elemental
analysis of total carbon, nitrogen, and sulfur.
The general analytical sample weighed ap-
proximately 400g to 500g and the elemental
analytical sample weighed approximately 30g
to 60g.
A preparation duplicate sample was split
from each field duplicate sample and also
from its associated routine sample. The
preparation duplicates served as measurement
quality samples for the purposes of evaluating
the precision associated with the homogeniza-
tion and subsampling procedures.
There were no deviations from the speci-
fied protocols for soil homogenization and
subsampling. A daily record of the riffle
splitting operation was made in a logbook for
each sample, specifying the initial weight of
sample and a description of the passes made
through the splitter to achieve the final weight
of the analytical samples.
The soil material remaining after sub-
sampling the analytical samples was placed
into a 2-L Nalgene bottle for archiving in long-
term cold storage.
Determination of Bulk Density
For the DDRP surveys, the bulk density
of a soil is defined as the weight of oven-dry
soil (minus rock fragments) per unit volume
including the pore space, expressed as g/cm3.
For most mineral soils, bulk density ranges
from about 0.6 to 2.0 g/cm3. With increasing
organic matter content, soils generally exhibit
a decrease in bulk density because organic
matter has higher porosity and lower density
than mineral particles of the same diameter.
As in past DDRP surveys, the clod meth-
od was the primary method for determining
bulk density in the MASS. Two alternate
methods were also attempted for soil horizons
that failed to yield satisfactory clods. The first
was a volume-replacement method which
utilized a known volume of small foam beads
packed into a cylinder to replace a selected
volume of soil excavated from a given horizon,
and the second was a volume-fill method that
involved a simple container filling procedure
which was used if the clod or volume-replace-
ment methods did not produce representative
samples.
Clod Method (BD_CLD)
Analysis of the clods was based on the
method described in USDA-SCS (1984). Repli-
cate soil clods, usually three per horizon, were
collected by the sampling crews whenever
possible (Kern and Lee, in press). The clods
were weighed at the laboratory and then oven-
dried. Each clod was dipped in a 1:5 by
weight sararracetorie mixture and was allowed
11
-------
to dry briefly. The dipping procedure was
repeated until each clod was assumed to be
impervious to water (usually three to six trips).
Each clod was submerged in a 2000-mL bea-
ker of water tared on a balance and the
weight of the clod was recorded. After two
hours in a muffle furnace at 400°C, the clods
were passed through a 2-mm sieve to detei-
mine the percentage by weight of rock frag-
ments used in calculating the bulk density.
There were no major deviations from
the specified protocols for the bulk density
determination. A number of floating clods
were identified, mostly when the bulk density
was less than 1.0 g/cm3. Such clods were
forcibly submerged to allow bulk density
determinations to be made. This was ac-
complished by fitting the clod into the lower
cavity of a test tube holder, such that the clod
would rest against the holder when submerged
and remain stationary within the beaker.
Extra care was taken in the use of the
Saran™ powder and acetone solution. When
mixed, the resulting solution has a tendency to
volatilize hydrogen chloride gas which can
cause deleterious health effects. The solution
was mixed and used only in an operating
vented fume hood, and a cartridge-type
organic vapor mask, gloves, and laboratory
coat were always worn.
Known Volume Methods (BDJCV)
Bulk density was also determined by two
known volume methods whsn clods could not
be collected in the field, as described in Kern
and Lee (in press). The initial and final vol-
umes of beads used for the volume replace-
ment samples were recorded in the field by the
sampling crews. The volume filling container
used was a 250-mL graduated plastic beaker,
which had an absolute volume of 300cm3 when
entirely filled. The known volume samples
were processed in a manner similar to the
method described in Blake (1965). The prepar-
ation laboratory oven-dried and weighed the
samples, sieved the rock fragments, and
calculated the bulk density.
A procedural change was made in the
calculation of the bulk density for known
volume samples. Instead of using the stan-
dard 2.65 g/cm3 particle density value as
stated in the PLSOP, actual rock fragment
estimates grouped by parent mrusra! cia'r;
were used. It waz assumed th3> Ur^o vnr ^.-n
volume samples gsnern'ly v-..-"irj c-'sV.r.l" a
higher than average per:.,-;?-;.;;--/,-? of ">"-•• •.. M-
ments and, therefore, ^ ->rj(' • • v^;.-* . >o. •
accurate information o?i ,, ".. :.:-j;. •? T>S''V '.?>.;
density was detenuirMK" • • .',••..•".'-;:•
submersion technique on •• ••, •;, •;) -<':-)>,,•: --;'. •- •<
from the bulk &,?mp!" r-' ••..-.:; ,i^-:'-'- J;--'
which a known, vo^ur-j ----.I,-. ;, .;-;> r ,-- ,.'-iv-.ci-"'.
The particle der^:'-5/ ».::, '•"-»,..'•- ~ •••> ;"r-.r.r: •:, :-
differing pys-.,:rt ?r:-3tr -; r, --;r; -• >::!" :^•';;•_
values YtGi j CCt.V" .'>'U :,.'r , ' '~ ;•"• .":?*, •'.''.'• '' -"-.
erial classes, she aifif-;a«"t r;itk;i-j t^'isS'ty
vabes that wore u>t'1 &>'~i Ji-- •?;,< ''ri r;1-/.^
Table 2-3. Partlclfc Density VaSu»*
Mat©rEaE«g
Parent
Parent Matehal
AO - Sandstone (unspecified)
A1 - Sandstone (noncalcareous)
BO - Tnterbeddod eedimentary
B5 -- Sandstone and shalo
B6 - Sandstone aiid siltstono
B7 - Shale and siltstona
MO - Matarncrphic (unspecified)
M9 - Quartets
Particla density (g/cm3)
2.52
y..$2
2.-r4
£.44
2A4
2.45
2.51
2.51
Sample Shipment
After the sample preparation and ana-
lyses were completed for a sufficient number
of sets of samples, batches were constructed
and readied for shipment to the analytical
iaboratories. A batch of samples consisted of
approximately 40 samples, including about 30
routine samples and 10 different measurement
quality samples (see Section 3) which were
used to control and assess the various com-
ponents of measurement uncertainty. The
preparation laboratory manager assisted the
QA logistic?, coordinator on packing and
shipping the batches of samples via overnight
carrier. The batches of samples were packed
in sealed heavy-duty cardboard boxes for
shipment to the designated recipient analytical
laboratories. The QA soil chemist routinely
called one or two days in advance to notify an
analytical laboratory that a b?tch of samples
was to be shipped.
12
-------
Sample Archiving
After e?inh batch of samples was
shipped, the remaining volumes of bulk soil
sample material were placed in long-term
archive. Samples were stored in 2-L Nalgene™
bottles, rffixed with a ODRP Label A (see
Appendix /') and were grouped in boxes by
their respective batch, set, and pedon codes.
Multiple bo'-.'Jes from the same sample code
were -so r.ov,.d en tho iabei. The boxes of
sample:; '• -c;c Ir,,-:,>-?diate!y placed in Song-term
co'o r/;, . :, ,.-•., •.';.-'-., oo,u~. warehouse facility
v v.;i ,r'.-••-.,- - ' •..,-•-. ;. ir...-:;.^?;1 ?:;? lo a^.-v
Precision and accuracy of the measure-
ment quality samples were checked while the
samples were being processed which per-
mitted the rapid identification of discrepancies
and allowed the QA manager and laboratory
manager to efficiently track progress at the
laboratory. Features of the computer entry
arid verification program are highlighted in
Appendix C.
The computer program was capable of
generating the thai batch data lorn"; ODRP
Form 101 (see Appendix B), whk.n vur>. ;-• fired
fw analytical data hv b;?;tch ;\ ih': c. .- •.•,.'
'
Seven raw data forms were designed by
EMSL-LV for use in sample tracking and for
entering data relating to sample receipt, bulk
sample processing, field-moist pH, air-dry
moisture, loss on ignition, clod bulk density,
and known volume bulk density. All raw data
were recorded ors these forms in indelible
black ink and the completed forms were or-
ganized in binders. An example of each form
is provided in Appendix B.
The preparation laboratory was provided
with logbooks for recording raw data on a
daily basis. The bulk sample processing
forms, plus sampling labels from the original
bulk sample, were placed in binders organized
by batch number.
Entry
A personal computer was used at the
laboratory for entry and verification of all
preparation and analysis data recorded on the
raw data forms. Entry of the raw data was
accomplished by selected laboratory personnel
who WP?S trained to operate a custom SAS-
AF™ menu-driven program for each of the
primary phases of sample preparation and
analysis. The program provided an efficient
way to track the status of the samples from
the time of field sampling through shipment to
the analytical laboratory. By comparing data
among different files, sample labeling dis-
crepancies were quickly identified and reported
to the QA manager c;r-d the ERL-C sampling
QA staff.
manually for ee.uh l>a'>r:h, >;v- C •'". ••,', ,y ;r ,-.
identified the batch as containing mineral or
organic soil samples and denoted each sample
only by batch and analytical sample numbers.
"this procedure disguised the originating pedon
and horizon of each routine sarnpie as well as
the identity of the measurement quality sam-
ples. The QA manager was provided with
copies of both forms to serve as documenta-
tion of the shipment.
Sample Tracking
As part of the data entry and verification
program, a sample tracking system was
implemented in order to document the transfer
of all samples between the field and the
preparation laboratory and, ultimately, their
transfer to the analytical laboratories. Before
sampling began, the DDRP soil sampling task
leader provided the EMSL-LV QA manager and
preparation laboratory manager with a list of
watershed and pedon identification numbers
for each pedon to be sampled. The informa-
tion was entered into the computer entry and
verification program at the preparation labora-
tory before samples were received. As the
samples arrived at the preparation laboratory,
the sample receipt form was filled out and
entered into the computer. If a particular
watershed/pedon combination had been previ-
ously used, e.g., the preparation laboratory
had previously received samples with that
watershed/pedon number or a particular
watershed/pedon number could not be found
on the ERL-C list, the soil sampling task leader
and the QA manager were notified immediate-
ly. Early in the survey, a series of samples
from nine pedons were determined to have
been labeled with incorrect pedon codes that
13
-------
were different from the codes initially provided.
The appropriate coding changes were made on
the sample labels by the laboratory staff.
Progress Reporting
The preparation laboratory data entry
and verification program, described in Section
3, produced summary statistics in order keep
the EPA management team informed on the
progress of field and preparation laboratory
operations. Upon request, the EMSL-LV techni-
cal monitor was provided with information on
the total number of sets, pedons, and samples
received by the preparation laboratory, and the
percentage of each laboratory analysis that
had been completed for the pedons specified
for sampling. The sampling information was
used to assess the progress of the sampling
crews. It was necessary to ensure a steady
rate of progress because crews were sampling
during late fall in areas that would not be
conducive to winter sampling. The data pro-
ved useful in showing the need for another
sampling crew to be added in order to com-
plete the sampling in the allotted period. The
analytical information was used to evaluate
preparation laboratory progress. With this
information, the QA manager and technical
monitor could determine where additional
effort was needed and take appropriate action
to improve the rate of progress.
14
-------
Section 3
Quality Assurance Program
A preparation laboratory QA program
was specified during the course of MASS
activities to ensure that data quality could be
evaluated in real time. The QA program con-
sisted of operational and assessment com-
ponents that were meant to aid in producing
data that would satisfy the quality require-
ments of the DDRP data users. The following
subsections explain aspects of the QA pro-
gram for preparation laboratory activities.
Operational Components
The operational components of the
preparation laboratory QA program included
training the laboratory personnel in the proto-
cols to be followed, establishing a communica-
tions network between the laboratory and
other MASS participants, setting measurement
quality objectives, and performing on-site
systems audits. In addition, preventive main-
tenance and corrective action were accom-
plished where necessary.
Training
The PLSOP (see Appendix A) served as
the protocols for sample preparation in the
MASS. Prior to the receipt of routine samples,
all laboratory personnel were initially trained by
the preparation laboratory manager at a two-
day hands-on workshop held at the laboratory
on September 9-10, 1988. The purposes of the
workshop were to familiarize the participants
with the laboratory protocols and to discuss
key activities and responsibilities. Training
established a basis for consistency among the
analysts, thereby increasing the likelihood that
the data would be of comparable quality.
Additional training sessions were held every
two weeks at the laboratory, where the labora-
tory manager clarified the preparation and
analysis protocols and trained the staff in
other aspects of soil genesis and morphology
that were needed to identify characteristics of
the many varied soil horizons the staff was
processing. Of particular importance was the
identification of weathered rock fragments so
that laboratory personnel would use care to
avoid crushing this material during the sieving
process.
Communications
Additional support was provided through
communications with other DDRP personnel.
The EMSL-LV QA manager periodically visited
the preparation laboratory to discuss the data
quality assessment procedure with the labora-
tory personnel. The laboratory manager called
the QA manager or other QA representatives
when additional information was needed
regarding a particular procedure. The soil
sampling task leader and QA manager were
notified whenever there was a problem con-
cerning the field sampling operation.
Weekly conference calls were helpful in
keeping the preparation laboratory operating in
an efficient and consistent manner by allowing
potential difficulties to be identified, discussed,
and resolved. The preparation laboratory
manager, EMSL-LV QA staff, ERL-C sampling
QA staff, and SCS state office representatives
participated in these calls. Issues discussed
during the conference calls included field
supply needs, clarification of procedures, e.g.,
sample labeling, and shipment of samples.
Routine calls, however, were more beneficial
for dealing with specific issues.
Preventive Maintenance
All laboratory equipment was maintained
in accordance with guidelines in the PLSOP
(see Appendix A). The analytical balances and
15
-------
pH meters were calibrated at the beginning of
every day of use. The instruments were
handled and stored according to the guidelines
identified in the manufacturers' instruction
manuals. The convection ovens, muffle fur-
naces and walk-in cooler were monitored to
ensure proper functioning. An air compressor
was used to clean the surfaces of the sample
preparation equipment, e.g., rolling pin and
sieves, after processing each sample. A
desiccant trap device was built into the airflow
conduit to collect any vapor or oily liquid
residues before reaching the nozzle. In addi-
tion to cleaning with compressed air, the
processing equipment was wiped with clean
tissues if soil was adhering to the surfaces.
The sample drying and sieving areas were
swept and damp-mopped once per week
during the exchange of samples in the drying
area. A laboratory safety plan was imple-
mented successfully during the survey.
Special Studies
Although there were no deviations from
the specified protocols for sample drying,
concerns were raised in the previous DDRP
surveys about soil homogenization techniques,
possible dust contamination of the samples,
and the potential for leaching of chemicals
from the kraft paper and nylon mesh into the
samples. Separate studies were undertaken
by the QA staff to investigate the effects of
these factors on bias or variability in the
resulting data. Copies of the internal reports
summarizing these studies are found in Appen-
dices D, E, and F, respectively. Brief sum-
maries of the special studies are provided in
the following subsections.
Soil Homogenization Study
During the planning stages of the MASS,
it was decided that a study should be initiated
to assess the best methodology for sample
homogenization and subsampling (see Appen-
dix D). The cone-and-quarter technique and
the closed-bin and open-bin riffle splitting
techniques were compared for four soil sam-
ples of different texture. Particle size analysis,
organic matter by loss on ignition, and pH in
0.01M calcium chloride were used to assess
differences in replicate subsamples from the
different techniques. The closed-bin riffle
splitting technique using five complete passes
through the device was determined to be the
optimal method.
Dust Study
An external dust study was initiated
before and during the construction of the
sample drying area in the warehouse (see
Appendix E). The study was intended to
assess the amount and constituents of dust
collected in the general warehouse facility and
in the enclosed sample drying room. The
study was conducted prior to the receipt of
samples for the survey. It was concluded
that the possibility of significant dust con-
tamination in the warehouse was low, par-
ticularly in the drying area. As additional
safeguards against dust contamination, a pair
of enclosed ventilated rooms were constructed
for sample drying and sieving.
Paper and Mesh Study
Another study assessed the possibility of
contaminants leaching into the samples from
the kraft paper and nylon mesh used during
sample drying (see Appendix F). It was con-
cluded that the mesh could not have served as
a contaminant of the samples, but that the
kraft paper could have had the potential for
detectable contamination by sulfur oxides and
sodium when a moist or wet sample made
contact with the paper.
To investigate this latter possibility, three
low-range audit samples that were identical to
the low-range laboratory audit (LAL) samples
used in the survey were saturated with deio-
nized water and were air-dried in the sample
drying area using the same procedure that
was used for the routine samples. Each of
these samples, coded "PAL" in the MASS
analytical data base, were placed in a batch of
samples sent to a separate laboratory for
analysis. The data for the PAL samples were
compared to the LAL accuracy windows and to
the LAL in their respective batches (see Appen-
dix F). In the first batch, the PAL was within
the accuracy windows for all parameters. The
PAL in the second batch satisfied the criteria
for all parameters except the sodium extracted
in calcium chloride. This parameter was above
the upper boundary of the window by 0.001
meq/100g. Within these two batches, the LAL
contained more values that did not satisfy the
accuracy criteria than the PAL sample. The
16
-------
third batch contained five parameters in which
the PAL samples were above the upper boun-
dary of the accuracy window. Three parame-
ters of particular concern were sodium ex-
tracted with calcium chloride, sulfate extracted
with deionized water, and the sulfate adsorp-
tion isotherm. However, the sodium value for
the LAL was also above the maximum ac-
curacy window value, and reanalysis of the
extractable sulfate parameter proved that the
original value was erroneous.
It was determined that no significant
differences existed between the LAL, which
was not placed on the kraft paper, and the
PAL. This finding suggests that the kraft
paper was not contributing a significant
amount of sulfur or sodium contamination to
the MASS samples.
Data Quality Assessment
A primary goal of the QA program was
to provide a detailed description of measure-
ment quality in accordance with objectives that
were defined prior to beginning the survey.
The following subsections describe the dif-
ferent phases of assessment.
Quality Attributes
The preparation laboratory data were
assessed in accordance with the following
quality attributes:
• Precision - The level of agreement among
multiple measurements of the same soil
characteristic.
• Accuracy - The level of agreement bet-
ween an observed value and the "true"
value of a soil characteristic.
• Representativeness - The degree to which
the soil data collected accurately represent
the population of interest.
• Completeness - The quantity of data that
is successfully collected with respect to
the amount intended in the experimental
design.
• Comparability - The similarity of data from
different sources included in a given set of
data, and the similarity of methodologies
from related projects across the regions of
interest.
Measurement Quality Objectives
To address the overall DDRP objectives,
conclusions must be based on scientifically
sound interpretations of the data base. Ac-
cordingly, the EPA requires all monitoring and
measurement programs to have established
objectives for data quality based on the an-
ticipated end uses of the data (Stanley and
Verner, 1985; USEPA, 1986). The utility of the
data is defined by the ability to confirm, reject,
or discriminate among hypotheses formulated
in the various DDRP computer models. The
primary purpose of the QA program is to
increase the likelihood that the resulting data
base meets or exceeds the overall data quality
objectives (DQOs). The actual quality of data
can be quantified in relation to the DQOs,
thereby allowing the data user to evaluate the
hypotheses with a known level of confidence.
In practice, DQOs are statements of the
levels of uncertainty that a data user is willing
to accept in the data. For the MASS QA
program, measurement quality objectives
(MQOs) were defined as surrogates for user-
defined DQOs. The MQOs for the sampling,
preparation, and analysis phases were speci-
fied in the MASS QA Plan (Papp et al., 1989).
The initial sample preparation MQOs were
established in accordance with the require-
ments of EPA data users and were based on
the selection of appropriate methods to obtain
the necessary data. The MQOs were reviewed
by scientists familiar with soil preparation and
analytical methods, including soil chemists and
laboratory personnel. Modifications to the
preparation laboratory MQOs and protocols
were implemented using information gained
from the DDRP Northeastern region and Sou-
thern Blue Ridge Province surveys and from
peer review comments, in accordance with the
limitations of a particular procedure or analyti-
cal instrument.
Soil sampling in the MASS included the
physical removal of soil samples from an
excavated pit as well as the characterization
of the soil pedon and sampling site. The
MQOs developed for sample preparation and
analysis were not applicable to the field sam-
pling activities; hence, specific sampling objec-
tives were developed to ensure that field
operations were conducted in a consistent
manner and that an estimate of variability
among sampling crews could be provided. The
17
-------
primary goal of the MASS sampling was to
describe and collect soil samples from repre-
sentative pedons of 15 established sampling
classes (Lee et al., 1989). The sampling clas-
ses were intended to represent the range of
soil types encountered in the Mid-Appalachian
region. The field sampling effort produced
qualitative soil data from pedon characteriza-
tion, e.g., soil color, and quantitative data from
soil analysis, e.g., soil pH.
To assess the sample preparation
MQOs, a series of measurement quality sam-
ples were analyzed together with the routine
samples in a manner that was statistically
relevant and in which conclusions concerning
the data quality could be made. There are a
number of assumptions upon which the run/
sample design was based:
• Complete attainment of the analytical
MQOs could be accomplished if each run
of samples satisfied the acceptance
criteria.
• The components of measurement uncer-
tainty at any given phase, e.g., within-run
precision, can be evaluated with data from
a minimum of 20 measurement quality
samples of a given type.
• The primary components of measurement
uncertainty to be assessed at the prepara-
tion laboratory were precision and ac-
curacy.
Table 3-1 presents the MQOs for the
preparation laboratory parameters of interest.
A brief description of the attributes as they
relate to the achievement of MQOs follows.
Precision is estimated in this report by
calculating the standard deviation (SD) or
relative standard deviation (RSD) from the
QE/QC samples for the seven parameters.
Additional precision estimates using analytical
data from the preparation duplicates are
documented in the QA report for the MASS
analytical laboratory data (Byers et al., in
press).
The accuracy characteristics of bias and
intralaboratory trends are assessed only for
the PH_MP and OM_LOI parameters because
reference values were not available for any of
the other parameters. The MOIST_P values
were assessed to determine if the MQOs of
2.5 percent moisture for mineral and 6 percent
for organic soils were satisfied. A consistent
air-dry moisture value was expected because
the field audit material was well homogenized
prior to subsampling.
The primary characteristics of represen-
tativeness at the sample preparation phase
are addressed in the QA report on the analyti-
cal laboratory data because the assessment
involves the use of data from the preparation
duplicates that were not analyzed at the
preparation laboratory.
Completeness is measured by the labor-
atory's performance of the analyses and
processing tasks for all samples satisfying the
sample receipt criteria.
Comparability of the DDRP preparation
activities is evaluated using the precision data
for rock fragments and bulk density from the
different surveys. The evaluation of prepara-
tion laboratory comparability is also docu-
mented in the QA report on the analytical
laboratory data (Byers et al., in press).
Uncertainty Estimation for Laboratory
Analyses
The QA staff was interested in assessing
the various components of measurement
uncertainty with respect to the overall uncer-
tainty confounded in the routine sample data.
The following subsections describe statistical
and assessment procedures that were used.
Statistical Model
Depending on its limitations or assump-
tions, each measurement procedure induced a
random error for each physical or chemical
characteristic of a soil sample. The sum of
these errors can be defined as data collection
error, which is treated as a random variable.
This variable can be placed in an additive
model (Cochran, 1977) and serves as an
estimate of uncertainty in a data set, such
that:
y = n + e
where: y is an observed sample characteris-
tic,
/j is the "true" sample characteristic,
and
e is the data collection error, which
is assumed to be the sum of the
errors generated by the independent
measurement phases.
18
-------
Table 3-1. Preparation Laboratory Wlthln-Run Measurement Quality Objective*
Parameter
PH_MP
OMJ.OI
MOIST.P
RF_FG
RF_MG
BD_CLD
BD_KV
Sample*
type
Buffer (QCCS)
Int dup'
FD
FAL(C)
FAP (B)
FAP (Bw)
FAO (O)
MS (Bs)
Int dup'
FD
FAL(C)
FAP (B)
FAP (Bw)
FAO (O)
MS (Bs)
Int dup*
FD
FAL(C)
FAP(B)
FAP (Bw)
FAO (O)
MS (Bs)
FD
FD
Reps'
Dups'
Reporting
format
0.01 pH units
II
II
II
II
II
II
0.01 wt. %
0.01 wt. %
II
II
II
0.1 wt. %
0.1 wt. %
0.01 g/cm3
0.01 g/cm3
Expected
range
3.90-4.10
2.50-7.50
2.50-7.50
4.80-6.00
4.40-5.40
4.70-5.70
3.70-4.50
4.10-5.00
0.00-90.00
0.00-90.00
0.00-3.00
5.00-15.00
1.00-9.00
65.00-85.00
5.00-15.00
0.50-6.00
0.50-6.00
0.00-2.50
1.00-4.00
1.00-4.00
4.00-8.00
0.50-6.00
0.0-60.0
0.0-60.0
0.60-2.00
0.40-2.50
Acceptable
accuracy
3.95-4.05
as specified
as specified
as specified
as specified
as specified
—
control limits
control limits
control limits
control limits
control limits
_
control limits
control limits
control limits
control limits
control limits
...
...
—
_
Acceptable
precision
0.15 SD
0.20 SD
0.15 SD
0.15 SD
0.15 SD
15% RSD
20% RSD
15% RSD
15% RSD
15% RSD
15% RSD
20% RSD
15% RSD
15% RSD
15% RSD
20% RSD'
20% RSD'
20% RSD'
20% RSD'
° Int dup = internal duplicate, FD = field duplicate, FAL = field audit low, FAP = field audit pair,
FAO = field audit organic, MS = manager's sample.
* The internal duplicate is a replicate subsample that was randomly selected from one soil sample per analytical run.
' The replicate bulk density samples are clod samples in which one to three clods per horizon were sampled.
d The duplicate bulk density samples are known volume samples in which one or two samples per horizon were sampled.
' Variability was anticipated because these duplicates or replicates are measurements of different samples within the
same horizon; however, when variability >20% RSD, these samples were checked for possible measurement error.
The identification of the error distribution
throughout the measurement phases requires
a large number of replicate measurements
which, from a budgetary and logistical stand-
point, can be restrictive. However, a relatively
small number of observations can be used to
estimate the first two moments, i.e., the mean
and variance, of the data collection error. The
moments are sufficient to estimate accuracy
and precision of the routine data in an additive
model. For this report, the standard deviation
(SD) in parameter reporting units or the rela-
tive standard deviation (RSD), i.e., coefficient
of variation, in percent, were used to assess
the precision components. Bias was consid-
ered to be the quantitative measure used to
assess the accuracy component. The sum of
these error terms represents the measurement
uncertainty in a QA sample data sets, and
includes both measurement and population
uncertainty in the routine sample data set. In
order to make effective use of the MASS prep-
aration laboratory data, the evaluation of mea-
surement uncertainty in relation to the overall
data uncertainty is essential.
The expected squared error was used
as an estimate of uncertainty and was calcu-
lated as the sum of the error variance and the
square of the bias term described below. The
error terms can be defined only for a given
analyte concentration range because the error
variance changes with analyte concentration;
however, the entire concentration range can be
19
-------
partitioned into intervals and a step function is
applied to those intervals in such a way that
each step value is the squared error, E^e2), for
a specific interval of the entire concentration
range. Details of the step function are con-
tained in the MASS QA report on the analytical
laboratory data (Appendix J in Byers et at., in
press).
For most data interpretation uses, it is
preferable to convert tha squared uncertainty
values into the standard reporting units of the
parameters of interest. This is performed by
calculating the square root of the squared
error. To evaluate the data uncertainty as-
sociated with the measurement system and in
the population of routine data, the VEj(^)
terms from each interval were pooled and
weighted by the proportion of routine sample
concentrations within the corresponding inter-
vals. The resulting uncertainty estimate, 6 or
delta, is defined as:
5 - I rVE,(f?) + 1 Pj
where: 5 is :n« pooieo uncertainly for a!i
intervals over the concentration range,
P, is the proportion of routine sam-
ples in the ith interval, and
E^e2) is the mean squared error for
the ith interval, and represents the
sum of the error variance and the
squars of the bias for the ith interval.
For the BD_CLD and BDJ
-------
were too numerous to allow statistically-based
sampling of all soil components. As an alter-
native, soils were grouped into sampling
classes belie ^d to have similar physical and
chemical properties in relation to soil respon-
ses to acidic deposition. Field sampling sites
were randomly selected frum the areas of
occurrence •:•>' each sampling class because
the object e ol tns sampling was to charac-
terize sampan ciasses rather than individual
'.vatershedjj or soil types. Prior to the WAGS,
:n«, ~'ii;a , -,v h:;tc;og>3rh;i*v of nhysicai and
chemical ;oi spercses for indHdua! sampling
•Ji . s&s ?stv! '•; ' ,' b ">.'*"'cn wr"3S -.'/as not know:';.
>' .•..<. lysis, ui tne two previous
ji ;nP surveys however, has shown that
oiasiiiert'tsiii i;j«v.$n.ainfy in the routine sam-
ples is cor>.K"o, ^to/'y parafnetar end quantitatively com-
i',.-.red vv ih", e."uf,si'>e of uncertainty stemming
:j''iounded fi&id sampling and prepara-
ror, T';are ara as many as 22 primary
' types associated with each of the 15
rrfj classes, forming a total of 148 dis-
tinct {>ai ^filing c!ass/S"iori7on groups. Taole
3-2 pffusems the number and percentrge of
primary horiion types and the number of
Table 3-2. Primary Horizon Type® for Sampling
Class/Horizon Groups
sampling
clause:-";; associated with each
The uncertainty estimates are generally
expected to increcse with increasing sources
of confounded error. For this report, two
types of deits values have been calculated: 83
and 84 values. The S( and 82 values could only
be calculated using data from the LAP/LAL and
PD sarfipws v;hk;S', were not analyzed at the
("""epacati'ito laboratory. ITia 83 values were
caiculetrd '-ssing data rrom the field duplicate
sampieK •;; :he bulk density replicates, and t;
-------
Table 3-3. Samples Placed In Mineral and Organic
Soil Batches
Samples per batch
Sample description Mineral Organic
FD Field duplicate sample
FAL Field audit sample, low-range
FAP Field audit samples, pair
FAO FielJ audit samples, triplicate
PD Preparation duplicate samples
LAL Laboratory audit sample, low-range
LAP Laboratory audit samples, pair
LAO Laboratory audit samples, triplicate
OCA Quality control audit sample
RS Routine samples
1
1
2
0
2
1
2
0
1
30
1
0
0
3
2
0
0
3
1
30
audit samples were stored in amber Nalgene
bottles identical to those used for the routine
samples and were not opened or disturbed in
any manner at the preparation laboratory.
Within each batch, the preparation laboratory
manager randomized the samples and labeled
each with a DDRP Label B (see Appendix A).
The only exception to the random placement
was the QCA sample which was always
designated the 15th sample in a batch.
Due to a limited volume of some audit
sample materials at EMSL-LV, e.g., the Bw
horizon audit sample, the volume of sample
provided for each audit sample was some-
times less than that provided for the associ-
ated routine samples in the batch. Therefore,
the identity of these audit samples may not
have been as anonymous to a laboratory
analyst as originally intended. An attempt was
made to disguise low volume audit samples by
purposely placing a few low volume routine
samples within these batches.
At the preparation laboratory, there were
two distinct groups of measurement quality
samples used in the sample analysis: quality
evaluation (QE) samples and quality control
(QC) samples. The QE samples allowed an
independent assessment of data quality by the
QA staff, while the QC samples enabled the
laboratory to control measurement error. In
order to assess the MQOs, both types of
samples were analyzed randomly with the
routine samples in a manner that was statisti-
cally relevant and in which conclusions con-
cerning the quality of data could be made.
During analysis at the preparation laboratory,
an attempt was made to place each type of
QE/QC sample within each analytical run.
However, due to the priority of the creation of
batches for the contract analytical laboratories
and the fact that not all types of QE/QC
samples were available at the time of analysis,
some analytical runs did not contain the full
suite of QE/QC samples defined below.
Quality Evaluation Samples
The preparation laboratory QE samples
are those samples which were submitted blind
to the laboratory staff, i.e., the laboratories did
not know the predetermined concentration of
the samples. The QE samples provided an
independent check on the analytical process
and were used at the preparation laboratory to
evaluate whether the MQOs set forth in the QA
Plan (Papp et al., 1989) were satisfied for all
analytical runs of samples. Every QE sample
had a specific purpose in the data assessment
scheme and each was similar to the routine
samples both in matrix and in analyte con-
centration. The QE natural soil audit samples
had been previously characterized by a suffi-
cient number of analytical laboratories to allow
the definition of accuracy windows used in
data assessment. The QE samples processed
and analyzed at the preparation laboratory
included field duplicate samples, field audit
samples, and a laboratory manager's sample,
as described below.
Field Duplicate Samples - One horizon per
sampling crew for each third day of sampling
was sampled in duplicate using an alternate
trowel-full method as specified in the field
sampling protocols (Appendix A in Kern and
Lee, in press). The first sample of the pair
was considered the routine sample and the
second sample was considered the field
duplicate. Both samples underwent the same
preparation steps as all associated routine
samples. Data from the field duplicate/routine
pair allowed an estimate to be made of the
combined uncertainty due to field sampling
and sample preparation and was used to
assess precision within each analytical run.
Field Audit Samples - Three types of field
audit samples were used in the evaluation
of measurement quality for mineral and or-
ganic sample runs of PH_MP, OM_LOI, and
MOIST_P. At least one pair of FAP samples
and associated FAL sample were placed in
each run of mineral samples. Where possible,
22
-------
a triplicate of FAO samples was placed in the
organic sample runs. The FAP/FAL samples
were processed by the sampling crews when
sampling every third mineral soil pedon, and
the FAO samples were processed whenever an
organic soil pedon was sampled. The sam-
pling crews treated the samples as if they had
just been obtained from the excavated pit,
hence, the samples were sieved and bagged at
the time of sampling. The field audit samples
were then handled as if they were routine
samples and underwent all of the subsequent
soil preparation and analysis steps. The
samples were used to assess accuracy and
precision within each analytical run.
Manager's Sample- A manager's sample was
included in each analytical run of PH_MP,
OM_LOI, and MOIST_P. The sample consisted
of natural audit material having a characterized
accuracy window for each parameter, although
the concentration range was unknown to the
preparation laboratory analysts. The mana-
ger's sample was used to assess accuracy
within each run. The samples were imple-
mented into the program in October 1988,
approximately six weeks after sample prepara-
tion was underway, as an additional check on
data quality.
Quality Control Samples
The preparation laboratory QC samples
are those samples which were known to the
laboratory staff and were used to control the
quality of analytical data produced. The QC
samples analyzed at the laboratory included
liquid QC check samples and internal duplicate
soil samples, as described below.
Quality Control Check Samples - Before begin-
ning a run of pH analysis, a quality control
check sample (QCCS) of pH 4.0 was analyzed
using a different batch of National Institute of
Standards and Technology (MIST) stock buffer
solution than that used for the meter calibra-
tion. Subsequently, after every 10 samples and
after the final sample of the run, another
QCCS of pH 4.0 was analyzed. The value for
each QCCS was required to be 4.00 ± 0.05; if
not, the electrode was recalibrated according
to the guidelines in the PLSOP. This procedure
was followed consistently during the operation
of the laboratory.
Internal Duplicates - Within each analytical
run of samples for the PH_MP, OM_LOI, and
MOIST_P analyses, a composite internal dupli-
cate subsample of a routine bulk soil sample
was obtained and analyzed for every tenth
sample in the run. The mean value of the
internal duplicate sample and its correspond-
ing routine sample was required to show a
standard deviation or a percent relative stan-
dard deviation less than that specified in Table
3-1. If these criteria were not satisfied, the
data were reviewed by the laboratory manager
and QA manager, and a decision was then
made whether to re-analyze any of the sam-
ples.
Technical Systems Audits
Formal technical systems audits of the
preparation laboratory were conducted im-
mediately before the initial receipt of samples
and twice during the operation. The audit
team consisted of the QA officer, QA manager,
and other representatives of the EMSL-LV QA
staff and representatives of the ERL-C sam-
pling QA staff.
During the pre-operational audit con-
ducted in September 1988, the audit team
toured the laboratory facilities, including the
sample drying room and the mobile laboratory
units. During the first operational audit in
October 1988, the audit team observed bulk
sample preparation including the sample
receipt and storage, sample drying and analy-
sis, disaggregation and sieving, and soil homo-
genization and subsampling operations.
During the second operational audit in March
1989, the team evaluated the bulk density
analysis and reviewed the concerns addressed
in the previous audits. Overall, the systems
audits did not identify any major flaws in the
methods or operations used at the preparation
laboratory facility. The audit reports are
presented in Appendix G.
A second type of systems audit was
instituted in which an EMSL-LV QA staff repre-
sentative visited the preparation laboratory
facility approximately every two weeks. This
allowed minor difficulties to be resolved before
they had an adverse effect on the preparation
activities.
23
-------
Data Verification
A computer program was developed that
allowed entry and verification of preparation
laboratory data from the raw data forms. The
program also tracked soil samples through the
various sample processing and analysis ac-
tivities, and assigned verification flags to
discrepant data. A remote modem assembly
allowed concurrent data assessment by the
QA staff. A summary of the preparation
laboratory verification procedures and flags is
presented in Appendix B.
After the laboratory analysts entered raw
data on the forms, an evaluation procedure
was initiated which identified coding discrep-
ancies and data that did not satisfy the MQO
criteria previously outlined in Table 3-1. The
flags identified in Appendix B were used to
qualify the discrepancies. All outliers were
confirmed or corrected before the data were
assigned to the verified preparation laboratory
data base.
The summary data form (DDRP Form
101) for each batch was reviewed by QA staff
to ensure that the sample codes corresponded
to those entered on the field data and raw
data forms. The inclusion of appropriate
QE/QC samples was checked and recorded for
each batch. Analytical data were checked for
completeness and correctness. The prepara-
tion laboratory manager was instructed to
acknowledge and correct any errors in the
data while reviewing the raw data forms,
control charts, or logbooks. The batch sample
shipping form (DDRP Form 102) was compared
with the summary data form to verify the
analytical laboratory destination and the num-
ber of samples shipped. The data base mana-
ger was notified by the QA staff of any
changes made to the data base as a result of
these checks.
Data verification and corrective action for
the preparation laboratory was accomplished
primarily through the application of a QA
reanalysis template (QART) for each parameter
of interest (see Appendix H). The QARTs
provided an assessment checklist of QE/QC
samples for each analytical run and were
designed to assess data according to prede-
fined precision limits and accuracy windows.
The templates were used to make decisions
on whether to request reanalysis of a par-
ticular parameter.
The QARTs were constructed as spread-
sheets, where the row entry fields specified
the evaluation criteria and the column entry
fields represented the severity of the measure-
ment uncertainty that had occurred. The first
column was used to identify major measure-
ment uncertainty and the second column
identified minor uncertainty. When major
uncertainty occurred, the QA manager informed
the preparation laboratory manager to check
for errors in the data for the suspect run and
parameter before initiating reanalysis. If the
laboratory confirmed the suspect values, then
reanalysis was requested for all samples in
the run. The following types of measurement
uncertainty were considered major for all
parameters measured at the preparation
laboratory and constituted the basis for re-
questing confirmation/reanalysis:
• Precision/accuracy - The pair of FAP
samples and the FAL sample (or the tripli-
cate of FAO samples) were grouped to
allow three evaluations. The FAPs and
FAQs were evaluated for precision and
accuracy, and the FAL was evaluated for
accuracy. Major uncertainty existed if any
two of these three evaluations failed to
satisfy the MQOs.
• QC check sample - Major uncertainty
existed if a QCCS exceeded the estab-
lished control limits.
• Cumulative minor uncertainty - Three or
more occurrences of minor uncertainty
were considered to be equivalent to major
uncertainty.
The following types of measurement un-
certainty were considered minor for all para-
meters measured at the preparation labora-
tory:
• Precision/accuracy - Minor uncertainty
existed if one of the three precision/
accuracy evaluations did not satisfy the
MQOs, or if the FD, internal duplicate, or
manager's sample did not satisfy the
MQOs.
For each QART, a table was constructed
to summarize the QE/QC samples analyzed
within an analytical run. For each type of
sample, an asterisk next to a precision value
or the letter "N" designates a failure of that
24
-------
value to satisfy the specific MQO. The final
two columns, one for the QA manager and
another for independent observation by a
representative of the Environmental Research
Center, University of Nevada at Las Vegas,
were used to designate pass or fail for the
run. Although some runs did not contain all
possible types QE/QC samples, it was con-
cluded that all runs passed the QART criteria.
Tables 3-4, 3-5, and 3-6 summarize the QART
data, by run, for the PH_MP, OMJ.OI, and
MOIST_P parameters, respectively. ~
Although the verification programs iden-
tified values exceeding the MQO criteria for
precision and accuracy, trends in the data
could not be discerned. Control charts were
used to detect trends, such as the data from
the B and Bw horizon FAR samples for PH_MP.
It was determined that a pH reference elec-
trode was replaced about the time that the
shift was detected (approximately October 30,
1988). Control charts spanning the duration of
the MASS sample preparation are presented in
Appendix I.
25
-------
Table 3-4. Summary of Achievements for Acceptance Criteria of Field-Moist pH (PH_MP)
Date of
run
9/17/88
9/20/88
9/21/88
9/23/88
9/24/88
9/28/88
9/29/88
10/1/88
10/4/88
10/6/88
10/8/88
10/13/88
10/15/88
10/19/88
10/22/88
10/27/88
10/31/88
11/3/88
11/7/88
11/11/88
11/15/88
11/19/88
11/23/88
12/3/88
12/10/88
12/15/88
12/20/88
12/28/88
1/3/89
1/10/89
FAP
[acc/prec]
(0.15 SD)
—
—
—
Y
0.08
Y
0.00
YY
.06 .05
—
YY
.04 .06
Y
0.04
Y
.01
YY
.04 .01
Y
.04
Y
.00
Y
.08
YY
.01 .01
Y
.01
Y
.04
YY
.01 .00
YY
.02 .01
YY
.01 .03
YY
.00 .02
YY
.01 .01
Y
.04
YY
.06 .09
Y
.04
YY
.01 ,01
Y
.01
Y
.01
FAL
[ace]
—
—
—
Y
Y
YY
—
YY
N*
Y
YY
Y
Y
Y
^Y
Y
Y
YY
YY
YY
YY
YY
Y
YY
Y
YY
Y
Y
QCCS
[ace]
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Int Dup
[prec]
(0.15 SD)
0.03
0.00
0.26*
0.07
0.04
0.06
0.06
.00 .01
.00
0.01
0.68*
.02 .01
.08
.04
.06 .01
.00
.02 .01
.01 .01
.01
.01
.02 .09
.01 .01
.05 .00
.00
.16* .01
.04 .01
.00 .01
.04 .01
.11 .09
.12 .00
.01 .02 .01
.04 .03
.01 .02
.03 .00
.05 .00 .06
.02 .01
.00 .06 .05
.01 .06
.06
.03
.01 .00
.03
.01 .00
.00
.01 .01
.02 .10
.00 .01 .01
.01 .01 .01
.01 .01
.12 .00
.00 .01
.01
FD
[prec]
(0.20 SD)
—
—
—
0.02
0.00
0.00
0.03
0.02
0.00
0.08
0.01
0.03
0.03
0.09
0.03
.29*
.13
.06
.04
.03
.03
.13
.07
.08
.01
.08
.04
.07
.16
.11
.04
.01
.03
.07
.06
.10
MS
[ace]
—
—
—
—
—
—
—
—
—
—
—
—
—
—
Y
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
Y
Y
Y
Pai
LESC
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
B
ERG
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
26
-------
Table 3-5. Summary of Achievements for Acceptance Criteria of Organic Matter (OM_LOI)
Date of
run
9/20/88
9/23/88
9/27/88
9/28/88
9/30/88
10/3/88
10/7/88
10/8/88
10/14/88
10/15/88
10/19/88
10/22/88
10/25/88
10/28/88
11/2/88
11/8/88
11/10/88
11/15/88
11/18/88
11/23/88
11/29/88
12/09/88
12/10/88
12/15/88
12/19/88
12/22/88
12/23/88
12/31/88
1/6/89
1/14/89
1/21/89
FAP
[acc/prec]
(15% RSD)
—
—
Y
1.0
Y
2.4
YY
3.7 0.0
YY
1.6 1.6
Y
3.8
Y
0.01
YY
0.8 1.5
—
Y
1.6
Y
2.2
YY
4.2 0.9
YN
39.6* 5.1
Y
1.2
YY
1.0 2.3
Y
1.9
YY
1.6 0.2
YY
0.7 0.9
YY 6.8
Y 0.9 1.5
Y
0.2
-
Y
5.8
Y
5.0
YY
2.1 1.6
Y
0.8
Y
3.0
FAL
[acc]
—
—
Y
YY
YY
Y
Y
YY
—
Y
Y
YY
YY
Y
YY
YY
YY
YY
YYY
Y
-
Y
N
YY
Y
Y
Int Dup
iprecl
5% RSD)
—
—
3.5 2.8
18.4* 1.1
0.1 2.9
2.4 1.1
0.2 1.6
7.5
0.3 1.4
4.0 1.9
0.4
2.0
2.4
2.0 6.5 8.1
7.0 1.9
11.6 2.4
10.7 1.3
1.0 1.8
0.5 7.4
21.7*
1.9 2.4
8.9
2.9
11.0
1.4
2.5 5.1
5.6
8.3 12.6
0.9
0.5
2.7 0.3
2.4
1.1 1.8
2.4 13.6
0.6 5.4
16.3* 6.9
2.1
4.1
2.1
FD
(20% RSD)
—
—
4.7
17.5
4.6
7.1
1.2
1.6
0.8
1.8
8.4
11.6
14.3
2.1
0.3 0.4
2.0
15.8
3.6
0.8
0.7
0.6
2.7 0.1
3.0
17.7
12.4
13.6
4.2
24.6* 7.7
2.4
3.7
1.1
2.6
5.9
3.8
11.5 5.7
1.0 0.3
2.4
10.9
MS
[acc]
—
—
y
Y
Y
Y
N
Y
y
y
Y
y
y
Y
y
N
y
y
Y
Y
Y
Pa
LESC
Y
Y
Y
Y
Y
y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
y
Y
y
y
y
Y
Y
y
18
ERC
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
27
-------
Table 3-6. Summary of Achievements for Acceptance Criteria of Air-Dry Moisture (MOIST_P)
Date of
run
9/20/88
9/22/88
9/24/88
9/27/88
9/28/88
10/1/88
10/7/88
10/8/88
10/12/88
10/18/88
10/21/88
10/22/88
10/25/88
10/29/88
11/3/88
11/7/88
11/11/88
11/15/88
11/19/88
11/23/88
12/3/88
12/10/88
12/15/88
12/20/88
12/28/88
1/3/88
1/10/89
1/18/89
FAP
[acc/prec]
(151 RSO)
—
—
—
Y
1.3
YY
1.5 1.1
YY
6.1 119*
Y
5.6
YYY 3 . 7
2.7 3.2
Y
1.5
YY
0.6 0.6
YY
4.6 2.2
Y
0.0
YY
3.6 2.1
YY
1.5 0.8
YY
1.4 0.8
YY
2.6 4.3
Y
0.5
Y
9.8
Y
4.5
Y
0.4
YY
6.6 5.4
Y
6.3
Y
3.4
PAL
[ace]
—
—
—
Y
YY
YY
Y
YYY
Y
YY
YY
Y
YY
YY
YY
YY
Y
Y
Y
Y
YY
Y
Y
Int Dup
[prec]
(151 RSD)
—
—
—
—
3.0 10.2
2.3 7.7
3.0 3.2
3.2 7.3
43.6*
3.3
3.3
1.0
3.6 4.9
3.4
7.1 2.1
0.5 0.6
4.4 9.7
1.7 1.1
4.9
3.1
20.3* 0.3
1.4 2.4
11.2 3.9
8.1
4.0
2.4
0.3 6.1
1.3
15.7* 3.1
2.7
0.6 13.9
5.7
2.9 9.3
2.8
3.0
7.2
FD
[prec]
(20% RSD)
—
—
—
9.0
2.3
6.0
2.9
2.1
8.5
19.1
14.8 3.8
0.8
23.6*
20.8*
12.1
17.4
14.4
7.8
3.5
23.9*
7.1
1.6
32.2*
4.5
10.1
3.2
2.2
10.9
6.5
18.7
24.0*
MS
[ace]
—
—
—
Y
—
Y
Y
—
Y
Y
Y
Y
Y
Y
Y
Y
Y
Pae
LESC
Y
Y
Y
Y
8
ERC
Y
Y
Y
Y
ONE SAMPLE
ANALYZED
Y
Y
Y
Y
TWO SAMPS
ANALYZED
THREE SAMPS
ANALYZED
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
28
-------
Section 4
Results and Discussion
This section provides the results and
discussion for data quality assessment of the
preparation laboratory data. Achievement of
the MQOs is expressed in relation to the
actual MQOs specified in the MASS QA Plan
(Papp et al., 1989).
Evaluation of Quality Attributes
The preparation laboratory data were
assessed in a systematic manner using statis-
tical results from the measurement quality
samples.
Precision
With the exception of the bulk density
and rock fragment data, it was determined
that concentration dependency would not
affect the presentation of pooled imprecision
estimates. The precision information for the
field audit samples are presented individually
for each horizon type, thereby eliminating the
concentration dependency issue. The field
duplicate and internal duplicate observations
were within very narrow ranges, therefore,
precision was not strongly affected by con-
centration.
Statistical analysis of precision was
performed for the PH MP, OMJLOI, MOIST P,
RF_FG, RF_MG. BD_CL.D, and BD_KV parame-
ters. Because bulk~density and rock fragment
data displayed some concentration dependen-
cy, these data were evaluated separately for
values above and below the mean. A root
mean square error statistic was used to
estimate the pooled standard deviation (SD)
or, where appropriate, the pooled relative
standard deviation (RSD) across each labora-
tory data set. The percent RSD was deter-
mined by dividing the SD by the mean of the
data set, and multiplying this quotient by 100.
Field-Moist pH (PH_MP)
The pH parameter was included in the
MASS for two reasons: (1) to compare field-
moist soil pH values with air-dry soil pH values
that were obtained during contract laboratory
analysis, and (2) to indicate the possible
presence of inorganic carbon. Rather than
perform an inorganic carbon test on every bulk
sample as was done in the NE and SBRP
surveys, the test was required only on sam-
ples of pH 7.0 and higher.
Summary statistics of the pooled im-
precision estimates for the measurement
quality samples are provided in Table 4-1. The
imprecision estimates for analysis of field-
moist pH are well within the specified MQOs
for the different measurement quality samples
that were analyzed. The data did not exhibit
concentration dependency because the pH
values occupied a narrow range of concentra-
tion.
Table 4-1. Imprecision Estimates for PH_MP
Sample
type
FD
FAP (Bw)
FAR (B)
Int Dup
n-pairs
34
20
19
92
Mean
4.84
5.14
4.93
4.93
SD MQO
— pH units —
0.08 0.20
0.03 0.15
0.03 0.15
0.05 0.15
Organic Matter by
(OMJLOI)
Loss on Ignition
This parameter was measured to distin-
guish more accurately organic soils from
mineral soils. Organic and mineral soils were
placed in separate batches to expedite analy-
sis at the analytical laboratories. All samples
containing 20 percent or more organic matter,
even though identified in the field with mineral
29
-------
soil designators, were placed in organic soil
batches.
Summary statistics of the pooled im-
precision estimates for the measurement
quality samples are provided in Table 4-2. The
imprecision estimates for analysis of organic
matter are well within the specified MQOs.
The higher RSD value for the FAR Bw can be
explained by one outlier pair. This pair had
individual values of 8.34 and 4.69, with the 8.34
value being suspect. Due to a lack of avail-
able sample in archive, this sample was not
reanalyzed although the raw data sheets and
entry were scrutinized for transcription errors.
If these two samples are removed from the
summary calculation, the RSD for the remain-
ing 19 pairs is 1.63 percent.
Table 4-2. Imprecision Estimates for OMJ.OI
Table 4-3. Imprecision Estimates for MOIST_P
Sample
type
FD
FAR (Bw)
FAR (B)
Int Dup
RSD MQO
n-pairs
46
20
19
64
Mean
9.02
4.72
10.16
7.56
— percent
6.68
12.99
3.48
9.48
20
15
15
15
Air-Dry Moisture (MOIST_P)
Summary statistics of the pooled im-
precision estimates for the measurement
quality samples are provided in Table 4-3. The
average RSD for the FD/routine pairs is greater
than the MQO. However, the only requirement
in the PLSOP was that the laboratories dry
samples to less than 2.5 percent for mineral
and 6 percent for organic samples. Once this
was accomplished the samples were con-
sidered ready for further processing. For
example, a routine sample containing a mois-
ture content of 1.0 percent and its associated
field duplicate containing a moisture content
of 2.0 percent were both acceptable even
though the pair had a calculated RSD of 47
percent.
Rock Fragments (RF_FG and RF_MG)
Summary statistics of the pooled im-
precision estimates for rock fragment content
are provided in Table 4-4. Since the audit
Sample
type
n-pairs
Mean
RSD MQO
— percent —
FD
FAR (Bw)
FAP(B)
Int Dup
46
20
19
64
1.85
1.84
3.37
1.09
21.57
3.73
4.36
11.71
20
15
15
15
Table 4-4. Imprecision Estimates for RF_FG, RF_MG
Parameter
RF FG"
> mean
< mean
RF MG°
> mean
< mean
n-pairs
46
24
22
46
24
24
Mean
7.86
12.50
2.85
15.74
27.25
5.66
RSD MQO
— percent —
19.04
13.76
42.69
16.61
13.33
22.87
20
20
" Estimates based on field duplicate/routine pairs.
samples had already been sieved, the only
samples for which imprecision could be mea-
sured were the FD/ routine pairs. The inten-
tion was to identify only those extreme outly-
ing values where sampling or preparation
errors might have occurred.
The RSD values less than and greater
than the mean were evaluated in order to
determine whether or not the statistical rela-
tionship of higher RSDs at lower concentra-
tions, in this case at lower rock fragment
content, would hold true for rock fragment
data. The RSDs presented in the table appear
to confirm this relationship. Because of the
simplicity of the method used for determining
the percentage of rock fragments in the bulk
soil samples, a greater amount of the im-
precision can be attributed to spatial within-
horizon variability or sampling error and a
lesser amount to preparation laboratory error.
Data for the two rock fragment size
classes were quite variable, especially at low
concentrations of rock fragments. For ex-
ample, one 20-mm rock fragment in one of the
two samples compared (when both samples
had low rock fragment contents) could show
an inordinate degree of variability. This situa-
tion is not unusual when considering the
30
-------
natural spatial heterogeneity of soil horizons.
Any pairs having larger than 20 percent RSD
were checked for transcription errors and the
fragments were reweighed.
Bulk Density (BDjCLD and BDJCV)
Summary statistics of the pooled im-
precision estimates for bulk density are provi-
ded in Table 4-5. Imprecision was determined
using replicates of each horizon sampled.
Clods were usually sampled in triplicate while
known volume samples were usually sampled
in duplicate. Due to the highly variable nature
of the bulk density samples, i.e., either struc-
turally intact or reconstituted, the preparation
of an audit sample was not possible. There-
fore, imprecision was estimated solely on the
replicate data.
Table 4-5. Imprecision
Parameter n-pairs
BD CLD" 705
< mean 347
> mean 358
BD KV* 145
< mean 75
> mean 70
Estimates
Mean
1.51
1.28
1.74
0.78
0.42
1.18
for BD_CLD, BD_KV
RSD MQO
— — osrcsnt -----
7.52 20
10.04
5.59
17.52 20
23.77
14.36
° Estimates based on clod replicates.
* Estimates based on known volume replicates.
The RSD values less than and greater
than the mean were evaluated in order to
determine whether or not the statistical rela-
tionship of higher RSD at lower concentrations,
in this case at lower bulk densities, would hold
true for bulk density data. The percent RSDs
presented in the table appear to confirm this
relationship. It is also evident that the known
volume sample mean values are considerably
lower than the clod mean values and that the
variability is higher. The lower means were
due to the collection of known volume samples
from horizons containing more organic matter,
unnatural voids, or rock fragments, all of which
generally tend to lower bulk density. The
techniques used in collecting known volume
samples are also considered to be less pre-
cise than the technique used to collect clods.
The data suggest a consistency of bulk
density values within a given horizon. Audit
reports indicated that the sampling crews
were able to select representative clods from
each horizon and that the preparation labora-
tory was consistent in its measurement
techniques. However, the exact percentage
of error in the data due to horizon variability,
sampling error, or laboratory error cannot be
determined because: (1) inherent within-
horizon spatial variability made it improbable
that identical clods could be sampled, or that
a clod audit sample for measurement of
potential sampling error could be provided, and
(2) audit samples that could theoretically allow
the estimation of preparation laboratory error
were not provided to the laboratory.
The following types of possible field
sampling errors could affect the data for the
bulk density replicates:
• Collection of replicates from transitional
zones or adjacent horizons,
• Mislabeling of clods,
• Inconsistent saran coating procedure, or
• Spatial variability relating to the coherency
of clods.
Based on the experience level of the SCS
sampling crews and the fact that the various
field audit reports did not indicate major devia-
tions from the protocols, sampling error is not
presumed to have been a significant factor
affecting the RSD values. Some variability in
the use of saran was noted, although the
imprecision estimates would not be affected if
the coating procedure was consistent for all
replicates within a horizon.
Measurement uncertainty may be intro-
duced at the preparation laboratory because of
measurement or method errors, such as:
• Transcription errors, e.g., misrecorded
weights or sample codes,
• Inconsistent saran coating procedures,
• Improper clod handling, e.g., compaction,
• Incomplete drying,
• Loss of soil or rock fragments during
sieving, or
• Faulty weighings, e.g., balance not calibra-
ted or tared.
There was no evidence to suggest that any
such errors occurred during the preparation
laboratory operations.
31
-------
Accuracy
The audit samples used to assess
accuracy were natural soil samples. Although
their "true" chemical composition and physical
characteristics were unknown, accuracy win-
dows were established using soil analytical
data from the previous DDRP surveys and by
preliminary analysis of new parameters added
to the preparation procedures. Accuracy was
based upon the data for FAP, FAL, FAO, MS,
and QCCS samples. Tables 3-3, 3-4, and 3-5
in Section 3 summarized the acceptance or
failure of the samples to satisfy the sample
run acceptance criteria. The audit samples
were also tracked on control charts to deter-
mine if certain trends could be detected (see
Appendix I).
Accuracy could not be assessed for rock
fragment or bulk density data. As the audit
samples were initially prepared and sieved,
there were no rock fragments in the samples
except those that entered the sample via
contamination from sampling or preparation.
Also, there were no audit samples used for
bulk density analyses.
Field-Moist pH (PH_MP)
The accuracy windows for PH_MP were
derived from confidence intervals placed
around a biweight estimation (Kadafar, 1982)
of the mean. In almost all cases, the within-
run inaccuracy for all audit samples satisfied
the MQO for the respective accuracy windows.
Only one FAL outlier and two MS outliers were
recorded. A trend towards higher pH values
was noticed in both the B and Bw horizon FAP
samples on or about October 30, 1988 (see
Appendix I). It was initially suspected that a
change in electrodes or a drop in ambient air
temperature might be responsible for such a
shift, although neither possibility was con-
firmed. However, the trend was not noticed in
the manager's sample nor the FAL sample.
The increase was not great enough to cause
the FAP samples to exceed the limits of their
respective accuracy windows for PH_MP, and
was well within the range of expected error for
multi-phase pH systems (Brezinski, 1983).
Organic Matter Loss-on-Ignition (OMJ.OI)
The accuracy windows for OMJ.OI were
based on a 15 percent RSD confidence limit
placed about the mean values of ten audit
samples at each concentration. Except for
one FAP, one FAL, and one MS, the inaccuracy
estimates for the audit samples satisfied the
QART criteria for each analytical run.
Air-Dry Moisture (MOIST_P)
The initial accuracy criteria for MOIST_P
were that the values should be below 2.5
weight percent for mineral soils and 6.0 weight
percent for organic soils. All audit samples,
with the exception of the manager's sample
and the B horizon FAP sample, satisfied these
criteria. Both types of samples were allowed
to dry for six days, i.e., twice the normal drying
period, with no significant change measured in
their mean moisture content of approximately
2.3 and 3.3 weight percent, respectively.
Control charts with 10 percent RSD warning
limits and 15 percent RSD control limits around
the mean values were defined for the samples.
Using these different criteria, all audit samples
satisfied the accuracy objectives for air-dry
moisture.
Representativeness
Representativeness is addressed in the
QA report on the analytical data (Byers et al.,
in press) by using data from the preparation
duplicates to assess the subsampling proce-
dure performed by each laboratory.
Completeness
The requested analyses and soil proces-
sing steps were performed on 100 percent of
the bulk samples that satisfied the sample
receipt and batching criteria. Samples that did
not satisfy these criteria included some groups
of field duplicates and audit samples that
were not sampled according to the field sam-
pling protocols or that were extraneous in
terms of the proportion of duplicate samples
and audit samples to the total number of
batches created. Preparation duplicates were
created from the field duplicate/routine sample
pair in each batch sent to the analytical labor-
atories, for a 100 percent level of complete-
ness. Hence, the MQO for completeness at
the preparation laboratory was satisfied.
Of the 844 routine horizons sampled
in the field, 705 horizons were sampled for
clods, resulting in 83.5 percent completeness.
32
-------
Therefore, the completeness objectives for
sampling were satisfied. Only three of the
2,021 clod samples received at the preparation
laboratory could not be analyzed, resulting in
99.9 percent completeness.
Comparability
Most of the preparation laboratory com-
parability issues are evaluated in the QA report
on the analytical laboratory data (Byers et al.,
in press). However, it was possible to inves-
tigate and compare the imprecision results
from data collected throughout the three DDRP
surveys for those parameters that were com-
mon to the surveys. Only two parameters,
rock fragments and clod bulk density, were
analyzed in all three surveys. Table 4-6 sum-
marizes the data from an inter-survey com-
parison cf the mean percent RSD pooled
across laboratories for the two parameters.
Clearly, the data are very comparable and
probably reflect the best precision that can
be achieved with the specified methods. It
should be noted that the variability in the NE
survey rock fragment data is probably under-
estimated because of differences in the field
duplicate sampling method (Byers et al., 1989).
Certain sampling crews in the NE survey
created a "split" duplicate sample, while others
sampled the duplicate using the (intended)
alternate trowel-full method.
Table 4-6. Multiple Survey Comparison of Imprecision
for Rock Fragments and Clod Bulk Density
Pooled RSD°
Parameter
Rock fragments*
Clod bulk density5
< mean
> mean
NE
13.21
8.32
8.97
6.38
SBRP
14.83
10.84
14.72
7.61
MASS
17.83
7.76
10.04
5.59
' Pooled relative standard deviation, in percent, across
laboratories (NE=4, SBRP=2, MASS=1).
* Data from field duplicates, including both fine and
medium gravel combined.
' Data from clod replicates.
Assessment of Data Uncertainty
Components
The following subsections describe
various data uncertainty components that are
estimated using data from the measurement
quality samples. The effect of measurement
uncertainty on overall data uncertainty is also
discussed.
Additive Components of Uncertainty
As described previously in Section 3, the
within-run imprecision estimates were pooled
over the concentration range using the propor-
tion, P(i), of routine data in each interval to
derive the within-run imprecision component of
the delta values. Also, it was assumed that
the laboratory analyzed the soil samples on a
run-by-run basis under conditions specific to
each run's analysis. Hence, the between-run
imprecision was calculated for the PH_MP and
OMJ.OI parameters using the data from the
FAR samples. Bias for these parameters was
calculated from the FAL samples and the
mean values of the FAR sample pairs. The
within- and between-run variances were added
to the squared bias term and were converted
into reporting units to produce delta values for
each data set and parameter. Table 4-7
shows the imprecision and bias estimates and
the contribution of the between-run and bias
components in relation to the within-run im-
precision estimates for the analytical com-
ponent.
The within- and between-run mean sums
of square and the squared bias component of
analytical uncertainty can be experimentally
compared with the three mean squares of a
randomized block design (Box et al., 1978). In
this scenario, both the expected value of the
squared bias and expected value of the bet-
ween-run mean square includes the expected
value of the within-run mean square as calcu-
lated from the field audit samples. Ratios are
used to highlight those parameters which
indicate a considerable contribution of the
33
-------
Table 4-7. Comparison of Measurement Uncertainty Components
Parameter
PH MP
OMJ.OI
Reporting
units
pH units
wt. %
Within-run SO
—..— units — — —
0.0285
0.5484
Between-run SD
linjtg .««.
0.0842
0.2619
Bias
- units -
0.0119
0.5223
Ratio
B-R/W-R
3.0
0.5
Ratio
Bias/W-R
0.4
1.0
NOTE: Comparisons were not applicable for MOIST_P, RF_FG, RF_MG, BD_CLD, and BD_KV parameters.
between-run and bias components to the
within-run component, as shown in the table.
If the between-run mean square is two times
larger than the within-run mean square, then
the between-run imprecision can be inferred to
be as large as the within-run imprecision.
Similarly, if the between-run mean square is
more than three times larger than the within-
run mean square, then the between-run im-
precision can be inferred to be significantly
larger than the within-run imprecision.
Hence, if the ratio of the between-run
component with the within-run component in
reporting units exceeds Vs, i.e., 1.73, then the
between-run imprecision can be considered
significant in relation to the within-run impre-
cision. The same criteria can be used to
examine the significance of the bias in relation
to the within-run imprecision. Under this
framework, there is no evidence of significant
between-run or bias uncertainty if the ratio is
less than 1. If the ratio is between 1 and 1.73,
the observed uncertainty is negligible and
could be due simply to random fluctuation. If
the ratio is greater than 1.73, then the bet-
ween-run or bias component could be as large
as 50 percent of the within-run component.
The ratio of between-run to within-run
was greater than 1.73 only for the PH_MP
parameter. Possible reasons for this parame-
ter to show substantial between-run error
could relate to the unique aspects of its meth-
od of analysis, e.g., composite subsampling of
a nonhomogeneous field-moist bulk sample, or
very precise within-run results which could
magnify the effect of between-run imprecision
and yield a high ratio. The PH_MP within-run
imprecision was so small, i.e., less than one-
sixth of the MQO, that the between-run com-
ponent appears to be substantial. In actuality,
both types of uncertainty are very low. There-
fore, although the contribution of the between-
run uncertainty to total analytical uncertainty
might be substantial, the magnitude of the
uncertainty is relatively small. The bias was a
relatively small component of uncertainty, and
all ratios of bias to within-run imprecision were
below the critical value of 1.73.
Effect of Measurement System on
Data Uncertainty
Table 4-8 provides the S3 and 84 values of
data uncertainty for the preparation laboratory
parameters. As expected, the 83 estimates
were much lower than the 84 estimates, which
are based on the field duplicate and sampling
class/horizon data sets, respectively (NOTE: 8,
and 82 values could not be calculated for the
preparation laboratory parameters). The 83
values contain the confounded overall mea-
surement uncertainty of field sampling and
sample preparation and analysis. The 84
values contain uncertainty due to the spatial
heterogeneity of the routine sample population
confounded with the overall measurement
uncertainty of field sampling and sample
preparation and analysis.
Table 4-8. Delta Values and Ratios for Assessment
of Data Uncertainty Components
Parameter
PH MP
OM LOI
RF FG
RF MG
BD CLD
BD KV
Reporting
units
pH units
wt. %
wt. %
wt. %
g/cm3
g/cm3
Mean"
4.72
11.93
7.36
13.47
1.42
0.78
Delta values*
83 84
0.1154
0.7820
1.5575P
2.0666P
0.1125P
0.2837P
0.5047
3.5038
5.8436
11.8267
0.1824
0.7431
Ratio
63
1
•)
1
1
1
1
S4
4
4
4
6
2
3
NOTE: Delta values were not applicable for MOIST_P.
• Mean value of routine samples.
* 83 calculated from field duplicates and 84 calculated
from sampling class/horizon groups of routine samples.
p Partial estimate because of incomplete information.
34
-------
It has been proposed that a measured
value can be considered as essentially error-
less for most uses if the uncertainty in that
value is one-third or less of the permissible
tolerance for its use (Taylor, 1987). The QA
staff was interested in determining the effect
of measurement uncertainty on overall data
uncertainty in the routine samples. The overall
measurement uncertainty was based on the S3
values calculated from the field duplicate
samples. The magnitude of this uncertainty in
relation to the associated S4 data uncertainty
values generated from the sampling class/
horizon groups of routine samples provides the
data user with a basis for assessing the
contribution of the measurement system to the
data uncertainty. For the purposes of this
assessment, the data uncertainty confounded
in the sampling class/horizon groups has been
considered to be a surrogate for user-defined
error tolerance values. Where 83 is one-third or
less of S4, measurement uncertainty is a negli-
gible contribution to the overall data uncertain-
ty.
Under this framework, measurement
uncertainty is negligible for all parameters
except BD_CLD. The relatively high measure-
ment uncertainty can probably be attributed to
naturally high spatial variability within a given
horizon in a given pedon due to the irregular
presence of rock fragments, pores, roots, or
localized compaction. Nevertheless, the mea-
surement uncertainty attributed to the data is
only a fraction of the allowable MQO for
BD CLD.
35
-------
Section 5
Conclusions and Recommendations
The conclusions and recommendations
discussed below have been summarized from
information supplied by preparation laboratory
personnel, QA staff, and other survey par-
ticipants. The recommendations are presented
for consideration and possible implementation
in future surveys.
Laboratory Operations
The MASS preparation laboratory opera-
ted in a very efficient manner and prepared
analytical samples of equal or higher quality
than those prepared in earlier DDRP surveys.
This quality can be attributed to the centralized
preparation facility and to the many QA proce-
dures that were implemented for the MASS,
including the ventilated sample drying room,
improved cleaning procedures, increased QA
oversight, and the skill level of the laboratory
personnel.
Laboratory Design
The laboratory design greatly increased
the efficiency of the preparation operations.
Sample receipt, storage, and supplies were
within the facility for quick access. The facility
was designed to allow samples to flow direct-
ly from one operation to another.
Equipment and Supplies
Originally, the saran:acetone clod dipping
solution was mixed at the preparation labora-
tory and shipped to the sampling crews. The
integrity of the mixture upon delivery to the
crews was not suitable. Much of the acetone
apparently volatilized during transport, thereby
increasing the viscosity of the saran:acetone
solution. It is recommended that pre- weighed
packets be sent to field crews who can then
purchase acetone locally and create adequate
supplies of the mixture as needed. Safety
instructions for the preparation of the mixture
should be provided.
Sample Receipt
The sample receipt procedure greatly
improved detection of samples mislabeled in
the field. // AS recommended that an accurate
list of the watershed and pedon identification
codes be provided, as was done in this survey,
to the preparation laboratory manager prior to
sampling. Laboratory personnel discovered
that pre-labeled field sampling labels, i.e., the
majority of sampling information were coded
on the labels before sample collection, general-
ly contained fewer errors. These labels were
usually more legible, making it easier to iden-
tify codes. // is recommended that the sam-
pling crews perform as much pre-labeling as
possible before going to the sampling site.
Preparation laboratory personnel were
available during the hours of the delivery
service (Saturdays included) to receive sam-
ples from the sampling crews and to check
the validity of the sample labeling using the
sample receipt criteria.
Sample Drying
Due to the low relative humidity and the
use of mesh drying tables, sample drying time
was minimal at the Las Vegas facility, thereby
improving the efficiency of the preparation
laboratory operation. However, it is recom-
mended that more attention be given to the
disaggregation of peds, especially in clayey or
organic soils, before they dry to a hard consis-
tency. These types of peds were extremely
difficult to disaggregate without using exces-
sive force.
36
-------
Field-Moist pH
This was the first of the DDRP surveys
to include field-moist pH as a preparation
procedure. The pH parameter was included for
two primary reasons: (1) to compare field-
moist pH values with values that were ob-
tained on air-dry samples during contract
laboratory analysis, and (2) to use as a pos-
sible indicator of the presence of inorganic
carbon in a soil sample. Rather than perform-
ing an inorganic carbon test on every bulk
sample (a procedure that was performed in
the previous surveys), the test would only be
run on samples a field-moist pH of 7.0 or
higher. The most crucial step in the procedure
appears to be the calibration of the pH instru-
ment. Strict adherence to calibration protocol
is required. // is recommended that several
extra electrodes be available prior to sample
analysis.
Organic Matter by Loss on Ignition
This was the first of the DDRP surveys
to include the determination of organic matter
content as a preparation procedure. Organic
matter was measured in order to more clearly
distinguish organic soils from mineral soils.
The two types of soils were organized in
separate batches for more accurate analysis
at the contract analytical laboratories. Sam-
ples containing greater than 20 percent organic
matter by loss on ignition, irregardless of their
field horizon designation, were placed in or-
ganic soil batches.
There were no major difficulties encoun-
tered with the procedure used to determine
organic matter. When analyzing organic
samples, it is recommended that only enough
samples be placed in the muffle furnace during
each run to prevent excessive amounts of
smoke that accrues as the organic matter is
ignited.
Disaggregation and Sieving
Determining the type and consistency of
soil material that must be disaggregated is
important before initiating the procedure. // is
recommended that saprolitic material or soil
containing highly weathered rock fragments
should be disaggregated with extra caution to
prevent the accidental disaggregation of frag-
ments.
It is recommended that the dispensation
of the rock fragments following analysis be
decided before soil processing begins. This
action will avoid lengthy or costly storage of
the fragments at the preparation laboratory.
Bulk Density
The bulk density procedure was modified
slightly from the procedure used in the two
previous surveys. All clods were oven-dried
before coating the clods with the saran:
acetone mixture at the laboratory. Drying
helped to stabilize the initial weight and the
weight after clod dipping that, in past surveys,
fluctuated due to evaporation of moisture from
the clod during the clod dipping process. A
2000-mL beaker is recommended for determin-
ing the clod weight in water. Three to six dips
in a 1:5 by weight saran:acetone mixture is
sufficient to seal a clod before submersion in
water. It is recommended that the modified
method for submerging floating clods, des-
cribed in Section 3, be added to the protocols.
It is recommended that the density of rock
fragments within each clod should not always
be assumed to be 2.65g/cm3, as is done many
bulk density studies. In areas with a variety
of lithologies or weathering characteristics, an
appropriately lower or higher density value
could be used where necessary. An average
particle density value of 2.47 was used for
clods collected from the Mid-Appalachian
region.
Sample Batching
Care should be taken to ensure the
anonymity of measurement quality samples
within a batch. These samples should be
interspersed in the batch and be identical in
weight and appearance to routine samples.
The laboratory manager inspected and labeled
all samples in each batch. // AS recommended
that all labels in a batch be coded by the
same person, and typed if possible.
Sample Archiving
Due to the inadequate quantity of soil
material available for some audit sample
types, the entire volumes of the field audit
samples were used in the preparation of the
analytical samples. // is recommended that all
audit samples sent to the sampling crews be
of sufficient volume to ensure that a portion of
37
-------
the sample can be archived upon receipt at the
preparation laboratory. This would allow the
QA staff to have a ready supply of additional
soil in the event that the analytical laboratory
depletes its supply.
Quality Assurance Program
Training
The preparation laboratory manager held
a training session every two weeks to review
protocols as well as help the laboratory per-
sonnel in their understanding of soil genesis
and morphology. The sessions increased the
staff's appreciation for the work performed in
the field and was a motivational strategy for
inviting participation in soil identification.
Measurement Quality Objectives
Because the sampling crews did not
collect field audit and field duplicate samples
until the third pedon, the initial runs of the
preparation laboratory parameters did not
contain all the samples necessary to evaluate
the measurement quality objectives using the
quality assurance reanalysis templates. This
was allowed only because of time constraints
to prepare samples for contract analysis. // is
recommended that a full suite of measurement
quality samples be maintained in each run for
consistency in data quality assessment. Field
audit and field duplicate samples should be
processed with the first pedon sampled and
every third pedon thereafter. It is recommend-
ed that the number of internal duplicates be
decreased from one per ten routine samples to
one per run, as was instituted at the DDRP
analytical laboratories.
A two-tiered precision assessment sys-
tem was used for the DDRP contract labora-
tory analysis parameters in which concentra-
tion dependent precision was evident. Depen-
ding upon the data users' requirements for
precision, two-tiered precision objectives may
be suitable for the rock fragment and organic
matter parameters. // is recommended that
the precision criteria for air-dry moisture be
dropped. Acceptance criteria should be based
solely on an allowable moisture content.
Precision of the bulk density and rock
fragment data for the Northeastern region and
Southern Blue Ridge Province surveys are
similar in magnitude to the Mid-Appalachian
survey data.
Technical Systems Audits
The technical systems audit was a very
important phase of the quality assurance
program because it provided documentation as
well as objective insights to the operation.
The pre-operation and two systems audits
conducted during preparation operations were
helpful in clarifying the protocols and correct-
ing minor discrepancies. // is recommended
that the quality assurance staff audits, which
were performed approximately every two
weeks, be conducted on a less frequent basis.
Data Verification
The Mid-Appalachian survey was the first
of the DDRP surveys which utilized a com-
puterized system for entry of all raw data.
The system greatly improved outlier detection
as well as helping to determine progress in
the field and in the preparation laboratory.
Data entry was the responsibility of the labor-
atory manager and one other laboratory ana-
lyst. // is recommended that a person know-
ledgeable of the entry program, other than a
preparation laboratory analyst, be responsible
for the data entry and verification activities.
The immediate entry of sample receipt
data and a subsequent verification of sample
codes helped to identify miscoded samples
within hours of sample receipt. The verifica-
tion program tracked the progress of all sam-
ples through the preparation and analytical
stages. Upon entry of data for any particular
procedure or analysis, the verification pro-
grams were run to identify any suspect data.
Any samples identified by this procedure could
be reanalyzed in the next run of samples. The
verification program improved sample flow and
provided a verified data base in a much shor-
ter timeframe than previous surveys. The
program also greatly increased the efficiency
of outlier detection and correction over the
past surveys.
38
-------
References
Blake, G. R. 1965. In Black, C. A. (ed.)
Methods of Soil Analysis, Part 1. Ameri-
can Society of Agronomy, Madison,
Wisconsin.
Brezinski, D. P. 1983. Kinetic, Static, and
Stirring Errors of Liquid Junction Refer-
ence Electrodes. Analyst, 108, 425-442.
Byers, G. E., R. D. Van Remortel, J. E. Teberg,
M. J. Miah, C. J. Palmer, M. L Papp, W.
H. Cole, A. D. Tansey, D. L. Cassell, and
P. W. Shaffer. 1989. Direct/Delayed
Response Project: Quality Assurance
Report for Physical and Chemical Ana-
lyses of Soils from the Northeastern
United States. EPA/600/4-89/037. U.S.
Environmental Protection Agency, Las
Vegas, Nevada. 216 pp.
Byers, G. E., R. D. Van Remortel, M. J. Miah, J.
E. Teberg, M. L. Papp, B. A. Schumacher,
B. L. Conkling, D. L. Cassell, and P. W.
Shaffer. In press. Direct/Delayed Re-
sponse Project: Quality Assurance
Report for Physical and Chemical Ana-
lyses of Soils from the Mid-Appalachian
Region of the United States. EPA/600/x-
xx/xxx. U.S. Environmental Protection
Agency, Las Vegas, Nevada. 337 pp.
Church, M. R., K. W. Thornton, P. W. Shaffer,
D. L. Stevens, B. P. Rochelle, G. R. Hol-
dren, M. G. Johnson, J. J. Lee, R. S.
Turner, D. L. Cassell, D. A. Lammers, W.
G. Campbell, C. I. Liff, C. C. Brandt, L H.
Liegel, G. D. Bishop, D. C. Mortenson, S.
M. Pierson and D. D. Schmoyer. 1989.
Future Effects of Long- Term Sulfur Depo-
sition on Surface Water Chemistry in the
Northeast and Southern Blue Ridge
Province: Results of the Direct/Delayed
Response Project. EPA/600/3-89/061, U.S.
Environmental Protection Agency, Wash-
ington, D.C., 887 pp.
Cochran, W. G. 1977. Sampling Techniques.
3rd Edition. J. Wiley & Sons, New York.
428 pp.
Kadaf ar, K. 1982. A Biweight Approach to the
One Sample Problem. Jour. Amer. Stat.
Assoc. 77(378):416-424.
Kern, J. S., and J. J. Lee. In press. Field
Operations and Quality Assurance Report
for Soil Sampling in the Mid-Appalachian
Region of the United States. EPA/600/x-
xx/xxx. U.S. Environmental Protection
Agency, Corvallis, Oregon.
Lee, J. J., D. A. Lammers, M. G. Johnson, M. R.
Church, D. L Stevens, D. S. Coffey, R. S.
Turner, L. J. Blume, L. H. Liegel, and G. R.
Holdren. 1989. Watershed Surveys to
Support an Assessment of the Regional
Effect of Acidic Deposition on Surface
Water Chemistry. Environ. Mgt. 13:95-
-108.
MacDonald, C. C. 1977. Methods of Soil and
Tissue Analysis in the Analytical Labora-
tory. Mar/times Forest Research Centre
Information Report, M-X-78. Fredericton,
New Brunswick, Canada.
Papp, M. L, R. D. Van Remortel, C. J. Palmer,
G. E. Byers, B. A. Schumacher, R. L.
Slagle, J. E. Teberg, and M. J. Miah.
1989. Direct/Delayed Response Project:
Quality Assurance Plan for Preparation
and Analysis of Soils from the Mid-Ap-
palachian Region of the United States.
EPA/600/4-89/031. U.S. Environmental
Protection Agency, Las Vegas, Nevada.
225 pp.
Soil Conservation Service. 1984. Soil Survey
Laboratory Methods and Procedures for
Collecting Soil Samples. Soil Survey
Investigations Report No. 1. U.S. Gov-
ernment Printing Office, Washington, D.C.
39
-------
References (continued)
Stanley, T. W., and S. S. Verner. 1985. The
U.S. Environmental Protection Agency's
Quality Assurance Program, in: Quality
Assurance for Environmental Measure-
ments. ASTM STP 867. American Soci-
ety for Testing and Materials, Philadel-
phia, Pennsylvania, pp. 12-19.
Steel, R. G. D., and J. H. Torrie. 1960. Prin-
ciples and Procedures of Statistics.
McGraw-Hill Book Company, New York.
481 pp.
Taylor, J. K. 1987. Quality Assurance of Che-
mical Measurements. Lewis Publishers,
Chelsea, Michigan. 328 pp.
U.S. Environmental Protection Agency. 1986.
Development of Data Quality Objectives:
Description of Stages 1 and II. Quality
Assurance Management Staff. U.S.
Environmental Protection Agency, Wash-
ington, D.C. 14 pp.
U.S. Environmental Protection Agency. 1989.
An Overview of the Direct/Delayed
Response Project. In: AERP Status, April
1989. EPA/600/M-89/005. U.S. Environ-
mental Protection Agency, Washington,
D.C.
Van Remortel, R. D., G. E. Byers, J. E. Teberg,
M. J. Miah, C. J. Palmer, M. L Papp, M.
H. Bartling, A. D. Tansey, D. L Cassell,
and P. W. Shaffer. 1988. Direct/Delayed
Response Project: Quality Assurance
Report for Physical and Chemical Ana-
lyses of Soils from the Southern Blue
Ridge Province of the United States.
EPA/600/8-88/100. U.S. Environmental
Protection Agency, Las Vegas, Nevada.
251 pp.
40
-------
Appendix A
Preparation Laboratory Standard Operating Procedures
for the Mid-Appalachian Soil Survey
The following protocols were used by preparation laboratory personnel during the Direct/
Delayed Response Project Mid-Appalachian Soil Survey.
41
-------
DIRECT/DELAYED RESPONSE PROJECT:
PREPARATION LABORATORY STANDARD OPERATING
PROCEDURES FOR THE MID-APPALACHIAN SOIL SURVEY
by
M. H. Bartling, M. L Papp, and R. D. Van Remortel
Lockheed Engineering & Sciences Company
Las Vegas, Nevada 89119
Contract Number 68-03-3249
Project Officer
L. J. Blume
Exposure Assessment Research Division
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Las Vegas, Nevada 89193-3478
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89193-3478
42
-------
NOTICE
This document is a preliminary draft. It has not been formally released by the U. S.
Environmental Protection Agency. It is being circulated for comments on its technical merit and
policy implications.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
43
-------
ABSTRACT
The Direct/Delayed Response Project soil surveys include the mapping, characterization,
sampling, preparation, and analysis of soils as part of a national program to assess watershed
response to acidic deposition within certain regions of the United States. Soil samples from the
Mid-Appalachian region are being collected from predetermined sampling sites and are shipped to
a preparation laboratory for soil processing and preliminary analysis.
The standard operating procedures contained in this manual will serve as the protocols for
operations taking place at the preparation laboratory. The manual is submitted as a final draft on
September 25, 1988.
44
-------
CONTENTS
Notice 43
Abstract 44
1.0 Introduction 47
1.1 Overview 47
1.2 General Laboratory Responsibilities 49
1.3 General Laboratory Support 50
1.3.1 Facilities 50
1.3.2 Equipment and Consumable Supplies 51
1.3.3 Technical Assistance 52
1.3.4 Personnel 53
2.0 Preparation Laboratory Logistics 55
2.1 Sampling Equipment and Consumable Supplies 55
2.2 Sample Identification 56
2.2.1 Types of Field Samples 56
2.2.2 Quality Evaluation Samples 57
2.2.3 Quality Control Samples 60
2.2.4 Use of the Sample Code 61
2.3 Sample Receipt 61
2.4 Sample Storage 63
2.5 Sample Tracking 64
2.6 Sample Batching 65
2.6.1 Batch Configurations 65
2.6.2 Batching Procedure 66
3.0 Sample Processing and Analysis 69
3.1 Sample Drying 69
3.1.1 General Considerations 69
3.1.2 Equipment 70
3.1.3 Procedure 70
3.2 Field-moist pH Determination 71
3.2.1 General Considerations 71
3.2.2 Equipment 72
3.2.3 Calibration and Standardization 73
3.2.4 Quality Control 74
3.2.5 Procedure 75
3.3 Organic Matter Determination by Loss on Ignition 76
3.3.1 General Considerations 76
3.3.2 Equipment 77
3.3.3 Quality Control 77
3.3.4 Procedure 78
3.3.5 Calculations 79
3.4 Air-Dry Moisture Determination 80
3.4.1 General Considerations 80
3.4.2 Equipment 80
45
-------
CONTENTS (continued)
Page
3.4.3 Quality Control 80
3.4.4 Procedure 81
3.4.5 Calculations 82
3.5 Disaggregation and Sieving 82
3.5.1 General Considerations 82
3.5.2 Equipment 83
3.5.3 Procedure 84
3.6 Rock Fragment Determination 85
3.6.1 General Considerations 85
3.6.2 Equipment 85
3.6.3 Procedure 85
3.6.4 Calculations 86
3.7 Homogenization and Subsampling 86
3.7.1 General Considerations 86
3.7.2 Equipment 87
3.7.3 Procedure 87
3.8 Bulk Density Determination 88
3.8.1 Clod Method 89
3.8.2 Known Volume Methods 95
References 98
46
-------
SECTION 1.0
INTRODUCTION
The Preparation Laboratory Standard Operating Procedures (PLSOP) serve as a guide for
personnel involved in the preparation of soil samples for the Direct/Delayed Response Project
(DDRP) Mid-Appalachian Soil Survey (MASS). All personnel must be trained by a laboratory
manager knowledgeable of the protocols detailed in this manual.
1.1 OVERVIEW
The DDRP is an integral part of the Aquatic Effects Research Program (AERP) of the
U.S. Environmental Protection Agency (EPA). The program is administered under the federally-
mandated National Acid Precipitation Assessment Program (NAPAP) to address concerns
relating to potential acidification of surface waters by atmospheric deposition within the United
States.
The overall purpose of DDRP is to characterize geographic regions of the United States
by predicting the long-term response of watersheds and surface waters to acidic deposition.
The DDRP has been designed under the concept of regionalized integrative surveys. According
to this concept, research programs initially are approached from a large region of study leading
to the selection of regionally characteristic systems. These systems can be assessed through
detailed, process-oriented research which will aid in the understanding of the underlying
mechanisms responsible for observed effects. The projected responses of watershed systems
typical of the regional population can then be extrapolated with confidence to a regional or
national scale.
47
-------
The EPA is assessing the role that atmospheric deposition of sulfur plays in controlling
long- term acidification of surface waters (EPA, 1985). Recent trend analyses have indicated
that the rate of sulfur deposition is slowly declining in the northeastern United States and is
increasing in the southeastern United States. If a "direct" response exists between sulfur
deposition and surface water alkalinity, then the extent of current effects on surface water
probably would not change much at current levels of deposition, and conditions would improve
as the levels of deposition decline. If surface water chemistry changes in a "delayed" manner,
e.g., due to chemical changes in the watershed such as increased cation activity due to mineral
weathering, then future changes in surface water chemistry (even with level or declining rates of
deposition) become difficult to predict. This range of potential effects has clear and significant
implications to public policy decisions on possible additional sulfur emissions control (EPA,
1985b).
Specific goals of DDRP are to (1) define physical, chemical, and mineralogical
characteristics of the soils and to define other watershed characteristics across these regions,
(2) assess the variability of these characteristics, (3) determine which of these characteristics
are most strongly related to surface-water chemistry, (4) estimate the relative importance of key
watershed processes in controlling surface-water chemistry across the regions of concern, and
(5) classify the watersheds with regard to their response to sulfur deposition and extrapolate
the results from the sample of watersheds to the regions of concern.
The MASS includes areas within the states of Pennsylvania, Virginia, and West Virginia.
The streams in this region were sampled as part of the National Surface Water Survey, which is
a NAPAP program designed and implemented by the EPA to conduct a chemical survey of lakes
and streams located in regions of the United States believed to be susceptible to the effects of
acidic deposition. A sampling design was applied to allow for unbiased characterization of
regional populations, and resulted in the selection of 37 watersheds.
48
-------
Soil types that were identified in the Mid-Appalachian sampling region are combined into
groups, or sampling classes, which are either known to have or are believed to have similar
chemical and physical characteristics. Each of the sampling classes is then sampled across a
number of different watersheds in which the classes occur. Using this approach, a given soil
sample does not just represent the specific watershed from which it came. Rather, each
sample contributes to a set of samples that collectively represent a specific sampling class for
all DORP watersheds within the sampling region.
The preparation laboratory is an integral component of the MASS. Personnel at the
laboratory are responsible for tracking soil samples throughout processing, performing some
preliminary analyses, formulating and shipping batches of samples to the analytical laboratories,
and documenting all activities.
1.2 GENERAL LABORATORY RESPONSIBILITIES
The preparation laboratory is designed to be the link between the sampling crews and
the analytical laboratories. The primary functions of the preparation laboratory are to prepare
homogeneous, anonymous subsamples from processed bulk samples and to transfer batches
of those subsamples to the analytical laboratories. For these tasks to be successfully
accomplished, the preparation laboratory must uniformly track, process, and store all samples.
The preparation laboratory manager assumes the responsibility for maintaining the
integrity of all samples upon their arrival at the laboratory facility. Samples are packaged and
shipped by the sampling crews utilizing an overnight express service.
49
-------
The preparation laboratory staff also performs certain soil analyses. These include the
determinations of percent air-dry moisture, percent rock fragments in the 2- to 20-mm fraction,
percent organic matter content by loss-on-ignition, field-moist pH in water, and bulk density.
After the samples have been processed and analyzed, they are assembled in batches
and shipped to designated analytical laboratories for further analyses. An important function of
batch formation is to place certain quality evaluation (QE) and quality control (QC) samples in
the batches without revealing their identity to the analytical laboratories.
A separate function of the preparation laboratory is to distribute equipment and supplies
to the sampling crews. Through adequate tracking of supplies, the preparation laboratory can
avoid inventory shortages by notifying the EMSL-LV QA manager of any equipment needs.
The preparation laboratory is required to keep thorough documentation of sample
tracking and processing. This information must always be available to the EPA project officer
or designee throughout the project. Documentation is submitted to the QA manager upon
conclusion of the survey.
1.3 GENERAL LABORATORY SUPPORT
1.3.1 Facilities
The MASS soil preparation and preliminary analyses are performed in a secure, climate-
controlled warehouse located in Las Vegas, Nevada. The warehouse contains a number of
mobile laboratory units that are used for specific preparation activities. Fume hoods and
exhaust fans must be operable for certain stages of sample processing and for the preparation
of the clod dipping solution. A source of compressed air must be available for routine cleaning
50
-------
of equipment. Deionized water is required in the preparation laboratory for various analyses
and cleaning. Cold storage facilities must be readily accessible with ample space to store at
least 600 bulk samples at a time.
1.3.2 Soil Processing Equipment and Consumable Supplies
The following hardware and supplies are required for the prescribed soil processing and
analysis procedures:
Log books
Raw data forms
Munsell color book
Pens, indelible black ink
Kraft paper, 36" wide rolls
Plastic gloves (unpowdered)
Dust masks
One-gallon paint cans
Aluminum weighing tins
Nalgene bottles: 2-L, 500-mL, 125-mL
Carboys, 13-gallon capacity
Crucibles, tolerance to 450°C
Evaporating dishes, tolerance to 450°C
Beakers, 1000-mL
Tongs
Ring stand
Thermometers, 0 to 100°C range
Sample drying tables, PVC / nylon mesh
Rolling pins, wooden
Rubber stoppers, No. 10 size
Brass sieves, square-holed: 2-mm and 4.75-mm
Riffle splitter, Jones-type, closed-bin, with 1.25-cm baffles
Top-loading balance, accurate to 0.01g, capacity to SOOg
Top-loading balance, accurate to 0.1g, capacity to 5Kg
pH meter, with proper electrodes
Desiccators
Desiccant
Hot plate
Convection oven
Muffle furnace
Shipping boxes
Packing material
Strapping tape
Saran® powder
Acetone
Plastic containers: 25mL, 50mL
51
-------
1.3.3 Technical Assistance
Seven raw data forms have been designed by EMSL-LV for sample tracking and for
entering data relating to: 1) sample receipt, 2) bulk sample processing, 3) field-moist pH, 4)
air-dry moisture, 5) loss-on-ignition, 6) clod bulk density, and 7) known volume bulk density.
All raw data are recorded on these forms in indelible black ink and the completed forms are to
be organized in binders. An example of each form is provided in Appendix A
A personal computer is available at the laboratory for entry and verification of all
preparation data recorded on the raw data forms. This provides an efficient way to track the
status of the samples from the time of field sampling through shipment to the analytical
laboratory. Through data comparisons, sample labeling discrepancies can be quickly identified
and reported to the Soil Sampling Task Leader at ERL-C and the QA Manager at EMSL-LV.
Final data values are calculated through a SAS-AF® computer program, allowing data to be
reviewed in real time. Precision and accuracy of the QE/QC samples are checked as the
samples are processed. This permits identification of any discrepancies and allows the QA
manager and laboratory manager to efficiently track progress of the laboratory staff.
The computer entry and verification system is capable of producing all of the final data
forms required by EPA. Selected laboratory personnel are trained to use the computer system.
Appendix B describes the preparation laboratory data entry and verification system.
Additional support is provided through communications with other DDRP personnel. An
EMSL-LV QA representative periodically visits the preparation laboratory to aid in training
personnel and answering questions that arise. The laboratory manager, however, should never
hesitate to call the QA representative or other QA personnel if a problem should arise. The soil
sampling task leader at ERL-C and the QA manager are to be notified whenever there is a
question concerning the field sampling operation.
52
-------
All information pertaining to telephone conversations and written correspondence are to
be thoroughly documented in a log book.
1.3.4 Personnel
The manager of the preparation laboratory is required to be knowledgeable in laboratory
methods and procedures, and have demonstrated ability to track large numbers of samples and
supervise laboratory personnel. Adequate staffing should be provided to ensure a fast and
efficient turnaround of samples from the field to the analytical laboratories. All personnel must
be thoroughly trained in the protocols and procedures by the laboratory manager before the
processing of the samples begins.
Ultimately, the laboratory manager is responsible for assigning duties according to the
specific project needs. The following division of responsibilities is tentative and may be
adjusted.
Laboratory Analyst A:
o Enters data into computer entry and verification system
o Runs verification programs
o Receives and logs samples at cold storage
o Organizes cold storage space
o Tracks all samples during processing
o Assists other analysts after other duties are complete
Laboratory Analysts B, C, D, and E:
o Spreads samples to air dry
o Performs field-moist pH determinations
o Performs loss-on-ignition determinations
o Performs air-dry moisture determinations
o Disaggregates and sieves bulk samples
o Performs rock fragment determinations
o Homogenizes bulk samples with riffle splitter
o Creates appropriate analytical and archive subsamples
o Performs bulk density determinations
o Completes necessary information on bulk sample processing sheet
o Assists other analysts after other duties are complete
53
-------
Laboratory Analysts F and G [Part-time]:
o Disaggregates and sieves bulk samples
o Performs rock fragment determinations
o Tracks all DORP equipment (both field and laboratory)
o Coordinates equipment procurement, preparation, and distribution
o Assists other analysts wherever needed
Laboratory Manager:
o Coordinates laboratory operations and time management
o Communicates with EMSL-LV QA manager and QA representative
o Communicates with ERL-C soil sampling task leader
o Oversees sample receipt and storage
o Oversees all computer data entry and evaluation procedures
o Oversees sample preparation and analysis activities
o Organizes analytical samples into batches
o Tracks all samples during processing
54
-------
SECTION 2.0
PREPARATION LABORATORY LOGISTICS
2.1 SAMPLING EQUIPMENT AND CONSUMABLE SUPPLIES
A designated area within the laboratory area is used to store field sampling equipment.
The laboratory manager is provided with a master inventory list that is updated each time
equipment is distributed to the field.
An initial package consisting of all sampling equipment and a portion of the necessary
consumable supplies is prepared and distributed to each sampling crew at the beginning of the
survey. Additional supplies that are anticipated to be necessary for the completion of sampling
are periodically sent to the sampling crew leaders at the appropriate Soil Conservation Service
(SCS) field offices. At a minimum, each initial field equipment package includes the following:
Equipment/Consumables Quantity
Hand pump 1
35-mm camera w/built-in flash 1
Clod boxes 30
Brass sieve, square-holed, 19-mm 1
Stapler, large 1
Stapler, small 1
Gel-pacs 80
Coolers 10
Hole punch 1
Straight-edged trowel 1
Volume-fill 250-mL cylinder 1
Equipment for bead method 1
Field pH kit 1
Random number table 2
SCS-SOI-232 field data form, dated 3/87 50
Orange flagging 10 rolls
Yellow marker flags 80
Indelible marker - thick 10
Indelible marker - thin 10
Golf tees (horizon delineation) 20
(continued)
55
-------
Equipment/Consumables Quantity
Photogray cards - 8" x 10" 20
Horizon depth tape 1
One-gallon metal can w/lid 2
Hairnets 1 gross
Plastic inner sampling bags 350
Canvas soil bags 350
DDRP Label A 200
Staples 2 boxes
Small plastic clod bags 600
Twist ties 300
Clod tags 300
Clod box labels 500
Log books 5
Saran® powder 10 packets
Acetone 5 gallons
Supply sheets are provided for the sampling crews. Crews are requested to log all
consumables taken from or returned to the SCS offices, which allows better estimates to be
made of any additional consumable items needed. Log entries should be written in black ink
and identify the crew, the equipment taken, and the date.
Equipment needs can be made known to the laboratory manager during weekly field
sampling conference calls. Any items nearing depletion are promptly shipped to the appropriate
SCS office or another convenient delivery point. Each crew leader is responsible for the return
of sampling hardware to EMSL-LV after MASS sampling is completed.
2.2 SAMPLE IDENTIFICATION
2.2.1 Types of Field Samples
Four types of soil samples are sent to the preparation laboratory during the field
sampling phase of the MASS. These are: 1) bulk samples (routine and duplicate), 2) field audit
samples, 3) clod bulk density samples, and 4) known volume bulk density samples.
56
-------
Bulk samples are those soil samples taken from a designated portion of a specific
horizon. A bulk sample contains approximately one gallon of soil and generally weighs about
5.5 Kg. The bulk samples are deposited in large plastic bags and placed within protective outer
canvas bags. Each bulk sample should arrive at the laboratory with a completed DDRP Label A
affixed to the inner plastic bag.
Field audit samples are QE samples (described below) sent by EMSL-LV QA staff to the
sampling crews for inclusion in the sampling of every third pedon by each crew. The samples
are sieved, packaged, and shipped in the same manner as their associated routine samples.
Each field audit sample should arrive at the laboratory with a completed DDRP Label A affixed
to the inner plastic bag. Both mineral and organic horizons are sampled; for the MASS, organic
soils are defined as having 20 percent or more organic matter as determined by loss on ignition.
Clod samples are fist-sized, structurally intact bulk density samples taken from
designated horizons. Each clod is wrapped in a hairnet and dipped briefly in a 1:5
saranracetone solution to help maintain the clod structure and reduce moisture loss from the
clod during transport and storage. Each clod is covered by a small plastic bag and placed in a
special clod box with dividers.
Known volume bulk density samples are taken from horizons where clods are
unobtainable. There are two types of known volume samples being used in the MASS: volume
replacement (VR) and volume filling (VF) samples. These bulk density samples are packaged in
small pre-labeled plastic bags in much the same manner as the bulk soil samples.
2.2.2 Quality Evaluation Samples
The QE samples are those samples which are known to the preparation laboratory staff
but are blind to the analytical laboratory, i.e., the laboratory does not know the predetermined
57
-------
concentration of the QE samples and cannot distinguish the QE samples from routine samples.
The QE samples provide an independent check on the analytical process and can be used to
evaluate whether the data quality objectives (DQOs) set forth in the DDRP-MASS QA Plan (Papp
et al., 1988) have been satisfied for any given batch or for all batches, i.e., overall measurement
uncertainty. Every QE sample has a specific purpose in the data assessment scheme. The
samples are similar to the routine samples both in matrix and in analyte concentration.
The two types of QE samples are variability samples and accuracy samples. The QE
variability samples are replicates of selected routine samples that are used to measure
precision of the analyses. The QE accuracy samples are audit samples that have been
previously characterized by a number of analytical laboratories, sufficient to generate accuracy
windows for data assessment.
2.2.2.1 Field Duplicate (FD) Sample --
The FD is collected at random by sampling crews using alternate trowelsful of soil to
produce a pair of samples (one routine, one duplicate) from one specific horizon within a
specific pedon. Individual pairs are used to assess system-wide within-batch precision. These
estimates are pooled to provide the within-batch component of overall system measurement
uncertainty.
2.2.2.2 Field Audit (FAP/FAO) Samples -
The FAP and FAO samples are median-range audit samples sent in pairs or in triplicate
to the sampling crews, who treat the samples as if they had just been obtained from the
excavated pit. The FAP samples are used when sampling every third mineral soil pedon and the
FAO samples are used whenever an organic soil pedon is sampled. The field audit samples are
sieved and bagged at the time of sampling. These samples are then handled as if they were
58
-------
routine samples and undergo all of the subsequent soil preparation and analysis steps. The
FAP and FAO samples are used to identify system-wide bias within a batch. The analytical
results are pooled from all batches to provide an estimate of all sources of measurement
uncertainty other than within-batch uncertainty.
2.2.2.3 Low-Range Field Audit (FAL) Sample -
The FAL sample is a low-range audit sample sent to the sampling crews and is handled
in the same manner as the FAP/FAO samples, but is included only in mineral soil batches. The
FAL is used to identifv any system-wide contamination sources and as an additional check on
system bias. Pooled data from all FAL samples provide an estimate of the system detection
limit.
2.2.2.4 Preparation Duplicate (PD) Samples -
The POs are split samples of a field duplicate and its associated routine sample that are
created by the preparation laboratory staff. The split samples are labeled PDF and PDR,
respectively. The PD samples are used to identify the within-batch preparation component of
the overall measurement uncertainty and as independent comparative checks on the within-
batch precision of the analytical laboratories.
2.2.2.5 Laboratory Audit (LAP/LAO) Samples --
The LAP and LAO samples are median-range audit samples identical to the FAP and
FAO samples except that they are sent in pairs or triplicate directly to the preparation laboratory
for inclusion in the analytical batches. The samples are used to assess analytical laboratory
bias within a batch and are used in combination with the FAP and FAO samples to identify
sources of bias, e.g., contamination or analysis. Within each batch, the FAP/FAO and LAP/LAO
59
-------
samples represent the same audit horizons (B, Bw, or 0 horizon). Data from the samples are
used to calculate interlaboratory differences and to assess between-laboratory and between-
batch components of measurement uncertainty.
2.2.2.6 Low-Range Laboratory Audit (LAL) Sample ~
The LAL is a low-range audit sample identical to the FAL sample except that it is sent
directly to the preparation laboratory for inclusion in mineral soil analytical batches. When
applied in conjunction with the FAL sample, the LAL is used to calculate instrument detection
limits (IDLs) as well as to identify the magnitude and, possibly, the source of sample
contamination. The IDL is estimated as three times the standard deviation of the LAL samples
for a given analytical procedure. Calculation of LAL data begins with the second batch and the
results are used to independently check the laboratory detection limits to ensure that the DQOs
for detectability are being met.
2.2.3 Quality Control Samples
The QC samples are non-blind samples established by contract to assist the
laboratories in meeting intralaboratory DQOs. The quality control audit (QCA) sample is the only
such QC sample batched at the preparation laboratory.
The QCA is an audit sample provided to the analytical laboratories as the 15th sample in
every batch. The laboratories are provided with the corresponding range of reference values
obtained by previous DDRP multi-laboratory analysis. The QCA sample is primarily used to
control bias and to reduce bet ween-laboratory and between-batch components of measurement
uncertainty. The QCA sample from each batch is tracked on a control chart to ensure that the
data are within acceptable limits set forth in the analytical laboratory contract.
60
-------
2.2.4 Use of the Sample Code
Each soil sample must be identified by a sample code, which is a 13-character
designation uniquely identifying that sample. The sample code also identifies those multi-bag
samples which are to be combined during soil processing. The first three characters of the
sample code represent the type of sample, e.g.:
R11 = routine sample, single bag (R12/R22 or R13/R23/R33 for multiple bags, if
necessary)
FDD = field duplicate sample (FD1/FD2 for multiple bags, if necessary)
FAL = field audit sample, low-range (mineral soils)
FAP = field audit samples, pair (mineral soils)
FAO = field audit samples, triplicate (organic soils)
The next two letters designate the state: PA for Pennsylvania, WV for West Virginia, or VA for
Virginia. Following this is a three-digit county designation. A subsequent dash (or zero)
separates the county code from the pedon/horizon designation, which is the final four-digit
code.
NOTE: For the field audit samples, the last two digits of the sample code reflects the audit
sample type. For the FAP samples, the B horizon samples contain the closing digits 01 and 02,
and the Bw horizon samples contain the closing digits 03 and 04. The FAO samples contain the
closing digits 01, 02, and 03. The FAL samples always contain the closing digits 01.
2.3 SAMPLE RECEIPT
The sampling crews should make every effort to inform the preparation laboratory when
samples are being shipped from the field. Bulk soil samples, bulk density samples, and field
61
-------
audit samples are to be placed in box-enclosed styrofoam coolers with two frozen gel-pacs.
The corresponding SCS-SOI-232 field data forms and sample shipment forms are sealed in a
plastic bag and taped to the underside of the cooler cover. The outer box of the cooler should
be firmly enclosed with strapping tape. The samples are then shipped directly to the
preparation laboratory via overnight carrier.
Because of their fragile nature, the soil clods (one type of bulk density sample) are
securely packed with vermiculite in coolers separate from the bulk soil samples. Clods are to
be stored in a well ventilated area while awaiting shipment.
When samples are shipped on a Friday, the sampling crew must notify the preparation
laboratory manager to ensure that personnel are available on Saturday to receive the samples.
All preparation laboratory personnel must become familiar with the DDRP Label A (see
Figure 1), as it is a primary tool used in sample receipt and tracking. There are several entries
on the sample label, including the date sampled, horizon designation, depth range of the horizon
sampled, crew ID, site ID, sample code, and set ID. The crew ID consists of four alpha-
numeric characters representing the state and the crew number assigned to each sampling
crew within the state. The site ID is a code assigned by ERL-C which designates a specific
watershed that is sampled. The set ID is a five-digit code relating to each day a sampling crew
samples a pedon. A range of set IDs are assigned to each crew. The laboratory personnel
must also become familiar with the DDRP Label B (see Figure 2) used to label and track
analytical samples within the batches.
The field data form is completed by the sampling crew at the sampling site. The form
lists all field data for a particular pedon, including descriptions of landform, vegetation, soil
climate, and a detailed pedon characterization. The field data form also contains codes, e.g.,
watershed ID, assigned by the soil sampling task leader.
62
-------
DDRP Label A
Date Sampled:
D D M M M Y Y
Crew ID:
Site ID:
Sample Code:
Horizon:
Set ID:
Depth: — _ _ _ ex
Figure 1. DDRP Label A
DDRP Label B
Batch ID:
Sample No: _ _
Figure 2. DDRP Label B
The samples received at the laboratory are checked for coding accuracy against the field
data form and sample shipment form. A sample receipt raw data form (see Appendix A) must
be filled out completely for all samples received and logged-in at cold storage. This information
is then entered and verified in the computer entry and verification system. Samples should not
undergo processing until they are verified under this procedure.
2.4 SAMPLE STORAGE
All samples are to be placed in cold storage as soon as possible after receipt. Known
volume bulk density samples, once checked through the sample receipt procedure, can be
placed in the drying area and allowed to air dry. Bulk soil samples and field audit samples
remain in cold storage until there is sufficient space in the drying area for the samples (grouped
according to set ID) to be air dried. After air drying, the samples are returned to cold storage
and must remain there when not being processed.
The cold storage room must be maintained at a temperature of 4°C. It is necessary for
the temperature to be monitored either by a continuous sensor or by daily thermometer
63
-------
readings. Any substantial deviations in temperature (± 2°C or more) should be recorded in a
log book; affected samples should be identified according to set ID and sample code.
Samples should be organized in cold storage so that a particular sample within a given
set may be easily located. Samples arrive at the laboratory in sets which must be maintained
during storage, processing, and shipping. Because organic and mineral analytical samples are
to be batched and shipped separately, it is advisable to keep organic samples separated from
mineral samples in the cooler.
2.5 SAMPLE TRACKING
An adequate tracking system must be initiated to document the transfer of all samples
from the field to the preparation laboratory and, finally, to the analytical laboratory. The
following protocols have been developed to ensure proper sample tracking.
Before sampling begins, the ERL-C soil sampling task leader provides the EMSL-LV QA
manager and preparation laboratory manager with a list of watershed identification numbers
(site IDs) and pedon numbers for each pedon to be sampled. This information is entered into
the computer entry and verification system at the preparation laboratory before samples are
received.
As the samples arrive at the preparation laboratory, the sample receipt raw data form is
filled out and entered into the computer. The information is checked against the site IDs and
pedon numbers that have been entered into the program. If a particular watershed/pedon
combination has been previously used, e.g., the preparation laboratory has previously received
samples with that watershed/pedon number or a particular watershed/pedon number cannot be
found on the ERL-C list, the soil sampling task leader and the QA manager are to be notified
64
-------
immediately. After the site IDs, pedon numbers, and other data on the sample receipt raw data
form are verified, soil processing and analysis can be initiated.
2.6 SAMPLE BATCHING
After the sample preparation and preliminary analyses are completed for a number of
sets of samples, batches are constructed and readied for shipment to the analytical
laboratories. To preserve the integrity of each sample during these steps, careful guidelines
must be followed.
2.6.1 Batch Configurations
A batch of samples assembled for chemical and physical analysis consists of a
maximum of 40 samples, including approximately 30 routine samples and 10 different QE/QC
samples which are used to assess or control various types of measurement uncertainty.
Typical mineral and organic batches consist of the following sample configurations:
SAMPLES PLACED IN MINERAL AND ORGANIC SOIL BATCHES
Samples per batch
Sample type Sample description Mineral Organic
FD Field duplicate sample 1 1
FAL Field audit sample, low-range 1 0
FAP Field audit samples, pair 2 0
FAO Field audit samples, triplicate 0 3
PD Preparation duplicate samples 2 2
LAL Laboratory audit sample, low-range 1 0
LAP Laboratory audit samples, pair 2 0
LAO Laboratory audit samples, triplicate 0 3
QCA Quality control audit sample 1 1
RS Routine samples 30 30
40 total 40 total
65
-------
Several sets are required to make up a batch. A set should never be split across two
batches. If it is not possible to get 30 routine samples without splitting a set, as many
complete sets as possible are batched. Within each batch, one set should contain a field
duplicate and associated field audit samples.
NOTE: Organic samples are to be separated into unique batches that contain no mineral
samples, and vice versa. This is the only case where samples are separated from their
respective set. All other batching and shipping procedures remain the same. An organic field
duplicate and associated field audit samples must also be included in the organic batches.
2.6.2 Batching Procedure
Batch numbers and a preparation laboratory ID are assigned by the QA manager for use
by the laboratory manager in batching samples. All batches must contain the types and
quantities of QE/QC samples described in Section 2.6.1.
Within each batch, preparation duplicates are split from the field duplicate and its
associated routine sample. The sample code for a preparation duplicate begins with the two
characters "PD" which identifies the sample as a preparation duplicate. The letter "F" in the
third space is used to identify the split from the field duplicate (PDF) and the letter "R" is used
to identify the split from the routine sample (PDR). The sample code continues with the final
ten alpha- numeric characters from the original sample, e.g., the split from routine sample
R11PA013-0405 would be designated as PDRPA013-0405.
66
-------
The sample code for a laboratory audit sample is comparable to that of its
corresponding field audit sample, with a letter "L" substituting for the letter "F" as the first
character of the code. The sample code for a QCA sample should read "QCAS", followed by a
dash and the 5-digit batch ID. The LAP, LAO, LAL, and QCA samples are provided to the
preparation laboratory directly from the EMSL-LV QA staff. The samples are stored in amber
Nalgene bottles identical to those used for the routine samples. The preparation laboratory
manager is responsible for labeling each of these samples with a DORP Label B and randomly
placing it within the proper batch. The only exception to the random placement is the QCA
sample, which is always designated the 15th sample in every batch.
NOTE: The audit samples are not to be opened or disturbed at the preparation laboratory for
any reason. If the laboratory audit samples are packaged differently from the routine samples
upon their delivery to the preparation laboratory, e.g., with a different type bottle or cap, or with
unusual markings on the bottle, call the QA manager immediately for further instructions. The
QE samples must be blind to the analytical laboratory.
With the exception of the QCA sample as the 15th sample, randomize all samples in the
batch and assign sample numbers from 1 to however many samples are in the batch, e.g., 40
samples in a batch would be numbered 1 through 40. Replace the DORP Label A for each
analytical sample bottle with the appropriate DDRP Label B. Staple the DDRP Label A on the
corresponding bulk sample processing raw data form and initial the label so that the writing
overlaps both the label and the form.
Enter the analytical sample numbers into the computer batching program. The program
then generates the DDRP Form 101 (see Appendix A), including all column headings and final
data for the processed soil samples.
67
-------
Place the 500-mL and 125-mL analytical sample bottles in separate shipping boxes and
prepare a Form 102 (see Appendix A) for each box. Upon approval of the QA manager, ship
each box of samples to the specified analytical laboratory using an overnight express service.
The QA manager should be provided with a clean copy of the Form 101 to serve as
documentation of the shipment. Arrangements can then be made to move the 2-L archive
sample bottles into long-term cold storage.
68
-------
SECTION 3.0
SAMPLE PROCESSING AND ANALYSIS
Specific areas of the preparation laboratory are designated for sample processing and
analysis. During all of the steps of sample handling, maintaining the integrity each sample is of
the greatest importance. Sample integrity is ensured by the use of detailed sample labels, by
documenting the status of each sample during the processing and analyses, and by avoiding
physical or chemical contamination during each processing step.
The sample preparation protocols are described in detail throughout this section of the
manual.
3.1 SAMPLE DRYING
3.1.1 General Considerations
Sample tables constructed of PVC and heavy nylon mesh are used to air dry the
samples. Use of the mesh enhances air circulation and increases the rate of sample drying.
Chemicals used in analyses are to be kept away from the drying area. Food, drinks,
and smoking are prohibited in the drying area. A separate pair of gloves should be worn when
handling each sample. The drying area should be damp-mopped at least once a week to
control dust.
69
-------
3.1.2 Equipment
Drying tables, PVC / nylon mesh
Kraft paper, 36" wide rolls
Plastic gloves (unpowdered)
Bulk sample processing raw data forms
3.1.3 Procedure
Label a bulk sample processing raw data form (see Appendix A) for each sample to be
air dried. Place two fresh sheets of kraft paper, approximately 1m x 1m in area, on the mesh
partition of the drying table. With gloved hands, slowly spread the sample on top of the sheets
of paper, taking care not to lose any soil off the paper or contaminating any adjacent samples.
While spreading the sample to dry, pay particular attention to the sample code on the
DDRP Label A. Where both the second and the third character of a routine sample code is the
digit "1", e.g., as in R11, do not combine the sample with any other sample. Where either the
second or the third character of a routine sample code is a number other than "1", e.g., as in
R12 or R22, combine that sample with those routine samples having the same state, county,
pedon, and horizon code. If a field duplicate sample has any designation other than "FDO", e.g.,
as in FD1 or FD2, combine it with its companion field duplicates having the same state, county,
pedon, and horizon number.
The combined samples are labeled with the sample code designating the final bag of the
sample and the total number of bags of the sample. For instance, if one had three samples
labeled R13 R23 and R33 all three samples would be combined on the drying table and
labeled R33 with the remainder of the sample code. Do not combine any samples that contain
a zero or an alphabetic character in the third space of the sample code.
70
-------
Disaggregate any large peds that impede the spreading of the sample over the entire
area of the paper. Place an additional sheet of kraft paper loosely over the sample. Attach the
sample's processing raw data form on the hook attached to the drying table. The original
canvas and plastic sampling bags should be kept on the floor beneath the sample as a second
check on sample identity.
NOTE: Immediately after a sample is spread to dry, field-moist pH is determined as outlined in
Section 3.2.
Daily, the soil sample is to be stirred with gloved hands to facilitate drying. For the first
few days that a wet sample is spread, the bottom sheet of paper may need to be changed
daily in order to alleviate excessive moisture accumulation. Any observations of fungal or algal
growth should be noted on the bulk sample processing raw data form.
NOTE: Soils high in clay may harden nearly irreversibly if allowed to dry without a preliminary
disaggregation of medium and coarse peds. An effort should be made during the air-drying
procedure to disaggregate these peds while they are still somewhat moist or friable, before they
reach an air-dry state.
Allow the sample to air dry until it is believed to be at or below the specified moisture
content. This process generally takes about two days, although drying time may vary from one
day to a week or more. At this time, an aliquot can be subsampled for the air-dry moisture
determination outlined in Section 3.4.
3.2 FIELD-MOIST pH DETERMINATION
3.2.1 General Considerations
71
-------
Immediately after a sample is spread to air dry, a field-moist subsample is collected and
pH in deionized water is determined.
Soils high in salts, especially sodium, may interfere with the pH reading and the
electrode response time.
Clay particles may clog the KCI junction of the pH junction, slowing the electrode
response time. Thoroughly rinse the electrode with deionized water between sample readings
to avoid this problem.
Do not wipe the electrode dry with laboratory tissue or similar materials, and do not
remove the electrode from solution when the meter is not on standby, as either practice may
lead to polarization of the electrode.
The pH can vary as much as one pH unit between the supernatant and soil sediment.
Always place the electrode junction at the same distance above the surface of the soil sediment
to maintain uniformity in pH readings.
3.2.2 Equipment
Digital pH/mV meter, capable of measuring pH to 0.01 pH units, voltage potential to 1
mV, and temperature to 0.5°C
National Bureau of Standards (NBS)-traceable pH buffer sets of pH = 4.0 and pH = 7.0
for calibration (two sets from different sources)
NBS thermometer
Combination pH and reference electrodes, high quality, low-sodium glass [Geltype
reference electrodes must not be used; an Orion Ross combination pH electrode
or equivalent with a retractable sleeve is recommended]
72
-------
125-mL Nalgene bottles
25-mL and 50-mL plastic containers
Carboy, 13-gallon, filled with deionized water
Field-moist pH raw data form
Field-moist pH log book
3.2.3 Calibration and Standardization
For storage and readings, the electrode need only be immersed to cover the liquid
junction of the reference electrode (typically about 2.5 cm). Rinse the electrode with deionized
water between each sample and each buffer to prevent solution carryover. Do not rub or blot
electrode dry because this may produce a static electric charge and thereby polarize the
electrode.
To prepare the pH electrode for use, move the sleeve covering the fill hole and fill the
reference reservoir to level of the hole with 3M KCI solution. Allow 5 minutes for the ceramic frit
to become wet with filling solution before immersing the electrode in the sample or the buffer
solution. The retractable sleeve junction allows easy cleaning of clay particles and insolubles
that clog the junction.
Using the NBS pH 7.0 and pH 4.0 buffer solutions, calibrate the electrode as specified in
the electrode manual for a two-buffer calibration. Place about 35mL of pH 7.0 buffer solution
into a 50-mL plastic container, stir, wait 30 seconds, and take the reading. Record in the pH log
book the temperature of the solution on the meter and also the reading on the NBS
thermometer. Adjust the temperature control on the meter, read the pH after equilibration, and
adjust the meter if necessary. Perform the steps again using the pH 4.0 buffer and the slope
control on the meter. Repeat measurements and adjustments until readings for both buffer
solutions are within 0.05 pH unit of the respective known values.
73
-------
Calibrate the pH meter before each run of sample analyses. All readings and
adjustments should be recorded in the pH log book and initialed by the analyst.
3.2.4 Quality Control
Before analyzing the samples, measure a quality control check sample (QCCS) of pH 4.0
using a different batch of NBS stock buffer solution than that used for the meter calibration.
Subsequently, after every 10 samples and after the final sample of the run, measure another
QCCS of pH 4.0. The field-moist pH raw data form (see Appendix A) contains an entry field for
these data indicated by "QCCS". The value recorded for each QCCS should be 4.00 ± 0.05. If a
QCCS value is not within that range, recalibrate the electrode according to the procedure
described in Section 3.2.3 above. If acceptable results still cannot be obtained, check: (1) that
the reference junction of the electrode is clean, (2) that the wiring straps into the meter are
firmly connected, (3) for static discharge from another instrument or from the operator, and (4)
to ensure that sufficient filling solution is contained within the electrode. If the problem
persists, replace the electrode, the meter, or both.
Within each run of pH analysis, an internal duplicate sample of a field-moist soil is to be
obtained and analyzed for every 10th sample. The letter "D" on the raw data form indicates
when a duplicate is to be subsampled. The "D" replaces the first character of the sample code
for the sample that was selected. The duplicate sample and its corresponding routine sample
pH values should have a standard deviation for the pair of 0.15 pH units or less. If the
standard deviation is higher than this value, the data should be reviewed by the laboratory
manager and the QA manager. A decision may be made to re-run the pH analysis on these
samples.
74
-------
Also within each run, a manager's sample (MS) is included. The MS is an audit sample
having characterized audit windows, but its concentration range is unknown to the analyst
performing the pH analysis. The sample is used to assess accuracy of the run.
If a run of pH analysis consists of fewer than 10 samples, the QCCS and internal
duplicate samples should be placed as the final samples in the run.
NOTE: Because the pH is measured immediately after a sample is spread to air dry, the
number of samples in each run varies. It is necessary to adhere to all calibration protocols and
QC specifications for all runs, regardless of size.
3.2.5 Procedure
After spreading and stirring the field-moist soil, scoop approximately 50g of mineral soil
or 15g of organic soil into a 125-mL Nalgene bottle and record the sample code on the label of
the bottle. Refrigerate the bottles at 4°C until such time as the analyst is ready to begin a run.
NOTE: Some soils may be in a saturated state when received. Follow the procedure as stated;
however, note the saturated condition of the soil on the bulk sample processing raw data form.
Place approximately 20g of mineral soil or 5g of organic soil into a pre-numbered 25-mL
plastic container. Add 20ml_ of deionized water for mineral soils or 25mL for organic soils.
Record the sample code and the container number on the field-moist pH raw data form.
Allow the sample to absorb the solution for approximately 60 seconds, then thoroughly
stir the soil-solution mixture for 10 seconds with a glass stirring rod. Rinse the rod with
75
-------
deionized water after stirring each sample. Stir again for 10 seconds after 15, 30, 45, and 60
minutes. Allow the suspension to settle for 30 minutes.
Place the pH electrode in the supernatant of the soil suspension. For mineral soils, the
electrode junction is to be below the solution surface and above the soil-solution interface.
Although organic soils normally absorb all of the free water available, an acceptable reading
generally is obtained if the electrode junction is below the meniscus of the organic material.
Record the pH of each sample (PH_MP) on the field-moist pH raw data form. Once the
pH analysis is completed for a run of soils, enter the data into the computer entry and
verification system. If there are no flags generated (see Appendix B), transcribe the pH values
onto the appropriate entry field of each bulk sample processing raw data form. Any sample
with a reading of pH 6.5 or higher is to be reported to the QA manager immediately.
After the pH measurements are completed, store the electrode in storage solution. Do
not allow the sensing element or reference junction to dry out. The level of the storage solution
should be one inch below the filling solution level. The electrode sleeve is to be closed during
storage. Check periodically to ensure that the electrode reservoir is full of storage solution.
3.3 ORGANIC MATTER DETERMINATION BY LOSS-ON-IGNITION
3.3.1 General Considerations
Loss-on-ignition (LOI) is the method used to determine percent organic matter of the
MASS samples. Because organic samples are oxidized at high temperatures, percent organic
matter can be calculated on a weight-loss basis. From the percent organic matter, the percent
76
-------
organic carbon can be estimated. A modified version of the method described in MacDonald
(1977) is used.
Laboratory personnel should use extra caution when working around the muffle furnace,
as temperatures of 450°C are common. The furnace should be activated only in an operable
fume hood.
Crucibles must be thoroughly cleaned after each use, as the crucible weights are pre-
determined and entered into the computer verification system. Tongs or finger cots should be
used when transferring crucibles from one location to another.
3.3.2 Equipment
Open-pan balance, accurate to O.Otg
Small crucibles, pre-numbered
Tongs
Finger cots
Desiccator, with desiccant and crucible plate
Muffle furnace in operating fume hood
Furnace gloves
Convection oven
Loss-on-ignition raw data forms
Loss-on-ignition log book
3.3.3 Quality Control
Within each run of LOI analysis, an internal duplicate sample of a routine soil sample is
to be obtained and analyzed. The letter "D" replaces the first character of the sample code for
77
-------
the sample that was selected. The duplicate sample and its corresponding routine sample LOI
values should have a relative standard deviation for the pair of 15 percent or less. If the
relative standard deviation is higher than this value, the data should be reviewed by the
laboratory manager and the QA manager. A decision may be made to re-run the LOI analysis
on this sample.
Also within each run, a manager's sample (MS) is included. The MS is an audit sample
having characterized audit windows, but its concentration range is unknown to the analyst
performing the LOI analysis. The sample is used to assess accuracy of the run.
If a run of LOI analysis consists of fewer than 10 samples, the internal duplicate and
MS should be placed as the final samples in the run.
3.3.4 Procedure
The LOI analysis can be performed at any time while the samples are air drying. The
data obtained are to be recorded on the loss-on-ignition raw data form (see Appendix A).
Fill a small pre-numbered and pre-weighed crucible about two-thirds full of soil. Record
the sample code and crucible number on the raw data form under "CRUC_NO". Oven dry the
crucible plus sample overnight, at 105°C for mineral samples or 60°C for organic samples.
The following morning, remove the crucible from the convection oven and place it in a
desiccator for 30 minutes. Weigh the crucible plus sample and record its oven-dry weight under
"OD WP.
78
-------
Carefully place the crucible plus sample in a muffle furnace and secure the door. Set
the furnace temperature at 450°C and turn on the unit. The sample should be left overnight at
this temperature.
The following morning, turn the furnace off and allow to cool for 30 minutes. Carefully
remove the crucible from the furnace and allow to cool in the desiccator with a dish of
desiccant.
NOTE: Because a vacuum may be created when cooling extremely hot samples, extra caution
should be exercised when opening the desiccator. If a vacuum has formed, the lid may be
extremely hard to remove and/or a rush of air may enter the desiccator upon opening.
Finally, weigh the crucible plus remaining soil and record this weight under "ASHED_WT"
on the raw data form. Enter the data into the computer entry and verification system. If there
are no flags generated, the sample can undergo further processing.
If the calculated organic matter content is 20 percent or more, the sample is assumed
to be an organic sample. If any samples which were initially labeled as mineral are determined
to have 20 percent or more organic matter, batch these with the organic samples, or vice versa.
Report these samples to the QA manager and the soil sampling task leader as soon as
possible.
3.3.5 Calculations
The following calculation is performed by the computer after the raw data are entered,
but it is presented here for internal use. It is advantageous to manually check a few samples
using the calculation in order to understand the procedure and to test the accuracy of the
computer program.
79
-------
Percent Organic Matter (OMJ.OI) = OD WT - ASHED WT
— — x100
OD WT - CRUC WT
3.4 AIR-DRY MOISTURE DETERMINATION
3.4.1 General Considerations
To ensure that each sample is at an acceptable moisture level for further processing, it
is necessary to perform a moisture analysis for each soil sample. The procedure for assessing
air-dry moisture is not to be performed, however, until the soil is believed to be air dry.
3.4.2 Equipment
Aluminum weighing dishes, pre-numbered
Forceps or finger cots
Open-pan balance, accurate to 0.01g
Convection oven
Desiccator and desiccant
Air-dry moisture raw data forms
Air-dry moisture log book
3.4.3 Quality Control
Within each run of moisture analysis, an internal duplicate sample of a routine soil
sample is to be obtained and analyzed. The letter "D" replaces the first character of the sample
code for the sample that was selected. The duplicate sample and its corresponding routine
sample moisture values should have a relative standard deviation for the pair of 15 percent or
80
-------
less. If the relative standard deviation is higher than this value, the data should be reviewed by
the laboratory manager and the QA manager. A decision may be made to re-run the LOI
analysis on this sample.
Also within each run, a manager's sample (MS) is included. The MS is an audit sample
having characterized audit windows, but its concentration range is unknown to the analyst
performing the moisture analysis. The sample is used to assess accuracy of the run.
If a run of moisture analysis consists of fewer than 10 samples, the internal duplicate
and MS should be placed as the final samples in the run.
3.4.4 Procedure
Data for this procedure are recorded on the air-dry moisture raw data form (see
Appendix A). Once a soil sample has been determined to be air dry, the information is also
entered on the bulk sample processing raw data form.
Thoroughly mix the air-dry sample with gloved hands. Transfer a subsample of
approximately 15g into a pre-numbered aluminum weighing dish of known weight. Because the
aluminum weighing dishes are manufactured to be a nearly-constant weight, the average weight
of ten labeled aluminum weighing dishes may be used. Enter the dish number on the air-dry
moisture raw data form under "TIN_NO". Handle the weighing dish with forceps or finger cots.
Weigh the dish plus sample to the nearest 0.01g and record this initial weight under
"INIT_WT". Place the dish in a convection oven which has equilibrated at 105°C for mineral
samples or 60°C for organic samples. Allow the sample to oven dry overnight at this
temperature.
81
-------
The following morning, remove the sample from the oven and allow to cool for 30
minutes in a desiccator. Weigh the dish plus sample to the nearest 0.01g and record this oven-
dry weight under "OD_WT".
Enter the data from the raw data form into the computer entry and verification system
and check for any flags (see Appendix B). If the calculated moisture content for a mineral soil
is above 2.5 percent for mineral soils or above 6.0 for organic soils, allow the bulk sample to
continue air-drying and repeat the procedure at a later time. However, if the sample is below
the cutoff percent moisture content for the soil type, then the air-dry sample may be rebagged
and placed in cold storage or it may immediately undergo the next stage of processing.
3.4.5 Calculations
The following calculations are performed by the computer after the raw data are entered,
but they are presented here for internal use. It is advantageous to manually check a few
samples using these calculations in order to understand the procedure and to test the accuracy
of the computer program.
Percent air-dry moisture (MOIST P) = INIT WT - OD WT
xfOO
OD WT - TIN WT
3.5 DISAGGREGATION AND SIEVING
3.5.1 General Considerations
After a bulk soil sample has been determined to be air dry, it is ready to be
disaggregated and sieved in order to remove rock fragments and to prepare the sample for
82
-------
homogenization and subsampling. The disaggregation and sieving areas should be covered
with protective layers of kraft paper and cleaned after each sample has been sieved. Dust
masks and protective clothing should be worn at all times. Some form of ventilation and
heating should be provided for the sieving area.
A source of compressed air is necessary to clean the surfaces of the equipment, e.g.,
rolling pin, sieves, and gloves, after processing each sample. However, because water and oil
tend to accumulate in the tank and conduit, the air must be passed through a trap to collect
the offending vapor and liquid.
In addition to cleaning with compressed air, the surfaces of the processing equipment
may be wiped with clean Kim wipes® if soil is adhering to the rolling pin or to the frame of the
sieve. Used tissues are to be discarded after each sample is processed in order to avoid
contamination.
3.5.2 Equipment
Rolling pin, wooden
Rubber stoppers
Brass sieves, square-holed: 2-mm and 4.75-mm
Clean sieving table and work area
Kraft paper, 36" wide rolls
Air compressor, with in-line filter
Gloves, unpowdered
Open-pan balance, accurate to 0.1g
2-L Nalgene bottles
DORP Label A rolls
Wide-mouth funnel
83
-------
Small cellophane bags with twist ties
Dust masks
Bulk sample processing raw data forms
3.5.3 Procedure
A bulk sample must be weighed in its entirety before it is disaggregated and sieved in
order to properly calculate the percentage of rock fragments in the sample. Weigh the bulk
sample while still in its plastic bag to the nearest 0.1g and record this weight under "TOTAL
BULK_WT" on the bulk sample processing raw data form. Large pieces of undecomposed
stems, moss, twigs, or roots are to be removed from a mineral sample before its bulk weight is
determined.
Place a 1-m by 1-m sheet of kraft paper on top of the sieving table. Next, place a 60-cm
by 60-cm sheet of kraft paper on top of the table and spread small portions of the bulk sample
within the confines of the sheet. Carefully examine the nature of the rock fragments within the
sample and determine the amount of pressure to apply to the sample in order to disaggregate
the soil peds without fracturing or crushing the fragments.
Place another 60-cm by 60-cm layer of kraft paper over the sample and gently roll across
the sample with a wooden rolling pin; enough force should be applied to break up the peds, but
not so much that weathered rock fragments are crushed. Place a portion of sample in the 2-
mm mesh sieve and gently push the soil through the sieve with a rubber stopper. Attempt to
include any soil adhering to rock fragments. Remove and discard any small remaining roots
that nest on the sieve. Continue sieving until all of the less than 2- mm soil has passed
through the 2-mm sieve.
84
-------
Save the less than 2-mm sample in a 2-L Nalgene bottle pre-labeled with a DDRP Label
A and place the bottle in cold storage. Continue processing the remaining rock fragments as
described in Section 3.6 below.
3.6 ROCK FRAGMENT DETERMINATION
3.6.1 General Considerations
The considerations for this procedure are the same as those outlined for disaggregation
and sieving in Section 3.5.1.
3.6.2 Equipment
As outlined in Section 3.5.2
3.6.3 Procedure
Save the rock fragments which could not pass the 2-mm sieve and place them on the
4.75- mm sieve. Retain in separate cellophane bags the material which passes the 4.75-mm
sieve and that which does not pass.
Weigh to the nearest 0.1g the fragments which passed the 4.75-mm sieve but were
retained by the 2-mm sieve. Record this weight on the bulk sample processing raw data form
under "2- TO 4.75-MM ROCK FRAGMENTS". Then weigh the fragments that were retained by the
4.75-mm sieve. Record this weight under "4.75- TO 20-MM ROCK FRAGMENTS".
85
-------
Place the rock fragment bags for each sample in the leftover plastic bag from the bulk
sample. Make sure that a DORP Label A with the appropriate sample code and set ID is
affixed to the bag. Save all rock fragment bags, organized by set ID, until instructed by the soil
sampling task leader to discard or transfer the fragments to another location. It is not
necessary to keep the rock fragments in cold storage.
3.6.4 Calculations
The following calculations are performed by the computer after the raw data are entered,
but they are presented here for internal use. It is advantageous to manually check a few
samples using these calculations in order to understand the procedure and to test the accuracy
of the computer program. The percentages of fine gravel (RF_FG) and medium gravel (RF_MG)
are calculated as follows:
RF FG = 2- TO 4.75-MM ROCK FRAGMENTS
x100
TOTAL BULK_WT
RF MG = 4.75- TO 20-MM ROCK FRAGMENTS
x100
TOTAL BULK WT
3.7 HOMOGENIZATION AND SUBSAMPLING
3.7.1 General Considerations
In order to obtain representative volumes of soil suitable for detailed analysis at the
analytical laboratories, it is necessary to prepare homogenous subsamples from the less than
2-mm soil fraction through the use of a riffle splitter.
86
-------
The riffle splitter is to be cleaned with compressed air after each sample is split. If the
operator suspects that the riffle splitter is still not thoroughly clean, the riffle splitter should be
washed with deionized water and allowed to air dry completely. The work area is to be cleaned
with a brush or hand vacuum after each sample is processed. The work area should be vented
with an operating fume hood.
The preparation laboratory prepares two different subsamples because different
analytical laboratories are responsible for the general analyses of 47 physical and chemical
parameters and for the elemental analysis of total carbon, nitrogen, and sulfur. The general
analytical sample should weigh approximately 500g, while the elemental analytical sample
should weigh approximately 50g. Each analytical sample is derived through homogenization and
subsampling by the procedure described below.
3.7.2 Equipment
Riffle splitter, Jones-type, closed-bin, two-pan, 1.25-cm baffles
Distribution pan
Open-pan balance, accurate to 0.1g
Nalgene bottles: 2-L, 500-mL, 125-mL
DDRP Label A rolls
Wide-mouth funnel
Oust mask
Gloves, unpowdered
Air compressor, with in-line filter
3.7.3 Procedure
87
-------
For each sample, pre-label one 500-mL and one 125-mL Nalgene bottle with the DORP
Label A information. Position the two receiving pans under the riffle splitter. Pour all of the
less than 2-mm soil from the 2- L Nalgene sample bottle evenly across the baffles of the riffle
splitter. Transfer the soil from each receiving pan into the distribution pan and replace the
receiving pans under the riffle splitter. Repeat this step five times in succession with the
material in each receiving pan.
Next, pour the sample evenly across the baffles and place the soil from one receiving
pan into the 2-L bottle. Transfer the soil in the other receiving pan to the distribution pan and
continue splitting as necessary until approximately 500g of soil occupies one of the receiving
pans. Place the entire contents of the pan into the pre-labeled 500-mL bottle (general analytical
sample). Repeat the splitting using the remaining soil until approximately 50g of soil occupies
one of the receiving pans; transfer the entire contents of the pan into the pre-labeled 125-mL
bottle (elemental analytical sample).
Transfer all leftover soil into the 2-L bottle which wilt serve as a DORP-MASS archive
sample. The analytical samples are grouped by set ID and placed in cold storage until they are
batched and shipped via overnight express service to the analytical laboratories. The archive
samples are grouped by batch ID and remain in cold storage indefinitely.
3.8 BULK DENSITY DETERMINATION
Density is defined as weight per unit volume expressed in units of g/cm3. The bulk
density of a soil is defined as the weight of dry soil per unit volume including the pore space.
For mineral soils, bulk density generally ranges from 0.6 to 2.0 g/cm3. With increasing organic
matter content, soils generally exhibit a decrease in bulk density because organic matter has
higher porosity and lower density than mineral particles of the same diameter.
88
-------
As in past DORP surveys, the clod method is the primary method for determining bulk
density in the MASS. Two alternate methods are also being attempted for soil horizons that fail
to yield satisfactory clods; one is a volume replacement (VR) method which utilizes a known
volume of small foam beads packed into a cylinder to replace a selected volume of soil
excavated from a given horizon, while the other is a volume filling (VF) method that is used if
the clod or VR methods do not produce representative samples.
3.8.1 Clod Method
3.8.1.1 General Considerations -
Where possible, three replicate clod samples are extracted from each horizon. The
average bulk density of the replicates is assumed to be the bulk density of that particular
horizon. Analysis of the clods is based on the method described in USDA/SCS (1984).
Laboratory personnel should use caution when working around the muffle furnace, as
temperatures of up to 450°C are common. The furnace should be activated only in an operable
fume hood.
Evaporating dishes must be thoroughly cleaned after each use, as their weights are pre-
determined and entered into the computer verification system. Tongs or finger cots should be
used when transferring the dishes from one location to another.
Extra care must be taken in the use of the Saran® powder and acetone solution (see
Section 3.8.1.3). When mixed, the resulting solution has a tendency to volatilize hydrogen
chloride gas which can cause deleterious health effects. The solution should be used only in an
operating vented fume hood. A cartridge-type organic vapor mask and laboratory coat must be
worn.
89
-------
3.8.1.2 Equipment -
Saran:acetone clod dipping mixture, approximately 1:5 by weight
Carboy, 13-gallon, filled with deionized water
Organic vapor mask, cartridge-type
Open-pan balance, accurate to 0.1g
Open-pan balance, accurate to 0.01g
Beakers, 1000-mL and 2000-mL
Ring stand
Test tube clip, adjustable
Convection oven
Evaporating dishes, pre-numbered, tolerance to 450°C
Furnace gloves
Tongs
Muffle furnace
Desiccator and desiccant
Sieve, square-holed, 2-mm
Paper bags, pre-labeled
Clod bulk density raw data form
3.8.1.3 Mixing the Clod Dipping Solution -
NOTE: The clod dipping solution must be mixed only under an operating vented fume hood. The
operator must wear a cartridge-type organic vapor mask and laboratory coat while mixing the
dipping solution.
The Saran® powder is packaged in 540g increments by EMSL-LV QA staff. One packet
mixed with approximately 3400ml_ of acetone provides a nearly 1:5 by weight mixture resembling
90
-------
light syrup. One-gallon paint containers are used for mixing and storing the solution. Slowly
add the contents of each packet to the acetone while stirring. Continue stirring to ensure
thorough dissolution of the powder. The solution normally is well mixed when it has the color
and consistency of amber syrup and there is no gummy saran residue on the edges or bottom
of the container. If the solution is a milky color, additional mixing is required. The powder has
a tendency to precipitate and clump around a container's lower rim when mixed with acetone.
To prevent this, use a long stir bar and thoroughly scrape the inner rims. If possible, use a
magnetic stirrer and a stir bar to mix the ingredients in a beaker until well blended.
After mixing the solution, the container should be capped as tightly as possible to
prevent leakage of the solution or evaporation of the acetone.
3.8.1.4 Procedure -
All raw data obtained from the clod analysis are recorded on the clod bulk density raw
data form (see Appendix A). For each clod to be analyzed, record the set ID, sample code, and
the replicate number on the form. Examine the label attached to the clod and record the
number of field dips performed by the sampling crew under "FIELD_DP" on the form. Weigh the
clod, without removing the label, to the nearest 0.1g and record this weight under
"FIELD_WT/MOIST".
Suspend the clod in a convection oven and dry overnight at 105°C for mineral samples or
60°C for organic samples. The following morning, transfer the clod into a desiccator for 30
minutes. Weigh the sample to the nearest 0.1g and record this weight under "FIELD_WT/DRY'.
Hang the clod on a suspended curtain rod within an operating vented fume hood. Dip
the clod into a container of 1:5 by weight saran:acetone mixture (see Section 3.8.1.3) for three
91
-------
seconds and then allow the coating to dry for approximately 15 minutes. Do not allow the
mixture to impregnate that portion of the hairnet hanging above the top of a clod.
Apply additional dips as necessary until the clod does not produce excessive bubbles
and a coating of saran is obtained that appears to be impervious to water. Usually three to six
dips is sufficient for most clods. Record the total number of dips made in the laboratory under
"LABJDP", then reweigh the clod to the nearest 0.1g and record this weight under "LAB_WT'.
Also, record the dry weight of a 500- g Class P balance weight immediately before and after a
run of samples. This weight is used to calibrate the density readings and is charted on a daily
basis.
Add approximately 1600-mL deionized water to a 2000-mL beaker. Place the beaker of
water on a balance and tare. Record the temperature of the water under "TEMP" so that the
density of the water can be calculated. Also, record the submerged weight of a 500-g Class P
balance weight immediately before and after a run of samples.
Suspend the clod over the beaker by attaching the top of the hairnet to the test tube
clip, then lower the clod gently into the water until the top of the clod is entirely submerged.
Promptly, after three seconds, record the weight displayed on the balance to the nearest 0.1g
under "CLOD_H2O". If the clod floats, forcibly submerge it by pushing down with forceps and
record the weight. Note on the raw data form, with a "V under "FLOAT1, that the clod floated.
Re- tare the balance before proceeding with other clods.
NOTE: When submerging, ensure that the clod is not touching the edge of the beaker and that
the clod label hangs freely. Occasionally, air bubbles may rise from the clod and the weight
reading on the balance does not stabilize but steadily decreases. This occasionally occurs in
clods sampled from relatively porous surface horizons and simply indicates that all of the
primary macropores along the exterior of the clod were not thoroughly water-sealed. By reading
92
-------
the weight immediately after submersion, error from this source of instability is minimized. Note
the bubbling on the raw data form under "COMMENTS". If the data for this clod do not pass
the verification, a "Y" can be entered under "PROB" that will direct the computer program to
delete that value from the average of the replicates. Also note any other problem, such as a
broken clod, under "COMMENTS".
After submersion, place the clod in a pre-numbered evaporating dish and record the dish
number on the raw data form under "CRUCJMO". Remove as much of the clod label and hairnet
as possible and place in a large plastic bag with others from the same run.
Place the evaporating dish plus clod in a muffle furnace equilibrated at 400°C. The
muffle furnace must be within an operating vented fume hood. Allow the saran coating to be
burned off the clod surface for two hours. Allow the furnace to cool, then carefully remove the
evaporating dish from the muffle furnace. After the clod has cooled thoroughly, place it in a
pre-labeled paper bag.
Disaggregate the clod through a 2-mm sieve. Weigh any rock fragments retained on the
sieve to the nearest 0.01g and record this weight under "R_FRAG". Place the rock fragments
into the paper bag and archive all such bags.
3.8.1.5 Assumptions -
It is assumed that the weight of each two-second saran coating applied in the field is
equal to the weight of each three-second coating applied in the laboratory, as some of the
dipping solution normally is absorbed into the clod when it is applied in the field.
93
-------
It is assumed that the specific gravity of air-dry saran is 1.30 g/cm3 and that the coating
loses 15 percent of its weight upon oven-drying. It is assumed that the particle density of the
rock fragments is 2.47 g/cm3.
It is assumed that the average gross weight of the hairnet and clod label is 1.88g and
the net weight of the submerged portion of the hairnet is 0.20g.
3.8.1.6 Calculations -
The following calculations are performed by the computer after the raw bulk density
data are entered. The calculations are presented here for internal use. It is advantageous to
manually check a few samples using these calculations in order to understand the procedure
and to test the accuracy of the computer program. A table of temperature- corrected water
density values is provided in Appendix C.
Use the following calculations to derive the weights and volume of the air-dry and oven-
dry saran coatings:
SARAN_WT (g) = {[(FIELD_DP + LAB_DP] x [LAB_WT - FIELD_WT/DRY]}
LABJDP
OD SARAN_WT (g) = SARAN_WT x 0.85
SARAN VOL (cm3) = SARAN_WT
1.30
Use the following calculations to derive the volumes of rock fragments, water
displacement, and soil/pore fraction:
R_FRAG VOL (cm3) = R_FRAG
2.47
94
-------
WATER VOL (cm3) = CLOD_H2O
DENSITY OF WATER
SOIL/PORES VOL (cm3) = [WATER VOL - (R_FRAG VOL + SARAN VOL)]
Finally, use the following calculation to derive the oven-dry bulk density (BD_CLD) for the
individual clod:
BD_CLD (g/cm3) = [LAB_WT - (R_FRAG + OD SARAN_WT + 1.88)]
SOIL/PORES VOL - 0.20
3.8.2 Known Volume Methods
3.8.2.1 General Considerations -
A volume replacement sample is designated by a "VR1" or "VR2" as the first three
characters of the sample code. The samples are subject to variable volume estimates which
are recorded on the SCS-SOI-232 field data form and transcribed into the sample receipt log
book. Subtracting the initial from the final volume yields the estimated volume of sample
collected.
A volume filling sample is designated by a "VF1" or "VF2" as the first three characters of
the sample code. The volume of this type of sample is based on the absolute height of a 250-
mL beaker, which is a constant 300cm3.
The known volume samples are processed in a manner similar to the method described
in Blake (1965).
3.8.2.2 Equipment -
95
-------
Small paper bags, pre-weighed
Open-pan balance, accurate to 0.01g
Convection oven
Brass sieve, square-holed, 2-mm
Known volume bulk density raw data forms
Known volume log book
Volume data from sample receipt log book
3.8.2.3 Procedure -
Known volume bulk density samples are shipped from the field in small plastic bags.
Upon passing the sample receipt verification, these samples can be placed out to air dry in the
drying area. The drying facilitates the transfer of the entire sample from its plastic bag into a
pre- labeled paper bag. Do not remove any material from the bag, even if the material is
thought to be morphologically different from the type of horizon sampled.
Record the set ID and sample code of the VR or VF sample on the known volume bulk
density raw data form (see Appendix A). Weigh the paper bag plus sample to the nearest 0.01g
and record this weight under "AIR_WT". Place the bag in a convection oven equilibrated at
105°C for mineral soils and 60°C for organic soils. Allow the sample to oven dry overnight.
The following morning, remove the sample from the oven and allow to cool for 30
minutes. Weigh the sample to the nearest 0.01g and record this weight under "OD_WT". Sieve
the contents of the bag through a 2-mm sieve to separate any rock fragments from the fines.
Weigh the rock fragments to the nearest 0.01g and record this weight under "R_FRAG". Place
the rock fragments in the paper bag and archive all such bags of fragments.
96
-------
Transcribe the computed volume (V, minus V() for a VR sample from the sample receipt
log book to the "VOL" column on the raw data form. For a VF sample, 300 cm3 can be
automatically entered under "VOL".
3.8.2.4 Assumptions -
The particle density for rock fragments in a VR bulk density sample is derived individually
by a submersion technique similar to that used in the clod analysis. It is assumed that the
particle density for rock fragments in a VF bulk density sample is based on four classes defined
by different soil parent materials.
3.8.2.5 Calculations -
The following calculations for known volume bulk density (BD_KV) are performed by the
computer entry and verification system after the raw data are entered. It is advantageous to
manually check a few samples using these calculations in order to understand the procedure
and to test the computer program.
FINES_WT (g) = OD_WT - R_FRAG
BD_KV(g/cm3) = FINES_WT
VOL - (R_FRAG / particle density)
97
-------
REFERENCES
Blake, G. R. 1965. In Black, C. A. (ed.) Methods of Soil Analysis, Part 1. American Society of
Agronomy, Madison, Wisconsin.
MacDonald, C. C. 1977. Methods of Soil and Tissue Analysis in the Analytical Laboratory.
Maritimes Forest Research Centre Information Report, M-X-78. Fredericton, New Brunswick,
Canada.
Papp, M. L, R. D. Van Remortel, C. J. Palmer, G. E. Byers, B. A. Schumacher, R. L. Slagle, J. V.
Burton, K. C. Shines, J. E. Teberg, and M. J. Miah. 1989. Direct/Delayed Response Project:
Quality Assurance Plan for Soil Preparation and Analysis in the Mid-Appalachian Region of the
United States. EPA/600/4-89/031. U.S. Environmental Protection Agency, Las Vegas, Nevada.
USDA/SCS. 1984. Soil Survey Laboratory Methods and Procedures for Collecting Soil Samples.
Soil Survey Investigations Report No. 1. U.S. Government Printing Office, Washington, D.C.
U.S. Environmental Protection Agency. 1985a. Direct/Delayed Response Project: Soil Survey
Data Quality Objectives (Draft). U.S. Environmental Protection Agency, Washington, D.C.
U.S. Environmental Protection Agency. 1985b. Direct/Delayed Response Project, Long-term
Response of Surface Waters to Acidic Deposition: Factors Affecting Response and a Plan for
Classifying that Response on a Regional Scale. U.S. Environmental Protection Agency, Corvallis,
Oregon.
98
-------
Appendix B
Preparation Laboratory Data Forms
The folbwing labels and forms were used by the preparation laboratory to label samples,
enter raw data, and calculated final data for the Mid-Appalachian Soil Survey.
99
-------
DDRP Label A
Date Sampled:
D D M M M Y Y
Crew ID:
Site ID:
Sample Code: —
Horizon: Depth: — cm.
Set I.D.:.
DDRP Label B
Batch ID:
Sample No.:
100
-------
SOIL SAMPLE RECEIPT FORM
SAX CODE
CREW
ID
SITE
ID
SET
ID
DATE
SAMPLED
DATE
SHIPPED
DATE
RECEIVED
RECEIVED
BT
SAMPLE CONDITION
WET/DRY (H/D) BAG SPLIT (B/S)
SIEVED/USSIEVED (S/D)
UNDER VOLUME (UV)
CLOD
NUMBER
KNOWN
VOLUME
CAVITI
SAMPLE
NUMBER
COMMENTS
-------
SAMPLE ID:
SITE ID:
SET ID:
BATCH ID:
BULK SAMPLE RAW DATA
DATE SAMPLED:
DATE REC'D:
PROCESS START:
PROCESS COMPLETE:
SOIL TYPE: M/0
Initials:
FIELD pH:
Initials:
AIR SAMPLE DRYING:
Date % Moisture Initials
Date:
TOTAL BULK WT:
Initials:
2 to 4.75 mm:
Date:
ROCK FRAGMENT WT:
. g 4.75 to 20 mm:
/ Initials:
ENTERED IN COMPUTER: Date:
Initials:
COMMENTS:
102
-------
pH RAW DATA
DATE / /
SAMPLE CODE
Init OCCS
D
QCCS
D
SAHP.
*
pH
SAMPLE CODE
QCC5
D
QCCS
Final QCCS
SAMP.
*
PH
103
-------
SAMPLE LOSS ON IGNITION DATA
ALL WEIGHTS REPORTED IN 0.01 GRAMS
SAM CODE
CROC.
NO.
AIR
ST.
0. D.
wr.
ASHED
HT.
DATE
ID
COMMENTS
104
-------
SOIL SAMPLE AIR DRT DETERMINATION
ALL WEIGHTS REPORTED TO 0.01 GRAMS
SAM CODE
TIN
NO.
TIN
WT.
AIR DRY
(TIN+SOIL)
OVEN DRY
(TIN+SOI1)
DATE
ID
COMMENTS
105
-------
CLOD BULK DENSITY RAH DATA
page /_
ALL HEIGHTS RECORDED TO 0.01 GRAMS
o
o>
REP
FID DP
moist/dry
CLOD H20
T/H
-------
10)0101 VOLUXE BULK DENSITY DATA
PAGE
ALL WEIGHTS RECORDED TO 0.0 GRANS
SAM CODE
AIR W 00
R FRAG
VOL
(CC.)
DATS
ID
COMMENTS
107
\
-------
DIRECT/DELAYED RESPONSE PROJECT (DDRP)
FORM 101
Date received
by Data Mgt.
D D M M M Y Y
Batch ID
Crew ID
Prep Lab ID
Lab Set Sent to
Set ID
Date Samj
Date Reef
Date Prej
No. of Se
Sample
No.
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Date Shipped
imples _ _
Site ID
Sample Code
Mineral or Organic Samples
Set
ID
Signature of Preparation Laboratory Manager
Comments :
Coal
Fragn
2- 4
4.7
se
tents
4.7-
20.
Moisture
%
Field
Air Dry
PH
LOI
Bulk
Density
g/cm3
108
-------
DIRECT/DELAYED RESPONSE PROJECT (DDRP)
SHIPPING FORM 102
DATE RECEIVED
BY DATA MGT.
D D M M M Y Y
0 D M M M Y Y
Prep Lab ID Dale Received
Batrh ID Dale Shiooed
Analytical Lab
SAMPLE NO
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
ID
BULK SAMPLES
SHIPPED
RECEIVED
Signature of Preparation Laboratory Manager
Comments
PH <: 85
RETURNED TO LEMSCO
(CHECK V IF YES)
H COPY TO SMO
M COPY TO EMSLIV
109
-------
Appendix C
Preparation Laboratory Computer
Entry/Tracking/Verification Procedures
The following text provides an overview of the computer tracking and verification program
designed for use at the preparation laboratory during the Direct/Delayed Response Project Mid-
Appalachian Soil Survey. The system is designed to track the progress of samples, calculate final
data values, perform verification checks, and assign flags to any detected aberrant values.
110
-------
PREPARATION LABORATORY COMPUTER ENTRY/TRACKING/VERIFICATION PROCEDURES
The following text provides an overview of the computer tracking and
verification program designed for use at the preparation laboratory during the
DDRP Mid-Appalachian soil survey. The system is designed to track the progress
of samples, calculate final data values, perform verification checks, and assign
flags to any detected aberrant values.
The program creates the Form 101 that summarizes final data reported by the
preparation laboratory. Hard copies of this form are sent to all DDRP
cooperating agencies and groups, in addition, the Form 101 is available on floppy
disk.
SAMPLE TRACKING
The preparation laboratory manager must track samples as they are received
from the sampling crews, during cold storage, through processing and analysis,
and during batch formation and shipment to the analytical laboratories. The
computer data are used to track the location and progress of each sample while
it is physically at the laboratory.
In order to track the sampling activities and shipment/receipt of bulk
samples, the following information is to be reported to the QA manager weekly:
number of samples received
— number of pedons sampled and percentage of total expected
— number of sets received
In order to track the progress of the preparation activities, the following
information is to be reported to the QA manager weekly:
number and percent of samples that have completed specific analyses
number of bulk samples which are completely processed (excluding bulk
density) and are ready to be batched
number of sets that have been completed
number of batches sent to analytical laboratories
DATA VERIFICATION
The system allows the entry and editing of data by preparation laboratory
personnel and the review of progress and data analysis by the QA staff. After
a run of data is entered into a "temporary" data file, the final values are
calculated and verified. A printout of this information is checked by the
preparation laboratory manager and by a representative from the QA staff. The
data that have satisfied the verification criteria are then moved into a "master"
data file which is used to track progress and percentages of the analyses that
have been completed. The master file cannot be edited by preparation laboratory
personnel.
The five raw data forms that are entered into the computer are: (1) sample
receipt raw data form, (2) clod bulk density raw data form, (3) known volume bulk
density raw data form, (4) loss on ignition raw data form, and (5) bulk sample
processing raw data form, in the descriptions of the forms below, a list of the
data calculations, verification checks, and pertinent verification flags are
included. At the beginning of each description is a list of the data from each
111
-------
particular form which must be transferred to the Form 101. The flags consist
of two characters, an alpha character and a numeric character. The numeric
character designates which data form the flag pertains to, e.g., a flag with the
numeric character "3" pertains to the known volume raw data form.
SUMMARY OF VERIFICATION FLAGS
Sample Receipt Raw Data Form
Al Missing data on form
Bl sample code mislabeled
Cl site ID mislabeled
Dl More than 2 known volume samples
El More than 3 clods
clod Bulk Density Raw Data Form
A2 Missing data on form
B2 Submerged clod weight > Lab clod weight, and clod did not float
C2 Clod floated
D2 Problem clod (deletes value from average)
E2 oven-dry clod weight > Lab clod weight
F2 oven-dry clod weight > Field-moist clod weight
G2 Coefficient of variation > 20 among replicates
H2 Bulk density > 2.0
12 Bulk density < 1.0
J2 Sample code not recorded same as on sample receipt form
K2 Number of replicates not recorded same as on sample receipt form
L2 Replicates > 3
Known Volume Bulk Density Raw Data Form
A3 Missing data on form
B3 Oven-dry weight > Air-dry weight
C3 Weight of fines + Rock fragments > Air-dry weight
D3 Sample code not recorded same as on sample receipt form
E3 Number of replicates not recorded same as on sample receipt form
F3 Bulk density > 2.0
G3 Bulk density < 1.0
H3 Coefficient of variation > 20 among replicates
Loss on Ignition Raw Data Form
A4 Missing data on form
B4 Organic matter > 100%
C4 oven-dry weight > Air-dry weight
D4 Ashed weight > Oven-dry weight
E4 sample code not recorded same as on sample receipt form
F4 sample analysis duplicated
G4 organic soil type not correlating
H4 Mineral soil type not correlating
14 Internal duplicate/routine pair > 15% RSD
J4 FAP samples > 15% RSD
K4 Field duplicate/routine sample > 20% RSD
L4 FAL sample outside accuracy window
M4 FAP sample outside accuracy window
N4 Manager's sample outside accuracy window
112
-------
Bulk sample Processing Raw Data Form
A5 Missing data on form
B5 Mineral soil air-dry moisture > 2.5%
C5 organic soil air-dry moisture > 6.0%
D5 Field-moist pH > 6.0
E5 sample code not recorded same as on sample receipt form
F5 Internal duplicate/routine sample > 15% RSD
G5 Field duplicate/routine sample > 20% RSD
H5 FAP samples > 15% RSD
15 pH Q.CCS buffer < 3.95 or > 4.05
J5 Manager's sample outside accuracy window
K5 FAP sample outside accuracy window
SAMPLE RECEIPT RAW DATA FORM
Form 101 Information;
SAM_CODE
SITE_ID
SET_ID
General;
ERL-C provides EMSL-LV with the pedon number (state, county, unit) and site
ID for each pedon to be sampled. This information is placed in the data base
before preparation activities begin.
All samples received from the field and their condition are recorded on the
sample receipt raw data form and immediately entered into the data base. A file
comparison procedure is initiated, comparing the pedon number and site ID with
the raw data from the sample receipt form. If any discrepancies are found,
processing of the affected samples ceases until the discrepancies are rectified.
All flags for the receipt data are unique and identified by the numeric character
" 1".
After all verification checks for the sample receipt data have been
satisfied, the data are sorted by set ID. All samples with the same set ID will
undergo processing simultaneously.
Calculations;
No calculations are to be performed on the sample receipt form.
Verification Flags
If any data are missing on the form; then flag Al
If sample code is mislabeled; then flag Bl
If site ID is mislabeled; then flag Cl
If more than 2 known volume samples are recorded; then flag Dl
If more than 3 clods are recorded; then flag El
When flags other than Dl and El are generated, processing of the affected
samples must cease until the problem is rectified.
Reporting;
The computer produces a printout, sorted by SET_ID and SAM_CODE, that lists
all flags assigned.
113 '
-------
CLOD BULK DENSITY RAW DATA FORM
Form 101 information;
BD_CLD
General;
The clod data are entered into a temporary working file. The verification
program should be run on a regular basis and the discrepancy printout generated.
All flags for the clod data are unique and identified by the numeric character
"2".
After all verification checks for the clod data have been satisfied, the
data are sorted by set ID and sample code. All samples with the same set ID will
undergo processing simultaneously.
Calculations;
SARAN_WT (g) = (FIELD_DP + LAB_DP) * (LAB_WT - FLD_WT) / LAB_DP
OD_SAR (g) = SARAN_WT * 0.85
SAJRAN VOL (cm3) = SARAN_WT / 1.30
R_FRAG VOL (cm3) = R_FRAG / 2.47
WATER VOL (cm3) = CLOD_H20 / WT_DEN
SOIL/PORES VOL (cm3) = WATER VOL - (R_FRAG VOL + SARAN VOL)
BD_CLD (g/cm3) = [LAB_WT - (R_FRAG + OD SARAN_WT + 1.88)] / SOIL/PORES VOL - 0.20
Verification Flags;
If any parameter field is blank, except for 'COMMENTS'; then flag A2
If CLOD_H20 > LAB_WT and FLOAT not marked 'Y'; then flag B2
If FLOAT is marked "Y"; then flag C2
If PROS is marked "Y"; then flag D2 (deletes value from average)
If FIELD_WT > LAB_WT; then flag E2
If CLOD_OD > FIELD_WT; then flag F2
If CV > 20; then flag G2
If BD_CLD > 2.0; then flag H2
If BD_CLD < 1.0; then flag 12
If a SAM_CODE does not match the receipt file; then flag J2
If REP does not match the receipt file; then flag K2
If REP > 3; then flag L2
Reporting;
For each SAM_CODE there are usually 3 replicates, although this number
varies. Therefore, a check for replicate consistency is necessary. If the PROB
column is marked "Y" for any replicate in a SAM_CODE, this replicate should not
be included in the following calculation.
114
-------
The computer produces a printout, sorted by SET_ID, SAM_CODE, and REP, that
lists all flags assigned. For each SAM_CODE, the mean, standard deviation, and
coefficient of variation of all replicates are listed.
The number of horizons and the percentage of the total number of samples
having clods are reported.
KNOWN VOLUME BULK DENSITY RAW DATA FORM
Form 101 Information;
BD_KV
General;
Known volume samples are taken from those horizons where clods cannot be
extracted.
Calculations;
FINES_WT (g) = (OD_WT - R_FRAG - BAG_WT)
BD_KV (g/cm3) = FINES_WT / (VOL - (R_FRAG / particle density)
verification Flags;
If any parameter field is blank, with the exception of 'COMMENTS'; then flag A3
If OD_WT > AIR_WT; then flag B3
If FINES_WT + R_FRAG > AIR_WT; then flag C3
If a SAM_CODE does not match the receipt form; then flag D3
If REP does not match the receipt form; then flag E3
If BD_KV > 2.0; then flag F3
If BD_KV < 1.0; then flag G3
If CV > 20; then flag H3
Reporting;
The computer produces a printout, sorted by SET_ID, SAM_CODE, and REP, that
lists all flags assigned. For each SAM_CODE, the mean, standard deviation, and
coefficient of variation of all replicates are listed.
The number of horizons and the percentage of the total number of samples
having known volume samples are reported.
LOSS ON IGNITION RAW DATA FORM
Form 101 Information;
OM_LOI
Calculations;
Each CRUC_NO will be associated with a pre-determined CRUC_WT.
OM_LOI = [(OD_WT - ASHED_WT) / (OD_WT - CRUC_WT)] * 100
115
-------
verification Flags;
If any parameter field is blank, with the exception of 'COMMENTS'; then flag A4
If OM_LOI > 100; then flag B4
If OD_WT > AIR_WT; then flag C4
If ASHED_WT > OD_WT; then flag D4
If a SAM__CODE does not match the receipt file; then flag E4
If the SAM_CODE is duplicated on the LOI form; then flag F4
If OM_LOI > 20 and SOILTYP on the bulk sample form is not "O"; then flag G4
If OM_LOI < 20 and SOILTYP on the bulk sample form is not "M"; then flag H4
If the internal duplicate/routine pair is > 15% RSD; then flag 14
If the FAP samples > 15% RSD; then flag J4
If the field duplicate/routine pair > 20% RSD; then flag K4
If a FAL sample is outside accuracy window; then flag L4
If a FAP sample is outside accuracy window; then flag M4
If a MS is outside accuracy window; then flag N4
Reporting;
The computer produces a printout, sorted by SET_ID and SAM_CODE, that lists
all flags assigned.
The number of horizons and the percentage of the total number of samples
that have been analyzed are reported.
BULK SAMPLE PROCESSING RAW DATA FORM
Form 101 Information;
PH_MP
MOIST_P (1,2,3)
RF_FG
RF_MG
General;
General information concerning the bulk sample is recorded on this form.
when the form is complete, the sample is considered ready to be batched and
shipped to an analytical laboratory.
Calculations;
PH_MP (no calculation required)
MOIST_P (1, 2, 3) = [(INIT_WT1 - OD_WT / (OD_WT - TIN_WT)] * 100
Each MOIST_P calculation will be conducted only if the previous MOIST_P is
> 2.5 percent for mineral samples or > 6.0 percent for organic samples.
RF_FG = (RF_FGRAW / A_BULK) * 100
RF MG = (RF MGRAW / A BULK) * 100
116
-------
Verification Flags;
If any parameter field is blank, with the exception of 'COMMENTS'; then flag A5
If SOILTYP = 'M' and MOIST_P (1,2,3) > 2.5; then flag B5
If SOILTYP = 'O' and MOIST_P (1,2,3) > 6.0; then flag C5
If PH_MP > 6.0; then flag D5
If a SAM_CODE does not match the receipt file; then flag E5
If the internal duplicate/routine pair > 15% RSD; then flag F5
If the field duplicate/routine pair for MOIST_P, RF_FG, or RF_MG > 20% RSD; then
flag G5 (or >0.20 SD for PH_MP)
If the FAP samples for MOIST_P, RF_FG, or RF_MG > 15% RSD; then flag H5 (or
0.15 SD for PH_MP)
If pH QCCS out of window; then flag 15
If MOIST_P is outside accuracy window; then flag J5
If PH_MP is outside accuracy window; then flag K5
Reporting;
The computer produces a printout, sorted by SET_ID and SAM_CODE, that lists
all flags assigned.
The number of horizons and the percentage of the total number of samples
that have been analyzed are reported.
117
-------
Appendix D
A Comparison of Soil Sample Homogenization Techniques
The following study was initiated to determine if the homogenization technique used at the
different Direct/Delayed Response Project preparation laboratories was a major source of
measurement uncertainty.
118
-------
Internal Report
A COMPARISON OF SOIL SAMPLE HOMOGENIZATION TECHNIQUES
by
B. A. Schumacher, K. C. Shines, J. V. Burton, and M. L Papp
Lockheed Engineering and Sciences Company, Inc.
Las Vegas, Nevada 89119
Contract Number 68-03-3249
Project Officer
L J. Blume
Exposure Assessment Research Division
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada 89193-3478
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89193-3478
119
-------
Notice
This document is intended for internal Agency use only. Mention of trade names or
commercial products does not constitute endoresement or recommendation for use.
120
-------
ACKNOWLEDGEMENTS
We greatly appreciate the assistance of Gregory A. Raab, William H. Cole III, Conrad A.
Kuharic, Gerald L Byers, Rick D. Van Remortel, and Robert L Tidwell in the technical review of this
document. Our thanks are also extended to Mohammad J. Mian for his aid in producing and
interpreting the statistical analyses.
121
-------
Introduction
The need for sample homogeneity prior to laboratory analyses has been long recognized by
geologists, pedologists, chemists, and members of other scientific disciplines. Homogeneity is the
degree that the material under investigation is mixed resulting in the random distribution of all
particles in the sample. Completely homogenous materials are so rare that they may be considered
nonexistent (Ingamells and Pitard, 1986) yet scientists must strive to obtain a homogenous sample
in order to obtain data exhibiting minimal error attributable to sample heterogeneity.
Care must be taken during the subsampling phase of soil preparation, during sample
transport, unpacking, and transfer to other containers in the laboratory to avoid particle sorting via
different particle densities, shapes, sizes, and resistance of certain minerals to mixing, such as
magnetite (Muller, 1967; Ingamells and Pitard, 1986; Reeves and Brooks, 1978). Numerous methods
have been used to obtain homogeneous soil samples. The various methods range from simple
grinding and sieving of the sample to a desired particle-size to various mixing and splitting devices
and machines. Each method will be discussed individually with both advantages and
disadvantages being presented.
One of the most common methods used to obtain a homogeneous sample is to grind and
sieve the soil to a desired particle-size (generally <2-mm) followed by random sampling. This
method assumes that the initial material is ground without preference to any given factor, such as
color, and that during grinding and sieving, the sample becomes sufficiently homogenized. To
further eliminate possible heterogeneity within the sample and to reduce the sample size to the
desired quantity for a given analysis, subsamples may be obtained by "spooning" (Carver, 1981)
or some other method which involves the random insertion of a spoon or other sampling device into
the previously ground and sieved sample. This method is preferably done rapidly and without
extensive visual examination of the sample that could lead to a processor preference in certain
cases, for example, light catching the shiny surfaces of mica flakes leading to preferential inclusion
or exclusion of that part of the sample. The "spooning" type of subsampling markedly reduces the
time of sample processing in comparison to multiple, successive splitting operations and has been
shown to produce equivalent correlation coefficients between observed settling velocities of sands
(median r-0.994 using other splitting methods and median r=0.9955 using the spooning technique)
when used to obtain 20 gram subsamples from initial 1 to 2 kg sand samples (Carver, 1981).
Two other methods have been devised which are similar to the "spooning" method except
they involve an intermediate step between grinding and sieving and subsampling procedures.
Gilliam and Richter (1985) used an intermediate step of stirring the sample with a spatula before
subsampling occurred, presumably until visual homogeneity was obtained. Some analysts content
themselves by merely shaking the sample in a bottle prior to subsampling and ignore the risk of
constituent segregation (Reeves and Brooks, 1978).
An additional method of sample subdivision with the goal of obtaining a representative
sample was presented by Allman and Lawrence (1972). Their method is similar to the "spooning*
method of Carver (1981) except that a scoop of ground and sieved materials was divided among
four containers. The process was repeated continually changing the filling order of the containers
until the sample had been quartered or an appropriate sample size had been obtained. As with the
spooning process, visual bias as to how much sample and into which container the samples were
placed is a concern.
Before proceeding to the more elaborate sample splitting schemes, a discussion of methods
used simply to mix (homogenize) the ground and sieved sample is warranted. The simplest of the
homogenization processes is tumbling the sample on a sheet of paper, cloth, or plastic. This
process involves the manual rolling of ground samples such that the sample must tumble upon
itself and not just slide along the surface of the sheet (Schuler, 1971; Van Johnson and Maxwell,
1981; Ingamells and Pitard, 1986). This method is effective on sample sizes less than 2 kg, yet
caution must be taken to ensure that the sheet material does not contain any element which is to
122
-------
Fig. 1. A twin shell V-blender.
be quantitatively analyzed and does not develop static charges that may lead to segregation of the
finer particle sizes.
Homogenization may also be achieved through the use of mechanical mixing devices
including a spiral mixer, a cement mixer, and a twin shell V-blender. The spiral mixer involves the
rotation of the bottled sample in both horizontal and vertical planes (Van Johnson and Maxwell,
1981). The cement mixer or similar devices involve the rotation of the sample in a chamber with
a series of internal baffles that cause the materials to be thoroughly tumbled and mixed. These
two methods are useful for samples ranging from less than one pound to several hundred pounds.
The twin shell V-blender involves the rotation of two hollow cylinders about an horizontal axis such
that the apex describes a circle in the vertical plane (Schuler, 1971). Twin shell blenders are
available commercially in sizes ranging from 4 to 16 quarts (approximately 10 to 40 kg of a mineral
soil) internal capacity (Fig. 1). However, Ingamells and Pitard (1986) expressed serious doubts as
to the value of mechanical splitters. They stated that, "In general, machines that use mechanical
violence and look and sound as though they are efficient are most likely to cause segregation of
heavy and light, large and small, and flat and round particles*. Further concerns have been
expressed concerning sample breakdown to finer particle sizes due to the violent tumbling in the
machines.
The riffle splitter (also called a chute splitter, Jones splitter, or just sample splitter) is
perhaps the most common mechanical method for sample homogenization and/or sample
sizereduction (Ingram, 1971, Mullins and Hutchison, 1982). The riffle splitter also provides one of
the best general methods of sample mixing to obtain bulk sample homogeneity (Ingamells and
Pitard, 1986). A riffle splitter is a device having an equal number of narrow sloping chutes with
alternate chutes discharging the sample in opposite directions into two collection bins (Figs. 2 and
3). Sample homogenization is achieved by repeated pouring of soil through the splitter and
combining the halves between passes. The use of the riffle splitter as a subsampling device is
done in a similar manner with the exception that after the sample is passed through the splitter,
123
-------
60° slope
on chutes/1
knife edges
alternate openings
/collecting bin
collecting
bin
Fig. 2. An open-bin riffle splitter.
Fig. 3. A closed-bin riffle splitter.
one collection pan is replaced with a clean pan. The material in the "replaced" pan, which contains
about one-half of the original sample, is then passed through the riffle splitter again thereby
reducing the volume in the clean pan to one-quarter of its original sample volume. This process
of sample reduction is repeated until the desired weight or sample size is obtained.
Many variations on the size and construction materials have been built on the principle of
the riffle splitter. For small samples (10 to 50 grams), an efficient microsplitter has been designed
by Newton and Dutcher (1970) using balsa wood and glass microscope slides. These authors
conducted experiments using 40 gram samples of the fine sand fraction and found only an 1.2%
average error when quartering the sample by a 2-step halving. The results of Newton and Dutcher
(1970) are supported by the earlier work of Griffiths (1953) who found that sample homogenization
using a riffle splitter produced an overall coefficient of variance (Cy) between median grain sizes
of less than 3% in 9 out of 10 samples run on various rock and sand samples. In the one case
where the Cv was greater than 3%, the error was attributed to operator differences (16 in the single
case vs. 5 or less in the other nine cases). A portable sieving and splitting device for field use has
been designed by Ibbeken (1974) which uses the riffle splitter to subsample and process coarse
grained sediments. During his research on unconsolidated sediments, no significant differences
were found between splits in terms of petrographic mineralogical composition, with the exception
of the 125 to 160 mm serpentine fraction (represented by only four pebbles), nor in grain size
distribution. Another variation of the riffle splitter found in the literature was the use of a single
piece of tin-plate bent several times to form a riffle splitter (McKinney and Silver, 1956). The
advantage of this construction method is that free grain flow in the chutes was obtained, without
the hindrance of unevenly soldered joints that may be present in other riffle splitters.
The use of riffle splitters and their variations are valuable in splitting samples which range
in size from less than 10 grams (Humphries, 1961) to several kilograms (Van Johnson and Maxwell,
1981). Ibbeken (1974) reported that 0.5 to 1 ton (455 to 909 kg) of sample can be processed daily
using a standard riffle splitter to reduce the initial quantity to 5 kg subsamples. Although most
authors find the use of the riffle splitter to be an effective, efficient method for sample
homogenization and sample splitting, several problems exist.
124
-------
The major source of error involved in using a riffle splitter is the loss of fine particle sizes
via "dusting" when processing air-dried or oven-dried samples which contain fine particle sizes (Van
Johnson and Maxwell, 1981). The process of riffle splitting requires that a uniform stream of
material be poured into the mouth or top of the splitter. Dust loss may occur through the chute
ends, in the collection bins, and through the mouth openings (Fig. 2) in an open-bin system or just
through the mouth if a closed-bin system is used (Fig. 3).
Several studies have been conducted comparing the effectiveness of the riffle splitter to
other homogenization techniques. Wentworth et al. (1934) compared the riffle splitter with a rotary
splitter (to be discussed) and found that the rotary splitter more accurately split an initial sample
of known grain size distribution into subsamples with similar grain size distributions than the riffle
splitter. Kellagner and Flanagan (1956) in a comparison experiment among a multiple-cone splitter
(to be discussed), cone and quartering (to be discussed), and a riffle-based microsplitter found the
microsplitter was the worst for both accuracy and precision of grain frequency percentages for
subsampling three different weights (5, 10, and 20 grams) of an artificially created very coarse and
coarse sand fraction mixture. In a similar type of study, Mullins and Hutchison (1982) compared
the Cv among sand fraction contents of several soils and ranked, in order of best to worst, the
rotary subsampling, riffle splitting, cone and quartering, and spoon sampling, in terms of their ability
to homogenize a sample. These authors did note, however, that the best and worst methods were
only significantly different at the 10 percent level.
Perhaps the best known sample splitting method is the classical cone and quarter
technique. This technique involves pouring the sample into a cone, flattening the cone, dividing the
flattened cone into four equal divisions (quartering), and then removing 2 opposite quarters (Fig.
4). The remaining two quarters are replied into a cone and the process is repeated until the desired
sample size is obtained. Variations on the process are possible which can enhance the speed of
sample size reduction by using just one quarter (chosen at random) to continue the splitting
process or which allow this method to be modified to homogenize a large sample. The use of the
cone and quarter method to homogenize a sample involves the removal of the first quarter and
repiling it into a cone followed by the subsequent repiling of the opposite quarter and then the
remaining two quarters to reform a single cone (Raab et al., 1990). This process is repeated
several times until sample homogeneity is achieved.
Several sources of error for this method have been identified. Van Johnson and Maxwell
(1981) reported that during the cone and quarter process on large samples (several kilograms),
there is the danger of unequal segregation of heavier materials during the flattening and coning
of the sample. Similar to the riffle splitting techniques, dusting is also a possible source of error
during cone formation. Sample loss from the inability to recollect all the soil from the underlying
material, the ability of the sample to "cling" to the underlying material via static charges, and
sample embedment are all further sources of error. Muller (1967) placed a limitation on the cone
and quarter technique to samples greater than 50 grams and stated that this method is most
successful in the field for sample processing of larger sample sizes. Raab et al., (1990) used the
cone and quarter method to homogenize large volumes of soil («150 kg) and then subsampled the
repiled cone with a 2-foot long plastic tube (2-inch internal diameter). They then sieved the sample
into 3 particle size classes (2-mm to 0.105-mm, 0.105-mm to 0.053-mm, less than 0.053-mm) and
subsampled using a riffle splitter. A resieving to check particle class weights of the three divisions
between the six subsamples resulted in the finding that no significant difference (error level = .05)
among the subsamples indicating a homogenous mixture had been obtained.
125
-------
Fig. 4. The cone and quarter technique.
Several other sample splitting/homogenizing devices found in the literature include a rotary
splitter (Muller, 1967; Schuler, 1971; Allman and Lawrence, 1972), a brass disc microsplitter (Brewer
and Barrow, 1972), and a multiple cone splitter (Kellagher and Flanagan, 1956). The rotary splitter
involves the pouring of the sample into a feeding hopper located above a rotating disc containing
sample bottles or pans (Fig. 5). The disc is mechanically rotated and the sample is divided among
the collecting bottles or pans. The brass disc microsplitter was designed by Brewer and Barrow
(1972) for subsampling small paniculate samples which range up to 1 gram in the particle size
range between 10 and 200 microns. This splitter divides the sample by passing it through a fluted
brass disc which contains several holes to separate the sample into different collection bottles.
The device also has a mechanical vibrator system to ensure complete collection of the fine
particles. The multiple-cone splitter of Kellagher and Flanagan (1956) consists of three funnels
and brass cones vertically mounted which are capable of splitting small samples (10 to 400 grams)
into 5 mg subsamples which are collected in one of four divisions in the base pan (Fig. 6). These
authors statistically compared the precision and accuracy of grain frequency percentages obtained
using their device with a riffle-based microsplitter and the cone and quarter technique and found
that their splitter was better in both categories than the riffle splitter while being better than the
cone and quartering in terms of accuracy only.
Due to the overwhelming concurrence (although not unanimous) in the literature as to the
value of the cone and quarter and riffle splitting techniques in sample homogenization and the
common use of random sampling after grinding and sieving of the soil, investigations were
undertaken to determine the effectiveness (degree of homogenization), efficiency (time
consumption), and the extent of the loss of fine particles by these methods during the
homogenization of large soil samples.
126
-------
Fig. 5. A rotary splitter.
i / >^
» I / >M»BIL |rr»li>f»
Fig. 6. A multiple-cone splitter.
-------
Materials and Methods
Bulk samples (approximately 15 kg) of surface horizons of the Overton clay (fine,
montmorillonitic, calcareous, thermic Mollic Haplaquepts), Gila silt loam (coarse-loamy, mixed,
calcareous, thermic Typic Torrif luvents), Calico loam (coarse-loamy, mixed calcareous, thermic Typic
Torrifluvents), and Jean gravelly loamy fine sand (sandy-skeletal, mixed, thermic Typic Torriorthents)
were collected in Clark County, Nevada (Speck, 1985). Surface horizons with different textures
were selected to represent soil systems with varying sensitivities to the loss of fines that may
occur during sample homogenization.
Moist bulk samples were split into three subsamples (approximately 5 kg each) and air-
dried. Samples were ground and sieved through a 2-mm (10 mesh) sieve with three random
samples being collected from each of the subsamples prior to any homogenization procedures.
The three original subsamplos from each soil series were "homogenized" by either passing the soil
seven times through the cone and quarter technique, an open bin riffle splitter, or a closed bin riffle
splitter. Three samples of approximately 60 g each were collected after the first, third, and fifth
passes. Seven samples were collected after the seventh pass through each of the homogenization
processes. Samples were collected from the riffle splitters by placement of the receiving bottle
under randomly selected chutes prior to the sample being poured through the splitter. Samples
from the cone and quarter technique were collected by pouring the sample over the collection
bottles during cone rebuilding and subsequently removing the bottles when they were full prior to
completion of the cone using all four quarters.
A timed experiment was conducted to determine the efficiency of each splitting technique
and involved the passing of the previously "homogenized" soil (from the seventh pass) through the
three homogenization procedures seven times without collection of soil samples after any
intermediate passes. Seven samples were collected after the timed experiment in which the soil
had now been passed fourteen times through the homogenization process.
Three parameters were selected to determine the effectiveness of the homogenization
method and loss of fines from the system, namely, particle-size analyses (for loss of inorganic
fines), loss on ignition (LOI) organic matter (for organic fine losses) and pH (for bulk chemical
changes). Particle-size distribution (<2-mm) was determined usjng the pipette method described
by Gee and Bauder (1986). Soil samples were oven-dried at 110° C overnight and loss on ignition
organic matter content was determined gravimetrically after heating at 450 C for a minimum of
6 hours. The pH values were determined in a 1:2 soil:0.01 MCaC\2 solution ratio (McClean, 1982).
Levels of confidence were determined by analysis of variance (ANOVA).
Results and Discussion
It should be noted that pass no. 0 represents the random samples collected prior to any
sample homogenization process other than preparatory grinding and sieving. Complete data tables
containing individual analyses of sand, silt, clay, sand fractions, pH (no standard deviations or
%RSDs presented), air-dry moisture contents, and LOI organic matter as well as means , standard
deviations, and %RSDs for replicate and cumultive (for all samples in the given treatemtn of a given
texture) data are presented in the Appendix
Observed Sources of Soil Loss
Loss of fines via dusting was most apparent during the use of the open bin riffle splitter.
Dust loss occurred from the mouth of the riffle splitter due to the air-dried soil hitting the baffles
and sliding down the chutes as well as in the collecting bin from the soil falling upon itself.
Apparent fine particle losses from use of the closed bin riffle splitter were less noticeable than for
the open bin riffle splitter. Dust losses occurred through cracks within the riffle splitter and through
the mouth into which the soil was being poured. However, although the overall dust loss appeared
to be less, additional soil loss was observed within the closed bin riffle splitter where soil collected
128
-------
on internal ledges and around the outside of the collecting bins. Little loss of fines, via dusting,
was observed using the cone and quarter technique. Dusting occurred only during the piling of the
quarters upon each other to form a new cone. Soil losses were observed due to an inability to
completely transfer all soil materials during new cone formation.
Influence of Sample Homogenization on pH
No significant differences were found in pH values among replicates regardless of soil
texture, splitting method, or the number of the pass from which the subsample was obtained (Table
1). All replicate pH values met the intralaboratory precision goals set for the Mid-Appalachian soi
survey (0.10 pH unit). The ranges in pH for the Jean gravelly loamy fine sand and Calico sandy
loam were 7.7 to 7.8. The Gila silt loam had pH values of 7.8 in all samples while the Overton clay
had pH values ranging from 7.9 to 8.0.
Influence of Sample Homogenization on Particle-Size Distribution
All replicate total sand, silt, and clay contents met the intralaboratory precision goals set
for the Mid-Appalachian soil survey (3.0 wt% standard deviations for sand and silt; 2.0 wt%
standard deviations for clay). Standard deviations for total sand, silt, and clay contents ranged
from 0.05 to 1.30, 0.05 to 2.59, and 0.00 to 2.00, respectively (Table 2).
The lowest standard deviations among replicates were almost always found after the
samples had been passed five times through the homogenization process regardless of
homogenization technique used or soil texture (Table 2). When exceptions did occur, the standard
deviations were not significantly different from the fifth pass and occurred in either the random
sample or after the first pass through the homogenization process in which the sample was simply
halved via riffle splitting or was effectively a random sample from the first flattened quarter for the
cone and quarter technique. At the 90% or greater confidence interval, the random (pass 0), first,
third, seventh, and timed experiment passes had 36, 33, 53, 86, and 97% of the samples having
significantly greater replicate variances than the fifth pass, respectively. If a 75% or greater
confidence limit was used, pass 0, 1, 3, 7, and T had 53, 56, 78, 98, and 100% of the samples,
respectively, had significantly greater variances than the fifth pass. Earlier passes (pass 0, 1, and
3) resulted in less samples with significantly different variances among the replicates than in
passes after the fifth, yet one-third to more than one-half of the samples had significantly greater
variances than found after five homogenization passes. These data indicate that any attempts to
further homogenize the soil after the fifth pass through the various homogenization methods
129
-------
Table 1. Soil pH values determined in a 1:2 soil:solution
of 0.01 M CaCI. for homooenization studv*.
Split No.
0
1
3
5
7
T
0
1
3
5
7
T
0
1
3
5
7
T
0
1
3
5
7
T
Open Bin Closed Bin Cone and Quarter
Riffle Solitter Riffle Solitter Techniaue
7.7
7.7
7.7
7.7
7.7
7.7
7.7
7.7
7.7
7.8
7.7
7.7
7.8
7.8
7.8
7.8
7.8
7.8
7.9
7.9
7.9
7.9
7.9
7.9
Jean loamy fine sand
7.8
7.7
7.7
7.7
7.7
7.7
Calico sandy loam
7.8
7.7
7.8
7.7
7.7
7.8
Gila silt loam
7.8
7.8
7.8
7.8
7.8
7.8
Overton clay
8.0
8.0
8.0
8.0
8.0
8.0
7.8
7.7
7.7
7.7
7.7
7.7
7.7
7.7
7.7
7.7
7.7
7.7
7.8
7.8
7.8
7.8
7.8
7.8
7.9
7.9
7.9
7.9
7.9
7.9
a = data presented are the means of 3 replicates for splits 1
through 5 and means of 7 replicates for splits 7 and T.
130
-------
Table 2. Standard deviations between replicate analyses for particle-size distribution and loss on
ignition organic matter*.
0
1
3
5
7
T
Open Bin
Riffle Splitter
Closed Bin
Riffle Splitter
Cone and Quarter
Techniaue
Split Sand" Silt Clav O.M. Sand Silt Clav O.M. Sand Silt Clav P.M.
Jean loamy fine sand
-)
0.15*"
0.12"
0.22"*
0.05
0.18"*
0.17*"
0.08
0.20*"
0.24*"
0.06
0.13"*
0.21*"
0.15*
0.08
0.08
0.08
0.16"*
0.1f"
0.05*"
0.01
0.03***
0.01
0.02"*
0.03*"
' 0.09
0.19*
' 0.73"*
0.09
' 0.25*"
' 0.22"*
0.05
0.16
0.63***
0.11
0.23**"
0.24*"
0.08
0.11
' 0.11
0.09
' 0.10*
0.13*"
0.07*
0.03
0.07*
0.04
0.14"*
0.16*"
0.33*"
0.10
0.19*
0.10
0.18"*
0.30"*
0.38*"
0.06
0.25"
0.10
0.15*"
0.35*"
0.08'"
0.04*
0.07"
0.02
0.07*"
0.10*"
0.05"
0.04*
0.03
0.02
0.03"
0.01
0
1
3
5
7
T
0.10
0.64"
0.89*"
0.23
0.30"
0.46"*
0.36"
0.63*"
0.79*"
0.15
0.32*"
0.74*"
0.41"
' 0.22
' 0.20
0.16
' 0.30*"
' 0.52*"
0.10*"
0.03
0.06"
0.02
0.04*"
0.10*"
0.34
0.52*
0.40
0.28
0.34"
0.33"
0.34
0.43
0.60*
0.33
0.36*
0.45"
0.00
0.10
0.21*
0.09
0.42*"
0.35*"
0.07*
0.07*
0.06*
0.03
0.11"*
0.06*"
0.27*
0.29"
0.16
0.12
0.47"*
0.54"*
0.48*"
0.13*
0.25*"
0.07
0.34*"
0.52"*
0.31*"
0.16*
0.11
0.08
0.34*"
0.24*"
0.08'
0.05'
0.02'
0.01
0.05'
0.07'
0
1
3
5
7
T
0.31*
0.20
0.31*
0.18
0.34*"
0.83"*
0.41*
0.19
0.26
0.19
0.51"*
0.88"*
0.21
0.19
0.39"
0.17
0.28*"
0.43*"
0.02*
0.04*"
0.06*"
0.01
0.07*"
0.08*"
0.37
0.48*
1.30*"
0.29
0.56*"
0.37"
0.30
0.17
0.96*"
0.25
0.49**"
0.42*"
0.26
0.32*
' 0.35*
0.21
' 0.22*
' 0.23*
0.09***
0.08"*
0.06"
0.02
0.06"*
0.07*"
0.67*"
0.23
0.50"
0.19
0.19
0.40***
0.38*"
0.06
0.51*"
0.11
0.34*"
0.28"*
0.46'"
0.28**
0.14
0.12
0.23*"
0.39*"
0.03
0.04
0.04
0.03
0.07'
0.07'
Overton clay
0
1
3
5
7
T
1.11"*
0.83"
0.61*
0.28
0.71"*
0.53*"
0.57
2.59*"
2.07*"
0.58
1.02"*
233*"
0.56*
1.85*"
1.46*"
0.31
0.57*"
200*"
0.05
0.07*
0.09*
0.04
0.07*"
0.05"
0.32
0.07
0.51*
0.29
0.32*
0.53*"
0.24*
0.21*
0.39"
0.13
0.37"*
0.65*"
0.30
0.26
0.90"*
0.24
0.47"*
0.59*"
0.24***
0.09*"
0.12*"
0.02
0.06"*
0.08*"
0.25" 0.17
0.59*" 0.66"
0.20" 0.50*
0.09 0.23
0.51"* 0.38*"
0.42"* 0.85"*
0.08
0.55"
0.44*
0.20
0.46*"
0.66*"
0.10""
0.06*
0.08"
0.03
0.07"*
0.10*"
a - data presented are the standard deviations of 3 replicates for splits 1 through 5 and means
of 7 replicates for splits 7 and T.
b - * = 75%; ** = 90%; *** = 95% confidence levels as determined by ANOVA. Significant
differences are compared to the fifth pass.
131
-------
markedly increased the variability among replicates and thus the heterogeneity of the sample
regardless of soil texture or homogenization technique.
Influence of Sample Homoqenization on Loss on Ignition Organic Matter
The lowest standard deviations among replicates for LOI organic matter content were found
after the fifth pass, with only two exceptions, similar to the results for standard deviations in
particle-size distribution (Table 2). The two exceptions were noted in the Jean gravelly loamy fine
sand (after the first pass through the closed-bin riffle splitter and after the timed experiment
homogenized by the cone and quarter technique), yet neither standard deviation was significantly
less than the standard deviation of the fifth pass. When a confidence interval of 90% or greater
was established, 58, 33, 50, 100, and 91% of the replicates from passes 0, 1, 3, 7, and T,
respectively, had significantly greater variances than those found in the fifth pass. At the 75% or
greater confidence level, two-thirds or more of the replicates (83, 66, 83, 100, and 91% for passes
0, 1, 3, 7, and T, respectively) had significantly greater variances. These results support our earlier
findings that the sample appears to have achieved the greatest homogeneity after the fifth pass
through the homogenization process.
Homogenization Technique Efficiency
The average time required to homogenize a sample using seven passes was 5.00 (range
4 to 5 minutes), 5.50 (range 4 to 7 minutes), and 30.75 minutes (range 24 to 44 minutes) for the
closed-bin riffle splitter, open-bin riffle splitting, and cone and quartering, respectively (Table 3). The
slightly longer time required to perform open-bin riffle splitting compared to closed-bin riffle splitting
was attributed to an initial lack of experience in the use of a riffle splitter by the technician.
Excluding the first time use of either riffle splitter, the average time required for open-bin riffle
splitting decreased to 5.00 minutes per sample compared to 4.3 minutes per sample required for
the closed-bin riffle splitter. Excluding the first use of the cone and quarter technique by the
technician, the average time required to homogenize the sample decreased from 30.75 to 26.33
minutes per sample. These results indicate that the use of the riffle splitter for homogenization
required markedly less time per sample than samples homogenized by the cone and quarter
technique and was thus more efficient.
132
-------
Table 3. Efficiency of various homogenization techniques.
Open Bin Closed Bin Cone and Quarter
Soil Riffle Splitter Riffle Splitter Technique
( minutes )
Jean
Calico
Gila
Overton
Mean
Mean-1b
4.
7.*
6.
5.
5.50
5.00
4.
4.
5.
5.*
4.50
4.33
24.
44.*
29.
26.
30.75
26.33
a - * = first sample homogenized by the given method.
b - Mean-1 = mean of timed experiment excluding first
sample homogenized by the given method.
Table 4. Loss of fines between the seventh pass and
timed experiment.
Open Bin
Soil Riffle Solitter
Jean
Calico
Gila
Overton
-0.03'
-0.02
+0.09
-0.36
Closed Bin Cone and Quarter
Riffle Splitter Techniaue
%fslf)\/« .
nc
-0.10
-0.22
-0.19
\
+0.06
-0.56
-0.57
-0.42
a - + = clay increase; - = clay loss; nc = no change.
133
-------
Loss of Fines
Loss of fines was determined by comparison of clay contents between the seventh split and
timed experiment. Clay contents generally decreased after the additional seven passes through the
homogenization procedures but the overall clay content loss was very small (less than 0.6%) and
could be attributed to expected variance in the homogenization and analytical methods (Table 4).
It was interesting, however, to note that the greatest clay content decreases were found when the
soils underwent cone and quartering as the homogenization process. This result was due to a two-
fold effect in which the inability to recover all the soil from the underlying paper led to a greater
clay loss from gap filling between the larger sand particles and perhaps due to static charge
development on the paper leading to a retention of the charged clay particles.
Conclusions
The use of riffle splitting to homogenize a bulk soil sample was more efficient and had less
loss of fines than cone and quartering and is therefore the recommended homogenization
technique. The use of a closed-bin riffle splitter was preferred to an open-bin riffle splitter due to
its greater apparent ability to contain and reduce the loss of fines from dusting. Only five passes,
instead of seven, should be used to obtain the most homogeneous sample, in terms of particle-
size distribution and loss on ignition organic matter, due to the overwhelming evidence that the
least variability among replicates occurred after the fifth pass using all three splitting techniques
and all four soil textures. Random sampling after grinding and sieving was the most efficient
homogenization method, since no additional sample preparation was involved, yet these samples
almost consistently had greater replicate variabilities than the other homogenization techniques
after the fifth pass reducing the value of this technique for soil sample homogenization.
134
-------
REFERENCES
Allman, M. and Lawrence, D.F. (1972). Geological Laboratory Techniques. ARCO Publishing Co., Inc.
New York. 335 pp.
Brewer, R. and Barrow, K.J. (1972). A Microsplitter for Subsampling Small Paniculate Samples. J.
Sed. Pet. 42:485-487.
Carver, R.E. (1981). Reducing Sand Sample Volumes by Spooning. J. Sed. Pet. 51:658.
Gee, G.W. and Bauder, J.W. (1986). Particle Size Analysis. In A. Klute (ed.) Methods of Soil Analysis,
Part 1. Agronomy 9:383-411.
Gilliam, F.S. and Richter, D.D. (1985). Increases in Extractabte Ions in Infertile Aquults Caused by
Sample Preparation. Soil Sci. Soc. Am. J. 49:1576-1578.
Griffiths, J.C. (1953). Estimation of Error in Grain Size Analysis. J. Sed. Pet. 23:75-84.
Humphries, D.W. (1961). A Non-Laminated Miniature Sample Splitter. J. Sed. Pet. 31:471-473.
Ibbeken, H. (1974). A Simple Sieving and Splitting Device for Field Analysis of Coarse Grained
Sediments. J. Sed. Pet. 44:939-946.
Ingram, R.L (1971). Sieve Analyses. In. R.E. Carver (ed.) Procedures in Sedimentary Petrology. J.
Wiley and Sons, Inc. New York. pg. 49-67.
Ingamells, C.O. and Pitard, F.F. (1986). Applied Geochemical Analyses. J. Wiley and Sons, Inc. New
York. 733 pp.
Kellagher. R.C. and Flanagan, F.J. (1956). The Multiple-Cone Splitter. J. Sed. Pet. 26:213-221.
McClean, E.O. (1982). Soil pH and Lime Requirement. In A.L Page (ed.) Methods of Soil Analysis,
Part 2. Agronomy 9:199.223.
McKinney, C.R. and Silver, LT. (1956). A Joint-Free Sample Splitter. Amer. Mineralogist 41:521-523.
Muller, G. (1967). Methods in Sedimentary Petrology. In W. Van Engelhart et a!., Sedimentary
Petrology, Part 1. Hafner Publishing Co., New York. 283 pp.
Mullins, C.E. and Hutchison, B.T. (1982). The Variability Introduced by Various Subsampling
Techniques. J. Soil Sci. 33:547-561.
Newton, G.B. and Dutcher, R.R. (1970). An Inexpensive Student Sample Splitter. J. Sed. Pet.
40:1051-1052.
Raab, G.A., Bartling, M.H. Stapanian, MA, Cole, W.H., Tidwell, R.L and Cappo, K.A. (1990). The
Homogenization of Environmental Soil Samples in Bulk. In M. Simmons (ed.) Hazardous Waste
Measurements. Lewis Pub. In press.
Reeves, R.D. and Brooks, R.R. (1978). Trace Element Analysis of Geological Materials. J. Wiley and
Sons, Inc. New York. 421 pp.
Schuler, V.C.O. (1971). Chemical Analysis and Sample Preparation. In R.E. Wainerdi and E.A. Uken
Ed. Modern Methods of Geochemical Analysis. Plenum Press, New York. 397 pp.
Speck, R.L (1985). Soil Survey of Las Vegas Valley Area, Nevada. USDA-Soil Conservation Service.
U.S. Government Printing Office, Washington, DC. 194 pp.
135
-------
Van Johnson, W.M. and Maxwell, JA (1981). Rock and Mineral Analysis. 2nd. ed. J. Wiley and Sons,
Inc. New York. 489 pp.
Wentworth, O.K., Wilgus, W.L and Koch, H.L (1934). A Rotary Type of Sample Splitter. J. Sed. Pet.
4:127-138.
136
-------
Appendix E
Determination of Dust Constituents
at the Preparation Laboratory Facility
The following study was initiated to determine if creating a preparation facility in the Las
Vegas area would lead to sample contamination due to the prevalence of many atmospheric dust
constituents that are common to the region.
137
-------
Determination of Dust Constituents at the
MASS Preparation Laboratory in Las Vegas, Nevada
Introduction
A proposal was suggested to locate the soil preparation laboratory for the MASS in Las
Vegas, Nevada. Concern was expressed about alkaline dust contamination of acidic soil
samples from the Mid- Appalachian region. Soils in the Las Vegas area are alkaline and have
abundant gypsum (CaS042H20) which could be a source of Ca and SO4 contamination to soils
especially those with low analyte levels (Speck, 1985).
Objectives
To determine the amount and content of dust at the site of the proposed preparation
laboratory and if the amount present is a source of significant contamination for soil samples
prepared at the laboratory.
Materials and Methods
The proposed soil drying area of the warehouse was divided into two areas and dust was
collected to represent typifying dust that has accumulated through multiple weather events over
time. One bulk dust sample was taken from the north end and one from the south end of the
warehouse. Both locations were near doorways (Figure 1).
The samples were passed through a 140-mesh sieve (0.01 mm) to remove human
contamination (sawdust, metal shavings, etc.) and to remove large organic materials (leaves,
twigs, etc.). Extractable SO4 was analyzed using 1C and 0.016 M sodium phosphate
(NaH2PO4.H2O) as an extracting solution (Johnson and Todd, 1983). Total elemental Ca, Mg, K,
Na, Al, Fe, and Mn contents were analyzed on an ICP after using the bomb method of digestion
with hydrofluoric acid (Bernas, 1968). The pH was determined in 1:1 DDI:H20 using a
combination electrode.
The amount of dust accumulated over a three day period (the average sample drying time)
was also determined prior to dust control measures implemented during three separate weeks.
Sheets used to collect the dust were placed throughout the proposed soil drying area were
grouped into three regions, namely, the north, south and middle regions. Vertical fallout was
collected by laying a 22 sq.in. piece of paper on a flat surface to simulate the condition of an
uncovered sample during air-drying. Lateral fallout was collected using two 22 sq.in. sheets of
paper and a 14-3/4 sq.in. by 1-1/2 in. frame (approximate height of a bulk soil spread to dry).
One sheet was laid flat, the frame was placed in the center, and the second sheet was placed
on top of the frame. This procedure simulates the covering of a bulk soil sample for drying as
outlined in the preparation laboratory standard operation procedures (Bartling et al., 1988).
The sheets were placed on tables in the proposed drying area of the warehouse. After
three days the dust was collected by brushing it to the center of the paper and transferring it
into a pre-weighed container. The weight of the dust was determined by difference.
After dust control measures were implemented at the preparation laboratory, including the
installment of four ceiling fans, and enclosing the soil drying area, twelve (12) uncovered sheets
of paper (similiar in dimensions to those use in routine soil drying) were placed randomly on
tables throughout the drying area for thirteen (13) days in an attempt to collect a measurable
quantity of dust.
138
-------
Results and Discussion
Results of the accumulated dust collected over a three-day period for three separate weeks
are presented in Table 1. The greatest amount of dust was recovered during the first week at
the south end from lateral fallout. During the first sampling period, a major wind storm with
gusts exceeding 50 miles per hour occurred and the south warehouse delivery door was opened
for approximately one hour. No major wind events occurred during the sampling period for the
following two weeks and the only samples with a measurable amount of dust were sampled at
the middle location with an average total dust collection of 0.015g.
The pH of the dust was moderately alkaline, ranging from 7.83 to 8.26 (Table 2). Due to the
trace quantities of the dust collected, the effect of the dust addition on soil pH will be
negligible.
Three major assumptions were made for the proposed dust contamination problem to
determine if the dust, prior to dust control measures being implemented, could cause
measurable contamination to Mid- Appalachian soil samples. The assumptions were that: 1)
the only cation was calcium and the only anion was sulfate, 2) all dust present was gypsum
(CaSO4.2H20), and 3) all material was in the exchangeable form.
Table 1. Weight of Dust Recovered (Grams) During a Three-day Period
Flow
Lateral
Vertical
Week 1 (location)
So. End
0.10
0.01*
Mid
0.01
0.04
No. End
0.00
0.04
Week 2 (location)
So. End
0.00
0.00
Mid.
0.01
0.01
No. End
0.00
0.00
Week 3 (location)
So. End
0.00
0.00
Mid
0.01
0.03
No. End
0.00
0.00
•Indicates disturbance to the sample
Results from total calcium and extractable sulfate were used in the assessment of
contaminant additions because these two elements were found in greatest abundance,
excluding total Si (Table 3), and would thus be of the greatest concern. The following
calculations show that only 0.01785 mg S/Kg and 0.9727 mg Ca/L would be added to each
sample in the prep lab assuming at this point that the sample was entirely composed of Ca
and SO4. It should be noted at that the added SO4, on the basis of assumption 1, is less than
the contract required detection limit (CDRL) set for the Mid-Appalachain soil survey and
therefore dust contamination prior to dust control measure implementation poses no problem as
related to contamination of the soil sample. However, the results for Ca are above the CRDL of
0.05 mg/L so Ca is still a concern under the first assumption.
Table 2. Dust pH Values and Extract Solution Concentrations (mg/L)
Sample ID
South End
Rep 2
Rep 3
MEAN
North End
Rep 2
Rep 3
MEAN
PH
7.83
7T8T3
8.26
8.26
Ca
12.35
13.01
15.02
13.46
17.39
28.07
15.49
20.32
Mg
6.83
6.69
6.91
£.81
6.78
10.47
8.32
8.52
K
3.82
4.12
4.98
4.31
3.11
4.78
4.98
4.29
Na
3.40
3.40
3.70
3.50
2.50
2.80
2.50
2.60
Pe
3.98
4.20
3.87
4.02
2.73
4.02
3.12
3.29
Al
5.28
5.15
5.42
5.28
4.18
«,19
4.88
5.08
Mn
0.089
0.087
0.085
0.087
0.060
0.086
0.075
0.074
S04
852
788
767
802
766
571
722
686
The second assumption stated that all the dust contaminant was qypsum. Petrographic
evaluation of the dust revealed that quartz is the dominant mineral comprising 75-80% with
minor constituents of 15- 20% organic matter, and approximately 5% hornblende, biotite,
plagioclase feldspars, gypsum, and olivine. Silica was the major elemental component and
comprised about 88% of the samples supporting the petrographic mineral analyses (Table 3).
From the results and the data from the first assumption, the following calculation is possible:
139
-------
Less than 5% of dust is gypsum - 0.09727 mg Ca/L x 5% = 0.0048635 mg Ca/L
This concentration of Ca (0.048635 mg Ca/L) is less than the CRDL for Ca (0.05 mg/L) set
for the Mid-Appalachain soil survey. Therefore, after correction of assumption 2, the Ca
content added to the soil samples from dust contamination prior to dust control measure
implementation poses no detectable contamination threat to the MASS samples.
Table 3. Total Oxide Percentage in the Bulk Dust Sample
Sample ID
South End
2
3
Mean
North End
2
3
Mean
CaO
3.46
3.64
4.20
3.76
4.87
7.85
4.34
5.68
MgO
2.27
2.22
2.29
2.26
2.26
3.47
2.75
2.82
K2O
0.32
0.99
1.20
1.04
0.75
1.15
1.20
1.03
Na20
0.92
0.99
1.00
0.94
0.67
0.76
0.67
0.70
Fe203
1.12
1.18
1.09
1.13
0.77
1.13
0.88
0.92
A1203
2.00
1.95
2.04
2.00
1.59
2.34
1.85
1.93
MnO2
0.028
0.027
0.027
0.027
0.019
0.027
0.024
0.024
Si02*
89.28
89.07
88.15
88.84
89.07
83.27
88.29
86.90
•Calculated by difference
Calculations
The average SO4 concentration in the dust = 744mg/L (Table 2)
[4g soil and 100ml extract]
744mg/L * 100ml/4g * 1L/1000ml * 1000g/Kg = 18,600mg SOJKg soil
= 18.60g SO4/Kg soil
- 0.0186g SOJg soil
0.0002678g SO4 per 0.0144g average collected dust (Table 1)
If a bulk sample = 5kg, then 0.002678g/5000.0144g
which is 0.05355mg SO«/Kg soil
To convert mg SCVKg -> mg S/Kg
.05355 x .33333 = .01785mg S/Kg
5.355 x IQ-'g SOjQ soil
The average total elemental Ca concentration = 16.89mg/L (Table 2)
[O.OSg soil and 100ml extract]
16.89mg/L * 100ml/0.05g * 1L/1000ml * 1000g/Kg = 33780mg Ca/Kg soil
= 33.78g Ca/Kg soil
= 0.03378g Ca/Kg soil
0.0004864g Ca per 0.0144g average collected dust (Table 1)
If a bulk sample - 5kg, then 0.0004864g/5000.0144g = 9.727 x
= 0.09727 ppm total Ca > CRDL of .05 ppm exchangeable Ca
Ca/g soil
The third assumption made during the analyses was that 100% of the total Ca would
dissolve in water to become exchangeable Ca. According to Brady (1974), only 25% of the total
Ca in soils is in the exchangeable form. Using this percentage reduces the Ca content in the
dust to 0.00122 mg Ca/L which is even further below the CRDL, supporting the conclusion that
the dust contamination potential is well below detectable limits and should not be considered a
problem at the prep lab.
Results presented represented the situation at the warehouse before dust control measures
140
-------
were applied. The following major dust control measures reduced the possibility of
contamination: (1) constructing a free-standing enclosure within the warehouse, (2) sealing off
the south delivery door, (3) coating walls and floors to prevent the addition of concrete dust,
and (4) sieving and homogenizing in an air-controlled environment. After these dust control
measures were implemented there was no measurable amount of dust collected from the
twelve sheets of uncovered paper in the soil drying area after thirteen (13) days.
Conclusions
The results show that the amount of dust recovered was negligible and that SO4 and Ca
addition were well below the contract-required detection limits of the instruments to be used in
the survey prior to control measures being enacted. After dust control measures were
implemented no dust was found collected after a significant length of time, greater than three
days (normal soil drying time); thereby, indicating a lack of dust for any form form of
contamination of the soil samples. Therefore, there will be no significant addition of any
targeted element to the bulk soil sample from dust in the preparation laboratory.
141
-------
References
Bernas, B. 1968. A New Method for Decomposition and
Comprehensive Analysis of Silicates by Atomic Absorption
Spectrometry. Anal. Cham. 40:1682-1686.
Brady, N. C. 1974. The Nature and Properties of soils. Macmillian
Publishing Co., Inc.
Johnson, D. W.. and D. E. Todd. 1983. Relationship Among
Fe, Al, C, and Sulfate in a Variety of Forest Soils.
Soil Sci. Soc. Am J. 47:792-800.
Speck, R. L 1985. Soil Survey of Las Vegas Valley Area, Part of
Clark County, Nevada. USDA-SCS. U.S. Government Printing
Office Washington, D.C.
142
-------
Appendix F
Kraft Paper and Nylon Mesh Analysis
The following appendix contains: (1) a table of data resulting from the analysis of the "PAL"
samples at the MASS analytical laboratories, and (2) an internal report that describes the analysis
of the kraft paper and nylon mesh used for air drying at the preparation laboratory. The table and
report address some data user concerns that there might possibly be contamination leached into
the MASS soil samples from these two materials.
143
-------
Table F-1. Comparison of LAL and PAL Sample Analytical Data from Three Soil Batches
Parameter
SAND
VCOS
COS
MS
FS
VFS
SILT
COSI
FSI
CLAY
PH H20
PH~002M
PH~01M
CA CL
MG~CL
K CL
NA_CL
CA OAC
MG OAC
K OAC
NA_OAC
CEC CL
CEC OAC
AC_BACL
CA CL2
MG CL2
K CL2
NA CL2
FE CL2
AL_CL2
FE PYP
AL~PYP
FE AO
AL AO
FE~CD
AL~CD
SO4 H20
S04~P04
SO4~~0
S04~2
SO4~~4
S04~8
S04 16
S04J32
C TOT
N~TOT
S~TOT
Reporting
units
wt %
II
II
n
it
n
li
II
11
n
pH units
II
II
meq/100g
II
n
n
meq/100g
H
II
"
meq/100g
11
n
meq/100g
"
n
n
11
11
wt %
«
n
"
n
ii
mg S/kg
II
mg S/L
n
n
n
11
"
wt %
II
n
30128
LAL
96.600
6.000
16.200
32.700
31.700
10.000
3.300
3.100
0.200
0.100
5.320
4.850
4.770
0.089
0.031
0.026
0.000
0.115
0.044
0.028
0.011
0.609
1.122
0.210
0.740
0.021
0.015
0.007
0.001
0.005
0.044
0.081
0.052
0.083
0.137
0.052
2.803
3.403
0.480
2.624
4.409
8.492
16.496
32.974
0.089
0.003
0.002
PAL
96.300
4.600
16.200
32.200
33.700
9.600
3.500
3.300
0.200
0.200
5.320
4.870
4.790
0.142
0.039
0.030
0.000
0.136
0.042
0.031
0.015
0.609
1.065
0.889
0.750
0.016
0.013
0.009"
0.000
0.004
0.035
0.080
0.041
0.098
0.210
0.055
2.603
3.003
0.641
2.572
4.438
8.586
16.560
32.479
0.076"
0.004
0.002
30129
LAL
96.900
7.500
15.500
34.100
32.600
7.200
3.100
0.900
2.200*
0.000
5.210
4.980
4.720
0.084
0.032
0.027
0.006
0.054
0.023
0.025
0.006
0.497
0.936
0.201
0.775
0.020
0.006"
0.004
0.001
0.006
0.036
0.056
0.200*
0.092
0.159
0.060
1.802
6.006
0.458
2.458
4.400
8.160
15.700
31.160
0.112
0.006
0.000
PAL
94.700
4.700
13.900
29.100
34.900
12.200
5.500
5.200
0.300
-0.200
5.240
4.950
4.730
0.064
0.025
0.026
0.009
0.073
0.027
0.029
0.010
0.463
0.958
1.126
0.756
0.022
0.014
0.007
0.001
0.005
0.030
0.047
0.057
0.063
0.143
0.050
1.602
5.005
0.431
2.353
4.225
8.080
15.880
31.240
0.094
0.004
0.000
30130
LAL
97.200
4.300
18.00
32.400
33.400
9.000
1.800
1.700
0.100
1.000*
5.460
5.150
4.800
0.057
0.019
0.025
0.006
0.049
0.025
0.020
0.005
0.741
0.741
0.405
0.761
0.028*
0.023*
0.009*
0.002
0.002
0.032
0.035
0.051
0.067
0.116
0.033
2.600
4.800
0.466
2.285
4.186
8.183
16.223
33.771
0.088
0.004
0.001
PAL
97.500
9.700
19.400
30.800
30.000
7.700
1.500
1.300
0.200
1.000*
5.370
5.150
4.750
0.064
0.020
0.027
0.011
0.061
0.022
0.024
0.009
0.741
0.836
0.486
0.410
0.014
0.011
0.024*
0.001
0.000
0.034
0.049
0.058
0.078
0.100
0.059
16.800*
4.000
1.980*
2.704
4.320
8.415
17.058
34.658
0.176
0.005
0.004
NOTE: No accuracy windows were established for MOIST, AL_CL, and SI_AO.
" Value below lower accuracy window limit.
* Value above upper accuracy window limit.
144
-------
Internal Report
DDRP Kraft Paper and Mesh Analysis
by
L.K. Fenstermaker
Environmental Research Center
University of Nevada-Las Vegas
Cooperative Agreement No. CR814701-01
Analysis performed by
J. Lindner, V. Letourneau, P. Nowinski,
S.H. Pia, DA Jadhon, B.A. Schumacher, and
R.D. Van Remortel
Lockheed Engineering and Sciences Company
Las Vegas, Nevada
Contract No. 68-03-3249
Project Officer
Louis J. Blume
Exposure Assessment Research Division
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada 89193-3478
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89193-3478
145
-------
NOTICE
This document is intended for internal Agency use only. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
146
-------
INTRODUCTION
The work described in this report was initiated as a result of peer comments on the review
of the EMSL-LV activities for the mid-Appalachian Direct Delayed Response Project (DDRP). The
peer review panel was concerned about the potential for contamination of the mid-Appalachian soil
samples during the drying process from kraft paper upon which the soil is laid to dry. External
auditors requested that the mesh table upon which the kraft paper was laid be analyzed as well.
The Lockheed Engineering and Sciences Company (LESC) preparation laboratory is responsible for
this procedure. LESC set up a special room in a warehouse for the drying process. The room has
limited access and is vented, and procedures had been established prior to operation to prevent
contamination from dust or other samples. Drying tables were constructed from nylon mesh and
polyvinyl chloride (PVC) pipe. At a minimum two sheets of kraft paper (approximately 1 x 1 m) were
placed on top of the mesh for each sample. If a sample contained free-standing water, then a
sheet of clean plastic was placed on top of the paper and the plastic sample bag with the sample
inside was cut open on top of the plastic sheet. All other samples were placed directly on the
paper for drying. In all cases, samples were covered with a sheet of paper to prevent any
contamination by dust.
The concern expressed by the peer review panel and external auditors was that, during the
course of drying, there might potentially be movement of ions from the paper and mesh into the
soil sample. To assess this potential, LESC was asked to analyze the mesh and paper for ions
for which the soil samples would later be analyzed, namely, SO4, Al, Ca, Mg, K, and Na.
The analytical work described in this report was performed by Lockheed Engineering and
Sciences Company for the EPA Environmental Monitoring Systems Laboratory (EMSL-LV) under
contract number 68-03-3249, and the report was written by the Environmental Research Center,
University of Nevada-Las Vegas, for EMSL-LV under cooperative agreement number CR814701-01.
METHODS
Samples of the nylon mesh and paper used for the drying process were prepared for
analysis in the following manner. Twenty 4 inch by 4 inch pieces of mesh were cut from the roll
used to make the drying table surface. Nylon mesh was used for the table surface to promote
rapid drying of the soil samples. The mesh had a border along the edge. While this border would
not be located directly beneath the drying soil samples, it was decided to analyze mesh samples
with borders to fully examine all possible situations. Therefore, half of the twenty mesh samples
had borders while the other half did not. The typical area of mesh utilized for drying a soil sample
was approximately 1 x 1 meter (1550 square inches), equivalent to 100 of the 4" by 4" samples. The
twenty samples were placed in 500 ml bottles and brought to volume with double deionized water
(DDI or Type II reagent grade water). The water and mesh remained in the bottle for three days
before separation and analysis.
Three kraft paper samples, each approximately 1 x 1 meter and weighing 100g on the
average, were placed in one-liter cylinders and one liter of DDI water was added. After a 24 hour
soaking period the paper and water were separated to prevent dissolution and physical breakdown
of the paper in the water.
The methods of analysis for both the paper and mesh leachates were ion chromatography
for S04, ICP for Al, Ca, and Mg, and flame atomic absorption spectrometry for K and Na. For the
paper leachates, analyses were run in triplicate except for K and Na which were run in duplicate.
Quality assurance samples consisted of blanks, drift blanks, standard reference samples, and
duplicates.
147
-------
RESULTS AND CONCLUSIONS
Tables 1 and 2 below present the analytical results. To examine a worst-case scenario, it
was assumed that all ions present in the mesh and paper leachate would go directly into the 5 kg
of drying soil. The maximum possible amount of ion addition to the soil from the paper and mesh
are presented in Table 3 in milliequivalents per 100 grams (meq/100g). Further calculations were
made to show how much of an increase could be expected during routine soil analysis (Table 3)
and whether this increase would be detectable, i.e., above the program defined detection limits of
.025 mg/kg for SO^ .1 mg/l for Al, and .05 mgfl for Ca, Mg, K, and Na. Only the highest
concentrations measured, presented in boldface print in Tables 1 and 2, were used for the
calculations after blank correction. The purpose of Table 3 was to provide two different means of
presenting and understanding what the maximum addition of ions from the paper and mesh to the
soil would be in common reporting units. Appendix 1 provides an example of how the calculations
were performed.
Any potential contamination from the mesh would be detectable for all ions except Al (Table
3). Even though this potential exists, movement of ions from mesh to soil is severely restricted due
to the absence of water transport from the mesh to the soil. The only mechanism which would
provide transport of ions into the soil would be a diffusion process between the paper and soil
during equilibrium after initial wetting of the paper by soil water. Convective movement (or water
transport) of ions is not a possible transport mechanism as there was not a continuous source
of water. LESC personnel very meticulously monitored the conditions of the drying soil and the
paper and mesh under the soil. They never noticed any staining of the bottom layer of paper by
the mesh or any signs of water movement onto the bottom sheet of paper from the soil.
Therefore, diffusion of ions from the mesh to the paper could not occur and contamination from
the mesh to the soil sample is not possible.
The potential for detectable contamination from the paper could occur for SO and Na if a
complete diffusion of the ions from the paper to the soil occurs. While this is not likely, this study
does not provide the necessary information to fully access this. If it is determined that the levels
reported in Table 3 for the extreme case are not acceptable, a follow-up study should be performed.
As suggested previously, the best approach to determine how much the paper is contaminating the
soil is as follows: moisten low-level audit soils to field moisture and allow them to dry on the
paper in the same manner as the mid-Appalachian soil samples; analyze the samples according
to the DORP Methods Manual; and examine the results for any increases that could be due to
contamination from the Kraft paper. In this manner the possibility of potential contamination can
be accurately assessed under normal routine sample conditions. Given that saturated soils were
either dried on plastic sheets or were not allowed extended contact with the paper, the paper
contamination issue should not be of concern for the mid-Appalachian DDRP soil survey. However,
extended contact between paper and soil may have occurred in the previous Northeast and
Southern Blue Ridge survey and therefore justify further research.
148
-------
TABLE 1. ANALYTE CONCENTRATIONS (MG/L) FOR NYLON MESH
Sample #
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
AVG
STD
Blank- 1
Blank-2
Blank-3
Blank-4
Blank-5
AVG
STD
SO4
0.22
0.47
0.52
0.30
0.24
0.28
0.24
0.21
0.23
0.22
0.20
0.19
0.26
0.35
0.20
0.31
0.32
0.18
0.28
0.26
0.24
0.13
<.01
<.01
0.01
<.01
<.01
__
—
A)
0.041
0.036
0.05
0.04
0.041
0.042
0.045
0.051
0.037
0.04
0.044
0.047
0.034
0.05
0.044
0.047
0.038
0.042
0.036
0.038
0.042
0.01
0.036
0.043
0.047
0.049
0.044
0.040
0.01
Ca
1.07
0.54
0.6
0.41
0.52
0.51
0.6
0.75
0.8
0.78
0.76
0.59
0.58
0.48
0.64
0.36
0.82
0.75
0.81
0.68
0.65
0.17
<.01
<.01
<.01
<.01
<.01
.-
—
Mg
0.088
0.047
0.058
0.042
0.052
0.049
0.049
0.066
0.071
0.067
0.066
0.051
0.049
0.042
0.051
0.037
0.1
0.065
0.067
0.058
0.059
0.02
<.01
<.01
<.01
<.01
<.01
__
—
K
0.172
0.018
0.031
0.034
0.021
0.021
0.023
0.021
0.09
0.032
0.021
0.023
0.026
0.023
0.031
0.013
0.194
0.042
0.058
0.019
0.046
0.05
<.01
<.01
<.01
<.01
<.01
„
—
Na
0.17
0.036
0.056
0.058
0.061
0.05
0.054
0.092
0.078
0.049
0.049
0.04
0.06
0.054
0.059
0.045
0.266
0.075
0.104
0.038
0.075
0.05
<.01
<.01
<.01
<.01
<.01
__
—
149
-------
TABLE 2.ANALYTE CONCENTRATIONS (MG/L) FOR DRYING PAPER
Sample #
Blank
1
2
3
S04
<.13
40.602
35.945
39.024
Al
0.016
0.090
0.091
0.079
Ca
0.008
2.811
3.053
2.616
Mg
0.001
0.298
0.304
0.261
K
<.10
1.34
1.45
1.35
Na
<.30
28.80
28.25
28.00
Al, Ca, Mg, and SO4 are means of triplicate measurements. Na and K are means of
duplicate measurements.
TABLE 3. ADDITIONS OF ANALYTE TO SOIL (MEQ/100G) AND TO EXTRACTION
Analyte
SO4
Al
Ca
Mg
K
Na
Mesh:
Addition to
soil (meq/100g)
0.0103
0.0010
0.0509
0.0078
0.0047
0.0070
Mesh:
Increase in
soil extraction
1.65 mg/kg
0.0079 mg/l
0.8476 mg/l
0.0792 mg/I
0.1537 mg/l
0.1347 mg/l
Paper:
Addition to
soil (meq/100g)
0.0169
0.0002
0.0030
0.0005
0.0007
0.0250
Paper:
Increase in
soil extraction
2.71 mg/kg
0.0012 mg/l
0.0509 mg/l
0.0050 mg/l
0.0242 mg/l
0.4800 mg/l
150
-------
APPENDIX 1
Since all analytical measurements were reported for the volume of mesh or volume and
weight Kraft paper only it is necessary to do some calculations to convert the data into numbers
which clearly indicate the maximum analyte addition to the soil could be expected if contamination
occurred. This appendix provides an example of the calculations which were performed to generate
Table 3. The example calculation will use the result from the calcium analysis.
Knowns:
* 4 x 4 inches of mesh were leached in 500 ml DDI water and the measurement
results were reported in mg/l;
* 1 x 1 meter of Kraft paper were leached in 1 liter DDI water and the measurement
results were reported in mg/l;
* the average area of paper and mesh used for soil drying was 1 x 1 meter;
* the average amount of soil per sample for the mid-Appalachian survey was 5 kg;
* and for all analyses done in this study, except SO4, 5 grams of soil are extracted in
approximated 60 ml of extract (SO4 analysis requires 4 grams of soil and 80 ml
extract).
Assumptions:
* to examine the maximum amount of contamination possible it was assumed that all
of the ions from the paper and mesh would go into the soil sample.
The first step is to figure out how much analyte would come from a 1 x 1 meter area of
mesh from the results for the 4x4 inch mesh sample. The maximum amount of calcium measured
in a mesh sample was 1.07 mg/l. Since 4x4 inches was extracted in 500 ml, two times 4x4
would be extracted in one liter and therefore 1.07 mg/l Ca is equivalent to 1.07 mg Ca per 2(4 x 4)
inches.
mg/l Ca = mg Ca/2(4x4); total mesh = 1x1 m = 1521 inch2
1.07 mo/1 Ca = x ma/I Ca = 50.858 Ca mg/l (total mesh)
32 inch2 1521 inch2
Assuming all 50.858 mg Ca goes into the 5 kg soil sample, the increase in Ca to the soil
would be:
20.04 mg Ca per milliequivalent
50.858 mg Ca per 5 kg soil = 2.538 meq Ca per 5 kg soil
20.04 mg/meq
2.538 meq Ca = x meq = .0508 meq Ca
5 kg soil 100 g 100 g soil
To calculate the maximum increase to the analytical result that could be expected if 5 g of
the 5 kg soil sample were extracted in 60 ml NH4OAc or NH4CI and all of the Ca in the 5 g sample
originally from the mesh was extracted the following were performed. (This also points out whether
the increase from the mesh would be above the CRDL's.)
50.858 mg Ca = x mg = .0509 mg Ca
5 kg soil 5 g soil 5 g soil (or 60 ml)
.0509 mq Ca = x mg = .8476 Ca mg/l
60 ml 1000 ml
151
-------
Appendix G
Technical Systems Audit Reports
The following appendix consists of reports that serve to document the technical systems
audits conducted at the Mid-Appalachian Soil Survey preparation laboratory.
152
-------
PRE-OPERATION AUDIT REPORT
for
DDRP SOIL PREPARATION LABORATORY
MID-APPALACHIAN SOIL SURVEY
Environmental Research Laboratory.
Corvallis, Oregon
September 1, 1988
153
-------
PRE-OPERATION AUDIT REPORT
for
DDRP SOIL PREPARATION LABORATORY
MID-APPALACHIAN SOIL SURVEY
INTRODUCTION
A pre-operation technical systems audit was conducted of the preparation laboratory (prep
lab) facilities for the Mid-Appalachian soil survey on 30 August 1988. The prep lab is located at
4675 Valley View Drive, Las Vegas, Nevada and is operated by staff from Lockheed Engineering and
Management Services, technical support for the U.S. Environmental Protection Agency (EPA),
Environmental Monitoring Systems Laboratory, Las Vegas, NV (EMSL-LV).
Personnel Contacted:
Louis Blume EPA Project Officer
Michael Papp Quality Assurance (QA) Representative
Rick Van Remortel Prep Lab Manager
Dave Jadhon Assistant Lab Manager
Rob Tidwell Shipping/Receiving Manager
Daron Peres Lab Tech (part-time)
Additional prep lab staff will include:
Glenn Merritt Lab Tech
Stanton Offemson Lab Tech (part-time)
Kim Cutler Lab Tech
Joe Parolini Lab Tech
Audit Team:
Deborah Coffey ERL-C QA Specialist
Jeff Kern ERL-C QA Coordinator (Deputy)
Documents received during audit visit:
o Revised "Mid-Appalachian Direct/Delayed Response Project Preparation Laboratory
Standard Operating Procedures" including Appendix B of this document, "Preparation
Laboratory Computer Tracking/Verification Procedures" as a separate document. The
computerized sample tracking system was also demonstrated.
o Floor plan of the prep lab facility.
The cold storage facility used for archiving samples from this and previous soil surveys was
also visited. Las Vegas Ice and Cold Storage Co. (formerly Mr. Ice) is located on Master Road in
(north) Las Vegas.
The project is covered by a health and safety plan which has been submitted to project staff
for review.
Overall, the prep lab is organized and most operations are in place. The only major problem
is the lack of a cooler for sample storage upon receipt. New equipment has been purchased, data
sheets have been generated, standard operating procedures (SOPs) have been posted, lab staff
have been identified and the lab is nearly operational. Once lab staff come on board they will
require training. Provisions for training have been considered. Staff already working on the project
were aware of QA concerns and requirements and seemed capable of successfully completing the
154
-------
project. The sample drying area should not experience problems with dust contamination. The
design of the sample drying tables is innovative and is expected to facilitate sample drying. Prep
lab staff should be commended for their effort in preparing the prep lab to receive samples. An
audit is to be conducted after the prep lab is operational by the EMSL-LV QA Officer, Gene Easterly.
The audit report from the second review should be forwarded to ERL-C QA staff, the ERL-C
technical director, and the ERL-C sampling task leader. Based on the findings of this audit, it is
expected that no major shortcomings in prep lab operations will be identified.
SUMMARY COMMENTS AND RECOMMENDATIONS
o It is recognized by all DDRP project staff that it is a priority that the cold storage
cooler, to be built in the prep lab, is operational by the time samples begin arriving
from the field. The prep lab manager is aware of this problem and will be working
to resolve it to the best of his ability.
o Tests of the Weathershade" material used for the drying tables should be made to
ensure that the material is inert and does not leach contaminants. Two tests were
suggested; (1) Place a wet sheet of white paper on the table to check for visible
leaching of contaminants from the Weathershade", (2) Leach the Weathershade" in
a beaker and analyze leachate for the presence of at least cations, metals, and
sulfate (ideally all chemical parameters of interest to DDRP).
o Thermometers should be posted in the sample drying area to monitor the
temperature. Place thermometers at the level of the drying tables in at least two
locations. Document temperatures at least daily, and more frequently if necessary.
A thermograph (24-hour or weekly) would also be appropriate for documenting
temperature in the drying area.
o Initiate the use of calibration notebooks for all balances used in the project. The
notebook should document all calibrations performed in a format similar to that
suggested below:
Date Expected wt. Observed wt. Acceptable? Initials
o A trap needs to be added to the compressed air system. A filter is anticipated to
be added to this system which should eliminate condensation and oil contamination.
Alternatively a flask with cotton or absorbent will function as well.
o The decision not to seal air vents on the roof should be discussed with ERL-C staff
before a final decision not to do this is made.
o The final version of the quality assurance project plan (QAPP) (which includes prep
lab activities) should include a signed (by EPA project officer, EMSL-LV QA officer,
EPA Technical Director) signature approval page.
* A telephone logbook should be initiated by the lab manager to record all telephone
calls concerning sample receipt, labelling and shipping. (This was not discussed
during the audit and so, is identified here.)
o The samples placed in the archive area should be well-labelled for easy retrieval.
Boxes should list batch numbers or other codes that can be used to trace samples.
Information needed to trace samples and retrieve them from cold storage should
be available in written form, on file at the prep lab (or in a sample custodian's
office following completion of prep lab activities) and somewhere at the cold storage
facility. Ideally, a master list of all samples, the batch in which each sample is
packed, and in which survey the sample was collected should be prepared.
155
-------
It is the responsibility of ERL-C staff to determine how long samples will be
archived. Sample storage currently costs about $5,000/year. Sample disposal will
also require funds. If samples are to be accessible to other scientists, funds will
need to be available to retrieve, package and ship samples upon request.
SUMMARY OF AREAS REVIEWED
Laboratory facilities reviewed included:
Trailer #1 - pH, clod bulk density analyses
Trailer #2 - Loss on ignition, air dry determination
Trailer #3 - Sample preparation area (riffle splitter) and subsample preparation area
Sample drying
Sample disaggregation and sieving (rock fragment determination)
Front office
Sample receipt/ log-in
Sample shipping
Archive - Cold storage
Computerized Sample tracking/laboratory status program
Soil description computer entry and data verification system
156
-------
Ti
I UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
f OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY-LAS VEGAS
P.O. BOX 93478
LAS VEGAS. NEVADA 89193-3478
(7O2/798-21OO- FTS 545-21OO)
SUBJECT: Preparation Laboratory Audit Report in Support of the Mid-Appalachian
Direct/Delayed Response Project of October 6, 1988
FROM: / Da*v
•-T^QA Officer
TO: Files
The purpose of this report is to document the major issues that arose as a
result of an on-site audit of the EMSL-LV soil preparation facility. The audit
was conducted approximately two weeks after initiation of sample analyses. The
audit was unannounced and conducted by; the Laboratory QA Officer, Mr. Gene
Easterly; the Nuclear Radiation Division Science Advisor and QA Officer, Dr.
Stuart Black; the DDRP Field Sampling QA representative, Jeff Kern; and, the
DDRP EMSL-LV Project Officer, Louis Blume. Attachment 1 of this report contains
an audit questionnaire which documents the specific details of the audit.
The preparation laboratory operations were well organized and managed
by an enthusiastic and qualified staff. We thought that operations were
running very smoothly especially considering this was only the second week of
analyses. There were a few areas of concerns identified, many of which were
already in the process of being rectified.
After the laboratory tour the laboratory managers were debriefed regarding
these issues and recommendations. Areas of concern and recommendations were as
follows:
a. Subsampling for various analyses should be on a composite basis
after samples are mixed instead of taking the complete subsample from one
location.
b. Corrections or cross-outs in logbooks should not be obliterated.
They should be crossed out with one line, then initialed. All entries should
be made in ink. In addition, we recommended the laboratory manager should
review, sign, and date each logbook on a weekly basis.
c. Samples received that are very wet should be laid out to dry along
with the interior plastic bag inverted or opened up. After the fine material
adhering to the bag is dried, it should be added to the sample.
157
-------
d. It is acknowledged that the method of breaking up aggregates and
passing soil through the 2 mm sieve is somewhat operationally defined;
nevertheless, we asked the technicians to consider some way of standardizing
this procedure.
e. Tolerance limits for blank drift of analytical balances should be
established and controlled for precise weighings. In addition, servicing
of the balances and scheduled preventive maintenance of the balances and pH
meters should be conducted and assessed.
f. Some method of carrying flags or comments about anomalous handling
of samples during sampling, shipment, and preparation should be carried forward
through data verification.
g. Sample receipt information should be kept in a logbook, or if loose-
leaf forms are kept, they should be sequentially numbered and be permanently
bound.
h. Buffer solutions and QCCS solution should be dated upon opening.
i. Ventilation to the grinding area needs Improvement primarily from the
standpoint of technician safety.
j. Routine samples should not be placed underneath ceiling fans if It
can be avoided. Although there Is little probability of contamination due to
the Kraft paper covering the drying soils.
k. Organic soils should be air-dried as soon as possible considering
there is a stronger potential for microbial oxidation of sulfur relative to
mineral soils.
1. A minor study should be undertaken to determine if sulfur can be
leached from the Kraft paper onto the soil considering Kraft paper is made
through a sulfite process.
The bulk density analysis was not observed In its entirety because there
were no ongoing analyses. Therefore, no comments were made, and this should
be observed In a future audit. A final report was requested on the dust study
and on whatever action was taken to follow-up on a possible contamination from
the plastic mesh that holds up the soil and Kraft paper. In summary the review
team found no major problem areas and were quite impressed with the operations.
Attachment
158
-------
PREPARATION LABORATORY ON-SITE
EVALUATION QUESTIONNAIRE
Date: 10/5/88
Laboratory Manager: Rick Van Remortel
Quality Assurance Auditor: Gene Easterly
Name
Personnel Present: R. Van Remortel
D. Jadhon
Evaluation Team:
R.
G.
D.
G.
J.
S.
G.
S.
L.
J.
B.
J.
Kohorst
Satterwhite
Peres
Merri tt
Porolini
Offemson
Name
Easterly
Black
Blume
Kern
Schumacher
Burton
Title
M.S. Soil , 10 yrs. exp. as
Soil Scientist
B.S. Geology, 4 yrs. exp.
Oil Exploration
B.S. Biology, 4 yrs. exp.
Med Specialist
B.S. Aquat Biology, 8 yrs. exp.
QA and Biologist
1 yr college, 5 yrs. exp. as
Warehouse Manager
M.S. Fisheries Biology, 4 yrs.
Aquatic Field Studies
M.S. Gepscience, 4 yrs. as
Geohydrologist
B.S. Engineering, 2 yrs. exp.
Soil Scientist
Title
Laboratory QA Officer
Nuclear Radiation Division
Science Advisor, QA Officer
DDRP EMSL-LV Project Officer
Soil Scientist
DDRP Field Sampling QA
Representative
Soil Scientist
Preparation Laboratory QA
Liason
159
-------
EQUIPMENT AND SUPPLY
Are the following equipments and supplies available and are there sufficient
quantities?
Items
Yes No
Comments
Logbooks? X
— — —— — — — — — — — •.»»••^«•— — •«— — — »—•« — »»—••••— — •" — «•»"••• — — — — — « — — — — —•-"«••— — — «• — — — «••-• — •— — — * — — ™» — — •
Logsheets? X Raw data forms
Binders? X
_ — — — — — -••• « V» — — •B-B — — — — — a.^,^™ _ — — — ••••••••»»•••» — »» — »•••• — •••—»•* — — •••««™™ — —• W«_K«H«««
Munsell color book? X
Pens-permanent black ink? X
Kraft paper - 36" wide? X
Plastic gloves? X
— —_ — w— — — — — «.»»„•• — «P — — — •—*•••••**•«•»•• — •• — — — ^ — — — — — — —"• — ""•••« — •» — »-• — — •"—— — — — — •• — — ••— — •••
Furnance gloves? X
Dust Masks? X More on order
Gallon paint cans? X
Aluminum weighing tins? X
Convection oven? X
Lab coats? X
Nalgene bottles - 500 ml, X (and 2-L)
100 ml? and 2 liter
Carboys - 13 gallons? X
Crucibles? X
Evaporating dishes? X
Beaker - 1000 ml? X
Tongs? X
Ring stand? X
Thermometers 0-100°C? X Also NBS
160
-------
EQUIPMENT AND SUPPLY (CONTINUED)
Items
Yes No
Comments
Drying trays - constructed
of PVC and nylon mesh? X
Rolling pins? X
Rubber stopper - no. 10 size? X
Sieves, 2-mrn and 4.75-mm
mesh? X
Riffle-splitter, Jones-type
with 1.25cm openings? X 2
Scale or top-loading
balance? X 2
Analytical balance? X 2
pH meters? X 2
Desiccators? X 1 large, custom made
Desiccant? X 2 jars
Hot plate? X
Drying oven? X Same as convection oven
Muffle Furnance? X 2
S Standard weights? X 1, 500g
Shipping boxes? X
Packing material? X
Strapping tape? X
Mops, buckets, brooms? X
Are there copies of the
laboratory's equipment
inventory lists available? X
161
-------
GENERAL ASSESSMENT OF THE PREPARATION LABORATORY
Questions
Yes No
Comments
Is a personal computer X
available for data entry?
Are the logsheets organized X Black 3-ring binders
in loose-leaf binders?
Are there hoods and exhaust X For bulk density,
fans where needed and is there LOI, and riffle
adequate workspace within hoods? splitting
Is there a source of compressed X Air compressor w/
air? in-line filter
Is there a source of deionized X Carboys delivered from
water? EMSL-Lab
_ _ _ — W^ — »• _—,— -...•.— ——— — ».«—»•——•»— — — — — — ™— • — -» _« • M • •• V •• «*• W _*—*_•»_ _ _ • «».•..»_•..»«•»• ••—— — —— —
Is there an area designated for X Warehouse area
storing sampling equipment?
Is there a systematic way of X
disposing of all waste
i.e., chemicals, paper, etc.?
Are hazardous areas clearly X No specific hazard
posted and is there a fire areas, but fire extin-
extinguisher and is it guishers are available
accessible? in every trailer
Are chemicals and other X
combustible materials
stored in a safe area?
— — — — ^.^.^ — — — —_— — _„.«. — — — — — — •••.•• — •• — •»•• — ««™™»M»™— •— ^— *• — <•• — •—»•— ••—••-•*«— •»—»«••«•«— — •• — .•••*«»—• — ••••*•*—»•
Are all instruments X Logbooks for all
calibrated and standardize improvements in pre-
and in proper working condition? ventive maintenance
is needed
Is the balance checked with a X Daily calibration
class S standard weight before
each use, and is the result
recorded in a logbook?
Is the analytical balance located X Observed fluctuation on
away from draft and areas subject mettle and sartorious
to rapid temperature changes? control limits for drift
should be established
162
-------
STANDARD OPERATING PROCEDURES
Questions Yes No Comments
Does the laboratory have a standard
operating procedures (SOP) manual and
is it available to each analyst/
technician?
Is the SOP manual followed in
detail?
Does the SOP manual contain
quality control practices?
Does the SOP manual deviate from the
procedures required by this project?
X
X
X
If the SOP manual does deviate, are
the deviations documented in written
form?
163
-------
ORGANIZATION AND PERSONNEL
Questions
Yes No
Comments
Is the laboratory clean and well X
organized?
Do personnel assigned to this project X
have the appropriate educational
background to successfully accomplish
the objectives of the program?
Do personnel assigned to this project X
have the appropriate level and type of
experience to successfully accomplish
the objectives of this program?
Is there adequate staffing to meet X 6 Technicians
the commitments of the project in 6-day work weeks
a timely manner? Alternating shifts
Is there training provided for X
preparation laboratory personnel?
•••^•••*» ••••••••••••••^•• — ^••••.» •••••••••••••••••••••••.•-•.••.•^•aii,^.„„„^ — —,_^__ __ _ -r»— -^ —i
Are the responsibilities of each X
technician fully outlined?
Does the laboratory manager have X
his/her own copy of the Field
Sampling Manual and the Standard
Operating Procedure?
Before filling out Form 102, does X
the laboratory manager:
Review data values on Form 101?
— —--__ -_.._____—__.
Review raw data in lab notebooks? X Ask to review, sign,
and date on weekly
basis
Check for adequate and accurate X
ID of QC samples?
Does the laboratory manager have X 101 is computer-
Forms 101 and 102 on file? generated
164
-------
SOIL PREPARATION PROCESS
Questions
Yes No
Comments
Is cross-contamination between X Good
samples in the drying area avoided?
Are there separate workspaces for X
sample drying and for sample
preparation?
Is there adequate workspace for X Area needs
drying, sieving and crushing of ventilation
samples?
Is there adequate workspace for X
performing soil analyses?
Are riffle splitters and sieves X
cleaned between samples?
^ml^^^tff^^mm^m,.»»,• ••.•••••i-Ti-^r-i— — — — — — ••• — ••.•••••••••••••••••••••••••••••••••••••••••••••••••••••'•••••••••I*
Is drying area removed from X
reagent storage?
Is the moisture-content sample X
removed In an appropriate manner?
fm j, — — • -m-r nr— ••«•!• m-n-^ — — _ — __ — — — — •• — — — — •• — •• — — •• — — ••«•••••••••••••••• — •••••••••••••••••••••
Is the moisture-content sample X
returned to the bulk sample?
Is the field moist pH sample removed X
in an appropriate manner?
Are rock fragments archived properly X
after the sieving process?
Are adequate measures taken to X One routine sample was
protect samples from indigenous stored directly below
dust contamination? ceiling fan
165
-------
COLD STORAGE FACILITY
Questions
Yes No
Comments
Is the temperature of the cold storage X Recorded on the time
facility recorded daily in a logbook? wheel and in logbook
Is there adequate space for the X
storage of samples?
Is the unit set-up for easy access to X
samples and is there a systematic way
of shelving and locating samples?
Is a map or key showing the locations N/A Archiving not
of archived samples readily available? underway yet
Are sample identifications X
permanent and legible?
Are sample numbers cross-referenced X
with field data and filed?
Are the stored samples tightly X
closed?
Are there any open samples stored X
in the cold storage facility?
Is the cold storage facility X
secure?
Is access to the facility limited X
to only specified laboratory personnel?
166
-------
PREPARATION LABORATORY PROCEDURES
Questions
Yes No
Comments
Is there a digital pH meter and is X
it capable of measuring pH to +/- 0.01
pH unit?
Is the combination of pH and reference X
electrodes high quality, low-sodium glass?
Are there 50-ml plastic X
containers?
Are there disposable X
stirrers?
Are there NBS traceable pH buffers X Opening date is not
of pH-4.0 and 7.0? labeled on bottle
Is the electrode stored properly X
when not in use?
•..U1..1— 4i» — -»••••• "T»-i- —r •••••»•!• _,. — _•_ — •.•.•••••»••-••••••«•• — •»•••••••»•—••••-•••••••••••••••••«•»'•••'•••
Is the open pan balance accurate X
to +/- 0.01 g?
Is the convection oven equilbrated X
at the proper temperature for 24 hours?
Is the desiccator equipped with X
desiccant and plate?
167
-------
SUMMARY
Questions
Yes No
Comments
Do responses to the evaluation indicate
that project and supervisory personnel
are aware of QA and its application to
the project?
Very impressed with
staff experti se and
enthusiasm
Have responses with respect to QA/QC
aspects of the project been open and
direct?
Very good
Has a cooperative attitude been
displayed by all project and
supervisory personnel?
Some resentment to
analysis of draft paper
for sulfate contamina-
tion
Is the overall QA adequate to
accomplish the objectives of
the project?
Are any corrective actions required?
If so, list in detail below.
See report
Comments:
168
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY-LAS VEGAS
P.O. BOX 93478
LAS VEGAS. NEVADA 89193-3478
(7O2/798-21 DO • FTS 545-21OO)
/IPR26
SUBJECT: Preparation Laboratory Audit Report, in Support of the
Mid-Appalachian Direct/Delayed Response Project,
conducted March 16, 1989
From: David G. Easterly /W*'<^ &• ^a^ojJU
EMSL-LV Quality Assurance Officer °
To: Audit Team Participants
TECHNICAL SYSTEMS AUDIT REPORT
OF THE SOIL PREPARATION LABORATORY
FOR THE DIRECT/DELAYED RESPONSE PROJECT MID-APPALACHIAN SURVEY
Location: 4675 Valley View Drive
Date: March 16 1989
Audit Participants:
Gene Easterly
Stuart Black
Dan Heggem
Mike Papp
Jeff Kern
EMSL-LV QA Officer
EMSL-LV Radiation Division
Science Advisor & QA Officer
EMSL-LV BAD QA Officer
LESC QA Operations Supervisor
ERL-C DDRP QA Coordinator
Preparation Laboratory Personnel:
Rick Van Remortel
Dave Jadhon
Glenn Merritt
Joe Parolini
Bryant Hess
Daron Peres
Prep Lab Manager
Lab Analyst
Lab Analyst
Lab Analyst
Lab Analyst
Lab Analyst
Purpose:
The audit was conducted in order to determine if the preparation
laboratory operations were being run in a manner consistent with Standard
Operating Procedures, and good laboratory practices.
169
-------
General:
Vith the exception of clod bulk density analysis, all laboratory pro-
cesses, from sample receipt, analysis, to sample shipment had been completed.
A previous systems audit team (Oct 6) had observed these activities. There-
fore, this audit was devoted to the observation of the clod bulk density
analysis, examination of rav data, and a review of the items of concern listed
on the previous audit.
Audit:
A brief review of the sample receipt procedures was conducted. The
sample receipt sheets were reviewed and a suggestion was made to bind these
sheets in a hard cover binder in numerical order. It was also suggested that
a second copy of these sheets be made. The forms that were reviewed were
complete. The sample storage cooler was examined and the storage and handling
of the samples was explained. Temperature graphs were examined for fluctu-
ations that might compromise the integrity of the samples but it appeared that
the cooler remained at a constant and acceptable level. The prep lab manager
then began the review of the clod bulk density procedure. The following
stages of this procedure were reviewed and demonstrated:
o clod drying
o clod dipping (saran)
o clod muffle furnaces
o clod sieving and weighing
- balance calibration log books reviewed. It was
observed that some of the calibration readings
were not initialed by the lab analyst.
o clod weighing in water
o data review
- it was recommended that the data be protectively
bound.
Although there were no samples in the drying area and most of the drying
tables had been moved out, the prep lab manager was questioned about concerns
the auditors had on the last visit. It was established that the prep lab
staff had moved tables that were directly under the air intake fans in the
ceiling thereby reducing the exposure of samples to possible dust contami-
nation. The auditors also noticed the set up for venting dust from the
sieving room.
170
-------
The lab audit was concluded in the main office area where raw data was
examined. The audit team reviewed their notes and concerns from the last
visit and drew the following conclusions:
- The prep lab is functioning very well.
- The laboratory manager and his analysts have
adequately demonstrated their knowledge of the
laboratory methods and good QA practices. Their
enthusiasm and dedication is commendable.
- The bulk density procedure appeared to be well
structured and efficient.
The following issues and corrective actions identified during the
previous audit have been dealt with:
o The lab staff began sub-sampling for pH, LOI and content by
taking a composite sample from 6 or 7 places from the bulk
sample rather than from one place.
o The lab has been making entries in pen and crossing out with
one line and initialed. There were a few corrective entries
that were not initialed and this will be avoided in the future.
o The lab adhered to the recommendation that bulk samples that
were very wet would be laid out to dry on the plastic sampling
bag.
o The lab attempted to standardize the method of disaggregating and
sieving bulk samples. The lab manager paid close attention to
the soil characterization forms noting highly weathered rock
fragments that needed particular care in sieving.
o The lab set a tolerance limit of IX deviation from the type "S"
sample weights when calibrating balances.
o The sample receipt forms have not been bound and sequentially
numbered. This will be accomplished.
o The dust study report was delivered to the ERL-C staff.
o Reports on Kraft paper and mesh contamination were delivered to
ERL-C. Further study of the possibility of contamination is
underway.
o An attempt was made to dry wet organic samples as soon as
possible to avoid microbial action. Extra fans were set close to
these sample and they were turned over more frequently than
mineral samples.
171
-------
The following issues or concerns were identified on this audit:
o All raw data notebooks should be protectively bound. A second
copy should be made.
o The lab manager mentioned that the entry system did not have
enough space for free form notes for some data forms. The
manager will check if this problem can be rectified.
o All balance calibration entries and data corrections should be
initialed.
In summary, the auditors were impressed with the organization of the
preparation laboratory and staff and feel that the operation is completing its
tasks successfully and within the specified quality objectives.
Participants:
Michael Papp, Lockheed
Lou Blume, BAD'
Stuart Black, NRD
Jeff Kern
Northrup Services
Environmental Research Laboratory
200 SW 35th St.
Corvallis, OR 97333
172
-------
Appendix H
Quality Assurance Reanalysis Templates
The following tables represent the evaluation criteria used to to determine whether a run
of samples for an analytical parameter satisfied the measurement quality objectives.
173
-------
Parameter: PH MP
Field
Audit
Low & Pair
QCCS
Internal
duplicate
duplicate
Manager
sample
MAJOR
If any 2 of 3 relationships occur and 2 minor
flags occur:
1 ) FAL beyond accuracy window
2 ) Mean of FAP or FAO beyond accuracy window
3) FAP of FOA SD > 0.15
Out of expected range
None
None
None
MINOR
If 1 of 3 of the relation-
ships listed in the major
column occurs
none
Routine/duplicate SD > 0.15
Routine/duplicate SD > 0.20
Sample beyond accuracy window
NON
If 0 of 3 of the
relationships
listed in the
major column
occur
within expected
range
Routine /duplicate
SD < 0. 15
Routine/duplicate
SD <0.20
Sample within
accuracy window
NOTE: Reanalysis would be requested on the occurrence of 1 major or 3 minors
Appendix H-l. Quality Assurance Reanalysis Template for the Determination of Field pH in Water
-------
Parameter: OM LOI
Field
Audit
Low & Pair
Internal
duplicate
Field
duplicate
Manager
sample
MAJOR
If any 2 of 3 relationships occur and 1 minor
flags occur:
1) FAL beyond QC control limits
2) Mean of FAP or FAO beyond control limits
3) FAP of FAO RSD > 15%
None
None
None
MINOR
If 1 of 3 of the relation-
ships listed in the major
column occurs
Routine/duplicate RSD > 15%
Routine/duplicate RSD > 20%
Sample beyond control window
NON
If 0 of 3 of the
relationships
listed in the
major column
occur
Routine/duplicate
RSD < 15%
Routine/duplicate
RSD < 20%
Sample within
control window
NOTE: Reanalysis would be requested on the occurrence of 1 major or 3 minors
Appendix H-2. Quality Assurance Reanalysis Template for the Determination of Organic Matter by Loss on Ignition
-------
Parameter: MOIST P
a>
Field
Audit
Low & Pair
Internal
duplicate
Field
duplicate
Manager
sample
MAJOR
If any 2 of 3 relationships occur and 1 minor
flags occur:
1) FAL beyond control limits
2 ) Mean of FAP or FAO beyond control limits
3) FAP of FAO RSD > 20%
None
None
None
MINOR
If 1 of 3 of the relation-
ships listed in the major
column occurs
Routine /duplicate RSD > 15%
Routine/duplicate RSD > 20%
Sample beyond control window
NON
If 0 of 3 of the
relationships
listed in the
major column
occur
Routine/duplicate
RSD < 15%
Routine/duplicate
RSD < 20%
Sample within
control window
NOTE: Reanalysis would be requested on the occurrence of 1 major or 3 minors
Appendix H-3. Quality Assurance Reanalysis Template for the Determination of Air-Dry Moisture.
-------
Appendix I
Laboratory Quality Control Charts
The following control charts represent the measurement quality samples used to evaluate
accuracy within each run of samples at the MASS preparation laboratory.
177
-------
Sample Type=FAP Bw Variable=PH_MP
5.4-f
X
a.
"So
'o
5.2-
5.0
4-8
ACC
. WARN
•
- MEAN
5.2-
OT
'o
2
1 4.8
4.6
4.4
WARN
ACC
20SEP88 150CT88
Figure 1-2.
09NOV88
Date
20SEP88 200CT88 19NOV88 19DEC88 18JAN89
Date
Figure M.
Sample Type=FAP B Variable=PH_MP
ACC
WARN
MEAN
WARN
ACC
04DEC88 29DEC88
178
-------
Sample Type=FAL O Variable=PH_MP
6.2-
6.0-
5.8-
:r
Q.
In 5"6-
O
s
2 5-4'
IZ
5.2-
5.0-
4.8,
•
• ' ACC
. ... . . WARN
•
* • •.. ••*
,„.-» ...... ..... , -. , M£AN
• . '
:
WARN
•
ACC
Figure 1-3.
4.8T
4.6-
CO
'o
0)
4.4-1
4.2-
200CT88
Figure 1-4.
20SEP88 200CT88 19NOV88 19DEC88 18JAN89
Date
Sample Type=MS Varioble=PH_MP
ACC
WARN
MEAN
WARN
ACC
14NOV88
09DEC88
Date
03JAN89 28JAN89
179
-------
Sample Type=FAP Bw Variable=OM_LOI
8-1
-------
Sample Type=FAL C Variable=OM_LOI
CD .5 -
-+-*
-t—'
o
y
i_
o
c
o
i_
CD '
CL '-
.2-
Figure 1-7.
11 -r
10-
C
O
en
^
O
c
CD
O
i^_
CD
CL
9-
8-
7-
6-
Date
Sample Type=MS Variable=OM_LOI
200CT88
Figure 1-8.
14NOV88 09DEC88
Date
CONT
WARN
•
- MEAN
WARN
CONT
, , , , , ! ,
20SEP88 200CT88 19NOV88 19DEC88 18JAN89
CONT
* WARN
MEAN
WARN
CONT
03JAN89 28JAN89
181
-------
Sample Type=FAP Bw Vanable=MOIST_P
2.5-
2.0-
22 1.5-1
a>
L_
2 1.0J
0.5.
o.o.
! *
Figure 1-9.
Sample Type=FAP B Variable=MOIST_P
O
Q_
4.0-
3.0-
2.0-
1.0-
0.0-
20SEP88
Figure 1-10
ACC
- MEAh
20SEP88 200CT88 19NOV88 19DEC88 18JAN89
Date
:ONT
VARN
vIEAN
\CC
/VARN
:ONT
150CT88 09NOV88 04DEC88 29DEC88
Date
182
-------
Sample Type=FAL C Variable=MOIST_P
o.u-
2.5-
^ 2.0-
en
"o
•o 1.5-
«
L_
2 1.0-
0.5-
0.0-
ACC
.... «-^.-r,.—i-._. •_. l_...s_.LT. -,-— : , ^ . MEA^
20SEP88 200CT88
Figure 1-11.
Sample Type=MS Variable=MOIST_P
Figure M2.
9NOV88 19DEC88 18JAN89
Date
o.u-
4.0-
i — '
| 3.0-
_~0)
i 2.0-
o
Q_
1.0-
0.0-
200
CONT
WARN
. * . . . . *
' ' ACC ,
• • . - MEAN
•
• WARN
CONT
1 i | i | i 1 1 jJ
3T88 14NOV88 09DEC88 03JAN89 28JA
Date
183
U.S. GOVERNMENT PRINTING OFFICE: 1990—7 ^ 8 - 15 9/ 20"*59
------- |