United States
Environmental Protection
Agency
Environmental Monitoring Systems*
Laboratory
Las Vegas NV 89114
Research and Development
EPA-600/S4-84-012 Mar. 1984
^EPA Project Summary
Documentation of EMSL-LV
Contribution to Dallas Lead Study
K.W. Brown, W.F. Beckert, S.C. Black, G.T. Flatman, J.W. Mullins, E.P.
Richitt, Jr., and S.J. Simon
During the summer of 1982 the U.S.
Environmental Protection Agency (EPA),
conducted a comprehensive multimedia
environmental monitoring program in
the vicinity of two secondary lead
smelters located in Dallas, Texas. This
monitoring program, which was con-
ducted within a 1 mile radius of the two
smelters, included a major soils investi-
gation and the collection and analysis of
approximately 3,000 soil, 1,000 blood,
and 750 dust samples. Approximately
250 vegetation samples were also
collected.
For this study the Environmental Moni-
toring Systems Laboratory, Las Vegas,
NV (EMSL-LV) was requested to provide
documentation and guidance for the
quality assurance/quality control (QA/
QC) program, and to provide the
analytical methods and soil sampling
procedures protocols. The extensive
QA/QC program was applied to all
phases of the sampling procedures and
analytical methods for documenting
the precision and accuracy of the moni-
toring data. Also, strict chain-of-
custody procedures for the sampling and
analytical programs were employed to
assure the integrity of the monitoring
data.
EMSL-LV designed a soil monitoring
strategy that identified isopleths of
constant soil lead concentration within
each of the designated monitoring
areas. The results of this soil monitoring
strategy plus the protocols, reports and
pertinent documentation provided by
EMSL-LV for the Dallas Lead Study are
presented.
This Project Summary was developed
by EPA's Environmental Monitoring
Systems Laboratory, Las Vegas, NV, to
announce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
During the latter part of 1980, analytical
results of home soil samples collected by
an Environmental Protection Agency
(EPA) Contractor in residential areas in
the vicinity of two secondary lead
smelters, the RSR Corporation (RSR) and
the Dixie Metals Company (DMC), located
near the National Lead (NL) smelter in
Dallas, Texas, showed soil lead concen-
trations exceeding 10,000 ppm. Based on
these data, it was recommended that
before any remedial action be undertaken
in these populated areas, the following
conditions should be adequately demon-
strated:
O that a potential public health prob-
lem exists,
O that elevated blood lead levels occur
around the smelters as compared to
other similar areas of Dallas,
O that decreasing blood lead levels
occur with increasing distance from
the smelters,
O that positive correlation exists
between blood lead levels and soil
lead levels.
In September 1981, EPA's Office of
Waste Programs Enforcement requested
that EMSL-LV assist in the review and
evaluation of the previously collected soil
lead data and that EMSL-LV recommend
additional soil sampling, if required, to
the EPA's newly formed Lead Smelter
Study Group in Washington, D.C. EMSL-
LV undertook that review and ultimately
provided a variety of technical services
and support to the study effort.
-------
This report contains the critique of the
previously conducted environmental lead
monitoring studies, as well as the
strategy and protocols prepared by
EMSL-LV for soil sampling and-monitor-
ing, soil handling methods, soil and dust
analytical procedures, and the attendant
quality assurance plan for a study to
identify the geographical distribution of
soil lead concentrations around the two
secondary lead smelters and one reference
(REF) area. In addition, the report
contains the soil and dust quality assur-
ance/quality control (QA/QC) procedures,
the results of a soil homogeneity and
equipment decontamination methods
evaluation, the auditing procedures and
results, a land use analysis, and the
geographical distribution of soil lead in
the three areas.
During the course of this field monitor-
ing study, chain-of-custody and data
validation procedures were utilized. A
discussion of the requirements and the
methods employed for these are also
included.
So/7 Sampling and Monitoring
Strategy
Preliminary soil lead data collected
from the previous monitoring efforts were
used in designing the monitoring and
sampling strategy. The data points were
plotted on maps encompassing a two-
mile radius around each of the smelter
sites. Observation indicated that higher
lead concentrations were generally found
in those soils collected near the smelters
and that the levels decreased with
distance. Therefore, the lead levels were
ranked as a function of distance from
their respective smelter site. Geostatistics
and the statistical methodology of kriging,
which deals with regional and/or spatial
variables, were used to design the
monitoring strategy. The application of
geostatistics provided not only a means of
evaluating the spatial variability of the soil
lead levels in the vicinity of the two
secondary lead smelters but also a means
of estimating the variance at points lying
within the geographical boundaries of the
sampling network.
Asemivariogram was used todetermine
the optimum distance between sampling
points required to fully describe the
geographical lead distribution. The
sampling distance is calculated in the
following way. A soil lead observation is
considered a regional variable if the lead
levels of spatially nearby samples are
numerically closer than lead levels of
samples from remote areas. A regional
variable isautocorrelated in time or space
or both. The initial step calculates the
semivariance (1 /2 of the variance) by the
following equation:
n
(h)=1/2n- I (y,i+h.-y)2
i=1
where
n=the number of pairs of points
h distance apart
(h)=the semivariance
h=vector distance between pairs
of points
yi=the lead concentration value
at point i
y(i+h)=the lead concentration value
at a point h distance from i
The (h) values are identified along a line
to provide data for the construction of the
semivariogram. The semivariogram is a
graph (as shown in Figure 1)depictingthe
square of the difference between samples
at distance h apart plotted against
distance h. If the samples are uncorrelated
the semivariogram is a horizontal line.
Semivariogram A in Figure 1 shows the
plotted soil lead data originally collected
from the REF area. This horizontal line
semivariogram indicates that the soil lead
levels originally determined in the REF
area were not structured.
If the sample levels are correlated, the
nearest values of h-distance have rising
or increasing squares of differences. At
greater h-distances the squares of
difference become horizontal as shown in
Square of Difference
Between Points at
Distance h Apart
Dal/as Lead Study
Semi- Variogram for
REF Area
•
Distance h
Figure 1, semivariograms B and C. The
rising or initial portion of semivariograms
B and C measures the range of correlation.
The calculated distance between sampling
points for optimum monitoring benefit is
judged to be approximately two-thirds of
the measured range of correlation.
Semivariograms B and C constructed
from the original RSR and DMC soil lead
data (Figure 1), show that there is a
structure or autocorrelation to the soil
lead levels collected from these two sites.
The measured range of correlation at
each of the two sites was about 366
meters. Consequently the optimum
distance between soil sampling points
was calculated to be 228 meters.
A square sampling grid with squares
measuring 228 meters on each side was
overlain on recently flown aerial photo-
graphs of the three monitoring sites. The
grid intersections were used to identify
potential sampling locations- A total of
177 locations were identified for sampling
in the RSR and DMC monitoring sitesand
80 in the REF site.
Four sdil cores measuring 2 cm in
diameter and 7.5 cm in depth were
collected around the perimeter of a 10
meter diameter circle at or near each grid
intersection. After the four soil cores
were collected, they were placed in a
single container and pre-processed in a
laboratory before analysis. The pre-
processing was to provide the analytical
B
01
Cj
C >-
Qj *Q *-
I? I
S^
o C m
01 0> TO
act;
& ai -5
CO CQ Q
Da/las Lead Study
Semi-Variogram for
RSR Area
••...'*.
OK
Distance h
-------
laboratory with a homogenous sample.
This homogenization process involved
drying the soil at 100°C, pulverizing it
with a mortar and pestle, sieving it
through a standard 16 mesh stainless
steel sieve, and mixing it for 30-minutes
in a porcelain grinding jar.
Analytical Methods
The analytical method used to determine
the lead in soils was Atomic Absorption
Spectroscopy (AA). The method involved
additional drying, pulverizing with mortar
and pestle, sieving through a standard
100-mesh sieve, and digesting a 5 g
aliquot with 8N nitric acid (HMOs) before
AA injection.
Quality Assurance/Quality
Control Procedures
The quality assurance/quality control
(QA/QC) objectives were established to
ensure documentation of the precision
and accuracy of the monitoring data.
Precision addresses variability, specifically
the identification of the range of values
that may be encountered in repeated
measurements. Accuracy includes both
the qualitative and quantitative analyte
measurements. To achieve these QA/QC
objectives, to handle the large volume of
samples efficiently and economically,
and to assure high quality decision-making
data, a large number of QA/QC samples
were processed along with the field
samples, and a number of QA/QC
sampling and analytical procedures were
incorporated in the study design.
Geostatistical Results
The utilization of geostatistics in the
planning and in the data analysis of this
environmental monitoring study permitted
geographical identification of lead con-
centration isopleths that showed the
following:
O Among a few patterns of high lead
values, both the DMC and RSR
smelters have a single dominant
pattern that includes the smelter.
(Dominance is defined as both a
large area covered along with high
lead values.) The closure of the
pattern implies that the source is
inside. There is a steep gradient
demarking the polluted areas, indi-
cating that the smelters are probably
the primary contributor.
O The lead patterns of DMC and RSR
smelter areas are different from the
patterns of the REFarea in magnitude
of lead values (500 vs. 2,500, 3,000
ppm) and gradients (rate of change)
between inside and outside lead
isopleth (200 to 500 vs. 300 to
2,500, 300 to 3,000 ppm).
O The DMC smelter area has a second
high isopleth pattern that is caused
by one sample (10,400 ppm). This
value may reflect a local maximum
for a very small area and should be
verified by re-analysis or re-sampling
before any decision is made about
that area.
O The kriging error isomaps showed
that the field sampling was compre-
hensive enough to cover the areas of
interest and intense enough to
ensure an error of estimate ranging
from 1.5 times the estimated value
in the center of the study areas to a
maximum of 2.2 times the estimated
value at the outer boundaries.
O The use of geostatistics for identify-
ing and evaluating environmental
contamination is a useful and
powerful tool for the environmental
scientist.
The EPA authors K. W. Brown, W. F. Beckert, S. C. Black, G. T. Flatman, J. W.
Mullins, and E. P. Richitt, Jr., are with the Environmental Monitoring
Systems Laboratory, Las Vegas, NV 89114; S. J. Simon is with Lockheed
Engineering and Management Services, Inc., Las Vegas, NV 89114.
K. W. Brown is the EPA Project Officer (see below).
The complete report, entitled "Documentation ofEMSL-L V Contribution to Da/las
Lead Study, "(Order No. PB 84-145 564; Cost: $40.00, subject to change) will be
available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
P.O. Box 15027
Las Vegas, NV 89114
------- |