PROGRAMS AND PLANS--Quality of Existing Dissolved Trace-Element Data
In Reply Refer To: March 20, 1992
Mail Stop 412
OFFICE OF WATER QUALITY TECHNICAL MEMORANDUM 92.05
Subject: PROGRAMS AND PLANS--Quality of Existing Dissolved
Trace-Element Data
BACKGROUND
Recent studies indicate that results for a number of dissolved trace
elements reported in the Division data base include contamination.
These findings were previously reported in Office of Water Quality
(OWQ) Technical Memorandum 91.10. Next, a suggested caveat
for inclusion in State data reports was provided in OWQ Technical
Memorandum 92.03; the caveat was then revised in memorandum 92.04.
The purpose of this memo is to describe: (a) some implications of
OWQ Technical Memorandum 91.10, and (b) how the Division can deal
with issues of uncertainty in the validity and usefulness of
existing dissolved trace-element data.
IMPLICATIONS OF OWQ TECHNICAL MEMORANDUM 91.10
Results of the various studies reported in OWQ Technical Memorandum
91.10 indicate that unacceptable contamination was identified in
dissolved trace-element data for arsenic, boron, beryllium, cadmium,
chromium, copper, lead, and zinc. Because the examined data sets
were rather small, this does not imply that elements not listed are
free from contamination. Nor does the finding imply that the eight
trace elements can not be, or have not been, collected in specific
projects in a contaminant-free manner.
NASQAN data collection involves hundreds of people and sets of
equipment, including many different types of sampling devices. For
these reasons, collection of data for the eight trace elements cited
above has been suspended in NASQAN until OWQ can: (a) certify
selected sampling equipment as trace element "contaminant free" at
the part-per-billion (ppb) level; (b) provide a certified sample
collection and field processing protocol for trace elements at the
ppb level, (c) develop and deliver associated training, and (d)
provide guidelines for collecting the proper types and adequate
amounts of field QC data required to support the quality of
collected trace-element data.
In contrast to the situation for NASQAN, present data collection
efforts should continue for individual projects, provided: (a) the
field methods are appropriate to the stated project objectives and
to the levels of trace elements that occur in the sampled hydrologic
system, and (b) adequate QC data are generated to substantiate the
quality of the trace-element data and to identify problems with
contamination.
After suitable equipment, a proper protocol, and appropriate QC
guidelines are available, NASQAN will resume production of
dissolved data for the eight elements. The results reported in
Memorandum 91.10 indicate that most of the contamination existing
in ppb level trace-element data derives from sample collection and
field processing, rather than from laboratory handling and
analyses. Thus, considerable diligence will be required in sample
collection and field processing to reduce trace-element
contamination. The effort must cover all trace elements, not just
the cited list of eight.
When NAWQA begins to collect dissolved trace-element data, the
program will support the field effort with appropriate QC data.
In addition, at the beginning of fiscal year 1993, the OWQ will
provide QC guidelines to Districts for all classes of constituents
in all sample media. At that time, the OWQ will formally suggest
that all Division projects substantiate the quality of collected
environmental data by adding a field QC component.
QUANTITATIVE CORRECTIONS CAN NOT BE MADE TO THE
EXISTING DATA BASE
Most dissolved trace-element data in the existing data base were
collected without field equipment blanks and other collaborative
QC data. Thus, we can not: (a) assure that results are
contaminant free, or (b) estimate the amount of contamination
based on actual data. The studies reported in Memorandum 91.10
show that depending on the specific trace element, contamination
derives primarily from: (a) sampling, (b) field processing, or
(c) both sampling and processing. Specific studies on sampling
devices indicate that although the amount of contamination depends
to a considerable degree on the type of sampler, large variability
occurs between samplers of the same type. In other words, some of
the contamination is systematic (sampler type), whereas some is
random (variation within sampler type). This finding is
important, because once identified, sources of systematic
contamination can be removed, whereas random contamination can
only be identified by having adequate QC data collected at the
time of sampling.
The Bottom Line
There is no way to apply a blanket correction factor to historic
data sets for dissolved trace elements, even for individual trace
elements at individual sites. This inability stems from: (a) the
general lack of field equipment blank data, and (b) observed
variabilities in both the sources and nature of contamination for
a given trace element.
WHAT WE CAN SAY NOW ABOUT EXISTING DISSOLVED
TRACE-ELEMENT DATA
1. Concentrations below the reporting level--all results showing
less-than values are considered to be unaffected by
contamination at that reporting level.
2. Concentrations at or above the reporting level--
A. If adequate field QC data (especially field blanks) are
available for specific data sets, and no detections
(concentrations above the reporting level) occur in the
blanks, the environmental data are probably unaffected by
significant contamination from sampling and field
processing. Accordingly, the data are acceptable. In
WATSTORE/NWIS-I, codes exist to indicate the availability
and type of QC data associated with each sample. Although
these QC codes exist in the present data base, they have
rarely been used; consider employing these codes wherever
possible.
B. If adequate field QC data are available for specific data
sets, and random detections occur in the blanks, some or
all of the environmental data may have been contaminated
by sampling and field processing. In such cases, and
without additional QC information, the environmental data
must be viewed as questionable, and an appropriate caveat
should be added to the data base and to data reports (for
reports, see OWQ Technical Memorandum 92.04). Caveats
cannot be entered into WATSTORE/NWIS-I; however, NWIS-II
will accomodate such comments. In addition, interpretations
of such a data set must take into account the impact of
potential outliers (which might result from the random
contamination observed in the blanks). For example, if
upon testing there is an apparent time trend, does it
still occur if the potential outliers are removed?
C. If adequate field QC data are available for specific data
sets, and systematic detections occur in the blanks, all
of the environmental data must be assumed to be contami-
nated by sampling and field processing. Accordingly, the
data are unacceptable, and an appropriate caveat should be
added (See B above).
D. If adequate QC data are unavailable, environmental results
showing detectable concentrations are questionable for at
least arsenic, boron, beryllium, cadmium, chromium,
copper, lead, and zinc. Accordingly, an appropriate
caveat should be added (See B above).
HOW WE CAN BETTER DEFINE THE QUALITY OF EXISTING
DISSOLVED TRACE-ELEMENT DATA
As previously noted, the OWQ is developing: (a) a new protocol for
producing dissolved trace-element data that are uncontaminated at
the ppb level, and (b) guidelines for producing field QC data,
including equipment blanks, to certify the quality of data
produced by the protocol. The target date for implementation of
the protocol and guidelines is October 1, 1992. Once the new
protocol is available, Districts can use the following approaches
to better define the quality of data produced by the present
protocol:
1. Conduct a side-by-side comparison of the new and present
protocols at selected sites to determine the mean levels
and variability of contamination in the dissolved data for
specific trace elements.
2. After an adequate number of data are generated at a site
using the new protocol, statistically analyze the old and
new data sets to compare means and variabilities for
specific trace elements.
Neither approach will enable correction of the existing data base
because of the multiple sources and inherent variability of
contamination for each trace element that arises through use of the
present protocol. Moreover, approach 2 will not provide a clear
understanding of the levels of contamination in the historic data
base because of temporal variability in the actual environmental
concentrations of the dissolved trace elements. However, use of
either approach will enable Districts to gain some increased insight
into the general level of contamination in the existing data base
for specific trace elements. The OWQ will work with the Branch of
Systems Analysis to define: (a) proper designs for both approaches,
and (b) appropriate procedures for statistical analysis of the
resulting data.
David A. Rickert
Chief, Office of Water Quality
Key Words: Trace elements, contamination
This memorandum refers to Office of Water Quality Technical
Memorandums 91.10, 92.03, and 92.04.
Distribution: A, B, S, FO, PO