Guidance for Verifying and Interpreting Field Blank Determinations that Indicate High Pesticide or Trace Element Concentrations or that Show Large Numbers of Detections



In Reply Refer to:
Mail Stop 412                                 November 14, 2000

Office of Water Quality Technical Memorandum 2001.02

Subject:  Guidance for Verifying and Interpreting Field Blank
          Determinations that Indicate High Pesticide or Trace Element
          Concentrations or that Show Large Numbers of Detections

BACKGROUND

Good quality assurance (QA) for water-quality data typically involves
collection of quality control (QC) samples (field and laboratory blank and
spike samples) and the subsequent data analysis of these samples.  Multiple
sources of laboratory QC are available for both blank and spiked QC
samples, and several sources of QC data are available for field spikes.
However, only project-specific sources of data are available for field
blanks because of unique features of each project, although national
aggregations of project blank data have produced useful results.

Occasionally, through errors in QC sample handling, field blanks or spikes
may be misidentified, or environmental samples may be misidentified as QC
samples.  It is relatively straightforward to recognize misidentified
spikes because the array of analytes in field samples rarely looks like a
typical spiked sample.  However, it is more difficult to identify an
environmental sample mislabeled as a blank, and the consequences of such a
misidentification can be substantial for subsequent data interpretation if
the analytes of interest are present in the environmental sample.

PURPOSE OF MEMORANDUM

Guidance is provided to project hydrologists and others on when to question
blank results for pesticide and trace element determinations for field
samples.  The National Water-Quality Assessment (NAWQA) Program, in
conjunction with the Office of Water Quality (OWQ) and the National Water
Quality Laboratory (NWQL), developed the pesticide blank criteria.  If the
criteria for questioning blanks are met, then a series of checks and
reviews are recommended prior to accepting the questionable blank results
as truly representative of blank conditions.

GUIDANCE AND INTERPRETATION

     PESTICIDES

For NWQL pesticide schedules 2001, 2010, 2050, and 2051, the NAWQA program
has suggested that any field blank that contains an unusually large number
of pesticide detections, or pesticide detections at an unusually high
concentration, be carefully examined for possible misidentification errors.
For blanks associated with ground water, 3 or more pesticide detections or
concentrations of 0.015 micrograms per liter (ug/L) or greater are
reasonable limits to investigate the sample for misidentification or other
errors.  For blanks associated with surface water, 4 or more pesticide
detections or concentrations of 0.060 ug/L or greater are reasonable limits
to investigate the sample for misidentification or other errors.

These limits were derived from a statistical analysis of 363 ground-water
field blanks and 458 surface-water field blanks.  In all cases, the
suggested limits correspond to the 95th percentile of ground- or
surface-water blank numbers of detected pesticides or pesticide maximum
concentrations.  Thus, only 5 percent of the time should a true blank
exceed these upper limits of numbers of detections or concentrations.  For
comparison, set blank results are available to USGS employees from the NWQL
at: http://wwwnwql.cr.usgs.gov/USGS/Catalog/SearchQCSets.html.  Enter the
desired pesticide schedule number, select "Procedure/Schedule", search, and
then select your appropriate NWQL quality control set number.  The set
blank and set spike information will be displayed, along with the reporting
limits, for each analyte.

The NWQL is developing two new low-level pesticide methods, which are
currently known as Lab Codes 9002/9011 and 9060.  When approved, these will
be Schedules 2002/2011 and 2060.  Although the same logic can be applied to
blank determinations using these Lab Codes or Schedules, different numbers
of detections or maximum concentrations may be appropriate because the
analyte lists are different.  At this time, there are insufficient field
blanks to determine these limits.  When published, the Open File Reports
for these methods will provide laboratory blank information, which can
provide a starting point for hydrologists to identify anomalous field
blanks that may require verification.

     TRACE ELEMENTS

For trace elements, conductivity is determined on all samples at the NWQL
that undergo Inductively-Coupled Plasma/Mass Spectrometry (ICP/MS) and for
Inductively-Coupled Plasma - Optical Emission Spectroscopy (ICP-OES).  For
these trace element determinations, if a blank shows an unexpected
concentration or number of trace element detections, check the measured
conductivity.  Inorganic blank water should have a very low conductivity,
while a misidentified blank should have a conductivity more representative
of other samples collected at that site, typically much higher than blank
water.  For comparison, Inorganic Blank Water analysis results are
available to USGS employees from the Quality-Water Service Unit in Ocala
at:  http://qwsu.er.usgs.gov/qwsu_ibw.html.  If a simple switch has been
made, then one of the environmental samples may have a low conductivity
representative of blank water, and the blank data may be recovered from
that sample, along with the sample data from the switched blank.

     SUGGESTED ACTIONS FOR EXCEEDANCES

If any field blank samples exceed the above criteria for numbers of
pesticide detections, maximum pesticide concentration, or
higher-than-expected conductance values for trace element blanks, then all
aspects of this sample handling need to be reviewed with care, such as:

Ensure that the sample submitted for analysis was a field blank by
verifying that all sample data submitted in the cooler containing the
questioned blank are consistent with each assigned sample identification.
This checks for most possible misidentification of samples and blanks in
the field or at NWQL login.  Contact the NWQL and request that the sample
results be verified, along with the reason for the verification, such as a
high number of pesticide detections or high concentration for a blank, or a
high conductivity (or other anomalous analytical result).

     WHAT NEXT?

If a clear picture emerges of a sample handling error that can be repaired,
then the corrections should be made in the data base, along with
appropriate written supporting documentation.  If there is no clear
picture, then the hydrologist should not make data base corrections, but
should carefully consider the impacts of questionable blanks on data
interpretation, such as in the following examples:

Including a high blank will have the effect of calling into question all
environmental analytical results near the concentration of the high blank.
If these environmental concentrations are the basis of an important, and
potentially controversial finding, then keeping the high blank
concentration may be vital.  Collection of additional field blanks and
possibly field spikes near the concentrations of concern may be
appropriate.

If, on the other hand, the environmental results are more than an order of
magnitude higher in concentration than the high blank result, then keeping
the high blank result may be less important, and it could be rejected as an
outlier without substantial impact to the data interpretation.

These two examples lie at the extremes of accepting or rejecting the high
blank as an anomaly in data interpretation.  Most actual situations will be
less clear, and will require the hydrologist to carefully examine actual
data-quality objectives and evaluate the potential impacts of including an
anomalously high blank, and the impacts of rejecting a high blank that may
be real.

     CONTACTS AND FEEDBACK

Pete Rogerson, Senior Chemist, OWQ (rogerson@usgs.gov) will be glad to
discuss specific applications with project personnel and provide additional
guidance, if possible ((303) 236-1836).  Pete would also appreciate
feedback on unusual blank occurrences and how you resolved them.

Jeff Martin (jdmartin@usgs.gov) would appreciate feedback on how the
guidance works for identifying questionable pesticide field blanks.

Merle Shockey, NWQL, (mshockey@usgs.gov) would appreciate any information
relating to sample misidentifications or other problems occurring at the
NWQL.

                            Janice R. Ward /s/
                            Acting Chief, Office of Water Quality

This memorandum does not supersede any other OWQ Technical Memorandum.

Distribution: All WRD Employees

*********************************************
Janice Ward
Acting Chief, Office of Water Quality
US Geological Survey
Reston, VA
703-648-6871
jward@usgs.gov