The Baseline - Mining and the Environment: From Ore to Metal - page 376

Once sampling is completed, i eld sheets are the sole monitoring records, and it is good
advice to always maintain a copy of i eld data collection sheets in case originals become lost.
Baseline data will eventually be used for decision-making, evaluating impacts and/or to
fuli l regulatory requirements. Thus, data will need to meet quality objectives set by those
who will ultimately use the data. Developing a Quality Assurance and Control Plan (QA/
QC Plan) is good practice, and is required by some jurisdictions. A QA/QC Plan is a writ-
ten document that outlines the procedures adopted to ensure that the samples collected and
analyzed, the analyses themselves, the data that is stored and managed, and the reports
reported are of sufi cient quality to meet the required data needs. Developing a QA/QC
Plan requires developing Standard Operating Procedures (SOPs) for all i eld sampling and
i eld/laboratory analytical methods. SOPs are step-by-step directions, including calibration
and maintenance procedures for i eld and laboratory analytical instrumentation. A number
of manuals provide detailed method descriptions, SOPs and even i eld sheets (e.g. USGS
1998 and EQB 1994). Most laboratories maintain SOPs for all their analytical protocols.
The Quality Assurance and Control Plan needs to consider the full spectrum of activi-
ties related to sampling and analysis, from the maintenance of sampling equipment at the
ofi ce to the critical review of laboratory data ( Figure 8.7 ) .
Some parameters are measured in situ (such as climatic parameters, sound levels, pH
and conductivity of water, and bird sightings) but many samples are sent to a laboratory for
analyses. Using i eld meters for some in situ measurements is acceptable, in fact necessary for
some parameters such as temperature, dissolved oxygen (DO), and pH, which can change
during the period between sampling and laboratory testing. Instruments can also change, so
that it is important that instruments or meters used for in situ measurements are calibrated
each time they are used and that all calibration results are recorded. Parameters for which
meters are frequently used include: DO, temperature, pH, conductivity, and turbidity.
Field kits are also available for a wide variety of other analyses, including metals.
However, laboratory analyses are generally used for cations and anions as much of the
work is automated, detection limits are generally much lower, and quality control is easier
to achieve in a laboratory than in the i eld. Most i eld practitioners would agree that there
are enough problems involved in i eld sampling, that they would not wish to cope with the
multitude of additional issues involved in chemical analyses.
Notwithstanding the above, laboratory analyses probably cause more confusion,
uncertainty, and disputation than any other part of the baseline data collection process.
Reproducibility of results between laboratories is notoriously difi cult to achieve for a
wide variety of reasons including differences in the type and age of equipment, operat-
ing and quality control procedures. In response to this issue which may sometimes have
serious legal implications, some countries such as Australia have introduced laboratory
accreditation schemes. Australia's National Association of Testing Authorities (NATA - see
www.nata.asn.au ) provides this service which involves inspection of laboratories, conduct-
ing examinations for laboratory staff, and sending reference samples for analysis by mem-
ber laboratories.

The Quality Assurance and
Control Plan needs to consider
the full spectrum of activities
related to sampling and analysis,
from the maintenance of
sampling equipment at the
offi ce to the critical review of
laboratory data.

Reproducibility of results
between laboratories is
notoriously diffi cult to achieve.

CASE 8.8
Reporting Results of Numerical Calculations

When reporting results of calculations, excess decimal
places should be avoided. The following rule of thumb

may be used: scan all the values used for the calculation,
and identify the measured value with the fewest decimal

places. The fi nal answer should have that same number of
decimal places.