The Field Sample

A field sample set consists of the environmental samples and several kinds of quality control samples. The project manager must determine what kinds of quality control samples should be collected, based on the purposes of the sampling program.

Quality control samples

Some field samples should be collected for quality control purposes. Sampling procedures can contribute to both systematic and random errors. Many environmental waters are inherently not uniformly mixed. Field blanks, spikes, and duplicates will help estimate errors caused by sampling bias (usually contamination and/or nonrepresentative samples) and calculate sampling precision.

* Detailed guidance for these procedures can be found in Environmental Sampling and Analysis: A Practical Guide,15 and in reference works such as Standard Methods, 1995.

Field QC samples must be handled exactly the same way as the environmental samples, using identical sampling devices, sampling protocol, storage containers, preservation methods, storage times, and transportation methods. For a correct interpretation, quality control samples must be subject to the same holding time criteria as the environmental samples.

Blank sample requirements

There are several types of blank samples, each serving a distinct QC purpose. Wherever a possibility exists for a sample to become contaminated, a blank should be devised to detect and measure the contamination. The most commonly collected blanks include Geld, trip, and equipment blanks. Although it may be prudent to collect a full set of blanks, most of the blank samples usually do not need to be analyzed. Analysis costs can be reduced if the strategy of blank sampling is understood. First analyze only the field blanks, which are susceptible to the broadest range of contaminant sources. If these indicate no problems, the other more specific blanks can be discarded or stored. If a problem is suspected, other blanks can be analyzed for confirmation of the problem and discovering the source. Holding time limits must be observed with blanks as well as with environmental samples.

Field Blanks: At least one "clean" water sample should be exposed to the same sampling conditions as the environmental samples at each sampling site. The analytical laboratory can generally provide analyte-free distilled water for this purpose. Field blanks are transferred from one container to another, passed through automatic equipment, or otherwise exposed to the conditions at the sampling site. At a minimum, the field blank container is opened at the sampling site and exposed to the air for approximately the same time as the environmental samples. It is then capped, labeled, and sent to the laboratory with other samples. Field blanks measure incidental or accidental sample contamination throughout all the steps of transportation, sampling, analysis, and sample preparation at the laboratory. They help assure that artifacts are recognizable and are not mistaken as real data.

Trip Blanks: For each type of container and preservative, at least one container of analyte-free water should travel unopened from the laboratory to the sampling sites and back to the laboratory. Trip blanks serve to identify contamination from the container and preservative during transportation, handling, and storage.

Equipment Blanks: These are especially important with automated sampling equipment. Equipment blanks, sometimes referred to as rinsate blanks, document adequate decontamination of the sampling equipment. These blanks are collected by passing analyte-free water through the sampling equipment after decontamination and prior to resampling.

The analytical laboratory will run a similar set of blanks to document contamination and errors arising from handling and analysis procedures in the laboratory.

Field duplicates and spikes

Field duplicates are two separate environmental samples collected simultaneously at the identical source location and analyzed individually. Field duplicates are sensitive to the total sample variability, for example, variability from all sampling, storage, transportation, and analytical procedures. Where the goals of the sampling program warrant it, as in permit monitoring requirements, at least one field duplicate per day should be collected for each analyte. More than two replicates may be required in cases where it is difficult to obtain representative samples.

Field spikes are environmental samples to which known amounts of the analytes of interest are added. Ampoules containing carefully measured amounts of analytes can be purchased from chemical supply sources. Field spikes can identify storage, transportation, and matrix effects, such as loss of volatile compounds and analytical interferences caused by certain compounds that are present in the environmental source. In a spiked sample with no problems, the measured analyte concentration should be equal to the concentration present in the environmental sample plus the added spike concentration within the limits of the method's precision.

Understanding Laboratory Reported Results

When a laboratory reports that a target compound was not detected, it does not mean that the compound was not present. It always means that the compound was not present in a concentration above a certain lowest reporting limit. There always is the possibility that the compound was present at a concentration below the reporting limit. The laboratory might even have identified the compound below the reporting limit but not reported it because the concentration could not be quantified within acceptable limits of error.

Reported results of analyte concentrations are never exact. There always is some margin of error. The only kind of measurement that can be exact is a tally of discrete objects, for example, dollars and cents or the number of cars in a parking lot. When measuring a quantity capable of continuous variation, such as mass, length, or the concentration of benzene in a sample, there always is some uncertainty. Like an irrational number, measurements capable of continuous variation can always be expressed to more significant figures, and any answer with a finite number of digits is always an approximation.

In addition, many repeated measurements of the same quantity, even under essentially identical conditions, will always yield results that are scattered randomly about some average value, due to uncontrollable variations in environmental, experimental, and operator behavior. The person who interprets experimental data must always bear in mind that no reported experimental value necessarily represents the "true" value. In other words, it is never possible to be completely certain of a result. Nevertheless, we still try to answer questions such as "Does this sample indicate a violation of a discharge limit?" or "Does the data indicate that my remediation activities are beginning to work?" Answers to such questions must always acknowledge that there is a range of error. The purpose of QA/QC controls is to assure that the reported value lies within a small enough range of error to be considered useful data.

Part of laboratory QA/QC involves determining how much scatter in the data is to be expected from different analytical procedures. The precision inherent in a particular procedure is a measure of how much deviation may be expected in repeated measurements of the same sample. The standard deviation is a common way of quantitatively expressing the precision, or reproducibility, of a measurement. A more thorough treatment of the standard deviation can be found in elementary statistics texts or in Standard Methods.30 The procedure for using the standard deviation to expressing the reliability of measurements is described by the EPA in several documents (e.g., U.S. EPA, 1992b; 40 CFR 136, Appendix B). Here, it is only necessary to understand that, for many measurements of the same quantity, one standard deviation above and below the average value of the measurements will include about 68% of all the individual values. For example, suppose the concentration of benzene in a sample is measured repeatedly many times to yield an average value of 12.3 |j.g/L and a standard deviation of 2.7 |J.g/L. Then, 68% of all the individual measurements can be expected to lie within the range of 12.3 ± 2.7 |J.g/L, or between 9.6 and 15.0 |J.g/L. Two standard deviations will include about 95% of all the values.

The larger the standard deviation, the greater the scatter in the data, in other words, the poorer the precision, or reproducibility, of repeated measurements. The standard deviation and, hence, the measurement reproducibility, depend strongly on the matrix in which the analyte is being measured. Stream and groundwater samples can be measured more precisely than wastewater and soil samples. For this reason, it is not possible to make general statements about the reliability of measurements based only on the concentration values. The sample matrix is taken into account by determining the standard deviation.

In Figure 6.10, different types of measurement limits are defined in terms of the reported concentration expressed as the number of standarddeviationsabovetheinstrumentalzero.Only the method detection limit (MDL) is rigorously defined (40 CFR 136, Appendix B). A calculation of the MDL puts it about 3 standard deviations above the instrumental zero. The estimated

FIGURE 6.10 Statistical measures of data reliability, based on the measured concentration expressed in standard deviations above the instrumental zero. The standard deviation is determined by repeated measurements of a laboratory control sample. Other names for EQL are PQL (practical quantitation limit), LLQ (lower limit of quantitation), and LOQ (limit of quantitation).

FIGURE 6.10 Statistical measures of data reliability, based on the measured concentration expressed in standard deviations above the instrumental zero. The standard deviation is determined by repeated measurements of a laboratory control sample. Other names for EQL are PQL (practical quantitation limit), LLQ (lower limit of quantitation), and LOQ (limit of quantitation).

quantitation limit (EQL) is an estimation of the lowest concentration above which reliable quantitative results can be obtained under routine laboratory conditions, and it is arrived at by evaluating the performance of different laboratories. If performance data are not available, the EQL can be estimated from the MDL. The EPA believes that setting the EQL at 5 to 10 times the MDL is generally a fair expectation for routine operation at most qualified government and commercial laboratories.36 Many laboratories determine their own "routine performance" EQL and generally use 3 to 4 times the MDL as the lowest level for an EQL. The method quantitation limit (MQL) is an estimation of the lowest concentration above which reliable quantitative results can be obtained under research laboratory conditions. The MQL can be measured but is often estimated to be about 5 standard deviations above the instrumental zero. In addition, many laboratories include a reporting limit (RL) in their analytical reports, which represents the concentration above which that laboratory has confidence in its quantitative values. Depending on the laboratory, the RL for an analyte may be anywhere between its MDL and its EQL.

Rule of Thumb

When an "official" EQL has not been designated for a particular analysis, an unofficial EQL of 10 standard deviations above the instrument zero (or 3.3 times the MDL) is often used.

Project Management Made Easy

Project Management Made Easy

What you need to know about… Project Management Made Easy! Project management consists of more than just a large building project and can encompass small projects as well. No matter what the size of your project, you need to have some sort of project management. How you manage your project has everything to do with its outcome.

Get My Free Ebook


Post a comment