SAMPLE PREPARATION, ANALYSIS, AND QUALITY ASSURANCE/QUALITY CONTROL

Methods of sample preparation and analysis are only briefly presented below. Detailed discussions for water, plant, and sediment samples can be found in Arbogast , 1996; Gough and Crock, 1997; Holmes, 1998; Lichte and others, 1986; and Orem and others, 1999.

In the laboratory, sawgrass was removed from the sample bags, placed in Teflon beakers, submerged and rinsed in deionized water, and drained. This process was repeated at least three times. Plant material was then placed in plastic colanders, rinsed briefly with deionized water, and allowed to drip drain. Colanders were then placed directly into ovens and the material was dried for 24 hr. at about 40^oC. This temperature is near the maximum summer ambient field temperature and should not result in any important loss of Hg through volatilization. Samples were then ground in a Wiley mill to pass a 2-mm (10-mesh) sieve. Splits of the ground plant material were ashed at 450-500^oC over an 18 hr. period and ash yield was determined.

In order to insure adequate material and sample type for the various analyses conducted, replicate cores were commonly extracted from each study site. The cores used for the geochronology studies (²¹⁰Pb analyses) and pore water chemistry, were sectioned (extracted) at 2 cm intervals whereas the cores used in the trace metal geochemical studies were sectioned at 5 cm intervals. Because most core material below about 40 cm was several hundred years old, the interval for sectioning commonly increased to 5 or 10 cm for all cores. This was performed in order to economize on the total number of samples being analyzed.

For element analyses, subsets of the sediment core sections were dried, ground, and ashed in a manner similar to the plant samples (Arbogast, 1990; Balistrieri and others, 1995). For details on the ²¹⁰Pb sediment dating method see Holmes (1996; 1998).

One hundred milligrams of plant and sediment sample ash was digested with mixed acids. After complete digestion of the ash, 40 major and trace elements were determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) (Lichte and others, 1987). Mercury was determined directly on a subset of the dried, ground, unashed plant and sediment material by cold vapor atomic absorption spectrometry (AAS)(Kennedy and Crock, 1987). Total sulfur was determined in plant samples only on 250 mg of the ground material by combustion at 1370^oC in an oxygen atmosphere with infrared detection of evolved SO₂ (Jackson and other, 1985). Water samples were analyzed by inductively coupled plasma-mass spectrometry (ICP-MS) (Meier and others, 1994; Arbogast, 1996).

Table 2 lists the approximate lower limit of detection (LLD) for the concentration of elements in samples of water, plants, and sediments. In table 2 the LLD is defined as the lowest concentration level of the analyte that can be determined to be statistically different from the analytical blank–this approximately corresponds to a confidence level of 90 percent or 3 sigma above the measured average blank (Arbogast, 1996). The element analyses (except for Hg) for the sediment material were all performed in a non-government contract laboratory. Analyses for plant material and water were performed by the Denver Laboratories of the USGS.

Blind standard reference materials were submitted to the laboratories as part of each suite of samples. This included material from the National Institute of Standards and Technology (NIST), the National Bureau of Standards (NBS), and from internal USGS prepared materials. In addition, some of the material was sampled twice in the field (identified by a "Y"), and split in the laboratory for duplicate analysis (identified by an "X"). Results are presented for the duplicate samples and duplicate analyses in the data tables that follow. Quality assurance (QA) and control (QC) practices, for most of the analytical methods used, are provided in more detail in Arbogast (1996).

The QC evaluation for the samples submitted was as follows:

• Blind NIST (NBS) and/or USGS internal reference materials (labeled as SRM or SAR-M, respectively) were submitted for every forty (or fewer) samples.
• Reported laboratory values are considered accurate if their reported value is ±20% of the "target" value of the blind reference material.
• The reported laboratory percent RSD is also examined; however, this value can vary greatly between "jobs" because of matrix effects, the number of analyses near the LLD, and the quantity of sample material available.

The results of all analyses of the SRM's are summarized in tables 3, 4, and 5. Results are compared to SRM certified and/or non-certified concentration values. In addition to the laboratory QA/QC procedures, the field study quality control included submission of procedural blanks for the water samples and periodic splits of sediment and plant samples. All samples, blanks, and splits were analyzed in a randomized sequence relative to their duplicate, their geographic location, and their order of field collection.

The concentrations of elements in the procedural blanks for water samples are generally below detection limits. The laboratory analysis of duplicated splits of sediment and plant samples indicate very good reproducibility (precision). In general, there is also good agreement between our laboratory analysis values and the certified values reported for NIST, NBS, and USGS SRM's. The NIST SRM's list non-certified values for many elements; some of these values are reported in the data tables.

For the element analyses in sediment material, all analyses fell within the ±20% criterion except for As (+36%), Cd (+21%), and Cr (-24%). The poor QC for As and Cd is attributed to a majority of values at or near the 10 and 2 ppm LLD, respectively. The poor QC for Cr appears to be a laboratory bias that may be attributed to the analytical matrix.

Next: Results and Discussion

Return to Publications Page