Appendix A. Quality Control and Quality Assurance of Water Samples


In the present study, 17 percent of the 2007 microcystin samples and 20 percent of the 2008 and 2009 microcystin samples were collected for quality assurance. Each of the quality-assurance sample types, split samples and method replicate samples, were collected on alternate sample weeks in 2007 and 2009, but only method replicate samples were collected for microcystin analysis in 2008. Split samples were collected by dividing a single (composite) volume of lake water with a churn splitter, and were used to determine the variability in the laboratory method for microcystin analysis. Replicate samples were collected twice in rapid succession from the same location and analyzed to determine variability in the sample environment and analytical method. Quality-assurance samples were filtered, as were the primary samples, to create dissolved, small particulate (1.5–63 µm), and large particulate (> 63 µm) fractions, and the fractions were analyzed separately for microcystin concentrations. Therefore, the different fractions presented in table A1 within each year and each quality-assurance sample type are from the same environmental samples (for example, two split samples and three replicate samples were collected in 2007, and those samples were each divided into three fractions). 


Results of quality-assurance sampling indicate that the variability in concentrations attributable to analysis and sampling methods was less than seasonal variability in each sample year (table A1). High median percent differences between the primary and quality-assurance (split or replicate) samples were generally observed when microcystin concentrations were near the method detection limit. This occurred with the small particulate fractions of quality-assurance samples collected in 2007 and 2008; in 2009, 71 percent of the small particulate fractions of quality-assurance samples contained microcystin concentrations less than the method detection limit, which is why table A1 shows 0 median percent difference in samples collected that year. Of the 3 years, samples collected in 2008 exhibited the highest variability within the dissolved and particulate fractions. The relative percent differences between primary and split or replicate samples of chlorophyll a and phaeophytin a also were generally higher in 2008 than in 2009 or 2006 (chlorophyll a data collected in 2007 were not reported), in that more of the quality-assurance samples collected that year varied by 30 percent or more (Kannarr and others, 2010; Lindenberg and others, 2009). Similar results were obtained in quality-assurance analyses of total nutrient samples between 2006 and 2009. This suggests that much of the variability in microcystin quality-assurance samples collected in 2008 may be due to increased spatial variability (manifested as temporal variability at the sampling site) in the phytoplankton bloom that year and lower overall microcystin concentrations relative to 2007 and 2009. 


Median percent differences in large particulate microcystin concentrations between the primary and quality-assurance samples were greater than 20 percent when measured as micrograms per liter in all years. However, with the exception of the 2009 split samples, when these concentrations were expressed on the basis of mass per dry weight of suspended solids, the relative percent differences were lower by more than 10 percent. These values were obtained from the same environmental samples, so the observed concentration variability may be a result of the method used for measuring microcystin concentrations within the solid phase of the water samples (that is, in measuring suspended material concentrations). In 2009, a high median percent difference was observed in the large particulate fraction of split samples reported as either micrograms per liter or micrograms per gram. Relative percent differences between split and primary samples were greater than 20 percent that year on July 21 (85 percent), July 27 (28 percent), and August 24 (27 percent) when the phytoplankton density was very low (as indicated by chlorophyll a concentrations and observations of field crew); microcystin concentrations were well above the detection limit. It is, therefore, likely that the concentration variability in these samples was due to the low volume of biomass collected. 


Table A1 is available in a Microsoft© Excel workbook, which can be downloaded from http://pubs.usgs.gov/sir/2012/5069/.


First posted May 30, 2012