CORRESPONDENCE_1990-1993 - Laserfiche WebLink

Home Browse Search

STANDARD PROVISIONS AND REPORTING REQUIREMENTS September 1993 FOR WASTE DISCHARGE REQUIREMENTS FOR DISCHARGES REGULATED BY CHAPTER 15 AND/OR PART 258 -10- Upon receiving written approval from the Executive Officer, alternate statistical procedures may be used for determining the significance of analytical results for common laboratory contaminants (i.e., methylene chloride, acetone, diethylhexyl phthalate, and di-n-octyl phthalate). Nevertheless, analytical results involving detection of these analytes in any background or downgradient sample shall be reported and flagged for easy reference by Board staff. For any given data set, the Discharger should proceed sequentially down the list below of statistical analysis methods, followed by the non-statistical method, and use the first method for which the data qualify. If that analysis tentatively indicates the detection of a release, then the Discharger shall implement the retest procedure under Discrete Retest. 1. The Discharger shall use one of the following statistical methods to analyze Constituents of Concern or Monitoring Parameters which exhibit concentrations equal to or exceeding their respective MDL in at least ten percent of the background samples taken during the Reporting Period. Except for pH, which uses a two-tailed approach, the statistical analysis for all constituents and parameters shall be one-tailed (i.e. testing only for statistically significant increase relative to background). If the data are log-normally distributed, then the data shall be transformed, by replacing each data point with the natural log (In) of the data point, prior to performing the statistical test. a. The One-Way Parametric Analysis of Variance (ANOVA), followed by multiple comparisons, shall be used when the pooled background data for the parameter or constituent, obtained during a given sampling period, have not more than 15% of the data below the PQL. This test requires at least four independent samples from each Monitoring Point and Background Monitoring Point during each sampling episode. Prior to analysis, replace all"trace" analytical results with a value halfway between the PQL and the MDL values reported for that sample run, and replace all "non-detect" results with a value equal to half the MDL value reported for that sample run. The ANOVA shall be carried out at the 95% confidence level. Following the ANOVA, the data from each downgradient Monitoring Point shall be tested at a 99% confidence level against the pooled background data. If these multiple comparisons cause the Null Hypothesis (i.e., that there is no release) to be rejected at any Monitoring Point, the Discharger shall conclude that a release is tentatively indicated for that parameter or constituent. b. The One-Way Non-Parametric ANOVA(Kruskal-Wallis Test), followed by multiple comparisons, shall be used when the pooled background data for the parameter or constituent, obtained within a given sampling period, have not more than 50% of the data below the PQL. This method requires at least nine independent samples from each Monitoring Point and Background Monitoring Point; therefore, the Discharger shall anticipate the need for