ARCHIVED REPORTS_2014 - Laserfiche WebLink

Home Browse Search

V APPENDIX G STATISTICAL ANALYSIS METHODS FOR THE FORWARD LANDFILL In order to update concentration limits (CLs) for the Forward Landfill, a statistical analysis of water quality monitoring data was performed by Geo-Logic Associates (GLA) in accordance with protocols outlined in Order No. R5-2014-0006 issued by the California Regional Water Quality Control Board — Central Valley Region and Code of Federal Regulations, Title 40, Part 258 (40 CFR 258) and California Code of Regulations (CCR) Title 27, Section 20415. The statistical analyses were performed with historical statistical algorithms that recognize that groundwater quality data may be correlated through time (in contrast, non-historical statistics rely on the data not being correlated). Geo-Logic Associates performed the historical statistical analyses described below with the Sanitas° statistical program. This software performs user- determined statistical analyses using methods recommended by the California State Water Resources Control Board (SWRCB) and the United States Environmental Protection Agency (EPA). INTRAWELL TOLERANCE LIMITS In accordance with Waste Discharge Order No. 115-2014-0006, intrawell tolerance limits (TLs) were calculated to establish updated CLs for the Forward Landfill. A TL is a one-sided tolerance interval constructed from a background population with a specified confidence level. If a future observation exceeds the calculated upper limit, this is statistically significant evidence of contamination. For the Forward Landfill, intrawell TLs were developed based on a Type I error rate a = 0.01 that future observations will fall below the limit. Statistical calculations are based on the procedures described in the USEPA's 2009 Unified Guidance Document (USEPA, 2009). For a parametric tolerance limit analysis, the mean (X) and standard deviation (S) are calculated for the raw or transformed background data and a one-sided (two-sided for pH) tolerance factor (K) is specified at the 1-a (or 99 percent) confidence level. The prediction limit (TL) is defined as: TL= X+ KS For non-parametric prediction limit analysis, the highest value from the background data set is used to set the prediction limit. The false positive rate for non-parametric analyses were established from California standards and determined based on sample size, with all false positive rates falling below 3.8 percent. The calculated intrawell TLs for the routine indicator parameters for each groundwater monitoring well were calculated using screened background data from the first quarter 2009 through the fourth quarter 2013 (20 sampling events), TL calculations are included as an attachment to this appendix. For constituents that have not been detected in background database, the laboratory's practical quantitation limit (PQL) is used as the CL.