relatively small, indicating that the total PAH data from both Washington and California 
can be used quantitatively. The Oregon results are more problematic, as they had a 
greater difference between the measured and true values (43% vs. requirement of 30%) 
and 47% of the PAH analytes deviated by more than +35% from the true value. Also, for 
13 of the 22 PAH compounds, the CV from the replicate reference samples was >30%. 
Because of these deviations with both accuracy and precision, the total PAH data from 
Oregon needs to be interpreted cautiously. 
The QA results for sediment PCBs are summarized in Appendix Table 1.3. 
California met the requirements of deviation from the reference materials and the 
percent recovery of the matrix spikes. California did not have any duplicate non-zero 
reference values so it is not possible to evaluate this measure of precision. Washington 
slightly exceeded the requirements for the average deviation from reference materials 
(32% vs. 30%) and the percentage of analytes within +35% of the true value (67% vs. 
70%). Washington did meet the requirements for the percent recovery of matrix spikes 
and the RFP for duplicate samples. Because of these deviations, the Washington PCB 
results should be used with qualified caution. As with the PAHs, the PCB results for 
Oregon are problematic. The deviation from reference materials was 146% and only 
38% of the PCB congeners were within +35% of the true value. Because of the 
problems with accuracy, the total PCB data are best used qualitatively to identify 
locations with sediment PCBs. The subset of congeners that met the requirements for 
both accuracy and precision (PCB 28, 105, 110, 118, and 153) can be used to quantify 
differences in PCB concentrations among sites. 
The QA results for sediment DDTs and other chlorinated pesticides are 
summarized in Appendix Table 1.4. For California, LCMs were only analyzed for two of 
the DDT compounds (4,4’-DDD and 4,4-DDE) though all the pesticides were analyzed 
using recovery from spiked sediments. In the absence of certified pesticide 
concentrations in a sediment matrix with the complete suite of pesticides, the excellent 
recovery of matrix spiked pesticides will have to suffice as indirect evidence that the 
methods employed by California yield results that meet the requirements for accuracy. 
Washington met all the requirements for the chlorinated pesticides (Appendix Table 1.4) 
though several of the individual pesticides showed deviations of up to 72% in the spiked 
blanks. Because of these deviations with the spiked blanks, the Washington pesticide 
data should be used with qualified caution. Oregon had problems with the analytical 
surrogate coeluting with hexachlorobenzene (HCB), which resulted in a large average 
deviation from the reference material (127%). Excluding HCB reduced the average 
extent of deviation from the reference material (57%) but it still did not meet the QA 
requirement of 30%, although three DDT compounds (2,4’-DDD, 4,4’-DDD, and 4,4’- 
DDE) and alpha-chlordane were quantified within 35% of the reference values. Overall, 
the poor performance with the reference materials indicates that the Oregon pesticide 
results are best used qualitatively to identify locations with sediment pesticides, with the 
exception of the four compounds that quantified within 35% of the reference 
concentrations. 
11 
