blanks; 4) calibration standards; 5) analytical surrogates; and 6) laboratory and field 
replicates. The objective of this performance-based approach is to assist the 
laboratories in meeting desired Data Quality Objectives (DQOs) as defined in the EMAP 
Quality Assurance Project Plan (U.S. EPA 2001). 
A measure of whether the analytical procedure is sufficient to detect the analytes 
at environmental levels of concern is the Method Detection Limits (MDLs). Approved 
laboratories were expected to perform in general agreement with the target MDLs 
presented for NCA analytes (Table A7-2 in U.S. EPA 2001). Because of analytical 
uncertainties close to the MDL, there is greater confidence with concentrations above 
the Reporting Limit (RL), which is the concentration of a substance in a matrix that can 
be reliably quantified during routine laboratory operations. Typically, RLs are 3-5 times 
the MDL. In these analyses, concentrations between the MDL and the RL were 
included in the calculation of the means or cumulative distribution functions (CDFs), 
while values below the MDL were set to zero. 
One measure of accuracy of the analytical procedure is the “relative accuracy,” 
which is based on computing the percent deviation of the laboratory’s value from the 
true or “accepted” values in CRMs, LCMs, or SRMs. The requirements for PAHs, 
PCBs, and pesticides are that the “Lab’s value should be within ± 30% of true value on 
average for all analytes, not to exceed ± 35% of true value for more than 30% of 
individual analytes” (U.S. EPA 2001). For metals and other inorganic compounds, the 
laboratory's value for each analyte should be within ± 20% of the true value of the CRM, 
LCM, or SRM. Another measure of accuracy is the percent recovery from matrix 
spikes. High percent recoveries in matrix spikes indicate that the analytical method and 
instruments can adequately quantify the analyte but do not evaluate the ability of the 
analytical procedure to extract the compound from natural tissue or sediment matrices. 
Measures of precision are the “relative percent differences” (RPD) or coefficient of 
variation (CV) of replicate samples, with the objective that the RPD or CV should be 
<30%. 
A post-analysis assessment of the success of the analytical laboratories in 
meeting NCA QA/QC requirements was conducted by the QA manager of the Western 
Ecology Division. The percent recovery from certified/standard materials, recovery from 
matrix spikes, and the average RPD for non-zero sample replicates and matrix spikes 
are given in Appendix Tables 3a - 3c and summarized here. 
2.5.2 Metals in Sediments 
The recommended MDL (Table A7-2 in U.S. EPA, 2001) varies by metal, ranging 
from 0.01 pg/g for mercury to 1500 pg/g for aluminum. The MDLs for metals in 
sediment were met by each state with the following exceptions. Oregon had a MDL for 
antimony of 0.3 pg/g versus the recommended detection limit of 0.2 pg/g. Washington 
had a MDL for selenium of 0.84 pg/g versus the recommended detection limit of 0.1 
pg/g. Oregon had a MDL for tin of 0.5 pg/g versus the recommended detection limit of 
13 
