Volatile Organics Data Validation
Data validation for the new canister method is still under development. For 1999 and 2000, a two step process was used and will hopefully be automated in the future. The analyst is the first step and checks every chromatogram and peak for proper shape and ion profile.
Once the data set is complete for a quarter the values are compared to other samples at the same site and other sites of similar character. This step is performed by someone other than the analyst. Any apparent anomalies are sent back to the analyst for confirmation. The sampler's serial number is checked to look up the results of the cleanliness test prior to deployment. Field notes are also checked for comments that may show a reason for the anomaly. If no explanation can be found, the data is left in the database.
During 1999, a sampler was found to be contaminated with 1,1,1 Trichloroethane. The data was voided for each of the sites where this sampler was in operation. This loss of data is shown in the data completeness table.
The canister sampling and cryogenic analysis have shown a remarkable improvement in precision from the previous tube sample based monitoring. The confidence in the system has grown and the precision results listed show an excellent agreement between the replicate analysis of the same can. Collocated samplers will be instituted in the future to test the sampling systems for any bias.
Differences in concentrations of some compounds from collocated samples have been noted in the PAMS program. The different sample handling (Nafion dryer) and realtime analysis and Flame Ionization Detector (FID) versus Ion Trap MS are significant factors that make the resolution of these differences difficult. A difference in field samplers has also been noticed and our evaluation of sampling systems is still ongoing. The plan to collocate a very different type of canister sampler is very important to insure field sampler problems are noted quickly and resolved.