Links to January '98 SEA FTIR data Summaries: North FTIR. . . . South FTIR
For the first time since we started data disks from the refinery, we have reason to believe that the disk we were sent for January does not contain all of the data generated by the FTIR monitors during that month.
The data files in question were generated during the morning hours of Jan 31, 1998. There is no indication that the missing files are the result of "beam block" conditions or any other type of instrument "down-time". In the case of instrument down time, (any time the FTIR instrument is not collecting data) no new file names are generated. This is by design.
On the morning of 1/31 we can see that there were 17 FTIR spectra generated (representing 17 consecutive 5-minute periods) which were not included on the disk we recieved for January. We discovered the missing data files during our routine analysis of this data, we received no prior information, or subsequent explanation from the refinery, as to why these particular files might be missing.
|Date||Time||FTIR File Name|
|Jan 31||09:00||TON01611 (where is 01612, etc.?)|
After being off-line for repairs since mid-December, the North FTIR unit was placed back in service on the afternoon of 1/23. A review of Toscos databases reveals that by early the next morning this monitor was reporting an ongoing release of MTBE. This "excursion" of MTBE is seen to continue for several hours, reaching its peak above 300ppb just before 4:00am.
It is troubling that this release event did not make it into Toscos monthly report for January. In their "FTIR Summary Report" for January, Tosco indicates that there were zero "hits" for MTBE during the month. This same report also lists 39 "false positive" hits for this compound.
SEAs subsequent analysis of the raw spectral data generated by the North FTIR monitor indicates that the detections of MTBE originally reported in real-time were in fact legitimate. We found no evidence to support Toscos contention that these hits of MTBE represented "false positive" reporting.
The illustration at the left shows two FTIR "traces". The upper
trace is a portion of the absorbance spectrum we created from Toscos
raw spectral file from 3:53AM on 1/24. The lower trace is the "library
reference" spectrum for MTBE. It is difficult to imagine how Toscos
consultants could have failed to confirm the presence of MTBE in the upper
The next two graphs chart
the respective results of SEAs analysis and the real-time FTIR results
as reported in Toscos databases. The results are seen as being very
similar. As we have seen in previous months, because we are able to obtain
better (lower) detection limits in our post-analysis than Tosco receives
in real-time, we end up reporting a greater number of low-level "hits".
Please notice that for the reported hits above 0.1ppm, the two charts
are virtually identical. The one notable difference is that Toscos
real time concentration figures are consistently slightly higher (approximately
8% higher) than our own.
For the first few months that the fenceline system was on-line, the refinerys contractor was reporting something like 50 "hits" of MTBE each month. A year later, the same contractors are consistently reporting 0 "hits". On its face, this would seem like good news.
Unfortunately, we find no indication that MTBE levels have subsided from
our own analysis of the raw data output by the (North) FTIR monitor. We
see no evidence of a decrease either in the total number of "hits"
or in reported concentrations.
A look at the average minimum detection limits (MDLs) for MTBE being reported by the refinerys contractors may help to clarify the situation. Notice that the MDLs for July 98 are fully twice as high as what they were reporting in July '97.
Once again, even though we are behind in our own reports, we can find
no reason why these MDLs have worsened in this fashion. Notice that our
own MDLs have remained relatively consistent by comparison.
Links to January '98 FTIR data Summaries