This is the first in a series of reports by the Shoreline Environmental Alliance, in which SEA will be making public the results of our own analysis of the raw data generated by the new fenceline monitoring system at the Tosco SFR refinery in Rodeo, CA., and making observations about the ongoing progress of this landmark system. This report is designed as an attachment to, or supplement for, Toscos monthly reports, and is intended for persons who are genuinely interested in gaining more information about chemical emissions from the Tosco facility, and about the new monitoring system.
At the beginning of October, 1997 the Signatory groups to the Good Neighbor Agreement received the monthly status reports on the fenceline system from Tosco. This report, prepared by Toscos contractor Terra Air Services of Houston, is similar to the reports which we have been receiving over the past year, with two important differences: This August 97 report marks the first time these reports have included data from the UV and TDLS (Tunable Laser) systems. This report also marks the first time which these monthly reports have been delivered to us along with a CD ROM disk which contains all of the raw data which Terra used in preparing the reports.
Access to this type of raw emissions data by community groups is absolutely unprecedented, and will allow us to do a number of things. First, and perhaps most importantly, having the raw spectral data will enable SEA to attempt to verify the accuracy of the concentration figures reported by the refinery. Secondly, in the case of the infrared spectral data from the FTIR monitors, we are able to identify and track the presence of other pollutant gasses which are not being monitored at the refinery. Thirdly, through the process of re-analyzing this raw emissions data, we should be able to make subjective judgments about such issues as relative instrument performance, and whether or not the refinerys contractors are indeed operating the equipment in an optimal fashion.
We encountered no problems reading and analyzing the raw FTIR spectral files provided to us on the August disk. Because of software difficulties, however, we have thus far been unable to read any of the other data from this disk. The emissions databases (the detailed results of Terras analysis) from all three optical remote systems contained on the disk are readable only in Microsoft Access 97 version 8.0. We are currently using Access version 7.0. In practical terms, this means we have been able to analyze the FTIR spectra, but unable to compare our analytical results directly with the "real-time" results arrived at by the refiners contractor. Also, we have been unable to view the databases which contain the raw output from the UV and TDLS systems. The biggest problem this has presented is that, in the absence of summary plots, (like those provided for the FTIR and TDLS systems) we are not able to determine when and how often the emissions reported by the UV system occur during the month.
Beginning with this August Report, SEA will be making available the results of our own analysis of the infrared spectra provided by Tosco, as well as a set of comments on the months data. The fenceline monitoring system itself is extremely complex, and our ongoing communications with Tosco and their contractors regarding the operation of the system are typically at a very detailed level. We will not attempt to go into that level of detail in our comments here, however, we do feel it is important for us to report to the larger community about the general nature of any disagreements we may have with the refinery as regards the operations of the new monitoring system.
The System as it Stands
The fenceline monitoring system at Tosco SFR, as it is operating today, is the most comprehensive monitoring system of its type at an industrial facility in the world, and is without peer in terms of its ability to provide early warning for a wide variety of chemical release scenarios. We are currently receiving information from the refinery about the continuous emissions levels of 39 chemical compounds. Furthermore, the team of consultants at Terra Air Services who are administering the system for Tosco are among the best in their field. Still, we at SEA believe that there are some significant problems both with the physical operation of the system and with the management of the system at this time. The next few paragraphs will deal briefly with some of the issues which we consider to be problematic.
Data Quality \ Quality Assurance Issues
From our perspective, the most glaring problem with the operation of the new system at this time is the apparent lack of any formal Quality Assurance (QA) plan. In the August reports, some of the reported emissions concentrations from the South UV monitor appear not to fit well with the rest of the data set. When asked about this, representatives of both Tosco and Terra Air indicated that these results probably were not reliable; and that the reported concentrations of several chemicals was likely too high. In the reports, however, no mention was made of this lack of confidence in the published emissions data. Apparently, even though refinery management does not believe the monitors are reporting accurately, no attempt has been made to challenge the results of these monitors with any type of QA check.
Standard QA methods vary somewhat between the three different optical remote systems involved here, but all of these methods involve routinely "challenging" the monitors by diverting the beam through a "flow-through" cell which contains a known quantity of a particular (reference) gas. At Tosco SFR, the problem is not that QA procedures are not carried out often enough, but that to date, there has been no QA testing on the new monitors done at all. In the case of the UV monitors, we have learned from the equipment manufacturer that neither Tosco nor Terra Air has even purchased the equipment necessary to perform standard QA checks on their instrument.
Obviously, confidence in the accuracy of the information we receive form the refinery is of utmost concern to us; and we believe that for the system to continue to be operated in the absence any QA guidelines will be completely unacceptable. Our dialog with the refinery and its contractors is open and ongoing, and we remain confident and hopeful that Tosco will quickly adopt and make available a formal QA plan which will be effective and which will meet the criteria laid out in our various agreements.
Raw Data/Database issues
While we will be receiving the raw data output from all three types of optical remote monitors, it is only with the FTIR systems, where infrared spectra are electronically recorded, that additional analysis is possible after-the-fact. For the August set of data, we have been able to successfully analyze all of the FTIR spectra provided to us. Because of the software conflict which has prevented us from being able to read Terras FTIR databases, we have been unable to compare analytical results on a spectra-by-spectra basis. A glance at the result summaries, however, seem to indicate that the concentration figures arrived at by both parties will match very nicely. (see Summary Table)
SEAs analytical results -vs- Terras results
Perhaps the most noteworthy emissions reported by the FTIR systems for August, were the 54 detections (hits) of MTBE which occur toward the beginning and end of the month. For this months data, the primary focus of SEAs analytical work was on attempting to verify these MTBE hits, and attempting to identify other compounds that may have been present when MTBE concentrations were at their peak. Tosco reports 54 hits for MTBE in August with a peak concentration of 0.964ppm and an average detection limit of 0.056ppm. We report 111 hits, with a peak concentration of 0.962ppm and an average detection limit of 0.031ppm. (see Summary Table).
It is very encouraging to see that our peak concentration figures match so closely, especially considering that we used a completely different background (clean air reference) spectrum than the refinerys contractor did. (In fact, we have no idea what they used for a background reference). While we do report more than twice as many hits for MTBE than they do, this does not necessarily signify a discrepancy between the two sets of analytical results. The extra hits can easily be explained by the fact that we were able to achieve lower detection limits, and therefore detect MTBE at levels which went unreported by the refinery. Besides these extra hits of MTBE which we pick up, we are also able to identify the presence of n-hexane, 2-methyl pentane, 3-methyl-1-butene, and propane on various occasions throughout the month, all at low levels (mid-low ppb range).
Because of the constraints inherent in working with FTIR spectra in "real-time", one expects to see this type of improvement in detection limits whenever subsequent (not real time) analysis is performed on the spectra. (Much of this is typically due to the luxury of being able to update background reference files at will from within the data set.) It came as somewhat of a surprise to us that we did not have to update or otherwise change background reference spectra during the month of data in order to preserve good detection limits. The concentration figures we report here were all derived using a single background reference file (#00909 generated on 8\5 @ 11:24am). It is not immediately clear to us then, why it is that our reported detection limits (at least for MTBE and n-hexane) are considerably better than those reported by the refinery.
The Summary Table which immediately follows this report lists the partial results of our analysis of the August FTIR data. Concentration figures are listed for 6 compounds only, and only for the North fenceline. Raw Spectra from the South FTIR was only available from the beginning of the month, apparently before the cryocooler failure there, and some of this data was of poor quality. The available spectra from the South were thoroughly examined, and except for an unknown compound which appears on 8/5 @ 7:41am (see figure 6) and some hits of ammonia, nothing else was detected. Difficulties encountered in managing the Excel databases, which have since been resolved, prevented detailed record keeping of this early data analysis. Similarly, several additional compounds were monitored on the North fenceline (all non-detect data) which did not make it into the Summary Table due to difficulties encountered managing database files.
The 6 compounds reported on here are MTBE, n-hexane, 2-methyl pentane, 3-methyl-1-butene, propane, and isobutylene. With the exception of MTBE and n-hexane, these chemicals are not among those monitored by the refinery. MTBE was chosen both because it is a chemical of concern, and because the reported hits of this chemical on the North fenceline made it a convenient point of comparison. N-hexane was chosen because it was reported as "non-detect" by the refinery in August, after having been detected and measured almost routinely during the earlier phases of the monitoring project. Also, the reported detection limits for this compound seemed unexpectedly high. The other compounds which were chosen either had been identified in earlier spectra from the Tosco site, or were identified through an "unknown search" in the present data set.
As regards the MTBE concentrations encountered at the North fenceline, it is important here to note that the Tosco refinery may not be the source of these emissions. On previous occasions when MTBE has been detected by the North FTIR unit, refinery personnel have indicated to us that the wind direction at the time of the reported detections suggest that the source of the MTBE emissions was actually the Wickland Oil Terminal located directly to the North of the Tosco refinery. At this stage, our project is not considering the implications of meteorological data, as it is not presently within our scope of work to attempt to determine the source of the emissions we are measuring. The information on disk which we received from the refinery for August did not include meteorological data, although we anticipate that subsequent disks will include this information. While incorporating met. data into these emissions studies may make for very useful research in the future, our present study is limited to finding out as much as we can about what is in the air at the refinery fenceline, not necessarily where it came from.
Downtime issues: Stiriling Engine Cryocoolers -vs- Liquid Nitrogen Dewar
The two FTIR monitors at Tosco have been operating now for almost a year. Generally, the "up-time" percentage has been at or above the 90% range. This seems good, especially considering the newness and ambitious nature of this project. August was an exception to this. The South FTIR unit experienced a cryocooler failure, and reported only 30% "up-time" during the month. This is only the last in a series of cryocooler failures and cryocooler-related problems. In fact, it seems that problems with these cooling devices, and difficulty replacing them expeditiously, are responsible for most of the downtime we have seen, not only in the current permanent system, but in the "6-month test" and "interim" phases as well. We at SEA believe that the refinery should consider converting the units to use a liquid-nitrogen dewar, rather than the stiriling engine cryocooler, for the detector cooling functions necessary for their FTIR units.
The FTIR devices are available form the manufacturer in two versions: either with or without the stiriling engine cryocooler. The stiriling engine option was designed to make the units more easily portable, or for those situations where the use of liquid nitrogen was deemed impractical. There are two main downsides to choosing the stiriling engine option. First, they have a usable life span of around 6 months, and are extremely expensive and difficult to replace. (As a practical matter, few of these units have been lasting anywhere near 6 months). Secondly, they generate noise in the infrared spectra which can adversely affect detection limits (although the manufacturer now claims to have alleviated this problem).
In permanent industrial applications, most facilities have chosen to use the liquid nitrogen dewar as a less-expensive and more reliable method for cooling the MCT detectors in their FTIR units. We have never understood the refinerys original decision to opt for the stiriling engines. Considering that there is now liquid nitrogen present and in use at both fenceline monitor installations for the other optical-remote systems, it seems to make very little sense to continue with the stiriling engine cryocoolers at this stage. We believe that converting the FTIR units to use liquid nitrogen would both save the refinery money, and make the units more dependable. The refinery has indicated they are willing to consider this change, but want to try the current platform a little longer.
The UV monitors
The ultraviolet monitors in place on the fencelines as Tosco are manufactured by Sci-Tech from Saskatchewan. This installation marks the first ever sale of these monitors for this company. The units have been on-line since June 97, and while they are apparently working fairly well, there have been some problems; both with the monitors themselves, and with integrating them into the overall fenceline monitoring system. One significant problem now seems to be a software glitch that is causing the monitors to report a greater "off-line percentage" than is actually the case. There also may be a problem with the quality of the data which these monitors are producing. In response to questions about some of the concentrations reported by the Southern UV monitor in August, representatives from both Tosco and Terra have indicated that the reported concentrations may be too high, and that the data may be unreliable.
The emissions activity reported by the UV monitors will continue to be of special interest to us, as the chemical compounds being monitored by these systems include some of those which we are most concerned about. This is especially true of the chemicals carbon disulfide, benzene, and toluene. These chemicals, all known to be extremely hazardous to humans, were all detected at the fenceline at low levels during either the "6-month test program" or the subsequent "FTIR interim program".
A review of the "UV Summary Report" on the last page August report reveals that of the 8 compounds monitored by this system; all 8 were detected on the North, and 7 of 8 were detected on the South fenceline during the month. In all cases, the detected levels were well below established alarm levels. It comes as a surprise to see that some of the chemicals, notably carbon disulfide and sulfur dioxide were recorded at much higher levels on the South fenceline. As this is the predominantly "upwind" fenceline for the refinery, we had expected to see lower emissions levels here than on the North. The reported peak for carbon disulfide on the South fenceline during August was .226ppm. The low-level (first stage) alarm for this compound is set at 1ppm. This appears to be the closest we have come to an alarm level detection thus far in the program. These readings from the South UV system in August are the very ones which the refinery indicates may not be reliable.
The August report came to us without a "summary plot" for the UV system results. Because of this, we have no way of knowing when, or how often, the emissions reported by the UV monitors were detected. Also, because of the aforementioned software compatibility problem, we have been unable to discern this information from the databases provided to us on disk. Due to these problems, we have thus far been unable to accomplish one of our goals, which is to attempt to determine if a correlation exists between emissions recorded by the new fenceline system, and odor complaints filed by members of the surrounding community.
As was discussed earlier in this report, no formal, or even informal Quality Assurance measures have been undertaken, either then or since, to validate the analytical results of any component of the fenceline system. We do not understand this. For the refinery to report on the concentrations of hazardous chemicals being released from their facility, then inform us that the reports may be unreliable, without even attempting to validate the data through conventional means is clearly unacceptable.
We believe that there are some basic steps, short of a formal QA schedule, which if followed, could give the refinery a very good idea if the numbers being reported by their new monitors are accurate. For instance; when the UV monitors on the South fenceline are reporting substantial concentrations of compounds like carbon disulfide or sulfur dioxide, as they did in August, why not take a conventional canister sample of the air somewhere along the monitor beampath during the same time period and have it analyzed for comparison? This is the type of basic QA step that we believe must become part of the standard operation of the fenceline system, and which must become part of the monthly reports.
The TDLAS Monitors
TDLAS stands for Tunable Diode Laser * System. The units at Tosco were manufactured by Boreal. Like the UV systems, these units have been on-line since June 97, and like the UV units, this is the first-ever commercial installation of these devices. These devices are set up to monitor two compounds only: hydrogen sulfide and ammonia. There are actually separate devices for each compound, four altogether on the two fencelines.
Of the three open-path systems, we know least about these TDLAS monitors. According to Tosco and its contractors, the units appear to be working very well. The August report indicates that neither of the H2S monitors detected anything over the course of the month. Both of the ammonia TDLS monitors did report low levels of ammonia during the month. ammonia is also monitored by the FTIR systems, and because it is the only compound which is monitored on two systems, it will be interesting to compare reported concentrations of this gas between the two systems.
A review of the August report indicates that there does not appear to be terrific correlation in the ammonia concentrations reported by the FTIR and TDLAS systems. For example, the North FTIR reports a peak concentration of 0.053ppm, while the North TDLAS reports a peak of 0.905ppm (Engineers at Terra inform us that some of the low-level "hits" on the TDLAS ((which go unreported because they dont meet "data quality parameters")) actually do compare very well with simultaneous FTIR hits; we should be able to check this ourselves once we have updated our software.) In any case, it does appear that the FTIR equipment is able to detect ammonia at lower levels than the TDLS. This runs contrary to what we had been told. It now appears that the sensitivity of the TDLS systems had been somewhat overstated prior to their installation.
Our agreement with the refinery does allow for them to discontinue monitoring for ammonia with the TDLAS monitors and use solely the FTIR units for this compound if they choose to do so. In the coming months, we will be looking closely at how the various concentration figures for ammonia compare between the two systems, and our findings will become a regular part of this monthly report.
SEA's data summary chart for 8-97
Attachments to SEA's August '97 report
SEA does not have a toxicologist on-staff. Except in such cases where a reported emission concentration represents a violation of Federal, State, or local standards, WE WILL NOT BE COMMENTING on the relative risks of any reported release.
By making these reports publicly available, both those emissions figures reported by SEA, and by TOSCO, we do not wish to imply that any of these chemical emissions pose ANY HEALTH THREAT WHATSOEVER to people in the surrounding communities.
SEA member Andy Mechling, who has been retained by SEA to re-analyze the raw emissions data from Toscos FTIR monitors, holds no formal scientific credentials. He holds Bachelors degrees in History and Communications. He has undertaken training by the Environmental Technologies Group, the manufacturer of the FTIR instruments, and now has more than two years experience producing and analyzing this type of data for a number of community groups in a wide variety of situations. SEA has every confidence in his abilities to accurately analyze and report this data. Still, considering of our staffs lack of scientific credentials, any chemical concentration figures reported by SEA should be regarded as our BEST ATTEMPT to arrive at the truth, rather than the ultimate truth regarding chemical emissions from the Tosco refinery.
Because of the nature of infrared spectroscopy, where the raw spectral data is saved electronically, all reported concentrations of chemical emissions are ultimately verifiable. SEA will make the raw spectra which make up our reports available upon request to any party for any reason. Furthermore, in the event that there is a discrepancy between our analytical results and those arrived at by Toscos FTIR contractor, SEA will make every attempt to notify the refinery, and to rectify the discrepancy BEFORE making any results publicly available.
We are undertaking this project, and making these reports public, because we believe that residents in our communities have a fundamental right to know as much as possible about the air that surrounds them. It would be much preferable to us if this task were being undertaken instead by one of the regulatory agencies, who possess infinitely better resources than our very small organization. Unfortunately, this is not the case at all, and the job is left for us to do. Simply put, we are making this data public in this fashion because we are presently the ONLY AVAILABLE SOURCE for this type of chemical emissions data.