combination of sampling, measurements, dispersion modeling, and estimates based on emission factors. Because this sampling program was a first of its kind effort, its scope was intentionally broad. Subsequent analysis showed that not all of the information obtained was necessary to identify significant sources and potential reduction options. For the Yorktown Refinery (and the petroleum refining industry overall), more general information, such as source specific VOC emissions, is adequate to identify many of the pollution prevention projects developed in this study. Total Voc emissions are a good indicator of overall emissions and can be used for tracking emissions reduction progress. B. A substantial portion of pollution generated at this refinery is not released to the environment. The release inventory process allowed a comparison of pollutant generation, on-site management and ultimate releases to the environment. The Refinery generates about 27,500 tons/year of pollutants. As a result of site hydrogeology, on-site wastewater treatment, and solid waste recycling practices, about 12,000 tons are recovered, treated or recycled and do not leave the Refinery site. Of the remaining 15,500 tons about 90 percent are released to the air. Figure 2.4 illustrates the transfers which take place between generation and ultimate release. Figure 2.5 characterizes pollutants released from the Refinery. This site-wide analysis of pollutant generation and release characteristics allowed the Workgroup to focus much of the remaining Project resources on the largest releases--airborne emissions. Modeling studies indicated relatively little naturally occurring transfer of hydrocarbon emissions from air into other media (Cohen and Allen, 1991). Most hydrocarbons are not very water soluble, and so are not easily removed from the air by rainfall. Section 2.0 includes a more detailed discussion of the potential for transfer to other media. Although the fate of criteria airborne pollutants (like NOX and S02) was not studied in this Project, they are known to be scavenged by rainfall and can contribute to nitrogen loads and pH changes in lakes and soil (See Appendix B). Measurements and modeling results showed small transfers from some surface water ponds to groundwater. Groundwater also enters the wastewater treatment system through the underground sewers, resulting in a net groundwater inflow. Transfers of pollutants between media do occur, particularly as a result of pollution management activities. Over 370 tons/year of hydrocarbons initially present in wastewater streams are volatilized into air from the water collection system. More than 2,000 tons/year of biosolids are produced by treating wastewater in the Refinery's activated sludge system. C. The TRI database does not adequately characterize releases from this Refinery. Title III of SARA, Emergency Planning and Community Right-to-Know Act, created the Toxic Release Inventory (TRI) in 1986. Title III requires regulated facilities in SIC Code 20-39 to submit annual release data on more than 300 chemicals manufactured, produced or otherwise used in quantities exceeding certain threshold values. Releases to all media must be reported. TRI is one way of focusing corporate attention on release reduction opportunities. The TRI reports are based on either emission estimates, direct measurements or a combination of both methods. Each facility is responsible for the accuracy of the data reported. Industrial facilities frequently file amendments to TRI reports to reflect improvements in the accuracy of the estimation and measurement techniques. The TRI database has become the de facto national release inventory. The quality and utility of data reported can vary widely. At a plant that uses a single solvent to wash manufactured parts, and that purchases extra solvent every year to make up for evaporative losses, the quantity of solvent emissions is well known and tracked through monthly purchasing records. A TRI report which included this solvent and plant should be quite accurate. However, at the Refinery, the TRI does not report total facility emissions because: The TRI is based on estimates rather than measurements. Estimating accuracy varies widely. During the measurement portion of this Project, several new sources were identified whose significance had been previously underestimated. One source was identified which had been overestimated. Figure 2.7 summarizes the results of this analysis. The measurement phase of this Project revealed substantially The TRI focuses on specific chemicals which account for only a portion of the total emissions. In the Refinery's case, the TRI report covers only 9 percent of the total hydrocarbons released, and only 2.4 percent of the total releases to all media. Criteria pollutants--CO, NOX, SO2, and PM-10--are not reportable in the TRI. Some activities and emissions are excluded by EPA from record Finally, TRI provides an approximate inventory of selected materials released to the environment. TRI data by itself does not allow for meaningful risk evaluation or comparisons on a facility basis, because it does not define the facility's relationship to nearby populations and ecosystems. D. Site specific features determined during the facility-wide assessment, affect releases and release management options. National programs, by design, address overall problems in specific media. But these programs seldom consider site-specific differences in developing standards. Other refineries, and indeed other industrial facilities, can use the general sampling approach developed here to obtain the facility-wide release inventory. However, each site will exhibit unique geophysical and process characteristics. Each assessment plan must include these site-specific characteristics in its design and focus. an example, the Yorktown Refinery does not have a hydrofluoric acid (HF) alkylation unit and HF was not measured. HF can pose a significant health risk if managed improperly, and may need to be tracked at facilities that use it. Groundwater: As a result of a clay soil layer, unique hydrogeology, the placement of the underground drainage system relative to the water table, and local climate, groundwater movement at this site is minimal. In fact, the underground drainage system is acting as a groundwater collection unit, sending groundwater to the Refinery's wastewater treatment plant. Thus, groundwater at this site is not leaving the property. Furthermore, sampling showed surprisingly low levels of groundwater contamination, compared to other refineries (LA Times, 1988). Marine Loading Emissions: Yorktown Refinery uses marine transportation for receiving all crude oil and shipping more than 80 percent of its products. Estimated releases from product loading operations are 784 tons/year of VOCs. Computer modeling analysis showed this source had the greatest impact on exposure of nearby residences to Refinery hydrocarbon emissions. Therefore, it would be useful to include marine loading emissions in this facility's environmental management plans. Many other refineries rely more on pipeline, rail and truck shipments to handle crude and products, and would thus not expect to find the same potential impact from marine operations. Airshed Status: As discussed in Appendix A, the Refinery is located in an airshed classified as an attainment area for all criteria pollutants including ozone. Therefore, relatively few hydrocarbon emission controls have been required or installed at this facility. The sampling program and release reduction options focused on hydrocarbon releases. Many other refineries in ozone non-attainment areas have already installed extensive hydrocarbon emission controls. Consequently, other facilities may have a significantly lower percentage of hydrocarbon emissions. Similarly, NOX, CO, PM-10 and SO2 emissions have been more tightly controlled in some other airsheds (such as the Los Angeles basin) which do not meet NAAQS for these pollutants. A. 1.3.2 Release Reduction Options A workshop approach, drawing on a diverse group representing government, industry, academic, environmental and public interests developed a wide range of release reduction options in a multi-media context more quickly than EPA or industry alone would do. The release inventory described in 1.3.1 above, served as the basis for identifying ways to reduce releases. A 3-day brainstorming Workshop, held in Williamsburg, Virginia generated more than 50 potential release reduction options for the Refinery. These ranged from producing a single grade of gasoline to specific technical options for particular equipment or processes. Table 3.1 lists all options identified. The Workgroup subsequently narrowed this list to 12 options for more careful, quantitative analysis. This winnowing process considered only those options that were technically feasible now, offered potentially large release reductions, addressed different environmental media, and posed no process or worker safety problems. Projects designed to comply with several current or anticipated regulations were also included. Table 3.2 lists engineering projects included for further analysis. The Workshop also addressed screening criteria to help prioritize the options, potential barriers and incentives for implementation, and permitting concerns. The diverse viewpoints brought to all these discussions helped guide subsequent Project activities. These views reinforced the Workgroup's desire to consider broader issues such as multi-media release management consequences, future liability impacts, etc. The Workshop was able to consider these issues more comprehensively than either government or industry alone would normally do. B. Release management frequently involves the transfer or conversion of pollutants from one form or medium to another. It is not at all unusual for pollutants to be converted and transferred from one form or media to another as part of a pollution control practice. For example, scrubbers used to remove acidic pollutants from many electric utility stacks generate large volumes of calcium sulfate sludge (EPRI, 1983) which must also be managed. For options developed at the Yorktown Refinery: Modifications of the underground drainage system and process water treatment plant (required under the Benzene Waste Operations NESHAP; Federal Register, 1990) will improve process water treatment and reduce air emissions, but produce more solid waste such as biosolids and fully spent activated carbon. The Refinery has limited sludge processing capacity. Keeping soils out of sewers would reduce the amount of sludge in the API Separator and thus allow for more on-site management of other solid wastes, reducing offsite disposal. Installing an electrostatic precipitator would reduce FCU particulate (PM-10) emissions (catalyst fines), but transfer the additional collected particulates to land disposal. Burning hydrocarbons that cannot be economically recovered generates other criteria pollutants which may also need to be managed. None of these transfers or transformations are bad, in and of themselves. The Project simply pointed out the need to recognize, plan, and manage these changes at an early stage of the release management cycle. C. Source reduction options were more cost-effective than most treatment and disposal alternatives. Nevertheless, source reduction alone was not adequate to achieve all the desired or legally required release reductions. The Workgroup agreed to consider the waste management hierarchy--source reduction, recycling, treatment, and safe disposal--as the basis for developing release reduction options. Technologies identified and analyzed fit into this hierarchy. Time and budget constraints limited technology choices to conventional, proven solutions rather than exploring innovative alternatives. However, less than half the options identified qualified as "source reduction." Had the options been limited to only source\ reduction, the scope of potential opportunities for reducing |