Brian Altonen, MPH, MS

[Note: The full report filed for the grant-funded portion of this work appears on this site at: https://brianaltonenmph.com/3-gis-environmental-health/report-for-grant-funded-research-2002/]

HIGH RISK SITES

How do we define “High Risk”?

Totals Reports for Toxic Release Sites (TRIs)

Pie Chart of results of Analyzing Chemical Release histories for 502 Confirmed Release Inventory (CRI) sites

The Chemistry of some of the Astoria Sites

Following a review of Superfund Sites, Toxic Release Sites (TRIs) and Confirmed Release Sites (CRIs), we can winnow this research down further in order to define a way to identify the most important sites to consider for spatial analysis aside from Superfund and Superfund Applicant/CRI sites. The reason we do this is simple–an assessment of jsut superfund and superfund applicant sites will no doubt exclude many of the most important sites in need of consideration when producing exposure risk maps.

For this reason, at one point in this analysis I developed a series of formulas used to define the toxicity and carcinogenicity of chemical release sites. These formulas incorporated parts of previously published formulas for determining risk and release site profiling tools produced by CERCLA/CERCLIS, EPA, NIH, AMA, CDC, American Cancer Society, IARC-CRG, etc. Their purpose was to be able to assign a relative risk value to each site based on its history, chemistry, and “seriousness” and “complexity” as a chemical release site.

This analyses of each site enabled me to produce a fairly useful algorithm for identifying the worst chemical sites for the state of Oregon. Following a rather unsophisticated, detailed review of all site releases, I came up with the following findings.

High Risk Sites are non-Superfund Sites that contain any of the following features:

Polycyclic Aromatic Hydrocarbons (PAHs), Aromatic (Benzen-containing) Molecules, Metals (self-explained), Petrols (oil/gasoline products and derivatives), agritoxins (pesticides and herbicides like paraquat and other cyclic/quaternium compounds), halogenic alipathics (non-ring bearers with highly reactive Chloride, Bromide or sometimes Iodide attachments), and halogenic aromatics (benzene 6C+ ring compounds with the above) all have chemical atomic and molecular bond features which make them qualify as potential carcinogens, co-carcinogens, or co-reactive agents capable of enhancing the likelihood of carcinogensis and/or teratogenesis from taking place, should exposure, contact, uptake, and intoxication occur.

A review of all of these features for a chemical release site resulted in the following array of data points on a graph depicting numbers of reports for a specific chemical group in relation to total numbers of reports filed by each site.

Site data, per site, per chemical, numbers of reports versus number of chemicals

The relationship between numbers of reports filed and numbers of chemicals for which reports were filed varies significantly with regard to how these industries are catagorized. Some types of sites like dry cleaning sites having the same few chemicals assessed (two sets of alkyl chlorides). Wood processing and metal industry sites have a great deal more testing done for a variety of chemicals (see late SIC Reclass analysis for more). Whereas we might first think that dumpsites and landfill sites will have similar chemical histories, we find instead that these two sites are quite different in terms of their chemistry based on reports filed with EPA.

The linear relationship between numbers of sites per SIC type and numbers of reports filed

When we look at the numbers of sites in relationship to number of reports filed and numbers of chemicals tested for each site, on a SIC-by-SIC/site by site basis for the 22 high risk sites, there is a fairly distinct linear relationship between these two features. This suggests that for the 22 sites selected for review, none of them really deviate too much from the norm, and so the same analysis can be performed on each site to try and detail its relationship to local cancer cases.

Numbers of reports, Numbers of Chemicals filed per Site, by Industry Type

In terms of industry type derived from SIC data, we find the sites have different relationships with regard to their frequency in the TRI/CRI listing and the numbers of reports typically produced as each site is evaluated for CRI classification. In some cases, industries are fairly rare statewide and have just a few toxins that have to be evaluated, like the battery reccyling centers and paper manufacturing facilities. Other times, sites have fairly unpredictable numbers of reports that have to be filed due to their varying chemical histories. The Electronics Industry is a fairly infrequent site with a large number of chemicals to be evaluated. Dry Cleaning facilities have primarily two chemicals that have to be tested for (2C alkyl halogens), but frequently have other chemicals evaluated as well in order to qualify for CRI approval.

Research Note: The identification of these sites and their chemical data are available in the datawarehouse where this information is housed, which at the time of this study was http://www.deq.state.or.us/wmc/ECSI/ecsicrl.asp.

Sites of highest risk (TRIs, CRIs, SFA, SF and HR)

Statewide Superfund Sites and High Risk Sites

Review of Chemical Release Data obtained from the Oregon State Toxic Release Site Database managed by Oregon State University (OSU). Portland area sites identified by OSU.

Twenty-two of the sites displayed on the above map were classified as “high risk” and added to the Superfund and Superfund Applicants (if they weren’t already on this list) sites for the remaining reviews of chemical release in the Northwest Portland Sector. Not all of the sites on this list were found to be toxic in the sense that they could be directly related to any cases displayed on the maps. Some areas did show significant areal features in need of more detailed well-localized reviews and small area studies.

The following two projects represent ways in which small area analysis can be performed over large areas using GIS to relate chemical release site location to case location. This step would be taken to define the areas of highest risk, which can then be analyzed in more detail. It therefore serves more as a helpful monitoring tool for use in dealing with the bulk of the information one is presented with when trying to document toxic release site release. As much time was taken developing the database for this site, and at times considerably more time with regard to developing the databases needed to investigate the characteristics of each of the chemicals and chemical groups reported for each site.

General Description of the 22 High Risks Sites

Superfund/Superfund Applicant Sites in relation to Population Density in the Portland Area (Site names are for Superfund or Superfund Applicant Sites)

Bar Charts depicting the chemicals, by group, released by High Risk Sites in Northwest Portland

The Application of High Risk Defining Features to Northwest Portland Sites (Pie Chart = HR Features; # = Count of Reports Filed)

Why develop a method to identify High Risk sites?

The main reason for identifying and defining high risk sites is that, based on the EPA/ECSI data and reviews of the site and land use history for each of these areas, these sites have a significant chemical release history in spite of their exclusion from the Superfund sites lists. The primary reason for this is typically related to how Superfund sites are selected for. It is not unusual for the selection process to be limited to a certain number of sites selected per period of time and for whatever reason these sites were not selected for the limited funding available for sueprfund clean-ups. Another reason some sites may have been excluded pertains to the filing deadlines imposed upon companies and agencies filing for superfund status. Whether or not these applications were filed on time is often indicated somewhere in either the superfund applicants information that at times may be found on the web, or simply because filing late allows for additional time to pass before the clean-up process has to be initiated. This information is also available on the web, primarily with the superfund applicants and at times the ECSI/TRI sites reports.

Like the superfund and superfund applicant sites, the high risk sties tend to be in the vicinity of the northern end of the Willamette valley, in close proximity to Portland. For this reason, studies performed of the Portland urban-suburban region provide some of the more complex examples on how such studies can be performed. Whereas we often are taken in by studies of a select few industry tyles situated of fairly moderate population density regions like the suburban-rural settings typical of some paper and lumber industries investigated over the years in the Pacific Northwest, the investigation of chemical release, exposure and risk-related events in the mdoerate to heavy Portland region makes it even more difficult to develop the way to perform such an analysis. Later portions fo this report define new methods that may be used to evlauate chemical density patterns amongst the population density environment of urban-suburban settings like Northwest Portland.