Crossroads The ACM Magazine for Students

Sign In

Association for Computing Machinery

Magazine: Features Mapping organized ignorance in environmental health

Mapping organized ignorance in environmental health

By ,

Full text also available in the ACM Digital Library as PDF | HTML | Digital Edition

Tags: Data management systems, Life and medical sciences

back to top 

The contamination of drinking water by perfluorooctanoic acid (PFOA) in Hoosick Falls, NY, was first discovered in mid-2014. It was not detected as part of a routine water quality test but as a consequence of efforts by resident Michael Hickey to understand why his father had died of kidney cancer with no apparent causal factors. Having a sense that manufacturing plants in Hoosick Falls might be a source of contamination, he began researching the health effects of the chemicals used at the plants. Hickey said it took just five minutes of Googling to find a connection between the PFOA used at the plant and the kidney cancer.

PFOA is a chemical compound that has been used to make nonstick and fire-resistant coatings in products like Teflon pans since the 1940s. Under an agreement reached by the U.S. Environmental Protection Agency (EPA) in 2006, major manufacturers in the U.S. began to phase it out. The agreement did not, however, extend to manufacturers outside the U.S. or to those using PFOA within the U.S. to make other products, as was the case in Hoosick Falls. That Hickey was able to connect the dots so quickly is not difficult to fathom. As the result of a 2001 class-action lawsuit settlement with DuPont, a chemical company that had contaminated ground-water with PFOA near a plant along the Ohio-West Virginia border, a panel was set up to study the effects of exposure to C8, another name for PFOA. The C8 Science Panel charged with this research, following years of studying the effects of exposure on thousands of people working at and living near the DuPont facility, found probable links between PFOA and six negative health outcomes, including kidney cancer. These results were first reported in 2011 and 2012. At the time of Hickey's initial research, they were publicly available online through the National Institute of Environmental Health Sciences. Given what Hickey found online, he sent local water samples to a private lab in Canada to be tested for PFOA. All were reported to be positive.

Since news of the contamination gained broader public visibility in late 2015, several individuals, agencies, and organizations have tried to map the contamination, including the New York State Department of Health and Department of Environmental Conservation, faculty and students at nearby Bennington College, Bennington, VT, and a geomorphologist hired through a grant from the local school district. But they have faced numerous challenges. Some of the primary ways such challenges manifest in public health mapping include the difficulty of relating incommensurate or incomplete datasets and attempting to represent them in meaningful ways. They also include the limitations of current regulatory knowledge infrastructures to reflect complex and dynamic ecological systems, in which water, soil, humans, politics, industrial legacies, and technologies interact. It is important to clarify for the purpose of this discussion that we conceive of mapping in two senses—first, in the traditional sense, of relating existing data points to geospatial units, and second, in a more critical sense of representing data in relation to knowledge-production systems across space, time, and scale in order to draw out the ways knowledge and ignorance are coproduced through these systems.

back to top  Organized Ignorance

Producing information about water quality has become a key strategy of water regulation in the U.S. Since the 1980s, the EPA and other regulatory bodies have turned to information strategies to plan for, regulate, and mitigate environmental disasters, a transition Kim Fortun refers to as "informating environmentalism." [1] As a result, the extent of data collection mandated by law in the U.S. has expanded significantly over the past three decades, and, overall, we know much more about pollution today than we ever did before.

Information strategies as a mode of regulation also have limits. Sociologists Scott Frickel and M. Bess Vincent examined them in their 2007 article. [2] Based on ethnographic work following that disaster, they noted how often experts, following standards for water-quality testing that are reinforced through institutions and regulations, can be oblivious to the complex ecological, social, and historical contexts that affect how information is produced. Water-quality tests are designed to detect certain contaminants at certain concentrations because they are programmed to do so ahead of time. Frickel and Vincent wrote, "We find what we seek, not necessarily what is there." As regulatory actors like the EPA legitimize testing results, certain risks are minimized, producing "organized ignorance."

Attempts to measure the effect of PFOA on Hoosick Falls faced and continue to face similar limitations.

Water-quality tests produce information only where data is collected. The Safe Drinking Water Act (SDWA) of 1974 requires the EPA to produce a list of potential but currently unregulated contaminants every five years in the form of the Contaminant Candidate List. The Unregulated Contaminants Monitoring Program was thus established as part of the 1986 amendments to the SDWA, requiring public water systems be tested for unregulated contaminants to which they were likely exposed. At the time, this testing was required of public water systems serving more than 500 residents. The 1996 amendments to the SDWA revised the program such that it pertained only to those public water systems serving more than 10,000 residents; in so doing, it created the Unregulated Contaminants Monitoring Rule (UCMR) under which the EPA must monitor no more than 30 potential contaminants in a given five-year cycle and outlined that the EPA had to make just five or more determinations about whether or not to regulate certain contaminants listed in the UCMR.

In 2012, under the Third Unregulated Contaminant Monitoring Rule (UCMR3), six perfluorochemicals were to be monitored between 2013 and 2015, including PFOA and PFOS. In addition to public water systems serving more than 10,000 people, this monitoring included 800 representative public water systems serving fewer than 10,000 people. Under UCMR3, less than 5% of public water systems in the U.S. were to be monitored, and no data about unregulated water contaminants in private drinking water wells would be produced [3]. Even leaving aside the larger universe of contaminants not included in the UCMR, but that may unevenly affect particular areas, it is still difficult to draw an exhaustive list of contaminants that places like Hoosick Falls, which has fewer than 10,000 residents, including a significant number of them with private wells, may host.

Water-quality tests produce information according to data standards. Many stakeholders, including the U.S. EPA and the New York State Department of Health, as well as independent researchers, began conducting water and soil testing in Hoosick Falls following the 2014 detection. In order to have the data considered legitimate by the regulations, each stakeholder follows particular water- and soil-testing standards outlined by either the Department of Health or the EPA. Standardizing water and soil testing enabled citizens and academic researchers alike to produce data considered readable, reliable, and legitimate by the EPA. In some ways, standardizing data collection thus helped render new stakeholders visible in the regulatory process.

However, standards also sometimes limit both the knowledge being produced and how it is regulated. For instance, the minimum concentration levels at which contaminants are reported, as well as the concentration levels at which health advisories are issued, are based, in part, on a measurement called the practical quantifiable level, or PQL. Labs across the U.S. can detect granular amounts of contaminants at different levels of precision, depending on their set up; that is, there are limits to which any given lab can confidently quantify the concentration of a substance using a given method. Comparing these limits across many different labs, the PQL is set to specify the precision at which most labs (not necessarily the most accurate ones) can feasibly detect contaminants. The PQL is set for regulatory purposes; standardizing this level ensures all labs can routinely produce results with good certainty.

Figuring out the level at which most labs are able to detect contaminants requires inter-laboratory studies, and, when data from such studies is not available, the PQL is estimated by increasing the lowest level at which the contaminant can be detected by a factor of 5 to 10. When defining drinking-water standards (such as the level at which contaminants must be reported to the EPA or the level at which a health advisory should be issued), the PQL is factored into the standard after health-based goals are determined, potentially diluting the rigor of the standards. Even if certain labs can reliably detect contaminants at levels much lower than the minimum reporting levels, these quantities do not get reported to the EPA. What the EPA and many other stakeholders know about water and soil contamination is shaped largely by efforts to standardize data collection. As with Frickel and Vincent, "We find what we seek, not necessarily what is there."

The way data gets reported and maintained thus places considerable limits on how it can be used to produce knowledge in the wake of a crisis.

Data standards also shape the possibilities for linking data. In order to investigate potential health risks associated with exposure, researchers and regulators rely on historic health data. The Department of Health, for instance, references data curated in the New York State Cancer Registry to investigate whether an elevation in cancer rates was seen in Hoosick Falls between 1995 and 2013.

There is little that academic researchers and citizen scientists can do with this data to examine how exposure might have affected public health. In order to protect the identities and locations of cancer patients, the New York State Cancer Registry reports cancer rates to the public by ZIP code only. Researchers cannot see if, say, there were spikes in cancer rates near water systems within Hoosick Falls reporting higher levels of PFOA; they can see only whether cancer rates spiked in the Village of Hoosick Falls as a whole that might be influenced by numerous other factors. This becomes even more complicated when considering that much environmental and public-health regulation requires evidence of statistical significance in order to mark a correlation between contaminant exposure and certain health outcomes. The way data gets reported and maintained thus places considerable limits on how it can be used to produce knowledge in the wake of a crisis.

back to top  Mapping Flows

Testing, registries, and datasets are often produced in ways that do not account for mobility. In tacking geographic units onto things so they are mappable, there is an assumption that people, non-human species, and material things stay in place. But water flows, as do people and knowledge. Such flows change not only across space but also across time. Mobility thus presents several challenges to mapping the level of contamination in the drinking water of Hoosick Falls with standard data-mapping practices. As researchers have tested for contamination, questions have been raised about how water and PFOA move through environments, how contaminants move between different materials, and why at times PFOA appears at very different concentrations in very proximate locations.

The effort to map the geomorphology of Hoosick Falls reflects the lack of available and easily usable data to adequately model the subsurface and distribution of different sediments and relate them to drinking water flows and sources. David De Simone, a glacial geomorphologist working on this mapping effort in Hoosick Falls, noted that in addition to a dearth of local geological data, "The existing well log database is one of the areas most fraught with error; well logs are, in some states, not in Excel files for easy analysis. Even when the logs are nicely cataloged, the actual location of the wells is problematic." Attempts to ascertain public health outcomes using National Cancer Registry data also face challenges. Such data does not account for how people move. It neither addresses the fact that residents might move into or out of an area during the period of time in which drinking water is contaminated nor the fact that some people work, but do not live, in places experiencing contamination.

Like water and people, knowledge also flows. As scientists conduct more research and produce more knowledge about environmental health, data models and standards shift. As with PFOA, standards can shift or become more expansive over time and may be contested by or vary across states. In May 2016, the EPA expanded its advisory levels for PFOA, from a 400-parts-per-trillion (ppt) shortterm exposure advisory level, to a 70ppt lifetime exposure advisory level. In September 2016, the New Jersey Drinking Water Quality Institute recommended a regulatory level of 14ppt for PFOA to New Jersey's own Department of Environmental Protection. This was lower than the state's previous level of 40ppt. Most saliently, however, the new recommendation of 14ppt fell below the EPA's minimum reportable level for PFOA of 20ppt. This means that, even though data at the 14ppt level could have been gathered through the UCMR3 process, at least at some labs, it was not collected and analyzed at this granularity. Using information strategies to plan for, regulate, and mitigate environmental crises can thus cause our understanding of what constitutes a crisis to change overnight or as we cross a geographic border.

To combat organized ignorance, researchers and regulators must develop more nuanced strategies for taking stock of how the assumptions built into data collection, data-reporting standards, and data mapping influence how we produce knowledge about crises.

back to top  Mapping Organized Ignorance

When relying on information strategies to regulate an environmental health crisis, recognizing how organized ignorance is produced and shapes knowledge matters a great deal. Data standards reinforced through institutions and regulations can render certain problems or populations invisible. New strategies are thus needed for visualizing data—strategies that can highlight rather than eclipse the limits of regulation through information and challenge dominant representations of space. Innovations in mapping organized ignorance can support this critical visualization.

The ongoing work of Bennington College faculty and students to measure PFOA levels in Hoosick Falls has begun to "problematize" these too-of-ten static representations. Collecting samples from the same water sources in an iterative manner, this research seeks to better understand if and how contamination of drinking water could shift over time, perhaps in nonlinear ways. In particular, it considers the juxtaposition of proximate spaces with very different levels of contamination. Such potential shifts would not only challenge the ways water contamination is mapped but also the ways standard agreements with polluters under state and federal Superfund legislation are deemed to be fulfilled and concluded.

back to top  Conclusion

Inspired by such work and its potential implications for informing public policy, we suggest mapping environmental health crises in ways that account for more than just the distribution of contaminants across spatial dimensions. Critical mapping practices can be designed to illustrate what researchers and regulators do not know and cannot know about a certain crisis due to the limits of the data and the limits of data analysis. Useful exercises in mapping ignorance might thus consider:

Data. Availability of data across space and time;

Standards. The way standards shift over time and across locations;

Scales. Clarity and obfuscation of working at different scales;

Lack of a common measure. (In)commensurability of units, formats, and ways of reading data;

Silos. Knowledge silos produced across regulatory structures;

Moving points. The ways certain data follows (or not) points that move; and

Actors. Inclusion or exclusion of certain communities or actors.

Such exercises could involve juxtaposing maps designed from the perspective of many different stakeholders. It could also involve juxtaposing maps of the location of contaminants before and after drinking water standards have changed. It could involve mapping the areas where water and soil quality are not tested or involve considering new forms of data (such as testimonial and mappable). In any case, it will require that researchers and regulators alike develop more critical ways of thinking about and visualizing environmental health crises, drawing attention to the complex ecological, social, and historical dynamics that shape how we know about contemporary problems.

back to top  References

[1] Fortun, K. From Bhopal to the Informating of Environmentalism: Risk Communication in Historical Perspective. Osiris 2nd Series Vol. 19, Landscapes of Exposure: Knowledge and Illness in Modern Environments. University of Chicago Press, Chicago, IL, 2004, 283–96.

[2] Frickel, S. and Bess Vincent, M. Hurricane Katrina, Contamination, and the unintended organization of ignorance. Technology in Society 29, 2 (2007), 181–88.

[3] U.S. Environmental Protection Agency. Drinking Water Health Advisory for Perfluorooctanoic Acid (PFOA), 2016; 822-R-16-005.

back to top  Authors

Laura Rabinow is a Ph.D. student in Science and Technology Studies at Rensselaer Polytechnic Institute in Troy, NY; her research examines the co-production of policy, politics, and regulatory science pertaining to drinking water.

Lindsay Poirier is a cultural anthropologist and Ph.D. candidate in Science and Technology Studies at Rensselaer Polytechnic Institute in Troy, NY; her research examines how scientists, technologists, and regulators are developing strategies to attend to the limits of big data.

back to top 

© 2017 ACM 0001-0782/17/03 $15.00

The Digital Library is published by the Association for Computing Machinery. Copyright © 2017 ACM, Inc.


There are no comments at this time.


To comment you must create or log in with your ACM account.