Alberto Giordano

Department of Geography

Maxwell School of Citizenship and Public Administration

Syracuse University

Syracuse, NY-13244

algiorda@mailbox.syr.edu

This paper presents a summary of the results of a research I conducted in the Emergency Planning Zone (EPZ) of the Nine Mile Point nuclear complex, located in Oswego County, New York. The research had two main objectives: to understand how people would react to a nuclear accident at the plant, and to explore the possibility of using a GIS as a tool for risk communication. The study showed that there are problems in the way risk is communicated in the county, and that a GIS facilitates the flow of information between people and emergency planners. The first part of the paper (Section 1) focuses on reactions to the nuclear emergency, whereas the second (Sections 2 and 3) deals with the use of GIS and frames my study in the context of public participation GIS.

On March 28, 1979, in the early morning hours, an accident occurred at the Three Mile Island nuclear power plant, located near Harrisburg, Pennsylvania. The accident, a coincidence of technical failures and human errors, had profound influences on the public perception of nuclear risk. It also prompted the Nuclear Regulatory Commission (NRC), the federal agency which regulates all aspects of nuclear energy production in the United States, to implement major changes in its policies.

An important part of the new set of regulations concerns the emergency planning procedures to be followed in the event of a nuclear accident. In the new policies, still valid today, great emphasis is placed on the notion of Emergency Planning Zone, or EPZ. In an EPZ, which covers a 10-mile radius from a nuclear power plant, special provisions are established for evacuation of the population in the event of an accident, and for the communication of nuclear risk to the public.

When I first decided to focus my research work on GIS and nuclear emergency response policies, I came across various studies (see, for example, Golding, Kasperson, and Kasperson, 1995) criticizing the concept of EPZ, as well as other aspects of the new policies implemented by the NRC. However, I realized that none of the critics, at least to my knowledge, had evaluated the effectiveness of the new planning procedures, and in particular the EPZ concept, by asking the opinions of those who live in the Emergency Planning Zone of a nuclear complex. How would these people react to a nuclear accident? How well do they know the implemented emergency plan? What can we learn from their attitude towards nuclear risk? How do these attitudes affect possible behavior in the event of a nuclear accident?

In order to answer these questions, I conducted an opinion survey in the Emergency Planning Zone (EPZ) of the Nine Mile Point (NMP) nuclear complex, located in Oswego County, New York. The objective was to understand how a sample of people living in the EPZ would likely react to a nuclear accident at NMP, compare their reactions with the directives given by the Oswego County Emergency Management Office (EMO), and determine if socio-demographic factors, distance from the nuclear complex, attitudes towards nuclear risk, and opinions on various social issues could explain their answers. I sent two different questionnaires, one approximately three months after the other, to a sample of 583 people. In the first survey, mailed in September 1997 and answered by 139 people (24%), respondents were asked to state how they would most likely react to nuclear accidents of different magnitudes. The accidents were described in four different scenarios modeled on messages likely to be broadcast by the Oswego County Emergency Management Office (EMO) in case of an emergency at the nuclear complex. Each scenario, written in collaboration with local emergency management officials, included directives given by the EMO for that specific type of accident. Each respondent received only one - randomly assigned - scenario, and his or her answers were compared with the EMO’s directives. The survey also included a set of questions aimed at building a socio-economic profile of the respondent, and at determining how close he or she was to NMP (in terms of driving, straight-line, or perceived distance.) The second questionnaire was sent out in mid-December 1997 to a subset of those who, in answering the first survey, had indicated that they were available for further participation. Of the 60 who received the second survey, 42 (70%) completed and returned it. The questions probed respondents’ attitudes towards nuclear energy production, their perception of nuclear risk, their trust of the media and other authorities, and their opinion on various societal issues (e.g., role of the government, limits to private industry.)

The results of the first survey highlighted problems in risk communication. For example, only 39% of the respondents could readily locate the booklet "Oswego County Emergency Planning and You," which contains instructions on what to do in case of an accident at the nuclear complex. Sent once a year to all people living in the EPZ, the booklet is the main risk communication tool used in the county. Equally worrisome is the fact that only 34% of the respondents know the Emergency Response Planning Area (ERPA) in which they live. ERPAs are the fundamental geographic units into which emergency planners have divided the Nine Mile Point EPZ,¹ and for some accidents the EMO might issue different directives for different ERPAs. Not knowing your ERPA, therefore, means not knowing which directives apply to you.²

Despite widespread unfamiliarity with crucial aspects of the plan, most (63%) respondents say they would follow the EMO’s directives in the event of an emergency at NMP. This proportion, however, increases or decreases depending on the action recommended. Almost all people (92%) would obey directives to leave the EPZ, but less than half (48%) would do so when the EMO issues different directives to different ERPAs.³ A closer look at the answers suggests that many respondents would act quite independently of the EMO’s directives. When a message is broadcast warning of a potentially dangerous event at NMP, they would assess the situation and decide whether to follow the order given. Respondents who indicated they would not follow directives, almost invariably say they would leave the area, suggesting that they are quite worried about the consequences of a severe malfunctioning at the nuclear complex.

Another element of concern for the EMO is the large population of people (54%) who, after leaving the area as directed, indicate that they would not report to the designated shelter in Syracuse. This finding is troublesome: those who are instructed to evacuate need to go to the shelter for a medical check, even if they do not intend to stay there for long. The reasons for this behavior are, quite likely, lack of knowledge of the procedures that should be followed in the aftermath of an actual or potential radioactive contamination. This is probably the result of a risk communication strategy that stresses the very low likelihood of an accident and neglects to mention the possible seriousness of its consequences.

The next step in the analysis was to look for factors that might explain respondents’ reactions to the emergency. It turned out that socio-demographic and economic characteristics were not very important. Certainly, differences do exist. For example, women, NMP employees, and respondents who served in the military are slightly more likely than other respondents to follow the EMO’s directives, but the differences are small. More correlated with respondents reactions were several factors examined in the second questionnaire, and especially the respondents’ views on society and nuclear risk perception. Those who would not follow the EMO’s directives strongly support the rights of individuals to use their property without outside interference, oppose government regulations of business, and distrust NMP officials in case of an emergency. They also appear to be more concerned about the safety of NMP, are less willing to accept the hazards of nuclear energy, and more strongly agree with the statement that they are exposed to nuclear risk without their consent.

In order to communicate the results of my study to the county’s emergency planners, I decided to use a GIS. In the rest of this paper, I will explain why I chose a GIS and how I accomplished my objective.

In the early 1990s, commentaries by Peter Taylor (1990) and Stan Openshaw (1991) sparked a wide (Taylor and Overton, 1991; Goodchild, 1991; Openshaw, 1992; Smith, 1992), sometimes heated,⁴ but long overdue debate on the nature of GIS. The questions raised were not entirely new - some had been anticipated by Chrisman in a 1987 paper, - but the debate had the merit of initiating a systematic discussion on issues such as the role of GIS within geography, the intellectual content of "doing" GIS, and the nature (a tool or a science?) and social implications of GIS. They also contributed to heighten awareness of the fact that the GIS is a technology rooted in a social and political context, not a value-free tool. Faced with accusations of perpetuating the social order, of being an instrument for the elite, and of resuscitating positivist quantitative geography, several researchers started developing GISs programmatically geared towards promoting equity, collaborative planning, and community empowerment. To cite just a few examples, limited to the context of GIS applications in developing countries,⁵ Yapa (1991) showed how to incorporate indigenous local knowledge in the GIS, Hutchinson and Toledano (1993) proposed guidelines for achieving participatory GIS, and Weiner and Harris initiated a pioneering community-oriented GIS project in South Africa (1995).

Theoretical insights and practical applications generated widespread interest on issues of GIS and society, which culminated in the publication of special issues of academic journals, in at least one book devoted to the topic, and in the establishment of an NCGIA Initiative on "GIS and Society."⁶ While the discussion continues, focusing more on how, rather than on if, we should build participatory GISs, some key factors seem to be generally accepted. For instance, it is recognized that developing a GIS is a social process (Sheppard, 1995), and that GISs can be built to empower communities (and, as Harris and Weiner note in a 1998 paper, also to marginalize them.) Most researchers would also agree that current GISs are limited in their ability to represent space, and that we should try to incorporate in our GISs alternative views of reality and modalities of representation (NCGIA, 1996, pp.5-14). Finally, we have all been warned that a GIS, with apparently sound spatial analysis and nice-looking maps, can make policy reports look more authoritative and reliable than they really are (Obermeyer, 1998).

The objective of my work was to build a GIS for the communication of risk. Recognizing that multiple perspectives on nuclear risk are present in the county, I concentrated on the views of the planners as well as those living close to the nuclear plant. Currently, the communication is one-way - from the planners to the people - but the surveys provided a way of bringing the views of the people to the planners, and in doing so initiated the flow of communication in the opposite direction. The GIS served primarily as a visualization tool and as a mean of exploring spatial patterns.

An unwritten assumption of nuclear emergency plans is that people will respond to an accident as directed by the authorities, and that risk communication strategies such as the booklet used in Oswego county will effectively educate citizens on the risks they face and on how to respond to critical situations. In the Nine Mile Point EPZ, as I have shown in Section 1, this is not always the case. A close look at my results suggests that work needs to be done on the communication of nuclear risk. As noted earlier, this is currently a one-way process, in which local authorities and plant managers try to educate people living in the EPZ. Basically, this means reassuring people that NMP is safe and telling them to follow the EMO’s directives should the impossible (a serious accident) occur. People in the county are not actively involved in planning for a nuclear emergency. If they had been involved in developing the emergency plan, they would probably understand its functioning and limitations, comprehend its stages and the efforts that go into its design, understand why they should be checked for possible health problems in the event of an accident, and (one would hope) achieve the highest possible level of trust for the policy-makers. Planners too would benefit from a more participatory approach to the management of nuclear risk; most important, they could achieve a better understanding of the public’s possible reactions in the event of an accident, and be better able to explain why actions like sheltering are sometimes more appropriate than more drastic measures, such as evacuation.

It is in this framework that I see a potential benefit in the adoption of the GIS. Although issues of data availability, data quality, training, and overall costs of implementation still limit its diffusion in emergency management offices,⁸ GIS is likely to play an important role in the work of future emergency planners. The flexibility and interactivity of the GIS, as well as its capability for quickly creating different scenarios, makes it a useful tool for a democratic, participatory nuclear emergency planning, a process in which the public and the authorities work together to achieve consensus on the plan. Already, products like CAMEO and EIS are used in conjunction with GIS software (such as Arc View) by some New York State EMOs for traditional risk assessment applications (see Monmonier and Giordano, 1998). But the GIS, with its powerful visualization functionalities, can be an appropriate and efficient tool for showing different, alternatives views on nuclear risk, and for the collaborative development of emergency plans. Of course, it is too late to propose this approach for Oswego County: the plan is already there. But alternative views on nuclear risk can be brought to the planners, and even if this is not direct communication yet, making the planners understand the rationale behind people’s attitudes towards nuclear risk is an important objective. This rationale often escapes local planners, who seem to believe many people are irrationally alarmist, or otherwise biased against the work of the local Emergency Management Office and the utility companies operating the nuclear complex.

Operationally, I developed the GIS and presented the results of my surveys to the county’s emergency planners using LandView III and Marplot 3.2. These two products, available together on 10 CD-ROMS, were developed by different branches of the U.S. federal government as a "community right-to-know" tool. As such, they can be found at many libraries and are available to the public at a low cost. LandView is produced by the Environmental Protection Agency (EPA), the Bureau of the Census, and the National Oceanic and Atmospheric Administration (NOAA), and is basically a database management systems that contains TIGER data, 1990 census data, EPA site locations and information, and other data from various government sources. Marplot is a mapping software application originally developed for CAMEO by the EPA, the Census bureau, NOAA and the U.S. Coast Guard. Marplot allows the user to display, query, and to some extent manipulate the data contained in LandView. The software does not deliver what most users would expect from a contemporary GIS: it has few analytic functionalities and allows for very limited graphic choices. For example, font and size of characters in the legend cannot be modified. However, Marplot has several advantages: it is inexpensive and comes with an impressive geographic database. Moreover, its very simple menu and limited range of functionalities makes it easy to learn and use. The CDs can be purchased separately (the one I used included the Northeastern states only), and they all contain Marplot. There is also another reason I used LandView and Marplot for my study. As I said, the County’s EMO has a copy of the software but currently does not use it; however, various EMO’s employees have expressed interest in learning more about its potential. My demonstration could show how the GIS would better inform their work as well as show a variety of relevant maps they might want to produce.

In order to show the results of my research, I required three main databases: a county base map, the location of the respondents to my survey, and the nuclear emergency plan. The base map came with LandView and was easy to create: it only includes the county’s main settlements and all roads. Finding the exact location of the respondents in the EPZ required driving around with a GPS receiver (I used a Garmin 12), locating all the addresses, and recording their locations in latitude and longitude. I then entered the coordinates in Marplot and created the respondents database, which also included the answers each respondent gave to the two questionnaires. The final step involved entering relevant parts of the nuclear emergency plan into my database. To do so, I first traced directly on the base map the 22 ERPAs that comprise the EPZ, and then added the evacuation routes that the population should use in the event of an accident. A few additional elements (e.g., the population of the ERPAs) completed the system.

In this way, I was able to cartographically represent the results of my survey and search for spatial patterns. For example, a cluster of respondents in ERPA 11 intrigued me as well as EMO personnel. It turned out that this is a relatively new residential area in the outskirts of Oswego, quite isolated and separated from the rest of the city, relatively affluent, and home to a few nuclear complex employees. Evidently, the residents were quite interested in my survey, and in fact I received a couple of phone calls from people living in ERPA 11 inquiring about my research. Another interesting finding was the apparent minimal influence of distance from the plant on the respondents’ reactions to the accident. This indicate that access to escape routes and closeness to the hazard do not have, as it might have been expected, a decisive effect on the respondents’ decisions.

On September 18, 1998, the GIS implementation was formally presented to EMO’s employees and to a representative of Niagara Mohawk (the utility company running two of the three nuclear reactors at NMP). A short questionnaire was distributed to the participants and collected after the presentation. The objective of the questionnaire was to gather opinions on the usefulness of my study and the appropriateness of the GIS as a tool for communication of my research findings. The results were generally positive. There was concern for the small number of people who would go to the designated shelter in case of an accident, a long discussion on the opportunity of broadcasting messages suggesting sheltering and evacuation (as in the second scenario), and, most important, the realization that the risk communication tool used in the county (the booklet) was not effective for a significant proportion of the population. As regards more specifically the use of the GIS, four of the six participants thought that it was useful to see where people were located in the area and that it helped their overall understanding of the results. One in particular said that the GIS helped her understanding that people’s concerns are not limited to an area, but common to people throughout the EPZ. This lack of evident spatial patterns was held against the usefulness of GIS by the remaining two participants. Finally, interest was shown for acquiring my GIS implementation, and for further collaborating in more general GIS-related issues.

My research had two objectives: a) understanding how a sample of people living in the Emergency Planning Zone (EPZ) of the Nine Mile Point nuclear complex would react to a nuclear accident; b) exploring the use of a GIS as a tool for communicating people’s opinions and concerns to professional emergency planners in Oswego County. The research identified problems in the communication of nuclear risk to EPZ residents: they seem poorly informed about key aspects of the emergency plan, and might tend to ignore the EMO’s (Emergency Management Office) directives. The research also showed that socio-economic characteristics and distance from the nuclear complex do not explain respondents’ reactions, but that nuclear risk perception, opinions on various societal issues, and attitudes towards nuclear energy production do. As regards the use of GIS, my attempt at using the technology to communicate risk, presenting multiple perspectives on the hazard of nuclear power production, and bringing to the planners the opinions and concerns of those who live in the shadow of Nine Mile Point generally worked. Of course, the application I described is still quite distant from a truly collaborative and participatory GIS, but I hope to have shown that it is a step in the right direction.

__________________

Carrara A., Guzzetti F., eds., Geographical Information Systems in Assessing Natural Hazards, Dordrecht, Kluwer, 1995.

Chrisman N.R., Design of Geographic Information Systems Based on Social and Cultural Goals, Photogrammetric Engineering & Remote Sensing, 53-10(1987), pp.1367-1370.

Golding D., Kasperson J.X., Kasperson R.E., (eds.), Preparing for Nuclear Power Plant Accidents, Boulder, Westview, 1995.

Goodchild M.F., Just the Facts, Political Geography Quarterly, 10-4(1991), pp.335-337.

Harris T., Weiner D., Empowerment, Marginalization, and "Community-Integrated" GIS, Cartography and GIS, 25-2(1998), pp.67-76.

Harris T. et al., Pursuing Social Goals through Participatory GIS: Redressing South Africa’s Historical Political Ecology, In: J. Pickles, ed., "Ground Truth. The Social Implications of Geographic Information Systems", New York, Guilford Press, 1995, pp.196-222.

Lake R.W., Planning and Applied Geography: Positivism, Ethics, and Geographic Information Systems, Progress in Human Geography, 17-3(1993), pp.404-413.

McMaster R.B., Leitner H., Sheppard E., GIS-Based Environmental Equity and Risk Assessment: Methodological Problems and Prospects, Cartography and GIS, 24-3(1997), pp.172-189.

Monmonier M., Giordano A., GIS in New York State County Emergency Management Offices: User Assessment, Applied Geographic Studies, 2-2(1998), pp.95-109.

National Center for Geographic Information and Analysis (NCGIA), GIS and Society: The Social Implications of How People, Space, and Environment Are Represented in GIS, Initiative 19 Specialist Meeting, March 2-5, 1996, South Haven (Minnesota).

Nyerges T., Robkin M., Moore T.J., Geographic Information Systems for Risk Evaluation: Perspectives on Applications to Environmental Health, Cartography and GIS, 24-3(1997), pp.123-144.

Obermeyer N., The Evolution of Public Participation GIS, Cartography and GIS, 25-2(1998), pp.65-66.

Obermeyer N., Special Issue on "Public Participation GIS," Cartography and GIS, 25-2(1998).

Openshaw S., A View on the GIS Crisis in Geography, or, Using GIS to Put Humpty-Dumpty Back Together Again, Environment and Planning A, 23(1991), pp.621-628.

Openshaw S., Further Thoughts on Geography and GIS: A Reply, Environment and Planning A, 24(1992), pp.463-466.

Pickles J., ed., Ground Truth. The Social Implications of Geographic Information Systems, New York, Guilford Press, 1995.

Scott M.S., Cutter S.L., Using Relative Risk Indicators to Disclose Toxic Hazard Information to Communities, Cartography and GIS, 24-3(1997), pp.158-171.

Sheppard E., Poiker T., eds., Special Issue on "GIS and Society," Cartography and GIS, 22-1(1995).

Sheppard E., GIS and Society: Towards a Research Agenda, Cartography and GIS, 22-1(1995), pp.5-16.

Smith N., History and Philosophy of Geography: Real Wars, Theory Wars, Progress in Human Geography, 16-2(1992), pp.257-271.

Taylor P.J. GKS, Political Geography Quarterly, 9-3(1990), pp.211-212.

Taylor P.J, Overton M., Further Thoughts on Geography and GIS, Environment and Planning A, 23(1991), pp.1087-1094.