Development and implementation of a right-to-know web site that presents estimated cancer risks for air emissions of large industrial facilities

Authors


Abstract

A publicly accessible web site was developed that provides environmental information on air emissions of local facilities in 3 regions in the Netherlands. Compared with other environmental right-to-know initiatives, the main innovative aspect of this web site is the presentation of emissions data in terms of estimated increased cancer risk. The basic idea is that this type of information will help people 1) to put the health effects of air emissions of the selected facilities into perspective and 2) to form a personal opinion about the acceptability of these risks. In the 1st 5 months, the web site was visited more than 77,000 times, with more than 33,000 visits in the 1st month (March 2004). The web site triggered a discussion on the accessibility and presentation of environmental risk information in The Netherlands, which even reached the Dutch parliament. To improve and maintain this type of right-to-know initiative, it is important that 1) responsible authorities establish and maintain a detailed and up-to-date database of national, regional, and local emissions data, 2) the initiative is implemented in a legal framework or adopted by an independent body, and 3) more knowledge becomes available about the understanding and effects of environmental risk information on the general public.

INTRODUCTION

Public disclosure of environmental information figures high on the political agenda of many states. The main drivers behind this trend are 1) the belief that the public has the “right to know” how they are affected by activities of 3rd parties (Sarokin and Schulkin 1991) and 2) indications that this information could help improve the environmental performance of the private sector (Khanna et al. 1998; Kleindorfer and Orts 1998; Tietenberg and Wheeler 1998; Graham and Miller 2001; Stephan 2002; O'Rourke and Macey 2003; Wang et al. 2004). The most widely recognized and copied disclosure program of environmental information is the Toxics Release Inventory (TRI) of the US Environmental Protection Agency (USEPA). The TRI is a publicly available database (www.epa.gov/tri/) that contains information on toxic chemical releases and other waste management activities reported annually by industrial and federal facilities. The TRI was established under the Emergency Planning and Community Right-to-Know Act of 1986 and expanded by the Pollution Prevention Act of 1990. The database takes a ZIP code and provides an overview of the facilities in that area and the reported emissions.

In Europe, public disclosure of environmental information is currently a hot topic. On 28 June 1998, the Aarhus Convention was adopted by 36 United Nations Economic Commission for Europe states, which recognizes that citizens must have access to information to be able to assert one's right to live in an environment adequate to his or her health and well-being. It imposes clear obligations on public authorities to provide and facilitate access to environmental information to the general public. In the United Kingdom, Friends of the Earth launched the Factory Watch web site in February 1999, which was closed down after the UK Environment Agency published a more extensive web site on environmental information (“What is in your backyard?”; www.environment-agency.gov.uk). Recently, the European Union launched the European Pollutant Emission Register (EPER; www.eper.cec.eu.int/eper/). EPER was established under the EU directive on Integrated Pollution Prevention and Control (96/61/EC). It covers emissions of 50 pollutants that must be reported by member states if a threshold is exceeded.

Although databases such as TRI, Factory Watch, and EPER are comprehensive, interpretation of the emissions data in terms of environmental impact is lacking. This led Environmental Defense in the United States to develop Scorecard (www.scorecard.org). The web site combines the emissions data of the TRI with information on the potential human health hazards of toxic chemicals. Scorecard indicates which chemical releases in an area might be of potential health concern and identifies the highest priorities for pollution prevention (Hertwich et al. 1998). This is a considerable step forward in providing the public with meaningful information, but it still does not answer the question of primary concern to most people: “How do the pollutant emissions affect our well-being and health status?”

Examples of web-based initiatives that seek to answer this health question are rare, especially in relation to the emission of toxic substances. One example is the National-Scale Air Toxics Assessment of the USEPA (www.epa.gov/ttn/atw/nata/), which predicts air quality and its possible effects on human health nationwide on the basis of emissions data of 32 toxic air substances from the National Toxics Inventory. However, this initiative is based on relatively old emissions data (1996), and it does not provide reliable information on the level of a ZIP code area. In other risk assessment areas (e.g., safety and flooding), publicly accessible risk maps are more common, although they are currently controversial because of the potential for misuse by terrorists. Examples are the safety maps of the Provinces of Friesland (www.fryslan.nl/risicokaart2/startrisicokaart.htm) and Gelderland (geodata. prv.gelderland.nl/km/risicokaart/) in The Netherlands and the flood maps of the UK Environment Agency (www.environment-agency.gov.uk).

In this paper, we outline the development of a Dutch pilot web site (www.rechtomteweten.nl) that provides information on estimated cancer risks for air emissions of major facilities in 3 industrialized areas in The Netherlands (Rijnmond, IJmond, and Nijmegen West). The web site was a private initiative of a consortium of Dutch environmental organizations that was funded by the Dutch Ministry of Public Housing and the Environment. The main aim of the initiative was to demonstrate that emissions data of industrial facilities could be presented in a more meaningful way than current right-to-know initiatives. This paper describes the methods underlying the development of the web site (data gathering, processing, integration, and interpretation), summarizes the responses to the web site, and discusses the lessons learned.

METHODS

The procedure followed to develop the right-to-know web site is similar to the USEPAs National-Scale Air Toxics Assessment (www.epa.gov/ttn/atw/nata/). It consisted of the following steps:

  • 1.Selection of the study areas
  • 2.Selection of carcinogenic substances and unit risk factors
  • 3.Selection of facilities for inclusion in the pilot
  • 4.Gathering of emission data and stack specifications
  • 5.Air dispersion calculations
  • 6.Calculation and aggregation of cancer risks
  • 7.Development of a framework for interpretation of the results.

Study areas

Three relatively highly industrialized areas in The Netherlands were selected (Figure 1). Rijnmond is a 20 × 20 km square area with a high density of petrochemical industry at the mouth of the river Rhine and close to the city and harbor of Rotterdam. IJmond is a 10 × 10 km square area dominated by an iron and steel production facility close to the city of IJmuiden. Nijmegen West is a 10 × 10 km square industrial area at the western side of the city of Nijmegen in the eastern part of The Netherlands. The latter area was chosen because an increased cancer incidence in certain parts of this area raised considerable concern among the population about the possible influence of industrial emissions (Van Dijck et al. 2004).

Substances and unit risk factors

The selection of substances for inclusion in the pilot project was based on the International Agency for Research on Cancer (IARC 2003) classification of carcinogenic substances. All substances classified in IARC classes 1 (carcinogenic to humans; 88 substances), 2A (probably carcinogenic to humans; 64 substances), and 2B (possibly carcinogenic to humans; 240 substances) were included in the selection. This selection of 392 substances was narrowed to those for which air unit risk factors were available in the Air Quality Guidelines for Europe 2000 of the World Health Organization (WHO 2001) or the IRIS database of the USEPA (2003). If both databases reported different air unit risk factors for the same agent, preference was given to the WHO unit risk factors. The resulting selection was further narrowed by deleting substances that were not emitted by the facilities considered (e.g., asbestos and azathioprine) or for which emission data were unavailable (e.g., radionuclides). This procedure resulted in 18 carcinogenic substances with corresponding unit risk factors (Table 1). Table 1 also lists the chronic exposure concentration that corresponds to a theoretical cancer risk of 1 × 10−6/y. This concentration is referred to in The Netherlands as the maximum permissible concentration (MPC) and is used as a guideline for assessing air quality (VROM 1989). MPCs were derived from air unit risk factors assuming an average individual lifespan of 70 y,

equation image((1))

where MPC; is the maximum permissible air concentration of substance i (μg/m3) that corresponds to a cancer risk of 1 × 10−6 y−1, MPR is the maximum permissible risk level for individual substances in The Netherlands (1 × 10−6 y−1), and URFi is the unit risk factor for substance i (m/μg).

Figure Figure 1..

Overview of the location of 3 selected study areas (Rijnmond, IJmond, and Nijmegen West) in The Netherlands.

Facilities

Because of limited time and resources, it was considered unfeasible to gather emissions data on all facilities in the 3 selected areas. On the basis of 1998 emissions data of the Dutch National Emission Registration (www.emissieregistratie.nl), a selection procedure was adopted to identify 8–15 facilities for each study area that contributed most to the cancer risk. This procedure was based on a priority indicator (PI) that represents the maximum volume of air that can be polluted each year by a facility up to the concentration level that corresponds with a cancer risk of 1 × 10−6 y−1,

equation image((2))

where PLx is the priority indicator of facility x (m3/y), n is the number of carcinogenic substances emitted by the facility, and YELi,x is the yearly emission load of substance i by facility x reported in the national emission registry (kg/y).

Emissions data and stack specifications

Emissions data of large facilities in The Netherlands are registered by the Dutch National Emission Registration (www.emissieregistratie.nl). Although this database is quite comprehensive, it is incomplete and not up-to-date. The last year for which the database contains relatively detailed and complete emissions data is 1998. Data from that year were used as a starting point for the emission inventory of the pilot. These data were updated as follows:

  • Publicly available documents (environmental permits, annual reports, monitoring reports) were reviewed for more recent emissions and stack data.

  • Local and regional public authorities were approached and asked to provide more up-to-date emissions and stack data. All authorities approached (Province of South Holland, Province of Gelderland, DCMR Rijnmond, and the municipality of Nijmegen) responded, and the data they provided were adopted.

  • The selected facilities were approached and asked to provide more up-to-date emissions and stack data. Of the 28 facilities approached, 11 responded, and 5 of the respondents provided new emissions data, which were adopted.

The result was a comprehensive database of up-to-date emissions data for the selected facilities.

Air dispersion calculations

Dispersion calculations were performed with the Gaussian plume model STACKS 4.1 for Windows (KEMA 1999). STACKS calculates dispersion parameters (i.e., σy and σz) directly from measured wind and turbulence data. In 1997, STACKS was accepted as the Dutch National Dispersion Model for regulatory purposes after completion of several validation studies (Erbrink 1995; Cosemans et al. 2001).

Table Table 1.. Overview of the 18 selected carcinogenic substances, their Chemical Abstract Service (CAS) numbers, unit risk factors (URFs), and the maximum permissible concentrations (MPCs)
SubstanceCAS NraURF (m3/μg)bMPC (μg/m3)cSourced
  1. a For metal compounds, the CAS number of the pure metal is listed in parentheses.

  2. b The unit risk factor (URF) is a cancer risk estimate for lifetime exposure to a concentration of 1 μg/m3.

  3. c The maximum permissible concentration (MPC) is the chronic exposure concentration that corresponds to a theoretical cancer risk of 1 × 10−6 y−1.

  4. d IRIS = USEPA (2003); WHO = WHO (2001).

  5. e The URF of benzo[a]pyrene was also used to estimate the cancer risk of PAH emissions under the assumption that 1.5% of the PAH mixture consists of benzo[a]pyrene (Slooff, Janus, et al. 1989).

  6. f The URF of chromium [VI] compounds was also used to estimate the cancer risk of chromium emissions that were reported as total chromium emissions under the assumption that 30% of this total emission consisted of chromium [VI] compounds (Slooff, Cleven, et al. 1989; Huggins et al. 1993; Wong et al. 1996).

Acetaldehyde75–07–02.20 × 10−63.18 × 101IRIS
Acrylonitrile107–13–12.00 × 10−53.50 × 100WHO
Arsenic and arsenic compounds(7440–38–2)1.50 × 10−34.67 × 10−2WHO
 71–43–26.00 × 10−61.17 × 101WHO
Benzidine92–87–56.70 × 10−21.04 × 10−3IRIS
Benzo[a]pyrenee50–32–89.00 × 10−27.78 × 10−4WHO
Beryllium and beryllium compounds(7440–41–7)2.40 × 10−32.92 × 10−2IRIS
Bis(chloromethyl)ether542–88–16.20 × 10−21.13 × 10−3IRIS
1,3-Butadiene106–99–03.00 × 10−52.33 × 100IRIS
Cadmium and cadmium compounds(7440–43–9)1.80 × 10−33.89 × 10−2IRIS
Chloroform67–66–32.30 × 10−53.04 × 100IRIS
Chromium [VI] compoundsf(18540–29–9)4.00 × 10−21.75 × 10−3WHO
1,2-Dichloroethane107–06–22.60 × 10−52.69 × 100IRIS
Formaldehyde50–00–01.30 × 10−55.38 × 100IRIS
Nickel compounds(7440–02–0)4.00 × 10−41.75 × 10−1WHO
1,1,2,2-Tetrachloroethane79–34–55.80 × 10−51.21 × 10°IRIS
Trichloroethylene79–01–64.30 × 10−71.63 × 102WHO
Vinyl chloride75–01–41.00 × 10−67.00 × 101WHO

For each facility, the dispersion of a hypothetical pollutant emission of 1 kg/s in the study area was simulated. The resulting pollutant ambient concentrations were used for calculating the local cancer risks (see Cancer risks). This calculation procedure is based on the linear relationship between emissions data and pollutant ambient concentrations in the STACKS dispersion model. Furthermore, it is assumed that the influence of substance characteristics on plume dispersion in the vicinity of the stack (up to 10 km) is negligible compared with the influence of environmental characteristics such as wind speed. Each study area was divided in 20 grid fields, which is the maximum spatial resolution of the STACKS model. This resulted in 0.5 × 0.5 km grid fields for the IJmond and Nijmegen West areas and 1.0 × 1.0 km grid fields for the Rijnmond area. Concentrations were calculated at the intersections of the grid lines. Meteorological data for Schiphol Airport were used for the dispersion calculations in Rijnmond and IJmond, whereas data from Eindhoven Airport were used for Nijmegen West. Meteorological data were gathered over a 5-y period (1 January 1990–31 December 1994), and it was assumed that this period is representative for current meteorological conditions. Surface roughness was determined with the Dutch standard map for surface roughness developed by Wieringa and Rijkoort (1983). Avalue of 0.25 (rough surface) was used for Rijmond and IJmond and 0.5 (very rough surface) for Nijmegen West. Pollutant concentrations were calculated at the standard reference height of 1 m.

Cancer risks

The cancer risk for the sum of carcinogenic emissions of a facility was calculated for each grid intersection,

equation image((3))

where is the estimated yearly cancer risk at grid intersection gi for the air emissions of facility x (y−1), Cgi,x is the predicted concentration at grid intersection gi corresponding to a hypothetical pollutant emission of 1 kg/s by facility x (kg-m−1/kg.s−1), and Eix is the emission of substance i by facility x (kg/s).

The equation above is based on the assumption that cancer risks of different substances are additive (USEPA 1986). The overall cancer risk for each grid intersection was calculated by adding the cancer risks of different facilities. Finally, the relative contribution of each facility to the overall cancer risk in a study area was calculated as

equation image((4))

where RCx,a is the relative contribution of facility x to the overall cancer risk in study area a (%), m is the number of grid points in study area a, and l is the number of facilities in study area a.

The calculation procedure resulted in a 20 × 20 grid file for each study area with risk predictions for all grid intersections (21 × 21 = 441 risk values). These values were imported into Microsoft Excel®, and the standard contour plotting function (2D Surface Plot) was used to produce risk contours. The grid files of the 3 study areas were also used to allocate cancer risks to 5-digit ZIP codes by means of a map overlay. Each ZIP code was allocated the risk value of the grid intersection that was closest to the center of the ZIP code area.

Framework for interpretation

To facilitate the interpretation of the calculated cancer risks, a frame of reference was developed that consisted of:

  • An overview of cancer risks related to background concentrations of carcinogenic substances in 3 common exposure situations: 1) a rural area, 2) a city area, and 3) a high-intensity traffic road in a city area. These so-called background risks were calculated on the basis of annual average concentrations of 4 carcinogenic substances (arsenic, cadmium, benzene, and benzo[a]pyrene) that are measured on a regular basis within the national air quality monitoring network of The Netherlands (Table 2; RIVM 2002).

  • An overview of risk estimates for activities or events encountered in daily life (Table 2).

  • An indication of the Dutch risk standards for exposure to a mixture of carcinogens, i.e., the MPR (1 × 10−5 y−1) and negligible risk (NR; 1 × 10−7 y−1; VROM 1989).

  • An extensive overview and explanation of the different sources of uncertainty involved in the emission inventory and risk calculations.

RESULTS

The results of the cancer risk calculations were presented as risk contours in maps of the 3 study areas (Figures 2 to 4). The overall cancer risks were highest in Rijnmond (1.3–83 × 10−8), which was closely followed by IJmond (1.9–56 × 10−8). The risks in Nijmegen West were approximately a factor of 10 lower (1.6–71 × 10−9). The estimated cancer risks in the 3 areas on the basis of the air emissions of the facilities included in this study did not exceed the Dutch MPR guideline (1 × 10−5 y−1), but the negligible risk guideline (1 × 10−7 y−1) was exceeded in Rijnmond and IJmond. In all study areas, 1 or 2 large facilities were responsible for more than 50% of the calculated cancer risks. Table 3 presents the relative contributions of the individual facilities in the Rijnmond region to the overall cancer risk calculated for industrial emissions. Table 4 gives an overview of the different sources of uncertainty in the emissions inventory and the cancer risk calculations that were explained and discussed in the interpretation framework.

The emissions inventory, risks maps, and interpretation framework were published on the web site www.rechtomteweten.nl (“recht om te weten” is Dutch for right-to-know) and was launched on 29 February 2004. The web site allowed users to enter their ZIP code and subsequently provided a quantitative estimate of the extra cancer risk attributable to air emissions of the selected facilities in the study area. Furthermore, the user could retrieve

  • a risk contour map for each study area (Figures 2 to 4);

  • information on the facilities included in the study and their emissions to air and water;

  • general information on the carcinogenic substances included in the study (use, effects, standards);

  • an explanation of the methods applied for gathering emissions data and calculation of cancer risks;

  • a Frequently Asked Questions (FAQs) section;

  • a forum in which people could send and publish their opinion about the web site; and

  • an overview of links to other right-to-know initiatives, emissions databases, and relevant environmental web sites.

Table Table 2.. Risk estimates that were presented on the web site to facilitate the interpretation of the estimated cancer risks. Background cancer risks were calculated on the basis of air monitoring data for carcinogens (RIVM 2002). Other risk estimates were obtained from different sources (VROM, 1989; Van Dijk-Looijaard 1993; Ragas and Leuven 1997). Most figures reflect the situation in The Netherlands around 1990. Risk standards and background risks in italics
ActivityYearly death risk
Smoking cigarettes (1 package a day lifetime)1 in 200
Riding a motor cycle1 in 1,000
Riding a moped1 in 5,000
Driving a car1 in 5,700
Asbestos (occupational exposure)1 in 16,000
Radiation exposure (general public)1 in 20,000
Cycling1 in 26,000
Pedestrian1 in 54,000
Fire accidents1 in 100,000
Maximum permissible risk level for carcinogenic substances1 in 100,000
Airplane accident1 in 814,000
Hit by lightning1 in 2 million
Calculated background cancer risk for high-intensity traffic road in Dutch city area1 in 2 million
Calculated background cancer risk for Dutch city area1 in 2.5 million
Bee sting1 in 5 million
Calculated background cancer risk for Dutch rural area1 in 6 million
Breaching of a dike1 in 10 million
Negligible risk level for carcinogenic substances1 in 10 million
X-ray in a hospital1 in 700 million
Meteorite impact1 in 1,000 million

People were encouraged to develop a personal opinion about the acceptability of the estimated cancer risks based on the information of the interpretation framework.

The launch of the web site triggered a considerable media response. Announcements and background articles appeared in local, regional, and national newspapers. Several news programs on radio and television announced and discussed the web site. The initiative was also discussed in the Dutch parliament during a debate on the implementation of the United Nations Economic Commission for Europe Aarhus convention.

In the first 5 months, the web site was visited more than 77,000 times. The 1st month was the busiest, with more than 34,000 visits. After that, the number of visits gradually decreased to an average of approximately 100 visits a day. In the first 36 d after the launch of the web site, 189 comments of 164 individuals were placed in the public web site forum. After this period, the web site forum was closed because it took too much time to maintain it. Of the 164 individuals that posted a comment, 39 opposed the initiative and 87 explicitly or implicitly supported the initiative (38 unclassifiable). Requests to extend the initiative to other areas or agents (e.g., radio activity) were considered supportive. Important arguments used by the opponents can be summarized as follows:

  • The web site would strengthen the general misconception that air emissions of some major industrial facilities are the main source of pollutant-related cancer risks.

  • The cancer risks were considered negligible compared with cancer risks from other sources (smoking and food) and too small to justify the investment.

  • The presented risk estimates should not be published because they were too uncertain and based on subjective assumptions.

  • The calculated cancer risks were difficult to interpret for laymen.

  • The web site was considered intimidating and the presentation of cancer risks would cause unnecessary unrest among the general public.

  • The presented cancer risks were considered a severe underestimation of the real cancer risks caused by air emissions of the selected facilities.

Proponents argued that

  • People have the right to know how surrounding facilities can influence their health.

  • The web site presented information on environmental emissions in a meaningful way.

  • The initiative should be broadened to include 1) emissions from other sources (e.g., traffic, agriculture, and households), 2) health effects of noncarcinogenic substances, 3) other areas in The Netherlands, 4) other exposure pathways, and 5) sensitive subgroups within the general population (e.g., children).

  • Withholding information would cause more unrest than the presentation of uncertain information.

The forum responses included 98 individuals (mainly proponents of the web site) expressing some kind of concern about their health or the state of the environment; 53 individuals referring to scientific arguments, facts, or data to motivate their opinion; and 26 individuals referring to risk comparisons.

Four companies responded actively to the initiative, 3 directly and 1 indirectly through the media. One company was frankly positive and argued that the publication of this type of information should be financially supported by companies and should be part of their corporate responsibility strategy. Another company was furious that they were on the web site because they did not emit carcinogenic substances into air. They requested to be removed or they would start a lawsuit. It turned out that a mistake had been made in the selection procedure of the facilities. The company was removed and a rectification was published on the web site. Another company called the initiative “needlessly agitating” in a national newspaper. Finally, 1 company contacted the organization because they did not understand why their facility contributed most to the calculated overall cancer risk in one of the study areas. A consultation round followed and the company provided updated emissions data. These new data were much lower and resulted in a significant reduction of this company's contribution to the overall cancer risk.

Figure Figure 2..

Contours representing annual cancer risk, calculated on the basis of the emissions of 18 carcinogenic substances by 15 selected industrial facilities in the Rijnmond region (20 × 20 km area; The Netherlands).

On 19 July 2004, a new introduction page was added to the web site, stating that the web site would no longer be maintained. The main aim was realized (i.e., to demonstrate that emissions data of large industrial facilities can be presented in a more meaningful way). The web site can still be consulted as a right-to-know demonstration project, but it is stated explicitly on the 1st web page that the emissions and risk data presented are outdated.

Figure Figure 3..

Contours representing annual cancer risk, calculated on the basis of the emissions of 18 carcinogenic substances by 8 selected industrial facilities in the IJmond region (10 × 10 km area; The Netherlands).

Figure Figure 4..

Contours representing annual cancer risk, calculated on the basis of the emissions of 18 carcinogenic substances by 14 selected industrial facilities in the Nijmegen West region (10 × 10 km area; The Netherlands).

DISCUSSION AND LESSONS LEARNED

The main aim of the pilot project was to demonstrate that emissions data of large industrial facilities can be presented in a more meaningful way than raw emissions data. On the basis of user statistics and media attention, the web site can be called a success. The web site triggered a discussion of the accessibility and presentation of environmental risk information in The Netherlands, which even reached the Dutch parliament. The assistant secretary of state on environmental affairs called the web site an incentive to intensify other national and regional initiatives for the disclosure of environmental information.

The development and launch of the web site raised several issues that require consideration for future initiatives. The most difficult phase in the development of the web site was the collection and selection of emissions data. The review of public documents and the consultation of authorities and facilities was a very laborious process. The workload involved could be reduced significantly if the responsible authorities would establish and maintain a more detailed and up-to-date database of national, regional, and local emissions data. Part of the problem was also caused by the limited cooperation of some companies (i.e., only 11 of the 28 facilities responded to the request to check their emissions data). Their reserved attitude might be because the web site was a private initiative of environmental organizations. The companies were not obliged to cooperate, and other priorities might have dominated their agenda. One option to solve this problem is the adoption and legal implementation of this right-to-know initiative by public authorities. Another option is the establishment of an independent body (e.g., a cooperation between environmental organizations, public authorities, and companies) that further develops and maintains the right-to-know initiative (Cohen 2001).

Table Table 3.. Relative contribution of individual industrial facilities to the estimated increased cancer risk caused by the selected industrial facilities in the Rijnmond region. The Netherlands
FacilityActivityRelative contribution (%)
Shin-Etsu BVProduction of PVC39
Shell Nederland Chemie BV (location Pernis)Oil refinery17
Exxon Chemical Holland BV (RAP)Production of aromatic hydrocarbons15
Shell Nederland Raffinaderijen BVOil refinery14
Aluminium&Chemie Rotterdam BVProduction of carbon anodes8
Odfjell Terminals Rotterdam BVStorage of oil products3
Esso Nederland BVOil refinery2
NV Afvalverwerking RijnmondWaste incinerator1
Other facilities 1
Table Table 4.. An impression of the different sources of uncertainty that influence the calculated cancer risks
Source of uncertaintyEffect on risk estimatesIndication of importance
Selection of substances
 Limited number of substances included in study (e.g., no radionuclides)UnderestimationUnknown
 Limited number of standards availableUnderestimationUnknown
Selection of facilities
 Limited number of facilities included in studyUnderestimationUnknown
Collection of emission data
 Only point sources includedUnderestimationUnknown
 Data may be outdatedOver- or underestimationUnknown
 Stack specifications sometimes lacking, assumptions madeOver- or underestimationSmall
 Estimated emission data tend to overestimate real emissionOverestimationUnknown
Air dispersion calculations
 Meteorological dataOver- or underestimationSmall
 Various assumptions (e.g., surface roughness, reception height)Over- or underestimationSmall
Exposure assumptions
 Temporal variations in exposure and risks ignored (yearly averages)Over- or underestimationSmall
 Continuous exposure (24 h/d) assumedOverestimationUnknown
 Averaging over ZIP code areasOver- or underestimationUnknown
 Additive risks for exposure to multiple substances was assumedOver- or underestimationUnknown
 Only exposure through inhalation wasUnderestimationSmall
Standards
 Toxicological and/or epidemiological data used to derive the standardOver- or underestimationLarge
 Interspecies variabilityOver- or underestimationLarge
 Intraspecies variability (e.g., genetic factors and age-related differences)Over- or underestimationLarge
 Low-dose extrapolationOverestimationLarge
Background risks
 Limited number of substances included in the calculationsUnderestimationUnknown
 No site-specific information availableOver- or underestimationUnknown
Risk calculations and spatial presentation
 Adding the cancer risks of different substancesOver- or underestimationSmall
 Interpolation of cancer risks (21 × 21 grid points) to risk contoursOver- or underestimationSmall
 Allocation of cancer risks to ZIP code areasOver- or underestimationMedium

An important issue in the public disclosure of information is whether the internet is a suitable medium to inform the target group. The target group of this pilot project can be loosely described as that part of the population that is interested in or concerned about potential health risks of industrial emissions. Internet access in The Netherlands is relatively high, with 68% of the Dutch population having access through a home PC in the year 2003 (Statistics Netherlands 2005), indicating that a substantial part of the population could potentially be reached. Indeed, the number of web site visits over the first 5 months (77,000) seems relatively high compared with the total number of inhabitants in the 3 areas (-850,000; Statistics Netherlands 2005). Furthermore, that the majority of the respondents on the web site forum expressed some kind of concern about their health or the environment seems to indicate a demand for this kind of information. However, more data are necessary (e.g., about the motivation and backgrounds of all web site visitors) to allow unambiguous conclusions about whether the target group was reached.

Another important issue, which was raised by several respondents to the web site, is whether it is appropriate to confront people with estimated cancer risks, which are highly uncertain because they are based on theoretical models and assumptions. The WHO (2001, p. 28) states that “… risk estimates … should not be regarded as being equivalent to the true cancer risk.” Detailed quantitative estimates on the extent of the uncertainty in low-dose carcinogen potency estimates are scarce, but a factor of 100 does seem a reasonable guesstimate (SCFSC 1980; Crump and Howe 1985; Ragas 2000). However, this considerable uncertainty never withheld regulators and risk assessors to widely use URFs for the estimation and comparison of cancer risks. The main reason is that low-dose cancer risk estimates, with all their uncertainty, reflect our best standard of knowledge. Why then, should this information be withheld from the general public? The imaginary risk someone creates in the absence of information might be much worse than an uncertain risk estimate. Although cancer risk estimates should not be regarded as true cancer risks, this does not imply that the estimates cannot be communicated and explained to the public. The challenge is to communicate the risk estimates so that they are understandable and that the risks and associated uncertainty can be put into a personal perspective (Slovic 2001).

It was beyond the scope of this pilot project to perform a detailed analysis of the understanding of the estimated cancer risks by web site visitors, let alone the general public. Nonetheless, some remarks can be made. In general, probabilistic statements are relatively well understood by the general public (Durant et al. 1989), but risk comparisons involve many limitations, such as failure to address uncertainty, failure to address the broad quantitative dimensions that define and measure risk, and failure to consider the broad set of qualitative dimensions that underlie people's concerns about the acceptability of risks and technologies (Covello 1989). That 26 of the respondents in the web site forum referred to risk comparisons indicates that the information provided on the web site was used by some visitors to put the estimated risks into perspective, but it does not indicate whether the risk information was understood. Given the numerous assumptions and uncertainties involved in the calculation of the cancer risk estimates (Table 4) and the aforementioned limitations of risk comparisons, it seems highly unlikely that all or most visitors understood the information provided. However, Tesh (1999) argues that citizens express their perception of risks largely through organized citizen groups, and that these groups employ and have access to many experts. If this is true, it might not be necessary to develop a right-to-know web site that is understandable for all members of the general public (in all its diversity). Additional research in this area is necessary to unravel the complex processes that underlie the understanding and perception of environmental risks—individually as well as in a broader social context.

Some practical improvements in the presentation of the web site and the estimated cancer risks can be identified on the basis of comments posted on the web site forum. Several respondents considered the pictures of smoking stacks presented on the web site intimidating. To avoid the impression of intimidation, this type of picture should not be used. Other forum respondents complained that the risk estimates and interpretation framework were presented on separate web pages, which complicated the interpretation of the results. This could be improved by presenting the risk estimates, associated uncertainties and the risks of the interpretation framework (background, standards, and daily life risks) in an integrated risk ladder (Lipkus and Hollands 1999), as illustrated in Figure 5. The integrative nature of this graph facilitates the comparison of risks, although it still contains some elements (e.g., the logarithmic scale) that could be difficult to grasp.

Figure Figure 5..

Example risk ladder for the integrated presentation of the estimated annual cancer risk, associated uncertainties, and the risks of the interpretation framework (background, standards, and daily life risks). The 95% confidence interval of the estimated cancer risks is indicated by arrow-shaped error bars. Maximum permissible risk (MPR) and negligible risk (NR) indicate Dutch risk standards for exposure to a mixture of carcinogens and correspond to an annual cancer risk of 1 in 100,000 and 1 in 10,000,000, respectively.

CONCLUSIONS

The development and launch of this right-to-know web site seems a promising step forward in the public disclosure of environmental risk information. To further improve and maintain these types of right-to-know initiatives, it is particularly important that 1) responsible authorities establish and maintain a detailed and up-to-date database of national, regional, and local emissions data; 2) the initiative is implemented in a legal framework or adopted by an independent body; and 3) more knowledge becomes available about the understanding and effects of environmental risk information on the general public.

Acknowledgements

This research was partly funded by the Dutch Ministry of Housing, Spatial Planning, and the Environment. The authors thank Els Kranenborg and Frank ter Beek.

Ancillary