Abstract

In recent years, the concern for protection of urban populations against terror attacks involving radiological, biological or chemical substances has attracted increasing attention. It sets new demands to decision support and consequence assessment tools, where the focus has traditionally been on accidental exposure. The aim of the present study was to illustrate issues that need to be considered in evaluating the radiological consequences of a ‘dirty bomb’ explosion. This is done through a worked example of simplified calculations of relative dose contributions for a specific ‘dirty bomb’ scenario leading to atmospheric dispersion of 90Sr contamination over a city area. Also, the requirements of atmospheric dispersion models for such scenarios are discussed.

INTRODUCTION

In the event of a terrorist attack leading to airborne dispersion over a large city area of chemical, biological or radiological (CBR) contaminants, it is essential to decision makers, planners and operational units on different levels to effectively and rapidly obtain an overview of the situation. This will ensure that resources are applied optimally to reduce hazards. It is equally valuable in advance of any contaminating incident to know what the hazards might be, so that a targeted operational preparedness can be developed and undue anxiety in the population can be avoided. This requires the use of reliable and detailed consequence assessment models.

Plausible CBR scenarios often have a number of parametric similarities, e.g. with respect to particle sizes and environmental migrations. For instance, bacteria typically have a size ranging from a few hundred nanometres to a few tens of micrometres, which is practically identical to the size range of particles that might be dispersed over a large area after a well-executed ‘dirty bomb’ explosion. Also nebulisation of harmful chemicals would be expected to generate particles of about those sizes (1,2).

To illustrate the complexity of required considerations and parameterisation in modelling, a simplified example of estimation of the relative contributions to radiation dose that might be received outside the immediate blast area of a ‘dirty bomb’ dispersing a strong 90Sr source in a ceramic matrix, taken from a radioisotope thermoelectric generator, is given. Also, the needs for detailed atmospheric dispersion models to adequately describe contaminant plume concentrations are discussed in relation to this scenario.

METHODS AND RESULTS

Scenario description

In the past, European standard decision support tools such as RODOS(3) and ARGOS(4) have practically exclusively focused on prediction of the radiological implications of accidental atmospheric releases from nuclear power plants. The attention here is, in general, on a limited number of radionuclides known to be present in a given reactor with well-defined initial physicochemical forms, even though actual release fractions and characteristics of dispersed contaminants could vary considerably depending on the type and extent of the accident. In contrast, in the event of a ‘dirty bomb’ explosion, a very large number of different radionuclides could, in principle, be dispersed. However, in reality, the possibilities for using a source that would have a significant harmful effect if dispersed over a wide city area would probably be very limited due to various reasons.

A first requirement is the availability of applicable sources. Acknowledging that the likelihood of terrorists gaining access to facilities for production of strong radioactive sources would in general be small, the existing sources would need to be applied. Such sources might be stolen from an authorized user or manufacturer. They might possibly also be bought, either by successfully pretending to be a legitimate user with a peaceful industrial purpose(5) or in a black market that might exist, considering the number of strong sources that are reported lost each year over the world. For instance, in the former Soviet Union alone, it is estimated that thousands of comparatively high-risk category sources are currently ‘orphaned’ or lost without a trace(6). Even so, it is very likely that only few of these are strong enough to potentially cause real harm through atmospheric dispersion over an inhabited area. Table 1 shows a series of examples of existing strong sources(1,6,7).

Table 1.

Some important radionuclides that might be of concern in connection with ‘dirty bombs’, including typical physicochemical forms of large existing sources and maximum activity estimates.

Radionuclide Typical physicochemical form of large existing sources Existing strong sources and their activities 
60Co Metal Sterilisation irradiator (up to 400 000 TBq). Teletherapy source (up to 1000 TBq) 
90Sr Ceramic (SrTiO3)—insoluble, brittle, soft (Mohs hardness: 5.5) Radioisotope thermoelectric generator (1000–10 000 TBq) 
137Cs Salt (CsCl), soluble Sterilisation irradiator (up to 400 000 TBq). Teletherapy source (up to 1000 TBq) 
192Ir Metal—soft—Mohs hardness 6.5 Industrial radiography source (up to 50 TBq) 
226Ra Salt (RaSO4) very low solubility Old therapy source (up to 5 TBq) 
238Pu Ceramic (PuO2)—insoluble Radioisotope thermoelectric generator (up to 5,000 TBq) 
241Am Pressed ceramic powder (AmO2Well logging source (up to 1 TBq). 
252Cf Ceramic (Cf2O3)—insoluble Well logging source (up to 0.1 TBq). 
Radionuclide Typical physicochemical form of large existing sources Existing strong sources and their activities 
60Co Metal Sterilisation irradiator (up to 400 000 TBq). Teletherapy source (up to 1000 TBq) 
90Sr Ceramic (SrTiO3)—insoluble, brittle, soft (Mohs hardness: 5.5) Radioisotope thermoelectric generator (1000–10 000 TBq) 
137Cs Salt (CsCl), soluble Sterilisation irradiator (up to 400 000 TBq). Teletherapy source (up to 1000 TBq) 
192Ir Metal—soft—Mohs hardness 6.5 Industrial radiography source (up to 50 TBq) 
226Ra Salt (RaSO4) very low solubility Old therapy source (up to 5 TBq) 
238Pu Ceramic (PuO2)—insoluble Radioisotope thermoelectric generator (up to 5,000 TBq) 
241Am Pressed ceramic powder (AmO2Well logging source (up to 1 TBq). 
252Cf Ceramic (Cf2O3)—insoluble Well logging source (up to 0.1 TBq). 

In any case, the original application of the source would, to a great extent, determine its characteristics, including the activity and initial physicochemical form. The physicochemical forms, as well as the elemental properties of the radioactive matter, are evidently important in determining the dispersibility of contaminants from a ‘dirty bomb’. According to recent experimentation by Harper et al.(1), for instance extremely little of a metallic 60Co source would be aerosolized in an explosion. However, if a source is in ceramic form, aerosolization fractions ranging between 2 and 40% have been reported, e.g., depending on the construction of the explosive device. This was in general found to produce a particle size spectrum with much of the contaminant masses are in the range of 30–100 µm, and a smaller peak in the range of a few micrometres.

This type of size distribution was also measured in US blast experiments impacting on soil(8), and after the Thule accident in 1968. The Thule explosion was essentially similar to that of a ‘dirty bomb’: a conventional explosion dispersing a solid, perhaps ceramic, radioactive material with a very high melting point. In the Thule case, only 1.3% of the particles were larger than ca. 18 µm, but these carried nearly 80% of the activity(9).

In the context of atmospheric dispersion over larger areas, it is the smaller particles that are of concern, since gravitational settling of large particles will occur very rapidly and over short distance. As an example, Hage(10) performed a series of experiments dispersing glass microspheres with a mean mass diameter of ca. 50–100 µm from 15 m above a flat prairie area. These experiments were conducted under carefully monitored conditions and showed that with an average wind speed at the point of release of about 5 m s−1, practically all the contaminants (>99 %) would deposit within a downwind distance of <300 m. This means that the air concentration of these large particles would decline greatly over the first minute. In contrast, small particles, on which gravity would have considerably less influence, would remain airborne over much longer periods of time and could reach correspondingly greater distances from the release point.

An example of strong and carelessly ‘orphaned’ sources is the 90Sr elements (typically in the range of 1000–10,000 TBq), originally applied in radioisotope thermoelectric generators. These have in recent years on several occasions unwittingly been retrieved by village inhabitants in the former Soviet Union(11). For comparison, it has been estimated that the total release of 90Sr from the Chernobyl accident was about 8000 TBq(12). However, it should be mentioned that such strong sources might be exceedingly difficult to handle and bring to dispersion, even if involved terrorists may not consider it important to survive themselves. Even though the beta particles from 90Sr/90Y might require less shielding than the gamma radiation from other strong sources, the Bremsstrahlung problem could also be substantial.

For the purpose of defining an illustrative scenario for a demonstration example of estimation of dose contributions after a ‘dirty bomb’ attack in a city area, it is assumed that the contaminants originate from a 1000-TBq 90Sr source in a ceramic matrix. The contamination within the immediate vicinity of the blast would, to a great extent, be dominated by large shrapnel. Therefore, the contamination pattern within that small area, which would anyway be very extensively monitored, would be impossible to model. However, as discussed above, a proportion of the radioactive matter would become aerosolised as fairly small particles, which might contaminate a large and complex city area.

Monitoring and mapping of beta activity over a large city area would be an extremely demanding and time-consuming task. Consequently, a reliable case-specific modelling would thus be very valuable, not only for pre-incident planning, but also for real-time impact evaluations. This is addressed in new developments for the ARGOS system. This study will therefore focus on doses received from the atmospheric dispersion of contaminant aerosols over a large city area.

On the basis of experience described above with explosions involving ceramic material, it is assumed that it would be the peak in the lower end of the aerosol size spectrum that would be important to follow over this area. Since the largest of the particles in that peak would carry the greatest contaminant mass, it is, for simplicity, assumed that the particles of concern are in the range of 5–10 µm. Such particles would in general be inhalable(13). According to the results presented by Harper et al.(1), it would seem reasonable to assume that 5% of the total activity could be dispersed as inhalable aerosol.

It is further assumed that the weather is dry on the day of the attack. Internal doses from inhalation of resuspended contaminated dust as well as external doses from the passing contaminated plume would be likely to be of comparatively little significance(14–17) and are not dealt within this paper. Also, dose contributions from ingestion of contaminated food are left out, as food is normally produced outside city areas, and often kept sealed in shops and homes.

It should be noted that it is assumed in the various dose calculations that no countermeasures are implemented. Such calculations can provide the required knowledge to understand the severity of a given emergency scenario and the possible need for intervention. A wide range of countermeasures might be considered to reduce different dose contributions. The extended ARGOS system will enable the evaluation of the effect of different strategies for reduction of doses after a radiological attack.

External dose from contamination on outdoor surfaces

The dose rate received at the depth of the basal layer of the epidermis of human skin from a large ground surface uniformly contaminated with 90Sr/90Y has, probably highly conservatively, been estimated by HPA-RPD(18) for decision-making purposes to be some 4 × 10−11 Sv h−1 per Bq m−2. This estimate is applied for the calculations, although the exact assumptions regarding geometries are not quite clear, and more detailed calculations for different distances and thicknesses of clothing are called for.

A requirement to reach as high dose rates to the skin as this would be that the contamination lies on the very surface of the ground. If it were 1 cm down in soil, the shielding effect would be so great that the dose rate to the skin would be about 3 orders of magnitude lower(19). The dose rate contribution would thus probably be reduced to a negligible value at the time when the first heavy rain shower arises, and the contamination is washed slightly down in the soil(20). For the purpose of this example, it is assumed that this happens 10 d after the deposition. Doses to inner organs would typically be at least 3 orders of magnitude lower, and thus are likely to be of little significance(19).

On the basis of a number of independent surveys in Western Europe and in California(21–23), it seems reasonable to assume that people on average spend some 15% of the time outdoors. This would give a skin dose contribution estimate of 15%×4 × 10−11 Sv h−1 per Bq m−2×10 d×24 h d−1 ∼ 1×10−9 Sv per Bq m−2.

Contaminant particles will naturally also deposit on other outdoor surfaces. As gravitational settling will largely govern the process of deposition for these supermicroneous particles, horizontal surfaces will receive much higher levels of contamination than vertical. Also the limited time normally spent outdoors very close to a wall would make the contamination on these surfaces less important from a beta dose perspective.

On paved horizontal surfaces, large and insoluble particles will much more rapidly be removed by natural weathering processes than will smaller particles(24,25). This is at the extreme illustrated by the very high decontamination factor of 50, obtained by Clark and Cobbin(26), when hosing water on a street contaminated by insoluble 44–100 µm particles. If the contamination had instead been in the form of submicroneous soluble Cs particles, as seen after the Chernobyl accident, hosing at the same pressure would be expected to only remove about half of the contamination(27). Routine street cleaning together with natural weathering processes would therefore be likely to result in low street contamination levels over only days or weeks.

For simplicity, only dose contributions from contamination on the soil are considered here, but an assessment of the post-deposition behaviour of various types of supermicroneous particles on other surfaces in the urban environment is called for to provide an adequate background for decision making.

External dose from contamination on indoor surfaces

It is here assumed that windows and doors are closed during the contaminating episode. For supermicroneous particles, it is not unreasonable to assume that nearly all the deposited indoor contaminants will be distributed on the horizontal surfaces, primarily the floor(28).

On the basis of the findings of Roed and Cannell(29), it can be shown that the deposition, Di, per unit indoor area can be estimated from the formula:  

formula
where D0 is the corresponding deposition on an outdoor reference surface (e.g. a cut lawn), vd,0 is the deposition velocity to that reference surface, h is the room height, f is the filtering factor, λd is the rate coefficient of deposition indoors and λv is the rate coefficient of ventilation of the dwelling. For particles in the relevant size range, λd would be high compared with λv, so that Di/D0hfλv/vd,0. With typical values for ca. 5–10 µm particles used in the new version of the ARGOS system, Di/D0 is ∼0.09.

A reasonable estimate of the dose rate can be obtained using the same dose conversion factor as for outdoor horizontal surfaces. In reality, indoor floor surfaces will be likely to be considerably smaller than open outdoor horizontal surfaces, but beta doses come from contamination within short distance.

A likely half-life of the natural removal process of the smallest relevant particles on the floor would be of the order of a few months(14). Here, a value of 60 d was assumed.

The total external dose to the skin from the contaminated 5–10 µm particles deposited on indoor surfaces (taking into account the fractions of time spent indoors and outdoors) would then be estimated to be of the order of (60 d/10 d) · 0.09 · 0.85/0.15 = 3 times that from contaminants deposited on outdoor surfaces. In other words, this would amount to 3 × 10−9 Sv per Bq m−2 on the horizontal outdoor reference surface.

It should be noted that fabric can shield significantly against beta radiation (although the high-energy beta rays from 90Y can penetrate thin layers). In general, the most critical situations would be those where unshielded skin comes into direct or close contact with a contaminated surface. One important situation is at night, when, e.g., the face is in direct contact with a possibly contaminated surface for hours. For this scenario, the applied dose conversion factor might well be much too low. However, by washing the pillow case regularly, these doses would be limited to a short period of time after the contamination took place, as ordinary machine washing is efficient in removing the contaminants(30).

External dose from contamination on human skin

The most critical exposure of human skin would occur to persons who are outdoors during virtually the entire period of deposition, and this is what is assumed. If people were indoors with windows and doors closed, the deposition of 5–10 µm particles would typically be at least one order of magnitude less(14).

Deposition velocities of 5–10 µm particles to human skin, as parameterised in ARGOS, are high, but thorough washing will be efficient in removing these particles from skin. However, if nothing is done to actively remove the particles (e.g. when no early warning is given), the natural removal of 5–10 µm particles would be expected to occur with a half-life of ca. 0.2 d(31).

Using this clearance half-life together with a data library value for absorbed dose rate at a depth in tissue corresponding to that of the basal layer of the epidermis, the dose from contamination with 90Sr particles on freely exposed skin would here amount to 8 × 10−6 Sv per Bq cm−2(14). This, multiplied by the relationship between deposition velocities on skin and the grassed reference surface, gives a skin dose estimate in ARGOS of 4 × 10−9 Sv per Bq m−2 on the reference surface.

Corresponding skin doses from contamination on clothing would be considerably smaller, both because clothing can, dependent on thickness, shield very well against the beta radiation, and because the typical deposition velocity to clothing of aerosols of this size is about a factor of 5 less than that to skin(32). Although the natural clearance half-life of these contaminants from clothing would typically be somewhat longer than that from skin, the dose contribution from the 90Sr particles deposited on clothing would be expected to be at least one order of magnitude less than that from deposition to uncovered skin(14).

Committed dose from inhalation during plume passage

For people staying outdoors during the plume passage, the inhalation doses can be estimated by multiplying time-integrated contaminant air concentrations by a dose conversion factor(33) and an inhalation rate. The particles in the range of 5–10 µm are all assumed to be inhalable(13). The time-integrated air concentration outdoors, Coint, that would lead to a deposition of 1 Bq m−2 on an outdoor grassed reference surface can be found by dividing the 1 Bq m−2 by the deposition velocity to the reference surface. By multiplication of this time-integrated air concentration with an inhalation rate (3.3 × 10−4 m3 s−1)(13), the total amount of inhaled contamination can be found.

According to ICRP(33), the committed dose by inhalation of 1 Bq of 90Sr would be 1.6 × 10−7 Sv if the ‘slow’ absorption class ‘S’ is assumed. By multiplying the inhaled amount by this factor, it is found that inhalation in this case gives a committed dose of 3 × 10−9 Sv per Bq m−2 on the reference surface. It should be stressed that this dose would be much smaller if people stayed indoors during the time period of elevated air concentrations, and if the aerosol had been more readily soluble. Also, if it rained during the plume passage, this could significantly deplete the local plume concentrations. Rain would instead enhance the deposition on outdoor surfaces, as observed after the Chernobyl accident(20).

Atmospheric dispersion modelling

The sections above give a demonstration of some considerations required for estimating dose contributions from a ‘dirty bomb’ explosion, relative to a given contamination level on a reference surface in the environment. This type of considerations are used to define parametric requirements for the extension of ARGOS to estimate consequences of ‘dirty bomb’ explosions. The creation of libraries of relevant parameter values has been initiated.

Also, a new highly detailed model for urban atmospheric dispersion and deposition of the different types of particles from ‘dirty bombs’ is being created for implementation in ARGOS. For this dispersion model, a methodology has been derived, primarily on the basis of experimentation, to determine the likely initial contaminant cloud dimensions from a ‘dirty bomb’ as a function of the applied explosive mass. Further, new data libraries have been generated describing the deposition of particles of the different relevant sizes on all different urban surfaces for five different weather categories at deposition: light rain, heavy rain, dry weather, dry weather with snow cover and snowfall(34).

Other than accommodating a wider range of particle sizes, the new ARGOS dispersion model, urban release and dispersion (URD), also includes a new high-resolution atmospheric transport module developed at the Technical University of Denmark. The module is in some of its features inspired by the British code urban dispersion model (UDM)(35). This new implementation was necessitated since it has become apparent that it would be problematic to use the mesoscale transport modules in decision support systems like RODOS and ARGOS for the estimation of contaminant dispersion after a comparatively low-altitude in-town release, which would be expected from for instance a ‘dirty bomb’ blast. In larger scale atmospheric dispersion models, such as ATSTEP(36) and RIMPUFF(37), inhabited areas are simply modelled as areas with enhanced surface roughness and different deposition rates compared with open areas. This is adequate for higher altitude longer range dispersion, e.g. following a large nuclear power plant accident. However, over areas near the point of a comparatively low-altitude release, the plume interaction with buildings and other obstacles and dispersion through street canyons can be important factors in determining the plume shape and dispersion pattern.

This is illustrated by the example shown in Figures 1, 2 and 3. All of these figures show URD calculations of time-integrated air concentration at ground level in an urban area (Frederiksberg, part of Copenhagen, Denmark) of aerosols of size relevant to the scenario, released at a height of 2 m above ground. In this case, the total released amount of activity is 1 TBq. The amount of dry deposition of aerosols is proportional to the time-integrated concentration, the proportionality factor being a function of the land cover, while for larger particles gravity becomes dominant. In URD, the presence of buildings has three effects: it limits the size of the horizontal eddies present in the atmosphere and thereby limits the large scale horizontal dispersion, it enhances the smaller scale dispersion by creating more small scale turbulence and it delays part of the dispersion by retaining aerosols in more or less stationary recirculation zones on the downstream side of the buildings.

Figure 1.

RIMPUFF estimate of the time-integrated concentration (Bq s m−3) at ground level. No building influence. Frederiksberg case. Total release: 1 TBq.

Figure 1.

RIMPUFF estimate of the time-integrated concentration (Bq s m−3) at ground level. No building influence. Frederiksberg case. Total release: 1 TBq.

Figure 1 shows an estimate made with the RIMPUFF mesoscale model. Figure 2 shows a corresponding estimate with URD, where buildings are only modelled to limit the large scale dispersion, whereas the other two building effects are not accounted for. In Figure 3 all building effects are included, and it is seen that especially the retaining of matter behind the buildings is of importance. Figure 4 is an excerpt from Google Earth of the area in which the calculation area is located.

Figure 2.

URD estimate of the time-integrated air concentration (Bq s m−3) at ground level, with the buildings just limiting the large scale dispersion. Frederiksberg case. Total release: 1 TBq.

Figure 2.

URD estimate of the time-integrated air concentration (Bq s m−3) at ground level, with the buildings just limiting the large scale dispersion. Frederiksberg case. Total release: 1 TBq.

Figure 3.

URD estimate of the time-integrated air concentration (Bq s m−3) at ground level, with all three building effects modelled. Frederiksberg case. Total release: 1 TBq.

Figure 3.

URD estimate of the time-integrated air concentration (Bq s m−3) at ground level, with all three building effects modelled. Frederiksberg case. Total release: 1 TBq.

Figure 4.

Google Earth picture of part of Frederiksberg, the calculation area indicated by the white quadrant.

Figure 4.

Google Earth picture of part of Frederiksberg, the calculation area indicated by the white quadrant.

Doses from dispersed contaminants

Figure 3 shows that each teraBecquerel release of contaminants would result, under the given circumstances and assumptions over a small area (about 250 m by 50 m), in a time-integrated ground level air concentration exceeding 108 Bq s m−3. With the assumed 1000 TBq release and the simplifying assumption of an aerosol fraction of 5%, this would correspond to a level exceeding 5 × 109 Bq s m−3. However, in this small area, contamination levels would be likely to be dominated by the inhomogeneous and unpredictable deposition of shrapnel and very large ballistically dispersed particles. Such an area would undoubtedly rapidly be evacuated and carefully monitored.

However, in areas stretching over ca. 2.5 km distance, the time-integrated ground level concentration would exceed 5 × 107 Bq s m−3, according to the scenario calculations. It must here be stressed that in reality, the initial spatial distribution of the contaminated cloud created by an explosion would particularly depend on the explosive mass applied, and that only a limited fraction of the aerosol would effectively be released at an altitude of about 2 m. The assumed release from a point is a very simplified approach, which will, as indicated above, not be applied in the new version of the ARGOS system. A low effective release height would generally limit the area over which the contaminants are dispersed.

In an area with a time-integrated ground level concentration range of 5 × 107–5 × 109 Bq s m−3, the total deposition on a grassed reference surface would amount to some 0.1–10 MBq m−2, according to the above assumptions regarding particle size(38). With the given dose conversion factor estimates, this would result in the dose contributions shown in Table 2. It is here for simplicity assumed that people are present at a given distance throughout the entire period of exposure.

Table 2.

Estimated maximum dose contributions in the wind direction at different distances.

Exposure contribution Assumed effective time period of exposure Max. dose at 250 m in wind direction (mSv)a Max. dose at 2500 m in wind direction (mSv)b 
External exposure from outdoor contamination 10 d 10 0.1 
External exposure from indoor contamination Half-life: 60 d 30 0.3 
External exposure from skin contamination Half-life: 0.2 d 40 0.4 
Exposure from inhalation of contaminants Few hours 30 0.3 
Exposure contribution Assumed effective time period of exposure Max. dose at 250 m in wind direction (mSv)a Max. dose at 2500 m in wind direction (mSv)b 
External exposure from outdoor contamination 10 d 10 0.1 
External exposure from indoor contamination Half-life: 60 d 30 0.3 
External exposure from skin contamination Half-life: 0.2 d 40 0.4 
Exposure from inhalation of contaminants Few hours 30 0.3 

Illustrative calculations based on the much simplified scenario involving a 1000 TBq 90Sr ‘dirty bomb’ detonation in Frederiksberg.

aGround contamination level: ca. 10 MBq m−2.

bGround contamination level: ca. 0.1 MBq m−2.

DISCUSSION AND CONCLUSIONS

The above simplified calculations might give the impression that skin beta doses from contamination on outdoor surfaces, skin beta doses from contamination on indoor surfaces, skin beta doses from contamination on human skin and committed doses from inhalation during the passage of the plume from a ‘dirty bomb’ explosion are of the same order of magnitude. However, it should, of course, be noted that the skin is not the most radiation sensitive of human organs (dose weighting factor of only 0.01(39)). The threshold for acute radiation effects by exposure of large areas of skin is generally assumed to be some 20 Sv, even though acute tissue breakdown has been reported for a dose as low as 15 Sv(40).

Evacuation might well more or less automatically be effectuated in an early phase for the people living within the ‘ballistic’ range of the blast (i.e. a few hundred metres). However, the main concern should be to keep inhalation doses down, if at all possible. Inhalation doses exceeding 30 mSv certainly give cause for alarm, and demonstrate a need for action. Here, sheltering is effective and comparatively uncomplicated. However, in the most contaminated areas, it requires extremely early implementation to be helpful. If the aerosols of concern were, as assumed, in the range of 5–10 µm size, the relationship between indoor and outdoor contaminant air concentrations would typically be about 0.02 at equilibrium(34), if windows, doors and forced ventilation systems are closed off.

According to these simplified scenario calculations, individual doses of several millisievert could also be received at considerably greater distances, which merit sheltering of much larger areas and possibly even countermeasures for reduction of external beta dose. Again, it should be noted that in reality, much of the contamination would be likely to have a significantly higher effective release altitude after an explosion, which would mean that the contamination would be dispersed over a greater area(41).

Also an early warning of inhabitants to wash thoroughly would be useful in reducing doses from skin/hair contamination. Further, it should be stressed that the justification and optimisation of countermeasures is a complex task, since social disruption and other indirect costs must be taken into account(39). Considering the limited dimensions of contaminated areas, compared with, for example, what was seen after the Chernobyl accident, authorities may well in a longer term decide to follow a public demand to decontaminate areas with even modest levels of contamination. Features of decontamination techniques for both indoor and outdoor surfaces in inhabited areas have been described in detail in standardized templates(27,42). However, a problem with existing data compilations is that much of the information was derived from the Chernobyl experience, where particles were substantially smaller and more readily soluble, and thus more difficult to remove from surfaces. An effort is thus needed in this context.

The most important lesson from the above considerations is that the proper estimation of each dose component depends on the values of a number of parameters, which would vary considerably between different conceivable ‘dirty bomb’ explosion scenarios.

The values calculated for the specific demonstration scenario reflect a wide range of simplifying case-specific assumptions, and a completely different level and distribution of dose contributions would be envisaged for other ‘dirty bomb’ scenarios. For instance, the dispersion of gamma emitting radionuclides could led to a much more long-lasting and severe outdoor external exposure problem.

Importantly, however, the scenario evaluations and considerations demonstrate the need for a number of data sets to properly estimate the different dose contributions that would be received after the detonation of a ‘dirty bomb’ in a city area. For instance, as mentioned above, a data library of parameters describing the deposition to the different surfaces in an inhabited area of particles of different relevant sizes, for different weather conditions, has already been created for use in a new version of the ARGOS system for evaluating the dose implications of ‘dirty bomb’ attacks. Also a compilation of factors determining the post-depositional migration in the inhabited environment of contaminants with different relevant physicochemical characteristics (notably particle sizes) is needed. Further, there is a need for a series of dose conversion factors for inhabited environments, for instance, for beta doses from 90Sr/90Y on different surfaces in different geometries, but also for gamma doses from radionuclides that are not traditionally envisaged in connection with nuclear power plant accidents, but would be considered of relevance in connection with a terror attack.

It is also important to note that the dose distribution following a ‘dirty bomb’ attack would depend on factors such as physicochemical forms of dispersed materials, radionuclides, weather, season, human behaviour, area type and protection offered by dwellings and ordinary clothing, which can all be ideally addressed in a computerised decision support system.

Without the use of a detailed and complex assessment tool, it is exceedingly difficult to judge the significance of the direct health implications of a given ‘dirty bomb’ attack scenario. In any case, it is essential for decision makers and authorities to be equipped with as good a background material as possible for understanding the implications of different types of ‘dirty bomb’ scenarios. This would help in issuing early and reliable information that could prevent undue anxiety and social disruption in an emergency situation. As mentioned above, also real-time prognoses of the contamination of a complex city area could be very useful for decision makers, planners and operational units in the event of an attack. This is targeted in an extended version of the ARGOS decision support system.

The implementation in ARGOS of a series of state-of-the-art data sets based on a combination of experimentation and theory, together with the described high-resolution dispersion features eliminating errors arising from the interaction of a contaminated low-altitude plume with structures in the environment, will be important instruments in securing reliable consequence prognoses.

Disclosure of potentially security sensitive parameters in ARGOS, such as plume dimensions and deposition relations on surfaces, will be subject to strict clearance procedures.

REFERENCES

1
Harper
F. T.
Musolino
S. V.
Wente
W. B.
Realistic radiological dispersal device hazard boundaries and ramifications for early consequence management decisions
Health Phys
 , 
2007
, vol. 
93
 
1
(pg. 
1
-
16
)
2
Byrne
M.
Aerosol exposure
Chem. Britain
 , 
1998
(pg. 
23
-
26
1998
3
RODOS Homepage
Realtime Online Decision Support System for Nuclear Emergency Management
2008
 
4
ARGOS CBRN Whitepaper
Decision Support for Emergency Management
2008
 
5
Anon
Fake business buys dirty bomb material
Chem. Eng. News
 , 
2007
, vol. 
85
 
29
pg. 
24
 
6
Ferguson
C. D.
Kazi
T.
Perera
J.
Occasional
 , 
2003
Monterey, CA
Center for Nonproliferation Studies, Monterey Institute of International Studies
 
Commercial radioactive sources Paper no. 11 ISBN 1-885350-06-6
7
Argonne Radiological dispersal device (RDD)
Human Health Fact Sheet
 , 
2005
EVS, USA
Argonne National Laboratory
8
Pinnick
R. G.
Fernandez
G.
Hinds
B. D.
Explosion dust particle size measurements
Appl. Optics
 , 
1983
, vol. 
22
 
1
(pg. 
95
-
102
)
9
Eriksson
M.
On Weapons Plutonium in the Arctic Environment
2002
Denmark
Riso National Laboratory
 
(Thule, Greenland), Riso report Riso-R-1321 ISBN 87-550-3006-8
10
Hage
K. D.
On the dispersion of large particles from a 15-m source in the atmosphere
J. Meteorol
 , 
1961
, vol. 
18
 (pg. 
534
-
539
)
11
King
G.
Dirty Bomb: Weapon of Mass Disruption
 , 
2004
Romford
Chamberlain Brothers, Penguin Group
 
ISBN 1-59609-000-6
12
Sohier
A.
A European manual for ‘Off-site Emergency Planning and Response to Nuclear Accidents’
 , 
2002
Belgian Nuclear Research Centre
 
SCK•CEN Report R-3594
13
International Commission on Radiological Protection
Human Respiratory Tract Model for Radiological Protection
 , 
1993
Pergamon Press, Oxford, New York, Tokyo
 
Publication 66 ISBN 0-08-041154-1.
14
Andersson
K. G.
Roed
J.
Byrne
M. A.
Hession
H.
Clark
P.
Elahi
E.
Byskov
A.
Hou
X. L.
Prip
H.
Olsen
S. K.
Roed
T.
Airborne Contamination in the Indoor Environment and Its Implications for Dose
Risoe-R-1462(EN)
 , 
2004
Roskilde, Denmark
Risoe National Laboratory
 
ISBN 87-550-3317-2.
15
Wagenpfeil
F.
Paretzke
H. G.
Peres
J. M.
Tschiersch
J.
Resuspension of coarse particles in the region of Chernobyl
Atmos. Environ
 , 
1999
, vol. 
33
 (pg. 
3313
-
3323
)
16
Andersson
K. G.
Roed
J.
Estimation of doses received in a dry-contaminated living area in the Bryansk Region, Russia, since the Chernobyl accident
J. Environ. Radioact
 , 
2006
, vol. 
85
 
2–3
(pg. 
228
-
240
)
17
Jones
J. A.
, et al.  . 
Description of inhabited area model
2008
Chilton, UK
Health Protection Agency, Centre for Radiation, Chemical and Environmental Hazards
 
Final report of the EC-EURANOS project activity CAT2RTD02, EURANOS(CAT2)-TN(05)-04
18
HPA-RPD
UK recovery handbook for radiation incidents
HPA-RPD-002
 , 
2005
Chilton, UK
Health Protection Agency – Radiological Protection Division
19
Eckerman
K. F.
Ryman
J. C.
External exposure to radionuclides in air, water and soil
1993
 
Federal Guidance Report 12, US EPA, EPA-402-R-93-081
20
Roed
J.
Deposition and removal of radioactive substances in an urban area
1990
Nordic Liaison Committee for Atomic Energy
 
Final Report of the NKA Project AKTU-245 ISBN 87 7303 514 9
21
Jenkins
P. L.
Phillips
T. J.
Mulberg
E. J.
Hui
S. P.
Activity patterns of Californians: use of and proximity to indoor pollutant sources
Atmos. Environ
 , 
1992
, vol. 
26A
 
12
(pg. 
2141
-
2148
)
22
Andersson
K. G.
Evaluation of early phase nuclear accident clean-up procedures for Nordic residential areas
1996
 
NKS Report NKS/EKO-5(96)18 (Nordic Nuclear Safety Research (NKS), Roskilde, Denmark) ISBN 87-550-2250-2.
23
Kousa
A.
Kukkonen
J.
Karppinen
A.
Aarnio
P.
Koskentalo
T.
A model for evaluating the population exposure to ambient air pollution in an urban area
Atmos. Environ
 , 
2002
, vol. 
36
 (pg. 
2109
-
2119
)
24
Sartor
J. D.
Boyd
G. B.
Agardy
F. J.
Water pollution aspects of street surface contaminants
J. WPCF
 , 
1974
, vol. 
46
 
3
(pg. 
456
-
467
)
25
Roed
J.
Andersson
K. G.
Clean-up of urban areas in the CIS countries contaminated by Chernobyl fallout
J. Environ. Radioact
 , 
1996
, vol. 
33
 
2
(pg. 
107
-
116
)
26
Clark
D. E.
Cobbin
W. C.
Removal of simulated fallout from pavements by conventional street flushers
US Naval Radiological Defence Laboratory
 , 
1964
 
USNRDL-TR-797
27
Brown
J.
, et al.  . 
Generic handbook for assisting in the management of contaminated inhabited areas in Europe following a radiological emergency, parts I–V
2007
UK
Radiation Protection Division, Health Protection Agency
 
Final report of the EC-EURANOS project activities CAT1RTD02 and CAT1RTD04, EURANOS(CAT1)-RP(07)-02
28
Lange
C.
Indoor deposition and the protective effect of houses against airborne pollution
1995
Denmark
Risø National Laboratory
 
PhD Thesis
29
Roed
J.
Cannell
R. J.
Relationship between indoor and outdoor aerosol concentration following the Chernobyl accident
Radiat. Prot. Dosim
 , 
1987
, vol. 
21
 
1/3
(pg. 
107
-
110
)
30
Andersson
K. G.
Fogh
C. L.
Byrne
M. A.
Roed
J.
Goddard
A. J. H.
Hotchkiss
S. A. M.
Radiation dose implications of airborne contaminant deposition to humans
Health Phys
 , 
2002
, vol. 
82
 
2
(pg. 
226
-
232
)
31
Hession
H.
Byrne
M. A.
Cleary
S.
Andersson
K. G.
Roed
J.
Measurement of contaminant removal from skin using a portable fluorescence scanning system.
J. Environ. Radioact
 , 
2006
, vol. 
85
 
2–3
(pg. 
196
-
204
)
32
Fogh
C. L.
Byrne
M. A.
Andersson
K. G.
Bell
K. F.
Roed
J.
Goddard
A. J. H.
Vollmair
D. V.
Hotchkiss
S. A. M.
Quantitative measurement of aerosol deposition on skin, hair and clothing for dosimetric assessment.
1999
 
Final report, Risoe-R-1075(EN) ISBN 87-550-2446-7
33
International Commission on Radiological Protection
Age-dependent doses to members of the public from intake of radionuclides: part 5 – compilation of ingestion and inhalation dose coefficients
1995
Pergamon Press, Oxford, New York, Tokyo
 
Publication 72 ISBN 0 08 042737 5.
34
Andersson
K. G.
Tables of Parameter Values to Describe the Deposition of Particles Released from a RDD (‘dirty bomb’) on Outdoor and Indoor Surfaces in an Inhabited Environment and the Relationship Between Outdoor and Indoor Aerosol Concentrations, Confidential Internal Project Report
2008
Roskilde, Denmark
Risoe National Laboratory
35
Hall
D. J.
Spanton
A. M.
Griffiths
I. H.
Hargarve
M.
Walker
S.
The urban dispersion model (UDM): version 2.2. Technical documentation
2002
Porton Down, UK
Defence Science and Technology Laboratory
 
DSTL/TR04774
36
Päsler-Sauer
J.
Description of the atmospheric dispersion model ATSTEP
1997
Karlsruhe, Germany
FZK
 
RODOS (WG2)-TN(97)01, Progress report of RODOS, Draft version, November 1997
37
Mikkelsen
T.
Larsen
S. E.
Thykier-Nielsen
S.
Description of the Risø puff diffusion model
Nucl. Technol
 , 
1984
, vol. 
67
 (pg. 
56
-
65
)
38
McMahon
T. A.
Denison
P. J.
Empirical atmospheric deposition parameters—a survey
Atmos. Environ
 , 
1979
, vol. 
13
 (pg. 
571
-
585
)
39
International Commission on Radiological Protection
The 2007 recommendations of the International Commission on Radiological Protection
 , 
2007
 
Publication 103. (Elsevier, Oxford, New York, Tokyo) ISSN 0146-6453.
40
International Commission on Radiological Protection
The biological basis for dose limitation in the skin
 , 
1991
Pergamon Press
 
Publication 59 ISSN 0146-6453.
41
Astrup
P.
Thykier-Nielsen
S.
Mikkelsen
T.
In-town dispersion calculations with RIMPUFF and UDM
2005
Roskilde, Denmark
Risø
 
Risø-R-1539(EN)
42
Andersson
K. G.
Brown
J.
Mortimer
K.
Jones
J. A.
Charnock
T.
Thykier-Nielsen
S.
Kaiser
J. C.
Proehl
G.
Nielsen
S. P.
New developments to support decision-making in contaminated inhabited areas following incidents involving a release of radioactivity to the environment
J. Environ. Radioact
 , 
2008
, vol. 
99
 (pg. 
439
-
454
)