Health Implications of Fragments of Irradiated Fuel at the Beach at Sandside Bay

RPD-EA-03-2006

Health implications of fragments of irradiated fuel at the beach at Sandside Bay

Module 6: Overall results

B T Wilkins, J D Harrison, K R Smith, AW Phipps, P Bedwell, G Etherington, M Youngman, T P Fell, M W Charles, P J Darley and A Sh Aydarous

ABSTRACT Small numbers of fuel fragments continue to be located and retrieved from the beach at Sandside Bay. Some originate from the Materials Test Reactor and some from the Dounreay Fast Reactor. SEPA commissioned a major project dealing with topics such as the capability of the monitoring system, the probability of encountering a fuel fragment while on the beach and the possible health effects. The work was carried out as a series of modules. Detailed reports of the results of each module have already been published by SEPA. This final report in the series brings together the overall results.

The HPA contribution to this study was funded by SEPA, the contribution from the University of Birmingham being supported by UKAEA.

This work was undertaken under the Environmental Assessments and Emergency Response Group’s Quality Management System, which has been approved by Lloyd's Register Quality Assurance to the Quality Management Standards ISO 9001:2000 and TickIT Guide Issue 5, certificate number 956546.

© Health Protection Agency Approval: Month Year Centre for Radiation, Chemical and Environmental Hazards Publication: Month Year Radiation Protection Division £00.00 Chilton, Didcot, Oxfordshire OX11 0RQ ISBN 0 85951 xxx x

This report from HPA Radiation Protection Division reflects understanding and evaluation of the current scientific evidence as presented and referenced in this document.

CONTENTS

1 Introduction 1 2 Radiological protection philosophy 2 3 The current system of monitoring and retrieval 3 4 Monitoring data for Sandside Bay 6 5 The effectiveness of the Groundhog systems 7 6 The probability of encountering a fuel fragment 9 7 Potential doses and risks from fuel fragments 13 7.1 Characteristics of fuel fragments 13 7.1.1 Radionuclide composition 13 7.1.2 Fragment solubility and intestinal absorption of radionuclides 15 7.1.3 Self-attenuation 16 7.2 Dose assessment methodology 16 7.3 Doses to the skin and the eye 17 7.4 Doses from inadvertent ingestion 19 7.5 Doses from inhalation 20 7.6 Fragments containing only 60Co 21 7.7 Equivalent and effective doses 21 7.8 External doses from fuel fragments remote from the skin 23 8 Final remarks 23 9 Acknowledgements 24 10 References 24

iii

INTRODUCTION

1 INTRODUCTION

Sandside Bay lies about 3 km to the west of the licensed nuclear site at Dounreay in Caithness. There is a large beach within the bay that is some 3 km in length. The beach is currently privately owned and is open to the public. A fragment of irradiated nuclear fuel was discovered on the beach at Sandside in 1984. Two further fragments were discovered in 1997. Also in 1997, fuel fragments were found on the seabed offshore from Dounreay.

As a result of these discoveries, the then Scottish Office imposed restrictions on the removal of marine foodstuffs under the Food and Environment Protection Act 1985. These restrictions applied to the area within a 2 km radius of the outlet of the liquid effluent pipeline at Dounreay, and are still in force. At that time, the Scottish Environment Protection Agency (SEPA) commissioned the then National Radiological Protection Board (NRPB)* to carry out an assessment of the likelihood of an individual consuming a fuel fragment incorporated into marine foodstuffs, and what the resultant doses might be. This work was based on the limited data available at that time, and necessarily adopted a cautious approach. The results of the study were published by SEPA in 1998 (SEPA, 1998).

SEPA subsequently placed a requirement on the site operator, the UK Atomic Energy Authority (UKAEA), to carry out a more comprehensive monitoring programme on the beach at Sandside Bay, as well as at other public beaches in the area. In each of the years from 1999 to 2002, typically about 5 fuel fragments were retrieved from the beach at Sandside. These fragments were broadly similar in size to sand grains and so would not be visibly distinguishable from general beach material. The majority of the fragments retrieved from the beach at Sandside Bay came from the processing of fuel from the Materials Test Reactor (MTR). This type of fragment contains 137Cs together with 90Sr and its decay product 90Y. The activities of these three radionuclides had been found to be broadly similar, and if such a fuel fragment came into contact with tissue most of the dose came from 90Sr and 90Y (SEPA, 1998; COMARE, 1999).

As the research and monitoring programme as a whole developed, there was an increase in the proportion of the total number of fragments found that originated from the Dounreay Fast Reactor (DFR). Some of these were retrieved from the beach at Sandside. This type of fragment also contained 137Cs, 90Sr and 90Y. However, the limited data available at that time indicated that DFR fragments contained much more 137Cs relative to 90Sr and 90Y. SEPA therefore considered that a further study was warranted, focusing more on the finds at Sandside, considering the health implications in more detail than had been possible in the earlier study and taking account of differences between fragments of MTR and DFR origins. There was also a need to provide an independent evaluation of the monitoring equipment and a perspective on the likelihood of individuals coming into contact with fragments as a result of using the beach.

* NRPB became the Radiation Protection Division of the Health Protection Agency with effect from 1 April 2005.

1 HEALTH IMPLICATIONS OF FRAGMENTS OF IRRADIATED FUEL AT THE BEACH AT SANDSIDE BAY

The technical specification for the work was developed following comment from the Scottish Executive, the Food Standards Agency, the Highland Council and the Highland Health Board. SEPA commissioned the then NRPB to carry out the work. The project considered the capability of the monitoring equipment used by UKAEA, the probability of an individual coming into contact with a fuel fragment on the beach at Sandside and, if they did, what the acute health effects might be. At the same time, UKAEA commissioned the University of Birmingham to carry out a programme that included the measurement of surface dose rates from fuel fragments and associated dosimetric assessments. It was agreed at the outset that these two projects would proceed collaboratively. The work at the University of Birmingham specifically complemented that carried out by NRPB on doses and health effects, and the two pieces of work were reported together (Harrison et al, 2005) (www.sepa.org.uk/radioactivity/dounreay/particles).

The overall work programme was set up as a series of modules, some of which were later divided further because of changes in the monitoring equipment during the course of the project. These modules or sub-modules have formed the subject of individual reports, each of which has been published by SEPA (Etherington and Youngman, 2005; SEPA, 2005a; Walsh et al, 2005; Smith and Bedwell, 2005a; Smith and Bedwell, 2005b; Harrison et al, 2005) (www.sepa.org.uk/radioactivity/dounreay/particles). One further report contains several short accounts of work undertaken subsequently to support particular aspects (Smith et al, 2005). The results from all of the earlier work are drawn together here to provide an overall assessment of the current radiological situation at Sandside Bay.

2 RADIOLOGICAL PROTECTION PHILOSOPHY

The International Commission on Radiological Protection (ICRP) is the primary international body concerned with the formulation of recommendations on radiological protection standards. Its most recent recommendations for an overall system of protection were issued in 1990 as ICRP Publication 60 (ICRP, 1991). The present system distinguishes between two categories of exposure, practices and intervention.

Practices are situations where the exposure of individuals is being increased. In terms of protecting people, emphasis is placed on the control of the source of the exposure. Generally, this can be planned before the practice commences. Examples of practices are the generation of electricity by nuclear power and the production of radioisotopes. ICRP recommends an annual limit on effective dose of 1 mSv for members of the public as an overall result of practices that are subject to control. In the UK, discharges are regulated by the environment agencies and licensed nuclear sites are required to carry out environmental monitoring to demonstrate compliance with the dose limit.

Interventions are situations where the sources, pathways and exposed individuals are already in place when a decision on control is taken. In such situations, control can only be achieved by intervention, ie, by removing or modifying the existing sources or exposure pathways or by reducing the numbers of people exposed. A decision on the

2 THE CURRENT SYSTEM OF MONITORING AND RETRIEVAL

most appropriate form of intervention is a process of optimisation, with the aim of doing more good than harm. For this reason, dose limits do not apply directly in intervention situations.

In the case of the occurrence of fuel fragments in the environment around Dounreay, the fuel fragments have already been discharged from the site and so control at source is no longer possible. Protection of people can be achieved only via an appropriate level of intervention. Consequently, under the ICRP system of protection the dose limits that apply to practices are not applicable.

3 THE CURRENT SYSTEM OF MONITORING AND RETRIEVAL

The report issued by SEPA in 1998 included an assessment of the public health implications of the finds of fuel fragments offshore from Dounreay (SEPA, 1998). This report and its recommendations were considered by the then Scottish Office. The Secretary of State for Scotland subsequently wrote to SEPA asking that:

“SEPA ensure that there is sufficient monitoring in place to ensure that any particles finding their way to the beach at Sandside Bay are promptly detected and removed.”

Within the site authorisation, SEPA has specified the frequency and extent of monitoring and other performance criteria as part of a Technical Implementation Document (TID). The specification has evolved since it was first imposed in 1999, and the requirements set in 2001 and 2003 are summarised in Tables 1 and 2. Emphasis is clearly placed on monitoring at Sandside Bay, because for public beaches this is where nearly all of the fuel fragments found since widespread monitoring began in July 1999 have been located. Monitoring is currently carried out by RWE Nukem under contract to UKAEA. From the radiological protection point of view, the periodic monitoring of beaches around Dounreay and the requirement to remove promptly any active fragments that are detected should be regarded in combination as an intervention strategy.

3 HEALTH IMPLICATIONS OF FRAGMENTS OF IRRADIATED FUEL AT THE BEACH AT SANDSIDE BAY

Table 1. Frequency and extent of beach monitoring for fragments of irradiated fuel - Technical Implementation Document requirements, September 2001 Beach Extent of monitoring Grid references Frequency of Detection criteria (GRs) Monitoring Sandside All of the sandy areas 295700, 966280 & Monthly > 105 Bq of 137Cs Bay that can be accessed by 296690, at 100 mm depth a vehicle from MHWS* to 965780 low water ** between GRs in column 3 Sandside Accessible sandy areas 295700, 966280 & Monthly > 105 Bq of 137Cs Bay which do not permit 296690, at 100 mm depth vehicle access including 965780 north beach, harbour, sandy areas below Fresgoe House, bands of sand north east of the beach below the public lavatories and the sandy areas north of Isauld Burn Sandside Strandline that can be 295700, 966280 & Fortnightly > 105 Bq of 137Cs Bay accessed by vehicle 296690, at 100 mm depth between GR’s in column 965780 3 Thurso Bay All sandy areas that can 311360, 968960 & Three times > 105 Bq of 137Cs be accessed by a vehicle 312070, per year at 100 mm depth from MHWS* to low 968850 water*between GRs in column 3 Scrabster All sandy areas that can 310040, 970180 & Three times > 105 Bq of 137Cs Bay be accessed by a vehicle 310605, per year at 100 mm depth from MHWS* to low 969170 water** between GRs in column 3 Crosskirk All accessible sandy 302860, 969900 & Twice per year > 105 Bq of 137Cs Bay areas from MHWS* to 302970, 970250 at 100 mm depth MLWS*** between GRs in column 3 Brims Ness All accessible sandy 304250, 971270 & Twice per year > 105 Bq of 137Cs areas from MHWS* to 304410, 971030 at 100 mm depth MLWS*** between GRs in column 3 * MHWS Mean High Water Springs. ** Low water means as reasonably practicable to low water springs, but at least to neap low water. *** MLWS Mean Low Water Springs

4 THE CURRENT SYSTEM OF MONITORING AND RETRIEVAL

Table 2. Frequency and extent of beach monitoring for fragments of irradiated fuel - Technical Implementation Document requirements, October 2003 Beach Extent of monitoring Grid references Frequency of (GRs) monitoring Sandside All of the sandy areas that can be 295700, 966280 Monthly Bay accessed by a vehicle from MHWS* & 296690, 965780 to low water** between GRs in column 3 Sandside Accessible sandy areas which do not 295700, 966280 Monthly Bay permit vehicle access including north & 296690, 965780 beach, harbour, sandy areas below Fresgoe House, bands of sand north east of the beach below the public lavatories and the sandy areas north of Isauld Burn. Sandside Strandline that can be accessed by 295700, 966280 Fortnightly Bay vehicle between GR’s in column 3 & 296690, 965780 Thurso All sandy areas that can be 311360, 968960 Three times per year Bay accessed by a vehicle from MHWS* & 312070, to low water** between GRs in 968850 column 3 Scrabster All sandy areas that can be 310040, 970180 Three times per year Bay accessed by a vehicle from MHWS* & 310605, to low water** between GRs in 969170 column 3 Crosskirk All accessible sandy areas from 302860, 969900 Six times per year Bay MHWS* to low water** between GRs & 302970, 970250 in column 3 Brims All accessible sandy areas from 304250, 971270 Six times per year Ness MHWS* to low water** between GRs & 304410, 971030 in column 3 Dounreay All accessible sandy areas from 298190, 967029 Fortnightly except East Foreshore MHWS* to low water** between GRs & 298340, 967095 during the period 1 in column 3 May to 31 August. Dounreay All accessible sandy areas from 298190, 967029 Fortnightly except West MHWS* to low water** between GRs & 298340, 967095 during the period 1 Foreshore in column 3 May to 31 August. Melvich All accessible sandy areas from 288246, 965662 Once during 2004 Beach MHWS* to low water** between GRs & 289109, 965028 in column 3 Dunnet All accessible sandy areas from 320336, 968460 Once during 2004 Bay Beach MHWS* to low water** between GRs & 321440, 970870 in column 3 * MHWS Mean High Water Springs. ** Low water means as reasonably practicable to low water springs, but at least to neap low water.

Monitoring of the strandline on beaches had been carried out using handheld equipment for many years. The widespread monitoring that began in 1999 was carried out using the Groundhog Mark 1 system, which was vehicle-mounted. In May 2002 this was replaced by new equipment referred to as Groundhog Evolution. At that time detection limits were still specified within the TID (Table 1). It should be noted that the information in this Table has been reproduced in the manner issued by SEPA. The term "> 105 Bq 137Cs at 100 mm depth" implies that the equipment should be capable of detecting at least this activity at this depth in the beach sediment.

5 HEALTH IMPLICATIONS OF FRAGMENTS OF IRRADIATED FUEL AT THE BEACH AT SANDSIDE BAY

SEPA subsequently requested a preliminary theoretical evaluation of the capability of the Evolution system, and as a result the TID was revised as shown in Table 2. The main difference between the requirements set in September 2001 and those in October 2003 was that detection limits were no longer specified. Instead, the vehicle was required to maintain a mean operating velocity of 1 ms-1 with no account being taken of any measurement made at velocities of greater than 1.2 ms-1. This was because a limitation on the operating velocity should imply conformance with the earlier criterion on detection limits. In addition, any surveying at velocities in excess of 1.2 ms-1 would no longer be included in the assessment of the area covered.

For the beach at Sandside Bay, the total area bounded by the grid references in Tables1 and 2 is about 230,000 m2, although parts of the beach are covered by rocks and are not routinely monitored. In addition, there will clearly be times during each tidal cycle when some parts will not be accessible. The TID specifies that the total area to be monitored each month should be 250,000 m2. Consequently, parts of the beach must be monitored more than once during each month.

4 MONITORING DATA FOR SANDSIDE BAY

UKAEA maintains a record of fuel fragments found at the various monitoring locations on its website (www.ukaea.org.uk/dounreay/particles). In the period to March 2005, a total of 55 fuel fragments had been retrieved from the beach at Sandside Bay. Monitoring has not been continuous since vehicle-based surveys began in 1999, a consequence of access to monitoring vehicles being periodically withdrawn by the owners of the beach (DPAG, 2003). As a result, it is not possible to use the numbers of finds to make any predictions about rates of arrival of fuel fragments from the sub-tidal zone. Typically, around 5 fuel fragments have been retrieved in each of the years since vehicle-based monitoring began. However, in 2003 a total of 24 was recovered, mostly in the period between 26 February and 12 May. This coincided with the introduction of the Groundhog Evolution system, and it is tempting to suggest that the increased number of finds was due to an improved monitoring system. However, it is important to remember that beaches are dynamic systems, and that large amounts of sediment can be transported within and between the intertidal and subtidal zones over time. The relatively large number of finds in the first half of 2003 followed a period of stormy weather when a great deal of beach sediment had been eroded. Anecdotal evidence indicated that at this time at least part of the beach was at a very low altitude, that is, a great deal of material had been eroded and transported seawards, although no confirmatory measurements were made. Nevertheless, it is possible that the removal of sediment meant that fuel fragments that had previously been deeply buried might then be near enough to the newly-revealed beach surface to be detected by the Groundhog Evolution system. Subsequently, the rate at which fragments were found decreased and it would be imprudent to use the data for 2003 to infer any estimates of rates of arrival or the likely numbers of fragments in the beach at a given time.

The majority of fuel fragments can be characterised by their 137Cs content, because this can be readily determined in a non-destructive manner from the gamma-ray emissions

6 THE EFFECTIVENESS OF THE GROUNDHOG SYSTEMS

of the decay product 137mBa. Fragments that are of DFR origin may be characterised by the presence of 94Nb, although the absence of this radionuclide does not necessarily mean that the fragment came from the MTR. Scanning electron microscopy is needed to confirm the origin of fuel fragments, but this is a more labour intensive process and so is not applied to all finds. At least 5 of the fragments found to March 2005 at Sandside Bay were of DFR origin. Eight fragments have so far been retrieved from the seabed that contain 60Co and no other gamma-ray emitting radionuclides. Radiochemical analysis of one such fragment indicated that, apart from a very small amount of plutonium, no other radionuclides were present (Harrison et al, 2005). A further four such fragments have been retrieved from the Dounreay foreshore, but none has been found on the beach at Sandside Bay.

The most active fragment retrieved so far from the beach at Sandside Bay contained about 3 105 Bq 137Cs, but most were in the range 104-105 Bq. For comparison, the 137Cs activity in some of the fuel fragments retrieved from the Dounreay foreshore was over two orders of magnitude greater. This difference has been ascribed to the greater mobility of the less active and physically smaller fragments (DPAG, 2003). It should be emphasised however that the 137Cs activity can only be used as a broad surrogate for mass, one of the properties that affect particle mobility in the marine environment. Until recently, the limited published evidence indicated that there was some variability in the activity : mass quotient by factors of around 5 for MTR fragments (SEPA, 1998)*. More data are now available (SEPA, 2005b). These have yet to be analysed in full, but much of the data are consistent with a proportional relationship between activity and mass. For a given mass however the 137Cs activities can in a few cases still range over about an order of magnitude. The implications of these results for the beach at Sandside Bay are discussed in Section 5.

5 THE EFFECTIVENESS OF THE GROUNDHOG SYSTEMS

Theoretical evaluations of the effectiveness of the Groundhog Mark 1 and Groundhog Evolution systems have been carried out as part of this project (Youngman and Etherington, 2003; Etherington and Youngman, 2005; SEPA, 2005a). The earlier study evaluated the detection limits that could be achieved under different circumstances and compared these with the criterion specified at that time in the TID (105 Bq 137Cs at 0.1 m depth). The findings were that in many cases the criterion could be met in terms of activity to a depth of 0.05 m, and in some cases to 0.1 m. In worst case conditions, the detection limit that could be achieved was about 3 times greater than that specified in the TID. However, requirements for the detection capability were not closely defined and for practical purposes the equipment met the criteria. Recommendations for improvements included a need to assess more rigorously the effects of changes in natural background, a requirement to monitor the speed of the vehicle more closely and a need to operate at lower speeds.

* Note that there is an error in the relevant data shown in this publication, and the true range of values was 6.6 108 - 3.9 109 Bq g-1.

7 HEALTH IMPLICATIONS OF FRAGMENTS OF IRRADIATED FUEL AT THE BEACH AT SANDSIDE BAY

The development of Groundhog Evolution had taken many of these recommendations into account. The evaluation carried out within this study made use of calibration data and measured backgrounds supplied by RWE Nukem, the operators of the system. The main findings for the beach at Sandside Bay were as follows.

• The system could reliably detect fragments containing 105 Bq 137Cs to a depth of 0.1 m at vehicle speeds of up to 1.6 m s-1 in most circumstances, and in all circumstances at a speed of 1.2 m s-1.

• The system could not reliably detect fragments of this activity at a depth of 0.2 m.

• For a fuel fragment of 105 Bq 137Cs at a depth of 0.1 m, there is only a small probability (less than 1%) that it would be undetected on a first scan and then detected on a second scan.

• There is a small decrease in the probability of detecting fuel fragments of this activity if the passes made across the beach do not overlap.

In addition, the evaluation indicated that a reduction in the vehicle speed from 1.2 m s-1 to 0.5 m s-1 would improve the detection limits by less than a factor of 2.

The observed variations in activity : mass quotients, noted in Section 4, means that there is a possibility that fragments more active than those found currently could be deposited on the beach at Sandside Bay. Some perspective can however be obtained from an extrapolation of the data used in the evaluation of Groundhog Evolution (Smith et al, 2005). This indicated that fragments containing 106 Bq of 137Cs should be detected reliably at depths of up to 0.3 m. The corresponding value for fragments containing 108 Bq was about 0.6 m. Had they been present near the surface of the beach, fragments containing these levels of activity should have been detected very easily with either of the Groundhog systems. The possibility of fragments containing 106 Bq of 137Cs or greater being present at Sandside must therefore be very unlikely because none has been detected since widespread monitoring began in 1999.

The beach at Brims Ness was also considered within this study because fuel fragments have been located offshore. In this case, the results suggested that the multi-detector vehicle-borne system could detect fragments of 105 Bq 137Cs at a depth of 0.1 m at scan speeds of up to 1.2 m s-1 but not at 1.6 m s-1. However, for the single detector system that is handled manually, fragments of this activity at this depth would not be reliably detected at any speed in the range 0.5 - 1.6 m s-1.

The independent evaluation was confined to the detection capability of Groundhog Evolution. However, the system also includes a positioning capability, so that the areas covered by the surveyors can be identified accurately (http://www.rwenukem.co.uk/products_and_services/products/groundhog/).

As noted in Section 3, so far only a few fragments containing 60Co have been retrieved from the seabed and the Dounreay foreshore and none has been found on any other beaches. The small numbers found on the seabed imply that the likelihood of such fragments being deposited in the future in intertidal areas away from the Dounreay site should be small. No data were available on which to assess the capability of Groundhog Evolution to detect such fragments in the beach at Sandside. Although in

8 THE PROBABILITY OF ENCOUNTERING A FUEL FRAGMENT

principle the capabilities should be similar to those for 137Cs, the system currently in use is set up specifically to detect fragments containing 137Cs (Etherington and Youngman, 2005); it would not detect fragments containing only 60Co in real time (Smith et al, 2005). Data could be analysed retrospectively. However, because of the continual movement of sediment within the beach, this would need to be done very shortly after the original survey to provide the best chance of retrieving any fragment that was located.

It is worth noting that evaluations of the performance of the Groundhog systems have so far had a significant theoretical element. UKAEA has recently commissioned a large scale laboratory trial. In addition, COMARE has requested that field trials are conducted on the beach at Sandside Bay, and is taking a leading role in designing and carrying out the work (COMARE, 2005). These field trials are expected to take place in the spring of 2006. It is recommended that, once the results of the practical work are available, the implications for the present study are distilled. The uncertainties are expected to be greatest for the low activity fuel fragments present in the beach, ie, around 104 Bq 137Cs.

6 THE PROBABILITY OF ENCOUNTERING A FUEL FRAGMENT

The evaluations of the two Groundhog systems have been used to estimate the numbers of fuel fragments that might be present on the beach at Sandside Bay (Walsh et al, 2005; Smith and Bedwell, 2005a). These estimates encompassed beach material to a depth of 0.2 m. On the basis of the evaluation of the Groundhog systems, it was assumed that fuel fragments at the lower end of the range of activities would not be detectable at depths of more than 0.1 m. In such cases, it was simply assumed that the distribution of any fuel fragments in the 0.1-0.2 m layer would be the same as that in the 0-0.1 m layer. This seemed reasonable given that the intertidal areas of beaches are frequently well-mixed to at least this depth by wave action (Green and Wilkins, 2005), and that the fuel fragments are expected to behave in a similar manner to the sand particles within the beach. A set of representative 137Cs activities was chosen and the probability of detection was estimated for each of them, based on the actual fragments located by the Groundhog Mark 1 and Groundhog Evolution systems. Fragments containing less than 10 kBq of 137Cs were not considered because so far only one fragment containing activity below this value has been detected.

The representative values chosen differed between the two studies, but in both cases the total number of fragments estimated to be present was dominated by those at the lower end of the range of activities. This was to be expected because fragments containing 1 104-5 104 Bq of 137Cs had only a low probability of being detected by the monitoring system. Consequently, the actual number of finds in this activity range represented only a small fraction of the total expected to be there. It should be emphasised however that the uncertainties in the estimated total were considerable. For fragments containing < 2 104 Bq of 137Cs, the range in the estimated values was about a factor of 20, the corresponding value for fragments in the activity range (2 – 5) 104 Bq being a factor of about 6 (Smith and Bedwell, 2005a). In contrast, the number of actual finds containing 105 Bq 137Cs or greater essentially represented all of the fuel fragments of this level of activity in the beach to a depth of 0.1 m.

9 HEALTH IMPLICATIONS OF FRAGMENTS OF IRRADIATED FUEL AT THE BEACH AT SANDSIDE BAY

A cautious estimate of the number of fuel fragments in the beach to a depth of 0.2 m at any one time gave a value of less than 15, with a best estimate of less than 10 (Smith and Bedwell, 2005a). In both cases, it was estimated that about 70% of the fuel fragments present would have 137Cs activities of less than 2 104 Bq. In terms of the number of fuel fragments per unit area of beach, the best estimate was about 3.2 10-5 m-2, again with the major contributors being fragments at the lower end of the range of activities.

Estimates of the likelihood that people will come into contact with a fuel fragment require information on the usage and occupancy of the beach, ie, what individuals do while on the beach and how long they spend doing it. This information is commonly referred to as habit data. In support of its general radiological assessment capability, SEPA commissions habits surveys on a regular basis around each nuclear licensed site in Scotland. The most recent survey around Dounreay took place in 2003 (Tipple et al, 2004), and the results from this and previous surveys were used in the present study.

Estimates of the likelihood of coming into contact with a fuel fragment have been made for a range of habits, including time following leisure pursuits such as walking and beachcombing and time spent digging for bait (Smith and Bedwell, 2005b). It should be emphasised that these probabilities refer to direct contact with a fragment, rather than being within its general proximity. This is because any potential hazard to health requires an individual to come into contact with a particle. As noted in Section 7.8, the external dose from a fuel fragment to a person standing a short distance away is extremely small (Smith et al, 2005).

The probability of coming into contact with a fuel fragment was estimated as a function of the 137Cs activity, the parameter by which the fragments are usually described (Section 7). This approach was adopted so that these probabilities could subsequently be combined with the results of the study on health effects, in which all of the radionuclides of interest were taken into account (Harrison et al, 2005). Where appropriate, the age of the individuals (adult, child or infant) was also considered. The overall results for the population groups of interest are summarised in Table 3. The results are presented in terms of an annual probability, and for this reason values are given to one decimal place. Probabilities are however more easily placed in context when expressed in terms of chance (that is, for example, 1 in one million). In this report, results in this form are given as rounded values, which is a reflection of the uncertainties associated with them.

10 THE PROBABILITY OF ENCOUNTERING A FUEL FRAGMENT

Table 3. Annual probabilities of coming into contact with a fuel fragment on Sandside Beach for each potentially exposed group for different 137Cs activity rangesa Exposed Group Annual probability of encountering a fuel fragment on Sandside Beach (y-1) for various fuel fragment activity ranges (137Cs activity) < 20 kBq a 20 kBq-50 kBq 50 kBq-100 kBq > 100 kBqb Total Adult Bait Digger 2.6 10-7 6.7 10-8 5.2 10-8 8.9 10-9 3.9 10-7 Adult Leisure 3.6 10-7 9.3 10-8 7.2 10-8 1.2 10-8 5.4 10-7 Adult Winkle 5.5 10-9 1.4 10-9 1.1 10-9 1.9 10-10 8.2 10-9 Consumer Child Leisure 1.2 10-7 3.2 10-8 2.4 10-8 4.2 10-9 1.9 10-7 Child Winkle 3.9 10-9 9.9 10-10 7.6 10-10 1.3 10-10 5.8 10-9 Consumer Infant Leisure 1.5 10-8 3.8 10-9 2.9 10-9 5.0 10-10 2.2 10-8 Notes: a. Lowest activity fuel fragment 8.4 kBq b. Highest activity fuel fragment 280 kBq

The values given in Table 3 relate to higher than average times spent on the beach in all cases. Data specific to Sandside Bay were used whenever possible. However, making use of habit data from all of the local beaches would have made little difference to the results. The assumptions made about the time people spend on the beach are summarised in Table 4.

Table 4. Assumptions made about annual occupancies at Sandside beach Age Group Occupancy rates (h y-1) High rate Distributiona Adult Bait Diggers 330b 470-39b Adult Leisure 300 410-24 Child Leisure 85 125-2 Infant Leisure 13.5 30-2 Notes a The distributions are lognormal in all cases. The upper and lower bounds given represent the 97.5th and 2.5th percentiles respectively b These values apply jointly to angling and bait digging. To apply to bait digging only, a scaling factor of 0.13 was applied.

The times spent on the beach are relatively short, even for members of the adult leisure group that includes habitual dog-walkers. However, for comparison the rounded high rate value derived here for adult leisure, 300 h y-1, is the same as that derived previously for a range of beaches in West Cumbria (Wilkins et al, 1994).

11 HEALTH IMPLICATIONS OF FRAGMENTS OF IRRADIATED FUEL AT THE BEACH AT SANDSIDE BAY

With the exception of consumers of molluscs, for all of the cases considered the exposure pathway of greatest importance was contact with the skin. In terms of contact with the skin alone, bait diggers were the people most likely to come into contact with a fuel fragment of any activity on an annual basis, the estimated probability being 3.7 10-7 y-1 (about 1 in 3 million per year). This population group also had the highest probability of coming into contact with a fuel fragment during a single visit to the beach (1 in about 60 million). The probability of a bait digger coming into direct contact with a fragment containing greater than 105 Bq of 137Cs was 8.9 10-9 y-1 (about 1 in 100 million per year).

For people who used the beach for leisure purposes, some other potential exposure pathways contributed significantly to the overall likelihood of coming into contact with a fuel fragment. Notable examples were the lodging of a fuel fragment under a fingernail or entrapment in shoes or clothing, although entrapment might not result in direct contact with the body. The estimated likelihood of ingesting a fuel fragment inadvertently was around 3-4 orders of magnitude less than the corresponding values for direct contact on the skin. Taking all potential exposure pathways into account, the group most likely to come into contact with fuel fragments was adults who used the beach for leisure purposes at higher than average rates. For this population group, the estimated probability of encountering any fuel fragment was 5.4 10-7 y-1 (1 in around 2 million per year). However, the probability of the same group encountering fragments containing more than 105 Bq of 137Cs activity was 1.2 10-8 y-1 (about 1 in 80 million per year). An individual in this group was estimated to be 30 times more likely to encounter a particle containing less than 2 104 Bq 137Cs than one with an activity of greater than 105 Bq 137Cs.

The possibility of sand being taken from the beach for use in a child's sandpit was also considered. For the amount of sand that might be required, the probability of a fuel fragment of any activity being present was estimated to be about 1 in 70,000. In turn, the probability of an infant coming into contact with the fuel fragment while using the sandpit was estimated to be about 1 in 1.5 million per year. The corresponding value for fragments containing > 100 kBq of 137Cs activity was about 1 in 70 million. Direct contact with the skin was the dominant contributor to these probabilities. However, this assessment was based on cautious assumptions. Moreover, the evidence that sand has been removed from the beach for this purpose is only anecdotal, and it is understood that the owner of the beach has stipulated that sand should not be removed for such uses.

The annual probability of a fuel fragment becoming trapped in the eye is very low, less than about 1 in one million million, ie, 1 in 1000,000,000,000. The probability of a fuel fragment becoming trapped in the ear is expected to be lower still (Smith et al, 2005). In view of the very low probabilities of coming into contact with fuel fragments, the probability of exposure by multiple pathways, for example via both skin contact and inadvertent ingestion, has not been considered.

12 POTENTIAL DOSES AND RISKS FROM FUEL FRAGMENTS

7 POTENTIAL DOSES AND RISKS FROM FUEL FRAGMENTS

7.1 Characteristics of fuel fragments

7.1.1 Radionuclide composition Generally, fuel fragments can be conveniently characterised by their 137Cs content via the photon emissions from its short-lived decay product 137mBa. The exception is those deriving from stainless steel, which essentially contain only 60Co. This radionuclide also emits gamma-rays and so can be identified in a non-destructive manner. However, 137Cs-containing fuel fragments originating from both MTR and DFR also contain 90Sr and its decay product 90Y. These radionuclides emit beta particles, and for MTR fuel fragments are the most important contributors to contact doses (SEPA, 1998; COMARE, 1999). However, they can only be determined by destructive radiochemical techniques. The 90Sr and 90Y contents of a range of fuel fragments have been determined by various laboratories. The results are discussed in detail in Harrison et al (2005) and are summarised in terms of radionuclide activity ratios in Table 5. Data for 238Pu and 238,239Pu are also shown.

Table 5. Radionuclide activity ratios Particle 90Sr : 137Cs 238Pu : 137Cs 239,240Pu : 137Cs MTR 1.0 0.01 0.004 MTR 0.9 0.003 0.008 MTR 0.8 0.0009 0.0005 MTR* 2.0 0.003 0.001 MTR 101 0.9 0.007 0.0004 MTR 109 0.9 0.001 0.0002 MTR 154 0.9 0.003 0.0004 MTR 157 0.9 0.006 0.0006 MTR 002 1.4 0.003 0.0006 MTR 113 0.5 Not calc’d Not calc’d MTR 132 0.04 0.00001 0.00005 MTR 138 0.05 0.000003 0.00004 DFR 055 0.2 0.000002 0.00001 DFR 082 0.2 0.00009 0.0005 DFR 098 0.2 0.00002 0.0001 DFR 106 0.1 0.00003 0.0001 DFR 111 1.3 0.00002 0.00007 DFR 107 2.9 0.00004 0.0002 DFR 125 0.3 0.000004 0.00002 DFR 128 0.2 0.00006 0.0002 DFR 134 0.04 0.0000004 0.000001 DFR 135 0.3 0.000002 0.000007 DFR 136 0.5 0.0005 0.005 * Mean of data for three fuel fragments

13 HEALTH IMPLICATIONS OF FRAGMENTS OF IRRADIATED FUEL AT THE BEACH AT SANDSIDE BAY

The 90Sr:137Cs ratios for MTR fragments were generally reasonably consistent. The two exceptions, MTR 132 and MTR 138, had Pu:137Cs ratios that were more consistent with those observed for DFR fragments, and may well have been wrongly attributed. As part of the present study, the University of Birmingham carried out measurements of skin dose rates for a separate set of MTR fragments. The results are summarised in Figure 1 (reproduced from Harrison et al, 2005). Most of the measured values were reasonably consistent with the calculated values of dose rate based on a 90Sr:137Cs ratio of about 0.9 (0.89 in Figure 1). Taking the results in Table 5 and Figure 1 together, a ratio of 0.9 was used in the assessment of doses from MTR fragments (Harrison et al, 2005).

From Table 5, the 90Sr:137Cs ratio for DFR fragments was generally lower than 0.9. The data for DFR fragments in Figure 1 are consistent with this view, since the measured values were lower than those calculated for MTR fragments that assumed a 90Sr:137Cs ratio of 0.9. Consequently, dose assessments based on the 137Cs content and the default radionuclide ratios for MTR fragments are likely to be cautious if applied to DFR fragments.

The calculations of doses from MTR fragments summarised in Sections 7.3-7.6 were based on the assumptions that the fragments were spherical with a homogenous elemental composition of uranium and aluminium (15% U) and a uniform specific activity for all fragments of 2 109 Bq 137Cs g-1 (Darley et al, 2003). The amount of data on mass and activity has increased since the work was carried out (SEPA, 2005b), and the value assumed here is consistent with the more recent results. The assumed activity ratios were 0.9 for 90Sr:137Cs, 0.003 for 238Pu:137Cs, and 0.001 for both 239Pu:137Cs and 241Am:137Cs.

14 POTENTIAL DOSES AND RISKS FROM FUEL FRAGMENTS

1000

Linear extrapolation 100 from low activities ) -2 m c g m

, 7 Wilkins et al 1998 MCNP MTR 2

m 10 S r/Y-90:Cs-137 =0.89 c (1

-1 h y G /

e t

a MCNP MTR r

e Cs-137 only s 1 do n Ski

0.1

0.01 1E+04 1E+05 1E+06 1E+07 1E+08 137 Cs Activity /Bq DFR Sandside M TR Sandside DFR Foreshore MTR Foreshore DFR Seabed MTR Seabed MCNP MTR sphere (Sr-Y/Cs=0.89) Cs137 Only Wilkins et al 1998 Linear (Extrapolation)

Figure 1. Skin dose rates for MTR and DFR fuel fragments (reproduced from Harrison et al, 2005).

7.1.2 Fragment solubility and intestinal absorption of radionuclides SEPA commissioned the National Nuclear Corporation to carry out an in vitro study to evaluate the potential solubility of fragment associated radionuclides in the gut. The experimental work was carried out by the then Scottish Universities Research and Reactor Centre. The results have been summarised as part of the present study, and a detailed account of in vivo experiments carried out by NRPB has also been given (Harrison et al, 2005). In most cases, for all of the radionuclides studied the percentage taken into solution was very small. The data from both sets of experiments were used to derive default values for absorption to blood for use in the dose calculations (Table 6).

15 HEALTH IMPLICATIONS OF FRAGMENTS OF IRRADIATED FUEL AT THE BEACH AT SANDSIDE BAY

Table 6. Values derived from in vitro and in vivo data for radionuclide absorption to blood following ingestion of a fuel fragment Fragment type % of ingested activity absorbed to blood 137Cs 90Sr 239Pu MTR 1 (0.02 - 2) 0.01 (0.001 - 0.03) 0.001 (5x10-5 – 5x10-3) DFR 0.03 (0.001 - 0.05) 0.003 (0.001 - 0.02) 0.001 (5x10-5 – 5x10-3)

Although the majority of the fuel fragments exhibited only low solubility, there was one notable exception. Fragment MTR 113 dissolved readily under the conditions of the in vitro extraction, and for all of the radionuclides around 50% of the activity went into solution in simulated gut fluids. This fragment has been regarded as unusual, because with this exception all of the fragments shown in Table 5 required rigorous conditions before they dissolved. The values in Table 6 have been used to estimate doses to body tissues from radionuclides absorbed to blood following the ingestion of a fuel fragment (Section 7.6). Fragments with the characteristics of MTR 113 have been considered separately.

7.1.3 Self-attenuation The earlier study published by SEPA in 1998 cautiously took no account of attenuation of beta particle energy within the matrix of the fragments themselves - the process of self-attenuation. Figure 1 shows measurements of dose rates to skin from selected fuel MTR and DFR fuel fragments. These were made as part of the present study by the University of Birmingham and are represented by points on the graph. The figure also shows values of dose rate as a function of 137Cs activity, calculated using the Monte Carlo N-particle (MCNP) radiation transport computer code. These are represented by lines on the graph. In each case, the dose rate is estimated to 1 cm2 at a depth of 70 µm (see Section 7.2). For higher activity MTR fragments, measured values fell between those calculated taking account of self-attenuation of beta particle energy within spherical particles (central solid line) and those calculated taking no account of self- attenuation (dotted line). This illustrates the importance of taking account of self- attenuation for the physically larger and more active fuel fragments. Measured dose rates for DFR fragments were lower than those for MTR fragments having the same 137Cs activity content and were closer to calculated values assuming no 90Sr/90Y content (dashed line). For comparison, the Figure also shows the more conservative dose rates estimated in the previous study (Wilkins et al, 1998; SEPA, 1998). These were based on limited early data that gave a 90Sr:137Cs activity ratio of 2 and cautiously took no account of self-attenuation.

7.2 Dose assessment methodology

The dose assessment methodology has been described in detail elsewhere (Harrison et al, 2005) and will not be repeated here. Briefly however local doses to skin were estimated for an area of 1 cm2 and a depth of 70 µm, in accordance with the recommendations of ICRP. Doses to the ear and to the cornea of the eye were

16 POTENTIAL DOSES AND RISKS FROM FUEL FRAGMENTS

assessed in the same way. Estimates of doses following ingestion made use of a new ICRP Human Alimentary Tract Model (HATM), which is not yet published but has been approved by the ICRP Main Commission. An important development in this model is the explicit calculation of doses to a target layer of tissue in the wall of the various regions of the alimentary tract. Doses were calculated for the rectosigmoid region of the large intestine, which receives higher doses than other regions because of its longer transit times. The possibility of inhalation was also considered, making use of the ICRP human respiratory tract model (HRTM) (ICRP, 1994). In each case, the main emphasis was the possibility of the occurrence of deterministic effects after exposure to fragments of different activities; that is, the possibility of acute tissue damage. Self-attenuation within fragments was taken into account in the estimation of potential doses to the skin, to the gut following ingestion, and to the lungs following inhalation. In addition, equivalent and effective doses were also estimated to assess risks of stochastic effects; that is, cancer and hereditary effects.

7.3 Doses to the skin and the eye

When considering hot particle irradiation of the skin, the effect of importance is acute ulceration. Most of the available information on the effects of hot particle irradiation of skin comes from studies using pigs; limited human data are also available. Together, * 2 these data allow the estimation of an ED50 value (measured for 1 cm of skin, at a depth of 70 µm) for acute ulceration of about 10 Gy and a threshold of about 2 Gy. It is clear from these data, together with data for larger area skin exposures, that toleration of radiation will be increased when a particle moves during skin contact, by even a few mm, and when dose rates are low. Taking no account of this amelioration of their possible effect, Table 7 provides a summary of the time taken for fragments of MTR origin to deliver doses of 2 Gy and 10 Gy. The ranges in time for the larger particles are based on differences in fragment shape, with shorter times for non-spherical particles because of reduced self-attenuation (see Section 7.1.3 and Figure 1).

Table 7. Estimates of time taken for MTR particles to deliver doses corresponding to the threshold and ED50 for acute ulceration Activity Dose rate 137 -1 Threshold: 2 Gy ED : 10 Gy Bq Cs Gy h 50 104 0.03 3 days 2 weeks 105 0.3 7 hours 33 hours 106 2-4 0.5-1 hour 2-5 hours 107 15-30 4-8 minutes 20-40minutes 108 70-140 1-1.5 minutes <10 minutes

* The ED50 is the dose that would be expected to produce an observable effect in 50% of cases. The likelihood of producing an observable effect depends on dose rate as well as on the cumulative dose.

17 HEALTH IMPLICATIONS OF FRAGMENTS OF IRRADIATED FUEL AT THE BEACH AT SANDSIDE BAY

For MTR fragments containing 105 Bq 137Cs, ie, typical of the most active particles found at Sandside Bay, no ulceration would be expected to occur for stationary contact

periods of less than 7 h. An exposure corresponding to the ED50 value would imply stationary contact with fragments of this level of activity for 33 h. Such an exposure might produce a small lesion that would become visible within 2-3 weeks, but this would be expected to heal within a further 2-4 weeks with normal personal skin care.

For MTR fragments containing 106 Bq 137Cs, the period of stationary contact needed for

the ED50 value to be approached is a few hours. In the unlikely event that an individual did come into contact with such a fragment on the beach at Sandside Bay, exposure periods of a few hours would be credible. Consequently, for the purposes of this assessment MTR fragments with a 137Cs content of 106 Bq have been considered to be broadly the lower level at which deterministic effects from contact with the skin might be expected.

The most active MTR fragments found in the environment around Dounreay contain about 108 Bq 137Cs. The short exposure period needed for these fragments to cause ulceration illustrates the high probability of such damage in the unlikely event of contact. The dose rates produced by these fragments are such that exposures of a few hours duration would give rise to serious ulceration. Lesions would occur that would be visible within 1-2 weeks. These would extend over areas of up to 1 cm2 and take several weeks to heal, probably with some scar formation. In such cases infection would be a possibility and medical treatment might be needed.

The above estimates of the time taken to cause skin damage and considerations of the possible severity of ulceration apply to all sites including the ear, for which entry of a fragment is considered to have an extremely low probability (Section 6) but for which residence times could be long.

A fragment entering the eye requires separate consideration because it is necessary to consider the possibility of induction of cataracts in the lens as well as ulceration of the exterior corneal surface. For a fragment on the corneal surface, the average dose to the equatorial region of the lens will be at least two orders of magnitude less than the dose rates to the skin given in Table 7. Damage to the cornea is therefore considered to be of primary concern.

The most active fuel fragment found at Sandside Bay (3 x 105 Bq 137Cs) would deliver a dose rate to skin or to the cornea of ~1 Gy (1 cm2, 70 µm) per hour. Such a fragment would have a diameter of about 0.4 mm, similar to that of a medium size grain of sand. It would seem reasonable to expect that toleration of particles of this physical size would not usually extend for more than a few hours. Consequently, early biological effects from fuel fragments of the activity found so far at Sandside Bay would be unlikely. In a thorough review of hot particle effects, The United States National Committee on Radiation Protection, NCRP, concluded that protection of the cornea should be considered on the same dose criteria as protection of skin (NCRP, 1999). Dose rates to the cornea are likely to be reduced due to particle movement around the eye, and movement of eye lids and eye ball: this movement will increase the threshold for observable effects, in the manner described earlier for damage to the skin. Extended corneal exposure to higher activity particles could produce corneal ulceration, which

18 POTENTIAL DOSES AND RISKS FROM FUEL FRAGMENTS

may require medical intervention and treatment. As discussed in Section 6, the probability of a fuel fragment entering the eye is very low.

7.4 Doses from inadvertent ingestion

Estimated doses to the rectosigmoid region of the large intestine are summarised in Table 8. Again, values have been estimated as a function of the 137Cs activity and based on the properties given in Section 7.1, assuming no loss of activity from fragments due to dissolution in gut fluids (see Section 7.1.2). Two sets of values are given in each case. The first is an expectation value corresponding to random movement of the fragment through the lumen of the rectosigmoid; and the second is a maximum value, based on the cautious assumption that the fragment remained in contact with the wall of the rectosigmoid throughout transit.

The effect of using the new ICRP HATM model and taking account of self-attenuation of beta particle energy within fragments can be seen from a comparison of the results in Table 8 with those from the earlier study (SEPA, 1998). For fragments containing 108 Bq of 137Cs, the estimated dose for an adult male based on random transit was about 0.3 Gy, with a maximum value of 1.2 Gy for movement in contact with the intestinal wall. The corresponding value in the earlier study was about 7 Gy. At the lower end of the activity range, the differences between the maximum values in Table 8 and those in the earlier study are less because self - attenuation in smaller particles is less important. However, for a fragment containing 105 Bq of 137Cs, the difference between the value for random transit in Table 8 and that in the earlier study is a factor of about 7.

Table 8. Estimated doses to the rectosigmoid, mGy Fragment Fragment Adult male Adult female One year old child activity, diameter, Random Maximum Random Maximum Random Maximum Bq 137Cs µm transit transit transit 103 67 0.01 0.08 0.01 0.1 0.04 0.2 104 150 0.1 0.7 0.1 1 0.4 2 105 310 0.9 6 1 8 3 16 106 680 7 40 10 60 27 110 107 1300 46 230 67 340 185 640 108 3100 290 1200 420 1800 1200 3500

Consideration of the possible effect on doses of changes in the assumptions made in the calculations (Harrison et al, 2005), including fragment specific activity, depth of target cells in the intestinal wall and colon dimensions, showed that variations in transit times are of greatest importance. Transit times for the colon might typically increase by factors of 2-3 in individuals suffering from constipation, but increases of a factor of 10 might occur in extreme cases. Doses to the rectosigmoid would be directly proportional to the transit time. In these extreme cases, for a fragment containing 108 Bq of 137Cs, it

19 HEALTH IMPLICATIONS OF FRAGMENTS OF IRRADIATED FUEL AT THE BEACH AT SANDSIDE BAY

is possible that doses for adults could be in the region 10-20 Gy, with a corresponding value of around 35 Gy for one-year old infants. For a fragment containing 105 Bq of 137Cs, the corresponding doses would be less than 100 mGy for adults and about 200 mGy for one-year old children.

On the basis of available animal data, the threshold for acute damage to the colon resulting in death, following protracted irradiation from ingested radionuclides passing * through the gut, has been estimated to be about 20 Gy, with an LD50 of 35 Gy . Thus it appears unlikely that ingestion of even a fragment containing 108 Bq of 137Cs by an adult would result in death, although in extreme cases the possibility, however small, cannot be ruled out for a one-year old child. It should be emphasised however that so far only one fragment of this activity has been retrieved from the Dounreay foreshore and a further two have been found on the seabed. Doses from the most active fragments found so far on the beach at Sandside Bay would be at least 100 times less than the threshold for lethality.

Since the movement of material through the rectosigmoid region in particular occurs via periodic muscular contractions, it was also important to consider localised doses (Harrison et al, 2005). The approach adopted was similar to that used for skin. The results indicated that, in the unlikely event that a fragment containing 108 Bq of 137Cs remained stationary at the surface of the lumen for 6 h, the resultant dose would be likely to cause ulceration that might not be easy to repair. Under the same conditions, a fragment containing 105 Bq of 137Cs would be likely to cause localised sterilisation of the crypts, but these should be replaceable by natural regeneration.

7.5 Doses from inhalation

Only particles having aerodynamic diameters of less than about 20 µm or less have some probability of penetrating sufficiently into the deep lung, referred to in the ICRP human respiratory tract model (HRTM) as the alveolar-interstitial (A-I) region (ICRP, 1994). The fragments found around Dounreay are very much larger than this and have a negligible probability of reaching the A-I region. Therefore, based on observed activity : mass quotients, those fragments that would be small enough to reach the A-I region would not be sufficiently radioactive to produce the high doses associated with acute effects in the lung (Harrison et al, 2005).

Larger fragments could however deposit in the extrathoracic (ET) airways, and so doses to the anterior nasal passages have been estimated for fragments of various activities. The possibility of localised doses was considered important, in a manner analogous to that adopted for the skin. The results indicated that a fragment of MTR origin containing 105 Bq of 137Cs would, if held against the same point on the ET epithelial lining for 12 h, deliver a dose to 1 cm2 of tissue of about 1 Gy. The corresponding value for a fragment containing 108 Bq would be about 500 Gy, although for practical purposes it is unlikely that a fragment of this physical size could be inhaled and remain stationary for this length of time. Again, the doses do not vary in a linear manner because of the

* LD50 - the dose that would be expected to result in death in 50% of cases.

20 POTENTIAL DOSES AND RISKS FROM FUEL FRAGMENTS

increasing importance of self-attenuation in the physically larger and more active fragments. Local doses of around 500 Gy would cause acute local ulceration, while doses of around 1 Gy may cause only imperceptible damage.

7.6 Fragments containing only 60Co

As noted in Section 4, only a small number of fragments containing only 60Co have been located so far, and so on this basis the likelihood of such fragments being deposited on the beach at Sandside Bay is expected to be low. For this reason, and because the emissions from 60Co are dominated by energetic photons, a less rigorous dose assessment was undertaken in which the self-attenuation of beta particles within such fragments was not taken into account. Typically, the estimated doses to the rectosigmoid for a given activity of 60Co were about 70% of the corresponding value for an MTR fragment having the same 137Cs content. MTR fragments are characterised by their 137Cs content, although 90Sr and 90Y are present in similar amounts (Section 7.1.1) and are important contributors to the overall doses. For the purposes of the present study, in the unlikely event that fragments containing only 60Co were present on the beach and were ingested, then the resultant doses could conservatively be taken to be around the same value as those from MTR fragments characterised by a similar 137Cs activity.

7.7 Equivalent and effective doses

The main health effects of concern in considering exposure to Dounreay fuel fragments are those discussed above, involving acute tissue damage. However, it is also important to assess doses relevant to the risks of cancer and hereditary effects (termed stochastic effects). This assessment focused on the inadvertent ingestion pathway and distinguished between doses from fragments having the solubility of MTR 113 and those from other fuel fragments. Potential exposures due to inhalation and skin contact were also considered (Harrison et al, 2005).

Radionuclides differ in their modes of decay, their radioactive half-life and their biokinetic behaviour, ie, their distribution and retention in body organs and tissues. In addition, individual organs and tissues differ in their sensitivity to radiation. To provide a method for the interpretation of absorbed dose in different organs in terms of the total risk of cancer and hereditary effects, ICRP uses the concepts of equivalent dose and effective dose (ICRP, 1991). These have units of sieverts (Sv) to distinguish them from absorbed dose in Gy. The first stage is to calculate absorbed doses to all of the important tissues, using biokinetic and dosimetric models to take account of the distribution and retention of radionuclides and their radioactive emissions. These are then converted to equivalent doses, taking account of differences in the effectiveness of different radiation types in causing cancer. For example, alpha particles are taken to be 20 times more effective in causing stochastic effects per unit of absorbed dose than beta particles and photons. Doses to different tissues and organs are then summed, taking account of their different radiosensitivities, to give a single value of effective dose.

21 HEALTH IMPLICATIONS OF FRAGMENTS OF IRRADIATED FUEL AT THE BEACH AT SANDSIDE BAY

The use of effective dose allows the summation of doses from external radiation and from radionuclides having different distributions and emissions.

Effective doses were estimated for the ingestion of MTR fragments containing 105 and 108 Bq of 137Cs, based on the intestinal absorption factors for typical fragments given in Table 6. For an adult male, the resultant doses were 0.1 and 80 mSv respectively, the corresponding values for a one year-old being 0.5 and 300 mSv. In each case, the equivalent doses to the alimentary tract contributed around 70% of the effective dose. As in the case of acute effects, these values would be sensitive to the transit time through the alimentary tract. For a fragment containing 105 Bq of 137Cs, an increase in the transit time by a factor of 2-3 would increase the committed effective dose for a one- year old child to about 1 mSv. An increase in transit time by a factor of 10 would give a committed effective dose of about 3 mSv.

In the case of ingestion of a fragment of 105 Bq of 137Cs with the solubility exhibited by MTR 113, the committed effective dose to an adult male was estimated to be about 2 mSv. In this case, the high solubility of the fragment and intestinal absorption of radionuclides meant that the contribution to the overall committed effective dose from the alimentary tract was only about 15%, while skeletal tissues contributed around 60%. Doses to a one-year old child would be about 2-4 times greater than that for an adult male.

Total detriment and the risks of fatal cancer can be broadly estimated by combining the committed effective doses with the risk factors published by ICRP (ICRP, 1991). Total detriment is a measure of harm adopted by ICRP that takes account of both fatal and non-fatal cancers. The risk factors relate to averaged values for the whole population, ie, both genders and all ages, and are 0.05 Sv-1 for fatal cancer and 0.07 Sv-1 for total detriment. Thus an effective dose of 1 mSv corresponds to an overall fatal cancer risk of 5 10-5. The estimated dose of 0.5 mSv to a child that ingested an MTR fragment of typical solubility containing 105 Bq of 137Cs would therefore correspond to a risk of around 2-3 x 10-5. The corresponding value for a fragment with the solubility characteristics of MTR 113 would approach 5 x 10-4. These estimates of risk should be considered together with the estimated annual probabilities of individuals ingesting a fragment containing this amount of activity, which were of the order of 1 in 1 million million (Smith and Bedwell, 2005a). In addition, fragments having the characteristics of MTR 113 must be regarded as atypical, because of the 25 fragments so far taken into solution for analysis, only one has been readily soluble (Section 7.1 and Table 5).

Local doses to the skin can be converted to an equivalent dose and thence to an overall risk of cancer using published information (Harrison et al, 2005). Equivalent and effective doses from a fuel fragment on skin are very small. As an example, a local dose of 2 Gy over an area of 1 cm2, which is considered to be the threshold for observable effects, corresponds to an equivalent dose of 0.1 mSv. A risk estimate of 1.6 x 10-4 Sv-1 has been derived for low dose rate exposure of the skin. This estimate applies to the general population and takes account of the differing sensitivities of UV- exposed and UV-shielded areas of the skin (Muirhead et al, 1993). This risk estimate can then be combined with an estimate of equivalent dose to give a risk of fatal cancer. Taking the above example, an equivalent dose of 0.1 mSv to the skin of an adult would on average imply a risk of fatal cancer of 2 x 10-8. If the area of skin irradiated was also

22 FINAL REMARKS

exposed to natural ultra violet (UV) radiation, eg the hands, face or arms, then the risk would be around 10-7. Skin cancer risk associated with possible exposure to Dounreay particles can therefore be regarded as of low importance compared with considerations of local dose and the possibility of skin ulceration. Again, this risk should be considered together with the low probability of coming into contact with a fuel fragment and thereby of incurring the skin dose.

7.8 External doses from fuel fragments remote from the skin

People making use of Sandside beach may spend some time in the proximity of a fuel fragment without actually being in contact with it. The potential importance of this exposure pathway has been evaluated using a predictive modelling approach (Smith et al, 2005). The evaluation considered an adult standing close to one of the most active fuel fragments found so far on Sandside beach, ie containing 105 Bq 137Cs. The estimated dose rates were less than the expected value due to natural background radiation. Fragments containing 60Co have not been found on the beach at Sandside Bay and so were not evaluated in this part of the study.

This pathway has not been considered further in this report.

8 FINAL REMARKS

Small numbers of fuel fragments continue to be found on the beach at Sandside Bay, with a total of 55 having been retrieved from the beach to March 2005. The situation should be regarded as an intervention because the sources, exposure pathways and potentially exposed individuals are already in place. The potential exposures are assumed to be the result of releases in earlier years; the fuel fragments now reside in the natural environment and cannot be controlled at source. The application of dose limits is not therefore appropriate. When taken in combination, the current approach of monitoring and retrieval can be considered as an intervention strategy

MTR fragments containing 106 Bq of 137Cs could give rise to short-term observable effects with contact periods that would be credible for people spending time on beaches. For the present, the possibility of such fragments being deposited on the beach at Sandside Bay cannot be ruled out. However, such a situation can be considered to be very unlikely because fragments of this activity would have been detected very easily in surface sand using either of the Groundhog systems, and none has been found since widespread monitoring began in 1999. In addition, most of the fragments containing this level of activity that have been retrieved from the seabed offshore have been within 1 km to the north east of the diffuser outlet at Dounreay. In the westerly direction towards Sandside Bay, the more active particles have generally been found within a few hundred metres of the diffuser outlet (Clayton, 2005)

23 HEALTH IMPLICATIONS OF FRAGMENTS OF IRRADIATED FUEL AT THE BEACH AT SANDSIDE BAY

In terms of the 137Cs activity, the most active fragment found so far on the beach at Sandside Bay contained about 3 105 Bq. This would be unlikely to give rise to observable effects if in contact with the skin, because to do so it would need to remain completely stationary for many hours.

On the basis of the existing monitoring data and current information about the occupancy and usage of the beach, the estimated probability of an individual coming into contact with such a fragment is extremely small. The population group most likely to come into contact with a fuel fragment would be an adult using the beach frequently for leisure purposes. The probability of such a person coming into contact with a fragment containing more than 105 Bq 137Cs was estimated to be about 1 in 80 million.

9 ACKNOWLEDGEMENTS

We are grateful to colleagues in our own and other laboratories who provided data, comment and advice. We thank SEPA, especially the project officer Dr Paul Dale, and the members of the SEPA Dounreay Particles Advisory Group (DPAG) and the COMARE Dounreay Sub-Group for permission to use material in advance of publication.

10 REFERENCES

Clayton K (2005). Personal communication. COMARE (1999). Committee on Medical Aspects of Radiation in the Environment. Sixth Report: a reconsideration of the possible health implications of the radioactive particles found in the general environment around the Dounreay Nuclear establishment in the light of work undertaken since 1995 to locate their source. National Radiological Protection Board, Chilton. COMARE (2005). Personal communication. DPAG (2003). Dounreay Particles Advisory Group, 2nd Interim Report. Etherington G and Youngman MJ (2005). An evaluation of the sensitivity of the Gro8undhog EvolutionTM beach monitoring system. Chilton, RPD-DA-01-2005. Green N and Wilkins BT (2005). Transport of sediment in a Cumbrian beach – implications for radiological assessments. Chilton, RPD-EA-8-2005. Harrison JD, Fell TP, Phipps AW, Smith T, Ellender M, Ham GJ, Hodgson A and Wilkins BT, Charles MW, Darley PJ and Aydarous A (2005). Health implications of Dounreay fuel fragments: estimates of doses and risks. ICRP (1991). 1990 recommendations of the ICRP. ICRP Publication 60. ICRP (1994). Human Respiratory Tract Model for Radiological Protection. Ann ICRP, 24 (1-3). Elsevier Science Ltd, Oxford. NCRP (1999). Biological effects and exposure limits for “hot particles”. NCRP Report No 130. National Council on Radiation Protection and Measurements< Bethesda, Washington DC. Muirhead CR, Cox R, Stather JW, MacGibbon BH, Edwards AA and Haylock RG (1993). Estimates of late radiation risks to the UK population. Doc NRPB, 4(4). Scottish Environment Protection Agency (1998). Fragments of irradiated nuclear fuel in the Dounreay local environment.

24 REFERENCES

Scottish Environment Protection Agency (2005a). An evaluation of the sensitivity of the Groundhog EvolutionTM beach monitoring system. Scottish Environment Protection Agency (2005b). Private communication. Smith KR, Bedwell P, Etherington G and Youngman MJ (2005). Public Health Implications of Fragments of Irradiated Fuel. Module 4: External dose rates on Sandside beach and other miscellaneous information. Smith KR and Bedwell P (2005a). Public health implications of fragments of irradiated fuel. Module 2b. Estimated numbers of fuel fragments per unit area of Sandside beach based on information on fuel fragments detected using Groundhog Evolution. Smith KR and Bedwell P (2005b). Public health implications of fragments of irradiated fuel. Module 3. The likelihood of encountering a fuel fragment on Sandside beach. Tipple JR, McTaggart KA and Clyne FJ (2004). Radiological Habits Survey, Dounreay, 2003. Environment Report RL 05/04, Lowestoft: CEFAS. Walsh C, Jones KA and Haywood SM (2005). Public health implications of fragments of irradiated fuel. Module 2b. Estimated numbers of fuel fragments per unit area of Sandside beach based on information on fuel fragments detected using Groundhog Mark 1. Wilkins BT, Simmonds JR and Cooper JR (Eds) (1994). Assessment of the present and future implications of radioactive contamination of the Irish Sea coastal region of Cumbria. Chilton, NRPB-R267. Youngman MJ and Etherington G (2003). Review of the procedures currently used for the monitoring of Sandside Bay.