Engr. Sajjad Ali Khan
Radiation Physics Division
Pinstech, P.O, Nilore, Islamabad

Human intrusion (HI) at a waste disposal site is one of the important scenario having a potential for release of the waste from geosphere to the biosphere. This scenario is comparatively little discussed, analyzed and understood, although it has the consequences comparable to those resulting from ground water transport of radioactive waste. In this study two limiting release scenarios based upon human intrusion (HI) are studied viz. waste release under the intact and degraded institutional control. It is concluded that under normal geological evolution process and intact control, the major contributor to the doses to the future generations are the geological processes specially faulting, ground water attack and erosion. Overall failure probability of a repository is found to be 1.94E-5 in 1.0E5 years under this scenario. However if institutional control degrades (scenario 2), then chances of release of waste are about 5.22E-3 in the same time period ie. about 269 times higher.

Risk assessment of geological waste disposal systems require identification of all normal and disruptive features, events and processes (FEPs) and generation of release scenarios [1]. Human intrusion (HI) is one of the important scenario having the potential for release of the waste to the environment [2-10]. In the past, major attention has been focused on the evaluation of releases of contaminants to ground water (GW) and their subsequent migration to the biosphere. Human intrusion (HI) is comparatively little discussed, analyzed and understood topic. It has lot of uncertainty and impact the overall safety assessment predictions. However recently it is recognized that consequences of human intrusion (HI) at a waste disposal may be equally comparable to those resulting from GW transport. The earliest analyses of the HI were mostly performed for disposal facilities for low level waste (LLW) and specially for salt formation [2,5].

Human intrusion impacts all the elements of performance assessment of waste repositories viz. development of scenarios, modelling, consequences assessment, uncertainties analysis, comparison with regulatory criteria and overall risk assessment. Various approaches have been used for the analysis of HI at waste disposal site [5-7]. For performing the sensitivity study for HI release scenario as presented in this article, fault tree analysis (FTA) approach is adopted using the computer code IRRAS-PC [11] to predict the release probability of waste to the biosphere (top event) in terms of failure of barriers in a waste isolation system (components). The code has already been tested and validated, and results are reported in ref. [1].

Uncertainty and sensitivity analysis of the results to different input assumptions encompass an important aspect for better understanding of HI problem. The sensitivity analysis presented in this article involve the releases under two limiting case scenarios. Under the limiting case scenarios the work presented in this paper aims at the four main objectives viz. 1. evaluation of the impact on the overall failure probability of the facility. 2. detailed evaluation of the impact on various risk importance ratios and 3. identification of the major contributors to the release phenomena under both release scenario.

Human intrusion assessment at waste disposal facilities is a multifaceted problem and have different dimensions in time, society, and environment. Therefore it has a broad scope and involve a number of combination of parameters and as a result other representative release scenario may be developed. Therefore this study is not a complete overview of the HI scenarios, rather it evaluate only the limiting cases.

Human exposure, because of deliberate or unintentional human activities at a waste repository site, is called human intrusion (HI). Various human activities play important role in the long term performance assessment of geological waste repositories. Human intrusion (HI) is frequently considered as the dominant contributor to the top event in a fault tree ie. releases to the biosphere and resulting doses to humans [2-5,10,13].

The HI at a repository site may be inadvertent human actions (eg. various mining operations, large engineering operations, drilling for water or exploration, underground explosions possibly nuclear tests etc.) while intruder is unaware of the presence of waste or of its potential hazard. The HI may also be due to deliberate actions like sabotage, terrorism, war etc. Different scenarios may be imagined for the intrusion at a disposal site [5].

Over very long time periods, human intrusion (HI) at a waste disposal site is a major threat to the repository integrity, specially when institutional control are no longer effective and warning signs are difficult to decipher by the future generations. Some of the situations which may lead to radiation exposures due to human interaction at a repository site may be the following eg. undetected past intrusion as undiscovered boreholes, mine shafts at the site; improper operation of the repository eg. improper waste emplacement; transport agent introduction in the repository eg. irrigation, reservoirs, intentional artificial GW recharge or withdrawal; chemical liquid waste disposal at the site etc [1-8].

Definition of a release scenario is the important first step in repository risk assessment exercise [1]. Various combination of events may be imagined for the release scenarios development. The basis for HI scenario development is normally a combination of historical and present practices coupled with human imagination and computers and their extrapolation over long time periods as presented in Fig.1 [5]. To determine the consequences of the HI, a limited number of representative scenarios can be analyzed based upon direct contact of the intruder with the waste material, contact to a widely dispersed material and normal evolution scenarios of the repository [2,3,5].

Once a repository has been sealed and closed, it will still be protected from human intrusion by legal or institutional restrictions, on the use of repository site. Detailed information of the buried waste, the repository and the site characteristics will be kept available in archieves and computer data bases. Some monitoring of radioactivity in the surrounding environment might also continue for some period. Over time, however as the generations come and go there is possibility that control

would be less stringent or discontinued, and that knowledge and information about the repository will be wholly lost. Therefore strict control needed to prevent any intrusion cannot be guaranteed for more than at most a few hundred years. Some EPA regional offices do not allow credit in PA for active institutional control beyond 100 years after disposal [2,3,5].

Following are some typical situations for radiation exposures due to human intrusion at a repository site [3].

1. Intrusion at the site causing approach to intact-buried waste..eg. due to drilling at a site and the resulting slurry of waste causing exposure to drilling crew and laboratory persons by inhalation and ingestion of dust etc. This is normally referred to as the "drilling" scenario.
2. Exposure to waste dispersed by previous intrusion eg. construction or farming on site containing waste extracted by these activities . This is referred to as the "resident scenario".
3. Human induced intrusion disturbing the normal evolution of a repository eg. modification of ground water (GW) flow by an open borehole.
4. Contact initiated by human action with contaminated GW. eg. Intersection of contaminated GW by a drinking water well. (GW Release Scenario).
5. Contact with materials contaminated by the GW.

For assessment of sensitivity of human intrusion, two limiting case scenarios have been used in this study , to assess the maximum probability of the failure of waste containment barriers. In scenario 1, it is assumed that current institutional control are made more effective with the passage of time (limiting case 1). There is remote chance that site existence is forgotten and no deliberate or inadvertent human intrusion take place. In scenario 2, it is assumed that current institutional control degrades in 1.0E5 years interval and almost no control remains effective (limiting case 2). There are maximum chances that human intrude at the site in the form of mining the site or any other way.

Long term safety assessment of a waste disposal system is normally performed by using the probabilistic safety assessment (PSA) approach [13-20]. Fault trees are developed and usually computer models are used to analyze the parameters for the long term safety analysis. A number of such analyses have been reported in literature for salt lithologies [21-28]. However the sequence analyzed in this paper is based upon the ref. [12,15]. The tree is developed for a bedded salt repository placed at a depth of 300 m and consists of three basic branches, showing the geological containment failure mechanism:

1. Exhumation
2. Flooding by water
3. Meteorite impact.

Details of each of the sections are shown in transfer gates (trees) in Figs. 2, 3, and 4. In these figures, the three types of gates are used, OR gate (with "+" sign), AND gate (with "x" sign) and transfer gate (with triangle sign).

Data issues are a matter of grave concern in every safety assessment study. HI is one of major area of uncertainty in this context. Various approaches have been used for the quantification of HI probabilities eg. event trees, product integral approach [6] and data based upon subjective/expert judgment [2,3,4,6]. Typical HI data for a salt and plastic clay formation is reported in Tables I and II respectively. It is obvious from these tables that loss of memory is almost certain for time period greater than 1.0E4 years.

Besides the HI data, other failure data as reported in Table III, are based upon ref. [15], using the expert opinion, historical records, and literature values. However due to poor quality of past records, there are large gaps in the available data. Therefore a variation of some order of magnitude may not be uncommon for probability values depending upon the source on which it is based.

Results for sensitivity analysis of two limiting HI scenarios are reported in Tables IV to VIII over a very long time periods ie. upto 1.0E5 years. The results include dominant cutsets, important risk ratios, and uncertainty analysis. However results regarding the first two parameters are discussed in detail. The definition and detailed description of these two parameters can be found in references [29,30]:

Detailed results for cut sets analyses, for HI scenario 1 and 2 are presented in Tables IV and V. For scenario 1, under normal operation of repository and intact institutional control, it can be seen from Table IV that the release probability of waste from the disposal system is 1.94E-5 and first five cut sets account for about 93% of the top event probability. The first cut set (E1-E10-E8) has a weight of about 41.3%. This cut set comprise of the faulting phenomena (E1) in the presence of ground water above (E8) or below the salt formation and which after saturation is continuously replaced by fresh water (E10).

The second cutset (E1-E11-E9) comprising of faulting phenomena (E1) coupled with confined ground water from below (E9) and simultaneous replacement of this water (E11) with fresh one, has a weight of about 31%. Both of these cutsets are major contributor to the repository failure and indicate the critical

potential of faulting phenomena coupled with the ground water attack and its subsequent transport, as a waste dispersion medium. The third dominating cut set (E12-E5-E8) contribute about 10.3 % to the repository failure and comprise of large fault (E5) coupled with replacement of saturated salt water in fractured layers (E12) from the fresh ground water above (E8).

The observations from Table IV under scenario 1, regarding the cutsets generation and their weightage etc., indicate that under the intact institutional control over very long time periods (although highly improbable), the normal geological evolution processes dominate the releases to the biosphere. This indicate the consideration of the detailed geological and geohydrological investigation of a candidate site before its final selection. These observations are also in agreement with the common opinion that under intact institutional control, the ground water attack and transport are the critical events in a repository safety.

From the Table IV, it is observed that besides the first three dominating cutsets, the alternate ways of release from the repository are cutsets no.4 (erosion), 5 (a large fault coupled with attack from confined ground water below, and its replacement in the fractured layers with fresh water) and 6 (very large displacement fault). However these cutsets have about 12% contribution in total.

Table V indicate the results of important cut sets under the assumption of total loss of memory of the existence of the repository as a result of failure of institutional control. Here the direct exhumation of the waste (E15) is the dominating cut set having a weight of 95.8%. This indicate that under scenario 2, the probability of release occurrence with in 1.0E5 years after the repository closure is related to the HI and it may be via direct exhumation of waste. It indicate that after having forgotten the existence of the repository, the future generations may start mining the salt and thereby may receive doses via drilling, resident, inhalation and other pathways [3].

Table VI and VII indicate the risk importance measures for various events. Under intact institutional control (scenario 1), first three events in the Table VI have the major risk reduction potential ie. small displacement fault (E1), ground water above the salt formation (E8) and replacement of saturated ground water with fresh water (E10). These events have a risk reduction potential of 3.6, 2.2, 1.8 respectively. This implies that if the repository site is selected in such a way that the chances of developing small faults, movement of ground water above or below salt and exchange with overlying/below fresh water are nil, then risk of waste release to the future generation may be reduced by a factor of 3.6, 2.2 and 1.8 respectively.

In limiting scenario 2, effective institutional controls are assumed to be lost and this value is taken as two order of magnitude greater than those reported in literature

Risk achievement worth ratios from Table VI under intact institutional control (scenario 1), indicate that if sufficient efforts are not made to control the degradation of the repository in a salt formation due to the events like small displacement faults (E1), large fault (E5), exhumation by human actions (E15), erosion (E22), etc. then the risk to the future generation may be increased by a factor of 5.16E4 at maximum. This fact underlines the importance of taking into account these factors in a repository site selection process.

Similarly in Table VII, under the complete loss of institutional control (scenario 2), first three events has the major risk reduction potential viz. small displacement fault (E1), ground water above the salt formation (E8) and replacement of saturated ground water with fresh water (E10). These events have a risk reduction potential of 2.38, 1.04, and 1.04 respectively. This implies that if the repository site is selected such that future improvements on various activities reduce the possibility of developing small faults, ground water above or below salt and exchange with overlying/below fresh water, then risk of waste release to the future generation can be reduced by a factor of 2.38, 1.04 and 1.04 respectively in the time period of 1.0E5 years.

Risk achievement worth ratios from Table VII under degraded institutional control scenario, indicate that the events like exhumation by human action (E15), erosion (E22), very large displacement fault (E6), slumping (E18), etc. may increase the risk to the future generation by a factor of 1.90E2 at maximum.

The important point emerging from Tables VI and VII is that if a poor site is selected, having faults and ground water problems, then further improvement and strengthening the institutional control would not significantly reduce the risk of failure over a time period of 1.0E5 years. Whereas increasing the probability of human intrusion at a repository site (scenario 2) due to degradation of current practices (eg. loss of institutional control, isolated or deserted sites selection, site reservation for public park, records in the archieves, redundancy in keeping records etc.) at a similar site would increase the risk of repository failure by a factor of about 269. This fact underlines the importance of properly maintaining the current control practices if no further improvement is possible at a selected site. It also indicate that improvement of the current practices has less risk reduction potential, rather they have high risk achievement worths.

As a matter of fact data uncertainties (of some order of magnitude) may be expected in some of the events and needed to be resolved to maximum possible extent. However this study has demonstrated uncertainties in HI scenarios and present some useful insights for two limiting cases. Prior to important decision making, based upon the results from such analyses, it is advisable that fault tree model should be adequately described, the event dependencies should be adequately understood and the processes involved in the repository failure should be well defined thereby eliminating the chances of any modelling errors [12].

Uncertainty associated with the minimal cut sets is also evaluated. Results for the same are reported in Table VIII under the complete loss of institutional controls (scenario 2), for maximum, minimum, point estimate, 5th percentile value, median value and mean value. A Monte- carlo procedure is used for the calculation of the probability distribution of the repository failure using the probability of each basic event in the minimal cut sets [11].

On the basis of this analysis, following conclusions are drawn:

1. Under the scenarios of both the intact and the complete loss of institutional controls, the probability of release of waste from the repository is 1.94E-5 and 5.22E-3 respectively in 1.0E5 years.
2. Human intrusion scenario dominate the releases from the geosphere to the biosphere, over long time periods provided institutional controls are lost.
3. If institutional controls are intact, then geological evolution processes specifically faulting, erosion, and attack of ground water, dominate the releases to the biosphere.
4. The quantified values of cut set under the intact institutional controls, indicate that the dominating cut set consisting of faulting with water movement contribute 41.2 % to the repository failure.
5. Under the complete loss of the institutional control, the single event cut set dominating the release, is human exhumation and this event alone can contribute 95.8 % to the repository failure.
6. The maximum risk reduction potential is for small faulting events. Avoiding the areas having the existing faults or any future potential, risk to the future generations may be reduced by a factor of greater than 2.
7. The maximum risk achievement potential is for the events like human intrusion, erosion, and very large faults. These events have potential of 3.8E4.
8. Due to some real limitations on the collection of data relating to the human intrusion, the sensitivity analysis presented here in this paper implies that for some important scenarios like human intrusion or ground water, sensitivity and uncertainty analyses may be the best approach for increasing the public confidence in the risk assessment studies.

1. The possibility of human intrusion over very long time periods need to be given due consideration in the site selection process.
2. The major implication from this study is the need of detailed geological and geohydrological considerations in the site selection process besides giving the priority to the mitigation measures to the human intrusion at a specific site.
3. Priority should be given to those areas in the site selection process which have the minimum, structural disturbances, underground water movement etc.
4. Risk achievement potential and risk reduction potential of various events indicate that future repository siting process should avoid the localities with underground moving water, faults, erosion or have a potential to develop ones in the future.
5. Human intrusion issue is a multifaceted problem. It needs detailed investigations using various computer models. Besides a better communication of the human intrusion risk to the public in the light of the uncertainties in current models, data etc. is needed.
6. Detailed uncertainty and sensitivity analysis is recommended to be performed under a variety of other human intrusion release scenarios for specific sites.
7. Nuclear waste disposal is a global environmental issue. Better exchange of knowledge, experiences and ideas on international level would be more fruitful.

The author would like to express his gratitude to Messr. A.A.Qureshi and I.E.Qureshi for providing facilities during the preparation of this paper. He would also like to thank the Dr. H.A. Khan for entrusting to initiate the waste related R&D activities at PINSTECH. He also wish to thank UK NIREX LTD., and NEA, PARIS, for providing useful literature on the PSA of waste repositories. Thanks are also rendered to reviewers for their valuable comments on this work

1. "Radionuclide Release Scenarios for Geological Repositories" Proc. of the NEA Workshop PARIS, 8-12 September, 1980.

2. NEA, Nuclear Energy Agency, "Risks Associated with Human Intrusion at Radioactive Waste Disposal Sites", Proc. of NEA Workshop, Paris, 1989. Nuclear Energy Agency, Organization of Economic Cooperation for European Countries.

3. WUSCHKE. D.M, "Assessment of the Long-term Risks of Inadvertent Human Intrusion into a Nuclear Fuel Waste Disposal Vault in Deep Igneous Rock". AECL-10279, presented at Third International Conference on Nuclear Fuel Processing and Waste Management....RECOD91(RECOD...91), Proceedings, Vol.1, P: 450-455 SENDAI, JAPAN, April 14-18, 1991.

4. NORDMAN, R. VIENO T., "Consideration of Human actions in the Finnish Performance Assessment of Nuclear Waste Disposal", [Ref.1], PP:94-99, (1989).

5. Thegerstrom. C, " Human Intrusion at Waste Disposal sites"., NEA, Newsletters, Fall, 1989, Vol. 7, No. 2, pp:16-18

6. Helton, J.C. "Drilling Intrusion Probabilities for Use in Performance Assessment for Radioactive Waste Disposal", Journal of Reliab. Engg. & System Safety, 40/3, (1993) ,PP:259-275.

7. Saunders, P.A.H.,"Safety Studies...NIREX Radioactive Waste Disposal..Research and Safety Assessment, United Kingdom Atomic Energy Authority (UKAEA), NSS/G100, September 1988.

8. Persson, L. "The Scientific Base for the Disposal of Spent Nuclear Fuel", Health Physics, Vol.64, No.4, April, 1993.

9. Persson, L., "Ethical Aspects of Nuclear Waste", Journal of Health Physics, Vol.58, No.3, March, 1990.

10. Alessandro,d. M. and Bonne, A., "Fault Tree Analysis for Proababilistic Assessment of Radioactive Waste Segregation: An Application to a Plastic Clay Formation at a Specific Site", Proc. of Papers 26th International Geological Congress in PARIS, July 1980, (1982).

11. K.D. Russel, etal, "Integrated Reliability and Risk Analysis System (IRRAS)-User Guide-Version 1.0 (draft), NUREG/CR-4833, Idaho National Engg. Laboratory USA (June 1987).

12. Sajjad A. Khan, "PC-Based PSA Study for a Geological Waste Repository Placed in a Bedded Salt Formation", Nucl. Safety, Vol.35, No.1, January-June, 1994. PP:128-141.

13 C. D. Marsily, D.F. Merriam, "Predictive Geology with Emphasis on Nuclear Waste Disposal". (1982). Pergamon Press.

14. Can Long-term Safety be Evaluated..An International Collective Opinion, NEA, OECD, PARIS 1991.

15. Bertozzi, G.M. D'Alessandro, F.Girardi and M. Vanossi, "Safety Assessment of Radioactive Waste Disposal into Geological Formations: A Preliminary Application of Fault Tree Analysis to Salt Deposits. Commission of the European Communities Joint Research Centre, Ispra Establishment, Chemistry Division. Ispra Italy (1978) EUR-5901.

16. "Disposal of Radioactive Waste....The International Probabilistic System Assessment Group...Background and Results..Nuclear Energy Agency". Paris, 1991.

17. PSACOIN Level 1a Intercomparison. Published by OECD/ NEA, PARIS, 1990.

18. Risk Analysis in Nuclear Waste Management. Proceedings of an ISPRA Course Held at JRC Ispra, Italy, May-June 1988.

19. Disposal of Radioactive Wastes: Review of Safety Assessment Methods. Published by OECD/NEA, Paris, 1990.

20. PAGIS...Performance Assessment of Geological Isolation Systems for Radioactive Waste....EUR-11775-EN to EUR-11779-EN, 1988.

21. USA technical Details of a Geological Repository in Salt for the Disposal of Radioactive Wastes, INFCE/PC/2/7, IAEA, Vienna (1979) Appendix 2.

22. Report of the Feasibilities of Radioactive Waste Disposal in Salt Formations in the Netherlands, Interministerial Nuclear Energy Commission, Ministry of Economic Affairs, The Hague (April 1979).

23. International Fuel Cycle Evaluation, "Technical Details of a Geologic Repository in Salt for the Disposal of Radioactive Wastes", Waste Management and Disposal (Proc. INFCE working group 7, 1979), IAEA, Vienna (1980), Appendix, 1.

24. L.R. Hill, "Characterization of a Site in Bedded Salt for the Isolation of Radioactive Wastes", IAEA-SM-243/43, IAEA Proc. series, "Underground Disposal of Radioactive Waste, Organized by IAEA, OECD NEA, at Otaniemi, Finland, 2-6 July , 1979. Vol.I.

25. Donald G. Schweitzer and Cesar A. Sastre, " Long-term Isolation of High-level Radioactive Waste in Salt Repositories Containing Brine". Nuclear Technology, Vol.86, PP: 305-312, Sep. 1989.

26. I.J.A.Carr and J.C.Cunnane, "Salt Repository Project Waste Package Program Plan, "Draft U.S. Nuclear Regulatory Commission, Office of Nuclear Waste Isolation , (May 10, 1986).

27. D.R. Diercks, et al, "Interaction of Waste Package Materials with Simulated Salt Repository Environment", ANL-IWPM-4/87, Argonne National Laboratory (1987).

28. Claiborne, H.C. and F. Gera. "Potential Containment Failure Mechanisms and their Consequences at a Radioactive Waste Repository in Bedded Salt in New Mexico." Report ORNL-TM-4639, Oak Ridge National Laboratory, Oak Ridge, Tennessee, (1974).

29. R.R. Fullwood, etal, "Probabilistic Risk Assessment in the Nuclear Power Industry-Fundamentals and Applications". BNL, Newyork, N.Y. USA (1988).

30. W.E. Vesely, etal, "Measures of Risk Importance and Their Applications", NUREG/CR-3385, BMI-2103, (1983).

BACK