TL: Analysis of Risk Associate With Nuclear Reactors in Turkey SO: John Taylor, CRES, For Greenpeace Australia (GP) DT: 1998 CRES Working Paper - please report any viewing problems. An Analysis and Visualization of the Risk Associated with the Potential Failure of Nuclear Reactors in Turkey John Taylor, Centre for Resource and Environmental Studies, Australian National University. email: taylorj@cres.anu.edu.au Stuart Ramsden, Australian National University Supercomputer Facility, email: Stuart.Ramsden@anu.edu.au Abstract An analysis of the risk to surrounding countries from a potential release of radioactive gas from a possible reactor failure in Turkey has been performed using the Australian National University Chemical Transport Model (ANU- CTM). Turkey is about to build its first of up to ten nuclear power generators. The modelling results indicate that Turkey and the countries of the middle East would be at greatest risk from a release of radioactivity from the proposed nuclear power reactors located near Akkuyu, Turkey. A computer visualization simulating the dispersion of a radioactive tracer released from near Akkuyu, Turkey was prepared using meteorological data for 1993. * Introduction. * Global tracer transport model. * Meteorological data. * Experimental method. * Visualization. * Model results. * Discussion. * Conclusions. * References. Commissioned by Greenpeace --------------------------------------------------------------------------- Disclaimer: Greenpeace does not necessarily endorse all of the opinions expressed in this report. Copyright - Australian National University, 1998. Introduction Turkey is currently planning to build up to ten nuclear power generators the first of which is to be located near Akkuyu on the southeast Mediterranean coast. This study uses the Australian National University Chemical Transport Model (ANU-CTM) to investigate the risk, that in the event of an accident at a nuclear power plant which resulted in the release of a radioactive gas, it would impact on Turkey and on other countries in the region. The study does not attempt to determine the likelihood of whether such an accident is possible or even probable. The approach taken in this study was to simulate the release of a unit amount of tracer, which in this case would represent a radioactive gas, and examines the proportion remaining after transport downwind from the point of release. Therefore this study does not reflect the result of a particular release scenario - it is more general and hopefully more useful than that - with results simply being scaled to reflect the magnitude of a potential release. The key conclusion of this study is that Turkey and the countries of the middle East would be at greatest risk from a release of radioactivity from the proposed nuclear power reactors located near Akkuyu, Turkey. Countries of the middle east are nearly always at substantial risk whereas countires to the west have much lower risk. During the winter months Westerlies prevail taking air masses from Southern Turkey to Syria, Israel, Lebanon, Iran, Iraq, Saudia Arabia and the gulf states, Uzbekistan and Kazakhstan. During the summer months the NE trade winds dominate with air masses travelling from Southern Turkey towards the middle East and North Africa including countries such as Syria, Israel, Lebanon, Jordan, Eygpt and Libya, Saudi Arabia and the gulf states. Tracer Transport Model An important tool for improving our understanding of the biogeochemical cycles of atmospheric trace gases that cause climate change and for studying the release of radioactive gases at the regional scale is a high resolution three-dimensional global atmospheric circulation model which can be used to investigate the sources, removal processes and atmospheric chemistry of these trace gases. In the past, even low resolution three-dimensional models have placed enormous computational demands on supercomputers. This continuing project has involved the development of a computationally efficient high-resolution simulation model for atmospheric transport and chemistry, the Australian National University Chemical Transport Model (ANU-CTM) (Taylor, 1989, 1991; Taylor et al., 1991) with parameterised interactions between the oceans and the biosphere. Models of the sources and sinks of the trace gases CFC-11, CFC-12 (Taylor, 1991), methyl chloroform (Taylor et al., 1991), methane (Taylor et al, 1991), CO2 (Taylor, 1989, 1993, 1995; Taylor and Lloyd,1992), radon (Taylor, 1991), CO (Erickson and Taylor,1992; Taylor et al 1995), N2O (Taylor, 1992; Bouwman and Taylor, 1996) have been developed and incorporated into the transport model. Key Physical Processes represented in the model include * Lagrangian tracer transport which conserves the mass of tracer by definition thereby avoiding the need for mass fixing * Wind fields that are based on the monthly mean and variances of 12 hourly ECMWF data calculated at the United States National Center for Atmospheric Research by Kevin Trenberth. The method of calculation is as described in Trenberth (1992). The data are available on a T42 grid approximately 2.8 degrees in longitude and latitude. Dr Jay Larson processed the data into a format suitable for application with ANU-CTM and transferred this data to ANU. * Surface topography defined by the surface pressure data obtained from ECMWF. The model vertical coordinate is pressure. * Time varying boundary layer using monthly mean estimates of the boundary layer height derived from the NCAR Community Climate Model Version 3 by Dr David Erickson III of Atmospheric Chemistry Division of NCAR. * Cloud transport of trace gases is based on a modified Tiedke scheme developed at the Max Planck Institute for Meteorology by Dr Martin Heimann and employed in the TM2 tracer transport model The basic approach of the stochastic Lagrangian model is to divide the troposphere into air parcels of equal mass. Trajectories for these air parcels are calculated using observed wind field data obtained from the European Centre for Medium Range Weather Forecasting. While the simulated air parcels are being transported around the globe, they can exchange chemical compounds with the oceans, the biosphere and one another and take up industrial emissions of trace gases. The model runs on the Silicon Graphics Power Challenge and VPP300 supercomputers at the Australian National University Supercomputing Facility. Model runs with up to 1 000 000 air parcels, giving an effective model resolution below 1 x 1 degree, have easily been achieved. Model runs at higher resolution are possible and will be required in the future to answer questions about sources and sinks of greenhouse gases and to investigate the release of radioactive tracers at the regional scale. Meteorological data This study employs meteorological data derived from data compiled by the European Centre for Medium Range Weather Forecasting in Reading, England and is based on observations collected around the world every 6 hours. This wind field data set is considered one of the best available for studying the transport of tracers and atmospheric chemistry and is employed by research groups worldwide for this purpose. Wind fields used in ANU-CTM are based on the monthly means and variances of 6 hourly ECMWF data and were calculated at the United States National Center for Atmospheric Research by Kevin Trenberth. The method of calculation is as described in Trenberth (1992). The data are available on a T42 grid approximately 2.8 degrees in longitude and latitude. Dr Jay Larson processed the data into a format suitable for application with ANU-CTM and transferred this data to ANU. Wind field data sets are available for the period 1980-1995. More recent data sets include data at 15 pressure levels in the atmosphere whereas data from the early 1980's include data at only 7 levels. Accordingly data from a more recent year was selected, in this case 1993. During most months of the year wind speeds at the surface are low, conducive to maintaining high concentrations of radioactive tracer in the region near the potential release point in Turkey. The low wind speeds also complicate the prediction of where the impacts associated with a particular radioactive release event will occur, making it difficult to respond in an emergency. During the winter months (Figure 1) Westerlies prevail taking air masses from Southern Turkey to Syria, Israel, Lebanon, Iran, Iraq, Saudia Arabia and the gulf states, Uzbekistan and Kazakhstan. During the summer months (Figure 2) the NE trade winds dominate taking air masses from Southern Turkey to such countries as Syria, Israel, Lebanon, Jordan, Eygpt, Saudi Arabia and Libya. [IMAGES N/A IN THIS VERSION: Figure 1 - The wind direction and average wind speed at 1000 hPa for January. Figure 2 - The wind direction and average wind speed at 1000 hPa in July.] Experimental Method The atmospheric transport model was configured to facilitate the assessment of the risk from the release of a radioactive gas for each day of the year. The site of the release was located at latitude 37 degrees North and longitude 33 degrees east near Akkuyu, Turkey. The model run included 100 000 air parcels. The advection of each air parcel was computed every six hours. The concentration on each air parcel was calculated every 24 hours. Estimates of the concentration were saved on an Eulerian grid at 2.5 degrees resolution at seven vertical levels. The visualisation was developed using the individual air parcel trajectories and their respective concentrations. The model run generates nearly 610 Mbytes of concentration data per year of simulation. Visualization To better understand the mechanisms being modeled by the ANU-CTM, sophisticated visualization techniques were utilized to produce an animated graphical representation of the data. Air parcels contain varying radioactive tracer concentrations as they disperse from the release point in Turkey and mix with other air parcels. In this visualisation only the 500 most concentrated air parcels are plotted, from the simulation set of 100,000. The air parcels are rendered as spheres with concentration mapped to colour, size and transparency. The daily data is spaced 10 frames apart, with the positions of the air parcels being interpolated through intermediary frames. This accounts for the regular bouncing motion seen in the animation. The height scale is exaggerated to show the topographic and field data more clearly. We used the animation package Houdini produced by SideFX to visualize the simulation results. The data was precalculated and the model was previewed in real- time on a Silicon Graphics Oynx Reality Engine workstation provided the ANU Supercomputer Facility Vizlab. Final images were rendered on the ANUSF SGI Power Challenge. The resulting animation was output to video from Avid Illusion and a smaller, shorter version of the simulation was produced as well with one day represented as a single frame, suitable for viewing on the World Wide Web. Results The results of the model run using ANU-CTM with ECMWF wind field data for 1993 are presented in Figures 3 and 4. Figure 3 presents individual concentration estimates at times throughout the year whereas Figure 4, an animation, contains the model results for each day. Both figures show the concentration of radioactive gas relative to the concentration at the point of release. Air parcel concentrations are considered high when they are close to the original concentration at the point of release near Akkuyu, Turkey and low when they are approaching background levels. Occasionally high concentrations can remain in air parcels transported over great distances. Eventually the release of a trace gas mixes throughout the global atmosphere while slowly being removed by rain, contact with the land surface, plants and the ocean and through chemical transformations in the atmosphere. [N/A IN THIS VERSION OF THE REPORT: Image Series... Figure 3 - Shows the air parcel concentrations and risk (high, medium and low) for every 15 days during the one year model run. Figure 4 - Shows the air parcel concentrations (high, medium and low) as an animation for every day during the one year model run.] Discussion The model results show that a major failure of a nuclear reactor sited in Turkey could be a disaster for Turkey and many countries in the region. The modelling results indicate that Turkey and the countries of the middle East would be at greatest risk from a release of radioactivity from the proposed nuclear power reactors located near Akkuyu, Turkey. Countries of the middle east are nearly always at substantial risk whereas Western Europe always have a much lower risk. During the winter months Westerlies prevail taking air masses from Southern Turkey to Cyprus, Syria, Israel, Lebanon, Iran, Iraq, Saudia Arabia and the gulf states, Uzbekistan and Kazakhstan. During the summer months the NE trade winds dominate with air masses travelling from Southern Turkey towards the middle East and North Africa including countries such as Syria, Israel, Lebanon, Jordan, Eygpt and Libya, Saudi Arabia and the gulf states. It should be noted that at all times during the year a release of radioactive gas is highly likely to impact countries other than Turkey. Turkey is of course at greatest risk throughout the year. This study employed one year of meteorological data derived from ECMWF by Trenberth (1992) for the year 1993. The conclusions are therefore dependent on this data set. It is clear however that the 1993 data set does not represent an unusual wind flow pattern. It is in good agreement with the long term average wind patterns (Trenberth 1992 ). Further investigations with each available year of meteorological data from 1980-1995 would help refine these estimates. Again, without prior knowledge of the nature of an actual release, and the prevailing meteorological conditions at the time, the results presented here can only be considered to represent the most likely conditions and not what may take place during an actual release. To facilitate an early warning of the arrival of radioactive fallout a high resolution regional meteorological forecasting and chemistry model would need to be developed if reliable estimates of the dispersion of a cloud of radioactive gas are to be obtained. The risk analysis demonstrates that the siting of nuclear power reactors is a potential hazard for the entire region, not just Turkey. Countries in the Middle East, the former USSR and Africa, should be consulted with regard to the potential siting and operation of nuclear power reactors in Turkey. Clearly alternative energy sources should be explored. Conclusions An assessment of the risk of fallout associated with the failure of a nuclear power plant located in Turkey was performed using the ANU-CTM transport model. The results of the the assesment were visualized to aid in the interpretation and communication of the model predictions. The study has found that Turkey and countries of the Middle East are at high risk of receiving fallout and that, for each country, this risk varies during the year. It is clear from the model results that regional environmental security considerations need to be taken into account when siting nuclear power reactors in Turkey and the Middle East. To eliminate these risks alternative sources of energy generation should be considered. References Ball, D (1991) Building blocks for regional security: An Australian perspective on confidence and security measures (CSBMs) in the Asia/Pacific region, Strategic and Defence Studies Centre, Canberra Papers on Strategy and Defence No. 83, Australian National University, Canberra. Bouwman, A.F. and Taylor, J.A. (1996) Testing high resolution nitrous oxide emission estimates against observations using an atmospheric transport model, Global Biogeochemical Cycles, (in press). Erickson III, D.J. and Taylor J.A. (1992) 3-D Tropospheric CO modelling: The possible influence of the ocean. Geophysical Research Letters, 19, 1955-1958. Taylor, J.A. (1989). A stochastic Lagrangian atmospheric transport model to determine global CO2 sources and sinks. A preliminary discussion. Tellus, 41B, 272-285. Taylor, J.A. (1991). Tracer transport model for the study of sources and sinks of trace gases important to climatic change. Computer Modelling in the Environmental Sciences, Farmer, D.G. and Rycroft, M.J. (eds.), IMA Conference Series, Oxford University Press, 233-242. Taylor, J.A. (1992) A global three-dimensional Lagrangian tracer transport modelling study of the sources and sinks of nitrous oxide. Mathematics and Computers in Simulation, 33, 597-602. Taylor, J.A. (1993) The mutable carbon sink? Nature, 366, 515-516 Taylor, J.A.(1995) Global Carbon Dioxide Sinks: An investigation of the role of the biosphere in the tropics and middle to high latitudes in the Northern Hemisphere. In Proceedings of the International Congress on Modelling and Simulation, Newcastle, 27-30 November, Vol. II, 39-44. Taylor, J.A., Brasseur, G., Zimmerman, P. and Cicerone, R. (1991) A study of the sources and sinks of methane using a global 3-d Lagrangian tropospheric tracer transport model. Journal of Geophysical Research, 96, 3013-3044. Taylor, J.A. and Lloyd, J.L. (1992) Sources and sinks of atmospheric CO2. Australian Journal of Botany, 40, 407-418. Taylor, J.A., Zimmerman, P.R. and Erickson, D.J. III (1995). A 3-D modelling study of the sources and sinks of atmospheric carbon monoxide, Ecological Modelling, (in press). Trenberth, K.E. (1992) Global Analyses from ECMWF and atlas of 1000 to 10 mb Circulation Statistics. NCAR Technical Note, NCAR/TN-373+STR, NCAR Boulder CO, USA, 191pp.