September 1999
1. Introduction
In December 1996, Turkeyís national state-owned electrical utility, the Turkish Electricity Generation and Transmission Company (TEAS), released bid specifications for construction of a nuclear power plant at Akkuyu Bay. The site is on the southern Mediterranean coast of Turkey, 45 km south-west of the city of Silifke, directly inshore from Cyprus. Three international consortia are bidding to build the Akkuyu nuclear plant: (1) Atomic Energy of Canada Limited (AECL) -- a Canadian government corporation with partners including Kvaerner-John Brown (UK), Korea Electric Power Corporation and Hanjung (South Korea), and Hitachi (Japan); (2) Nuclear Power International (NPI) -- a partnership of Siemens (Germany) and Framatome (France); and (3) a partnership of Westinghouse (USA) and Mitsubishi (Japan).
After several delays, the bids were opened on October 15, 1997. A decision was to have been made on the winning vendor by June 1998, but that decision was repeatedly delayed. Political instability in Turkey and the controversial nature of the Akkuyu decision have made it difficult for succeeding governments to select a nuclear vendor. Prior to the national election held on April 18, 1999, it was announced that a decision on the Akkuyu vendor would be made after the election. In mid-July it was announced that a decision would be made no later than October 15, 1999 -- a date clearly tied to the opening of the bids two years previously. However, a major earthquake struck Turkey on August 17, 1999. The governmentís failure to enforce building standards and respond effectively to the Izmit earthquake may make it politically difficult to proceed in the near future with a decision on the Akkuyu nuclear plant. One of the most controversial issues about the nuclear plant is whether the proposed site will be acceptably safe from earthquake damage. Recent research has indicated that an active fault line, known as the Ecemis Fault, runs close to the plant. Ignoring this strong evidence, the government of Turkey and Atomic Energy of Canada Limited (AECL) have maintained that there is no active faulting in the vicinity of the plant, and that there is negligible danger of a nuclear accident being caused by an earthquake.
On July 14, 1999, Turkish Energy Minister and Deputy Prime Minister Cumhur Ersumer announced that the government coalition would select a vendor by October 15th for the Akkuyu nuclear plant. [1] There are two reasons why the Akkuyu proposal and the Izmit earthquake should cause concern. First, the Turkish government and foreign reactors vendors are engaged in a mutual conspiracy of silence to conceal the real risk of earthquake damage to a nuclear plant at Akkuyu. Earthquakes can simultaneously destroy multiple operating and safety systems in nuclear reactors [2], leading to a catastrophic accident that could cause an unimaginable disaster in the eastern Mediterranean region. One independent team of nuclear experts has already determined that for the Darlington nuclear generating station in Ontario, Canada (which has much less earthquake risk than the Akkuyu area), the most likely cause of a catastrophic accident is an earthquake.[3] The same finding would also likely hold true for a CANDU nuclear station built in an area of higher earthquake risk such as Akkuyu in Turkey.
The second reason for concern is that government corruption and the failure to enforce building standards may be an indicator of what can be expected in nuclear safety regulation if Akkuyu is built. Incompetence by the Turkish Atomic Energy Commission (TAEK) has already been condemned in a January 1999 incident which saw a gamma radiation source sold illegally to scrap dealers in the Ikitelli district of Istanbul, causing the hospitalization of fourteen people, including five children.[4] Another disaster -- even worse than the Izmit earthquake -- could result.
2. The Izmit Earthquake
On August 17, at 3:01 a.m. local time, Turkey experienced another in a long line of devastating earthquakes. Known as the ÎKocaelií or ÎIzmití earthquake, it took place in north-western Turkey about 11 km south-east of the city of Izmit near the Sea of Marmara. The earthquake registered a magnitude of 7.4 on the Richter scale, and took place at the western end of the Northern Anatolian Fault. This fault has seen a series of seven massive earthquakes (over 7 magnitude) moving progressively westward from 1939 to 1999, which led to predictions of another serious earthquake in the area.
Despite its predictability, the Izmit earthquake has resulted in over 15,000 deaths[5] and over 42,000 injuries[6]. Over 600,000 people have been left homeless[7] and property damage is estimated at up to $30 billion.[8]
There has been widespread outrage in Turkey that deaths have been needlessly caused by failure of local and national governments to enforce existing building standards in one of the worldís most earthquake-prone countries. The causes are a combination of corruption and incompetence. Despite repeated warnings of a large earthquake in precisely the area in which it occurred, the government also failed to mobilize a prompt and effective emergency response effort.
Instead of admitting its mistakes and trying to rectify them, the government has lashed out at its critics. On August 25, 1999, Prime Minister Bulent Ecevit attacked the media for criticizing the government. On August 24, state regulators had shut down the private TV station Channel 6 for a week for criticism of the government and corrupt contractors who had built sub-standard buildings that collapsed in the earthquake.[9] Private Turkish aid groups sponsored by Islamic organizations have also been harassed and shut down by Ecevitís secular government. These actions are not surprising in a country where democracy is weak and the military retains ultimate power. Civil and human rights have been routinely violated in Turkey and corruption and incompetence in state activities are not uncommon.[11] There have been four military coups from 1960 to 1997. In 1997, the democratically elected government of Necmettin Erbakan was forced out of power.
3. Seismic Background on Akkuyu
The basic geological research used to estimate the earthquake risk at the Akkuyu site is in a six volume study by TEK (the former Turkish Electrical Authority) and its consultants, dated 1983.[12] Turkish authorities and Atomic Energy of Canada Limited (AECL) have refused to release these studies, so a thorough evaluation is impossible. AECL commissioned a geological study on the Akkuyu site, dated December 11, 1985, by the Canadian consulting firm Terratech/Canatom.[13] The report noted the poor quality and deficiencies in the research conducted by TEK. It is not clear if TEAS and/or the Turkish government have acted to remedy the deficiencies noted in the Terratech/Canatom report. An update to the 1983 study was apparently conducted in 1990, but is thought to be only a probabilistic study looking at past seismic events.
AECLís Terratech/Canatom report pre-dated independent research that took place in the area in the late 1980s. In 1988 and 1989, two cruises were undertaken by the research vessel Piri Reis, of the Institute of Marine Science and Technology (IMST) of Dokuz Eylul University, in Izmir, Turkey. The research team consisted of Turkish geophysicists Sungu Gokcen, Attila Ulug, Nuran Gokcen and Erdeniz Ozel of the IMST and British geophysicist Gilbert Kelling of the University of Keele. A paper published by this team in 1991 identified three structural sub-basins in the area, separated by the Ecemis Fault Complex in the east and the Anamur Komakiti high in the west. The researchers demonstrated that the Ecemis Complex is an active fault.[14] The fault runs from the north-east to the south-west into the Mediterranean Sea, about 20 to 25 km southeast of Akkuyu Bay.
In June 1998, a workshop was conducted on the Ecemis Fault at Nigde University in Turkey. The workshop concluded that the northern parts of the fault are active, classifying the region as seriously prone to earthquakes, and recommending that building codes be changed to prevent future disasters.[15] In contrast with this recent evidence, AECL still maintains categorically that ìThere is no evidence of active geological faulting within the vicinity of the Akkuyu site.î[16]
Dr. Attila Ulug, one of the authors of the 1991 study, and head of the Geophysics Department at the Institute of Marine Science and Technology at Dokuz Eylul University, has stated, ìAt the very least, the Turkish government must conduct further investigations of the Akkuyu Bay area to determine the threat posed by the Ecemis Fault. To go ahead and build a reactor at Akkuyu Bay without further study would be a totally irresponsible, if not criminal, decision.î[17]
AECL has accepted the seismic design parameters in the 1983 TEK studies. Based on those studies, it is assumed that a magnitude 8 earthquake could only occur at the boundary of the African and Eurasian plates at a distance of 60 km from the Akkuyu site. Prof. Arsalan Mohajer has noted that depending on the angle of dip of the African plate as it passes beneath the Eurasian plate, a magnitude 8 or higher earthquake could actually occur much closer to Akkuyu Bay.[18]
Another model used by AECL assumes that, at worst, an earthquake of magnitude 6.5 could occur at a depth of 30 km directly below the Akkuyu site. Based on California data from the early 1980s, TEAS and AECL assumed that such an earthquake would produce horizontal ground shaking of no more than 0.25g (horizontal acceleration of 25% of the force of gravity). It is likely, however, that by using California data, TEAS assumes much greater attenuation (weakening) of the earthquakeís force than is actually the case in Turkey. Moreover, an earthquake of similar magnitude occurring at a depth shallower than 30 km could actually produce a greater level of shaking.[19]
The United States Nuclear Regulatory Commission (USNRC) has set a standard Design Basis Earthquake (DBE) of 0.3g for all commercial nuclear plants in the central and eastern United States, with a few exceptions.[20] According to the USNRC, the DBE (sometimes known as a Review Level Earthquake or RLE) is a level of shaking set sufficiently higher than a Safe Shutdown Earthquake (SSE) to ìdemonstrate sufficient margin over SSE to ensure plant safety and to find any Îweak linksí thatmight limit the plant shutdown capability to safely withstand a seismic event bigger than SSE.î[21] For several American nuclear plants on the south shore of Lake Ontario (Fitzpatrick, Ginna, and Nine Mile Point), the estimated frequency of exceedance for a 0.3g DBE is estimated at 2 to 3 X 10E-5 (i.e. an earthquake with horizontal ground motion exceeding 30% of the force of gravity is likely to occur 2 to 3 times in 100,000 years).[22]
In the Canadian province of Ontario, there are 20 CANDU reactors owned and operated by Ontario Power Generation (formerly Ontario Hydro). A recent exercise to upgrade the seismic qualification of the oldest nuclear station, Pickering A, has set the Design Basis Earthquake at 0.235g. The estimated frequency of exceedance for a 0.3g seismic event at the station is 5 X 10E- 5 (i.e. an earthquake with horizontal ground motion exceeding 24% of the force of gravity is likely to occur 5 times in 100,000 years).[23] As a tectonic plate boundary area, the Akkuyu Bay region has a much higher earthquake risk than central and eastern United States, or southern Ontario. Yet TEAS and the nuclear vendors have agreed to a DBE of 0.25g -- roughly equal to that being applied at the Pickering nuclear generating station, and an exceedance frequency roughly an order of magnitude lower (i.e. one in ten thousand years, as opposed to 5 in 100,000 years).
On June 27, 1998, an earthquake of magnitude 6.6 occurred with its epicentre about 150 km east of the Akkuyu site.[24] The nearest cities affected were Adana and Ceyhan, which suffered 144 deaths and over 1,000 injuries.[25] Another nine earthquakes followed in the same area on the same day, with magnitudes ranging from 3.5 to 4.4.[26] A number of earthquakes have occurred closer to the Akkuyu site. The earthquake data base of the United States Geological Survey notes that just since 1973, 31 earthquakes have taken place within 100 km of the Akkuyu site (Appendix A). The distance of these earthquakes from the Akkuyu site has ranged from 13 to 99 km. The magnitude of the earthquakes ranged from 2.9 to 4.7.
It should also be noted that AECL does not have a good reputation for veracity on this issue. In Canada, AECL President and CEO, Allen Kilpatrick has claimed publicly that CANDU reactors have withstood earthquakes. In letters to the Canadian newspapers[27] Kilpatrick stated that the Gentilly-2 reactor in the Province of Quebec survived a serious earthquake in November 1988, and that the Point Lepreau reactor in the province of New Brunswick survived a serious quake in 1983. In fact, the epicentre of the November 1988 earthquake was 230 km from Gentilly-2. There has been no earthquake near Point Lepreau since the plant began operation in 1983.[28]
4. The AECL Design for Akkuyu
By 1993, there was an active debate in the AECL reactor design team regarding the options and trade-offs of various design alternatives to withstand higher earthquake risks. The Îreferenceí plant for Akkuyu is the Wolsung-2 reactor in South Korea[29], which is only qualified to a DBE of 0.20g.[30] It was assumed then, as now, that Akkuyu would only require 0.25g, but that other possible CANDU purchasers would require higher seismic qualification -- specifically, 0.34g to 0.6g for Indonesia; 0.6g for the Philippines; 0.4g (or higher) for Taiwan; and 0.3g to 0.35g for Thailand.[31]
The modifications to upgrade the CANDU-6 from from 0.2g to 0.25g DBE for the Akkuyu plant were described as ìconceptually simple but extensiveî -- design changes that were estimated to cost $35 million.[32] These design changes included: changes to civil engineering; process equipment and piping; control and instrumentation; and reactor fuel channels and fuel handling equipment.[33]
Essentially these changes represented a ìstiffening and strengtheningî of the earlier CANDU-6 design. Even in 1993, however, it was recognized that the upgrade to 0.25g DBE represented the seismic design limit of the CANDU-6... ìThe 0.25g represents the envelope design limit of the CANDU-6 with moderate modifications.î[34] It was thought that favourable soil conditions, combined with ìstate-of-the-artî seismic technologies might extend the limit to 0.26g or 0.27 g. However, higher seismic qualification would require conceptual or lay-out changes.
AECL engineers concluded that a Îfixed-baseí CANDU-6 with ìmajor redesign involving features similar to those used in the CANDU-3 designî could be qualified to 0.3g. However, this was characterized as a ìdrasticî design change, and ìits similarity with current CANDU-6 will be only in name. It would appear that the 0.3g represents, perhaps, the envelope limit for a fixed base completely redesigned CANDU-6.î[35]
The alternative to a fixed base design was a reactor with so-called ìbase mat isolationî or ìseismic isolationî -- essentially floating a foundation pad using alternating steel and rubber disks to absorb the energy of an earthquake. AECL engineers considered base mat isolation as a necessity for any DBE over 0.3, enabling the CANDU-6 design to be used on sites requiring up to 0.6g DBE or more.[36] AECL engineers argued successfully that seismic isolation should be used for the Akkuyu design, since the design cost would be no more than the cost of the fixed base modifications, and it would be a one-time design investment that could be used in proposals to other countries with a high earthquake risk.[37]
There are a number of serious engineering problems associated with seismic isolation. Solutions must be found for a movement of ± 12 to 15 inches between isolated and fixed base structure connections such as pipes and conduits. Flexible bellows or articulated couplings will be needed for large pipes such as the steam mains and feed water lines. There are major licensing issues since there is ìunfamiliarity and lack of long term experience with base mat isolation particularly in the nuclear industryî.[38] Thus there are risks associated with seismic isolation, which has not been used in any CANDU plant to date. With no reference plant using this design feature, there is a real risk that untried design modifications will cause problems, possibly resulting in higher costs and ultimately the risk of poor performance under accident conditions. Moreover, since the floating foundation represents a fundamental change in the CANDU-6 design, AECL is in violation of the original TEAS terms of reference for the Akkuyu bid. The terms specified that bidders should designate a reference plant ìsimilar to the one to be offered for Akkuyuî, and that there should be ì5 year operation experience for the type [of nuclear plant] offeredî.[39] That is clearly not the case with AECLís Akkuyu bid. The Wolsong-2 reactor in South Korea, designated as the AECL reference plant, is a fixed base design with a lower DBE.
5. Conclusion
The Izmit earthquake was a tragic example of how deadly these natural disasters can be. The Turkish government ignored warnings of this earthquake. The same government is ignoring the real earthquake risk at the proposed nuclear plant site at Akkuyu Bay.
The Turkish government failed to enforce building regulations despite the well-known earthquake risk in north-western Turkey. If the government proceeds with a nuclear power program, a similar failure in nuclear regulation could result in another disaster with even worse consequences -- the radioactive contamination of the eastern Mediterranean region.
The western nuclear vendors bidding to build the Akkuyu nuclear plant are conspiring with the Turkish government to cover up the real earthquake risk at the Akkuyu site. An acknowledgment of the real risk would result in more delays in construction; increased design costs for the vendors; and increased construction costs for Turkey. Recognition of the real earthquake risk at Akkuyu could result in the cancellation of the Akkuyu nuclear plant. Trade-offs are being made in the nuclear power industry between cost and safety. These trade-offs are even more worrisome in countries such as Turkey with a weak regulatory environment, and a nuclear regulator possessing no experience with large-scale commercial nuclear power projects.
There is little doubt that the earthquake risk at Akkuyu is higher than Turkish authorities first claimed in 1983. Specifications issued by TEAS for the Akkuyu site called for a Design Basis Earthquake (DBE) of 0.25g (i.e. a conservative assumption that the worst earthquake that can be expected will cause horizontal shaking equal to 25% of the force of gravity). However, it seems likely that the DBE for the site should be at least in the range of 0.5g to 0.7g (i.e. horizontal shaking equal to 50% to 70% of the force of gravity). Thus the proposed nuclear plant designs from all three bidding vendors at 0.25g are inadequate and if built, could result in a catastrophic nuclear accident caused by an earthquake.
Further independent geo-physical research needs to be conducted in the
Akkuyu Bay area, and specifically on the Ecemis Fault. The government of
Turkey should put the vendor selection process on hold pending the achievement
of a broad scientific and lay consensus on earthquake risk at the Akkuyu
Bay site. The government must recognize that this may mean a change of
the bid specifications to require a higher seismic qualification. This
will undoubtedly entail further costs for the bidding vendors as well as
further delays. As the costs and the risks of nuclear power become more
apparent, it is to be hoped that Turkey will abandon the nuclear power
option and consider the benefits of opting for a truly sustainable energy
future -- a non-nuclear energy future that is based upon efficiency and
renewable energy. High efficiency natural gas plants can provide a low
cost and relatively clean bridging technology to this sustainable future.
Appendix A -- Earthquakes within 100 km of Akkuyu since 1973
[available only in hard copy by fax or mail]
Endnotes
1. Reuters, ìTurks cannot postpone nuclear bid decision - Ministerî,
July 14, 1999.
2. Known as a ëcommon modeí or ëcommon causeí accident. Common cause accidents can be either external (such as earthquakes or a bombing of the nuclear plant), or they can be internal (such as a fire).
3. Gordon Thompson, Institute for Resource and Security Studies, Risk implications of Potential New Nuclear Plants in Ontario: Summary (Vol. I), Coalition of Environmental Groups, December 1992, p. 15.
4. Ann MacLachlan, ìScrapped Co-60 source forces 14 people to hospital in Turkeyî, Nucleonics Week, January 21, 1999, pp. 14-15.
5. Peter Calamai, ìEarthquake prediction takes a joltî, Toronto Star, September 12, 1999, p. F8.
6. Reuters/AP, ìHope Fades for Missingî, Toronto Star, August 25, 1999, p. A14.
7. AP, ìAftershock kills one, injures 166 in west Turkeyî, Toronto Star, September 1, 1999, p. A12.
8. Selcan Hacaoglu, AP, ì146 hours in darknessî, Toronto Star, August 24, 1999, p. A12.
9. AP/CP, ìPM says media source of low moraleî, Toronto Star, August 26, 1999, p. A12.
10. Lori Montgomery, ìIslamic quake aid hamperedî, Toronto Star, August 27, 1999, p. 10.
11. Amnesty International, Turkey: No security without human rights, AI Index: EUR/44/84/96, October 1996.
12. AECL claims that these studies were ì...carried out by Turkish experts (Middle Eastern Technical University) and international consultants (Swiss and German consultants) including the International Atomic Energy Agency to determine the Design Basis Earthquake (DBE) for the Akkuyu site. As well, a micro-earthquake monitoring program was carried out for about 10 years.î Source: Undated and unreferenced AECL document excerpt, Technical Summary of Akkuyu Seismic Design.
13. Ralph Kall, Supervising Geologist, Review of Interim Report on Cross-Hole Geophysical Survey at Akkuyu NPP Site, Turkey, Montreal, Quebec, December 11, 1985. AECL released this study on December 11, 1998.
14. S.L. Gokcen, G. Kelling, A. Ulug, N. Gokcen and E. Ozel, ìNeotectonic Structural Features in the Alanya-Mersin Shelf Area (Southern Turkey), Jeofizik, Vol. 5, No. 1, March 1991.
15. Greenpeace Mediterranean, ìTurkish reactor to be built next to active fault line, according to 1991 studyî, News Release, July 2, 1998.
16. AECL, Undated and unreferenced document excerpt, Technical Summary of Akkuyu Seismic Design, received 1998. See also: AECL Briefing Note: Seismic Safety of the Proposed Nuclear Power Plant Site in Turkey, received September 1998.
17. Ibid., Greenpeace Mediterranean.
18. Mohajer is a professor of geology at the Scarborough Campus of the University of Toronto. He had a brief opportunity to peruse the TEK documents at a meeting with AECL officials on November 27, 1998. See his brief summary letter: ìSeismic Risk Assessment for the Akkuyu Nuclear Plant, AECL Meeting of Friday, November 27, 1998î.
19. Ibid., Mohajer.
20. Acres International, Seismic Assessment Systems and Components at Pickering A, Atomic Energy Control Board, January 11, 1999, p. 2-3.
21. USNRC, Procedural and submittal guidance for the Individual Plant Examination of External Events (IPEEE) for severe accident vulnerabilities, NUREG-1407, 1991. Cited in Ibid., Acres International, Appendix A, Letter from Geomatrix to AECB, February 9, 1998.
22. Ibid., Acres International, Appendix A, Letter from Geomatrix to AECB, February 9, 1998.
23. Ibid., Acres International, Appendix A, Letter from Geomatrix to AECB, February 9, 1998.
24. US Geological Survey Earthquake Data Base. The quake took place at Latitude 34.35 N and Longitude 32.15 E.
25. Reuters, ìEarthquake rocks Turkeyî, Toronto Star, June 28, 1999, p. A14. See also: ìTurkey Quakeî, Globe & Mail, July 2, 1999, p. A14.
26. Ibid., USGS Data Base.
27. See: Le Droit, August 24, 1998; Fredericton Daily Gleaner, August 27, 1999; Patrides, November 20, 1998.
28. Karl Buckthought, Selling CANDU Reactors to Turkey: Seismic Risks, February 17, 1999.
29. Personal communication from Melda Keskin, May 27, 1998. The reference plant for the NPI bid is Neckarwestheim-2 in Germany, and the reference plant for the Westinghouse/Mitsubishi bid is Ohi-3 in Japan.
30. S.A. Usmani, Seismic Design Envelop Limits of CANDU-6 and Recommendations to Enable Offer of CANDU-6 to Higher Seismic Sites, AECL July 14, 1993, p. 1.
31. Ibid., S.A. Usmani, p. 1.
32. Ibid., S.A. Usmani, p. 1.
33. Ibid., S.A. Usmani, pp. 2-3.
34. Ibid., S.A. Usmani, p. 3.
35. Ibid., S.A. Usmani, p. 3.
36. Ibid., S.A. Usmani, p. 5.
37. J. Biswas & R. Ricciuti, Memo to A. Usmani, AECL, July 15, 1993, p. 3.
38. Ibid., S.A. Usmani, p. 6.
39. Nevzat Sahin, Head of Nuclear Power, TEAS, TEAS Akkuyu Nuclear Power
Plant Project, p. 2.
Nuclear Awareness Project
P.O. Box 104 Uxbridge, Ontario, Canada L9P 1M6
Tel/Fax 905-852-0571
E-mail: nucaware@web.net