ML20197E170
| ML20197E170 | |
| Person / Time | |
|---|---|
| Issue date: | 04/21/1975 |
| From: | Gossick L NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| To: | Anders W, Gilinsky V, Kennedy R, Mason E, Rowden M NRC COMMISSION (OCM) |
| Shared Package | |
| ML20197D284 | List: |
| References | |
| NUDOCS 9712290202 | |
| Download: ML20197E170 (24) | |
Text
- - - - ' - -
NUCLEAR REIULATORY COMMISSION
("ASHi%CT!N, D. C. 20s55 April 21, 1975 gf p y t ti1 2 t G *i't h 5 HENORANDUM FOR: Chairman Anders yscs s Commissioner Gilinsky /
Commissioner Kennedy Commissioner Mason
- Commissioner Rowden FROM: LeeGossick[
Ben Huberma i
SUBJECT:
NSSM 216: NATIONAL SECURITY ASPECTS OF RELEASING SAFEGUARDS INFORMATION This memo responds to your request at the April 17 Policy Session on NSSH 216 for a follow-on paper on the issue of classifying safeguards information. This paper is in three parts:
a comparison of the several approaches we could take to protect MUF data and sensitive safeguards information as they apply to the various types of facilities a brief discussion of the analytical shortcomings of the NSSH report .
-- four illustrative letters to the President
- l. POSSIBLE APPROACHES TO PROTE" TING SAFEGUARDS I NFORf!AT I ON For purposes of analysis, the issue of protecting safeguards information breaks down into two separate questions:
What is the need for special measures to protect information, over and above the protection afforded this Information today?
If the decision is that the information needs additional special protection, what is the best way of protecting it? ,
in answering the first question we will consider three categories of Information:
HUF data
- go onto+ --
the physical security and material accounting plans of installations having significant amounts of fissile material and, f u w n f : ? ' . ,:: 5*
+
. ::3 a
!?re. & s p a.4 ss
. g 3,,
yy De of Dec!stWa # '"/1s/ ;;/&.Q .
^
9712290202 971222 PDR ORO MEACAM PDR
.o e i o '2 o ';...
77iu '
The Commissionors ,
tho physical security and material accounting plans of -
installations, such as LWRs and LEU fuel fabrication facilities, where there may be some risk of sabotage but where there are no significant amounts of fissile material which could be readily made into a weapon.
In dealing with the second question, we will compare the implications of using.the national security classification system to protect information with the impilcations of using an NRC classification program along the lines the staff is developing.
Is additional protection needed?
Various types of information are subject to different threats, different consequences if the threat is realized and different degrees of protection under current practic.es. The NSSM report distinguishes between MUF Information and physical security and material accounting plans. In our analysis we have further separated the physical security and material accounting plans into those of facilities containing 1
significant amounts of Pu or HEU which might be stolen and those of facilities where such significant amounts are absent but where there might be a possibility of sabotage.
MUF Information. The threat seen to arise from the release of current MUF Information where an investigation,if necessaryc of the cause of unaccounted material had not be completed, is that a crank or terrorist could claim that he had managed to steal ur divert material shown to be missing in a MUF report and would use it in a bomb or dispersal device. To minimize
~ the number of such hoax threats,.the MUF data should be held back from public release for a certain period of time, to allow t.ompletion of any investigation into the cause of the Inventory discrepancy. In our view this is the most speculative of the various threats that need to be protected against. On the other hand, the amount of effort needed to hold the I,nforma t ion back for a period of time -- say six months -- is generally deemed Justifiable to guard against repeated false alarms by would-be terrorists or hoaxters, in future the hoax type of threat will assume greater proportions as fuel cycle facilities have greatly increased throughputs of strategic materials, hence, much larger levels. Inventory discrepancy data is currently withheld from public disclosure as proprietary data. Since there are two requests for disclosure of material discrepancy from The New York Times and Senator Ribicoff, which have been pending for several months, the problem of dealing with MUF is not theoretical. In fact, the MUF problem was the underlying reason that the national security study was conducted. A final point to consider regarding MUF is that it has become a highly controversial issue, has reached an emotional level, and reasonable men can be expected to disagree on whatever course of action is followed. 3 6
The Commission:rs .
-- Physical Security and Material Accounting Plans for FaciIitles Having SignifIcant Amounts cf Fissile Material. The unique threat in the case of facilities licensed to possess significant amounts of Pu or HEU is theft. This problem not only applies to fixed facilities such as fuel fabrication and fuel reprocessing plants, but also the transportation systems in which these materials are carried. At present, only a small number of facilities (23) with a small amount of material are in this category.
However, future licensing of HTGRs and fast breeder reactors and the use of recycled pitJtonium in LWRs would lead to substantial increases in both the number of facilities and anounts of material involved. There is general public appreciation that adequate security measures must be taken to keep this material from being stolen. And it follows that the measures used to guard this material from theft should not be disclosed to would-be thieves. The present method of protecting this information, through reliance on its proprietary character, is inadequate for three reasons.
First, the " proprietary" rationale is not the real reason we withheld the information; second, this method does not allow the NRC to assure the reliability of industry personnel-responsible for the security of the material or facilities; and third, by treating this information as proprietary we are relying on the good Judgment of the plant personnel for its protection. However, it must be acknowledged that the present system has been successful and has not been attacked through Freedom of Information claims or otherwise, let alone tested in the courts. But on balance we believe that the present measures to prevent disclosure of the security plans are at 1 cast potentially inadequate and that more certain means of protecting this information would be justifiable both in terms of the threat involved and public perceptions.
-- Physical Security and Material Accounting Plans for Facilities Which do not Contain Significant Amounts of Fissile Material.
This category includes LURs, some fuel cycle facilities such as those for fabrication of LWR fuel elements, and the trans-portation links in the LWR fuel cycle. While the danger of theft exists in these cases, it does not significantly differ from the danger of theft in industrial facilities as a whole.
The chief threat involved is that of sabotage. The consecuences of sabotage vary in turn,wlthin this general category. In some cases, such as fuel fabrication of LEU fuel elements, the consequences of sabotage present a relatively low threat to the general public safety. In other cases, such as sabotaging an LWR plant to cause meltdown, a significant dispersal of
The Commissicn:rs w l dangerous radioactive material could result. Blowing up a tru d containing spent LWR fuel elements is an intermediate case; some dlspersal of plutonium could result, but the very large amounts of high explosive needed to blow up the heavy fuel rod containers make the threat scenario somewhat speculative, it is our view that measures to protect the physical security plans of LWR power plants are of great importance from the point of view of public health and safety. For example, the loss to life and property resulting from the effective sabotage of a large nuclear power plant near a population center could be at least as great as the damage predicted in the more credible home-made bomb scenarios. Admittedly, there are other industrial facilities, such as dams, which pose comparable threats to public safety.sHowever, given NRC's responsibility to regulate the nucel r power industry from the view point of health and safet nd the clear threat which sabotage of a power plant could pose to the general pubile, we believe NRC should concern itself not only with the adequacy
'of the physical security plans in effect at power plants, but also with the adequate concealment of those plans from potential saboteurs. The present degree of protection afforded is the same as that described earlier; that i s , we t rea t the plans as proprietary information. As we pointed out above, there are drawbacks to this practice, in our view, therefore, if special measures are justified to prevent the unauthorized disclosure of the security plans at facilities which contain significant amounts of Pu or HEU, similar measures are justified for protecting the security plans of LWR power
. plants. However, in the case of LEU fabrication facilities and transportation links in the LWR fuel cycle, no measures in addition to the normal industrial security procedures now being practiced should be taken to protect security plans, since the security plans themselves are not now subject to HRC licensing controls.
To sum up, we believe MUF data ne.eds some additional degree of protection and that this issue should be faced promptly because of the requests for the data now pending. In the meantime, we think Senator Ribicoff's request should be at least partially satisfied so that any move to protect this data will not appear to be a direct response to his letter. The security plans of fuel cycle facilities (and eventually HTGRs and fast breeders) and transportation links where large amounts of fissile material are present should also be protected because of the consequences of theft. Similarly, the security plans of LWRs should be protected, because of the potentially serious consequences of sabotage to public health and safety. In none of these cases is the present system for protecting this information adequate.
. . _ .. _ _. . j
~-
The Commission:rs In the next section we will consider the two plans f protecting -
this information that have been proposed and analyze differences between them. ---
Alternate Ways of Protecting Sensitive Safeguards Infor.a ion Two methods have been suggested to provide a sound legal basis for withholding sensitive safeguards information from the public.--
classification under Executive Order 11652 or an NRC generated clearance and classification system which would be implemented in conjunction with a statutory exemption to the Freedom of Information Act, in general the relative merits of these two systems are not related to the type of information being protected. That is, whatever advantages or disadvantages an NRC classification system has compared to national security classification apply across the board to MUF information, the security and accounting plans of fuel cycle facilities, the sccurity plans of transportation links, or the security plans of power plants. The characteristics of the two systems are as follows:
National Sec,urity Classification Under Executive Order 11652. On the grounds that the release of fiUF data or security and accounting plans could damage the national defense, the material to be pro-tected would be classified " Confidential" under Executive Order 11652.* It would then have to be stored in safes, divulged only to cleared personnel with a need-to-know and transmitted and destroyed under established procedures. The clearance of the personnel involved wouid be done through a national agency check (NAC) or background investig6 tion (BI). The material would be classified both in the plants and in the NRC files and would be
- protected from divulgence under the existing exemption for classifled information in the Freedom of Information Act. Inter-venors and parties in Interest in NRC proceedings could be given access to these security plans upon being cleared through an NAC.
-- NRC Classification System.
~
The material in the hands of Ticensees and in NRC files would be handled according to NRC regulations with limited access to the material to those with a need-to-know. The material in the hands of licensees would be stored in safes, registered and numbered, and subject to required procedures of transmittal and destruction. It would be handled similarly in our own internal files. In accordance with a separate regulation, personnel having access to SNM would be subject to an NRC clearance procedure on the basis of an NAC or Bl. For the most part it would be the same individuals who would have access to safeguards information.** Because neither the NRC regulations nor the NRC clearance procedure; would prevent this information f rom being divulged under the AThe Executive Legal Director is sending you a legal brief concerning classification of privately produced information.
- The Executive Legal Director is sending you a legal brief concerning NAC's authority to require clearances of non-licensee personnel.
r .
..- W Frc: dom of Information Act, o specici cxempticn to that Act would have to be granted by Congress. Intervenors and other parties in NRC proceedings could have access to the sensitive material if they agreed not to divulge it.
Theoretically, intervenors would not require a national agency check; however, there would be practical problems in distinguishing between licensee employees and Intervenors in terms of the precautions taken to protect the information.
It should be noted that, under an NRC system, Information control and personnel clearances are separable. It would be l possible to set up an information control system without '
personnel clearances, though the degree of protection would then be lower. ,
in terms of practical impact there is very little difference between .
these two systems of preventing the sensitive Information from being I divulged.
Both systems appear to be effective. In the case of national security information, the system is in place and functioning.
Although the NRC system has not been set up yet, it does not appear to have any obvious drawbacks other than the need to obtain a specific congressional exemption for safeguards information from the Freedom of information Act. Either system could be fine-tuned with respect to the number of people cleared or the amount of information covered. Either system would cost about the same and create about the same personnel demands on NRC, as discussed last Thursday. 1 As to public impact, a national security classification system
' may be perceived differently by the public than an NRC generated classification system. However, the major impact would result from the imposition of some system to prevent public disclosure of the information, regardless of which system was used. The fact that a classification system was being imposed several years after the fact (security plans for power reactors were first required in 1967) would further focus public attention on the adequacy of prior government practices in regulating the commercial nuclear Industry, in our Judgment, it is not possible to predict the public perception of national security clasrification versus an NRC system. ble fe e l i t is likely that the Commission will be vocally criticized for either course of action. However, the majority of the public will be unconcerned by either course of action.
An NRC classification system would be more flexible. As mentioned above, personnel clearancesare separate from a document control system. Moreover, an NRC system could be replaced by national security classification at some time in the future, if that were found necessary or desirable. (Replacing national security classification with an NRC system would be much harder.)
.~
.' i, i
t The Commissicners .
The institutional Irnpac'. ,pf,the two systems on NRC's relations to the national security agencies probably would not differ a grent deal in practice. An NRC cic25irication system muld emphasize our separation from the national security r;iencies in the Executive.
Branch, while the use of national security classification system would tend to create some nexus. However, in ort not to exaggerate this problem two factors should be kept in mind. First, even if NRC sets up its own system s r.e degree of coordination with the Executive Branch will probably be necessary so that the criteria ERDA uses and the criteria NRC uses to protect the same information will not be grossly out of phase. Second, the normal practice is that each agency administers its national security classification program autonomously. Thus, once the President had established that the information must be classified, the NRC would be responsible for imposing and administering a system of controls through the regulatory process, just as in the case of an NRC classification program.
With regard to International implications, there is little difference between the two systems. In implementing the U.S. safeguards offer we would be guided principally by the sensitivity of the U.S. safeguards information, not by the formal modality of classitying it.
In sum, we think the case favoring either system over the other is not everwhelmingly strong. However, we come out differently in our recommendations.
Gossick believes we should set up ar. NRC classification and clearance system. This would be a re n propriate to NRC's role as an open civilian agency, with regard both to our pubilc image and our relations with the Executive agencies. This recommendation is based on the premise that an NRC classification program could be made as effective as a national security classification system and, if it becomes necessary at some later date to classify some or all safeguards information based on national security, that the NRC could and would be prepared to take that action.
Huberman favors national security classification. The system is in place and operc+.I g, the national defense criteria are appropriate to the nature of the threats involved -- including the threat of sabotage to LVRs --
and classification is consistent with the measures we require licensees to take to protect the material and facilities. Beyond that, the relative advantages and disadvantages of a national security versus an NRC classi-fication system is marginal. Each would be effective.
1
The Commission rs W public perceptions of the two systems would not differ
~
greatly and procedures for giving Intervenors necessary access to the informatlod'could be worked out almost as easily for the national security system as for the NRC system.
Id. OlSCUSSION OF THE ANALYTICAL SHORTCOMINGS OF THE NSSM 216 REPORT -
The following major issues do not appear to be adequately anal' zed in the report subn.itted in response to NSSM 216. These deficiencies may have little impact on the final conclusions of the report that:
certain sensitive safeguards information generated by NRC licensecs should be withheld from public disclosure, and Inventory discrepancy data can be released after some appropriate time period However, satisfactory treatment of these issues may alter the recommendation of the report that sensitive safeguards information should be protected utilizing the national security classification system.
The Commission may want to draw on some or all of these issues to support the conclusions or recommendations in its letter to the President or might wish to separately transmit the list of issues to the NSC staf f for whatever action is appropriate.
Adequate Protection of information The report assumes that the present degree of protection is inadequate without rigorously analyzing the question. In addition, the report falls to distinguish between MUF statistics and physical security and other accounting information. A further important distinction --
not addressed in the report -- concerns the differences between light water reactors and low enriched fuel fabrication facilities and facilities with, and transportation of, SNM.
MUF Information. The NSSM analysis supports the need to establish a system to withhold MUF information for a certain minimum period of time. In developing specific plans to protect MUF data, the six-month delay should be analyzed-along with other possible periods in terms of inventory and accountin9 Practices now in use.
.' .' 1 d The Commission:rs .
F,uel C/cle Facilities and' Transportation Links Containing Special Nuclear Material. There is general appreciation that the theft of SNM could constitute a significant public threat.
It therefore makes sense that the measures intended to thwart theft should not be disclosed to potential thieves.
Light Water Nuclear Power Plants and Low Enriched Uranium Fuel Fabrication Facilities. There are obvious distinctions, not made in the report, between light water reactors and their support fuel cycle facilities and SNM fuel cycle facilities. Since LWP, plants and fuel fabrication facilities do not contain fissile material in a form that could be used to make nuclear devices, the problem of protecting security plans of fuel fabrication facilities and transportation of fresh and spent fuel is closer to che general issue of industrial sabotage. However, security plans for power plants should be protected because of the potential danger associated with a massive release of fission products.
Trustworthiness of Individuals.
The report does not discuss the issue of trustworthi,ess of individuals involved in the nuclear industry. Currently, no security clearances are required for personnel in the civilian nuclear industry. Thus, treating the issue of control of information without addressing the trustworthiness or loyality of people who will make any administrative control system for sensitive information work, falls to recognize that any system is only as good as the people who operate it.
Feasibility of Classification The report suct_rily disposes of two legal obstacles in reaching its conclusion in favor of classification -- the fact the the information has not been classified before and the question of whether it is
" official." However, not even summarily treated is the question of whether the information meets the requirements in the Executive Order relating classification to " national Jefense or foreign relations."
Exemption to the Freedom of Information Act The report concludes - without supporting analysis -- that it would not be possible to obtain a legislative exemption to the Freedom of information Act to enable MUF data, physical security and accounting Information to be withheld from the public. However, the report fails to mention that the FAA was recently successful in obtaining Congres-sional approval to exempt airport security plans from Freedom of Information provisions.
4
l l
The Commission 3rc ,
IAEA Problem The report correctly concludes that classified safeguards information could be provled to the IAEA with an appropriate understand 1og of its protection. However, the report does not draw the necessary connection between sharing this information with the I AEA and the need to protect it from irresponsible governments which support terrorist groups.
The relationship between a policy decision to classify information and the decision to place the U.S. nuclear industry under international inspections must be considered. .
Interim Measures A problem not covered in the report which must be addressed by the NRC is the measures to be taken to protect information while the decision on NSSM 216, including follow-on work and implementation, is completed. With regard to physical security and accounting plans, NRC now has procedures in effect. The problem of MUF data is more difficult. Senator Ribicoff and The New York Times have requested MUF figures for all licensed facilities and those requests are still pending. It would appear at a minimum that some data should be provided to Senator Ribicoff, perhaps excluding the most recent information, along with an explanation of the delay.
Impact on Public and Industry The underlying concern about classification rests on two questions --
its impact on the public and its burden on the industry. The report inadequately deals with both issues.
Public Imnact. Where significant amounts of plutonium or HEU are involved, there is a general public perception that the industry should be organized in a way that will deny unauthorized access to that material or information about the protection c' that material.
There are no readily usable accumulations of fissile material in low enriched uranium nuclear power plants, transportation of fuel and fuel fabrication facilities. Classification may be perceived as a qualitative step toward closing the industry.
Burden on Industry. There is little if any additional impact on the industry to classification versus an NRC system of control.
The economic costs under either program seem to be relatively small and the incremental dif ference negligible.
^
~. .. . . . .- ~ .-
9 m
The Commissioners -
111. ILt.0STRATIVE LETTERS TO THE PRESIDENT We have prepared four illustrative" letters to the President, which might be useful to you in discussing the alternatives:
-- The letter at Tab A essentially accepts the recommendation of the NSSH report that sensitive safeguards information be classified. .
The second letter (Tab B) urges the President not to classify on three grounds: the information to be protected is not properly national defense i,nformation; public perception of classification smid be adverse; and the NRC already has the authority to protect the Information adequately through regulation. The letter suggests that the Administration and the HRC make a joint effort to obtain Congressional approval for an F01 exemption. The letter mentions that the HRC retains the option of classifying some of the information at a later time if that becomes necessery.
The third letter (Tab C) accepts the need to classify MUF data during the investigation of a possible loss. However, the President is - n' to send the study back to the Working Group for furthe- t;>;is of the question of classifying security plans an. erials accounting procedures.
-- The final letter (rab 0) recommends no action en any portion of the report until the Working Group has done more work. It also informs the President that NRC in the meantime is planning
, to take its own actions on controlling safeguards Informat'on.
^~
7 ~~* -ae ,
- - - ~. .. _._. p
,__ _ m,
-(Accept NSSM recosmCndOtiCn)
\-
Dear Mr. Presidents v..
I wish to convey to you the Nuclear Regulatory Commission's views on the study of the national security aspects of releasing safeguards informa-tion carried out under NSSM 216.
- We'Connission has reviewed the conclusions and recommendatians of the study and supports the working group i re wnendation that sensitive safeguards information be classified under Executive Order 11652. We are of the view that the minimum degree of classification and personnel clearance necessary to protect the information from disclosure should be brought into effect.
We will be pleased to participate in the neces-sary follow-on work to identify specific safeguards _information that might be classified and to alalyze the further practical implications of classification. Et:for to making our determinations on what specific safeguards information will be classified, we will provide the NSC staff with the opportuniti to comment on our proposed determinations.
I wculd like to thank you for the opportunity afforded the Commission -
staff to participate in the study. The participation of our staff was
?
of great assistence to the Commission in its diiberations on this issue. '
r Respectfully yours, r
k
, ',' .' i, 5l W/7 V (Rocammpnd again:t cla-aific: tion)
Dear Mr. Presidents o...
- I welcome this opportunity to give you the views of the Nuclear Regulatory Commission on the study of the national security asp 9 cts
-of releasing safeguards information carried out under NSSM 216.
~
It is tha view of the Commission that effective means to protect domestic
~
safeguards information are essential to guard against theft or diversion of nuclear materials or sabotage of civil nuclear installations.
However, we do not believe that classification of thin information under Executive Order 11652 is aither a practical or desirabic means to achieve the needed degree of protection.
o our first concern is that the criteris for classification set out in the Executive Order do not clearly encompass the kind of information that must be protected. It is not clear that the divulgence of information on the physical protection of licensed nuclear facilities or accountability data on licensed nuclear materials would cause " damage to our national defense or foreign relations." Criminal exploitation of this information could damage the domestic peace and security of the country, but this type of damage is not explicitly contemplated in the Executive Order. We are concerned, therefore, that a challenage to classification under the Freedom of Information Act might succeed in the courts,..thereby nullifying the protection of the information which we are trying to achieve.
A second problem relates to public perception of classit. cation. In view of the concerns for openness and public accountability that gave rise to the creation of the NRC, we are concerned that classification
Mr. Prc;0ident- would be viewed as an effort, to withhold from the public the type of information essential to an infomed"public debate on the licensing and regulation of the nuclear power industry.
Finally, the Atomic Eneegy Act gives NRC the authority to institute a document control and per sonnel clearance system. We are prepared and Planning to institute such a system, which would be more appropriate to NRC's role as an open civilian agency than a system of national security classification. We are convinced that an NRC classification program can be as effective as a national securi classification system. b # s w it W === h V M LLAWFhIYibh/h W LA-fd, M %q to U W T Luytp.Jd .NvY Whf ws%E /JU 4%4M//VMei Mh (A
For the above reasons, we consider that classification is not the pre-ferred alternative for affording the necessary protection to safeguards information. We would urge as an alternative that the Administration and the NRC make a joint effort to obtain from Cor.gress an explicit exemption to the Freedom of Information Act for domestic safeguards informa-tion. If we are successful in obtaining such an exemption -- which we believe is a realistia goal -- classification, with its attendant uncer-tainties and disadvantages, would not be necessary. In the event Congress would not act favorably on our request for an exception, the question of classification could then be reopened.
On behalf of the entire Commission I would like to thank you for inviting the NRC staff-to participate in this study. The participation of our staff has greatly facilitated our deliberations on this issue.
Respectfully yours,
- - _ ~
4/21/75 , !
(Class,1fy .10F, further study on rest).
1 1
s 1
Dear Mr. Presidents a...
I wish to _ convey. to you the Nuclear Regulatory Comunission's views on the' Estudy of.the national security aspects of releasing safeguards 4nforma-tin carried out under NSSM 216.
The Commission has reviewed the conclusions and recommendations of the
~
study and supports the working group's recommendation that any report on " material enaccounted for" which gives rit.e to an investigation be classified- for a period of time sufficient to permit the investigation to be concluded.
With regard to the portions of the study addressing the protection of the security plans and materials accounting procedures of civil nuclear facilities, we are of the view that the analysis contained in the report does not give a sound basis for deciding whether such information should be classified. We therefore urge that, before any decision is taken on the classification of such data, the working group be requested to analyze L further the issues involved in this -question, particularly as they relate to the authority the Commission now has to establish a perscnnel clearance and document security system through its own regulatory authority.
On- behalf of the entire Commission, I would like to thank you for inviting the NRC staff to participate in this study. The participation of our staff has greatly facilitated our deliberations on this issue.
Respectfully yours,
4 .' . . 4/21/75 (Furthirstudybeferod2 cision) i
Dear Mr. President:
I wish to convey to you the Nuclear'hegulatory Commission's views on the study of the national security aspects of releasing safeguards information carried out under NSSM 216.
The Commission agrees with the conclusions of the study that certain sensi-tive safeguards infortnation generated by NRC licensees should be withheld from public disclosure and that inventory discrepancy data can be released
.after some appropriate time perioc. However,.the Commission is concerned that the report fails to adequately analyze several major issues. Corree-tion of these deficiencies (further analyzed in the attachment)* could change the recommendation of the report that sensitive safeguards informa-tion-should be protected utilizing the National Security Classification system. Consequently, the Commission recommends that no decision be made on classification of certain sensitAve safeguards information until further ntudy on the subjecc is conducted. The Commission is ready to assist in any further analysis undertaken.
Regardless of the outcome of this study, the Commission la planning to issue regulations that will require licensees to limit dissemination of safeguards information to those with a need to know and implement the authority.obtained recently to require personnel clearances of individuals with access to Spacial Nuclear Materials. Neither of these actions will conflict with whatever final decision is made on classificatior. and they will provide an-immediate improvement in the level of protection provided Special Nuclear Materials and information about the protection of that 4 - material.
Not included in this illustrative draft.
'8* 6 H1 5 S,. get g 4N gg
--@a 4-g WM *-
- 4 (49m -
~' '
- Mr.:Prc0ident- -
2-
. I would like to thank you for the opportunity afforded the Commission staff to participate inL the study. Jhe participation of our staff was of great assistance to the Commission in its deliberations on this issue.
Respectfully yours,
.J :
in N
t i . .
9 d
I l
l ,
i e
1
- ' a w: _ ,,
j
FT 4/18/75
,. MAJOR ISSUES NOT ADEQUATELY ADDRESSED IN NSSM 216 #5S+f de .
- The-following major issues do not appear to be adequately addressed in g, J--
- the report submitted in response;to NSSM 216. The impact of these m>7/yp deficiencies may have little impact on the final conclusions of the report that:
. certain sensitive safeguards information generated by NRC licensees should be withhcld from public disclosure, and
. inventory discrepancy data can be released after some appropriate time period.
However, satisfactory treatment of these issues may cause a change in the reconrnended methods of achieving the objective of the safeguards program to insure that the common defense and security and public health and safety are protected from malevolent acts or the threats of such acts associated with nuclear materials or facilities.
RELATIONSHIP 0F PROTECTION OF SAFEGUARDS INFORMATION AND TRUSTWORTHINESS OF INDIVIDUALS Adequate Protection t
The NSSM report assumes that the present degree of protection is inadequate without rigorously analyzing the question. It is important to distinguish clearly between MUF statistics on the one hand and physical security and other accounting information on the other. A further important distinction --
not addressed in the report -- is between light water reactors and low enriched fuel fabrication facilities, and strategic nuclear material (SNM) fuel cycle facilities and transportation of SNM.
t -- MUF Information - Some f arm of protection for MUF data is
- ;m < required if the public is to be protected against repeated
.. . felse alarms from would-be terrorists. The NSSM analysis supports the need to establish a system to withhold this information for a certain minimum ceriod of time. It is.not
~
'y '$$ 4: D clear why the period should be six months, as the report suggests. In developing specific plans to protect MUF data, L. either through classification or otherwise, the six-month
\ delay should be analyzed along with other possible periods in terms of concrete inventory and accounting practices now in
.- _. sg use.
OL '9 db -- Strategic Nuclear Material Fuel Cycle Facilities and Transportation
- - p g ; j. inks - The rationale for withholding security plans for the pro-
% Of . tection of substantial quantities of fissile material is' clear. T e SUBJECT-TO GENERAL DECLASSIFICATION SCHEDULE 0' TCUTIVE ORDER 11652 AUTOMATICALLY DOWN- ,
j l L J AT 2 YEAR INTERVALS AND GECLASSIFIED ON uECEMBER 31 iW ~
Y 3..m.iATIOF y p- = - ----
- 0. Copy f ofj
.- 2-is general public appreciation that the theft of such nuclear material would constitute a significant public threat. It therefore makes sense that the measures intended to thwart theft should not be disclosed to potential thieves. On the other hand, the report correctly identifies the general need to keep the amount of protected information to a minimum both to facilitate the job of protecting it and to permit the maximum amount of data on nuclear power to be available to the public.
The present method of protecting this information, through reliance on its proprietary character, is inadequate for two reasons. The " proprietary" rationale is not the- real reason we withhold the'information and by treating this information as proprietary we are relying on the good judgment of the plant personnel for its protection. On the other hand, it should be acknowledged that the present system has been successful and has not been attacked through Freedom of Information claims or otherwise, let alone tested in the courts. Nevertheless, it is probably sensible to seek surer and better justified means of protecting this information before any challenges are encountered.
-- Light Water Nuclear Power Plants and Low Enriched Uranium Fuel Tabrication - There are obvious distinctions, not made in the report, between light water reactors and their support fuel cycle facilities and SNM fuel cycle facilities. Since LWR plants and fuel fabrication facilities do not contain fissile material in a form that could be used by terrorists to make bombs, the
, problem is closer to the general issue of industrial sabotage.
But the short and long-term consequences of sabotage of a nuclear power plant dictate surer protective measures than those practiced in ordinary industrial operations. If it is decided that security plans of LWR plants should be protected, relying on the " proprietary information" rationale has the same drawbacks as in the case of fuel cycle facilities. If so, a judgment must be made balancing the various methods available to prevent sabotage against their impact on a large-scale civilian industry.
-- Trustworthiness of Individuals - The report does not mention the issue of trustworthiness of individuals involved in the nuclear industry. Currently, no security clearances are required for personnel in the civilian nuclear industry. Thus, treating the issue of control of information without addressing the trustworthiness or loyality of people who will make any administrative control system work fails to recognize that any administrative system for protection of sensitive information is only as gcod as the people who make it operate.
_ m , m
Copy / of[
o, Classification of safeguards infomation will require that security clearances be granted those individuals with access to the infor-mation. . Implementation of the report's recommendations; however, may not require clearances for_the very people who have. access to SNM. The NRC has under preparation a regulation that would require
, clearances for personnel with access tc or control over SNM. This effort must be integrated into any total protective package to be developed and implenented.
Is Classification Feasih,le In reaching its conclusion in favor of classification, the report disposes of two possible legal obstacles -- the fact that the information has not been classified before and the question of whether it is " official."
However, no t>aatment is given to the question of whether the definition in the Executive Order, which relates classification to " national defense or foreign regulations", can logically be applied to MUF data and physical security and accounting plans. In the question of MUF data and physical security plans to protect fuel cycle facilities and transportation links against theft of fissile material, the standard of " national defense" appears to be logical, particularly in an era of widespread political terrorism.
Here again, however, the case for using national defense grounds to withhold the physical security plans of power plants might be weaker. Certainly, if a challenge to classification under the Freedom of Information Act succeeded-in the courts, this would nullify the protection of the information which is the basic purpose of this effort. On the other hand, determination by the President that MUF data and physical secucity plans are classifiable under the Executive Order would carry a strong presumption of authority since the President also has the power to modify the Executive Order. Therefore, the conclusion of this report that classification is legally defensible (apart from whether it is desirable) seems to be cor. rect.
Exemption to the Freedom of Information Ac.t The NSSM report concludes -- without supporting analysis -- that it would not be possible to obtain an exemption to the Freedom of Information Act to enable MUF data and physical security information to be withheld from the public. It is difficult to evaluate this conclusion, given the absence of data. Certainly, recent Congressional action on the Freedom of Information Act seem2 to point toward a more open attitude on the release of official facts and figures. On the other hand, the FAA was recently successful in obtaining Congressional approval to exempt airport security plans from Freedom of Information provisions. More data on this FAA action, which is not mentioned in the NSSM report is required. The NSC should find out as much as possible from FAA. including the amount of time-it took for them to get the exemption, who they dealt with, and what tactics they used.
= -
' ^ ~~ ^ - -
4 One important factor should be borne in mind. If we attempt to gain Congressional approval for a Freedom of Information exemption and are turned down on substantative grounds. it will be very difficult to classify the information.
The IAEA Problem The NSSM report correctly concludes that classified safeguards information could-be provided to the IAEA legally with an appropriate understanding on
. its protection. Classified information is shared with international-organizations and foreign governments routinely -- even in some cases with the Soviets. However, the report does not draw the necessary connection between sharing this information with the IAEA and the need to protect it from irresponsible governments'which support terrorist groups. In implementing the Presidential safeguards offer, it will be necessary to give the IAEA inspectorate access to U.S. plans, along with some infonnation on our physical security methods. This problem may be manageable if we are careful to identify possible areas of sensitivity in working out the details of the safeguards offer with IAEA. However, the solution cannot be taken for granted. Our participation in the voluntary safeguards offer should permit IAEA to do its job and set a standard for the physical security measures we want to see in other countries. On the other hand, it may be difficult to justify giving sensitive safeguards information to radical Arab states which support terrorism. The relationship between the policy decision to classify infor-mation and the decision to place the U.S. nuclear industry under international inspections must be considered.
Interim Measures A problem not covered in the NSSM report which must be addressed is the measures to be taken to protect infonnation while the decision on NSSM 216, including follow-on work and implementation, is completed. With regard to physical security plans, NRC now has procedures in effect (the " proprietary information" theory) which are not under challenge. In the event a demand for publication were made under the Freedom of Information Act while the NSSM project is in train, NRC may be able to persuade the court that since a Presidential study is in process, the matter should be deferred until a decision on the study is reached. The problem for MUF data is more difficult.
4 Senctor Ribicoff and The New York Times have requested MUF figures for all licensed facilities and those requests are still pending. The matter has been outstanding since November in the case of The New York Times and since January with Senator Ribicoff. (Incidentally, the history of this issue as presented in the NSSM report fails to point out that part of the delay was caused by an original AEC decision to release the data, followed by a reversal F
4 I
.s~_
Copy o e
m.
- and the AEC request to the White. House for the present study.) It would appear at a minimum that at some point the data should be provided to Senator Ribicoff, perhaps with some appropriate time lag -along with an explanatien of the delay. Continued refusal of The New York Times request risks a' Freedom of information action.
Impact on Public and Industry The underlying concern about classification seems to rest on two questions --
its impact on the public and its burden on the industry. The NSSM report inadequately deals with both issues.
-- Public Impact - To analyze the question of public impact it will be useful once more to break the problem into its parts -- MbF data, fuel cycle and transportation' safeguards, and nuclear plant protection. The protection of MUF data and information on fuel cycle and transportation safeguards would seem to present fewer problems with regard to public perception of classification simply because the materials involved are inherently dangerous and the processes are' clearly associated with similar processes in the national security sector. It can be argued that classification might not be the most effective or most desirable method of pro-tecting this information, but it is difficult to assert on the one hand that we need a more effective means than classification to deny access, while arguing on the other hand that these more effec-tive means will somehow result in a more open nuclear industry.
The fact is that where significant amounts of pluton 1um or HEU are involved, there is a general public perception that the industry should be organized in a way that will deny unauthorized access to that material.
Low enriched uranium nuclear power pl' ants are a different story.
Here' the situation we are trying to create is that of inherent safety. Except for the future cases of HTGR's and fast breeders, there are no readily usable accumulations of fissile material present. Classification, it could be argued, would be perceived as a qualitative step toward closing the industry. Even on the question of power plants, however, no one argues that the physical security plans should-be publicly available. They are protected now and we should have some means to protect them in the future.
But there may be a stronger argument in the case of power plants for finding alternative means to classification, if available.
,. /
~, l i n f, --
Copy / og $ ,,,
Q e
- . A final point with regard to public image is the position of
~
intervenors. In this case, classification versus some other means of denying public access seems to make e distinction without a difference. The information will not be publicly available in any event yet it should be available under appropriate safeguards to intervenors. Making it available to intervenors, whether it is-classified material, proprietary data, or infonnation protected through a Freedom of Information exemption, is an administrative problem that could be readily handled for any of these categories.
-- Burden on Industry - There is general agreement that information that would be of assistance to would-be thieves and terrorists should somehow be protected, and that as small an amount of information as possible should be kept from the public. From this premise,
.there is little if any additional impact on the industry to classification versus a system of unclassified control. The NRC is now working on a procedure to clear most of the licensee personnel in both fuel cycle facilities and power plants. The chief advantage of a national security classification system is that it would permit clearance of certain essential non-licensee personnel (architect-engineers, transportation executives, and so on) whom we may not be ,
authorized to clear under the current legislative authority. In this sense, classification could be more effective. The ecor.omic costs under either program seem to be relatively small and the incremental difference negligible. The only way to have a significantly lower impact on the industry is to decide to exercise little control over the dissemination of physical security plans and MUF data and to hold off on implementing the clearance authorization NRC now has.
RECOMMENDATIONS The report recommends classifying certain saleguards i information then developing a plan for implementation. It might be more appropriate to release some MUF information for example that from October 6,1974, and before to reiteve the immediate pressure then to develop a detailed plan for implementing
, a national security based classification program. Then, after the full impact of classification on the industry, public, and government has been evaluated, a decision as to need for and feasibility of classification of safeguards infonnation could be intelligently made.
N
., ,.,y.