Opinion for the Court filed by Circuit Judge STARR.
STARR, Circuit Judge:This petition for review challenges the decision of the Nuclear Regulatory Commission (NRC or Commission) granting operating licenses for Units 2 and 3 of the San Onofre Nuclear Generating Station (SONGS Units 2 and 3). These facilities are situated near San Clemente, California within the perimeter of the United States Marine Corps Base at Camp Pendleton. The challenge is mounted by August S. Carstens, an individual residing in San Diego, California, and Friends of the Earth, a California nonprofit corporation. Mr. Carstens and Friends of the Earth were permitted in 1977 to intervene and participate in the NRC proceedings culminating in the issuance of the operating licenses under review here.
I
A
The issues before us relate primarily to the location of the SONGS facilities in an area of seismic activity in Southern California. These issues are by no means new. They were raised before the Commission by prior intervenors, who did not participate in the operating license proceeding, at the time the utility-applicants sought construction permits to build these facilities. Those permits were granted by the Atomic Energy Commission (now the NRC) in 1973.1 In March 1977, applications were filed for operating licenses for the two SONGS Units. Following petitioners’ successful motion to intervene in May 1977, the NRC’s Atomic Safety and Licensing Board (Licensing Board) admitted for discovery purposes petitioners’ contention that the seismic design of the facilities was inadequate:
The seismic design basis for SONGS 2 and 3 is inadequate to protect the public health and safety and does not comply with [applicable regulations] in that the earthquake which could cause the maximum vibratory ground motion has not been assigned as the safe shutdown earthquake.
Prehearing Conference Order of January 27, 1978 (reprinted in Petitioners’ Brief at 6-7).
Discovery was undertaken with respect to the seismic issues pursuant to the Licensing Board’s August 27, 1980 order. Written testimony was filed by the petitioners, and answers to interrogatories were provided by the applicants and the NRC staff.2
The hearings began in June 1981 and consumed 25 days. Twenty-eight witnesses testified before the hearings concluded on August 4, 1981. Testimony at the hearings addressed a number of issues raised by the intervenor-petitioners with respect *396to the adequacy of the seismic design for SONGS Units 2 and 3. Of especial relevance was testimony with respect to three fault areas located in the vicinity of the SONGS Units. The first area is a zone known as the “Offshore Zone of Deformation” (OZD), situated off the coast of Southern California. One hotly contested issue before the NRC’s Licensing Board was whether the OZD was a single, thoroughgoing fault or, instead, was a zone of linked segments of folds and faults unlikely to rupture along its entire length. The second fault zone situated even closer than the OZD to the SONGS facility is the Cristianitos fault, located at the closest point of its approach less than one mile from the plant.3 Throughout the lengthy prior proceedings involving nuclear power facilities at San Onofre, the Cristianitos fault was deemed by the NRC to be inactive. Petitioners vigorously contested this long-held assumption by virtue of two earthquakes that occurred near the San Onofre site in 1975.4
The third geologic feature examined in the Licensing Board hearings was the offshore Cristianitos Zone of Deformation, not to be confused with the on-land, Cristianitos fault. Of specific concern at the Licensing Board hearings was the relationship of the Cristianitos Zone of Deformation (CZD) to the OZD and the Cristianitos fault.
Following the lengthy hearings on the seismic issues, the NRC’s Licensing Board rejected petitioners’ challenge in a decision rendered in January 1982. Consistent with governing procedures,5 petitioners took an administrative appeal from the Licensing Board to the NRC’s Atomic Safety and Licensing Appeal Board (Appeal Board). Rejecting petitioners’ various challenges, the Appeal Board in March 1983 affirmed the Licensing Board’s determination.6 In June 1983, the NRC declined to review the Appeal Board’s decision, thereby upholding the grant of operating licenses for the two Units. This petition for review followed.
B
Before turning to petitioners’ challenges, we pause to examine briefly the regulatory framework within which this dispute has been addressed by the Commission.
The Atomic Energy Act (the Act), 42 U.S.C. §§ 2011, et seq. (1982), establishes a two-step procedure for the issuance of licenses to operate nuclear power plants. First, an applicant must seek from the NRC a construction permit to build the proposed facility. In such instances, section 189 of the Act requires that the Commission hold a public hearing before granting a construction permit. Id. § 2239(a). After obtaining a construction permit, an applicant must subsequently return to the Commission to secure an operating license. The Act requires the NRC to hold a second hearing at this stage only if requested by an interested person. Id.
Both construction permit and licensing hearings are conducted by an Atomic Safe*397ty and Licensing Board. Appeals from decisions of the Licensing Board are taken to the Atomic Safety and Licensing Appeal Board. Both the Licensing and Appeal Boards determine only whether issuance of a license is authorized. It is the Commission itself that actually issues the license. 10 C.F.R. §§ 2.760, 2.760a (1983). Before issuing a license, the Commission must reasonably assure itself that “the applicant will comply with the regulations ... and that the health and safety of the public will not be endangered.” Id. § 50.40(a). However, as we observe below, the NRC is given broad discretion to determine what technical specifications are necessary to protect the public health and safety.
In the licensing process, NRC regulations mandate administrative consideration' of the physical characteristics of proposed sites for nuclear facilities, including the locations’ geologic and seismic makeup. See 10 C.F.R. § 100.10(c) (1983). The governing portion of the regulations, found in 10 C.F.R. Part 100, Appendix A, “describes the nature of investigations required to obtain the geologic and seismic data necessary to determine site suitability and to provide reasonable assurance that a nuclear power plant can be constructed and operated at a proposed site without undue risk to the health and safety of the public.” Id. § 100.10(c)(1). The siting criteria and the investigations prescribed in Appendix A are, of necessity, broad and leave wide discretion to the Commission.7 The investigations mandated in the regulations are designed with the end of arriving at a “Safe Shutdown Earthquake” to be used as a basis for the seismic design specifications imposed in the construction of the plant.8 The regulations define “Safe Shutdown Earthquake” as:
that earthquake which is based upon an evaluation of the maximum earthquake potential considering the regional and local geology and seismology and specific characteristics of local subsurface material. It is that earthquake which produces the maximum vibratory ground motion for which certain structures, systems, and components are designed to remain functional.
Id. § 111(c). The regulations mandate that the procedures prescribed for arriving at the Safe Shutdown Earthquake “be applied in a conservative manner.” Id. § V(a)(l)(iv).9 Petitioners do not contend *398that these specific procedures and investigations were not performed; rather, they take issue’ with the substantive outcomes of those investigations.10 Unfortunately for them, it is in these substantive areas that this court’s deference to the agency justifiably reaches its zenith.
II
The issues presented in this case arise in the context of a regulatory scheme that this court has recognized is “virtually unique in the degree to which broad responsibility is reposed in the administrative agency, free of close prescription in its charter as to how it shall proceed in achieving the statutory objectives.” Siegel v. Atomic Energy Commission, 400 F.2d 778, 783 (D.C.Cir.1968). See also Westinghouse Electric Corp. v. Nuclear Regulatory Commission, 598 F.2d 759, 771 (3d Cir.1979). Section 182 of the Atomic Energy Act vests broad discretion in the NRC to establish qualifications for licensees of nuclear facilities:
In connection with applications for licenses to operate production or utilization facilities, the applicant shall state such technical specifications ... and such other information as the Commission may, by rule or regulation, deem necessary in order to enable it to find that the utilization or production of special nuclear material will be in accord with the common defense and security and will provide adequate protection to the health and safety of the public.
42 U.S.C. § 2232(a) (1982) (emphasis added). Recognition of the breadth of this delegation of authority to the Commission necessarily informs our examination of petitioners’ challenges.
Indeed, the Supreme Court has made it crystal clear that our review of NRC licensing decisions is “limited.” In reversing a prior decision of this court, the Court stated:
We are acutely aware that the extent to which this Nation should rely on nuclear power as a source of energy is an important and sensitive issue. Much of the debate focuses on whether development of nuclear generation facilities should proceed in the face of uncertainties about their long-term effects on the environment. Resolution of these fundamental policy questions lies, however, with Congress and the agencies to which Congress has delegated authority, as well as with state legislatures and, ultimately, the populace as a whole. Congress has assigned the courts only the limited, albeit important, task of reviewing agency action to determine whether the agency conformed with controlling statutes.
Baltimore Gas & Electric Co. v. Natural Resources Defense Council, 462 U.S. 87, 103 S.Ct. 2246, 2252, 76 L.Ed.2d 437 (1983). We conclude that in issuing the operating licenses at issue in this case the NRC acted well within its broad statutory authority. The voluminous record persuasively evidences the care with which the NRC discharged its statutory duties. Without hesitation, we find that its decision is supported by substantial evidence, and that its actions were not arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law. See 5 U.S.C. § 706 (1982).
Ill
Petitioners mount various arguments against the NRC’s determination that the seismic design of SONGS Units 2 and 3 provides a reasonable assurance of safety against earthquake hazards, and against the procedures employed by the Commission in arriving at that conclusion.11 First, petitioners challenge the Licensing Board’s *399decision “foreclosing” from consideration the issue of the “capability” of the Cristianitos fault, and the Appeal Board’s decision that the Licensing Board’s foreclosure of that issue constituted nonprejudicial error. Second, petitioners claim that the Licensing Board wrongfully redecided an issue foreclosed years before the construction permit stage as a result of a stipulation by the parties, namely the seismic nature of the OZD. Third, petitioners allege that the “Safe Shutdown Earthquake” magnitude and peak ground acceleration designations reached by the NRC are unsupported by substantial evidence. Fourth, petitioners contend that the Licensing Board committed prejudicial error when it admitted into evidence the 30-volume Final Safety Analysis Report, prepared by the utility-applicants, for the truth of the matters contained therein. Fifth, petitioners assert that the NRC’s licensing decision is rendered arbitrary and capricious by the Commission’s failure to.define with particularity its broad regulatory requirement that the specified seismic siting procedures be “applied in a conservative manner.” 10 C.F.R. Part 100 App. A § V(a)(l)(iv) (1983). Finally, petitioners press several points which, in their view, demonstrate the overall unfairness of the hearing afforded them and the inability of the NRC to decide the pertinent issues in an unbiased manner. We examine each of the arguments in turn.
A
During the hearing before the Licensing Board, petitioners sought to introduce evidence to establish that the Cristianitos fault, located within one mile of the Units, was “capable,” as defined in the NRC’s regulations, and should be considered as the controlling geologic structure for arriving at the “Safe Shutdown Earthquake” for the SONGS facility. After hearing petitioners’ evidence in its entirety, the Licensing Board struck the proffered evidence on the grounds that (1) it lacked' probative value because of “the witness’ sketchy qualifications as an expert, the superficiality and questionable accuracy of his pre-filed evidence, and his demeanor upon cross-examination,” and (2) petitioners were foreclosed from litigating the capability of the Cristianitos fault because that issue, while the subject of much investigation at the early stages of the proceedings, was not contested at the construction permit hearing. See Southern California Edison Co. (San Onofre Nuclear Generating Station, Units 2 and 3), 15 N.R.C. 61, 76-82 (1982) [hereinafter Licensing Board Decision], Joint Appendix at 548-51 [hereinafter J.A.]. The Licensing Board ruled that the Cristianitos-capability issue was foreclosed notwithstanding the fact that, since petitioners were neither parties nor in privity with any party to the construction permit proceedings, foreclosure was not warranted under judicial preclusion doctrines. Id. at 80, J.A. at 550.
On administrative appeal, the Appeal Board concluded that the Licensing Board had erred in determining that petitioners were foreclosed from litigating the capability vel non of the Cristianitos fault, but nonetheless found the error nonprejudicial. See Southern California Edison Co. (San Onofre Nuclear Generating Station, Units 2 and 3), 17 N.R.C. 346, 351, 355 (1983) [hereinafter Appeal Board Decision], J.A. at 639, 641. The Appeal Board, after fully reviewing the record, including the evidence stricken by the Licensing Board, based its conclusion on the Licensing Board’s independent ground that the evidence proffered by petitioners lacked probative value and on petitioners’ failure to make any offer of proof as to what more they could have presented but for the foreclosure ruling. See Southern California Edison Co. (San Onofre Nuclear Generating Station, Units 2 and 3), 15 N.R.C. 688, 697 (1982) [hereinafter Appeal Board Stay *400Decision], J.A. at 619.12 Petitioners now challenge the Appeal Board’s finding of nonprejudicial error; we find it more than adequately supported by the record.
Initially, we must note that it is misleading to refer to the Licensing Board’s ruling as a “foreclosure.” The Licensing Board in no wise precluded petitioners ab initio from presenting evidence concerning the capability of the Cristianitos fault. Rather, as we have seen, the Board received petitioners’ proffered evidence in full before ordering it stricken as lacking in probative value. The Board’s actions demonstrate a full recognition of the importance of a potential showing of the fault’s capability to the ultimate determination of the appropriate safeguards for public health and safety. See Licensing Board Decision at 78, J.A. at 549 (“If the Cristianitos were shown to be a capable fault, it would certainly be significant, and perhaps crucial, to the safety of the San Onofre facility”). In addition, the Licensing Board expressly indicated its willingness to reconsider its prior finding of incapability upon a showing of changed circumstances. Id. After receiving all of the petitioners’ evidence, the Board simply ruled that such a showing had not been made. Id.
It is, of course, beyond peradventure that the Licensing Board enjoys authority to exclude evidence it finds unreliable or otherwise lacking in probative value. See 5 U.S.C. § 556(d) (1982); 10 C.F.R. § 2.743(c) (1983). Since the Board is obviously best situated to assess the credibility and demeanor of witnesses, this court must defer to that judgment so long as it is reasonable. See generally Universal Camera Corp. v. National Labor Relations Board, 340 U.S. 474, 496-97, 71 S.Ct. 456, 468-469, 95 L.Ed. 456 (1951); Second Taxing District of Norwalk v. Federal Energy Regulatory Commission, 683 F.2d 477, 485 (D.C.Cir.1982); Penasquitos Village, Inc. v. National Labor Relations Board, 565 F.2d 1074, 1076-80 (9th Cir.1977). The reasonableness of the Licensing Board’s decision striking the testimony of petitioners’ expert witness, Richard Simons, is amply supported by the record. First, Mr. Simons’ specific credentials in this field were reasonably found wanting. Mr. Simons received a B.S. degree in geology and geophysics in 1959 from MIT, which was plainly an auspicious beginning. But he has since performed no graduate work and, more fundamentally, is not and does not purport to be a recognized expert in what petitioners themselves recognize as an extremely specialized science — seismology. Second, his methodology was deemed highly dubious, on grounds expressly set forth in detail in the Licensing Board’s decision.13 We cannot conclude that the Licensing Board erred in striking this testi*401mony as lacking in probative value, especially in light of the evidence already on the record relating to the capability of the Cristianitos fault.
The relationship between the Cristianitos fault and the location of the San Onofre nuclear facility has been thoroughly examined since the construction of the facility was first contemplated over twenty years ago. Throughout this long history, the Cristianitos fault has been considered inactive. See Southern California Edison Co., 2 A.E.C. 366, 376 (1964), J.A. at 427. Indeed, until petitioners entered the proceedings at the operating license stage, no party in the long and spirited debate over plant safety had apparently considered the capability of the Cristianitos fault even worth contesting. See Licensing Board Decision at 78, J.A. at 549. That is, frankly, not surprising. Uncontested evidence on the record indicates that the fault has been inactive for at least 125,000 years, which the Appeal Board found to be persuasive evidence of incapability. Appeal Board Stay Decision at 699, J.A. at 621. The NRC has carefully re-examined this determination whenever any new evidence was proffered to bring. the conclusion into doubt. See, e.g., Appeal Board Decision at 356, J.A. at 642 (noting that the evidence with respect to the 1975 earthquakes was inconsistent with any possibility that those earthquakes were related to the Cristianitos fault). And, consistent with its past policy, the NRC allowed petitioners fully to present their evidence before ruling on its admissibility.
Based upon the record as a.whole, we conclude that the NRC properly excluded petitioners’ evidence. Petitioners can demonstrate no harm resulting from the Licensing Board’s ruling “foreclosing” from further consideration the capability of the Cristianitos fault, particularly in light of their undisputed failure to make an offer of proof.
B
Petitioners next argue that the Licensing Board committed reversible error when it “redecided” the geological nature of the Offshore Zone of Deformation, the controlling geologic formation relevant to SONGS, in contravention of a stipulation entered into by the parties at the construction permit stage.14 Despite what the Licensing Board described as “an honest difference of opinion ... between the experts” as to the actual geological nature of the OZD, the parties to that earlier proceeding agreed prior to the construction permit hearing to have the adequacy of the proposed “design basis earthquake” adjudged in the framework of the geological *402model proposed by the United States Geological Survey (USGS). See Southern California Edison Co. (San Onofre Nuclear Generating Station, Units 2 and 3), 6 A.E.C. 929, 943 (1973) [hereinafter Construction Permit Decision], J.A. at 454. The USGS model described the OZD as “an extensive linear zone of deformation, at least 240 km long, ... [which] must be considered potentially active and capable of an earthquake whose magnitude could be commensurate with the length of the zone.” Id. at 942, J.A. at 453. The parties disagree over the proper interpretation of this description. Petitioners contend that the Licensing Board violated the stipulation when it ruled, based on the evidence presented to it, that “the OZD is a segmented, branching system of faults and folds and that the assumption of a rupture along its full length is speculative and unreasonably conservative.” Licensing Board Decision at 106, J.A. at 563. Respondents and intervenors, on the other hand, argue that this finding is fully consistent with the actual understanding of the parties a decade ago.
Petitioners’ reading of the description set forth in the USGS model as contemplating related faults capable of rupturing along the entire length of the zone is entirely understandable, particularly in light of the fact that petitioners are not and do not purport to be trained seismologists. The Licensing Board itself conceded that the USGS’ language was “ambiguous [and could] be read to describe the OZD as a single, ‘thoroughgoing’ fault.” Id. at 71, J.A. at 545. The most persuasive evidence to clarify the meaning of the USGS model, and consequently the stipulation of the parties, comes from USGS itself. That evidence supports entirely the Licensing Board’s interpretation. USGS witness James F. Devine testified at the licensing hearing:
[The USGS model] specifically avoided the term “fault zone.” We called it a zone of deformation because there are indeed segments which are not faulted, but instead deformed, folded, for example.
And so when attempting to describe then the earthquake potential one should assign to such a feature, we argued that the three discrete zones should not represent individual fault zones and earthquake magnitudes dependent on each of those individual segments, but instead should consider them all in one segment, for the purposes of estimating earthquake size.
Licensing Hearing Transcript’ (Tr.) at 5333, J.A. at 357. When asked whether the USGS model contemplated a fault “capable of rupturing at the same time in a single event,” Mr. Devine responded:
No. As I recall, none of us had the opinion or the position that the entire length could rupture at once, but only that there was indeed some relationship, probably at depth, of these three segments, such that it all should be considered one zone.
Id. (emphasis added).
In sum, the Appeal Board’s ruling that the Licensing Board did not violate the stipulation of the parties when it received evidence concerning the geological composition of the OZD is adequately supported by the record. Indeed, it seems curious, to say the least, to suggest that the utility-applicants would have stipulated to the USGS model if it were interpreted in the broad manner petitioners contend, since the net effect of such a stipulation would have been to foreclose almost all seismic issues from the licensing hearing and to doom the utilities’ licensing applications to certain failure.15
C
At the conclusion of its hearings, the Licensing Board ruled that the seismic *403design of SONGS Units 2 and 3 — based upon a “Safe Shutdown Earthquake” of a magnitude of 7 (on the Richter scale) and a maximum vibratory ground motion of .67g —was adequate to protect the public health and safety. This ruling was upheld by the Appeal Board, and was relied upon by the NRC itself in issuing the operating licenses for the San Onofre facilities. Petitioners challenge these findings on the ground that they are not supported by substantial evidence. Petitioners’ arguments focus on siích a limited portion of the evidence presented in this case that, even were we to accept their specific contentions, we would have no difficulty in concluding that the NRC’s findings are supported by substantial evidence. Petitioners point to the testimony of only 3 of the 28 witnesses who testified before the Licensing Board during 25 days of hearings. They emphasize only a miniscule fraction of the 7,000 pages of hearing transcript. More fundamentally, however, they ignore the Appeal Board’s express finding that even excluding the testimony assailed by petitioners, “[t]he choice of a Ms 7.0 safe shutdown earthquake for San Onofre is amply supported by other expert testimony in the record,” particularly the testimony of Drs. Smith and Ehlig. Appeal Board Decision at 709 n. 40, J.A. at 631 (emphasis added).
First, petitioners attack the testimony of Dr. David Slemmons, a witness for the NRC staff, with respect to the appropriate magnitude for the “Safe Shutdown Earthquake.” Petitioners contend that the magnitude arrived at by Dr. Slemmons through his study is a mean, or average, value and therefore cannot be considered “conservative” as a matter of law. They vigorously contend that the standard should in fact be the mean value plus one standard deviation. Application of the standard advanced by petitioners would result in a “Safe Shutdown Earthquake” of Ms 7.8. See Pet. Brief at 44.
Acceptance of a mean value as an appropriately “conservative” measure of adequate safeguards for the public health and safety in this critical area seems, at first blush, to be lacking in stringency. But upon analysis, we conclude that the NRC’s reliance on the Slemmons study was entirely reasonable and that its ultimate findings were supported by substantial evidence. The Commission has adequately explained the “conservative” assumptions inherent in Dr. Slemmons’ model, see Licensing Board Decision at 121-22, J.A. at 570-71, and, as noted above, his findings are corroborated by other expert testimony. Specifically, the NRC found that the underlying data base utilized by Dr. Slemmons was “already overly conservative”:
Dr. Slemmons used only the largest percentage rupture reported for each fault to obtain the average rupture length, which also adds conservatism to his estimate.
Id. (emphasis added).16
Petitioners’ specific claim that Dr. Slemmons repudiated the methodology of his study upon cross-examination and adopted petitioners’ mathematical expression of “conservatism” is utterly without merit. An examination of Dr. Slemmons’ entire testimony reveals at least three occasions when the witness expressly rejected the measure of “conservatism” proffered by petitioners’ counsel. See Tr. at 6230-31, J.A. at 375-76; Tr. at 6265-69, J.A. at 381-85; Tr. at 6272, J.A. at 388.
Second, petitioners challenge the Licensing Board’s discounting of the téstimony of two other expert witnesses, Drs. Boore and Luco. In each case the Board simply concluded, as was its province to do, see Universal Camera Corp. v. National Labor *404Relations Board, supra, 340 U.S. at 495-97, 71 S.Ct. at 468-469; Penasquitos Village, Inc. v. National Labor Relations Board, supra, 565 F.2d at 1076-80, that their testimony was not as credible as other testimony or was outweighed by other evidence. Petitioners cannot be heard to challenge the Commission’s findings as unsupported by substantial evidence merely by pointing to the evidence against the NRC’s ruling, while ignoring the substantial evidence in its support.
Petitioners’ Reply Brief indirectly addresses this defect in their argument.' See Pet. Reply Brief at 21-26. The gist of petitioners’ argument in this respect appears to be that the uncertainty of the science of seismology, coupled with the regulatory requirement that the inquiries be conducted in a “conservative manner,” counsels the judiciary to require the NRC to accept the worst prediction of earthquake potential proffered to it, régardless of its perceptions of the relative credibility of the testimony.
In advancing this argument, petitioners fundamentally misperceive the judiciary’s role in complex regulatory matters. The uncertainty of the science of earthquake prediction only serves to emphasize the limitations of judicial review and the need for greater deference to policymaking entities. This is no startling or novel development. Just last year the Supreme Court in an analogous context observed that “[a] reviewing court must remember that the Commission is making predictions, within its area of special expertise, at the frontiers of science. When examining this kind of scientific determination, as opposed to simple findings of fact, a reviewing court must generally be at its most deferential.” Baltimore Gas & Electric Co. v. Natural Resources Defense Council, supra, 462 U.S. at 103, 103 S.Ct. at 2256.17
In addition, the courts have long accepted the Commission’s definition of its statutory mandate to “provide adequate protection to the health and safety of the public” as requiring not a risk-free environment, but a “reasonable assurance ... that the reactor could be safely operated at the proposed location.” Power Reactor Development Co. v. International Union of Electrical, Radio & Machine Workers, 367 U.S. 396, 414, 81 S.Ct. 1529, 1538, 6 L.Ed.2d 924 (1961). See also North Anna Environmental Coalition v. Nuclear Regulatory Commission, 533 F.2d 655, 665-67 (D.C.Cir.1976); Nader v. Nuclear Regulatory Commission, 513 F.2d 1045, 1052 (D.C.Cir.1975). These cases are fundamentally at odds with petitioners’ argument that uncertainty, coupled with the mandate of “conservatism,” requires the NRC to adopt wholesale the worst case scenario that a party may gloomily frame. Indeed, petitioners’ arguments when extended to their logical conclusion would suggest that, as a matter of law, no conservatively acceptable level of earthquake risk exists. See Reply Brief at 24-25 (“It is clear that, with the burden of proof to show that the design basis earthquake ground motion could not be exceeded, the burden could not be scientifically held, given our present lack of knowledge”) (quoting written testimony of Dr. James Bruñe). After examining the entire record, both the evidence for and against the Commission’s findings, we find those findings amply supported by substantial evidence.
D
Petitioners also argue that the Licensing Board committed prejudicial error *405when it admitted into evidence the multivolume Final Safety Analysis Report (FSAR) prepared by Southern California Edison with respect to the SONGS Units. NRC regulations require applicants for operating licenses to prepare a FSAR. See 10 C.F.R. § 50.34(b) (1983). Under Commission precedent, the FSAR may be admitted into evidence at the licensing hearing in order to prove compliance with this regulatory requirement. See Arkansas Power & Light Co. (Arkansas Nuclear One Unit 2), 6 A.E.C. 25, 32 (1973). However, in this case, the Licensing Board admitted the FSAR into evidence for the truth of all the matters asserted therein. Petitioners argued before the Appeals Board that the FSAR, as unauthenticated hearsay, was improperly admitted and that petitioners were prejudiced in their ability to conduct cross-examination because of the applicants’ failure to identify the authors of the specific sections of the report. The Appeal Board agreed with petitioners, finding that the Licensing Board had erred in admitting the FSAR for any purpose save that of demonstrating compliance with the regulatory scheme. Appeal Board Stay Decision at 366-68, J.A. at 647-48. However, the Appeal Board ruled that this evidentiary error was harmless inasmuch as only limited reliance had been placed by the Board on the FSAR. See id. at 365 n. 31, J.A. 646 (“the Licensing Board mentions the FSAR only twice in its seismic decision and cites independent authority for the same conclusions”).
Petitioners challenge this finding of nonprejudicial error. Specifically, petitioners argue that, since it is impossible to determine the actual extent to which the Board relied on the FSAR, prejudice should be presumed because “[t]he absence [thereof] does not affirmatively appear.” Pet. Brief at 57. Petitioners’ description of the burdens of proof is, however, quite in error; the onus is squarely upon them to establish prejudice. In Braniff Airways, Inc. v. Civil Aeronautics Board, 379 F.2d 453 (D.C.Cir.1967), this court noted:
A court will not reject an agency finding that is supported by substantial evidence merely because the agency also incidentally mentions incompetent or irrelevant material. Thus, in dealing with an agency’s “official notice,” in an incidental or verifying reference, of matter outside the record, the courts do not “make a fetish of sticking squarely within the four corners of the specific record in administrative proceedings ... unless substantial prejudice is shown to result.”
Id. at 466 (quoting United States v. Pierce Auto Freight Lines, Inc., 327 U.S. 515, 530, 66 S.Ct. 687, 695, 90 L.Ed. 821 (1946)). See also Consolidated Gas Supply Corp. v. Federal Energy Regulatory Commission, 606 F.2d 323, 328-29 (D.C.1979), cert. denied, 444 U.S. 1073, 100 S.Ct. 1018, 62 L.Ed.2d 755 (1980); National Labor Relations Board v. Seine & Line Fishermen’s Union of San Pedro, 374 F.2d 974, 981 (9th Cir.), cert. denied, 389 U.S. 913, 88 S.Ct. 239, 19 L.Ed.2d 261 (1967). As noted above, the NRC’s findings are amply supported by substantial evidence, other than the FSAR. The Appeal Board properly ruled that the Licensing Board’s erroneous admission of the FSAR constituted harmless error.
E
Equally unpersuasive is petitioners’ penultimate argument — that the NRC’s failure to provide a quantifiable definition for the key regulatory phrase, “conservative manner,” see 10 C.F.R. Part 100, App. A § V(a)(l)(iv) (1983), renders all of its licensing decisions arbitrary and capricious. We find the Commission’s approach entirely reasonable, notwithstanding the absence of a specific definition of the word “conservative,” especially in light of the broad policymaking authority conferred by Congress on the agency. Many of the investigations prescribed in the regulations, and most of the infinite number of judgment calls required in conducting those investigations, are not mathematically expressa*406ble and, therefore, do not lend themselves to rigid statistical definitions. Certainly, the quantifiable “mean plus one standard deviation” definition propounded by petitioners is not the only reasonable option the Commission could arrive at in exercising its delegated authority.
More fundamentally, the Commission has adequately enumerated the “conservative” assumptions inherent in the methodologies of the studies upon which it relied. Its proceedings have been exhaustive, and the record generated thereby contains substantial evidence in support of each of the NRC’s conclusions. “We see no reason why we should not accord to the Commission’s interpretation of its own regulation and governing statute that respect which is customarily given to a practical administrative construction of a disputed provision.” Power Reactor Development Co. v. International Union of Electrical, Radio & Machine Workers, 367 U.S. 396, 408, 81 S.Ct. 1529, 1535, 6 L.Ed.2d 924 (1961).
F
Petitioners’ final area of argument consists of several contentions aimed at setting aside the NRC’s licensing decision because of alleged unfairness infecting the entire hearing process. A dispassionate review of the record simply lends no support whatever to petitioners’ claims; rather, the record persuasively shows that petitioners were afforded a full and fair opportunity to participate before the Commission. It cannot be said that “the entire hearing was permeated by ... pervasive impatience — if not hostility — ... which made fair and impartial consideration impossible.” Office of Communication of the United Church of Christ v. Federal Communications Commission, 425 F.2d 543, 548 (D.C.Cir.1969). This is simply not a case such as United Church of Christ where the agency’s “entire handling of [the] case,” id. at 550, is brought into doubt because of indications of agency bias and prejudicial error. The NRC has not taken action inconsistent with its prior findings. Id. at 545-46. It did not erroneously allocate the burden of proof. Id. at 548-49. Indeed, all the actions taken here are well within the Commission’s authority to control its own proceedings. See 10 C.F.R. § 2.718 (1983).
In their opening broadside against the Commission's procedures followed here, petitioners vigorously contend that, by virtue of an NRC general directive to expedite licensing proceedings, petitioners were not given adequate time to prepare their case, thereby depriving them of a fair hearing. Specifically, petitioners contend that the Licensing Board imposed “a truly impossible schedule, thereby making it as difficult as possible for petitioners to present such a truly massive case in a manner best suited to protect the public health and safety.” Petitioners’ Brief at 16.
This claim, upon analysis, must fail. Dispassionately considered, the record utterly fails to support as a factual matter the contention that the schedule was “impossible.” This case was not rushed to hearing in 1981 out of the blue. To the contrary, petitioners were permitted to intervene in 1977, and their contention with respect to the adequacy of the SONGS Units’ seismic design was formally established in January 1978, over three years before the hearings began. The Licensing Board order directing that discovery be completed by February 1981 was entered in August 1980. It was almost a year after entry of that order before the hearings began in June 1981. This schedule, notwithstanding the preparation demands imposed on all parties in the spring of 1981, is not asserted to violate any statutory or regulatory requirement.
Petitioners do mount a more specific argument that the NRC’s regulations were violated, independent of the Commission’s general direction to expedite licensing proceedings. They contend that they were wrongfully required to participate in a prehearing conference prior to the completion of discovery, allegedly in violation of 10 *407C.F.R. § 2.752.18 This argument is patently frivolous. Petitioners omit from their citation of this regulation the critical language “or such other time as the Commission or the presiding officer may specify.” 10 C.F.R. § 2.752(a) (1983). The NRC reasonably expedited these proceedings, consistent with the public interest, without unduly implicating the demands of fairness. The Commission’s deliberations over SONGS Units 2 and 3 spanned nearly ten years. Petitioners were permitted to intervene in the agency’s proceedings over 3V2 years before the licensing hearing was held. The relevant seismic design issues were well-known throughout that period. The long history of the NRC’s proceedings in this case evidences deliberate decision-making, not hurried rubber-stamping of prior tentative decisions. That petitioners may have been understaffed in comparison to the applicants, and thereby have sustained a comparative disadvantage, cannot and does not affect our examination of the reasonableness of the Commission’s actions. We are quite unable to conclude that the timing of the proceedings deprived petitioners of a fair hearing or rendered the Commission’s actions arbitrary and capricious.
Nor can we conclude that the Commission committed reversible error when it required petitioners to examine one of their witnesses a day earlier than they had planned. Indeed, the witness was put on the stand during the very afternoon preceding. the day on which the petitioners had planned. The suggestion that the acceleration of a single witness’ appearance in a long and protracted hearing evinces agency bias and hostility is simply too extravagant a claim to be seriously maintained.
Finally, petitioners make the imaginative, but ludicrous argument that the NRC is biased as a matter of law and must accordingly be disqualified from presiding over the proceedings relating to the issuance of an operating license because of the Commission’s alleged potential liability to the nuclear applicants were it to reverse its prior seismic design decision. Even assuming arguendo the extraordinary proposition that the NRC could be held liable,19 petitioners’ disqualification suggestion is specious. The Atomic Energy Act, as we have seen, establishes a two-step licensing process — the construction permit phase and the operating license phase. Under the logic of petitioners’ analysis, the NRC would always be disqualified from rendering decisions in the licensing phase because of prior decisions in the construction phase. Such a result is fundamentally at odds with the statute and common sense. Congress has not seen fit to designate an alternative decisionmaker to conduct licensing proceedings once the NRC has determined to issue a construction permit. That decision must be made by Congress’ delegatee — the NRC. It cannot be made by the courts.
*408Moreover, the premise of petitioners’ argument — that recipients of construction permits are entitled to rely on NRC findings at the construction permit hearing— has expressly been rejected by the Supreme Court. See Power Reactor Development Co. v. International Union of Electrical, Radio & Machine Workers, supra, 367 U.S. at 415, 81 S.Ct. at 1538 (applicant on notice “that it proceeds with construction at its' own risk, and that all its funds may go for naught”). The regulations do provide a mechanism whereby an applicant may obtain Commission approval of the safety of any particular design feature at the construction permit stage, see 10 C.F.R. § 50.35(b) (1983); however, the applicants in this ease sought no such approval.
An examination of the entire record in this case demonstrates that petitioners were given a full and fair opportunity to present their arguments before the Commission, and that the Commission has fully complied with the procedures mandated in the applicable statutes and regulations. The orders of the Commission are therefore
Affirmed.
. Construction permit proceedings with respect to SONGS Unit I began in the early 1960’s. During those proceedings, the issue of the facility's location in an area of seismic activity was also raised and adjudicated by the AEC.
. Petitioners note that the interrogatories were resisted by the NRC staff and the applicants, necessitating motions to compel which were not resolved by the Board until very shortly before the hearings began in 1981. Intervenors vigorously protest that discovery deadlines imposed upon them were "onerous" and resulted in "great cost to their ability to prepare the written testimony of volunteer witnesses and ... great cost to their being able to prepare for cross-examination.” Petitioners Brief, at 8-9. This asserted burden constitutes one factor pointed to *396by petitioners in their argument that the NRC throughout this proceeding evidenced hostility and bias toward these intervenors, a contention we address below.
. The NRC represents that the Cristianitos fault is "about one mile" from the SONGS facility at the point of its closest approach, whereas the applicants-intervenors state that the fault is situated "slightly over one-half mile from the site.” Petitioners represent that the fault is "within one-half mile of the plant.”
. Petitioners represent that the two earthquakes were of the magnitude 3.3 and 3.8 bn the Richter scale. "The preliminary locations were near the central portion of the Cristianitos fault.” Petitioners Brief at 19.
. A summary of NRC licensing procedures with respect to nuclear power plants is set forth in Part B, which follows immediately below.
. The Appeal Board did determine, however, that the Licensing Board had committed non-reversible errors in several respects, as we will discuss more fully infra in treating petitioners’ contentions.
. The regulations note that "[t]he criteria are based on the limited geophysical and geological information available to date concerning faults and earthquake occurrence and effect. They will be revised as necessary when more complete information becomes available." 10 C.F.R. Part 100, App. A § I (1983).
. Among other things, the regulations require that "[f]or faults, any part of which is within 200 miles of the site and which may be of significance in establishing the Safe Shutdown Earthquake, [a] determination [be made] whether these faults are to be considered as capable faults." Id. § IV(a)(7). According to the regulations:
A "capable fault” is a fault which has exhibited one or more of the following characteristics:
(1) Movement at or near the ground surface at least once within the past 35,000 years or movement of a recurring nature within the past 500,000 years.
(2) Marco-seismicity instrumentally determined with records of sufficient precision to demonstrate a direct relationship with the fault.
(3)A structural relationship to a capable fault according to characteristics (1) or (2) of this paragraph such that movement on one could be reasonably expected to be accompanied by movement on the other.
Id. § 111(g). Each capable fault is then examined to determine whether it should be used as the controlling geological formation for arriving at the Safe Shutdown Earthquake.
.Although, as petitioners note, no specific definition of the phrase "applied in a conservative manner” is provided, the regulations do mandate the application of several conservative assumptions. For example,
[t]he magnitude or intensity of earthquakes based on geological evidence may be larger than that of the maximum earthquakes historically recorded. The accelerations at the site shall be determined assuming that the epicenters of the earthquakes of greatest magnitude or the locations of highest intensity related to the tectonic structures are situated at the point on the structures closest to the site.
Id. § V(a)(l)(i).
. As we will see in Part III, below, petitioners do mount various procedural attacks against the Commission's decision here, but those contentions go principally to the manner in which the Licensing Board hearings were conducted.
. This appeal is limited to the Commission’s seismic design findings. Challenges filed by other petitioners to the emergency preparedness of SONGS Units 2 and 3 are pending before the *399court in another case transferred from the Ninth Circuit. GUARD v. N.R.C., No. 84-1091.
. At oral argument before the Appeal Board, counsel for petitioners informed the Board that the only effect of the Licensing Board’s foreclosure ruling on the record available for their review was the absence of cross examination of two staff witnesses — Dr. Reiter and Mr. Car-done. The Appeal Board directly addressed the potential impact of this lack of cross-examination in finding the error harmless:
We have reviewed the record material .(including that which was formally stricken) and do not find the gap in cross-examination prejudicial. [Petitioners] did in fact cross-examine Mr. Cardone and Dr. Reiter as to post-1973 evidence dealing with the potential capability of the Cristianitos fault. What they were precluded from pursuing by virtue of the Licensing Board’s foreclosure ruling was pre-1973 information bearing on the fault’s capability. But as to that, [petitioners] had had virtually no questions to ask when cross-examining Dr. Biehler, the applicant’s consultant, whose testimony covered the Cristianitos fault in its full historical range. And intervenors do not quarrel with the scope of their cross-examination of Dr. Biehler.
Appeal Board Stay Decision at 697, J.A. at 619 (references to hearing transcript omitted). Petitioners’ argument that deprivation of cross-examination constitutes per se reversible error is without merit. See United States v. Gambler, 662 F.2d 834 (D.C.Cir.1981) (harmless error rule applied in criminal case); Wasson v. SEC, 558 F.2d 879, 884 (8th Cir.1977).
. A comparison of the methodologies employed by Mr. Simons and Dr. Biehler, a scientist hired by the applicants and relied upon by the Licensing Board, provides a telling indictment of the relative probative value of Mr. Si*401mons’ testimony. The Licensing Board describes Mr. Simons’ methodology as follows:
What [Simons] did, essentially, was transfer earthquake location data covering the period 1932 to 1980 from the epicenter catalogue published by the California Institute of Technology to a map of the vicinity of the Cristianitos fault. He then drew error circles of different sizes around the estimated epicenters, the size depending upon the presumed accuracy of the location.
Licensing Board at 77, J.A. at 548. In contrast, the Appeal Board described the studies conducted by Dr. Biehler in order to determine whether two small earthquakes that occurred in January of 1975 were related to the Cristianitos fault as follows:
The applicant's investigations included a geomorphic study, an evaluation of microseismic events, a study of focal mechanisms, the construction of a subsurface contour map, an updating of historic seismicity, and geophysical surveys____ Through calibration blasts, Dr. Biehler developed a model to locate more accurately the epicenters of the small earthquakes and to fix limits on their hypocentral depths. The difference in faulting style and spatial separation from the Cristianitos fault led him to conclude that the events could not be associated with that fault.
Appeal Board Stay Decision at 699-700, J.A. at 621-22. It is little wonder that this comparison caused the Licensing Board to “question whether any useful conclusions could be drawn about the seismicity of the Cristianitos fault itself from this circle drawing exercise." Licensing Board Decision at 77, J.A. at 548.
. It will be recalled that petitioners, of course, did not participate at the construction permit stage.
. As petitioners’ brief explains, a fault rupture of 240 km, the entire length of the OZD, would be occasioned by an earthquake of a magnitude of Ms 7.87. See Pet. Brief at 38 (quoting testimony of Dr. James Bruñe). That the Licensing Board would issue a construction permit on a plant designed to withstand an earthquake of just over l/10th the magnitude of the earth*403quake hypothesized under the model allegedly stipulated to by the parties defies reason.
. The Commission further noted that Dr. Slemmons could have lowered the mean value of averaging the percentage rupture of only those faults of less than 400 km length. He did not, however, do so, thereby further increasing the "conservatism” of the data base.
. See also FPC v. Florida Power & Light Co., 404 U.S. 453, 463, 92 S.Ct. 637, 644, 30 L.Ed.2d 600 .(1972) ("A court must be reluctant to reverse results supported by such a weight of considered and carefully articulated expert opinion. Particularly when we consider a purely factual question within the area of compefence of an administrative agency created by Congress, and when resolution of that question depends on 'engineering and scientific' considerations, we recognize the relevant agency's technical expertise and experience, and defer to its analysis unless it is without substantial basis in fact.").
. The relevant portion of § 2.752 reads:
The Commission or the presiding officer may, and in the case of a proceeding on an application for a construction permit or an operating license for a facility of a type described in §§ 50.21(b) or 50.22 of this chapter or a testing facility, shall direct the parties or their counsel to appear at a specified time and place for a [prehearing] conference____ A prehearing conference held under this section in a proceeding involving a construction permit of operating license shall be held within sixty (60) days after discovery has been completed, or such other time as the Commission or the presiding officer may specify.
10 C.F.R. § 2.752(a) (1983).
. The district court case cited by petitioners in support of their contention that the NRC may be held liable, General Public Utilities Corp. v. United States, 551 F.Supp. 521 (E.D.Pa.1982), is — assuming its holding to be correct — inapposite. That case involved the duty of the NRC to "monitor nuclear experiences and to disseminate appropriate warnings,” id. at 526, not any potential liability for licensing decisions themselves or the discretionary judgments made therein. The respective duties are fundamentally dissimilar. The explicit duty to warn discussed in General Public Utilities clearly inures to the benefit of the utility, while the Commission’s duty to make the seismic design decisions involved in this case is directed to the public in general. The issuance of warnings involves little if any exercise of discretion, while the decision to issue a license involves the exercise of judgment in the midst of uncertainty.