Oral History Transcript — Dr. Robert Frosch
This transcript may not be quoted, reproduced or redistributed in whole or in part by any means except with the written permission of the American Institute of Physics.
This transcript is based on a tape-recorded interview deposited at the Center for History of Physics of the American Institute of Physics. The AIP's interviews have generally been transcribed from tape, edited by the interviewer for clarity, and then further edited by the interviewee. If this interview is important to you, you should consult earlier versions of the transcript or listen to the original tape. For many interviews, the AIP retains substantial files with further information about the interviewee and the interview itself. Please contact us for information about accessing these materials.
Please bear in mind that: 1) This material is a transcript of the spoken word rather than a literary product; 2) An interview must be read with the awareness that different people's memories about an event will often differ, and that memories can change with time for many reasons including subsequent experiences, interactions with others, and one's feelings about an event. Disclaimer: This transcript was scanned from a typescript, introducing occasional spelling errors. The original typescript is available.
See the catalog record for this interview and search for other interviews in our collection
Interview with Dr. Robert Frosch
Robert Frosch; May 28, 1998
ABSTRACT: The interview focuses on Frosch’s involvement in issues related to the seismic detection of underground nuclear weapons tests during the 1960s. He describes his time as director of the Advanced Research Projects Agency’s Nuclear Test Detection Office from 1963 to 1965. In this position Frosch helped to manage the Department of Defense’s “Project Vela Uniform,” which aimed at the improvement of seismic detection capabilities. He played a major role in the realization of the Large Aperture Seismic Array (LASA).
Frosch’s background before he came to ARPA:
• Frosch came to Hudson Labs in fall 1951, just finishing his PhD which had nothing to do with Hudson Labs (he needed a job); Frosch came as theoretician, worked on classified topics, such as long range underwater acoustics for submarine detection
• started to do theory, then involved in experiments and equipment and Frosch went to sea; Frosch became combination geophysicist, oceanographer, and physicist
• Frosch became director of the lab in 1956; about 100 people;
• involved in ARTEMIS, submarine detection, very big engineering projects, hundreds of hydrophones, theoretical work on stability of sound in the ocean and long-range detection; stability of acoustic paths in the ocean over very long ranges, the ocean stability temperature measurements; probing the oceans with acoustics
• out of blue sky Frosch got call from Jack Ruina sometime in spring of 1963, talking about LTBT [Limited Test Ban Treaty 1 (was just signed). Ruina asked whether Frosch would come to Washington, DC to do the R&D for detection system. Frosch had not heard before of Jack Ruina or ARPA; Frosch guesses that Harvey Brooks brought up Frosch’s name
• Frosch checked out the offer, expecting not to take it, but Ruina talked him into it. In August 1963 Frosch moved to DC; Ruina left and Frosch didn’t know Robert Sproull or Charles Hertzfeld (new ARPA director and deputy ARPA director then) either; 15 Sept. 1963: Frosch started in Pentagon
• when Frosch started as director of NTDO [ARPA’s Nuclear Test Detection Office] he knew nothing about seismic detection; but he knew detection of underwater blasts (2.5 pound block of TNT was standard tool, and he had thrown a lot of TNT); underwater detection of lkt trivial; furthermore, acoustics and seismometry were kind of the same general subject, though different frequencies and shear waves play a role in seismology, but not in the ocean; but the whole business of detection was similar, and the noise problem was analogous; using some arrays in submarine detection: hydrophones hooked up with phone lines; with submarine detection ahead of the engineering art: problems of signal processing; some unclassified literature on this stuff, by Fred Fisher of Scripps
• Harry Sonnemann built ARTEMIS, the receiving array: Frosch has high opinion of Sonnnemann
• Frosch came out of group of people who thought about coherent arrays; Frosch came in thinking about signal detection, signal-to-noise etc.; how do you know it’s a submarine and not a surface ship or three wales: discrimination problem
• Frosch never had a job interview [for the ARPA position], Ruina made sure he wanted to recruit Frosch, and they talked for an hour
• Frosch understood that underwater tests were no problem: simply use existing Navy system; he never did anything with infrasonic air acoustics, that was all in hand -> sampling clouds; two problems left: Vela satellites (when Frosch came the first satellites were already designed), and big unsolved problem was Vela Uniform: how small a seismic event can you detect? Would decoupling work and how would this interact with detection capabilities? What’s the right detection system given that? And finally, discrimination between explosion and earthquake
• how much did Frosch know about decoupling before he came to ARPA? He followed that in the New York Times as “oh, that’s interesting”; Frosch was active in the Acoustical Society and Society of Exploration Geophysicists, and probably in American Geophysical Union AGU, but he did not read about seismology before his time at ARPA; Frosch knew people at Lamont: [Joe] Worzel, [Maurice] Ewing, [Frank] Press; knew Press from graduate school; Frosch talked to Bruce Heezen and people who did the early stuff in continental drift: ocean-bottom mapping: rift valleys etc.; Frosch knew in a general way about earthquake detection, but he hadn’t followed it closely
• Who was NTDO director before Frosch? Frosch not sure, but apparently German sounding name; guy who was deputy NTDO and left because he was not made director of NTDO (not Canton Beyer, not Rathjens); Frosch remembers Ted George; Frosch says that [Charles C.] Bates tried to convince Frosch that Bates wants to write a biography of F
• Frosch became deputy director of ARPA the same day Charly Hertzfeld became director of ARPA; when Bob Sproull left, Frosch moved offices; Steve Lukasik took over NTDO after Frosch, but Frosch doesn’t know who came after Lukasik; Frosch served in ARPA until the end of June 1966, when he moved to become Assistant Secretary of the Navy R&D; Peter Franken came in after Frosch as [ARPAJ deputy (Franken and Frosch knew each other from college at Columbia college, also together in graduate school);
• Frosch’s early years at ARPA: how was interaction with Sproull and Bates, how often would he see them, how much influence on contracting, on Vela Uniform in general? Frosch: we ran it every day; ARPA was a small place, doing things in a hurry if it needed to; two guys in procurement prided themselves with if a well thought of guy called up in the morning with a good idea you could have an authorization for that guy to spend money the same day; you couldn’t do this ten times a day, but occasionally. ARPA was fast; up to about $500,000 or so Frosch could authorize it; for bigger sums he would go to Hertzfeld or Sproull; just stick your head in the office, informal; personal secretaries in Pentagon very protective to the point that he couldn’t get anything done; Frosch asked what should he read, he became member of SSA and started looking at the journals and reading all the classified and unclassified material, finding out about Latter brothers and Mississippi [nuclear test] shot (Shoal?); doing some traveling around: checking out seismometer vault, wanted to see Wiechert [seismograph], went to Ottawa with Sproull: both quite amazed how far behind the seismic instruments were; Frosch: “My god, this is ridiculous, I mean this thing is scribing little hen tracks on a slight that had been smoked with a candle?! What year are we in!?” Frosch felt he had gone back in time; Geotech [The Geotechnical Corporation] build more modem instruments though, amplifiers, modifying some instruments from the oil industry; but generally the feeling was that there is an under supported piece of science; and practitioners too cautious: arrays are crazy, electronics not better than mechanics etc.; Frank Press got involved, he knew Frosch; Frosch: “[Press] and I conspired to drag seismology into the 20th century.” Charly Bates had started at least WWSSN. The only person who had thought about arrays was Hal Thirlaway in UK: but he was focusing on teleseismic P waves and it was natural for Frosch, he says, to build filters; Carl Romney, says Frosch, is very good scientist, honest guy, but kind of conservative: he didn’t believe in arrays; Frosch: Romney had right to be a little defensive, because suddenly all these outsiders were telling him what to do; Frosch: Romney and he had a kind of pleasant battle over the years over arrays;
• Bates and Frosch did a little census and concluded that they supported every seismologist in the US but two (two priests at Fordham University, who didn’t believe in taking that kind of DOD money; [when I mentioned the name “Lynch”, that sounded right to him] .
• theoretical seismology: on the geophysical wave side very sophisticated, but from the point of view of signal processing very primitive; oil people much more sophisticated then, but not much communication between these two, because of different frequency range
• there were few small arrays [Wichita Mountain Seismological Observatory etc. but Frosch asked then, why are they so short? Why are you building fraction of a wavelength arrays? Then discussion about coherence, because of the complicated geology: Frosch replied that the rate of the change of geology very slow compared to the waves, but it can’t be incoherent, where could the scattering come from?
• Frosch was the outsider; both Frosch and Bates thought let’s try large array; Frosch says that [Armed] Services too clumsy bureaucratically, that’s why ARPA got the job; AFTAC [Air Force Technical Applications Center] very upset that they wouldn’t get it; tension between ARPA and AFTAC: uneasy alliance; some AFTAC people felt that they had done a good job, and now suddenly they felt that they were told that they hadn’t done a good job; Romney: professional and friendly, but clearly disapproving; and AFTAC was intelligence agency with focus on Soviet testing: mindset that everything is classified; Frosch says that even today he doesn’t know everything AF1’AC was doing during that time, although Frosch had clearance;
• question: did Frosch know about secret AFTAC operation and data? Frosch doesn’t recall. Remembers that the arrays were very short; so what was Romney’s argument against LASA [Large Aperture Seismic Array]? Frosch: Romney’s argument was, a) Frosch didn’t understand the seismology and the variability, and it would not really going to work as an array, that we spend all the money for a large system that doesn’t give any significant array gain; b) Frosch talked to Press and [Hal] Thirlaway and other people to build a P wave array and Carl [Romney] was a strong influence on making sure that ARPA would build a 3-component shear wave array as well; Frosch would have gone for the simpler, cheaper version with just short-period [SP] seismometers for P waves, but then convinced to put in good shear wave instruments; Carl Romney had a strong influence: he thought it was not worth it, but if ARPA builds it anyways, then try a good shot at it
• Romney part of LASA design team [Romney and Frosch published a paper together in the IEEE proceedings]; Paul Green was part of design team: signal processing and electronics; nowadays probably against procurement rules that design team was partly government people and partly contractors and party anybody ARPA thought was any good and then just sit down and do the job; ARPA had all sorts of exceptions: we didn’t have to go on open bids for design etc.; conflict of interest? Frosch: it it’s better to know the people, knowing where they came from
• LASA design: ideas to run it from central location, no paper output, all electronically; doesn’t remember how much design had to be done for seismometers: Frosch thinks that Geotech had stuff; sending data through wires was done by AFTAC earlier; 200 vaults hooked together was a wild though, and nobody knew whether computers could do the job
• ARPA made proposal, about 10 pages, no ARPA Order, internal “This is what I want to do” report; Frosch can’t remember problems in getting the $8M authorized; with this plan he walked to Sproull, Bates would cook over it with the procurement guys, and he might have presented it to Harold Brown, DDRE (Harold was just ahead in graduate school from Frosch in Columbia); Frosch doesn’t recall that he briefed McNamara on LASA; Frosch briefed McNamara once: test ban treaty stuff; for that he received advice: “Don’t make any sounds that sounded like you need a decision or you may get one whether you want it or not.”
• Frosch remembers briefing the Joint Chiefs on the whole disarmament business, everything that ARPA did on nuclear test detection; Air Force general with cigar was the tough one [name??]: he had pointed questions; also quite a lot of contact with CIA guy [Scoville?], and George Rathjens (International Security Affairs, ISA, or ACDA)
• WWSSN [Worldwide Standard Seismograph Network]: question: why do we find deletions in congressional record when WWSSN is discussed? Wasn’t WWSSN only for unclassified stuff? Frosch: recollection and reconstruction [so to take with caution]: two motives for WWSSN: trying to build up seismology and seismological knowledge, because that was needed for detection job; but at the same time WWSSN could be used for detection purposes; but for diplomatic reasons the nuclear test detection aspects was not publicly emphasized; Congress not wildly enthusiastic about dragging seismology into the 20th century as such: they are interested in what comes out of WWSSN; what was classified was what we are getting out of it as compared with what we otherwise know about Soviet testing; Frosch expects that what was missing from the WWSSN discussion in the Congressional Hearings was how the WWSSN has helped to understand Soviet tests, about which other pieces of intelligence were available, information which was classified
• Frosch doesn’t find concept of dual use very helpful: Frosch: “I have never seen a defense technology for which I couldn’t instantly see civil use for. The usual problem is the price isn’t right... And I have never seen a piece of science for which I couldn’t think of a military use.”
• question: to what extent was Vela Uniform classified? Frosch: what was classified was either stuff that compared unclassified with classified material, like separate intelligence estimates; or system estimates for the classified purpose of saying, if the Russians try to cheat we can detect down to x kilotons. Frosch: at a certain point detection capability is policy and political problem; Frosch lost patience with part of the arms control people because they argued that detection is possible and of course evasion won’t work; Frosch amused himself by inventing evasion schemes: testing in earthquake area, testing in aftershock sequence; big lunchtime arguments over whether these evasion schemes mattered or not; Frosch’s point was that in the end it was not a technical problem; he didn’t say to just send it to the politicians but he didn’t see any way in which any set of technical facts that are plausible can stop this argument about whether detection is good enough; Frosch says in the end it was a battle of the optimists and the pessimists, which is always portrayed as a battle between the militarists and the peacemakers, but the job of the militarists was to take not too much risks; Frosch says that he lost patience with both sides; Frosch believes now that discriminants have been found (PS, mB vs. Ms, etc.)
• other Vela Uniform programs: how much influence did Frosch have on VU programs in universities: Frosch: he came out of ONR tradition: don’t tell anybody what to do, but to raise questions; with VU came support to most seismologists and also new instruments; most seismologists were so poorly supported before that a $1,000 amplifier was just not an option; giving them instrumentation, then they would begin to be interested; but most seismologists would do straight seismology; some seismologists, who were leaders, got in fact interested in detection problem: Frank Press and his friends would do earthquake stuff and some specifically sponsored discriminant stuff
• Frosch: then they were other people who did straight seismology and then Frosch would throw in an extra question, like distortion at the source etc.; Frosch wouldn’t tell somebody to work on what and with who: style was more like “We just heard that x at University y has done z, and we thought you might be interested in talking to him directly; and if it turns out that you guys want to work together, let us know, maybe ARPA can help.” ONR style; Frosch says he was not a seismologist, but he went to the meetings and got to know some people, especially from Lamont and specifically Frank Press was obviously involved. Press and [Carl] Romney knew who was any good, so they helped ARPA involve people: Jack Oliver (Frosch knew him from Lamont days and he was a seagoing guy), and somebody at Dartmouth [?], Eugene Herrin
• VESIAC meetings: helped to facilitate interaction; ARPA also organized sessions at the SSA [Seismological Society of America]; Frosch: there was a virtual club of senior seismologists who were interested in the detection problem scientifically and politically, who were in communication with each other;
• Anchorage earthquake 1964: SSA meeting, hallway discussions about whether arrays are going to be important etc. and whether that was the right place to spend money; but then the earthquake happened, and every seismologists focused on that (like stepping on an anthill): they all wanted to be in Anchorage; in general: ARPA program closely connected with the seismological community, sort of part of it, everyone understood what ARPA was doing;
• how were program decisions made? Did some seismologists take the Vela Uniform money and run? Frosch: sure; part of the view was that seismology had to become much better if we want to solve detection problem; therefore ARPA perfectly willing to support those who didn’t have detection application in mind; Frosch: we would have been unhappy if everybody would have taken the money and run, what ARPA “was seeding a science that was gonna have to get better and we had views of how it had to get better and Frank Press was sort of the kitchen cabinet adviser and a formal adviser.” But that was not where the money was going: most of the money went to AFTAC and in Space Vela [Vela satellites] ; therefore Frosch felt comfortable spending this money [for the universities] and thought it was clearly defensible; cannot remember any big budgetary problems
• Anchorage earthquake drew President Johnson’s attention: great deal of action in earthquake prediction: Sonnemann and Frosch were commissioned with Press to design an earthquake detection network system for San Andreas fault: and ARPA did! Would cost more than LASA (more than about $8M: LASA budgeted at $8M, and opened at $7.6M: Frosch says that that might have been a factor why ended up as Ass. Seer. of the Navy R&D, plus old friends..) so ARPA designed system for San Andreas fault, but was not built: Vietnam took over. What was built later in California has a lot of similarity with ARPA’s plan from the mid- 1960s
• LASA grew out of the conversation with [Hal] Thirlaway about P waves detection: Frosch went to UK, stayed with Thirlaways (saw Beatles on TV); Frosch thought about LASA as translation of underwater acoustic array into seismic arena; started out with P wave, with Romney’s input added low frequency shear wave array, which turned out to be important
• Harry Sonnemann and Paul Green then build LASA; was a mess: all the land was owned, you have to go to 200 ranchers and get permission etc. (“get the feds of my property” attitude quite often); ARPA was prime contractor; power system in Eastern Montana: about 17 local power cooperatives: with each of who you had to do a treaty; same with phone system: 15 local phone companies; biggest problem in digging the ditches for the cables were rattlesnakes; Harry Sonnemann was the contracting manager and Paul Green was the one for the electronics: Harry was Frosch’s agent for getting LASA build
• ARPA had no contracting authority and had to go through agents: Frosch: but that didn’t make any difference, because procurement people could just find a way to cover this under existing agreements with the Air Force etc.; sufficiently transparent system that Frosch didn’t worry who the agent was; Sonnemann and Bates probably knew the details; ARPA was small agency, they partied together, had dinner with each other; Frosch: ARPA a place that actually does work; Frosch: ARPA was invented to be not clogged by existing bureaucracy and it worked this way
• contact with other program directors: Frosch: yes, informally with Licklider; question: formal connection between signal processing and computing at ARPA? Frosch: vague recollection: Vela Uniform [VU] used some of the computing peoples’ contractors (and university people); Ivan Sutherland; Frosch: the place worked as one entity: Frosch knew the guys who did BMD [Ballistic Missile Defense], they had suggestions about how to do things in Vela; Frosch says that ARPA not a turfy place at that time: everybody helped out
• Frosch: maybe ten people in VU at that time, including secretaries; lots of traveling to contractors; going to Montana all the time; someone would pick the brains of the oil people; Frosch: it was a lot of fun, very hands on; always feedback from Carl Romney and Frank Press, and asking Hal Thirlaway
• LASA worked: simple compared to other projects at that time; Paul Green had just worked on first ground radar based air battle control system; in seismology one had hours to do signal processing, not microseconds; LASA big and complex, though;
• Was Frosch satisfied with LASA? Frosch: satisfied enough; problems in array gain in shear waves, regional effects: this might have been the start of the regional correction business for all of seismology; but it was clear that LASA brought enough real gain; other arrays: NORSAR, ALPA; for NORSAR: location was problem: Sweden didn’t want to be involved, Norwegians did it themselves: made improvement; in total not good enough to build another 20 of these large arrays; good enough for signal processing ideas; most influential arrays for LASA: Tonto Forest (was Carl Romney’s array) and British array at Eskaldemuir; Tonto Forest left Romney saying that it was probably not worth the trouble, whereas Eskaldemuir left Thirlaway with the feeling that this was the way to go. Frosch: the whole computer thing worked in general; nobody was coming back from Montana sounding worried about LASA: more of a “Look, we did it”
• control center [in Billings] was in the back of a store front, looked like a retail store gone out of business; but in the back room you could listen to the earth through an ear which spread over half of Montana; Frosch mentioned again that LASA came in under budget
• LASA funding: Frosch thinks that it was above what an ARPA director could decide, estimate was under $10M; but Harold Brown could authorize it (McNamara probably not necessary: if Sproull contacted McNamara, then just a short note, and McNamara signs it, but Frosch thinks it never went to McNamara: in fact, political manipulating necessary to get the Secretary’s attention for LASA; with Brown different: you could just call him up, with Foster even easier
• recollection of other VU projects? Nuclear tests for VU? [Nuclear test] SHOAL shortly after he began at ARPA; Frosch quite involved in SALMON and LONGSHOT. Frosch remembers less of SALMON (Hattiesburg) shot then preparation for the shot: decoupling pretty crude stuff then; arms control guys desperately didn’t want decoupling to succeed and the others... And ARPA in the middle. He went down to Hattiesburg to check on the project (same time as civil rights movement and Freedom Rides): flew to New Orleans and rented a car for three hours’ drive to Mississippi: discovered that it was safe, because guy behind counter asked where he was going: with Hattiesburg and DOD mission he was safe from attacks, Frosch guesses (no evidence for that); he was present during the test, not a big deal, just a blip on the control display; people in Hattiesburg perfectly comfortable; LONGSHOT (Aleutians) different: ARPA vs. state of Alaska; nothing went wrong with the shot; Frosch gave assurances that everything was safe; embarrassing because he gave assurance that that was the only shot out there: later two others: MILLROW and CANNIKIN; critics: LONGSHOT will trigger an earthquake, will start a new volcano, it makes tsunami etc., Frosch: all that didn’t make sense
• major reasons to test in the Aleutians? Frosch not sure: he thinks there was an argument if you did a shot in the place like that, that because of the deep slope into the trench and general volcanic geology, that you would get a very small signal; island arch trench might be a reasonable place to go if you want to hide an explosion; the seismologists also had some scientific reasons; argument went that everything we know about seismic detection might be wrong if tests were conducted in this environment; argument that there might be funny places where people could hide tests
• his time after ARPA (after 1966): Frosch did not keep up with VU; he spent with his Ass. Sec. of Navy job, where he spent his next six years, fully occupied with that; talked occasionally with Steve Lukasik; Frosch wouldn’t read the ARPA reports; biggest connection with ARPA afterwards more on projects like DEFENDER, AGILE; more connected socially then professionally, saw Bates, Hertzfeld etc.
• presented BMD to Paul Nitze, became then Ass. Sec. of Navy R&D