Notice: We are in the process of migrating Oral History Interview metadata to this new version of our website.
During this migration, the following fields associated with interviews may be incomplete: Institutions, Additional Persons, and Subjects. Our Browse Subjects feature is also affected by this migration.
Please contact [email protected] with any feedback.
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.
In footnotes or endnotes please cite AIP interviews like this:
Interview of Kolumban Hutter by William Thomas on 2012 October 2,Niels Bohr Library & Archives, American Institute of Physics,College Park, MD USA,www.aip.org/history-programs/niels-bohr-library/oral-histories/40452
For multiple citations, "AIP" is the preferred abbreviation for the location.
In this interview, Kolumban Hutter discusses topics such as: his work at ETH Zurich; his research in glaciology; graduate degrees at Cornell University in theoretical and applied mechanics; Hans Ziegler; Hans Rothlisberger; Peter Kasser; ice plates; Daniel Vischer; John Nye; John Glen; thermodynamics; Andrew Fowler; Leslie Moreland; International Glaciological Society; hydrodynamics; Richard Sebass; fluid mechanics; physical limnology; visiting professorship at University of Arizona in Tucson; Terry Hughes; ice sheets and shelves; teaching at Darmstadt University of Technology; Ernst Becker; Reinhard Calov; Mary Williams; cold-temperate transition surface (CTS); global climate models; and working at Academia Sinica, Taiwan.
This oral history is based on a conversation between myself and Prof. Kolumban Hutter. Prof. Hutter was offered the opportunity to edit the interview transcript, and he generously took the time to make extensive clarifications and additions to it, resulting in a document of much greater historical value. However, readers should be aware that, although the basic interview format has been preserved, this transcript no longer closely reflects the original conversation and may be more accurately thought of as a written memoir. For the most part, new content is indicated with square brackets, while minor changes have not been indicated at all. But in making changes, it has not been possible to follow any rule strictly. Please note, very little original content was removed in the amendment process, and no point of major historical significance was removed. The original audio files are available.
This is William Thomas. It’s the 2nd of October 2012, I’m at the ETH in Zürich with Professor Kolumban Hutter. Did I pronounce that correctly?
Yes, and we are here to talk today about his career in glaciology and its intersection with his other interest in mechanics.
So why don’t we start a little bit with how you got into the field that you're in, in the first place. Where did you grow up and where did your interest in science and mathematics get started?
Well, I grew up in the northeastern part of Switzerland and went to school in Rorschach at Lake Constance. I did my high school degree at Sankt Gallen, the capital of Canton Sankt Gallen, where I got an education with an emphasis in ancient languages, [a high school degree with Latin and Greek as main languages].
Then I went to ETH, became a civil engineer in 1964 and wanted to do a Ph.D. degree in the field of Engineering Mechanics, but my first attempt failed.
I had to leave and then worked in this institution, the Laboratory of Hydraulics, Hydrology and Glaciology [VAW, abbreviated for Versuchs-Anstalt für Wasserbau, Hydrologie und Glaziologie] as a hydraulic engineer, involved with cooling of atomic power plants, not by cooling towers but with river water used as the cooling medium.
I did that for two years and was very unhappy. The institute at that time was not in the best shape [a lot of mobbing was going on]. So, I decided to go to the United States and did a Ph.D. degree in the Department of Theoretical and Applied Mechanics at Cornell University with a Chinese advisor, Professor Yih-Hsing Pao [1930-2013].
I worked there for a Master’s degree in structural engineering and, after that, for a PhD degree in the electrodynamics of moving solid media. That’s why I have minors in physics and in mathematics, which I needed and also wished to pursue in that application.
How did you decide to get into civil engineering in the first place?
I actually never wanted to become a civil engineer. My first choice would have been music, but my violin teacher said that I’m not talented enough. Then I leaned towards mathematics and physics, which was a unified education at that time. My parents said, “You are not smart enough for that.” So my mother picked civil engineering and I obeyed, as was the custom at that time.
Okay, so that was your undergraduate degree at the ETH?
That was my undergraduate degree up to the diploma level, as one said, which corresponds roughly to a Master’s degree, but when I went to Cornell, I wanted to make a Master’s degree as well because you never know what happens if you are abroad for a couple of years, away from the family. So, I wanted to have at least one ticket to come back if anything would happen that did not allow me to finish the PhD degree. So, that was the only reason for my Master’s degree.
How did you decide to go over to Cornell?
Well, that had to do with my first attempt to do a PhD degree here in Zurich in mechanics. That was in 1966.
Perhaps we’d better go back and decide how you settled on mechanics —
I was always interested during my civil engineering education in the more rational topics of the profession, and this was structural mechanics. Before that we had classes in fluid mechanics, rigid body mechanics, and strength of materials during the basic education, roughly at the freshman-junior level.
[In my last year of studies, I got the chance to become a student tutor of mechanics, and afterwards a job as an assistant of mechanics, but after two years in 1966], I was told by the leading mechanics professor [Hans Ziegler] that I was incapable of getting a degree, and essentially he fired me.
Okay. Was it your original intention to do that academically or just —
It was intentional, yes. I wanted to do this academically, not in VAW but in the mechanics institute at ETH. It so happened that a Cornell mechanics professor, later Dean of Engineering at Cornell University, spent his sabbatical in the mechanics institute at ETH. He had directly observed what was done to me and came to me in a moment where we were the only two in the room and said, “This is not correct what has happened to you. If you ever need me, you can contact me,” and this is why I approached Cornell.
I see. Sorry, that person was from Cornell or…
He was Chairman of the Department of Theoretical and Applied Mechanics (TAM), when I started as a PhD student, then Dean of Engineering of Cornell University, and later became President of Worcester Polytech, [and at last professor at the University of New Hampshire in Manchester, New Hampshire].
What was his name?
Edmund T. Cranch. He’s still alive, probably about close to 90 now, I would say [and living in Ithaca where he also grew up]. So I was rescued by him.
You were rescued by him.
[Yes — in the true sense that somebody’s life is rescued by an American], and I could tell a long story how all this happened, how Professor Ziegler even tried to make me fail in Cornell, [and how I performed in Cornell. In fact it so happened that I performed well in the Department of TAM at Cornell and was respected, but the Zurich mechanics professor tried to interfere. I was told, he made an attempt to convince the faculty at TAM to make me fail the A-exam (This is the comprehensive exam that one needs to pass before starting working on a dissertation). I was secretly told this after I had passed the Ph.D. exam in ‘72.]
You see, my start was very negative, but it turned positive.
I see. So you had your family before going over to Cornell and then —
Yes, my wife Barbara and I had one child [Bettina 1967, before Cornell] and the second child [Katja, 1971] was born in Ithaca, New York.
Okay. So then how did you settle on, it was Yih-Hsing Pao?
It was Yih-Hsing Pao, [whom I picked as my advisor], yes.
Did you intend to work with him originally or did you decide that when you —
When I arrived in Ithaca, I knew nothing; actually I wanted to work with Prof. Cranch, but he had just become Associate Dean at that time and was no longer in the department, so he was out as a principal supervisor.
So, I was looking for someone else, and Prof. Yih-Hsing Pao was, at that time, the person whom I was most attracted by.
Okay, and what was he working on?
His specialties at that time were waves in elastic solids, stress concentrations, and he had just newly started with electromechanical interactions, more specifically with mechanics of solid bodies, mainly subjected to electromagnetic forces and responding to those.
Yes, I noticed that you worked on similar things, naturally I suppose.
Yes [and no!] I chose this topic as my PhD work because I had won a fellowship, when I had passed the A-exam. The engineering school at Cornell University had a fellowship at that time, called Last Year Fellowship for the best student in the engineering school. Just when I had passed the exam, I happened to be number one, and so they told me, “You can get this fellowship and then you’re free to do as your thesis work whatever you want to.”
So I picked the new topic in which Professor Pao was concentrating his research efforts as my PhD subject; it was a very fascinating thing, and I had also a great interest in enlarging my knowledge of physics and using mathematics, and, at the same time, thermodynamics, basically mathematics in the context of thermodynamics, and this I could achieve.
Was this purely theoretical work or was there an investigative or experimental —
Well, a physicist or mathematician would say it was not purely theoretical, but it was top theoretical with my background.
There was no laboratory work or anything like that?
[For another student of Professor Pao, yes], but not for me. I had done laboratory work at VAW-ETH for two years, so I felt I had experience in engineering laboratories and laboratory work. [I was in fact happy that I did not have to pursue laboratory work for my Ph.D. Degree.]
You said you were, at that time, more attracted to the mathematical, rational —
I was always more attracted by that, and I wanted for a while to become a mathematician, I had thoughts to study mathematics here in Zurich after my failure in mechanics. However, the Dean of the Math School, whom I had consulted explained the severe conditions: I would have needed to do the complete undergraduate education in mathematics. We concluded and he said, “That’s stupid, go to America!”
So you were there for four or five years, was it?
My education as a civil engineer was a four-year training, fall ‘60 to late ‘64, and my work in the fields of mechanics and the laboratory work in hydraulics in Zurich was four years. Then I went to Cornell, and that was another three and a half years.
This was, just to get the chronology correct, after your work on the cooling of atomic power plants?
That was after my work on the cooling of these reactors by river water, yes.
Is there anything we should discuss with respect to that, or should we just, was it —
Well, my laboratory work on the cooling of nuclear reactors by river water became insignificant or illusory. [The laboratory experiments dealt with the dispersion or mixing of river water that was heated by the reactor and re-introduced into the river. The purposes were (i) to mix the 10°C hotter water as quickly as possible with the remaining river water, and (ii), to avoid recirculation of heated water in the power plant.] I worked on this for four atomic power plants of which three were not yet in construction, one already existed, and they want to shut it down now in 2019. The three others were not done with cooling by river water.
The reason is very easy, because ca. 10 power plants were planned along the River Rhine, from Lake Constance here to Karlsruhe. I estimated that in winter at low discharge the River Rhine would roughly be 30°C hot; fish would die. So the government soon stopped this cooling technology by river water. At the time when I was in Cornell they changed everything to cooling towers. [The technique was also popular in the US with strong research activities in the 60s at MIT (Prof Harleman).]
Okay. So should we discuss any further your research at Cornell or shall we move to your return here?
We should, perhaps only briefly mention, that during my education as a PhD student, I wrote only one paper with my advisor. That was my very first paper in a peer-reviewed journal and had to do with beams on skew supports, more specifically bridges on skew supports. It used a mathematical technique called singular perturbations or matched asymptotic expansions to get the right stress distribution in the bridge cross-sections. [This problem was my self-suggested topic for my M.Sc.-thesis, with a nice short story about how it came about, but it is off the road of my scientific life.
In short: in my youngest years as a civil engineer I had occasionally done consulting work. One of these was statics for a bridge with skew supports. Later, after the measurement of the strains and stresses in that finished bridge, it was found that large errors between the computations and measurements existed close to the supports, while errors away from them were small, then a mystery. In a reading seminar at Cornell on singular perturbations during my first semester, when cables with little bending rigidity were treated, I realized where I had made the error in my statics of the torsion in the bridge calculations. I worked three days and two nights over the following weekend, presented the result to Prof Pao, and asked whether he would accept this as a basis of a M.Sc.-thesis. He did three days later. It became my first publication in spring 1971.]
You sent me your publication list ahead of time, that’s in the International Journal of Solids and Structures.
Yes, this is a rather good journal. Back to the main discussion, since electrodynamics and thermodynamics were also basically new for me, my stay at Cornell was a tremendous educational experience with publications done after the degree.
I came back here to ETH to this institute [VAW] and had fairly generous conditions, so that I could, beyond the work, which I had to do in the field of glaciology, finish my PhD thesis work. With Prof. Pao, I could write three papers, and that lasted about a full year. Generally, I spent half a day every day during a one-year period writing three papers with him and one or two by myself. [The second half of the day was then devoted to VAW work. I never got complaints that I was insufficiently productive.]
Maybe I want to say something else. There was a former PhD student and then assistant of theoretical mechanics at Eindhoven University of Technology with the name Alfons (Fons) van der Veen. He had written in a note added to his PhD dissertation that what I had done with Professor Pao in my dissertation had some mistakes in it. So I contacted him, we had a long correspondence between ‘73 and ‘75, where we wrote about at least 100 hand-written pages, each, and we had through this correspondence accumulated enough work and extensions of it that we decided to write a book on the subject. I spent my vacations in ‘76 and ‘78 working with Fons on the electrodynamics of solid bodies.
I have that list in here, too, that’s Field-Matter Interactions in Thermoelastic Solids.
Exactly. That book even got a second enlarged edition in 2006.
Yes. This area of integrating, I guess, electrodynamics and thermodynamics and material science, was that a fairly novel thing to do at the time? Could you give me a sense of the field?
In engineering professions, it was brand new, and the new stuff of it was on the analytical side. It was recognized that nonlinear theories, which became important with better electronic computational facilities, were not done correctly before. In fact, Alfons van der Veen and I were the only people worldwide who, then, did it correctly. In fact, Alfons did it correctly, and in my computations he had found a mistake.
But we were the first ones who tried a systematic approach to write down physical equations, of which every physicist would say, “Well, that’s obvious.” But, to find approximate equations, from which we could, for instance, find the electromagnetic vibrations of a tuning fork, which, at the beginning of the electronic watch, was not mastered yet.
I saw that you did a little bit of consulting with the watch —
Exactly! This has marginally to do with it. But I did this work before I went to the United States, before my Ph.D. degree. I did that consulting assignment here with a mathematician together, later Prof., H. Schwarz of the University of Zurich. Indeed, this work may have had some influence that I picked electrodynamics with Professor Pao.
Right. As far as the day-to-day work goes, is it mainly working with equations or are there computers involved at this point in time?
It was mainly developing equations for fairly complex electromechanical interaction properties of different solid materials and later fluid materials.
However, in my second part of the dissertation, I had also attacked a seismic wave problem. The problem was: can we, from a balloon, measure an earthquake passing, say, in the Earth below? So, the balloon would not be affected by the seismic waves of the solid; it would be stably suspended in the atmosphere and sense the electromagnetic oscillations. I did the electronic computations for this problem as my last chapter of the dissertation. [Technically it was a seismic Rayleigh wave problem of a magneto-elastic half-space and a half-space atmosphere with possibly measurements of the magnetic field from a balloon. It meant solving a boundary (eigen) value problem. I wrote a FORTRAN program, punched it in cards, and ran, during approximately three months from my office to the computer center in the neighboring building at Cornell until debugging was completed, and then wrote my last thesis chapter. Measured in labor it was the toughest part of my PhD work.]
The work was never published. My advisor had the gut feeling, “Probably it’s not correct,” so that’s why we refrained from publishing it.
I see. Okay, then your Habilitation is through Vienna?
This choice has also to do with the electromagnetic part of my research activities. When I came back from Cornell, I sensed I could not do my Habilitation degree [at ETH because of my experience with Prof. Ziegler, the head of the mechanics department at ETH. It was obvious for me that my application would be blocked.] Are you familiar with the Habilitation degree? This is a form of higher-level PhD degree which you needed at that time to apply for professorial positions, if I ever would do that.
So coming back from Cornell, I always had the intention to get into an academic career. I knew, because of my experience with the mechanics group in Zurich, this would be hopeless here in Switzerland. They would all block it, and indeed they tried to block everything subsequently. So, I was looking for professors in Europe who would do research in the electrodynamics of solid materials. One was in Vienna, that was Professor Heinz Parkus [1909-1982], and another one was in Eindhoven, this was Professor Johannes Bartholomeus Alblas [1915-1998], the boss of Alfons van der Veen.
I assumed that these two professors would understand my background [and my work in electro-magneto-mechanical interactions] and would, perhaps, be willing to accept my candidacy. I started asking Professor Parkus, and he said after a while, “Yes, you can submit a dissertation, if you want to.” So, eventually I did that. [The contact was established by Prof. Udo Gamer, a mechanics professor in Vienna, whom I met in spring ‘73 in the mechanics department at Stanford University where I delivered a lecture on floating ice (see later) and visited Prof. Pao (who was there on his sabbatical) for collaboration on a thesis paper.]
I should have asked you right after I asked about numerical modelling, the equations involved, this is mainly solving differential equations?
Yes, these are initial-boundary-value problems of partial differential equations [(in my electro-magneto-mechanical Rayleigh problem, an eigen-value problem), with focus on electromechanical oscillations in the air].
Okay. So then you do end up back here at the ETH?
Well, when I was returning from the US, I was looking for a job; at those very high economic times, finding jobs would be rather easy. But for me it was difficult because of my experience with the mechanics prof., Hans Ziegler. Switzerland is a very small country; so, if somebody with academic intentions is looking for a job like me, the former employer — [here Prof. Ziegler, who was a well-known and at many places a well-respected figure] — would be consulted. [I cannot prove that he interfered. The fact is that all my attempts failed, in which the slightest connection with research was involved. In fact, the chief engineer of the hydraulics section of VAW, Eng. Ernesto Bisaz [1918-1979], had strongly recommended me to the newly appointed director of VAW, Prof Daniel Vischer [pronounced Fisher]. Luckily for me, I was hired. I never had met Prof. Vischer before.]
So, this was my only place where I could get a job.
This is the VAW?
Yes, this is the Versuchsanstalt für Wasserbau, Hydrologie und Glaziologie [Laboratory of Hydraulics, Hydrology and Glaciology], but following Wasserbau they cut it off: Versuchs-Anstalt für Wasserbau. [VAW in the early seventies was and still today is rather interdisciplinary; it had in 1972 three sections: (i) Hydraulics, which meant hydraulic power plants and morpho-dynamics of rivers, (ii) Hydrology and (iii) Glaciology, mainly alpine glacier mechanics and glacier hydraulics.]
At that time, all three topics were researched at VAW. I had two years’ experience in this institute, and since my experience from an emotional personal side in hydraulics was not very positive before I went to Cornell, I talked to the director of VAW, Professor Vischer (he was the boss in this institute at that time), into letting me become a glaciologist. [This would bring me into a section that was likely free from mobbing, which I had seen in the hydraulics section before.]
Ah, okay, and how did you arrive at that goal?
Well, when I returned from the States, in my first interview which he gave me, I just told him that I would like to work in glaciology. I picked glaciology because I knew there were a few people who were known worldwide as glaciologists, there were at least two at that time: Hans Röthlisberger [1923-2009], Peter Kasser [1914-1989], and later Almut Iken. These were the scientists who had impressed me from looking through papers and reports and from discussions by word of mouth and conversations. I did not know them personally, not one.
So, Professor Vischer, who was a very ambitious person, agreed, and at that time the finances were incredible. He could, during his first 10 years, hire about 70 people. I mean it’s unbelievable how large this institute’s wealth was. I didn't know this in advance, and my desire to work with glaciology was not prompted financially. I wanted to avoid trouble and to do work, and my gut feelings told me the glaciology section would provide these conditions.
This is for the institute here as a whole?
For the institute as a whole, perhaps also for ETH. The school obviously was very rich. This was, and still is so, because it is a government school, annually financed by the Swiss parliament, which is by tradition generous to education. So, I got that job, and Prof. Vischer had a tremendously generous attitude. I forgot whether this was in our first or second meeting, he said, “I want you to make our institute known worldwide for mathematical glaciology,” and my knowledge of glaciology at that moment was zero, absolutely zero!
What was your first exposure to glaciological theory? Was this at Cornell or here?
It was here, and what you should perhaps know, in the winter of ‘62-‘63 — this was the very cold winter when Lake Zurich froze over, and the ice, the plate on top of it was about 50 cm thick, if I remember correctly. Even Lake Constance froze over entirely. One Sunday afternoon my mother walked from Rorschach (at the Swiss shore) straight north to Friedrichshafen (at the German shore) which is about 14km.
Across Lake Constance?
Across Lake Constance and she didn’t even inform us, we were scared to death! [We received her phone call from there when it was already dark, around 8 pm. Anyway, this ‘Seegfrörni’ = ‘lake-freezing’ was still in Tschoon’s (Hans Röthlisberger’s nickname, pronounced “John”) head], when I returned from the States in the beginning of January ‘73. It was still a problem for the glaciologists here; they had to perform consulting for the Zurich lake police and still didn't feel secure whether they knew enough as to when they should let the people safely and freely walk on the ice.
So, I was assigned by Dr. Röthlisberger to study floating ice plates. This was absolutely marvelous. I could document to my boss that I was competent in this field because I did my PhD with Professor Pao, and he was a former student of Professor Mindlin at Columbia University, a top specialist in plate and shell theories.
Mindlin was a second-generation American but originally a Russian. He was the American specialist in developing plate theories where material properties varied across the plate thickness. Since the atmosphere is cold and the water is warm, at least the temperature varied across the depth of the plate, and that influenced the structural properties of the plate.
So, when I asked Professor Vischer, I was immediately granted permission to study this problem. And furthermore, I knew from Professor Pao’s seminars how plate theories were developed in the Mindlin spirit. So, I just could sit down, work intensely, and generated in three or four months a report about 150 pages thick. [Technically, the three-dimensional equations of motion (essentially the momentum equations) had material coefficients which were temperature dependent that varied strongly across the depth (z-direction). By a procedure, known as ‘principle of weighted residuals’, paired with a function expansion of all variables in the z-direction, the three-dimensional original equations could be transformed to a larger number of equations in only two dimensions. One replaces the three-dimensional equation set by a larger, more convenient, set of equations in two dimensions. The coefficients in these two-dimensional equations now depend on how the coefficients in the original three-dimensional equations vary across the depth.]
[The emerging plate model was rather general, more general at least than just to be useful for floating ice plates, but it was phrased to be convenient for ice plates.] Of course, the water was playing a role, the atmospheric influences of the climate or the meteorology had some influence, and depending on the temperature, the ice had inside some water bubbles and some air bubbles, which had an influence on the material properties as well.
So this was the theme, which I had started in Dec. ‘72, and it was June or July ‘73, I guess, when it was published as a VAW report.
Were there any other publications on this sort of thing at this point or was this just fresh?
Yes, in the mechanics community, in civil and mechanical engineering, there were many people who derived plate and shell theories in the spirit of Mindlin. [In fact there are even new models coming out today.]
Right, but not for ice?
[Also for ice, to a limited extent. In April-May ‘73 I presented my model at the Canadian Congress of Applied Mechanics in Montreal, I believe, and met a graduate student of mathematics, Mary Williams from Simon Frazer University in Burnaby, who was attracted by the model. She extended it to viscoelastic ice response with aging relaxation functions and submitted that work as part of her PhD dissertation. Later, in ‘75 or ‘76, Prof. Vischer paid her a research stay of a few weeks at VAW, when we wrote our first paper on ice shelves that was published in Acta Mechanica and in a conference proceeding of an ice materials conference in Lyngby, Demark in ‘79.
Moreover, some impetus came from Russian scientists, who had a conjecture, partly supported by their experiments, that the Poisson ratio, which is one major coefficient of isotropic elasticity, depended on temperature, and this variation would have an influence on the elastic wave propagation in floating ice plates.] The US-Army had in Polar Regions air strips on floating ice and needed to be sure that their planes could land on floating ice. The Russians seemed to say: “Be careful in computations, don’t use a constant Poisson’s ratio”. [There was no plate model around in ‘73, which could resolve this unproven conjecture, but my plate equations had the potential to do so. In sea ice the proportion of brine inclusions is also temperature-dependent and this influences the bending stiffness of the composite plate. The influence of the temperature variation across the plate thickness is quite a bit larger in sea ice than in fresh water ice; but there was no proof of this Russian claim that this variation was significant in loaded floating sea ice plates.]
Who specifically were the Russians who are working on this, if you can recall?
I’m not sure any more. It’s too long, I’ve forgotten the name.
Is there an institute where they worked?
Yes, they were in St. Petersburg; well, at that time in Leningrad. In the USA it was the collaboration of earth scientists and engineers at SIPRE.
Right, the Snow Ice and Permafrost Research Establishment which became CRREL.
SIPRE was located north of Chicago in Wilmette, and then they moved to Hanover, New Hampshire and became CRREL [Cold Regions Research Establishment Laboratory] of the United States Army. [My contacts were with Andrew Assur, Günther Frankenstein and Don Nevel.]
The scientist who was chiefly involved with this was a former Russian who was working there, and I got all this information in detail from CRREL through Andrew Assur, who left Russia around 1920. Actually he was from a region in Russia where people spoke German. He told me these interesting materials’ facts and documentation. So, I was likely the first person within this very limited, little ice research community who could account for the influence of the variation of Poisson’s ratio across the thickness of the ice plate to estimate its bearing capacity.
Were you in correspondence with them or did you actually meet with them?
[The first contacts with CRREL were established by Tschoon Röthlisberger, who had worked in the ‘50s and early ‘60s as a geophysicist at SIPRE and knew all these people. He arranged for me a lecture at CRREL when I travelled to the Canadian Congress of Applied Mechanics in Montreal in April ‘73, and Prof. Vischer generously paid. On the same trip I also went to the Canadian Research Council in Ottawa, the Department of Building Research, where Dr. Lorne Gold led an ice research group. Moreover, I lectured on “my” plates also in the Department of Applied Mechanics at Stanford University, where my thesis advisor, Prof. Pao, spent his sabbatical, and we worked for a week on a paper from my thesis.]
On return, I attacked the question of the temperature-dependent Poisson ratio because the CRREL people were concerned so much that they might miss something, and I proved that the effect of this variation is “peanuts” for fresh water ice plates, very small, and somewhat larger for sea ice plates. [This is published in a Royal Soc. London paper in 1975], and it killed the problem as a research topic.
But it was good work, and I could submit this work in Vienna as my Habilitation dissertation, and I got permission to teach there in ‘76. I took the exam in November ‘74, but, since I was not an Austrian, they needed permission from the Swiss government, whether I was allowed to get this degree in Austria, and that took up to ‘76.
[I lectured in Vienna in blocks of two weeks each day 4 lectures making a 3 credit hour course from ‘76 to ‘87. Topics were continuum thermodynamics, ice mechanics (chapters from my draft book) and stability theory.
One note seems worth being stated here. I was not straight away permitted by Prof. Vischer to attempt a Habilitation degree in Vienna. He first made efforts to see whether this would be possible at ETH. After a few weeks his permission came. I have no knowledge how he came to this decision; all I know, he made inquiries at the Mechanics Institute and then gave his permission. Moreover, Prof. Parkus at the Technical University in Vienna visited Prof. Hans Ziegler in Zurich before my oral exam in ‘74 on one of his business trips and asked Prof. Ziegler, whether he had anything against me getting a Habilitation degree through him in mechanics at the Technical University in Vienna. Apparently not, as Prof. Parkus informed me about this.
A further spin-off was a series of lectures on floating ice plates, which I held at Memorial University, Saint Johns, Newfoundland, Canada in summer 1978, prompted by an invitation by Prof. Fritz Legerer, an Austrian physicist and former Ph.D. student of Prof. Parkus. In those lectures one of the PhD students was Vernon Squire, then a graduate student at Scott Polar Research Institute in Cambridge, UK. He took up one part of my work: waves in an ocean layer covered by an ice plate, and extended it in his PhD dissertation. Vernon later emigrated to New Zealand and became a professor [now Deputy Vice Chancellor at the University of Otago, in New Zealand.]
Okay. So then after that, your first publication, you had this large report…
Well, I made two papers of that.
Right, I saw those in your CV. But there was also the large report. Was that submitted, too?
No, that was one of the institute’s reports, which have been published by VAW since Professor Vischer started the series of reports in 1970; about 250 have been published since the inception. [These ‘Mitteilungen’ are often less densely written and also contain additional information beyond the mere central facts. They are, thus, easier to read than the more formal papers.] By the way, these reports were all covered in a green cover [now they are dark blue]; so my report, because it was full of equations, with which glaciologists were not so familiar, they called my report “the green monster.” So if you ask MacAyeal what is the “green monster…”
He will likely know what it is.
Your name came up originally in my interview with Doug MacAyeal because he said he was modeling something — I forget the specifics — and that there was actually a much more rigorous way of doing it, and that he had gotten that methodology from your work.
Yes, that is possible.
So returning to your work in materials science, you seem to have an interest in foundational problems and in the unification of different approaches. I think you may have covered this a little bit in your work with Alfons Van der Veen. So was that an explicit interest, trying to get at the foundations of a subject?
That is maybe the red line running through my vita. I always want to study a subject from its basis, from the foundations from which one can derive the equations with which one works in their description. I did this at the beginning of my post-Ph.D. work when I worked on the “green monster” and shortly later, I still was in the learning phase, [using papers from the pile, which Tschoon Röthlisberger had given me in ‘72-‘73 as initial studies].
So, I tried to understand the formulations of continuous ice sheets and glaciers, which were in a miserable state, if I may say so. [Authors seemed to study isolated facts, motivated by observation, and these were mostly only being described by explanations, which I did not understand. Most papers were dealing with motion of ice masses in a rather ad-hoc fashion. There were only a handful of scientists guiding me to an overall description based on principles of classical physics.]
So my aim was to generalize and unify these apparently disparate glimpses of idealized explanations of observations.
Yes, if I can go back and through the field of glaciology as I understand it. My knowledge is mainly from the British and American perspective and I assume that people like John Nye and John Glen, Johannes Weertman were important.
In the piles of papers, which I received through the recommendation of Tschoon Röthlisberger, I identified John Nye and John Glen from the UK, and Lliboutry in France, of course. John Glen and John Nye had Cambridge degrees and had worked at the Cavendish Laboratory. Nye was a student of the son Brack and John Glen was Orowan’s student, if I'm not mistaken. [I could expect good physics, mathematics and materials science. Lliboutry also had a degree from a materials science Nobel Prize winner. Scientists who equally impressed me were also Hans Weertman at Northwestern University and Barclay Kamb at the California Institute of Technology, and Bill Budd in Australia.
For me, the writings of Nye were the best to understand the theory, and those of Glen made me understand the materials science of ice. But reading the many papers was hard and at the beginning diffuse, and therefore confusing. I had to separate in my brain ‘the chaff from the corn’ as we say in German.
I went through many of these people’s glaciological writings. John Nye’s papers were written in a spirit which I liked most. He followed in his papers best what I called earlier the “red line,” and arrived in a rational way at his results. John Glen was writing in a similar British spirit, more focused on rationally explaining the experiments that led to the creep law of ice, later called Glen’s flow law. This law is — or perhaps was — handled by most glaciologists as if it were one of God’s glaciological commandments. Weertman’s and Kamb’s works were harder to read, but the hardest was Lliboutry. His intention was essentially the same as that of the others whom I have mentioned, using fundamental laws of classical physics and deducing from this knowledge through analysis and computation results that could be verified by observation, but there is in Lliboutry’s work for a newcomer always a touch of irrationalism or a glimpse of unexplainable invention to get there.]
I had tremendous difficulties understanding Lliboutry even though I learnt with time that he was probably even the better glaciologist than John Nye was, but he was absolutely non-understandable to somebody who was inexperienced but had a good understanding of physics via equations.
[Much later, I told jokingly to my students: “Lliboutry shows you in a paper by what differential equations a glaciological problem is described and presents and discusses the solution, and in the conclusion of the paper he casually mentions the boundary conditions which he needed to write down the solution.”]
So you refer to him as the better glaciologist. What would characterize being a good glaciologist in this respect? I’ll let you articulate it if you can.
A good glaciologist is not defined through what my desire was, how I would understand the science. Nye, Kamb, Weertman, Glen, and also Lliboutry formulated the descriptions of equations and deductions of results in a way, which was rather rational, Lliboutry perhaps semi-rational. From these I could derive a rational approach how to get from the physical axioms to a level where equations emerge which are useful to explain the glaciological observations. [The difficulty was to proceed from a basis which allowed derivation of facts already known and derived by the mentioned scientists, but which also indicated paths of extensions, which I was missing in many of the works.]
These equations explain phenomena of all kinds?
Yes, of whatever kind.
Right, whether it be mountain glaciers or sea ice or…
Well, at the beginning it was focused on mountain glaciers.
Okay, which we have not gotten to in our discussion.
In our discussion not, but I encountered them between ‘73 and ‘75, and mostly in a learning, not yet productive fashion. The productivity came, say, from ‘76 to roughly 1980. In the years before I had worked enough to believe that I was ready to derive this systematic deduction of equations from the fundamentals.
So I decided in my head to write a book on glaciology which shows this. This has become the book, published in ‘83, but it was finished in early ‘81. A critical, likely British, reviewer had requested revisions which took me more than a year for amending the text.
Paterson’s book was as early as ‘69. Did you view this as a competitor book or as a supplement, or as something else?
The Patterson book is a beautiful book and it is very good for students who do not have a deeper mathematical physical education. And it provides a wealth of information in all its three editions, all well updated. I think it’s the master book at the lower level, and Paterson is a very good scientist. I haven't met him for many years now. Is Paterson still alive, do you know?
I’m not sure actually.
I do not know, the fourth edition, published by Tuckey and Paterson, has only become available very recently. [Patterson (1924-2013) died only recently; an obituary can be found in the Glaciological Society’s magazine Ice]
Anyway, I wanted to have my book as a competing book for pupils, more advanced in mathematics. I saw it also as a book from which one could learn to go beyond Nye, Kamb, Weertman, and Glen, and others. Glen was an experimentalist, but materials science was more than just what Glen’s law expresses.
Glen had made a very important step in the ‘50s, his Ph.D. thesis was from ‘53, published in the Proceedings of the Royal Society in ‘55. At the same time there was another physics student, called Steinemann, who did similar, in my view better, work than that of Glen in his dissertation. But the dissertation of Steinemann was sitting on the professor’s desk, and was not looked at for approximately five years, as it is told.
So, his thesis came out ‘58, and Glen’s came out ‘55 in the Proceedings of the Royal Society, and that’s likely why the creep law is called Glen’s flow law. It should actually be the Steinemann flow law, and in my draft for the second edition of my book, if I still can do it, I will call it the “Glen-Steinemann” flow law.
The thesis by Steinemann is written in German, [that’s another disadvantage]. He was later professor of physics in Lausanne until about 10 or 15 years ago, and he must be beyond 90 now.
The other thing, as long as we’re on the subject of a foundational approach, I noticed you had a paper in 1977, “The Foundation of Thermodynamics: Its Basic Postulates and Implications. A Review of Modern Thermodynamics.”
I completely forgot to mention this earlier. In the Habilitation procedure in German speaking universities, you have [i] to submit a thesis, [ii] to pass an exam, and [iii] you have to deliver a lecture, the intention of which is more like demonstrating your ability to make your knowledge flow across.
So, it may be a talk of a more didactic nature. For this talk you have to propose three different titles, and the Habilitation committee picks one for presentation — [not so for me!] Professor Parkus in Vienna called me and said, “I’ll be in Zurich then, and then I want to talk to you.”
Then he came to my home and told me, “We are not satisfied with those three topics. You have to do thermodynamics,” and he knew that I had some special interest in thermodynamics because I had offered talks in Vienna (not within this competition for the Habilitation degree), I had given lectures on modern axiomatic approaches to thermodynamics.
I had already done some foundational studies in Cornell in my Ph.D. thesis, with contact to the big scientists in thermodynamics at that time. They were at Johns Hopkins University: C. A. Truesdell [1919-2000], J. Ericksen, I. Müller; at Carnegie Mellon, W. Noll and M. Gurtin, to name just a few.
I had done work during my stay at Cornell, some of which had not entered the dissertation, but Ingo Mueller [later in Paderborn and Berlin], who is German, was reading it and he also sent me his papers which I was carefully studying during my education with Professor Pao.
Professor Pao never touched this, he was scared of this kind of thermodynamics, it was not his topic, [but he was using the simpler Coleman-Noll approach to the entropy inequality (the second law of thermodynamics)]. So this was left to me, but I was really thrilled by the logic behind this approach to thermodynamics. When I returned to Zurich, during nights and spare time, I wrote a few small things and gave lectures; Prof. Parkus was attracted by that. So he gave me the assignment, “Instead of the three topics, you speak about the foundations of thermodynamics.”
As a historian of science, I generally think of the foundations of thermodynamics being in the middle of the 19th century, here at the ETH, especially with Clausius, and Boltzmann. In the late 20th century, what does the topic look like? Is it possible to describe what the differences were from what we learnt in school about the very early work?
Well, [the research on the fundamentals today concentrates on a proper axiomatic structure of the Second Law. Carnot, Kelvin, Clausius, Duhem, Gibbs, Planck, Caratheodory, and others of the 19th and 20th centuries came step by step to irreversibility statements, which had to be generalized because earlier ones were essentially violated by new facts, which were not encountered in the studies before. Today we are more or less still in the same stage, merely at a higher level of complexity. Can we come up with an entropy inequality, paired with axioms on how to handle it, so that all processes which we encounter in our universe are in conformity with these postulates?] Today there are basically two dominant schools of rational thermodynamics. One has been led, say by Truesdell at Johns Hopkins University in Baltimore [C. C. Wang was Truesdell’s student and Ericksen was the other dominant figure at Johns Hopkins at that time], and at Carnegie Mellon University in Pittsburgh [Noll, Coleman, Gurtin, Williams]. [The followers of the postulations introduced by these leading scientists use the entropy inequality in the form called the Clausius-Duhem inequality with a procedure of its exploitation according to Coleman-Noll. In short, they assume that some equations may possess arbitrarily assignable external source terms.]
The second school is led by Ingo Müller [(Technical University Berlin), who was at Johns Hopkins between ‘67-‘68 and ‘70-‘75.] The difference of this school with the first has to do with treating a physical system as an isolated system with no interacting boundaries with the outside world. The Coleman-Noll approach instead regards a system subject to free sources acting from the outside. Truesdell, Coleman, Noll, etc. — I could add 50 names — went this second route and even assume that the source terms from the outside can have any value we please, but Ingo Müller and his followers say, “Such outside sources cannot affect the material behaviour.”
So, the basic axioms in the two branches of thermodynamics are different. The results deduced from the Second Law with those distinct basic axioms are also different, but only in a very subtle way, and only when systems are very complex. So, in Truesdell’s branch of rational thermodynamics from, say, the late ‘50s to, roughly, about the ‘80s or ‘90s of the 20th century, exploitation of the entropy inequality is based on the fact that there are free sources, which may assume any values we please. This condition leads to restrictions of the constitutive relations when the Second Law is exploited.
The other route is to assume that there are no such free sources. So, this is a totally different approach and it makes the second law more flexible, and, as a result, one may deduce more flexible conditions from the equations. When one works with complicated systems where one finds deviations of the mathematical results from observations, Müller’s axioms have always proven to be closer to observations than the others.
[However, followers of the Clausius-Duhem-Coleman-Noll approach in the exploitation of the entropy inequality] refuse to follow Müller’s approach, because, perhaps, the amount of computations to get to the results is a factor of 10, 50, 100 bigger than with the other approach, so scientists shy away from the excessive labor. Thermodynamics in the approach with the Müller axioms can be tremendously hard work to reach results, and I fell in love with those axioms.
I can tell! You must be very enthusiastically in love. And the two branches — it’s very interesting to me, as I say, as a historian of science because too often the historians will just look at the origins of a topic and then let it go when, in fact, there are a lot of fascinating things that continue on, and I was totally unaware of this history myself. So it’s good to have at least a peek into it.
[I should say, I loved the axioms because the fundamental assumptions are less stringent. For instance, in the Coleman-Noll approach the existence of the absolute temperature (the Kelvin temperature) is assumed ab initio even though it is an established fact only in the thermostatics of adiabatic systems (Caratheodory proved it in 1909). In Müller’s approach its existence is a provable result of the analysis or else the concept is generalized (to the coldness function, which may depend on absolute temperature and a time derivative of it).
The lecture, which Professor Parkus wanted me to hold as my Habilitation talk, was this! I worked a full month on its preparation, had no notes, just chalk and an empty blackboard, and impressed the audience — there was a 2 1/2 hours discussion by the (large) audience after it — and Prof. Parkus invited me to publish it in Acta Mechanica in ‘77; it is my second most-quoted paper of my whole career.
I should also add to this that I profited immensely from a two-month stay in July-August ‘74 in the Natural Philosophy Group (Profs. Truesdell, Ericksen, Müller), where I delivered lectures on “thermodynamics of para- and soft-ferromagnetic thermo-elastic solids” and worked under the leadership of Ingo Müller on “thermodynamics of mixtures of relativistic fluids,” which both were published later.
Perhaps I should say how I became so attracted to continuum thermodynamics (CMT). In my first year as graduate student I took a course on elasticity taught by Prof. Constantine Dafermos, a former student of J. Ericksen at Johns Hopkins University. His course was actually an introduction to CMT, in which all essentials of what I tried to explain above were touched on. I was enthusiastic about the subject and how it was taught, and I performed rather well. I wanted to work in such a way in future. An event of equal importance occurred some weeks after the course had ended. I received a parcel in the mail. It contained the book Nonlinear Field Theories of Mechanics by Truesdell & Noll as a present from Prof. Truesdell with a polite dedication. I was extremely touched. For us as students with some inside knowledge, this book had at that time the reputation of a bible of CMT. Prof. Dafermos must have told something to Prof. Truesdell, but they never revealed their secret.]
Okay, so should we continue on then with glaciology or maybe talk a little bit about the environment here that you were working in? You mentioned you were working with Röthlisberger and Iken and some of the others.
Whatever you feel that is important. I might want to say a few words on the early work of mine beyond the “green monster” and spin-offs from it, [like the Poisson ratio issue]. I wrote three other papers in those times, from ‘75, say, to ‘82 — this is the time when they were researched and written. [They are the first manifestation of my desired approach, namely to derive the results from first physical principles using rigorous mathematics.] The topic was the continuum approach embracing glacier flows down steep glaciers to flows of ice masses sitting on solid more or less horizontal ground and, later, floating on the ocean.
Right, like Greenland or —
Like Greenland, Antarctica, and like honey on a breakfast plate. That’s the same thing. I occasionally made in my classes a demonstration: I take a plate and honey and pour it on the middle of the plate, and we watch how it moves. The motion depends on the material properties — Glen’s flow law, it is just a power law — and on the geometry of the basic topography, and on the climate input.
I worked on this, but [after about 1978] I was not alone. It should be said that at that time I was greatly influenced in my theoretical glacier work by two mathematicians, one from Oxford University, Andrew Fowler, and the other Leslie Moreland from the University of East Anglia in Norwich.
I’ve heard of Fowler, but Moreland?
Leslie was an external advisor to Fowler. He was in Norwich and Andrew was in Oxford. But we three were friendly competitors. We developed our models, Andrew Fowler in ‘78 in his PhD dissertation on gravity-driven thin-sheet flows. And Leslie Moreland was already a senior lecturer or a reader at Norwich, [who was doing his own work on ice motion on a bumpy bed, and, with his PhD student Ian Johnson, on gravity-driven flows down inclined or essentially horizontal beds, respectively.] We did very similar things and influenced each other. [I met both for the first time at a conference on ice dynamics in 1978 in Ottawa, Canada, when we recognized how close our work was. I spoke there on the three-author paper — Hutter, Legerer, and Spring — published later, dealing with the transfer of topographic variations to the surface of an ice layer.
Fritz Legerer was mentioned already earlier as host of my visit to deliver lectures on ice plates in 1978 at Memorial University, St. Johns, Canada. He spent a sabbatical at VAW in ‘78-‘79, with the then-PhD student Ueli Spring. The above-mentioned research article was written on how topographic bed variations are transferred to the free surface.]
There is one theory on polythermal ice caps, [by Andrew, which triggered my opposition.] A polythermal ice cap is an ice mass which is composed of cold and temperate ice in different sub-regions. So, such ice masses have domains of cold ice, of which the temperature is below the melting temperature, and patches in which the temperature is exactly at the melting point.
[When I started as a glaciologist more than 40 years ago at VAW, people seemed to believe that glaciers in the Alps were all wholly temperate. However, measurements indicated ice with temperature below the melting point. In polar glaciers, e.g. at Axel Heiberg, vertical borehole temperature measurements showed a kink in the temperature profile at a depth where cold ice changed to temperate ice. Moreover, it was more than a suspicion at that time that Antarctica was grounded at some places on water. So, the concept of polythermal ice was hanging in the air.] [Ed.: At this point Hutter draws diagrams to discuss how polythermal ice manifests itself in different environments; a brief portion of the conversation is not intelligible without the diagrams, which were not kept after the interview.]
In the Alps, [the original concept was that glaciers were only of one sort, temperate, and polar glaciers were also only of one sort, cold. Mathematically, they were treated very much alike, cold ice, as a single constituent medium, i.e., ice of which the temperature would vary according to the atmospheric temperature variation, and temperate ice with the temperature variations being ignored. Because of the strong thermo-mechanical coupling via Glen’s flow law, cold ice models must account for the temperature variations within the ice mass to properly predict the flow. Now, the viscous deformation, principally due to shearing, generates heat within the glacier, which changes the temperature.] So, this is pure ice. It’s material that is made out of ice grains which are cold; so, we have no water. While meso-scopically we may have something to say about how they stick together, [flow field and temperature are determined by Newton’s law and the first law of thermodynamics.]
If the ice is temperate, the viscous heat will generate water, which is preserved in bubbles and may diffuse through the temperate polycrystalline ice along grain boundaries. These water bubbles change the material property, and due to the climate input from above and the geothermal heat from below, the energy contents inside the ice in those specimens vary from position to position. This means those bubbles can grow or they can shrink depending upon whether the environment gets cold or hot. It follows that temperate ice is a mixture of ice and water. So, one is faced with the question, what are the fundamentals, what are the basic physical laws to derive for this kind of ice mass?
Assuming a continuous distribution of the two phases “ice and water,” the mathematical model must be at least a binary mixture of which the thermo-mechanical concept is known, but it is more complicated than for a single constituent material.
This is what Andrew Fowler did in a chapter of his dissertation: he derived the simplest possible mixture theory and found that one could not find, say, a cold patch of an ice mass in an environment of temperate ice or vice versa. His theory showed that in a glacier flowing downhill the surface separating cold ice from temperate ice had to connect the basal surface with the free surface and temperate ice would be on one side and cold ice on the other side of it. [A temperate patch sitting, e.g., as a blob on the basal topography close to the glacier snout would be impossible.] This is what is told by the theory of Andrew Fowler [and the crux is deeply rooted in the hyperbolicity of his moisture flux equation].
Okay, so just trying to explain for the recording what you’ve been illustrating here: you cannot have a patch of colder ice within the solid, well-below-freezing ice, is what you’re saying with that blob there. But that it will always be the case that if you have a glacier on a slope that there will be — could one say altitude-dependent division between the colder ice and —?
[Let me try again. Consider the following Gedanken-experiment. Let’s assume we have a wholly temperate glacier and some demon that can grow a cold ice mass of, say, 100 m3, and cut the same mass of temperate ice out from inside the glacier and replace it by a cold ice mass. So, we now have a blob of cold ice in the otherwise temperate glacier. This blob will in reality adjust its temperature to the temperate environment and eventually all ice may be temperate.
The original theory of Fowler says: this cannot be! This theory would not be able to predict the motion of this composed glacier correctly.] I caught this flaw in the theory of Fowler because Tschoon Röthlisberger and I discussed this and other issues at many coffee breaks. He was not a good mathematician, but he was a fantastic person with a tremendous physical understanding.
He’s just died in 2009, was it?
Yes, he was 86 years old. [In his old age he had tremendous difficulties breathing. That was said to possibly be due to work he had done in Thule, Greenland in the ‘50s when he was handling asbestos for the US army.]
I corrected this theory of Andrew Fowler, and we had friendly correspondences about it. This theory is now used worldwide in those climate models which describe the flow and changes of mass and geometry of ice sheets, ice shelves, etc. One incorporates the polythermicity by dividing the ice mass into temperate and cold sub-domains [with separate sets of differential equations and transition jump conditions across the connecting surface, now known as cold-temperate-transition surface — CTS. This system describes the evolution of the compound ice mass.]
This is contained in my formulation and it was not contained in Andrew’s formulation. [Actually, the difference is small and subtle; ignoring the moisture diffusivity destroyed the thermal jump across the CTS.] I don’t want to make too much fuss of Andrew’s mistake, because it was at the beginning, and with a little bit softer attitude we might have talked and agreed much more quickly than we did.
When was your publication, the corrected…? Did he publish the dissertation version of this? He must have.
Well, Andrew published two joint papers and a third from his thesis in the Proceedings of the Royal Society, together with a post-doctoral fellow at Oxford at that time, Dale Larson. The two papers are very good, they are really very good, and published in ‘78 and ‘80. As a PhD student, I would not have been able to do that. My paper was published ‘82.
Where was that published?
This was in Geophysical and Astrophysical Fluid Dynamics [GAFD]. It could have been published in ‘81 but it was rejected by Cold Regions Science and Technology — my first choice — as being nonsense; I was so furious that I sent the whole correspondence, the reviews, the editor’s letter, and my paper, without changing anything, to Prof. Paul Roberts [FRS] who was then at UCLA. Earlier, he was at Newcastle-upon-Tyne in the UK [and was founder and first chief editor of GAFD]. I received the return letter from Paul after 10 days, which was very fast because letters had to go across the Atlantic. Paul said he would publish it immediately, [and stated in that letter “After an afternoon’s joyful reading…” This story explains why I had about a one or two years’ delay of publication.]
[Looks for the paper on Hutter’s list of publication.] The paper’s title is “A Mathematical Model of Polythermal Glaciers and Ice Sheets”. Okay, there it is.
Yes, but there are several papers later that came out as well. I still had a mistake, which was later corrected by a student of mine, Ralph Greve, [now professor of Theoretical Glaciology at Hokkaido University, Sapporo, Japan]. So the final version is now the version that Ralph Greve wrote.
As I said, Andrew Fowler wrote a hyperbolic moisture flow equation, [I re-instituted its parabolicity and the Clausius-Clapeyron heat flux jump across the CTS correctly. This made the theory harmonize with observations].
Okay. You mentioned you developed your own approach in conversation with Röthlisberger over coffee. Did you have a lot of collaboration with non-mathematical people here at VAW? I notice you have one publication on “Deep Drilling with a Hot Water Jet” with him.
A topic, which was his expertise [and my co-authorship was an expression of appreciation for discussions.] We don’t have other joint papers, but we discussed a lot. The reason is most likely that I was far too mathematical for him and he was interested in fieldwork and problems related with fieldwork much more than I was able to absorb; my knowledge was then very poor in that part. But we had tremendous discussions [and lots of fun with them].
I’m just noticing a couple of other names that come up. There’s Vincent Olunloyo, is it?
Vincent Olusegun Olunloyo, [born 1943, now probably retired, professor of engineering mathematics at the University of Lagos, Nigeria,] was sharing the office room at Cornell University with me as a graduate student, and he had an adviser [Prof G.S.S. Ludford (1928-1986), a mathematician, who was also in my PhD committee], whose specialty was application of singular perturbation techniques to parabolic, elliptic and hyperbolic differential equations. You may recall my work for the master’s degree at Cornell; my very first paper had to do with matched asymptotic expansions of bridges on skew supports.
Matched asymptotic expansions are mathematical approaches of which a further engineering example is a drum. Basically, the drum is a membrane and the membrane has no bending stiffness, it’s like a soap film. This soap film is practically infinitely thin. It exerts no resistance to bending.
But a drum, with non-vanishing thickness of the membrane, exerts such bending stiffness close to supports. [These are mathematical problems exhibiting boundary layer effects], and these are dealt with by singular perturbation techniques. [I wrote a number of such papers with Vincent because we liked the method, which was somewhat “hot” in those days in theoretical engineering. I believe these papers helped him to advance in his career at Lagos University in Nigeria.]
Then you had a student in glaciology in the late ‘70s, Ueli Spring, is that correct?
Yes, the work with Ueli [Swiss dialect, for Ulrich or Uli in German] Spring was significant and became essential much later. [The topic is sub- and intra-glacial water flow. Such flows occur, when ice dammed lakes at the merging point of two temperate glaciers break out, are quickly emptied, usually later in the high summer of the year, through the glacier door and may cause floods in the valley below. The hydrograph of the water discharge in the torrent or river below peaks after 1 to 3 days and drops quickly off afterwards until the ice-dammed lake is emptied. This phenomenon occurs frequently in Switzerland. A much bigger version of such floods occurs regularly in Icelandic so-called jökulhlaups. In Switzerland, mitigation of the floods was Tschoon Röthlisberger’s duty. So, in 1976, after the first paper by Prof. John Nye on the turbulent flow of water through intraglacial channels had appeared, Tschoon and I formulated a research project to the Swiss National Science Foundation (SNF) for a PhD student. Tschoon was referee of Nye’s paper, and we had deep discussions about its content, disagreed with many details and felt compelled to research on it. The money of the proposal was granted and Ueli Spring, a finishing hydraulic engineering student, was hired. Ueli and I worked for three years very intensely together. The intellectual problem was to mathematically formulate (i) the growth of the diameter of the intraglacial pipe due to the melting of the icy wall by the turbulent heat of the water discharge, and (ii) its reduction due to ice overburden pressure. This was essentially already Nye’s concept, but we derived our equations from the three dimensional description and reduced the turbulent water flow to a one dimensional flow along a line in three-dimensional space and averaging the dynamical equations over cross sections. The model was deduced in the same spirit as mechanicians derive beam, shaft, or cord models for fluid flow. However, the energy equation along the curved and twisted but unknown channel was important to obtain the melting rate along the channel wall, which at last determined the evolution of the hydrograph of water at the end of the glacier. In a second part of his dissertation a much simplified version was presented and demonstrated, how Nye’s model of the hydrograph was improved.
We wrote two papers on the topic. The first, describing the theoretical derivation of the general model, was sent to the Royal Society of London, but was quickly rejected with only one review, written in a language, which I sensed to be British. My telephone conversation with the General Secretary of the Royal Society, asking for reconsideration, brought no help. With a delay of approximately two years, that paper was then published in the International Journal of Engineering Science (without revisions!) in ‘82.
The second paper on the hydrograph was submitted to Cold Regions Science and Technology and was accepted after a proper review process.
I never had any response through more than 20 years related to these two papers. I decided to never touch the topic again, but the papers became significant later in 2003. For they were used by (among others) the Awards Committee of the Glaciological Society to bestow the Seligman Crystal Prize upon me. Incidentally, since 1992 VAW had three peoples who won this prize: Tschoon Röthlisberger (‘92), myself (‘03) and Almut Iken (‘12)]
The Seligman Crystal is the major prize…
It’s the major honorary research prize of the International Glaciological Society for research achievement.
Okay. I’m also gathering that your interest in lakes begins around the same time?
Yes, this has to do with the intentions of Professor Vischer, director of VAW, who came from industry. If you work as an engineer in industry — and I know this from a limited number of consulting projects I have done — you are assigned something to do, and when it’s finished you go somewhere else. That was a dominant attitude of his.
So, in ‘75, after the green monster and after I had nothing done yet in fundamental glaciology which would have appeared in journals — the papers and the book Theoretical Glaciology appeared around ‘83 — he said, “I want you to change fields.” He gave me a chance and said, “I want to form a research group on ‘standing water’.” It turned out that he meant physical limnology, the oceanography of lakes.
He really put in a tremendous amount of effort, money, and initially he gave me tremendous freedom. The project operated within a national program [of the SNF] dealing basically with the dynamics of water in lakes. A few Swiss research institutes were encouraged also to submit proposals to the National Science Foundation, and I had to formulate that proposal, but I knew nothing, zero! Just like 3-4 years earlier when he said ‘Make VAW a world center of theoretical glaciology,’ so I was thinking: what’s physical limnology?
I had to learn, and eventually I learned it quite well with substantial help. The first thing which Prof. Vischer did, he said “You take a trip, three weeks roughly, through European institutes where this is studied and you keep your eyes and ears up.” This I did in February ‘76, [and was accompanied by the VAW engineer, Mr. Jürg Trösch, the person who later became the “mediator” of the granting committee of all the projects.]
Were there major centers working on limnology?
Yes, on bio-chemical limnology. If you don’t say “physical limnology,” it’s just like oceanography.
Right, it could mean anything.
Yes, ‘limnology’ can mean anything. It means biochemical, biophysical, and physical processes in general.
[Our assignment was physical limnology, more like oceanography or physical oceanography, or physics of lakes. All these denotations mean similar things and our focus is perhaps best called lake hydrodynamics.
The institutes which we visited were (i) The Institute of Electronic Computing in Engineering in Hannover, Germany, led by Prof. Jürgen Sündermann, actually an institute dealing with physical oceanography or estuarine hydromechanics, (ii) the Institute of Marine Sciences at the Hamburg University, then led by the late Prof. Hansen, (iii) Institute of Theoretical Oceanography of Kiel University, headed by late Prof Wolfgang Kraus (1930-2010), (iv) the Meteorological Institute of Sweden in Stockholm with contacts with Dr. Malin Falkenmark and Dr. Urban Svensson, who had done a dissertation at Imperial College, London, in lake hydrodynamics on turbulent parameterization of vertical momentum transport, (v) the Hydraulics Institute at the University of Lulea in northern Sweden, which was headed by the very young Prof. Lars Bengtsson, who had developed in his dissertation at Lund University, a theory of wind-induced currents in lakes. Before the trip I studied papers done by scientists of these institutions to be prepared for discussions. The information gained by the visits was overwhelming, actually too much to absorb in the short timespans of the visits].
Physical limnology was underrated and Switzerland made an effort to change that. They invested a total of 5 Million Swiss francs to support VAW, a corresponding institute at Ecole Polytechnique Féderale in Lausanne (EPFL), EAWAG, a large institute on environmental sciences and water sciences — sanitary engineering, water purification. It has a few hundred employees, [as well as a group of high school physics teachers in Lugano and Mendrisio, in the Ticino, the Italian part of Switzerland, where there was no University at that time.]
What is EAWAG?
[This is the Eidgenössische Anstalt für Wasserversorgung, Abwasserreinigung und Gewässerschutz (Swiss Federal Institute of Aquatic Science and Technology), situated in Dübendorf, near Zurich — and was annexed to ETH. It’s a huge institute and presently an American lady, Janet Hering, is the director. — At that time it was led by Prof. Werner Stumm, a former Harvard University professor of water chemistry].
So I formulated the proposal, [with inside help]; we received the money, about 500,000 SFR, and Prof. Vischer complemented it with [likely a comparable amount of ETH-money and lots of fringe support through VAW and its personnel.] We bought a vessel for field campaigns, an old fishing vessel from the Baltic Sea.
Here, a few words should be inserted concerning how the support from the SNF was granted. ETH applicants had to submit proposals to the school first, where it was judged by the ETH research committee and from there sent with recommendations to the SNF for final evaluation. This process was not smooth in this case, because Prof. Ziegler from the Mechanics Institute at ETH demanded (not as a committee member!) that my proposal be dropped. Rumors indicated that this interference was wild, and the proposal was granted because Prof Vischer had signed it as PI and I only functioned as something of a ghost writer with no true responsibility.
In fact I was “guilty” for the generation of this complication. In ‘76 I participated in the IUTAM (International Union of Theoretical and Applied Mechanics) congress in Delft and met there Prof. W. Schumann [1927-2014], the head of the mechanics institute in the French language at ETH. Once he was my teacher when I was a senior student. I enthusiastically told him about our intended proposal to SNF on physical limnology. That was obviously a mistake as it triggered negative activities of the professors of the mechanics institute in the German Language (i.e. Prof Ziegler and his colleagues).
Back to the formation of the physical limnology group! The intention, suggested by governmental and school officials was that VAW at ETH and the Hydraulic Institute at EPFL should be responsible for physical limnology about for a time scale of 20 years into the future or longer. My concept for VAW, supported by Prof. Vischer, was to form the group around three people with different, but interacting specialties: (i) A physical modeler, specialized in geophysical fluid dynamics, and primarily dealing with the foundations and the overall concepts. I saw my role there. (ii) A field experimentalist and possibly instrument-builder with good knowledge in planning field campaigns with a research vessel, familiar with electronic instruments like current meters, thermistor chains, acoustic density, current, pressure, sensors (ADCPs), etc., the recovery of data from them, and their graphical disposition for scientific exploitation (i.e., Fast Fourier Transforms (FFT)), cross spectral analyses, etc.). The expertise for this was physical oceanography.] Dipl. Oceanographer Wilfried Horn from Kiel, Germany was hired [with a 5 year contract, but he was promised prospects of a tenured position under adequate performance.] (iii) We wanted an additional specialist who should possess advanced knowledge in numerics. A candidate who could possibly fill this position was an in-house PhD student [later Dr.] Jürg Trösch. [A further PhD student was hired as modeler, (later Dr.) Gabriel Raggio. These people with me formed the core of the physical limnology group.]
When you say numerics, this is digital computation?
Digital, electronic computation, which started to grow at that time: [the computation of initial boundary value problems arising in geophysical fluid mechanics, especially solving eigenvalue problems for barotropic and baroclinic seiches, wind induced currents, and models for seasonal variation of the thermocline. We did time series analyses in ALGOL, not in FORTRAN, at that time.]
So, this was the basic structure, but Prof. Vischer provided much more support through VAW, [which, in the ‘70s, had, perhaps 140 employees, academic, administrative and workshop personnel, from whom, during high active times,] 20 or more people worked, to build instruments, [and field campaign equipment. During campaigns they aided in the preparation and organization of the field observations, logistically a substantial endeavor, and, finally, helping taking measurements. For instance, for more than a year, about 20 people or more were involved in craftsmen activities, from locksmith to fine-mechanics to carpenter and electronic and mechanical instrument-builders and administrative personnel.]
That’s the word. So my group had in the end about 15 people, and we conducted field campaigns in Lakes Zurich, Zug, Lugano, and Constance [of one to two months duration. Not all utilities, such as spherical glass buoys, current meters, thermistor chains, were built by VAW personnel; a part was purchased, another part leased. However, the three meteorological buoys were our own product. They were equipped with Aanderaa anemometers, thermistors, and radiation instruments at 2, 5, and 7 meters above lake level. Some Aanderaa current meters and thermistor chains were also bought from the Aanderaa Company in Norway, but costs were limited for a full coverage for the planned synoptic field campaigns. So, we additionally leased about 25 Aanderaa current meters and thermistor chains from the “Institut für Meereskunde” at the Hamburg University for the August-to-September 1978 Lake Zurich campaign, and some for the Lake of Lugano August-to-September 1979 field campaign.]
In short, these were tremendous programs. During campaigns experimentalists were working up to 15 hours per day in the field, 7 days a week. In the Lake Zurich campaign we used our own vessel, [Jan from the Baltic Sea; and also a smaller, more flexible vessel, including driver, owned by the “Amt für Wasserwirtschaft” in Bern, could be rented. Moreover, during 24-to-48-hour campaigns with continuous measurements two additional teams from the other supporting Institutes complemented our efforts with their own personnel, Dr. Imboden from EAWAG and his team, and Dr. Zimmermann and his team from the Zurich Water Supply Department.
Perhaps a few words should also be spent on the academic support, which we received from outside. First and foremost I should mention Prof. Clifford Mortimer [British and FRS, 1911-2010] [a zoologist with a PhD degree (1936) from Humboldt University, Berlin.] When he was young, he was working at an Institute in the Lake District in England, and then at Loch Earn in Scotland. [During World War II he was influenced by geophysical fluid dynamicists; they “made” him a physical limnologist. He was the world expert on data analysis and interpretation of water motion by graphs from synoptic field campaigns in the Alps, the lakes and lochs in England and Scotland, and the Great Lakes in the USA and Canada. In the ‘50s he had already explained surface and internal seiches in Lake Geneva (and other lakes) and likely also exercised some influence in the 1949-50 synoptic surface seiche experiments in Lake Geneva, done by the Swiss “Amt für Wasserwirtschaft” in Bern, with the nine limnigraph stations, specially made by a local Geneva watchmaker. Professor Mortimer knew the 1975 Swiss state of the art of physical limnology well. In 1963 he went on sabbatical to Madison, Wisconsin, USA, and a year later became founder and first director of the Great Lakes Research Laboratory, later called the WATER (Wisconsin Aquatic Technology and Environmental Research) Institute, and now the School of Freshwater Sciences] in Milwaukee, an institute with several hundred employees.
Is that an independent institute or is it associated with the university?
It’s associated with the University of…
Yes. It’s the University of Wisconsin in Milwaukee.
It should be mentioned here that Prof. Mortimer was well known to the upper academic community in Switzerland when the national program for lake research was started at SNF. He was one member of the committee deciding about the distribution of the five million Swiss francs and observing the performance of the individual groups, which I had mentioned above. I do not know the other members of the committee, but believe that Prof. Vischer was one. This committee appointed Mr. Trösch, then a beginning PhD student at VAW, as “mediator,” who had to put through their assignments to the individual groups. Professor Mortimer’s involvement was of immense value. So too was Mr. Trösch’s, less certainly, but he was important in his role as mediator.
I am under the impression that it was primarily Mr. Trösch’s doing that the Ticino group of gymnasium or high school teachers was funded. Their members, involved in lake research, were Dr. Febo Zamboni as leader; Mr. Giorgio Salvadè, a physicist for electronic computation such as time series and seiche analyses of Lake Lugano; and Dr. Spinedi for electronic computation of wind-induced water motion. The group formed the “Istituto di Physica Terrestre” (Institute of Earth Physics) in Lugano and the members were between 20% and 40% relieved from high school teaching. It was also decided that VAW should closely watch their performance through the five years of funding, a function to which I was assigned by Prof. Vischer. It became my most rewarding activity in the project with a number of nice joint papers on Lake Lugano in peer reviewed journals.
Back to Prof. Mortimer. I regard him as one of the mentors in my and our efforts to become competent in oceanography, and more specifically, as a designer of field experiments for the hydrodynamics of lakes. He gave us tremendous advice early in the program. From ‘80 to ‘82 he spent a few weeks each year during summer as a guest scientist at VAW, shared the office room with me and was working with Mr. W. Horn and drafting personnel — Chrispin Bucher, who drew most of the figures from data of our field campaigns — in exploiting the Lake Zurich current and temperature data of the August-September ‘78 field campaign. Two papers were directly related to this work; they were the result of this collaboration. I was not a co-author of any of them despite the tremendous work I had put in, which I did not view as fair. The papers were co-authored by C. Mortimer & W. Horn (Vierteljahresschrift der Zürcher Naturwissenschaftlichen Gesellschaft,1982) and W. Horn, C. Mortimer, and D. Schwab (Limnology and Oceanography, 1986).
I maintained contact with Prof. Mortimer throughout the years, kept him abreast of my limnological activities as Professor of Mechanics at Darmstadt University of Technology and after retirement until his death in 2010 at the age of more than 99 years.]
Additionally, we had tremendous luck with a second visitor, Professor Lawrence A. Mysak; [he stayed at VAW from August ‘81 to August ‘82, the extreme El Niño year. He has an applied mathematics degree from Harvard University and was at that time professor of applied mathematics and geophysical sciences at the University of British Columbia (UBC) in Vancouver. He specialized in “Waves in the Ocean” (recall his book with Paul LeBlond 1978/’81). Later in 1986, he went to McGill University in Montreal, Canada to the Meteorology and Oceanography Department as founder of the Centre of Climate Research, which became later a leading institute in Canada on climate studies. Professor Mysak also played a major role in the academic community in Canada, and was for two terms President of the Academy of Sciences within the Canadian Royal Society. In January ‘81, then still at UBC, he had applied for a visiting position at VAW during his sabbatical, with the intention to work with Mr. W. Horn on the Lake Zurich field data. On 28 January ‘81 Prof. Vischer asked me via an internal note, and a few days later in a personal discussion, whether this person should be invited for the period of September ‘81 to August ‘82. None of us knew Prof. Mysak personally, but I had some understanding of his work on oceanic waves, which I rated as superb. So, I highly recommended Prof. Mysak as a visitor to the limnology group. An offer was made by Prof. Vischer for a one-year appointment subject to the condition that Prof. Mysak would primarily work with Mr. W. Horn. This was done in September ‘81 for a month after Prof. Mysak’s arrival, but then was terminated because Mr. Horn did not adequately respond to Prof. Mysak’s initiatives. Prof. Mysak subsequently worked more closely with me and the remaining group members.] So I owe my knowledge of physical limnology to a large extent to Profs. Mortimer and Mysak, and they helped tremendously to make the whole project a success.
[I feel compelled to interrupt this description of tremendous cooperation with external academic experts, and to report on the developments of relationships between internal members with activities in physical limnology at VAW. They were less than ideal. For one, there was the assignment of Mr. Trösch as mediator of the funding committee at SNF and simultaneously as a Ph.D. student assigned by Prof. Vischer to develop software for wind-induced wave motion in stratified Lake Zurich, whom I was also assigned to supervise. This ended after about one year with the (granted) request of Mr. Trösch that my supervision assignment be stopped, and that I provide support whenever Mr. Trösch felt he required such help — an arrangement that would never work. Second, one early evening in ‘77, just before I left for home, Prof. Vischer called me to his office, donating a number of bottles of wine to the VAW-limnology group, which he had received when joining a Zurich Rotary Club. I put them by the desk of mine and Mr. Horn’s office and rushed home. The next morning, back in my office, I saw half of them empty. It was pretty clear who might have been involved in the possible evening party. However, nobody would have admitted participation. A very positive act of Prof. Vischer had turned into a painful event. Third, in February ‘78, one day early in the morning, a field campaign in Lake Zug was begun with the entire experimental limnological crew of VAW, and I knew nothing! I was completely stunned. This was perfect mobbing, obviously now with the tolerance of Prof. Vischer, who had approved the campaign but not informed me.
In the ensuing period I retreated somewhat and left the planning and performance of the field campaigns in Lake Zurich (Aug.-Sept . ‘78) and Lake Lugano (Aug.-Sept. ‘79) to Mr. Trösch, who was logistically very talented. Instead I restricted my own activities to the supervision of the three Ph.D. students in fluid mechanics (Kurt Hofer, turbulent jets and plumes in a stratified ambient), intraglacial channel flows (Ueli Spring, Jökulhlaups) and barotropic seiches in long and narrow basins (Gabriel Raggio), who all finished their degrees in late ‘79 and early ‘80 with a number of peer reviewed joint papers later whose writing was pretty much my own doing. Meanwhile, Mr. Trösch did not seem to make progress with his work on wind induced currents in barotropic and baroclinic lake basins; it appeared likely that we would hardly ever get an acceptable result by the end of the five years. And neither was there any hope to have software ready for computations of external and internal seiches in Lake Zurich. These facts exerted tremendous psychological pressure on me, and I began to have thoughts about leaving VAW, but did not have the courage to do so. Discussing the situation with Prof. Mortimer, he suggested a possible solution for the Lake-Zurich-seiche problem.] We should invite Dr. David Schwab from the Great Lakes Research Laboratory (GREL) in Ann Arbor, [a former PhD student of Prof. Mortimer].
At the University of Michigan?
[I do not know if there is an association. It is a NOAA (National Oceanic and Atmospheric Administration) Institute. Prof. Vischer granted Dave’s visit; he came with the program, which he had developed to compute the eigenvalues and eigen modes for some of the American lakes for his PhD dissertation and applied it to the stratified Lakes Zurich and Lugano. It so happened that Prof. Mortimer was also with us when Dave Schwab was here. So, we had a very active period, including a long weekend walk in the environs of Zurich. We kept Dave’s output but had to await the spectral analyses of the ‘78 data. The paper was only published in 1986 and without my co-authorship.]
So on the lower level I had a little bit bad luck, but I had tremendous luck with all the support I got from outside, and it became a success. I’m still working on the results of this research. [My interest in physical limnology, and in particular oscillations of the water in enclosed basins, continued after I had moved to the Darmstadt University of Technology, and also after my retirement.] Perhaps you have looked at the recent books [entitled Physics of Lakes (three volumes), Springer Verlag, co-authored by Profs. Yongqi Wang (Darmstadt), Irina Chubarenko (Kaliningrad) and myself]?
Those have just come out recently?
Yes, this is the work, which I did not have time to touch during my active time in Darmstadt. Many topics treated there went into MSc and PhD dissertations.
I did notice you have several papers with Raggio.
Yes, Gabriel Raggio was hired as a PhD student [within the physical limnology project, and my intention was to make him a specialist in geophysical fluid mechanics. My early and somewhat naïve understanding of physical limnology was barotropic and baroclinic seiche analysis. Numerical, finite element computations of barotropic seiches of Lake Constance had been done in ‘72 by Paul Hamblin and Eckart Hollan from the “Canadian Centre for Inland Waters” (CCIW) and the “Amt für Umweltschutz, Baden Würtemberg,” respectively. Paul Hamblin was the owner of the FE seiche software and willing to let Gabriel Raggio visit him at CCIW in Burlington, Ontario, Canada, were he numerically solved the barotropic seiche-eigenvalue problems for Lakes Zurich and Lugano. This was done in late ‘77 or earlier. Computations were repeated later by Giorgio Salvadè from Ticino with a finite difference eigenvalue solver. The measurements in Lake Zurich with the limnigraphs of the “Amt für Wasserwirtschaft” were done by me and Ch. Bucher, the draftsman in our group in ‘80; and in Lake of Lugano by the Ticino group in ‘81. Results were published in two multi-author papers in the Schweiz. Zeitschr. Für Hydrologie (now Aquatic Sciences).
The idea for Gabriel’s dissertation was triggered by the Chrystal equations, which are spatially one-dimensional equations of motion for water in barotropic processes in long and narrow lakes. The Chrystal equations cannot describe the effects of the rotation of the Earth; so, the challenge in this dissertation was to find a spatially one-dimensional formulation of the barotropic equations of motion, in which the Coriolis effects are accounted for. Such approximate theoretical models are known in beam and shaft theories, and it was in a way also used in our project to derive a theoretical model for intraglacial channel flows. Intellectually, the concept is to define a channel axis and orthogonal to it lake cross-sections with two axes, one across the lake, the other vertical, and to express all field variables in terms of function expansions of the cross-sectional variables—the buzz word is “multi-pole expansions” — and applying the “principle of weighted residuals” to extract a set of spatially one-dimensional field equations. Depending on the number of functions used in the function expansions a hierarchy of simplified models would emerge, which better and better approximate the three-dimensional flows in the long and narrow lakes.
Gabriel is smart and wrote a substantial piece of work as his dissertation, but he left immediately after graduation in January ‘81 without any written papers. [Those were done by me and became four papers, three of which appeared in the Journal of Fluid Mechanics in 1982. Actually, I had sent them to Sir James Lighthill, then provost at University College, London for the Proceedings of the Royal Society, London. He returned them within 10 days saying that they were good but were not quite at the level of the Proceedings of the Royal Society; I should send them to JFM, which I find interesting.] Upon Gabriel’s return to Argentina he was soon drafted to Patagonia during the Falklands War. That was, I think 1980, after Maggie Thatcher’s came in.
‘82, I believe. 1982.
Never mind, he went to Argentina and immediately was drafted.
[Before we continue, let us go back to 1979. The situation at the home front in VAW had not improved, and I was in a rather depressed mood. I needed some time off. In discussions with my wife Barbara we concluded that a sabbatical might be the right thing to look for. Not being eligible for that, it became a leave of absence without pay as Visiting Associate Professor of Mechanics in the Civil Engineering Department at the University of Arizona in Tucson with full, but low payment there. The Dean of Engineering, Prof. Richard Gallaher, who was my teacher of Finite Element Methods at Cornell, immediately hired me and I taught strength of materials at the sophomore level, and elasticity and shell theories in the graduate school. Unfortunately, for unexpected personal reasons, my family could not join me, but I spent a professionally beautiful ten months from mid-August ‘79 to May ‘80 there, and got two new friends, Profs. Sid Yakowitz (1937-1999) and Ference Szidarovsky (Szidar) from the Systems and Industrial Engineering Department, with whom a number of theoretical-numerical papers on glacier and ice sheet dynamics were written.
In 10 double-lectures in the applied mathematics seminar in Tucson on fluid dynamics of large ice masses, I received the chance from the organizer, Prof. Richard Sebass (1936-2000), my former teacher of compressible fluid flows at Cornell, to polish draft chapters of my book Theoretical Glaciology, published later in ‘83. Dick Seebass also invited me to write a review article on glacier flow in Annual Reviews of Fluid Mechanics, which was a write-up of the lectures. It became “Dynamics of Glaciers and Large Ice Masses,” published in ‘82.
I travelled extensively from Tucson to lecture on recent work and work in progress. Places I visited were: the Department of Civil Engineering at Stanford University; the Department of Geology and Geophysics at the California Institute of Technology; the Department of Mechanical Engineering at the University of Calgary, Canada; the Department of Biology at the University of Wisconsin at Milwaukee; the Canadian Centre for Inland Waters (CCIW) in Burlington, Ontario, Canada; and the Department of Theoretical and Applied Mechanics at Cornell University.
These visits were very fruitful. In Milwaukee I was the guest of Prof. Mortimer. We took the occasion to exchange opinions regarding the state of the VAW physical limnology group. Our concern was the lack of expertise in electronic computation of wind-induced currents in enclosed homogeneous or stratified basins [Mr. Trösch’s assignment for his PhD thesis]. With no progress in sight on the home front, we came to the conclusion that a young scientist needed to be hired who is familiar with the required software of water motion in lakes. I was pretty much lost on how to proceed, but Prof. Mortimer said that I should give him a couple of months to suggest something meaningful. The result — much later — was presented to us in the person of a physicist with an MSc but no PhD degree, Gordon Oman from the Physics Department of the University of Minnesota at Duluth. I’ve forgotten the details, but Gordon was hired and worked at VAW from mid-August ‘81 to mid-August ‘82. He had been under the guidance of Prof. Michael Sydor at Duluth, and developed a finite difference program for wind-induced currents in Lake Superior, and eventually delivered results for Lake Zurich (see Physics of Lakes, Vol. 3). We shall say more on this later.
Of the above-mentioned lectures that I gave, the one at CCIW was special. I was prepared for something on lakes, but had forgotten to prepare a lecture on ice dynamics. I had nothing with me on ice, not a single transparency. And the organizer insisted that I hold a lecture on ice. I was given 30 minutes to prepare myself. In a way I was lucky: the 10 lectures on fluid mechanics of large ice masses in Tucson gave me a background. I “only” needed to compress them to something less than 60 minutes. Chalk, eraser, and blackboard were my only tools. The reaction of the audience was positive.
My time off from VAW was extremely enjoyable and fruitful. My depression was gone and the critical issue I had left behind in Zurich was a good way toward being solved with the help of Gordon Oman.]
Could you tell me a little bit about the problems of physical limnology? Is it temperature distribution? Waves?
Physical limnology is the little sister of physical oceanography. [This is to say that the physical principles that are used are essentially the same for both. The difference lies in the size of the objects to which these principles are applied, and this implies differences in the responses to the external driving mechanisms.] The input mechanisms are the meteorological activities, the wind input in the form of momentum, and the solar radiation, [how that arrives at the ocean or lake water surface through the atmospheric processes. Because the lake rests on solid ground of the Earth, the latter also exercises a mechanical and thermal influence on the processes within the lake; these are generally of lesser influence than the atmospheric input. With the air, also gases (Oxygen, Carbon Dioxide, Hydrogen, Deuterium, Tritium, Methane, etc.) and nutrients (Nitrogen, Sulphur, Sulphate, etc.) enter the lake and are transported by the water motion within the water basin. This shows that the hydro-physical processes also chiefly affect the bio-chemical behaviour within the ocean or lake.
The annual variation of the weather leads, via the heat input, to a corresponding annual variation of the temperature within the lake (and the salinity in the ocean through diffusive processes). This leads in temperate regions in winter to nearly homogeneous density distributions of the water and in summer to a stratification with] a warm upper layer, called epilimnion [at 10°-20°C (depending on altitude, latitude and longitude and climate)] and a cold lower layer [of 4°-6°C], called hypolimnion; and then the metalimnion.
Is that the transition?
Yes, [the metalimnion is a transition region, connecting the water with (more or less) constant small density with that of larger density in the hypolimnion. So, the metalimnion density varies from small above to large below. The location with the largest vertical density gradient is denoted as thermocline surface, just the thermocline in brief. In popular simple mathematical models the metalimnion is often collapsed to zero thickness; such models then only consist of epilimnion and hypolimnion, separated by a surface of density jump, the thermocline. This density jump between the lower level of the epilimnion and the upper level of the hypolimnion is very small, in general, Δρ/ρ = 0.001 to 0.01, where ρ is the actual water density and Δρ the density difference between the hypo- and epilimnia. Thermocline depths vary, depending on the radiation environment and size of the basin of still water, in lakes from a few decimeters to meters — 12m in Lake Zurich at high summer. In the ocean the water density varies both with the temperature and salinity. The concept of the thermocline is not helpful from a hydro-physical point of view and the pycnocline separates the light upper-layer water from the heavier one at depth. Pycnoclines in the ocean can be as deep as 200m.
Whereas density differences between epilimnion and hypolimnion during stratification are small, the hydro-mechanical response of a lake basin with homogeneous mass distribution differs substantially from that under stratified conditions in nearly two or more layers.
In a long basin with homogeneous mass distribution and oriented east-west, say, a strong west wind, lasting a day or so, will push all the water towards the east, i.e., all the water from top to bottom. After wind cessation this water will swing back to be reflected at the west-end and so forth. An oscillation will ensue, in which all particles of a water column will move with a velocity.in the same direction. This motion is called barotropic. Contrary to this, the water in a two-layer stratified basin subject to the same wind input will eventually lead to an oscillation in which the epilimnion water moves forth and back, whilst the hypolimnion water moves simultaneously back and forth; the horizontal velocities in the epi- and hypolimnion are synchronized in opposite directions. They are in counter phase, as one says. Moreover, the magnitudes of the particle velocities in the epi- and hypolimnia are not the same, but so related that in each water column the mass flux of the epilimnion water balances that of the hypolimnion water. The wave speed of these so-called baroclinic oscillations is much smaller than that of the barotropic counterpart. Consequently, the periods of the oscillations are longer, depending upon the size of the water basin.]
What kind of a frequency are we talking about?
The periods [of barotropic waves in Lake Zurich] are about 45 minutes from here to the upper lake and back. [For the fundamental internal seiche it is roughly 44 hours.]
I know you’ve dealt a lot with elongated lakes, is that along the length or the width?
It’s along the length. [For a long rectangular basin with constant water depth a simple formula allows approximate evaluation of the periods. Let L, c, T be the basin length, L, the phase speed of the shallow water wave, c, and the period of the standing wave, T. The time for a wave to travel forth and back along the lake is then given by T =
Knowing the phase speed c then allows determination of an approximation of the longitudinal seiche period. The theory of shallow water of constant depth then tells us for barotropic motion cbt = √gh, for baroclinic motion of the two-layer model cbt = √gh′ in which cbt and cbc, are the barotropic and baroclinic shallow water wave speeds, h is the water depth, and h′ is the so-called equivalent depth, defined by
are the depths of the epilimnion and hypolimnion, respectively, h = h1+h2, and Δρ/ρ is the scaled density jump across the thermocline.
We now choose as an example L = 100 km, h = 100 m, h1 = 15 m, h2 = 85 m, Δρ/ρ = 0.001 and then obtain (we set g = 10 m/s2) cbt = 31.6 m/s, Tbt = 1 h 45 min, cbc = 0.29 m/s, Tbc = 191 h 07 min.
These results imply that the shallow water wave speed is in this case approximately a factor 100 larger than the corresponding two-layer baroclinic wave speed for the chosen Δρ/ρ and h1 and h2. Correspondingly the periods stand in the same ratio, Tbc ~100 Tbt.
This explanation is given using a long rectangular basin with very small width, and we have determined the period T by choosing the basin length L to be half the wave length of the oscillation. This choice delivered the fundamental longitudinal mode of the seiches with nodal lines at the boundary points and a hump in the middle of the basin. In principle, higher longitudinal modes are also possible. For constant depth basins, the fundamental and higher-order modes are given as listed in the following table:
|Mode||Period||Wave length||Nodal points|
|1||T = 2L/c||2L||2|
|2||T/2 = 2L/(2c)||2L/2||3|
|3||T3 = 2L/(3c)||2L/3||4|
|n||T/n = 2L/(nc)||2L/n||n+1|
Quite naturally, transverse seiche oscillations are also possible, and they may be excited theoretically for a basin with any width, practically, when the width of the basin is not very small. They are given as in the above table by replacing the length L by the width W.
These longitudinal and transverse seiche solutions are not all that can be constructed for constant-depth basins. Coupling modes where longitudinal and transverse motions interact are briefly touched upon in the book Physics of Lakes, Vol. 1 in Chapter 4. Moreover, for basins with variable water depth the above constructions are only qualitatively correct. The determination of the periods and mode structures — the so-called eigenvalues and eigen-modes — must be constructed by the associated eigenvalue problem; see also Physics of Lakes, Vol. 2.
In the above simple construction it was further assumed that the basin, for which the water motion was analyzed, is referred to an inertial frame. An Earth-fixed coordinate system can for many processes in geophysical fluid dynamics (GFD) not be assumed to be inertial. In such a case, assuming the angular velocity of the Earth to be constant in time and the Earth to be a rigid body, it is easily seen that any body moving with horizontal velocity v (in a observer plane tangential to the Earth) experiences a Coriolis force, pulling the body on the Northern hemisphere to the right when looking in the direction of the velocity v.] Have you heard about Coriolis forces?
[This force is given by FCoriolis = 2mΩsin(φ)v, where m is the mass of the body, Ω is the angular velocity of the Earth and φ (positive on the Northern hemisphere and negative on the Southern hemisphere) is the latitude angle of the position of the moving body. On the equator sinφ=0 and on the poles sinφ=1.] So, if e.g. this rectangular room, in which we now sit, would be a basin and I would be a water particle, then I, as a water particle, would be drifting to the right because of the Coriolis force. When moving initially parallel to the wall, I would sooner or later hit the wall [and would have to slightly correct the orientation of my motion to maintain my speed parallel to the wall. This would continuously happen and result in my motion along the wall around the room in the counter-clockwise direction.]
Vortical motion, yes.
Yes, and on the southern hemisphere, this Coriolis force forces you to the left, in the opposite direction, that’s why you have clockwise rotation. [The longitudinal seiching motion, which we were describing earlier, when effects of the rotation of the Earth are ignored, have now changed to a shore-bounded, counter-clockwise gyre due to the Coriolis force on the northern hemisphere, which, incidentally, is clockwise on the Southern hemisphere. The bounded layer width of this vortical motion can be shown to be given by the so-called Rossby radius of deformation r=c/f and, thus, depends on the shallow water speed c and the Coriolis parameter f=2Ωsinφ of the wave forming the oscillation.
We have seen that for barotropic processes, cbt, is much larger than for baroclinic motions, cbc. With rbt ≈500 km and rbc a factor 0.01 to 0.001 smaller, it is clear that the Coriolis effects must be accounted for barotropic oscillations only for very large basins, like e.g. the Baltic Sea, Caspian Sea and the Great Lakes in North America. Alternatively, for internal oscillations the Coriolis effects must be accounted for only for lakes of intermediate size, such as for many mountainous Alpine lakes (Lakes Geneva, Constance, but generally not, when their widths are just 1-5 km, e.g., Lakes Zurich, Lugano, and smaller lakes). Nevertheless traces of the Coriolis effects can be seen in temperature measurements of such small lakes.]
Internal oscillations, which have largest amplitudes in the metalimnion, are important for the evolution of the bio-chemical-physical processes. [Theory says that the water particles are in these motions displaced by large vertical distances. In Alpine lakes those amplitudes are one to several meters large under common wind input, but may reach during storms up to 15 meters.] In oceans, they may be as large as 200 meters. Often during storms they may become so large that wave breaking may occur. [This wave breaking depends on a competition between the stratification of the water, the vertical density gradient ¦dρ/dz¦ and the vertical velocity shearing ¦du/dz¦, of which, together with ρ and g, a dimensionless ratio is expressed by the Richardson number:
With increasing wind, ¦du/dz¦ will grow and Ri will decrease. Theory shows that for Ri≤1/4, internal waves will break. This wave breaking will mix the epi- and hypolimnion waters. This mingling will bring oxygen- and nutrient-rich upper layer water down to the deeper water. This hypolimnion ventilation depends largely on the momentum input by the wind. Now, since the (dρ/dz)- and ¦du/dz¦-profiles, and, correspondingly, pycnocline depths vary seasonally, so do the minimum Richardson numbers. Such seasonal variations also occur for the internal seiche oscillations. The phyto- and zoo-plankta like to stay at the upper metalimnion. Their depth positions will essentially follow the oscillations of the pycnocline. On the other hand, the daily light climate, to which the plankta and other creatures of the food chain are exposed, will equally vary. This should explain that these two different oscillations — of the seiche and light — will affect the biochemistry of the lake, for each lake individually. Hence the biochemistry cannot be decoupled from the hydrophysics!]
This obviously calls for a merger between biologists and physicists, but this is fairly difficult to achieve. [The reason is that classical limnologists are biologists and biochemists, whose knowledge of mathematical physics is limited. A similar statement holds, conversely with altered sign, for hydrophysicists.]
I failed in this merger, not so much because I didn’t want to work on it, but [because it was agreed at the higher level of VAW and EAWAG (two of the institutions involved in the project near Zurich) that VAW should concentrate on hydrodynamics, while EAWAG should emphasize biochemistry. In this way coming together was naturally avoided. [On the other hand I learned later that large groups must positively combine in joint proposals to funding agencies, in order to stand a chance of being supported. And decision makers of the latter, generally, stay away from funding such large but risky collaborations, simply because the science of physical limnology is too small. Oceanography is more promising!]
That was very fascinating. I’m from Minnesota originally, we’re a state in the United States with a lot of lakes. I really had no idea that the physical processes within them were so interesting and complicated.
[Let me insert here a few words on what I think would be necessary to bring physical limnology a considerable step ahead in the near and somewhat distant future. Restricting considerations to physics alone, today’s computational tools are apt to reproduce or forecast the three-dimensional response to direct wind forcing for fixed stratification kept constant over a time of a few weeks to a month at most. Already post-wind baroclinic water oscillations are not yet adequately reproduced. Part of the reason for this failure is the fact that changes in the internal stability structure through internal wave breaking cannot yet be mathematically described to allow prediction of the seasonal variation of the pycnocline (thermocline). Today’s models attempting simulation of this variation are vertically one-dimensional and applied to a mid-lake water column. These models show that the shadowing effect of possible suspended tracers significantly influences the vertical temperature profile, but, of course, horizontal tracer diffusion is not accounted for in such models. It follows that an adequate computational model should be a spatially fully three-dimensional turbulent model, in which the velocity, temperature and density distributions including the evolution and transport of tracer components (such as gases and nutrients) are concurrently determined by starting from an initial state. Such computations should be carried out over at least one, if not several years. Verification and validation by observation should then demonstrate the suitability and usefulness of the chosen procedure. It is not likely that much of this program will find the required support. It is simply too far reaching in its aims and goals for the present political timescales of decision makers.]
[Let me also insert here some comments pertaining to work on snow, and in particular avalanches, which became later a very active subject of my own research. In the late ‘70s — I’ve forgotten whether it was ‘78 or ‘79 — contacts were established between VAW and the “Eidgenössisches Institut für Schnee- und Lawinen Forschung” (SLF) (Swiss Institute for Snow and Avalanche Research) in Davos. Connections had always existed. VAWE, the precursor institute of VAW, was founded in 1930 and out of that institute SLF was founded in 1936. It devoted its early research primarily to snow properties in the stagnant snow cover. But that focus changed after the appearance of a significant paper on avalanches of snow in 1958. In 1951, a severe winter with a large number of disastrous avalanches triggered intensified work on forces exerted by moving snow on structures and moving trains. The breakthrough of these efforts consisted of a four-part paper in 1958, published in the Schweizerische Bauzeitung (Swiss Civil Engineering Magazine) by a mechanical engineer, A. Voellmy, working at the Swiss Institute of Materials Testing (Eidgenössische Material-Prüfungs Anstalt (EMPA)). The model was a spatially one-dimensional hydraulic model for the flow of snow down inclined planes, or, better, a mass point model for such sliding motions. The resistive force during motion was motivated by the Mohr-Coulomb and hydraulic Manning formulas, and it was stated that, by adjusting the frictional coefficient accordingly, the flows of both flow avalanches and powder snow avalanches could be adequately modeled.
Flow avalanches are dense snow avalanches of granular structure with speeds generally less than 40 m/s and thicknesses of less than approximately 10 meters. These avalanches form from heavy and wet snow covers with weak bottom layers. Powder snow avalanches constitute turbulent density currents of ice or snow particles plus air with gigantic size and speeds of up to 100-150 m/s. They occur less often than flow avalanches and are formed generally from looser and dry snow covers.
In the years following Voellmy’s paper, his model was used and slightly extended worldwide, mostly for use with flow avalanches. It was also recognized that for powder snow avalanches binary turbulent mixture models ought to be developed. With these prerequisites in the background I was commissioned to cooperate with a group of SLF scientists. Dr. W. Good, Dr. B. Salm, Dr. H.-U Gubler, Dr. O Buser, and a number of “lower level” snow scientists, to hold research meetings on avalanche dynamics. We regularly met in the SLF research center at Weissfluhjoch, 2600 m a.s.l., above Davos. The usual arrangement was that I was presenting a lecture on a topic of interest from which interesting discussions quickly emerged. This let me learn a lot of the fundamentals of granular dynamics, which at that time was a statistical theory much like the kinetic theory of gases, but not with an energy conserving collision operator. Different from molecules, encounters of grains have coefficients of restitution smaller than unity. The first such kinetic models are due to S. Savage, a professor at McGill University, Montreal, and Prof. J. T. Jenkins of Cornell University, whom I knew from my PhD times. Besides these theoretical formulations, which became significant in my later avalanche research, we were also thinking about first concepts of theoretical models for powder snow avalanches. It was such a formulation which SLF scientists were hunting for.
During my 10-month stay as Visiting Associate Professor at the University of Arizona in Tucson, I got the “chance” to make the first concrete research steps in this regard. I was asked by Professor Vischer to formulate a research proposal to the SNF for two PhD students on a model of powder snow avalanches; this consent, however, was only reached after I had suggested the topic in a private conversation to him, in which I pointed out the possible similarities of hydraulic flows, mud flows, and snow avalanche processes. Professor Vischer then agreed, and not without saying that “this would be a good occasion for him to document his competence in snow science.” I wrote the proposal in Tucson in November ‘79 and sent it to Professor Vischer and Dr. Salm at SLF. Minimal changes were implemented as a consequence of the criticism. Then it was signed by Profs. Vischer and De Quervain (1915-2007) (the director of SLF at that time). That led to some protest at SLF; Dr. Salm should be the responsible author for SLF and I correspondingly for VAW. In the end it was signed by Prof. Vischer and Dr. Salm, and I was, as earlier, reduced to a ghost writer. After all, Vischer and Salm were schoolmates at ETH and I was nine years younger than both. Modesty and humiliation were what you needed most as a scientist in Switzerland in those days, i.e. as someone pressed between PhD students, for whom you had to provide the otherwise absent supervision to make the bosses’ status justified to society.
More important, however, was and is, in general, the scientific standard of the proposal. It was quickly granted in this case. It stated that to derive the dynamic equations of a binary mixture of suspended ice (snow) particles under turbulent conditions in water (not air), to write a program for the electronic computations of the velocity and particle concentration fields in a chute flow of water and polystyrene particles (slightly heavier than water) down an inclined, water-submerged chute in a large water tank. Both the electronic computations of the model equations and the construction of the experimental site, and, more so, the experimental equipment of the Doppler measuring technique for the particle concentration and velocity distribution within the chute were then new. A physicist, Thomas Scheiwiller, was hired for the theoretical computational part, and an electronic engineer a few years later, who had provided student support in the electronic laboratory of VAW. Felix Herrmann was employed for the experimental part of the proposal. They started with their assignments in 1980 and 1984, respectively. Thomas developed a binary mixture hydraulic model for the turbulent particle-laden fluid chute flow and employed a spectral expansion in the cross section for the numerical solution of the flow. Felix built the electronic Doppler measuring technique to measure the particle distribution, which then was compared with the computational results. The progress of the work was relatively slow, but the important paper (among a few that came out) was published in Annales Geophysicae, and the degrees were granted in ‘85 and ‘90, respectively.]
[There is still an important incident, which exerted a tremendous impact on my life, both emotionally and physically. It happened exactly on my 40th birthday at late afternoon. Now I think, it was intentionally done on that day to maximize its impact. I was called by Prof. Vischer to his office for a discussion. I thought that he would want to know details about the state of the limnology group. Gabriel Raggio’s thesis was finished and four papers were well under way; Mr. Trösch’s thesis was in miserable condition, but Prof. Vischer had let him abandon my supervision assignment and had taken it over himself. Moreover, the work with guests — Prof. Clifford Mortimer, FRS, and Dr. David Schwab on Lake Zurich — was in good order, except that Mr. W. Horn, who was responsible for the time series and spectral analyses, was not proceeding well, and Mortimer and Schwab were waiting for his results. Moreover, the work of the Ticino group (Dr. Febo Zamboni, Mr. Giorgio Salvadè, and Dr. Spinedi), whom I visited roughly biweekly owing to my supervision assignment, was in very good shape; the time series and spectral analyses of the synoptic campaign in ‘79 were presented in a report; papers on internal waves were in preparation, and papers on the surface seiches of Lake Zurich and Lake of Lugano were in good shape.
Instead of being able to report on these facts, Prof. Vischer went forward straightaway with the statement that he was not satisfied with my performance, that my publication rate was too low (at that time it was around 10 peer-reviewed papers per year) and that my work was in general unsatisfactory. Briefly, he wanted me to look for a job outside. This aggression was totally unexpected; I was stunned and shocked and certainly unable to react properly, because I was essentially intellectually paralyzed. I had a tenured job, and, in the annual reports during previous years, I was always the most productive scientist in the institute in terms of publications and outside invited lectures. Professor Vischer refused my request to justify his claim; neither did he in the following days and weeks put his accusation in writing, as I had requested as a condition of resignation. But he had definitely effectively destroyed the rest of my 40th birthday and my attitude to VAW. Only 7 days later he had received an inquiry from Prof. Lawrence Mysak, a world-known physical oceanographer from the University of British Columbia, Vancouver, Canada (at that time, later and still now as emeritus at McGill University, Montreal) for a sabbatical from Aug./Sept. ‘81 to Aug. ‘82, and Professor Vischer had the guts to ask me, whether this person was any good and whether he should offer him such a position. I strongly recommended him, but had an uneasy feeling, because I said that we are not unlike in our scientific working methods and that he had just criticized these a week earlier. Professor Mysak got the job and the assignment to work with Mr. Horn on Lake Zurich. However, the collaboration was terminated after approximately one month, and Lawrence Mysak, Thomas Scheiwiller ,and I had a beautiful 11 months of cooperation on topographic waves in Lake Lugano with a publication in Phil Trans. Royal Soc. London. It turned into a long, still lasting friendship.
In the one and three-quarter years following Professor Vischer’s attack and accusations, I decided to emotionally retreat from VAW and the members of the physical limnology group. I worked daily in my office from 6 am to 6 pm without interruption and devoted this time to finalizing the papers and reports that needed to be done for SNF until 1982. This involved the summary of our work on Lake Zurich and (with the Ticino group) Lake Lugano. As for Lake Zurich, Mr. Trösch was still without success regarding three-dimensional computations of the wind-induced response of the stratified Lake Zurich. And there was certainly no hope to receive final results before the deadline. However, in 1980, Clifford Mortimer was successful in finding a physicist (with an MSc, but no PhD) from the Physics Department of the University of Minnesota at Duluth, who had developed his three-dimensional software integrating the hydrostatic approximation to the stratified hydrodynamic equations: Gordon Oman and his wife arrived in spring ‘81. He became the savior of this part of the Lake Zurich project. He very quickly implemented the digitized Lake Zurich topography and constructed solutions during, and in the days after, storms to the homogeneous and thermally stratified Lake Zurich. Two reports were written — in the German language at Gordon’s request. We sat daily together for quite some weeks to press out the text for two reports, one for barotropic and one for baroclinic responses to impulsive wind scenarios, the second even including the response of a föhnstorm, whose response was recorded during our ‘78 campaign. Gordon had done marvelous work in just about 15 months. He had asked me whether he could use this work as a thesis for a PhD degree, which I wholeheartedly supported. But Professor Vischer outright rejected it when I approached him. He did not even give me time for an explanation. This was a bitter pill for me to report to Gordon, and it was certainly a much bigger disappointment for Gordon. When I had told him, he had to fight against his tears. My respect for Prof. Vischer fell a bit further, but it was already close to the bottom.]
Okay, let’s see here. What should we talk about next? Should we move on to Darmstadt or is there more to talk about before, perhaps with glaciology? Let me ask you, is there any overlap between physical limnology and your work in glaciology? Similar tools, or approaches, or problems?
Well, there are similarities, but they are more on the foundations and less on the level of results. If you know the physics, then you know what basic equations, what basic laws the physics is described by.
So, the fluid dynamics of lakes and fluid dynamics of ice sheets and glaciers is the same thing from the fundamentals. The only difference is that the ice creeps and the water moves faster. One ignores in the description of the ice the acceleration terms. So, Newton’s Law — mass times acceleration equals the forces — reduces to the sum of the forces being equal to zero, and this simplification is the so-called Stokes approximation. One uses essentially the same equations and drops the acceleration terms in the ice problem but keeps these in most analyses of lake water flows. [The analysis goes hardly any further. Water and ice are material-technologically different material descriptions.] So then the studies of glacier ice and water in lakes and the ocean are two totally different sciences. [This is, for example, manifested in material equations relating stress and deformation measures. Even though ice and water in the glaciers and ice sheets and ice shelves are treated as fluids, ice still shows, in its response, many characterizations of solid bodies, and practically never exhibits instabilities or transitions to turbulence], which are frequent in lake physics, oceanography and meteorology. [So, in ice these latter processes need not to be studied.] There’s no turbulence in ice (laughs) except for Terry Hughes. He once wrote an article ‘To see or not to see.
Ah! I interviewed him as well.
He’s a very interesting person.
He is a very interesting person. In fact, since we’ve mentioned Terry Hughes, he keeps emphasizing this idea of thermal convection in ice sheets, and I was wondering if you had –
I think it’s not correct. Terry Hughes is a glaciologist who does the observational, non-mathematical part of the science and seems to formulate his conjecture with application [of cross knowledge from thermal convective instabilities in theoretical fluid mechanics with rather courageous extrapolation and insufficient corroboration whether the mathematical prerequisites of the convective instabilities are satisfied.] He believes in the existence of these vortical structures close to the bed, [but where is the observational evidence?]
Right, right, that’s exactly what he keeps pushing.
It cannot exist. [Let me try to explain. I guess that his basis is the Rayleigh-Bénard problem. To explain it, consider a fluid between two horizontal plains heated from the lower and cooled from the upper boundary. When the temperature difference between the plates is small, the fluid is at rest and the heat is transported from the lower to the upper boundary by conduction. If this ΔT reaches a threshold value, ΔTth, the mathematics of the convection sets in, which, because of mass balance, must merge into a circulation flow, which can manifest itself in rolls, hexagonal or more complicated cells. I think something like this is Terry Hughes’s model. Now, for such convective cells to form, the mathematics of the problem shows you that the thermal expansion of the fluid must be so large that the gravity force changes by an amount sufficient to trigger convection flow that sets in at ΔTth . I do not know a process in glacier ice which would achieve this. Bénard observed this in 1900 and Lord Rayleigh did its mathematical description in 1916, with many extensions later. Of course, the flow configurations at the base of glaciers are not simple shear, but that is not the key point. Buoyancy is too low to generate what Terry Hughes wants to see and can obviously not see.]
There are, however, some processes in which one can get some lifting. For instance, on a soft glacier bed which consists of water-saturated grains, not rock, the sliding motion of the glacier sole can grab individual grains from the sediment layer and incorporate these grains in the ice. These individual particles are subject to a lifting force analogous to the lifting force of an airplane wing, just more complex. Grains exposed to these forces will move very slowly upward, [but this motion will eventually die, because the flow will symmetrize farther away from the glacier bed, and, thus, the lifting force will die. I have seen photos from near-basal cliffs of ice, of which a few meters were contaminated by sediment grains; yet there is not a trace of evidence of vortical (rotational) motion.]
I wanted to do this problem once, but I didn’t have time. It is a non-trivial problem.
[I have since formulated the general equations to this. It is a mixture theory in which soft beds are treated as a sediment layer of a certain thickness, and hard beds are just an interface. Below the soft layer one assumes solid rock, above one has ice and water, enough water where the heat generated by sliding makes it slip better.]
Is this associated with your work on granule motion? We haven’t gotten to that yet.
[In principle, yes, but not for me.] I studied granular motion in the fast mode, such as landslides, debris flows, and avalanches as illustrated already earlier. The problem, which we discuss here, is slow flow, but I doubt that it can be dealt with in the Stokes approximation. [The lift force “of aerodynamic origin” can be traced back to the convective acceleration terms. Other forces can act as lifting elements as well; in short this problem is hydro-dynamically not at all trivial. I have not seen a paper, in which this slow motion of the particles into the ice is treated.]
One final question on physical limnology. You mentioned that it’s of strong interest in biological processes. But when Vischer suggested the problem, were there other practical interests in the subject aside from that?
Yes, and I was involved in one problem of practical relevance: sewage disposal into lakes. [Untreated sewage water is no longer allowed to be deposited; it may just be water with a different chemical composition. For simplicity we refer here simply to sewage disposal. Fluid-mechanically, the different density is the dominant factor determining the kind of local flow. If the sewage water is heavier than the surrounding lake water, a density current will form, and when this water is entering the lake by an open channel flow, it will move down the shore slope, roughly along the direction of steepest descent, because it is heavier than the surrounding water. During this motion some surrounding water will be entrained into the (turbulent) density current until its density will be the same as that of the surrounding water. At this level the sewage water will layer into the lake and spread out to cover eventually the entire extent of the lake at that depth. This level should not be too deep in the hypolimnion as the sewage might contribute to the oxygen deficit there.
If the sewage water has a smaller density than the surrounding water, the sewage pipe will be guided to a depth where its density is smaller than that of the surrounding water. In this case, one wishes to avoid the sewage water to rise to the surface. It should be below the level where people bathe, generally below the thermocline. This means that the exit cross section of the sewage pipe should lie below the thermocline or pycnocline, respectively.
A second practical problem to address is the withdrawal of drinking water from a lake.] Many communities draw their drinking water from lakes, so one has to select the position from which the water is drawn correctly such that no sediment, no undesired chemicals, say from other deposits, are drawn with the water, so that no additional efforts may have to be done in the purifying process. Of course, the goal is to do the minimum for economic reasons.
Or, if an oil accident from a pipeline occurs, as was the case once in the eastern part of Lake Constance. When an oil pipeline breaks, a growing oil spill is formed that spreads on the lake. [For such situations communities subject to potential danger are often well prepared. Nevertheless, it is advantageous to know rules of thumb: how fast the oil spill spreads and what scenarios one best applies in optimizing the procedures for such accidents. Generally, this requires individual general studies of oil dispersion prior to such accidents.]
Similar problems are encountered in environmental protection problems of the atmosphere, e.g. the spreading of pollutants from chimney exhaustion gas. [In this case the problem is dispersion of a turbulent jet into an external velocity field.]
It’s probably an opportune time to ask about your consulting activities with the Swiss government, what the nature of those are or were.
I have had virtually none. I was in a subordinate position. I was not in a position to exercise an influence. These were given to me by the director, and then I had to do them or I could avoid doing them. In one case, I did not manage anything because it was just outside my expertise.
But I had done lots of consulting work related to avalanche dynamics. There was an unofficial consultancy assignment of VAW to the Snow and Avalanche Research Institute [SLF] in the ‘80s, [where mutual help was given between these two governmental units, and there existed a semi-formal agreement for these. I had regular contacts with them — and I had recently privately been a consultant again to researchers of the same institute for the development of new avalanche models. Such contacts are useful as there are frequently still rather large deviations between predictions of today’s more sophisticated avalanche models than those of, say, the 80s, and observed avalanche deposits.
In lake and ice research I did not have consultancy assignments at VAW. In those fields collaborations with scientific specialists are performed by the usual international interaction, but all my work has centered on different aggregation forms of H2O.]
Yes. We haven’t talked a lot about glaciers and avalanches — we did with Fowler’s work.
Well, my later Darmstadt work was concentrating more on ice sheets and ice shelves. It was also there where I got contact with the climate people.
Oh, very interesting. Should we move on to Darmstadt then?
I went to Darmstadt in ‘87, and there I became a professor of fluid mechanics and continuum mechanics.
Was that an opportunity that arose, or was there any other reason that you went to Darmstadt?
[Well, I had attempted and passed the habilitation degree (in ‘76) at the Technical University in Vienna, and running for such a degree means in many European countries automatically that one will eventually run for a professorial position. From ‘76 onwards I applied to almost every position that was announced in the field of mechanics in German-speaking Europe and did not get anything, and this is most likely again related to my experience here in mechanics, probably, most certainly so, because in such competitions some committee members expressed this explicitly. More than one such member said confidentially that someone who is in conflict with Prof. Ziegler will never get a position in mechanics in Germany.]
On the other hand, I never got a sufficient amount of support from Profs. Yih-Hsing Pao or Edmund Cranch; and no other professors from Cornell University could do anything in Europe.
Your entire time here before 1987, then, your position — you mentioned on your CV that it was as a mathematician employed by Prof. Vischer.
[Yes, that was the description of my job, which I had suggested to Professor Vischer (as a want-to-be mathematician) when I was first hired. I was not part of any section of VAW (Hydraulics, Hydrology and Glaciology), but directly responsible to him, first with the assignment of “mathematical glaciology” (that was later amply fulfilled with the book Theoretical Glaciology). Then, about ‘75-‘76, he commissioned me to form a new group of VAW, researching Physical Limnology, whilst the PhD student supervision in glaciology for Ueli Spring continued and supervision of Ph.D. students in sewage disposal in lakes (Kurt Hofer, ‘77-‘79) and lake oscillations (Gabriel Raggio, ‘77-‘79) were newly commissioned. I was assigned as head of the new and growing group, a position which Mr. Trösch took on during my absence at the University of Arizona in Tucson during 10 months in ‘79-‘80. Upon my return the PhD student supervision responsibility was extended to include Thomas Scheiwiller (‘81-‘85) and Felix Herrmann, ~‘84-‘90) in the VAW-SLF proposal to SNF on powder snow avalanches without changing my function as head of the Physical Limnology Group, but extending my responsibility to rapid flow of granular materials. After the disastrous event at my 40th birthday (Jan. 22, ‘81) nothing officially happened, but later, after I had written all the papers and the final report of the activities on the lake research to SNF, in ‘82, I was removed in that function and appointed as Head of Research for the entire VAW, but with absolutely no decision-making capacity, thus it was a phony position. I was told that I had to look for support outside, provided the proposal was granted by the director and travel support would no longer be available. This was isolation in a golden cage without escape. I kept that position until I found a position as professor of fluid mechanics at “Technische Hochschule Darmstadt” (now Darmstadt University of Technology), Germany, as of October ‘87)]. But I had already been asked to change my emphasis into granular materials, when I was in Tucson, i.e. avalanche dynamics, where I wrote a proposal on avalanche dynamics that was founded. That is when my avalanche research started.
When was this?
1979-‘80, but the first work on snow avalanches came out in ‘82, first as a VAW-report by Thomas. Scheiwiller and me, and several years later as a number of formal peer reviewed papers.
Okay, you were at Tucson…
Yes, I took a leave of absence to Tucson at the University of Arizona, where I taught Mechanics, Statics, Strength of Materials, and Plates and Shells. This means leave without pay; however, I had a very small but continuous salary ($22,000 in 10 months). Professor Vischer had refused to count this as a sabbatical.]
Yes, for 10 months. There I formulated for this institute a research proposal for avalanche dynamics. Powder snow avalanches, in particular. This topic was very hot at that time. These are a form of boundary layer density currents, a mixture of air and snow or ice particles. By contrast, flow avalanches constitute dense layers of snowballs with very little air in it. They can be treated as granular materials. However, a powder snow avalanche is a mixture of suspended ice particles and air, perhaps best be called a particle laden fluid material and the fluid is actually air. It has similarities with temperate ice, which is a mixture of ice and water. So the fundamental equations are the same. [Both models can be derived from mixture concepts, but that is the only analogy.]
I see. But then you do go to Darmstadt.
I went to Darmstadt, but only as late as Fall 1987, yes, and I stopped submitting any applications for professorial jobs, [since it did not make any sense being in second or third or even lower position in the application ranking. My wife said I should stop being cannon fodder.]
Had you been doing any teaching here?
Yes, I did teach at the Technical University in Vienna after I had gotten my habilitation degree there, formally in ‘76. [With such a degree teaching is compulsory in the form of at least one 3-credit hour course per academic year. In Zurich, however, teaching without an individual assignment was not possible, and this assignment could only come from Prof. Vischer. So for teaching I needed a second habilitation degree from ETH.
So, it’s a long story how I got a second habilitation degree at ETH in ‘Theoretical Glaciology and Limnology’ in ‘82. It tells a lot about the stubbornness and inflexibility of the academic system, at least at ETH Zurich in the ‘80s.
In April ‘78 Prof. Ziegler from the Mechanics Institute retired and became an Emeritus. This was (erroneously!) the motive for me to try to transfer my habilitation degree in Mechanics from Vienna to Zurich. This was an often pursued procedure and it is generally straightforward. When successful, it meant that I could transfer my regular teaching (which was done in block form) from Vienna to Zurich. So, on April 24, ‘78 I sent a parcel with all the necessary documents (among those copies of my papers) to the President of ETH, Prof. Ursprung, applying for this transfer. His office acknowledged its receipt with a statement that the Department of Mechanical Engineering (the Institute of Mechanics was a unit of it) would handle the case, and that I would hear from them in due time. Fine! On April 24 ‘79, ‘80, ‘81 I was periodically asking the President, Prof. Ursprung, by letter about the state of my application, with no response, perhaps because my letters were not sent as registered mail. On 24 April ‘82 I wanted to know it clearly, and wrote to Prof. Ursprung my fourth annual reminder concerning the state of the transfer of my habilitation degree from Vienna to ETH, Zurich, and made it rather explicit that I would take further steps, if needed as far as Lausanne — where the Federal Court is located, everybody in Switzerland would understand what the expression meant.]
The response came extremely quickly after a few days with the request to submit my documents again, [apparently because they could no longer be found, updated, of course, and to be sent to the committee led by Prof. Thomann, a fluid dynamicist. It was pretty clear that Prof. M. Sayir, the immediate successor of Prof. Ziegler, did not have the guts to confront me. Professor Ursprung’s office indicated it would finalize the case within 2 to 3 months. I sent my registered parcel of documents, which, apart from the papers, written since ‘78, contained the finished draft of Theoretical Glaciology, submitted to Reidel Publishing Company. I asked in my accompanying letter “to acknowledge receipt of the documents,” which was done in a letter by Prof. Thomann with the sole statement, “This is to acknowledge receipt of 5.1 kg paper.” (I cannot judge how scornful this sounds in English; it does so very much in German.)
It took only about a month, when the presidential office organized a meeting between President Ursprung, his personal assistant or advisor, Professor Thomann, and I. It took place in the President’s office. To report on the discussion is not possible after more than 30 years. The upshot is that Prof. Thomann tried to prove that of the more than 50 peer reviewed papers, of which I was author or coauthor, those with coauthors had to be discounted, as had all those involving electrodynamics, glaciology, hydrology, and avalanches; they represent by no means mechanics, as Prof. Thomann said. He concluded that only two or three papers would fit the definition of mechanics; furthermore they were of far too low quality for a habilitation degree in mechanics at ETH. From the looks on Prof. Ursprung and his assistants’ and advisors’ faces and reactions, I sensed that he and they obviously must have done their own study of my qualification as a university-level scientist. I felt secure.
Professor Ursprung wanted me to withdraw the transfer of the habilitation in Mechanics from Vienna to Zurich. I would fail anyway, and that would be a dark spot on my vita. I indicated I would no longer care. After a somewhat painful discussion back and forth, I indicated agreement and the President promised to make my attempt smooth for a habilitation in Theoretical Physics.] I objected, because I am definitely not a theoretical physicist and suggested that I would accept a new habilitation in Geophysics. [I did not have to write a letter of withdrawal for my application in mechanics, but I wanted to submit the 5.1-kg paper with the book Theoretical Glaciology as a habilitation thesis. Professor Stefan Müller of the Institute of Geophysics was the decisive member of the habilitation committee. I was informed not to send Theoretical Glaciology (and, perhaps, other papers, I’ve forgotten). My publication record and my international scientific reputation would be sufficient for me to pass the habilitation degree. ETH’s bylaw had a paragraph stating that a habilitation degree could be granted on the basis of merits, if certain conditions are satisfied (which apparently was so in my case). I only had to present a lecture in the Department of Natural Sciences to demonstrate my teaching ability. I offered “Can one-dimensional models tell the glaciologist and limnologist acceptable inferences?” (translated from German). With this, my ordeal of transferring my “venia legend” from Vienna to Zurich came to a positive end. It still had a somewhat bitter-sweet postscript: which field of science should my habilitation represent? Rumors, which reached me and whose source I could not identify, said that the words “Mechanics” and ‘Geo-something’ are not allowed to enter the description. The committee decided for “Theoretical Glaciology and Limnology.” This is what they put down as my field of representation on the Habilitation Diploma.
1. I am dumb and especially talented in the same school, depending in which department you are (laughs).
2. I lost pretty much all respect for the academic system at ETH.
3. I needed to leave here as soon as I could.]
I taught theoretical glaciology (a two semester 3-credit hour course) (and for a while also wave dynamics in lakes) from 1983 to 2011, 28 years. My glaciology courses were the only ones (of all glaciology teachers) accepted as one of the emphasis subjects in the curricula of mathematics and physics, and I could educate a fair number of students who became respected glaciologists. The courses were mostly offered in block form and amounted to ~40 lectures per semester.
All right. So I keep trying to get to Darmstadt. I think we probably better move there.
Well, should we make a break of 10 minutes? I’m a little bit tired. [Break]
Okay, we are recoding again, after a brief break, with Kolumban Hutter. So we were just at the point of your moving to Darmstadt. So why don’t we talk a little bit about your move there, and then we can talk about your work there. While the recording was off, we were talking especially about ice sheets and ice shelves and some of your work and your students’ work and collaborations in that area, which is of great interest to me. And so we’ll want to talk about that on the recording as well. So Darmstadt.
Yes, I eventually went to Darmstadt. However, and roughly in 1982, Prof. Vischer asked me to change my field again and to go into sedimentology, [another field I knew nothing about], and I refused. I had changed my fields so many times, that I was not willing to do it again because I had the feeling that I needed to consolidate the research I had begun.
As a result I was taken away from the physical limnology group. [Professor Vischer coined a new working assignment for me “Head of Research” for the entire VAW, but this unit had only a “head” but no “body.”] Somehow I was put in a golden cage, in which I could do as I wished. If I was able to get the research money from SNF, I was free to do that but all support [for travel and conferences] was cut-off.
I was extremely fortunate in this deadlocked situation for two reasons:
1. Professor Sidney (Sid) Yakowitz from Tucson, his colleague Ference Szidarowski (Szidar) from Budapest, and I had just finished a research proposal to the U.S. NSF. Sid as principal investigator, and Szidar and I as foreign co-investigators, on the numerical investigation of the ice sheet equations. That was granted, so I could travel to Tucson and to a limited number of conferences with U.S. NSF finances. Out of this joint collaboration grew a number of the early papers on the steady behavior of ice sheets that exerted considerable impact among early climatologists. We three met at least three times between 1982 and 1987.
2. It so happened that Prof. Huw Davies, (a theoretical meteorologist) from the “Institute of Atmospheric Sciences” at ETHZ had in early ‘84 too many finishing students to supervise for their diploma (roughly equivalent to an MSc degree). He asked me to help him out with just one student: Thomas Stocker. With a habilitation degree, I was eligible for this. I assigned Thomas an MSc-thesis on “Topographic Waves in Enclosed Lake Basins.” This became a very successful piece of work, for he captured ETH-support for his PhD finances on the same topic. Thomas finished with distinction and a few papers, and, together with me, a book on the same topic.
3. In late ‘85 Prof. Stuart Savage from the Department of Civil Engineering, McGill University, Montreal, applied for a supported visit, during his sabbatical from fall ‘86 to fall ‘87 (he then stayed until fall ‘88, when I was already in Darmstadt). We worked out a granular avalanche model that was already started in ‘85. It was submitted to J. Fluid Mechanics in fall ‘86, but needed almost 3 years for revisions until it appeared in print, one of my hardest-to-publish papers; it became the most famous one. The paper made us famous among specialists of avalanche dynamics and became the “Savage-Hutter avalanche model.” I still believe that its first draft was better.
Thomas Stocker graduated with his PhD in ‘87, just about when I went to Darmstadt. He stayed for about 6 months with Ted Johnson in the Department of Applied Mathematics at University College London, then spent 2 years at McGill University in Montreal with Prof. Lawrence Mysak and a further two years with Prof. Wallace Broecker at Columbia University to build up his expertise in climatology, and then became successor of Prof. Hans Oeschger as Professor of Climate and Environmental Physics at the University of Bern. He occupies the co-chair of the working group 1 at the International Panel on Climate Change (IPCC) since 2008.
I have already mentioned earlier that I had applied to almost every professional position that was announced in the field of mechanics and did not get any offer, certainly because of my difficulties with Professor Ziegler in the ‘60s or his successor, Sayir, later. There was one exception to this attitude: the succession of Prof. Dr. Ernst Becker. He was, in my view, the leading fluid (and solid) mechanician in Germany, and tremendous prestige was attached to his research group. He died of cancer after a long illness in Nov. ‘84.] Soon afterwards his position was announced, which I could not resist to apply to.
You mentioned Becker?
Yes. [Professor Ernst Becker (1929-1984) was a mathematical physicist with education from Darmstadt (diploma in mathematics) and Göttingen (Dr. rer. nat.). He had done significant research in gas and fluid dynamics, rheology and had written well accepted lecture notes in book form on engineering fluid mechanics, continuum mechanics (with W. Bürger), and thermodynamics. He had a number of successful PhD students who became professors of mechanics in Germany. In such applications it is the custom in Germany that universities submit a list of the best candidates in ranking order to the “minister of education” of the Land, (i.e., state, Hessen in this case), who picks freely from this list. With another candidate I was number 3 or 4 (ex aequo = of equal standard). To cut a long story short, I was offered the position in December ‘86 and knew from word of mouth that the first two candidates had declined the offer long before.]
It was a modest position both financially and equipment-wise with only one secretary for three professors and three “hard money” positions [and a laboratory and workshop personnel that had to be shared with 7 other professors. Professor Vischer had at that time more than 120 personnel.] Anything beyond this had to be found from industrial and funding agencies and consisted of “soft money”.
[Because my wife with children were not willing to come to Darmstadt, I asked for a meeting with Prof. Vischer to discuss whether this full professor position in Darmstadt could be traded into a titular professorship at ETH. His answer was: “Das werde ich zu verhindern wissen” (“I would certainly know how to prevent this”). So, I became a professor of mechanics who was regularly commuting between Darmstadt and Zurich. This was the end of formal politeness.]
However, it was a tremendously challenging position and I could definitely continue with the work I had done at VAW, because otherwise I would not have been offered the position in fluid mechanics. My work was recognized as a branch of fluid mechanics, and I just decided to do whatever I had done before: electrodynamics, I could refresh that. I was working on ice, on lakes, and on avalanches (rapid granular flows), electrodynamics and thermodynamics. For these fields I tried to find money, and it worked very well. A proposal on ice research was the first, which I wrote and submitted to the German Research Foundation, Deutsche Forschungsgemein-schaft (DFG), because in the first 100 days of a new job whatever you submit is accepted, if it sounds more or less convincing — it is the same as with politicians.
This is exactly what I did in the first month. I formulated a proposal and I got the money for a PhD student whose name is Reinhard Calov. He came from the Meteorological Institute at Hamburg University, had done his diploma thesis there. With him I started ice sheet dynamics. [These equations are now referred to as “shallow ice approximation” (SIA). These equations were “in the air” 30 years ago and were deduced with more or less rigorous techniques from the equations of general three-dimensional fluid dynamics. Fowler, Herterich, Hutter, Huybrechts, and Morland knew them, but only Fowler, Hutter, and Morland (in alphabetic order) had systematic derivations.]
One early article I wanted to ask about, with F. M. Williams on “The Response of Unconfined Ice Shelves to Climactic Conditions.” Was this an early manifestation of this interest?
Yes. I met Mary Williams in ‘73 at the Canadian Congress of Applied Mechanics in Montreal, Canada, [when she was a PhD student of mathematics at Simon Frazer University, Burnaby, BC, Canada] and worked on an extension of my “green monster.” I was impressed by that and we had some discussions at the conference and later through letters, and out of this came then her PhD thesis in applied mathematics and her later paper. [If I remember correctly, Mary extended my visco-elastic floating ice plate model to include aging in the relaxation functions, an effect which I had ignored.
Our joint papers on the climate response of floating ice plates was motivated by matched asymptotic techniques, used by Grigoriyan and Shumsky from Moscow, which we simplified by selecting function dependences across the ice layer thickness. The paper was very much motivated by Weertman’s simple and first-ever ice shelf paper subject to outside climate. I recall that it received a belittling review by Bob Thomas and was then buried in a mechanics journal, Acta Mechanica. Such reactions by glaciologists often occurred with my somewhat mathematically motivated works.]
Okay, now we’re onto Reinhard Calov.
[Reinhard Calov was hired in 1990 as a PhD student and supported from my first granted DFG-proposal to construct three-dimensional solutions of the SIA-approximated ice-sheet equations, applied to the Greenland ice sheet. We were interested in finding Greenland’s response to an external climate driving, to the effects exerted by the geothermal heat, the deformation and temperature variation of the lithosphere, in short a workable model for the growth and retreat of the Greenland Ice Sheet to climate variations.
These equations were published in the late ‘70s and early ‘80s when I still was a member of VAW, and spatially two-dimensional solutions were promised in the granted U.S. NSF proposal, formulated by Sidney Yakowitz, Ference Szidarovski and myself. Sid and Szidar developed the numerical finite difference technique for the electronic computation of plane two-dimensional and rotationally symmetric steady ice sheet flow on solid ground. Subsequently, I constructed a large number of numerical solutions on Sid’s machine for varied equation parameters, plotted these and discussed them from a glaciological point of view. The three author paper appeared in J. Comp. Phys., J. Glaciology and Water Resources Research These results of 1984, ‘85, ‘87 were known to us when Reinhard’s PhD work began.]
And I see here that these are numerical studies, we’re more heavily into computation…
[Computations were not heavy when measured by present computational tools, but they were time consuming. I spent day after day at Sid’s machine; I forget how many weeks; it was boring, but important as can be seen in our paper published in the J. Glaciology, and we received tremendous help from the drafts personnel at VAW.]
Do you remember the sort of machine that it was?
(A PDF-11.) I spent two months in Tucson after I was visiting professor there in ‘85, and I wrote the manuscript, which became the first significant paper on numerical prediction of ice sheet flow. Our understanding was rather naïve, but the model was the first attempt to obtain numerically a rough physical understanding. With Reinhard Calov I wanted to attack the three-dimensional problem. He was very well educated with his three-year education at the Max Planck Institute in Hamburg under the guidance of Klaus Herterich, who was another glaciologist who started as a theoretical physicist. [He later became professor at the University of Bremen but took early retirement.
The arrangement with Klaus Herterich was very convenient for all. Klaus had a finishing diploma student but no money for continuation, and I had the latter; Reinhard completed his PhD degree in four years, worked hard and produced the first results on how Greenland would react to the climate variations through the Eemian intermission and the Pleistocene and Holocene to the present.] He started with the computations 175,000 years before present (BP), [about at the minimum of the Illinoisan Ice Age with the topography of today’s ice sheet and a steady state temperature distribution, the present topography, a geothermal heat flow of 42 x 10-3 Watts and a climate input thought to be reasonable. We needed about 100,000 years of integration prior to formal computations just to reach these steady initial conditions, from which computations could be started. The electronic file of the basal topography was given to us by Dr. Anne Letreguilly from the “Laboratoire de Glaciologie et Géophysique de L’ Environnement” (LGGE) in Grenoble.]
Is this in parallel to things that had been going on with, say, the associates of Hans Oerlemans, or is it not quite comparable?
It is comparable, but they had at that time a simpler [depth integrated model of the governing equations], and their model was in certain respects climatologically and in material-science behavior a little more general. [For instance, their climate model was not based on the positive degree-day model but energy considerations at the surface.]
Were you working with precipitation and accumulation on the ice sheets?
Yes, [the surface temperature was parameterized with the time series of Vostok ice core temperatures, as published by Barnola et al. in Nature in 1987, but simplified by us], and the accumulation/ablation rate at the surface was parameterized by Roger Braithwaite’s positive degree day model [as improved by Niels Reeh]. The temperature distribution of a 5 km thick lithosphere layer was equally determined by solving the heat equation in the rock subject to geothermal heat flow of 42 x 10-3 Watts [applied at the lower boundary of the rock layer. The driving “Vostok temperature data” were incorporated in a parameterization of the surface temperature accounting for its dependence on latitude and altitude.
The model was restricted, or incomplete, by the fact that only the cold ice was modeled. Wherever this model generated temperatures above ice melting, this temperature was artificially put equal to the melting temperature, an obvious imperfection.]
For you, was this motivated by the growing interest in the response of ice sheets to anthropological climate change?
It was definitely triggered by that. [I was interested in whether computational methods like the SIA-equations could be used to estimate the melting rate of the Earth’s ice sheets, and, thus, their contribution to the sea level rise in the coming years. My interest in the anthropogenic climate change received a tremendous impetus in early ‘88 by an invitation by the Alexander von Humboldt foundation (AvH-F) of Germany to present the main lecture at the 1988 annual meeting of the “fellows of the AvH-F.” This was within the context of the exchange program (since ‘72) of highly qualified American (and later also others) and German scientists. The 1988 meeting was held in Rotach-Egern in Bavaria,] and I was crazy enough to suggest “anthropogenic climate variations.” [In the ‘80s experts predicted from observations, e.g. an annual loss or gain of mass of the Greenland Ice Sheet. Under such conditions computational methods could perhaps bring more security.]
I prepared myself for two full months for this lecture [and learned a lot from the PhD dissertation of C. J. van der Veen, done in Utrecht, Holland], aside from my teaching, which I had to do. [I also read a number of other books on the influence of Greenhouse gases. All this gave a rather solid first knowledge on what was “hot” in climatology, which I wanted to pursue later on. The meeting in Rotach-Egern was a wonderful experience.]
It had a great influence on my own personal interests and an immediate impact on my activities.
[I could immediately organize an interdepartmental course at Techn. Hochschule Darmstadt on climate related topics that were of general interest. Internationally known specialists were invited to hold a lecture on their specialties: the speakers were willing to deliver their manuscripts which I edited. The lecture series was then published in book form as:
Hutter, K. (Ed. und Autor), Dynamik umweltrelevanter Systeme, Springer-Verlag, Berlin-Heidelberg-New York-Tokyo-Hong Kong, 1991
I consulted Dr. Heinz Blatter, a theoretical physicist and locally known glaciologist at that time at the Institute of Geography, ETH, Zurich and asked him, whether he would be willing to write a paper with me out of the Rotach-Egern lecture in their institute reports, “Geographische Schriften.” He added a number of important contributions. Unfortunately, to receive Prof. Ohmura’s consent for the publication took more than a year. In the meantime the editors of Physik in unserer Zeit had also asked for a printed version of an article covering this topic in the German language.
The two papers are:
Hutter, K., Blatter, H. and Ohmura, A., Treibhauseffekt, Eisschilde und Meeresspiegel, Physik in unserer Zeit, 20, Nr. 6 (1989): 161 - 171, Nachdruck in "Dynamik umweltrelevanter Systeme" (K. Hutter, Hrsg.), Springer Verlag, Berlin etc. (1991)
Hutter, K., Blatter, H. and Ohmura, A., Global ice sheet dynamics and the Greenhouse Effect, Zürcher Geographische Schriften, No. 37, (1990): 1 – 83
Had we published just a year earlier the general impact would have been bigger and we would not have had precursors.]
All this had a great influence on my own personal interest in anthropogenic climate dynamics. Thomas Stocker, who had just finished his degree in Zurich with me on topographic waves in lakes, went to McGill University, Montreal, Canada to work with Lawrence Mysak, who had become a climatologist. And later he worked with Wallace Broecker for two years and then became, at the age of 34, professor of climate and environmental physics in Bern, Switzerland, the successor of Hans Oeschger, an early follower of the Milankovich theory.
This activity opened for me a door to combine thermo-fluid dynamics with serious climatology. We had reasonably consistent fluid mechanical models by which ice sheet evolutions could be predicted, if the climate driving and geothermal heat were properly parameterized. Reinhard Calov’s software worked for wholly cold ice masses; it needed the irrational trick of setting the ice temperature back to the melting temperature for instances when it should, according to the cold ice theory, be above the melting temperature.] We had already derived in ‘81 the more general theory for polythermal ice masses [consisting of disjoint regions of cold and temperate ice. The field equations in the two regions were formally the same, parabolic systems of equations. In cold ice the field variables are the velocity, temperature, and pressure fields. For temperate ice, a significant field variable is the moisture production rate density in the mass balance equation of the water. It is due to the shear heating and can be obtained from the energy equation of the mixture as a whole via the latent heat of fusion. Across the cold-temperate transition surface (CTS), the surface separating cold and temperate ice, the jump conditions of water mass, mixture momentum and energy must hold. This leads to the jump conditions, relating the kink of the temperature profile across the CTS to the jump of moisture flux across it.
Just in time, when Reinhard Calov had sufficiently advanced in his dissertation, after his degree in ‘94 he went to Prof. André Berger at the Catholic University of Louvain, Belgium, for two years, and then, and up to the present, he is at the “Helmholtz Institute of Climate Impact Research” (Institut für Klima Folgenforschung) in Potsdam near Berlin] — a brilliant student, Ralf Greve, [who had done his diploma thesis on granular chute flow], wished to continue his PhD degree with me. [I told him, he should look for someone better, but he insisted. He had already won financial support for his graduate education from the “Studienstiftung des Deutschen Volkes,” which supported him for five years.] I assigned to him as PhD work and thesis to put the polythermal shallow ice approximation (SIA) on the computer [and to construct, if possible, the response of Greenland and/or Antarctica to realistic climate driving.] The model should also account for the deformation of the basal surface by including a model for the coupled lithosphere and asthenosphere.
Ralf’s PhD thesis consisted of the re-derivation of the equations of the polythermal ice masses [with an improved parameterization of the water production rate and Clausius-Capeyron equation as compared to mine.] He developed as part of the thesis his electronic code SICOPOLIS [Simulation Code for Polythermal Ice sheets), and fortunately made it an open source program. With this software it was not only possible to computationally determine the variation in space and time of the ice surface due to the varied climate input from the atmosphere, one could also vary the deformation of the solid earth and the evolving temperate patches in the ice, if these should form.
Since 2003 Ralf Greve has been professor of theoretical glaciology in the Institute of Low Temperature Sciences at Hokkaido University in Sapporo, Japan. In this capacity he and his students have extended the original SICOPOLIS software to include sheets and shelves. Ralf regularly updates SICOPOLIS, and he also generated a version in spherical coordinates for whole or semi-global applications. His software now also includes ice shelves with the shallow shelf approximation (SSA, which we have not touched yet).] In that form SICOPOLIS is used as a cryospheric module in a number of GCMs.
Global Climate Models?
Yes, [the early GCMs in the ‘80s and ‘90s had full theoretical modules of the atmosphere, the oceans, and simple versions for the pedosphere, but only a primitive version of the cryosphere, a simple parameterization of the ice extent only coupled via the albedo to the back-radiation into the atmosphere. The SIA of grounded ice masses could do this better, quasi with the complexity equal to the models for the ocean and the atmosphere.] A comparable shallow shelf model still did not exist. [A first model of an ice shelf, wholly floating on a still ocean was derived by Prof. Morland at the University of East Anglia in Norwich, UK, in ‘87, but did not account for the melting/freezing processes at the ice-ocean interface so that ice shelves could and did consist of meteoric ice underlain by marine ice with distinct thermo-mechanical properties. Evidence of the existence of a considerable layer of marine ice was given by Thyssen, Engelhard, and Determann (from the University of Münster, Germany), first by seismic shots by conjecturing a meteoric-marine ice interface from a reflection surface, and later by drilling a hole on the Ronne-Filchner Ice Shelf, corroborating a more than 100 m thick marine ice layer at the lower part of the hole. A physically deeper theoretical understanding was needed before a consistent Shallow Shelf Approximation (SSA) of the underlying equations could be developed.
It so happened that in the ‘90s a number of physics students attended my course “Analytical Dynamics,” covering Hamiltonian dynamics from rigid body motions to wave dynamics, ending with the Schrödinger equation. This was a course to prepare the students for a preliminary exam after 2 years of study. Several students of this course asked me later to guide them through the diploma (roughly equivalent to an MSc) degree, i.e. to give them topics for the diploma thesis, which they could write and submit after six months (with possible extension of two months, if necessary).] Magnus Weis was one of these students. [He was not only a superb physics student; he had unusual knowledge and talent in informatics. He performed as his diploma thesis three-dimensional computations with the SIA equations under the past climate driving, and later also generated, with support from Reinhard Calov, a video clip of the evolution of the ice cover of the northern hemisphere during the past 250,000 years, then using the SIA equations of cold (not polythermal) ice in spherical coordinates:
Weis, M., Hutter, K. and Calov, R., 250,000 years in the history of the Greenland Ice Sheet, Annals of Glaciology, 23 (1996) 359 – 363
The thermodynamic theory for the evolution of ice shelves incorporating meteoric and marine ice was not entirely straightforward. Two difficulties were:
(i) To introduce the correct scaling analysis that allowed non-dimensionalization and eventually the derivation of the approximate models — Shallow Shelf and Second Order Shallow Shelf Approximations, SSA and SOSSA, respectively — by perturbation analysis in the perturbation parameter, ε = H/L, H = typical thickness, L = typical horizontal length, to zeroth and second order. This was a mathematical problem that was eventually solved.
(ii) One weak point of all shelf-models is the still incomplete handling of the kinematic boundary condition at the ice front. The ice front is the ocean front of the ice shelf, about 200 m thick, where ice breaks off in the form of icebergs. In principle, ocean water could also freeze on the shelf front, but this contribution must be negligible small]. Its mathematical derivation as such is not difficult, but in the unsteady case it involves a mass loss/(gain) quantity, the so-called calving rate, which must be parameterized. These parameterizations are still in their infancies now, but proper propositions determine whether the retreat of ice shelves can realistically be predicted. The disintegration of the Ross and Ronne-Filchner Ice Shelves is the determining sub-process of the ensuing wastage of the West Antarctic Ice Sheet (WAIS). Our papers are:
Weis, M., Greve, R. and Hutter, K., Theory of shallow ice shelves, Continuum Mech. Thermodyn. 11, (1999), 15 – 50
Greve, R., Weis, M. and Hutter, K., Palaeoclimatic evolution and present conditions of the Greenland Ice Sheet in the vicinity of Summit: An approach by large-scale modelling, Palaeoclimates 2(2-3), (1998), 133 – 161
Magnus subsequently took his model equations and wrote a finite element program for the steady response of ice shelves. This way the calving rate did not have to be parameterized; it can simply be set to zero or evaluated as the flux quantity across the front by holding the latter at constant position. With explicit academic computations of rectangular shelves, held fixed at two opposite boundaries and one boundary providing the incoming flux of mass and the other as the outflow boundary, his thesis was completed.
Magnus was an exceptional personality and appeared to me as more discreet and more mature than others.] This was likely because his father in the mid ‘50s was seriously ill with Alzheimer’s, and Magnus, his mother, and sister had to take care of him in turns. [Magnus was free to come and go as it was possible until his father passed away.]
Just to clarify, the shallow-ice approximation, that also applies to a marine ice sheet?
[Well the answer to this is a bit complicated. The SIA equations are physically very close to simple shearing; the most important components of the stress deviator are τxz, τyz, whilst the deviatoric normal stresses σxx, σyy, σzz, and τxy nearly vanish. With a viscous stress-stretching law the SIA equations give rise to shearing in the z-direction, not plug flow. By contrast, the SSA equations describe membrane behavior just like a soap film or a drum with viscous behavior. Here, the deviatoric normal stresses σxx, σyy, σzz and τxy are dominant, whilst τxz, τxz are negligibly small. With τxz=0, τyz=0, the corresponding shearings equally vanish and the horizontal velocity components describe plug flow. The two equation sets cannot come together. Stresses which are dominant in the SIA are vanishingly small in the SSA and vice versa.
It follows that in the vicinity of e.g. the grounding line neither SIA nor SSA are acceptable approximations, since the SIA holds true in the inland ice whilst the SSA does so in the central shelf region. One way is to solve the complete full Stokes equations in this transition region: another way is to apply a different non-dimensionalization of the equations (i.e. to apply a different scaling!) whose solution can be identified with the SIA solution on the far-inland side with the SSA solution on the far ice-shelf side. This is the method of matched asymptotics. Yet another way is to view the SIA and SSA as lowest order approximations of regular perturbation expansions and to pursue these expansions to second order. This is possible, if Glen’s flow law (a power law) is replaced by a polynomial law, in which at small strain rates near-Newtonian behavior survives. (This regularizes the SIA equations and is a technical mathematical detail). Formally, this yields
Shallow ice approx. = SIA + …
Shallow shelf approx. = SSA + …
Here… are first and second order contributions to the shallow ice and shallow shelf approximations, respectively. It can be shown that the first order contributions are zero. Moreover, in these two equations ε = H/L is the so-called shallowness parameter; as explained before ε << 1. The terms on the right-hand side with the growing exponent of ε are of smaller and smaller influence. It also so happens that the terms involving ε linearly vanish. Moreover, the non-vanishing second order terms (ε2SOSIA*, ε2SOSSA*, where the asterisks indicate the second order corrections to the second order approximation) involve, in the second order shallow ice approximation, those stresses which were missing before in the zeroth order SIA and SSA. So, formally, the equations
SOSIA = SIA + ε2SOSIA*,
SOSSA = SSA + ε2SOSSA*
Are the minimum complexity needed to be able to model a smooth shelf-sheet transition. The two solutions must be patched together, say at the grounding line. To obtain these patching conditions a local analysis of the grounding line is needed, since the grounding line moves in general whose differential equation must be found.
With the second order SOSIA and SOSSA one is in principle also capable of analyzing ice streams and marine ice sheets, which have separately been looked at by Christian Schoof. A relatively simple description can be found in:
Kirchner, N., Hutter, K., Jakobsson, M. and Gyllencreutz, R. Capabilities and limitations of numerical ice sheet models: a discussion for Earth-scientists and modellers, Quaternary Science Reviews, 30, (2011) 3691-3704, doi:10.1016/j. quiascirev.2011.09.012
In addition, a paper on the SOSIA has been published already in 2001:
Baral, D. R., Hutter, K. and Greve, R., Asymtotic theories of large scale motion, temperature and moisture distributions in land based polythermal ice sheets. A critical review and new developments, Applied Mechanics Reviews 54 (3), (2001) 215-256
However, I have not seen a published version of the SOSSA except in Dambaru Baral’s dissertation:
Baral, Dambaru. Asymptotic theories of large scale motion, temperature and moisture distributions in land based polythermal ice shields and in floating ice shelves — A critical review and new developments (1999), TU Darmstadt.]
But this also applies to situations such as Greenland?
[The second order approximations apply in principle to ice sheets and ice shelves alike. So, they can be applied to Greenland and Antarctica. However, the reasons they do so are of unequal urgency. The Antarctic Ice Sheet has land-based portions, the marine WAIS, and the entire Antarctic Ice Sheet is surrounded by many ice shelves, of which the largest ones are the Ross and Ronne-Filchner Ice Shelves, which are nourished by the inland ice.] In this case the use of higher-order models is vital to be able to ‘hook’ the shelves together with the sheet. [In the Greenland Ice Sheet, which is wholly based on solid rock, the SOSIA equations improve the model results of the SIA in the vicinity of domes, ice divides, and the free surface.]
The computations of the derivation of the second-order sheet and shelf approximations from the thermo-mechanical equations were done between Feb. 1995 and Fall 1999 by Dambaru Baral, a Nepalese PhD student with considerable help from me. It was a dissertation of more than 200 pages, full of equations. I never dreamed that anybody would ever try to put them on a computer, and tried to forget it. This was around the years 2000-01, when I concluded this. I thought it was a nice exercise, but that it was most likely the most useless thing I’ve ever done in my life. But apparently it was not so useless.
There were really dedicated people who did more, and School was one of them. For marine ice sheets he did a related perturbation analysis. (Marine ice sheets are ice sheets which sit on solid ground but below sea level.) So, they have a little bit of friction. Shelves have no friction [however, in the SOSSA they can support a bit shear traction]. How can I draw this? So, these are shelves, OK? Below here they have no friction.
Right. Because they are on water.
[Because of the water the submerged part of the marine ice sheet is subjected to the gravity force minus the buoyancy force of the submerged portion of the ice. On the inland side of the grounding line, this gives rise to a basal friction of reduced magnitude (zero at the grounding line, and its maximum in positions where the basal surface is just above the sea level).]
This isn’t worrying about ice streams or anything like that?
[The SOSSA also applies to these but explicit calculations are not known to me. An indication is given in the paper by Nina Kirchner et al., which was already cited above. In that paper the flow of the ice in Svalbard is determined under a realistic present climate scenario using SIA, in which several ice streams (but not all) are rather well reproduced by the computations. These are promising results for an extension to SOSIA computations.
Nina Kirchner, now lecturer at Stockholm University, has a diploma as a mathematician at ETH Zurich from 1998, where she fell in love with my two courses on Theoretical Glaciology. She did her doctorate with me in Darmstadt (1998-2001) on thermodynamics of structured granular materials, and stayed until 2002, and then had positions at the University of Kaiserslautern (Germany) in mechanics and the Fraunhofer Institute on Industrial Mathematics. Since 2006 she has been at the Department of Geography and Quaternary studies at Stockholm University.
Grounding line behavior is another sub-process, where SOSIA will likely deliver promising results. I am told, but have not closely followed Andreas Vieli’s recent activities, and think that he developed a model for ice flow in the grounding line region that belongs to this category of models.]
I don’t know whether you know about Vieli?
He was in Durham. He was also our student here. [Vieli was in the ‘90s one of my students in the two courses Theoretical Glaciology at ETHZ, and did his PhD thesis with Dr. (later Professor) Heinz Blatter. He went to Bristol to work with Antony Payne and then “went up the ladder” at Durham University, UK, from lecturer to reader until he returned home to Switzerland to represent Glaciology in the Department of Geography at the University of Zurich. In 2012 he became Wilfried Haeberli’s successor there.]
This is all ongoing work, and my recollection is probably rather incoherent, [but SIA, SOSIA, SSA, SOSSA, and alternative approximations suggested by different scalings such as those by Christian Schoof and Richard Hindmarsh, will likely keep modelers busy for more than 10 years.
The extension of the SIA and SSA to the second order is actually rather obvious. Had we looked at the reduced equations of zeroth order in the two approximations, compared them and identified those terms which were present in one formulation but not in the other, it would have been quite clear how SOSIA and SOSSA had to look like. This would not have freed us from deriving the complicated equations from the dissertation of Dambaru Baral, but it would have made those complicated derivations more comfortable.] Unfortunately, I did not see this 10-12 years ago.
We were the first who did this but I did not have the courage to press a graduate student: “Why don’t you do it?” Danish geophysicists [Egholm et al (2011) derived a depth averaged version of the SOSIA and published their results in J Geophys. Res. More will have to come in the near future.] For now Schoof [and Hindmarsh] are the experts.
[Before I continue we should point at a peculiarity of such computational perturbation approaches as suggested by SOSIA and SOSSA, which might bring them back to the center of computational cryospheric climatology. It is at least for 20 years that numerical experts in ice sheet and ice shelf dynamics have been advocating for integration of the Stokes equations, i.e., the equations, which are the full balance laws of mass, acceleration-free momentum, and energy without imposition of the shallowness assumption. Several such open source programs are available, and extensive computations for ice flow have been performed, e.g. for the Greenland Ice Sheet and other ice masses. The disadvantage of these programs is that they are excessively time consuming. More specifically, computations over some 100 years, but no more, are economically possible. For climate reconstructions, at least 100,000 to several 100,000 years are required, which does not seem to be possible in the near future (say 20 years). This fact may well make SOSIA and SOSSA economically competitive alternatives.]
You mentioned while we had the recorder off your association with Richard Hindmarsh.
[I met Richard Hindmarsh in the mid ‘80s at the University of East Anglia, Norwich, when I was a visitor to Leslie Morland in the School of Mathematics and Physics, ‘84. He had planned to apply the SIA equations to near circ sheets and I had drafted a manuscript during my three weeks presence. For some undefined reason Leslie never found time to devote to this problem. The manuscript never got published. I was then still at VAW and could freely do my own research as long as no money was involved. Richard then was working with Leslie Morland and Professor Geoffrey Boulton (who had a strong interest in glacio-geomorphology), helping them to numerically integrate the SIA equations for irregular two-dimensional ice masses. I do not recall whether Richard had already done his PhD degree at that time.] He had already written a thesis on the use of the finite element method applied to ice flow problems. The cooperation of Leslie and Geoffrey with Richard was not optimal. So, it was decided that Richard should visit me at VAW for a few weeks. [I recall that we had intellectually a very hard time coming together in the mathematical formulation of a transformation, which made the integration domain time-independent (a so-called “fixed domain mapping”). We discussed for days, and eventually the disagreement disappeared and everything ended in harmony, we could start writing our ideas without major differences. Richard visited me once more, then in Darmstadt, about ‘93. He became a solid expert in the use of perturbation methods of shallow ice masses. Our joint publications are:
Hindmarsh, R. C. A., Morland, L. W., Boulton, G. S. and Hutter, K., The unsteady plane flow of ice sheets: A parabolic problem with two moving boundaries. Geophys. Astrophys. Fluid Dyn., 39 (1987): 183 – 225
Hindmarsh, R. C. A. and Hutter, K., Numerical fixed domain mapping solution of slow creeping free-surface flows coupled with an additional evolution equation, Int. J. Num. Analyt. Meth. Geom. 12 (1988): 437 – 459
Hindmarsh, R. C. A., Boulton, G. S. and Hutter, K., Modes of operation of thermomechanically coupled ice sheets, Annals of Glaciology, 12 (1989): 57 – 69]
Richard is now often attacking problems involving a critical point or critical regions in ice flow problems such as the dome region or the grounding line region.
Of the ice sheet —
Yes, where all applied glaciologists, the climatologists drill holes, and deduce the climate [from the composition of the trapped air: in Greenland (GRIP, GISP, NORTH-GRIP, DYE-3), in Antarctica (Vostok, Dome Concordia, EPICA (Kohnen station)) and many others. However, it is well known to the specialists of the SIA that computational results of the SIA are critical (singular, when Glen’s flow law is used) at domes calling for SOSIA computations…]
Richard Hindmarsh looked at the ice divide region, but he was not the first. More than 10 years earlier, Anne Mangeney presented in her PhD dissertation in Grenoble [a computational analysis of the flow of ice in the vicinity of a two-dimensional ice divide using essentially the SOSIA approximation, in which longitudinal stress effects become important. So, her work is likely the first account where the SOSIA was applied. I happened to be in her thesis committee, had helped her and invited her to Darmstadt for a few weeks. The related paper is:
Mangeney, A., Califano, F. and Hutter, K., A numerical study of anisotropic, low Reynolds numbers, free surface flows for ice sheet modelling,J. Geophys. Res. 102, No. B10 (1997), 22749 – 22764
Anne Mangeney was actually a PhD student within the seismology group in the Department of Geophysics at the University of Paris VI, but she and her husband did their PhD work at CNSR-LGGE, the glaciological institute in Grenoble. After her degree they returned to Paris and she started to teach in the seismology group. She then changed her specialty to rapid landslides, especially the numerical integration of the Savage-Hutter and related equations for use, e.g. at Mont Serrat and on Mars. In February 2005 she invited me for a month to their institute, where I lectured on avalanching flows and where we studied a significant numerical paper by Bouchut and Westdickenberg (two mathematicians), which became important for us for the computation of avalanching granular flow down arbitrary topography. Today, Anne is full professor and I believe head of seismology at the University of Paris VI.]
But, at the time, Grenoble was involved in the GRIP ice core —
The Greenland ice core?
[It actually applies to any ice core from top to bottom. The French glaciologists have done a lot of good work in climate reconstruction from ice cores. With the analysis of the polycrystalline ice down in the ice cores it became clear that the simplistic Glen flow law, or its replacement by a polynomial rather than a power-law parameterization, was not modeling the stress-stretching response of polycrystalline ice properly in its motion from its surface down into the ice sheet. The mechanical property of the ice near the surface could be assumed to be isotropic, i.e. the same in all directions, a typical fluid like behavior.] After all, the polycrystal was formed from snow crystals, which fell onto the surface with arbitrary orientation, each with the same likelihood. [With time, as a representative volume element (RVE), i.e., as a small ice cube consisting of a large number of ice crystals is moving along its trajectory through the ice mass, the individual crystals in the RVE perform their own rotations. The orientation distribution within the RVE will, in turn, determine the mechanical behavior of the RVE, and the latter will affect how fast the ice in the RVE will move through the ice mass. This, on the other hand, will determine the age of the ice at a certain location within the ice sheet, i.e., the time that elapsed since the ice in the RVE fell as snow on the latter’s surface. This description shows that the following items must be determined:
(i) The composition of the air in the samples at a given position,
(ii) The age of the ice in the sample.
To understand this properly, let me explain this under somewhat idealized conditions. Let us isolate in thought a crystal in the RVE and imagine the crystal to be a small elongated hexagonal cylinder and let the cylinder axis be given by an arrow of length unity.] This axis is called the c-axis of the hexagonal crystal. [Let, moreover, a plane perpendicular to the vertical bore hole axis be alled equator plane of a sphere, of which the center is the intersection point of the bore hole axis with the equator plane. Consider only the upper half-sphere and orient all c-axes such that their arrows point into this upper half space. (The hexagonal ice crystal does not differentiate between the c-axis pointing in either of these opposite directions, one of them can freely be chosen.) A point on the unit semi-sphere can be viewed as the orientation representation of a crystal, as is the projection of this point onto the equator plane]. The orientations of all crystals in an RVE define the distribution of points within the so-called Schmidt circle in the equator plane. [If all crystals are vertically oriented, all points collapse to the central point of the Schmidt circle, and if they are uniformly distributed they fill the entire Schmidt circle concentrically with largest concentration at the circumference and smallest concentration at the center, but well defined (by the geometry). Observationally, one has in each sample a large but finite number of crystals, but theoretically one assumes a continuous distribution.
The GRIP ice core, where the earliest analysis was made, showed that the Schmidt circles at the surface indicate isotropy, whilst at depth, but still in the Holocene ice, orientation points are concentrated close to the center.]
Is this a compression effect, or is it from some other source?
It is compression and shearing. Consider, e.g., a particle trajectory starting at a dome, where the velocity profile has a vertical tangent. At that position the state of stress is sidewise compression and the motion is vertically downwards. [As the ice now moves along its trajectory, the trajectory will turn to the side and eventually become practically horizontal as the ice moves essentially horizontally to the bed. In this regime the stress state is nearly simple shearing. So, as the ice particle (an RVE of ice!) moves through the ice during an ice age, it will be subjected to different forces, deformations, and the orientations of the crystals change.]
Okay, I understand that.
[The continuous orientation distribution function varies with time and space and will influence the material behavior. Its evolution is based on postulated statistical properties in orientation space as is the Boltzmann equation in velocity space. Certain moments will enter the constitutive relations as structure variables and the entire set of equations must be restricted to satisfy the Second Law of Thermodynamics.
These modern theoretical formulations of the description of the constitutive behavior of the ice in large ice sheets or ice shelves are significant improvements over the classical simple Glen flow law.] It is known that they exert a considerable effect on the prediction or reproduction of the age of the ice. [They require for proper application and theoretical extension a deep understanding of the material science of ice in the geophysical environment.]
This has just been developed within the last 10 years?
10, 15 years [as of Aug. 2014]
Ah, that’s very interesting.
[I had a number of further students working on microstructural effects and inclusion of recrystallization and other subscale processes:
Sérgio Faria, a Brazilian, who got his MSc Degree in physics at the University of Brazil in Curitiba and
Luca Placidi, a student of Engineering Mechanics from the University of Rome (La Sapienza)
Both did considerable work in the derivation of thermodynamically consistent creep models, in which structural and recirculation effects are accounted for as their PhD work. In this brief account it is not possible to give an objective description of their work. A summary of some of our work during the last 20 years can be found in:
Placidi, L., Faria, S.H. and Hutter, K., A critical review of the mechanics of polycrystalline ice. GAMM Mitteilungen 29, 1, 2006, 77-114
Okay. But you were also working on the granule physics. We don’t have a lot of time, but if you want to talk about it a little, please do.
[Yes, granular media became one of my active research fields in 1980 when I formulated a research proposal on snow avalanches for Prof. Vischer at VAW-ETHZ. That activity led to the PhD work on powder snow avalanches by Thomas Scheiwiller and Felix Herrmann at VAW. Work on dense granular avalanches that was scientifically really significant was the avalanche model, which Prof Steward Savage from the Department of Civil Engineering at McGill University, Montreal, Canada and I developed around 1986, and was published in the J. Fluid Mechanics (JFM) as:
Savage, S. B. and Hutter, K., The motion of a finite mass of granular material down a rough incline, J. Fluid Mech. 199 (1989): 177 – 215
With many significant continuations later on; see the book:
Pudasaini, S. P. and Hutter, K. Avalanche Dynamics – Dynamics of rapid flows of dense granular avalanches Springer Verlang, Berlin, etc. (2007)
Professor Savage had applied for a visiting position at VAW and received it for two years from August ‘86 to July ‘88. We knew each other a bit from the “US-Japan Granular Materials Conference” at Cornell University in summer ‘82 and had worked on ideas of a depth-integrated model, probably in ‘85. The first draft of the paper was finished at the end of ‘86 (I had spent three weeks of my summer vacation in August ‘86 in Montreal to write the paper, when Stuart — to my surprise — went on vacation on the day of my arrival). The paper, when submitted to JFM, was not well received by the referees of JFM and it took us until ‘89 to extend the text considerably and make the manuscript publishable. (I still prefer the first draft.) In this paper we derived the depth integrated balance laws of mass and momentum for a shallow moving incompressible two-dimensional medium down an inclined plane and used very limited laboratory data collected by Dr. Andreas Huber at VAW. The model and its extensions became known as Savage-Hutter (SH) model/equations and were further explored both theoretically and experimentally in several Diploma, PhD, and post-doctoral works (Chr. Plüss, VAW; Ralf Greve, Thilo Koch, M. Wieland, J.M.N.T. Gray, Yih-Chin Tai, S.P. Pudasaini, Min-Chin Chiou, all at Darmstadt). An account summarizing the work on avalanching theories is the book by Pudasaini and Hutter.
The research work on avalanching flow was continued after my retirement from Darmstadt University, when I spent a total of 8 months at Academia Sinica, Taiwan, between 2006 and 2009, in the research group of Dr. Chih-Yu Kuo in the Institute of Applied Sciences. He and Prof. Tai are still continuing this kind of research, now focusing on Typhoon-induced landslides.
Fundamental work on the thermodynamics of granular materials and solid-fluid mixtures kept me equally busy throughout the Darmstadt period with Diploma and Ph.D. students and visitors. Among these are Nina Kirchner, Georg Bauer, Chung Fang, and Prof. I. Luca from the Mathematics Department of the Polytehnica in Bucharest. All this work could be documented with articles in professional journals. An example is the book:
Schneider, L. and Hutter, K. Solid fluid mixtures of frictional materials in geophysical and geotechnical context based on a concise thermodynamic analysis. Advances in Geophysical and Environmental Mechanics and Mathematics, Springer Verlag, Berlin, etc. (2009)
A further research activity that received a prominent position among all research that was done in my group in Darmstadt was physical limnology and physical oceanography, best paraphrased as geophysical fluid mechanics, in particular linear and nonlinear waves in laboratory channels and in lakes and the numerical computation of wind-induced barotropic and baroclinic currents in Lake Constance and seiches and sediment movement in lakes. These topics involved quite a few Diploma and PhD dissertations of mostly physics students and post-doctoral work, most of which was published in peer-reviewed fluid journals. Since my retirement this work was summarized in the following books:
Hutter, K., Wang, Y. and Chubarenko, I. Physics of Lakes Vol 1: Formulation of the Mathematical and Physical Background, Springer Verlag, Berlin etc., 2011)
Hutter, K., Wang, Y. and Chubarenko, I. Physics of Lakes Vol 2: Lakes as Oscillators, Springer Verlag, Berlin etc. (2011)
Hutter, K., Wang, Y. and Chubarenko, I. Physics of Lakes Vol 3. Understanding lakes as components of the Geophysical Environment Springer Verlag, Berlin etc. (2014)
Hutter, K. (Editor and author) Nonlinear Internal Waves in Lakes (AGEM^2), Springer Verlag, Berlin, etc. (2012)
This last book was written as a report to the European Science Foundation for a joint collaboration with Russian, Ukrainian, and British scientists within a program, called INTAS. This is all I can report now.
Is there anything else that we would like to cover? You were a scientific editor of the Journal of Glaciology for a while. I don't know if there’s anything we ought to talk about with respect to that.
I’ve been editor of several journals and periodicals. Professor Ingo Müller, Berlin and I founded Continuum Mechanics and Thermodynamics, a Springer Journal. That was started in ‘88 and came into life in ‘89; I stopped being Chief Editor when I retired. [I was also member of the editorial board of Acta Mechanica for more than 20 years.] Then I was Scientific Editor of the Journal of Glaciology for 14 years, and Chief Editor of four Annals of Glaciology. The Annals are peer-reviewed conference proceedings. Twice I was Scientific Editor of two other annals. [For some years I did a lot for the International Society of Glaciology] and I was vice president of the society for 6 years and — this is interesting — in 2003 they bestowed the Seligman Crystal [the most prestigious science prize of the Int. Glaciological Society] upon me for work, which was essentially rejected as papers at the beginning in glaciology and published elsewhere. Papers on mixture formulations were not well accepted — or ignored, perhaps that is the better expression. And work on intraglacial channel flows or jökulhlaups was equally not well received; this work was done with Ueli Spring — it covered Spring’s dissertation with a tremendous involvement of myself — and was another piece of work, [which was vital for us when we were doing it. I put lots of emotion into it. The publication was hard, a humiliating process. I never touched the topic again afterwards.] But much later, I was told that this work was important in bestowing the Seligman upon me.
Are you familiar with the jökulhlaups?
Okay, a jökulhlaup is water flow through an intraglacial channel, [which is under water availability and water pressure conditions so that the flow of water through the channel starts harmlessly, but quickly increases to a tremendous maximum that creates beneath the glacier or ice sheet a damaging flood. This growing phase of water discharge has a duration of only a few days, or at most one to two weeks. Once the maximum discharge is reached, this flow generally drops off quasi exponentially. In the Swiss Alps, e.g. at Gorner Gletscher, the formation of “Lake Gorner” is a prerequisite condition in which the phenomenon can be initiated. This lake is created at the merging junction (Y) of the upper Gorner Gletscher and neighboring Grenzgletscher in spring and early summer, when the melt water is collected in the lake. When the lake level has reached a certain threshold level, the water at depth and at the lower Gorner Gletscher may break through and find its way through cracks and channel-like conduits,] and generates enough viscous heat that the turbulent channel water can melt ice from the channel walls and at the base or within the ice. Only if the turbulent flow generates enough viscous heat, can the channel water melt ice from the channel walls, and, thus, enlarge the effective cross section. [As a consequence, the flow of water through the system will increase as long as the discharge from the lake can supply the discharge. However, the lake level will lower, and, correspondingly, the outflow will decrease; think of the Toricelli formula]. So, initially the discharge through the intraglacial channel will increase and reach its maximum when the inflow into the channel from the lake will start to decrease because of insufficient availability of water until the lake is emptied.
There is an additional process, which operates as an accelerating process of the closing operation. The intraglacial or basal channels, close to or at the base of the ice mass, are exposed to a tremendous overburden pressure. This pressure contributes to a closure rate of the channel cross section, which reduces the amount of discharge as well. [In the case of the Gorner Gletscher this process, however, cannot be large since the overburden height is not very large and Lake Gorner is relatively small.]
Is that the same sort of thing that happens in Greenland, or is it different from that. It’s not that scale?
The scale is different but the foundations are the same. [In Iceland at Vatnajökull, the water in the ice sheet is produced by the volcanic heat from below, the ice there is at least partly floating on melt water, and depending on this melting rate, the growing subglacial lake is lifting the ice. A “riegel” in the basal topography prevents the water from flowing off as long as the top surface of the water level in the gigantic cavity is below the level of the “riegel.” The operating mechanism of the sub-glacial channels is then again the turbulent heat generated by the water flow. However, because the water cavity is so huge and the overburden pressure large, the discharge at the “riegel” down the ice sheet to the ocean is primarily reduced by the closing rate of the channel walls. Both effects mean that the time and space scales of this jökulhlaup are larger than in the Gorner Gletscher.
Such events are of short duration and there is an immense discharge in the river or low land below the glacier door. For Gorner Gletscher the entire phenomenon lasts about 3 days, with maximum discharge of Q = 45 m3/s. For Vatnajökull the duration is about 3 weeks.] Ueli Spring and I have obtained a maximum discharge of a quarter of that of the River Amazon. When such a jökulhlaup occurs at Vatnajökull, the low land is always heavily affected. Highways and streets are destroyed and bridges washed away, if they have not been removed prior to the event.
I have an impolite way of characterizing a jökulhlaup event: I say, the glacier is flushing its toilet. It’s really a very dramatic process. It starts with almost nothing, but picks up very large momentum and very large discharge masses, which even in the Alps can generate damaging floods.
Are they at all predictable, or do they occur more or less randomly?
Not exactly randomly. They have a certain period of reoccurrence, and that is in Iceland 5 to 10 years. In our glaciers it’s practically one, seldom 2 years, and happens in summer. [Not every year has a dramatic outburst, rather simply large discharge flow of longer duration.]
Through an understanding of the physical processes, you can understand when the cycles happen, and at different scales?
Here in this institute the glaciologists work continuously on this, and they had several dissertations come out. The work with Ueli Spring was done in ‘77 to ‘79 and I had tremendous difficulties in publishing it. The theories derived from the fundamentals is very complicated, and when I submitted the paper on the foundations to the Royal Society — probably Phil. Trans. at that time — it was reviewed by a prominent glaciologist who instantly rejected it and said something like, “This is nonsense, absolutely nonsense, such complicated things.” I don't know who this was; I have a suspicion who it could have been. [In a phone call to the Secretary of the Royal Society asking for re-consideration I was informed that a very prominent glaciologist was a referee. The relevant papers are:
Spring, U. and Hutter, K., Conduit flow of a fluid through its solid phase and its applicaton to intraglacial channel flow, Int. J. Engr. Sci., 20 (1982): 327 – 363
Spring, U. and Hutter, K., Numerical studies of Jökulhlaups, Cold Regions Science and Technology, 4 (1981): 227 – 244]
But this is what the Seligman Crystal was ultimately rewarding?
Exactly this. I got the Crystal award for this and the mixture theory.
Right. Okay. I suppose we’re about finished then.
Yes, we have been talking almost three hours. Are you producing a written version, and then do I see this?
Yes, absolutely. You’ll have the opportunity —
Could I change or —
Absolutely. You’ll have every opportunity to do that. There’s no rush whatsoever. Sometimes these things can be frustratingly long. The transcription process has sometimes taken a long time.
So you don’t have to type it!
I won’t have to type it personally. I’ll read it over and make what changes I need to.
I’m in a position where I have to type everything, but that’s because I’m retired.
Yes. Was there any question when you retired of not coming back to Zurich, or was this always the place that you intended to be?
Well, my wife and my children never joined me in Darmstadt. So, we made a clear-cut decision at that time. There was no point to stay at VAW any longer in the years ‘82-‘87. True, I was in a golden cage, [but did not further wish to stay any further in an environment of selfishness, intolerance, and professional ignorance of those who made the decisions.] I knew I would never advance and so we came to this conclusion. My wife, Barbara, said she would not join me because our children were just getting ready for high school or university. So I commuted.
Okay. Well, that should do it then. Thank you very much for your time. It’s been a very informative interview and I’m sure we’ll be in touch about getting you a copy of the transcript.
I hope it is to your expectations.
It was terrific, thank you.