[an error occurred while processing this directive]
[an error occurred while processing this directive]

Oral History Transcript — Dr. Kirk Bryan

[an error occurred while processing this directive]

See the catalog record for this interview and search for other interviews in our collection


Interview with Dr. Kirk Bryan
By Spencer Weart
At Geophysical Fluid Dynamics Laboratory, Princeton, NJ
December 20, 1989

 
open tab View abstract

Kirk Bryan; December 20, 1989

ABSTRACT: Informal, unedited, exploratory interview. Covers work by Bryan and other groups of physical oceanography, particularly computer numerical modeling of the oceans.

Transcript

Weart:

Let's go back to when you started in the field, I guess in the sixties, and we'll talk about what were the main things going on then. What were the main groups?

Bryan:

Well, in oceanography, I feel that the main laboratories were perhaps Woods Hole on Cape Cod, and the Scripps Institute of Oceanography at La Jolla, and then of course, in geological oceanography, Lamont, which is now Lamont-Doherty Laboratory at Columbia University.

Weart:

Among foreign ones would that apply or are these still the main ones?

Bryan:

Of course, at that time everything you know was, Europe was still recovering from the war, and of course the Germans had quite a tradition going previous to the war, but their science took a long time to recover, and it was not up until the seventies and eighties really that they began to be productive again. England, on the other hand, the National Institute of Oceanography which is now the Institute of Oceanographic Sciences, they had some very good experimental work going on, and their experimental work was certainly on a par, I think, with Woods Hole at that time. The two groups worked quite closely together.

Weart:

In terms of people, who would you have said, around the time you started, would have been the people that you would have regarded as the leaders?

Bryan:

Well, I think in physical oceanography, the two you know people on the sort of theoretical side that really stood out were of course Henry Stommel, whom I had a chance to work with at Woods Hole when I was starting out for two years, and then Walter Munk at Scripps. And there really were a very very small number, I would say, of leaders in the field of physical oceanography. I mean, there were other very important people like Roger Revelle, but more on the chemical and geological side.

Weart:

Was he already important at that point? I guess he was.

Bryan:

Yes. I think shortly afterwards he became head of Scripps, and he did a lot to build it up, their staff, yes. He took over, you know, some time after Professor Sverdrup? stepped down, and so, he really built Scripps up first and then he founded the University of California at San Diego.

Weart:

What would you say were the main problems or discoeries, you know, the focuses or the exciting stuff that was going on?

Bryan:

Well, let's see. Stommel was of course working on building up, you know, a series of models of ocean circulation, and Professor Munk was of course involved in that too at that time. His interests later went into more smaller scale things. But I think that there was already an influence of the work that Charney and you know Rossby were doing on the atmosphere, on oceanography, to try to make things more quantitative. You see, that general — there is a great deal about oceanography that is in the way of, I would call, sort of qualitative interpretation of the observational evidence. You know, they would make these different sections, you know, through the ocean, and we had some of these great expeditions, like the Carnegie expedition, the Meteor expedition, and then of course the first thing when I was at Woods Hole they had the IGY expedition, which was an attempt to systematically survey the Atlantic Ocean. Of course many things, it didn't actually break new ground compared to what the Meteor you know had done earlier, but it sort of confirmed a lot of the features of the ocean. But the tendency was of course to describe those in more or less qualitative terms, tracing water masses around from one part of the ocean to another.

Weart:

Would you say then that the main problems were simply the equivalent of the general circulation of the atmosphere, describing the general circulation of the oceans?

Bryan:

Well, I think that it was very much more, I would say, in the way of exploratory science. Many features of the atmosphere, you might say, were pretty well explored, I would say, you know, by that time, by the fifties, by certainly the end of the fifties, because the radiosonde network had gotten up, you see, all over the world. And so I'd say that the atmosphere was explored, except for the very high atmosphere. But at that time there was still a feeling, and of course many oceanographers still feel that it's, you know, primarily exploratory science. You get a ship and go out and try to find some new part of the world that hasn't been sampled very much before, and describe the features, you know, in the area. But I think even at that time, they were interested in developing a more, what I would call a synoptic view of the ocean, of getting out enough ships in one place to make a survey. And when I was at Woods Hole they made one of the first surveys of the Gulf Stream, this Gulf Stream 60 survey, Fritz Fuglister, and that's described in Stommel's book on the Gulf Stream there.

That was the first time that you got a view of the ocean which is more familiar, I think, to a meteorologist. You know, meteorologists are used to looking at weather maps, you see, and something like that just wasn't available to oceanographers. No. Most of the interpretation was based on looking at two dimensional sections, you know, through the ocean. And oftentimes the interpretation would also be two dimensional because that's all they had. That's all they had. They didn't have that, you know, three dimensional view of things. And of course this is the very exciting thing about oceanography in the present, is that I think we're getting away from that, with the satellite altimeters and other means that we have to map ocean features and mesoscale eddies in the ocean, that we're finally breaking into a more synpotic view of the ocean. At that time, for instance, a lot of data that people would gather were really just kind of an addition to the climatological archives, because you'd be out in the middle of the ocean, take a few measurements, and without any context, it was impossible to interpret what it was.

Weart:

You didn't know if it was an eddy or (crosstalk)

Bryan:

—sure, that's right, that's right. Sometimes you know I've heard that two ships would be rather close to each other taking measurements. They'd have a completely different interpretation of what they were —

Weart:

— and completely different measurements.

Bryan:

— that's right, just because of sampling. (crosstalk) Yes, and I think that to a certain extent, the Navy was putting a lot of money into making hydrographic measurements, you know, partly in support of underwater sound projects, you know, and I think that gradually we were realizing that a lot of that money was just completely wasted, because it was sort of under sampling really.

Weart:

I see, just a lot of random measurements (crosstalk)—

Bryan:

— yes, yes, that's right. The sampling just wasn't adequate to measure what it was. And I think only gradually developed a little more sophisticated viewpoint of how you sample something.

Weart:

At the IGY or only after the IGY?

Bryan:

Well, let's see, this was —you see, in addition to major —you see, IGY was a really organized measurement of sections. There was a great tendency to have a lot of cruises, you see, in between that, where people would sort of go out and —and I'm sure they had some objective in mind there, but oftentimes it wasn't that systematic.

Weart:

I must say, when I look and see the tracks of the ships on the maps, they're rather zigzag, and ... where they were going.

Bryan:

— that's right, that's right, and I think that — of course, all those measurements helped us build up quite an ecology of the ocean. But probably the efforts weren't as well organized, you know, as they really should have been. At that time, there were really very few people working in the sort of dynamics, you know, of the ocean. I remember Bert Bolin saying, "There's only Stommel and Munk in this field." You know. So that sort of made me feel it's a good area to go into.

Weart:

Really something to get into.

Bryan:

And I'd had some experience in numerical modeling, working with Ed Lorenz up at MIT. So that, when I got an offer to join the lab, seemed like a good deal too, to go into.

Weart:

Let's talk about what happened then in the sixties.

Bryan:

Well, in the sixties, I think one of the — there was a real interest in trying to determine more the whole spectrum of motions, you know, in the ocean. And what they noticed with that synoptic mapping of Gulf Stream 60, you know, the importance of mesoscale eddies, you know.

Weart:

They got the eddies already that soon?

Bryan:

Yes, the eddies actually —they had been noted by oceanographers, I'm sure, at a much, probably even back in the thirties, but during the war period, Athelstan Spilhaus developed what he called the expandable —no, at that time it wasn't expendable, it was just the bathythermograph. It was something connected with anti-submarine warfare. So that gave him an ability to make maps, and they made some of the first maps, I think, during the war or shortly after the war, of Gulf Stream eddies. And then this gave them the idea of trying to look at the eddy field which corresponded to cyclones and anti-cyclones in the atmosphere.

Weart:

They realized that it was an important part of the —(crosstalk) transport—

Bryan:

—yes, that's right, and so there was a project which Stommel was behind in the, around Bermuda, they made measurements of the eddies sort of in the deep field in the area where, you know, it's not visible at the surface, the eddy field there. But then this led to —well, there were two other main efforts. They were called the MODE Eddy, let's see, it was Mid-Ocean Dynamics Experiment. I think that's what it was called. And that was in the sixties there, the middle sixties. That was an attempt to pick out a certain area that was I think somewhat to the west of Bermuda, and they put an array of current meters which actually didn't work, actually they got stuck. They made a little problem of, they replaced some parts with a slightly different metal, yeah, and of course, you got corrosioni that led to the —

Weart:

Who was doing this, NSF?

Bryan:

NSF, yeah, and of course there's a lot of material on the MODE thing. I think, I'm sure just NSF has a lot of it, kept a lot of the material, and they made a movie.

Weart:

They kept a lot of the data. I'm not sure if they have the archives describing how it was set up and so on, that might be a good thing —

Bryan:

Yeah, I think there were —

Weart:

It's very spotty as to what —

Bryan:

Oh, really? I see. OK, Woods hole I'm sure probably would, I think they would have a lot of those documents up there, when it was set up. But I think the MODE—of course, there was a little bit —yes, the MODE was the first, and then that was followed by GEOSECS, which is the effort of the geochemists, and Stommel did a lot to work with them, and Broecker, and Harmon Craig also at Scripps were involved in that. That was of course a really global survey. And very important. Very important results.

Weart:

Yes, important results for Broecker and others.

Bryan:

Sure, yes. And it was looking at the transient tracers caused by the bomb tests, you see.

Weart:

Was that set up on purpose to look for those?

Bryan:

I think, well, I think that was — well, I think there was a lot of interest in the, you know, cosmic ray radioactive tracers in general. And then, yes, that gave them a thing that they could say, you know, the importance of going out at that particular time —

Weart:

(crosstalk) —that particular time —

Bryan:

—time, because you catch them, yes, you catch them. And so I think that, yes, it's sort of a motive, you know, they put out. And they were already of course very much thinking in terms of Greenhouse Effect, you know, at that time. The role of the oceans I think was seen mostly at that time as the sequestering of carbon dioxide, and I think people of course don't realize now how far back that people were worried about this you know Greenhouse Effect.

Weart:

That's true. Sure. Back at the turn of the century people were talking about the ocean as a sink for carbon dioxide.

Bryan:

Yeah.

Weart:

That goes way back.

Bryan:

And certainly both Revelle and Wally Broecker were very much you know concerned about this.

Weart:

Were there other players beginning to appear at this time in oceanography? You talked about the time when there were really just Munk and Stommel and so forth. Others were starting to come on the scene?

Bryan:

Oh yes. Well of course, in the MODE project, some of the early —Carl Wunsch of course was —and Alan Robinson, I think were co-chairmen of that group. And of course Wally Broecker and Harmon Craig were the ones getting — I think Stommel got sort of worn out with organization things about that time, and decided he'd had enough, I think, after GEOSECS, he thought other people could sort of take over.

Weart:

In terms of institutions, it would have still been Scripps, Woods Hole and maybe Lamont-Doherty?

Bryan:

Well, yes, now, of course, I'm sure this would depend, other groups though, if I were to say that you know, others, University of Washington, University of Miami, groups were you know playing an important role too. But in terms of having the most — still now, I mean, Woods Hole in physical oceanography, they have really a dominantly large group working.

Weart:

Depends on who has the ships and so forth.

Bryan:

Yes, that's right.

Weart:

What about as funding sources? How important was the role of the military, of NSF?

Bryan:

Well, let's see, when I first went to Woods Hall, it was almost entirely Navy funded, and block funded at that time, which was very helpful in some ways for pursuing a lot of basic science projects. When I was there, there was interesting work. Willem Malkus was working on convection, and with George Veronis was working with him on that field, pursuing all sorts of different things. I was doing some model experiments, I mean rotating tank experiments at that time.

Weart:

Oh, you were doing those experiments?

Bryan:

Yes, yes, and following on —

Weart:

(crosstalk) ?...not for ...

Bryan:

Well, I was sort of doing general, thinking in GFD, you know, in general.

Weart:

Just in general.

Bryan:

Of course, I might say that this was, I think, a very important concept came up at that time, was really through I think Professor Rossby, University of Chicago, with the idea of Geophysical Fluid Dynamics, and then the exploration of the other planets, you know, was quite important. It's an idea that's making a more unified science of the atmosphere of other planets, the atmosphere of earth, and the ocean of earth, all systems more or less guided by rotation. Of course it would include other systems like the earth's core of course and solar systems too, but at that time that's a more complicated physics of course in that, so —

Weart:

Were there any —this was at the time when you were beginning to do modeling. Who else was doing modeling?

Bryan:

Let's see, Bill von Arx actually. He was a pioneer there in ocean — doing laboratory models of the ocean there. And then of course Dave Fulks at the University of Chicago, that was mainly directed to thinking about atmospheric problems but of course —

Weart:

—I was thinking about the experimental models.

Bryan:

Yes, experimental models —

Weart:

—computer models —

Bryan:

—well, computer models were very little in the —the only work done in the fifties that I know of was done in the Soviet Union by a man named Sarkisyan, ...so the use of computers really hadn't you know, had started in in meteorology, but I think in oceanography, people were still pretty much thinking in terms of steady state models of ocean circulation, so they thought, well, and of course, a lot of the early ideas were kind of linear too, you see, so —

Weart:

— so it became a sort of a ?

Bryan:

—well, it became, I think it became a little bit of a playground of sort of applied mathematicians, you see, to come dream up these linear solutions for complicated geometries and things like that.

Weart:

Yeah, they'd have an idealized ocean basin, and —

Bryan:

—yes, that's right, that's right, that's right. Professor Carrier at Harvard became involved at that time. And Professor Morgan at Brown University, and —but then later people realized that nonlinear effects were important there, and they —

Weart:

What was the opening wedge to that? What was the first nonlinear effect?

Bryan:

Well, one was a problem which we now realize is a rather artificial problem, but it was really, if you took simple shallow water equations or tidal equations, and then you had the wind pattern blowing on it, you see, what would be — Stommel had a simple sort of linear solution — what would the nonlinear effects be? Yes, and people felt, well, gee, unless we understand that problem, we can't really proceed further.

Weart:

That's interesting, because there aren't that many parts of the world where that would be crucial, I would think.

Bryan:

Well, actually it really turns out to be a somewhat artificial problem, but it was, you might say, a kind of a very, a simple sort of model of the ocean, you see. And people thought, well, unless we solved that simple one, we shouldn't go on to the next step. And that's what sort of motivated their —

Weart:

— was there any connection here, I'm just guessing, because they haven't — was there any connection with things like modeling lakes, that kind of problem?

Bryan:

Well, now, a lot of the lake modeling was concerned with —it was a little bit more like tidal modeling, like ? set up on lakes, and of course, at the same time, the parallel development was of course trying to model the earth's tides you know, on the earth, which —ocean tides, yes. But that was somewhat independently, yes, because that is dealing with a more high frequency sort of thing, and that is more governed by linear equations.

Weart:

OK, so then as you started getting into numerical ocean modeling, nonlinear and so on, what are the groups, obviously one here.

Bryan:

Yes, yes, yes, that's right, I came from Woods Hole down and joined this group here.

Weart:

You came here with the intention of eventually coupling it with an atmospheric model or just?

Bryan:

No, that was really —Joe (Smagorinsky) —Joe's intention I think from the first. He really had very much the concept —which by the way is actually of course outlined by L.F. Richardson, you know, in his famous book.

Weart:

Right.

Bryan:

He really had the concept of the climate system as being integrated, an integrated system including the oceans, and actually I think, I wasn't — Joe had tried to get various people, including George Veronis, to come down here and so I was, I think I was very fortunate to get the opportunity to join the lab. And I'd met Manabe the summer before. He'd come up to Woods Hole to take the first summer GFD course that they had up there.

Weart:

So during that period, those first five years or so, were there any other groups that were attempting to do anything like that?

Bryan:

Yes, there was at the National Center for Atmospheric Research, they had first Gunther Fischer, who's just retired from the University of Hamburg, he was working on this same problem. And then later, Jim O'Brien who is now down at Florida State University. He was also working on it there.

Weart:

So your paper with Manabe was '69, was still the first, but there were other people who were working on this around that —

Bryan:

—yes, they were sort of gearing up to it, yes. Yes, I guess they weren't —I think they were just trying to maybe construct models of the ocean that worked.

Weart:

Ocean modeling.

Bryan:

Yes. But I think they also, I'm sure at NCAR they had the idea for a couple of models too.

Weart:

I suppose as soon as one gets an ocean model, one then tries to couple it with the atmosphere? Or do people just go off and do ocean models alone?

Bryan:

Well, I think there are quite a few that are, you know, just interested in it from the oceanographic applications too. There are certainly many applications to not only studying the physics of the oceans, but the chemistry, and of course we've done a lot of that kind of work here too.

Weart:

OK, then, let's move on. We're going over this very quickly, but let's move on to the seventies then. Or let's go the other way. First let's talk about modeling and then about the rest of oceanography. So what happened to ocean modeling in the seventies? There was your work, obviously, but what else was going on?

Bryan:

Well, the computers of course were gradually getting more powerful there, so that enabled us to — just at the end of the sixties, we'd gotten into some three dimensional models. Then we could go a little bit further into it, you know, and go from the very idealized basins, to try to get more, so that we could compare with observations. And that was really a major effort in, I'd say, in the seventies, to try to get more realistic simula, but of course, at that time there was really a lot of criticism of what we were doing, because the oceanographers were making this real effort to look into mesoscale eddies. So the validity of any kind of coarse scale ocean modeling was really called into question.

Weart:

The eddy would never show up.

Bryan:

That's right, it couldn't possibly resolve that well, kind of where you get ... (crosstalk)…you know, and weather, and of course people felt well, you might just be leaving out, you know, the most important part of the physics of the ocean. And so that really made it quite hard even to publish actually a lot of that work, at that time.

Weart:

Who was working on this besides yourself?

Bryan:

Well, let's see. There was a group. There were efforts going on at Lawrence Livermore Laboratory, a man named Crowley out there, and I think he was working with Chuck Leith, quite well known, who was at NCAR, now is back at Lawrence Livermore, and then at —in the seventies, oh, Bill Holland who was here at GFDL and later went out to NCAR, he is still very active at NCAR in doing this ocean modeling. Francis Bretherton was involved in certain local models, Bretherton, yes. By the way, he was one of the sort of leaders on that MODE, for the MODE. There was work on trying to model the details on mesoscale eddies there. And then Bert Semtner who was a student here went out to California, University of California at Los Angeles. He did some very interesting work.

Weart:

Did they take the code with them or did they go and build their own model?

Bryan:

Well, in the case of Bert Semtner, he completely kind of rewrote our code and then we got back from him really what he'd done. He really wrote the first code that was really suitable for the vector computers, vector operations there.

Weart:

Had to be done rewrite.

Bryan:

Yes. And the codes we have now are really based on what he had done. Also he made, while he was out at UCLA, he wrote the sort of first sort of detailed documentation of the code, and giving examples there, and we have used what he did as sort of a model of how to do the documentation for other people, and I think this has been a great help to make the code very popular with people around the world now, because they could learn, you know, he had a regular tutorial you know that they could —

Weart:

—they could learn and they could play with it and change it —

Bryan:

Yeah, that's right.

Weart:

—which is hard to do otherwise.

Bryan:

Yes, that's right. That's right.

Weart:

One of the things I'll be interested in doing as I get more into this is trying to track the genealogies of some of these codes.

Bryan:

Oh yes.

Weart:

Because they breed, they exchange genes —

Bryan:

—yes. Yes. I think this is very important now, is to get a much better system of exchanging these codes around, because that's where a lot of redundant work was done. I mean, I hear about people who wanted to go into this work and they would spend two or three years and then maybe they'd lose their funding before they even got off the ground, you know.

Weart:

Well, what else was going on in oceanography? That's the modeling.

Bryan:

Well, if course, I think that the results really coming in from the transient tracers were bringing in another point of view, that sort of counterbalanced the people that were kind of preoccupied looking at mesoscale eddies, and of course the MODE people really found that their main results that they were getting applied to gravity waves and things like that. They found that they set out to measure mesoscale eddies, but the results they were getting actually applied to, I mean their best results were really defining the spectrum of gravity waves. And this was also followed up by Professor Munk, who did a lot of work, kind of organized the, made a model of gravity waves in the ocean, which are amazingly, the spectrum of gravity waves in the ocean is amazingly sort of self-similar from one part of the ocean to another. People now see there are more differences than they thought before, but there's something to do with the thermocline, it sort of traps energy from going above, and so that it builds, it builds up energy to certain limits. And so this was really I think a beautiful work, to kind of organize, and so the final outcome, I would say, is a much better sort of statistical sort of grasp of the kinds of motions and the spectra of perturbations and temperature, things like that, in the ocean.

Weart:

Now, you're talking in terms of different scales, in terms of spacial scales, but when you talk about transient tracers, that's another thing, that's a very long time scale.

Bryan:

Well, the great thing about transient tracers was that before the people had a kind of a frozen concept of the ocean, because it was sort of a geological thing. You know, you would see these tongues of salinity moving, you know, appear to be moving down in the ocean, but there was no concept of what the rate, you know, of spreading really was.

Weart:

Because you couldn't measure it.

Bryan:

No, that's right, that's right. You could see sort of evidence of something going from A to B, but you had no idea of how fast.

Weart:

It could be two years or two thousand years.

Bryan:

Yes. And the thing with the material coming from the bomb tests into the ocean, you knew that it all had to go in about from '58 to '60.

Weart:

It was a wonderful experiment.

Bryan:

Yes. And then it only had that much time to penetrate down into the ocean. And it really showed that the upper thermocline of the ocean could be invaded by stuff coming from above, on the order of about ten years.

Weart:

Before that people just didn't have a clue as to the time scale.

Bryan:

Yes. That's right. That's right. And of course people were very eager to interpret that in terms of sequestering of carbon dioxide in the ocean there. And I guess quite early, the controversies over how much was going into the biosphere and how much going into the ocean started up. I think—

Weart:

— late sixties, I guess. There was a real controversy in the late sixties, early seventies.

Bryan:

Yes. Yes. It didn't seem to me it's progressed very far. They haven't resolved it.

Weart:

Yes, they just add more and more different possible mechanisms.

Bryan:

Yes. I think they are narrowing it down a little bit, you know, due to (crosstalk)...

Weart:

—the first few years, when the order of magnitude and the sine were totally —

Bryan:

—yes, that's right.

Weart:

People are still coming up with mechanisms.

Bryan:

Yes. I think there still could be some question of sine, I guess, really.

Weart:

Yes, well, but that's because the range has narrowed down. The (crosstalk)

Bryan:

—yes, yes, yes—

Weart:

—... order of magnitude has narrowed down a bit. But I think there's still a possibility of somebody coming up with another mechanism that hasn't

Bryan:

—that's right, that's right, I think that we're still you know, this still is very much a frontier thing. But —what did this, of course, the few chemical tracers really turned out to be a tremendous help for us in ocean modeling, because it gave a kind of integral type of constraint to our models, which was extremely valuable.

Weart:

You had to make sure that your models reproduced something.

Bryan:

Yes, that's right, and something we could handle with our models, which were still, you know, relatively crude, because they're not resolving these mesoscale eddies. See, a lot of people said, well, you know, we know in the atmosphere that cyclones, any cyclones are the main sort of driver of the whole circulation. How can you have the courage to try to model the ocean with a model where you're not really resolving something like that?

Weart:

Right. Well, you parameterize it, but the advantage of this slow stuff is that you're talking about things that happen very slowly over a long period of time.

Bryan:

Yes, that's right. That's right.

Weart:

You can just parameterize it and work it out. Whereas if you try to measure, try to match up with something that happens on a much shorter time scale, then you're much more—

Bryan:

Yes, that's right. That's right. But I think more —

Weart:

—ocean modeling has never attracted as many people as atmospheric modeling, is that right? (crosstalk) —of course, no weather —

Bryan:

—well, you see, actually, they are fairly limited, the places in which this kind of thing could be pursued. For instance, there's no real effort on modeling open ocean circulation at Woods Hole or Scripps, for example, and they're two of the major institutions in the field. And it's only, it's only really, although they do work in coastal modeling, of course, especially at Woods Hole. At Scripps I don't think they even do that. And generally, this field, I think was looked at with deep suspicion by many of the oceanographic colleagues, as a kind of a premature sort of thing.

Weart:

They still seem to be data gatherers.

Bryan:

Yes. As a matter of fact, I think the feeling in the WOCE thing, that's sort of the latest, you know, expedition for the period 1990-1995, major effort to measure the world ocean circulation, is that it's really somewhat premature to model the ocean until the results of that, you know, effort are in, so that would be 1997 probably at the earliest.

Weart:

I see. If you were talking about the eighties, have there been new planets, sorry, ocean modelers coming along in the eighties?

Bryan:

Well, of course, one thing that's happened now is that there's much more effort to do high resolution ocean modeling.

Weart:

— take a part of the —

Bryan:

—yes, that's right. Now, Mike Cox who is at our lab, he did some very important work in that area, and then out at NCAR, might be interested in going out there some time, they've, you know, Holland and Frank Bryan have just completed you know high resolution models of the Atlantic that include the mesoscale eddies, so it has that more the full range of significant scales in the ocean. More like meteorological models. We're getting really something comparable. And Bert Semtner, who's now out at Monterey, had worked with Bob Turbin at NCAR to do a global model with 50 kilometer resolution.

Weart:

That's a lot of computer.

Bryan:

Sure, yes, a lot of computer resources. But of course, now they're really a great deal compared to what's done, you know, in atmospheric modeling, you know, for something like the European Center Model in England, you know, has I think they're down to less than 100 kilometer resolution for their daily forecast there. So we're, because of the operational requirement in meteorology, they of course really use a vast amount of computer resources.

Weart:

What about ocean atmosphere models? What would your competition be?

Bryan:

Well, let's see, there's really the main, and for coupled models, they had a group at Oregon State University and at NCAR, really Washington and Bert Semtner of course were, in Washington, Bert Semtner, were working together out there at NCAR on coupled models. I think we've had more in the way of computer resources than a lot of these other groups, that has given us quite a significant advantage I think, in doing that.

Weart:

NCAR has the computer, but when it comes to time that's been—

Bryan:

—yes, that's right, that's right, that's right.

Weart:

Where has the funding flow mainly come from? I know about ? but — oceanography in general?

Bryan:

Well, a lot of these two other groups that I'm talking about —by the way, Hansen never really went in for the ocean part of it. I think now he has seen the light, I hope, I guess, but he's also of course working on extremely limited budgets, you know. It's really amazing what Jim Hansen has done really considering the resources available to him. There also of course has been work with coupled models on the El Nino problem.

Weart:

Yeah.

Bryan:

And this I think is very important, as it's developed a look at air-sea interaction. And of course partly it's the Greenhouse Effect, that is air-sea interaction too, a very complex one we don't understand yet, but—

Weart:

—El Nino manifestly can't be understood without the coupling.

Bryan:

No. No. No, that's right, and —yeah, and so that is certainly a very good test bed, you know, for our understanding, and the geochemistry is a very good test bed also, you know. I feel that, a lot of this work at Oregon State and at NCAR is supported by the Department of Energy.

Weart:

Because of the Greenhouse Effect?

Bryan:

Yes, I think that's the principal motivation for that. Kuminoff, Fred Kuminoff, you know, is head of that.

Weart:

Didn't we miss something? You've mentioned MODE and then WOCE.

Bryan:

And GEOSECS.

Weart:

GEOSECS, yeah.

Bryan:

But you have to, you know, it's rather interesting. I don't think — there was another expedition, TTO. Yeah, Transient Tracers in the Ocean, I think it was, which was about the end of the seventies, very early eighties, yeah, and that was sort of an organized expedition to get more data, and they also repeated some of the IGY sections there, which showed some very significant changes, by the way, in the Atlantic Ocean, which was I think a historic first because people had thought of the oceans as pretty much stationary things, the water masses of the ocean. And this —the work which was Karl Wunsch was —I should mention just the concept of WOCE was really, came out of the world climate research program, sort of a follow-up of GARP that I think it was very forward looking, the planning, it was felt that you know to really make further progress on the climate problem, we had to have a better grasp of the ocean. And they were more or less searching around for a way to put this program together, and Karl Wunsch came up with the idea of building it around the satellite altimeter, and that —I think it was a very important thing, you know, leadership really on his part, to really come up with that idea, because the satellite altimeter, you see, you can look at sea surface temperatures, but the altimeter gives you a kind of pressure map of the surface of the ocean, so that enables you almost to look down into the ocean, which — (crosstalk) — yeah, really gives us an idea of the pressure gradients within the thermocline itself. And so it's a very very important technical development, and I think it will sort of break the log jam of ocean modeling, because a lot of people felt, well, what sense was there in building models of the ocean, you have no data.

Weart:

Nothing ...?...

Bryan:

—yeah, and of course I felt that wasn't quite true, that we still had a lot of water mass data and geochemical data that we hadn't been able to explain yet. But nevertheless, the altimeter does allow us to check our models of course in quite a bit of detail.

Weart:

OK. Let me move on now to see if you can give me any ideas about where documentation of these kinds of things can be found. Obviously I can go to places like Woods Hole and NCAR, where I have been actually for data, and talk with some of these people, but more generally, how much does one rely on letters and reports? How much does one rely from the fifties up to now and how do you think it may have changed? Would there be a lot in letters, or has it, have you been isolated from other groups and haven't interacted with them?

Bryan:

Well, let's see. In terms of —now, I think that you could find quite a bit of documentation on some of these large organized efforts, because so many proposals you know were written, and then, well, I guess the proposals probably would be more important than the letters, actually, because I don't think that these —that the — I guess in the old days the power structure was such that one person would write to another person and get something going, but in the more modern times, it takes more of a group proposal is required.

Weart:

What about in terms of scientific results, ideas, criticisms? Do you correspond much or do you —?

Bryan:

— well, one thing that I think you might be interested in is that meeting we had sponsored by the National Academy in the early seventies, that was on numerical models of the ocean. And it was sort of the first kind of organized conference of that kind.

Weart:

The first time the guys all got together.

Bryan:

Yes, and what we had, the format was kind of interesting. I think it would be impossible to organize a meeting like this now, but we insisted that all papers be written up ahead of time and submitted, and we assigned two people to discuss each paper, and they had read the paper before coming to the meeting. This wasn't true 100 percent but it was, you know, almost 80 percent. And then the —I told them at the time we were organizing the meeting, get a court reporter you know to type it all up. They didn't do that. They tape recorded it. And then they spent about two years unraveling it all, and I think three or four secretaries quit their jobs at the National Academy while they were doing it. But I think we ended up with a very nice record of the actual comments, you know, made at that time. I'll show you the book after our interview, and I think you could follow it up maybe even with some telephone calls to some of these people, what they recall, yeah.

Weart:

One has the feeling that the essential interactions take place at things like this, at conferences and so on.

Bryan:

Yes, that's right.

Weart:

And not necessarily over the telephone, you wouldn't really talk to an individual?

Bryan:

We told everyone, you know, that we were going to record all the discussion, and at first people thought, well, boy, this is going to be very inhibiting. But it didn't turn out that way. People just said what they wanted. And they also, they had a chance later of course to edit their comments. So of course it turned out, there are some non sequiturs in there because people edited out their questions, you know, and left the answers.

Weart:

Especially two years later. I know. I've been in that situation.

Bryan:

But still I think it's quite interesting. The British Meteorological Society has meetings at which they record the discussion. You know, from the historical point of view it's very interesting.

Weart:

I was very interested by the discussion following Calendar's 1938 paper, it's published —and it was a very interesting discussion followed his paper. They all very patronizingly told him what a lot of work he had done.

Bryan:

Oh really? Oh yes.

Weart:

But he considered the urban heat island effect and everything.

Bryan:

Really? Oh yes. Yes.

Weart:

Well, OK. Proposals. You yourself don't have much proposals, right, they're just planted by the lab here or do you sometimes make proposals?

Bryan:

Well, let's see, of course we're now being a little bit forced into the proposal mode now, because that's the way the mill is sort of going. They want to go into much more like other organizations. There'll be proposals and peer review, different projects. But it is true, in the past we have been block funded there, so that I haven't got really you know, I'm trying to think of —

Weart:

—on the other hand, you may very well in the past have gotten proposals to review.

Bryan:

Oh yes. I've done a lot of reviewing.

Weart:

—? has not saved many of its review files, especially for proposals that were turned down. Let me ask you, do you have old files of things like that, of your correspondence?

Bryan:

I do have files of my correspondence. I've saved pretty much, yes, I've saved my correspondence.

Weart:

It's useful for historians, don't throw it out now.

Bryan:

Yeah. No, I'm a great believer in saving letters. I've got, as a matter of fact, my house is filled with family letters.

Weart:

One of these days I want to talk with you about —not right now, but one of these days I want to talk with you about what's to be done with that. One of the tasks—for example, we've recently arranged for King Hubbert's papers to be deposited.

Bryan:

Yes. As a matter of fact, there's one thing which I did in my life which I've always regretted. My father died when I was still, you know, fairly young, just in college, and you know, I was sort of grief-stricken at the time, and somebody gave me the job of cleaning out his office, you know. And at that time I wasn't, you know, old enough to be aware of the importance of saving everything. I threw out a lot of his correspondence. And you know, he was a geologist and corresponded with a lot of people all over the world. And I've always regretted that. It really was a stupid thing to do.

Weart:

Yes, but it is frequent, usually for the reasons you said. What about old computer codes? Does there exist the code from any of your early ocean models, for example?

Bryan:

Yes. Actually I have saved some of that.

Weart:

It could be —you know, these are artifacts too, you know. It's nice to save a piece of an old experiment. Would it be possible, is this in Fortran?

Bryan:

Yes, in Fortran.

Weart:

In principle it could be run on somebody's home computer, probably.

Bryan:

Oh yes. Yes, that's right. Actually you know, that would be actually, I have been thinking of doing that for our ocean course here, is to try to — unfortunately of course it takes a lot of work to kind of work all those, you know, the kind of course here that is trying to use various old model results, because they, I'm certainly, of course —and actually home computers, it's a little bit hard to do some of this work, but work stations are certainly becoming powerful enough.

Weart:

Yes, and in a few years, it depends on what mean by a home computer, but a 486 or 586 —or whatever —

Bryan:

Yes, that's right.

Weart:

And you know, if you can run it in background, then you don't mind any (crosstalk)

Bryan:

—sure, that's right, that's right (crosstalk)

Weart:

—just put it on ?... and nobody's charging you for the computer time—

Bryan:

—yes, that's right, that's right, of course. Of course for students you'd like more instant gratification, where you'd be able to get it a little bit quicker.

Weart:

Yes. But you could still do it. I think it would be very interesting. It's a way to preserve these artifacts, and in a way, just as you have downstairs here, you have some of the old artifacts, physical artifacts from early oceanograph or meterology or whatever, when one tries to think how to preserve a computer program, it's go to be preserved on a machine. It has to have something to go on. Then you have that old artifact.

Bryan:

Gosh, I wonder if I have any of my old? I ran my original codes for a run on the Whirlwind computer at MIT, but I don't think I saved any of my codes. Of course, that would have been a peculiar machine code.

Weart:

That's pretty hard to preserve. You can save the thing and look at it, but you can't feel it in the sense you can preserve a code that was in Fortran.

Bryan:

Yes, that's right. See, I did my thesis work with Ed Rench you see and ran up a simple low order model of some of the early work that he was doing, you know, in looking at the climates of simple systems.

Weart:

I see. So could some of this old Fortran from the sixties or early seventies be readily adapted?

Bryan:

Sure. Yes. Certainly I could Xerox it off, because it isn't, I think it's no more than 500 lines or something like that.

Weart:

Is that so?

Bryan:

Yeah.

Weart:

How does that compare with the atmospheric models of the same period?

Bryan:

They generally were a lot more lines there, because of course, what I was doing at that time was simple sort of two dimensional flows there, you know, very very simple. And they were already going three dimensional. Well, Joe, you see Joe pioneered in building up a three dimensional type of climate code, with the physics of radiation and water, and many things now that are in general use at forecasting centers all over the world.

Weart:

Well, let me move on to one more thing. It's a pity to be sitting here with you when you've just published this paper in NATURE, 7 December 1989, —

Bryan:

We were just looking at that yesterday. A friend of mine came down and said, "Oh, I hadn't seen this. So many things are published in NATURE these days!" I said, "What a put down."

Weart:

This has just happened, and it's fresh in your mind. First of all, who is Stouffer?

Bryan:

He's a long time staff member at the laboratory here, Ron Stouffer. He's been very much involved in this work, in a couple of models, you know, really, over the past—

Weart:

When was (crosstalk)... first?

Bryan:

Ten years. Well, you know, we—he's a very important member of the team there, and we change around, you know.

Weart:

Take turns. I see. So this is, well, it's an atmospheric model coupled with an ocean model. Does this mean you take responsibility for the ocean model or is there more sort of back and forth?

Bryan:

Well, actually, yes, the ocean model itself is not—you know, we keep putting in different improvements, but that is how we sort of divide up the work, you know, that we work on the ocean model and put in various improvements, largely based on comparing with the geochemistry, the water mass, and working with Keith Dickson. There are several of us up here, and Robbie ? the geochemist there, and then, that's coupled in with the meteorological model, and the rather simple model of sea ice too, that's another component.

Weart:

Simple model of what?

Bryan:

Sea ice. Sea ice, yes, see, that's another complicated component.

Weart:

Is that given to somebody separately to work on, or?

Bryan:

Well, actually, we had a very good person who was working in that field, Bill Hibler, who's now at Dartmouth there. As a matter of fact, he's sort of a pioneer in that field of sea ice modeling. And we worked on that together. Now, the model that we use isn't quite as complex as the one that he came up with, but I got a lot of good ideas from working with him.

Weart:

And what about the interaction? How do you and Stouffer work out the question of doing the interaction?

Bryan:

Of linking them?

Weart:

Linking models.

Bryan:

Well, that's really based on our experience, you know, going way back to when we first did this coupling, you know, back in the sixties really out there, that we worked out various ways. And one of the most difficult things is to bring the system into equilibrium to start with. There we want to get a good climate equilibrium, because you know the Greenhouse Effect is only a small perturbation in the heat budget, something like 4 watts per square meter, and so you have to work quite hard to get an equilibrium that's comparable to that.

Weart:

Equilibrium meaning, you can run it for a long time and it won't change?

Bryan:

Yes, because you see, usually, if you just have a small slow temperature change in the ocean, it can amount to quite a bit more than the transfer between the ocean and atmosphere per watts per square meter. And that's one thing that we've kind of stressed. I think that we're ahead of other groups, in that they've often tried to run these experiments, and they're still at a kind of transient stage, and so, they're trying to sort out the Greenhouse, bring out the Greenhouse Effect from a series of transients, you know, in there.

Weart:

I see, so you have to run it for quite a while before you start—

Bryan:

—well, we've also worked out a way of kind of speeding up the convergence there, because we found, that's the difficult thing about the ocean, it has very long time scales to it, you see, that are actually thousands of years actually.

Weart:

Right, it's still responding to the end of the Ice Age.

Bryan:

Yes. So with the atmosphere, it will come to adjustment in about a year's time, so just this kind of forward modeling that we do, just times ? the equilibrium, is perfectly feasible with the atmospheric model, but it's not quite, not really feasible for an ocean model, and it will be more and more true as we go to higher and higher resolution.

Weart:

You have to see which way it's going and anticipate it.

Bryan:

Yes. We have to work out a more sophisticated approach to it. We really haven't found, I would say, the most efficient way to do that yet.

Weart:

Right. Now, when you're making this model, what drives you to make adjustments? Do you always want to change your parameter here or change something there? What is the data or the effects or whatever that?

Bryan:

Yes. Well, the kind of data of course I feel that's most available, you know, for verifying the model, are what are called the water mass characteristics of the world. In other words, that's the temperature and salinity structure of the oceans which you know, we have these results from these expeditions, and we know that they are changing slowly, like in the Atlantic there, where we're close to sources. You know, there's water mass formation there. But in other parts of the ocean, they're probably changing far more slowly. So that we can consider these large scale features as something to try to reproduce in the ocean model. I have to admit we're getting closer to it, but we still haven't really reproduced... (off tape)

Weart:

... how your model differs from the salinity or whatever.

Bryan:

Temperature and salinity. Then of course another source of data which is very important as far as the Greenhouse modeling is concerned are these transient tracers which I mentioned. One of the things for instance we're doing now to kind of check out this model we just published in NATURE, we're running the same model with the two species of the freons, the chlorofluoro carbons in it, and we're gradually getting data measurements of how those have penetrated down into the ocean.

Weart:

Aha, so you know when they went in.

Bryan:

We know more or less when they went in, quite uniform globally apparently, and we know the physics of the transfer fairly well, and so I feel that this is going to be —yes, we don't have as much measurements as we'd like because it's fairly recently that they came up with a, Lovelock came up with the, you know, the idea of how to do that.

Weart:

Right.

Bryan:

But it's a beautiful measurement, because they get it, you see, the other radioactive tracers, it takes sometimes about five years to get the data back, because they're counted you know with a geiger counter. They take huge vessels of water, you know —

Weart:

I see, it's a long —

Bryan:

—yes, it's a very very tedious sort of thing. While the freon, they get the measurements right away. But this gives us an idea of how something received at the surfaces of the ocean, the vertical pathways going down, and that's precisely what we need to see how the oceans will buffer the Greenhouse Effect, you see, how heat will go down into the ocean.

Weart:

Right, so that obviously is one of the driving forces.

Bryan:

Yes, that's right.

Weart:

That's one of the things you have to worry about. The other thing that you talk about here is the North Atlantic thermohaline circulation.

Bryan:

That's right. That's right.

Weart:

To what extent is that a driving force that affects your model?

Bryan:

Well, the North Atlantic seems to be a very very important sort of engine of the whole engine circulation, because apparently lots of the deep water that appears in other parts of the world ocean really is generated at this one spot here. And then apparently a lot of the great climate variations of the past I think are linked, you know, associated very closely with changes of North Atlantic circulation. And we still don't understand, you know, exactly how that occurred. But that's kind of a frontier area. We know that if you just look at the Ice Ages very circumstantially, you know, the great ice sheets are all clustered right around the Atlantic. And you see less, much less in the Pacific area and Southern Hemisphere (crosstalk)

Weart:

And also the (crosstalk) ... very strong changes...

Bryan:

—yes, that's right, that's right.

Weart:

So being able to reproduce this is an important part of your model.

Bryan:

Oh yes, yes. I think that —but the critical thing that we of course like to verify is, you know, this very deep mixing between surface and deep waters in the Southern Hemisphere. See, some time ago, we had, I think it was called —Smagorinsky headed a committee of the National Academy looking in to the Greenhouse Effect, and it was sort of a follow-up of the early Charney Report, and Francis Bretherton was one of the members of the panel, and at that time he strongly believed well, you really wouldn't have a Greenhouse Effect because all the heat would go down into the ocean.

Weart:

When was this?

Bryan:

This was in the seventies.

Weart:

As recently as that.

Bryan:

Yes. Yes, I think it was in sort of the middle or late seventies, I think.

Weart:

So still then, people could believe that it was a —

Bryan:

—yes. Well, at that time we tried to talk him out of it on the basis of tracer information which at that time we had mainly from the Northern Hemisphere. See, we mentioned well, tritium is mainly up in the upper thermocline here, so you can see that only a little bit of it is getting through. But at that time we didn't have the tracer data for the Southern Hemisphere, because that sort of, the GEOSECS process was rather slow, in coming out, becoming available. So it really turns out that he was right, only for the Southern Hemisphere.

Weart:

There is this circulation there.

Bryan:

Right. But it's still very controversial.

Weart:

(crosstalk) ... is warming, the circulation goes much faster.

Bryan:

Yes. This is still very controversial, and other groups are not getting this result.

Weart:

There's a certain pull of the Southern —

Bryan:

—yes, that's right, that's right. The NCAR group sort of claims that they've got this result, but I think if you read their paper, it doesn't really —

Weart:

—quite —

Bryan:

Yes, yes.

Weart:

You say this is controversial and it's even of public interest. Another interest of mine or concern is, to what extent, the fact that this has all become of such public interest, to what extent do you think that this is now affecting the science?

Bryan:

Well, it is a problem, I think, in becoming —you know, becoming almost a political question. It's like, if you're conservative you don't believe in the Greenhouse Effect, and maybe if you're a liberal you do, or something, which is kind of stupid, isn't it? I mean, from the scientific point of view. I believe that people are kind of capitalizing on that a little bit, you know, to —

Weart:

— do you think it affects what problems people choose to work on or what questions they'll choose to ask their models?

Bryan:

Well, I guess that I really, you know, I'm sure it's bound to, yes. I don't know just in what way though, right at this point.

Weart:

Has it had any effect on you, the fact that there's all this public interest in the Greenhouse Effect?

Bryan:

Well, I think, to me, I do find it sort of deeply satisfying, that we had the opportunity to start on, you know, sort of the basic sort of background work in this area so long ago, and that I didn't have to, you know, switch my field because of some funding problem, or go off and do something else. I've been able to —it's kind of frustrating to me that I'm sort of at the end of my career instead of right at the beginning, but I guess I've got a few more years to go at it.

Weart:

That just makes it easier to pursue what you wanted to pursue and what you had been pursuing.

Bryan:

Yes, that's right. That's right. I think that we were just very fortunate really, you know, in the laboratory here. And then of course at Duke, that Dr. Smagarinsky really deserves a lot of credit for being so far-sighted really, in pursuing this thing, I'm sure with a lot of difficulties in getting the funding going and, you know, protecting the project at various critical junctures there.

Weart:

I think we'd better stop now.