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.
We encourage researchers to utilize the full-text search on this page to navigate our oral histories or to use our catalog to locate oral history interviews by keyword.
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 Joseph Farman by Steve Norton on 1999 October 11,
Niels Bohr Library & Archives, American Institute of Physics,
College Park, MD USA,
For multiple citations, "AIP" is the preferred abbreviation for the location.
In this interview Joseph Farman discusses topics such as: the British Antarctic Survey and the discovery of the Antarctic ozone hole; Brian Gardiner; Jonathan Shanklin; ozone layer depletion; stratospheric ozone; chlorofluorocarbons (CFCs); creating 2D and 3D models; Susan Solomon; working in the aerospace industry on guided missiles; World Meteorological Organization; F. S. Rowland; Mario Molina; Imperial Chemical Industry (ICI); Total Ozone Monitoring Spectrometer (TOMS); Adrian Tuck; Shigeru Chubachi; International Geophysical Year (IGY); Al Flagg; National Aeronautics and Space Administration (NASA).
Today is October 11, 1999. It's just after noon. My name is Steve Norton, and I'll be interviewing Joe Farman regarding the discovery of the Antarctic Ozone Hole. Just to go over what we had talked about on the phone, the interview will be taped, transcribed, and once it's transcribed you'll have a chance to make corrections and/or additions before it's archived at the Neils Bohr Library at the American Center for Physics. Is that okay?
Sure. No problem.
First off, could you go over what your educational background was and sort of like the path you took to get to the British Antarctic Survey?
Yes, sure. Basically I was in my first job, which was guided missiles, believe it or not.
So you were in the aerospace industry?
Yes, sure. And an advertisement appeared to work in Antarctica. I had climbed and skied, and I didn't have anything to do at that point, so I thought, "I might as well apply now." So I went down to a thing called International Geophysical Year...
Was this in 1956?
1956 when I was interviewed for it. And I guess we ran... say from during the IGY, which did everything. I mean, I was in Winchester [?] for a part of it.
And which station were you at?
This was Faraday. Actually, I called it Argentina Islands. It was Faraday, now especially because the Ukrainians have got it, which is a wonderful thing for our countries to do. Never mind. So we ran this, and then sort of came back. And, of course, pressure was on internationally having gone through all those exercises of setting up these sessions in Antarctica. Perhaps it was a bit short-sighted to stop, so it was decided to carry on. I was offered the job of overseeing the observatory physics program in Antarctica, which since has become the British Antarctic Survey.
When did you actually go to Faraday and how long did you stay?
I got there in March '57 and came out in March' 59.
Okay. That's how long you were at Faraday?
And you said after the first year you were offered —
When we came back from there, then when we were working up all the data and getting it ready to go to all the world data centers, it became obvious that we ought to keep these things going. And then I was offered the job of superintending the development of these observatories and so on. And this went on and on. It stopped in 1990 when they throw you out when you get too old.
They threw you out, did they?
Well, actually, you didn't get much choice. I mean, the idea that you might actually be able to help the young and encourage the young and tell them a few things what not to do as well as what to do, it doesn't seem to really enter into anyone's sort of thought. But luckily the people in the Chemistry Department were then setting up this thing called the European Ozone Research Coordinated Unit, which attempts to make sure that the money available in Europe is wisely spent. Or perhaps not unwisely spent, as we say. And they are very kind to let me sit here and tell them what they're doing wrong. It was too exciting to leave. I mean, all hell was breaking loose and it wasn't really quite clear what was going to happen.
Over at Bass?
No. Over in Antarctica and the ozone hole itself. I mean, opinions were still strongly divided on how permanent it was and what changes were taking place and so on. I feel I know a little about it now.
You actually came out in?
March '59, yes.
March of '59. So you were down there for?
Two and a half years.
Two and a half years. And so the entire time you were at Faraday Station? You didn't go to any of the other stations?
No. In those days it was difficult. You went through all the British Stations because you had to re-supply them.
On the way down there.
On the way down there and back.
But you were stationed at Faraday?
At Faraday, yes. I've been to the South Pole, but that was much later.
I've read some stuff where you were the Base Camp Director at the Falkland Island Station in 1957. That doesn't ring a bell?
I'm sorry; I was Base Commander in 1958.
In '58? And that was at the Falkland Island Station?
Well, it was called the Falkland Island Dependency Survey. Everything was, at that stage, run through the foreign office.
But there is a station in the Falkland Islands, right?
No. There used to be a meteorological station there. It's a dreadful business. South of the convergence, the politics changes. So we were originally a dependency of the Falkland Islands, and then that was all changed in '63 or something when it became British territory. There's a lot of name changing.
Yes. I was under the impression from one of these accounts I've read that there was a station actually in the Falkland Islands.
No. We did run a Dobson Bay fraternity [?] there for about six months one year, while I was waiting to go somewhere else.
Okay. But you were the Base Camp Director at Faraday?
High Commander in '58. Yes. Base Leader, what they called it in our own terminology.
And did you ever spend any significant amount of time at Halley?
No, no significant time. A couple of months. The earlier stories are very complicated.
It's interesting to hear what people are writing, right? But you were Head of the Geophysical Unit at Halley Bay Station?
No. What happened was we worked initially from London. Then we had a unit set up. The idea in those days was that each of the science disciplines would have a little unit set up in the University. I was a Senior Research Fellow in Edinburgh from '61 to '76, when the whole survey was reorganized and brought down to Cambridge.
Okay. So we got that straight. Actually, when was Halley Station —
The ground party went in in 1956. It was a separate organization developed by the World Society. The grounds party went in in ' 56. And Stan Evins made some ozone measurements in '56. Unfortunately, somewhere in between the World Society Station closing down in ' 59 and we took it over, some of his records got lost. So we've never been really able to put a good absolute scale on his measurements. But his measurements, as relative ones, did indeed show interesting things. But Dobson wrote that up. I've forgotten when. In was in '62 or '63. And he made some sensible remarks, which most people have misquoted.
Yes. So the British Antarctic Survey took over the Halley Bay Station in '59?
In '59. Yes.
Okay. The British Antarctic Survey's Dobson's are not calibrated against the International Standard Instrument 83.
What's the primary reason for that?
Well, we had a hard choice to make effectively. And basically it's when you pack up an instrument and send it 10,000 or 12,000 miles you probably do more harm than you do good, so you don't really end up with anything particularly worthwhile. I mean, (a), you can't be sure the instrument doesn't change between the calibration sight. And (b), you can't be sure it's not going to change between the calibration sight back to the same. So we really had to work on our own. And effectively. I mean, I think you can, provided that you do enough observations to actually go through the procedure, which they use to calibrate the 83-A anyhow.
Using the Langley Method?
Yes. And also, of course, it takes an instrument out of service for a hell of a long while. Antarctica is, of course, an evolution year, unless you've got lots of spare instruments. Eventually we got to a stage where we had a spare instrument, but we still didn't really think it was worth the effort of actually sending it to these places. When we came in, gosh, I've forgotten when, '86 was it, we compared it very carefully with TOMs and then agreed to within one or two-percent, which is all you can effectively do.
This was '83? '82?
' 86. The year after I was — and then we were sort of talking to TOMs people and exchanging data day by day. And you can overdraw the overpass of how they guessed the Dobson. And as I said, there were very few discrepancies. Basically, you're agreeing to within one or two percent, which is the best you can do effectively.
This raised a question I was talking with John Shanklin about, is that prior to 1985, he said he had been working with people at NASA Wallops and they had been taking measurements of ozone when the satellites were supposed to be going overhead for calibration purposes. Do you recall... ?
I don't remember that. The [???] was because we sent down some ozone signs to Argentina, and Jonathan was concerned with making sure we ran the ozone signs properly and according to the American ideas.
So specifically taking ozone profiles?
Yes, profiles. Which is a very different game from sitting with the Dobson.
Because he seemed to think that along with that there was also an issue of comparing with the satellites, too, on their overpasses. But you don't recall anything like that?
Well, no. We didn't really do that until '86. I mean, what I did, with a chap called Hamilton, who was an Antarctic explorer, was completely rework the whole series from' 56 to '76. We did the best we could effectively But it was very arrogant. I mean, it was only a couple of stations. I'm sure it was effectively done. So we were fairly confident with it. Limits existed because you always have a few problems when you're miles away from anywhere. And we reckoned that was a pretty good, homogeneous stasis there, which you could compare '76 to '57, within the limits of the Dobson ratio, which is one or two percent. And it was only later that other people started to — well, not too many as we know, but other people started looking at their own data sets in this way. And when you see what their revisions were, you just raise your hands in horror.
What their revisions were, after they —
Yes. I mean, there was one, I don't know if you've come across it, the danger on the Koshiva [?] used to update the ozone data of world. In one of their papers they produced the graph of Kadarken [?] Island in India showing a huge increase, during which anyone who knew anything about Dobson's just raised their eyebrows and said, "Sorry, mate. You don't know what you're doing. You've clearly got some instrumental problems here that you haven't come to terms with." And that was the whole trouble. But effectively, the idea really wasn't too bad. Because people were stimulated and doing things as well as they could. But that was really the first time that Dobson had actually published the recommendations for running them. They actually arrived from here Antarctica. We hadn't seen it before we went down. So everyone was still learning, in a sense, how to run Dobson's properly, to put it bluntly.
This is the manual they put out in '62?
Yes. The IGY manual, in essence. But that did actually come out rather late. But after IGY, and in particular in the UK MET office, which I am familiar with because they helped to train our people, ozone measurements were just handed too far down the chain, and the people making the measurements never used to get any feedback from the people who understood them, and were trying to interpret them from a meteorological point of view And so there was this sort of terrible dichotomy that things were being done in a rather sloppy way.
The measurements themselves?
The measurements, yes, because the people doing them had been so crudely trained and weren't given any feedback as to how they fit it into the world observations and so on and so forth. So it was really a rather dismal thing, despite having various sort of international committees to sort things being done in the right sort of way.
So how would the feedback between the observers and the people actually understanding — how would that have improved the measurements?
Because people make better measurements if they think they're doing something useful. And, you know —
Oh, okay. Just being more careful.
They look at their watch, and it's 10:00 o'clock, "I ought to go and make an ozone..." I mean, it got to the stage in the MET office where he would sort of run off and twiddle his dial and read off the dial. And they decided that they'd work up the observation in Brackethall [?]. So this was telegraphed to Brackethall or whatever. Well, it was telegraphed in the early days. But they wouldn't even bother to tell the man what the answer was. You understand? I mean, he just read something off a dial, which could be translated into an ozone reading. But he quite often didn't see that ozone reading ever.
So the measurements certainly meant nothing, and so not really concentrating —
Not thinking about what they were doing, and weren't really being interested. At least we managed to avoid that, I hope, in Antarctica. But it is difficult when you're running networks of this sort, and if you aren't careful that's what happens. Things just slip because you've split off the people who have tried to do the signs from the people who are actually making the fairly routine observation. That sort of goes the whole way through, because then you've got some preventive maintenance and all these sorts of things, your standard tests and everything else, and they just do them and plot it out and as far as who's jumping it, they just sort of blink a bit and forget about it. And what you should be doing is at least sitting down and putting the instrument right and finding out what the hell it all means. So it was a rather sad history, I'm afraid, the ozone measurements we took, until we re-stimulated it with the ozone model. Because, you know, then you can say quite definitely I think that if anyone found a trend before the ozone hole, that's a pretty good sign they were doing something silly. You know, there aren't any real measurements which show any huge trend in between that time, and various were reported. Well, I mean, indeed, when we came to publish, the first thing we had to decide was what the hell is going on. Is it real? Etc., etc. And it took some time.
In your own 1977 paper, the ozone measurements in the British Antarctic Survey was in the philosophical transactions.
In there you mentioned there was like slight downward trends in that. How did you interpret those? Were those interpreted as an instrumental error?
Basically, because they disagreed at Halley and Faraday, they were just nonsense and noise, it seemed to me. But, you know, you've got to have some supporting evidence.
So you interpret it as like an artifact of the station, or instrument —
It was within the noise error effect. Yes, sure. I've forgotten, but I think it was the Australians who were actually reporting some trends at that time. The [???] wasn't it, back in Brisbane [???], and going, you know, his suspicion was.
It was the noise.
He wasn't being naughty; he mucked up the retrieval message by making some silly error in how to eliminate the aerosols.
This was found after the fact? You said he mucked things up.
He wrote a paper which had absolute nonsense, pointing out how you could eliminate better the aerosol thing. But it turned out to be just a rather silly mistake.
Okay. So nothing ever came of that?
No. They have been re-worked, and they make reasonable sense now.
When you get the data back from the stations, what sort of errors do you — I mean, I assume that you get back all these measurements, you calculate what the ozone is, and you find some crazy measurements in there, so you start going through the data a little bit more carefully and looking for errors.
Well, both at Halley and Faraday one was really quite lucky. Because in the period when things are changing very rapidly, and you can get huge ozone changes — well, you could in those days, you can't now from day-to-day. But they were always correlated with 100 millibar temperatures. And we had upper air station most of the time at both Halley and Faraday. They closed the Faraday one, needless to say, just from the amount of interest. So you first of all, look at the 100 millibar temperatures, and if there's nothing bumping up and down in the 100 millibar temperature, you get a bit suspicious. And then you just simply have to go back and re-work the things effectively. You look at dial readings to see if they consist and so on and so forth. I mean, you can't be sure. But if you're sensible in the whole, if something stands out like a sore thumb, you ring it and leave it there for the moment.
So one of the first things you'd look for was the 100 millibar temperature. If that was changing a lot, you would expect the ozone to change?
Yes. In that season of the year, from October through December. Thereafter it's not quite so easy. But again, you still expect to see a slight correlation of it.
So if the temperature variations weren't that great and you were getting a lot of ozone changes, then...?
Then that makes you horribly suspicious.
Okay. And what sorts of specific errors did — how common where they?
Silly things would happen in those days. People wrote things down and you can transcribe wrong and that sort of thing. And if you look back at the original notebooks you can go back as far as you can.
So in processing the data you actually found transcription errors, things like that?
Sure. I mean, some you could sort of quite simply say, "There are 100." That's easy. When there are 10, that's much more difficult, and now you're in trouble. So you have to be cautious with how you reinterpret these things. But basically if you're in doubt you reject it, and if you think you can make a simple transcription you are to put it in. You're allowed to do this.
I'm actually kind of interested in this. Because there's a lot of people, like in political debates, that would like to make distinctions between data and models. Specifically it's stuff I'm thinking about what goes on with, as well as the ozone, but global warming, saying we can't rely on the models because the models are full of uncertainties. Whereas, you know, we can rely on the data. I've talked to Walter Komer and the people at Boulder, and talking with Brian and John, that the data itself seems to be modeled quite a bit too. I mean, you just don't get a measurement and say, "Oh, well, of course." So, I'm curious as to what your general feelings are about the distinctions between data and models. And to what extent, what is supposed to be raw data, has to be interpreted.
It's the price you pay for automation effectively. Once you automate then, yes, you build into that all sorts of funny presumptions. And you're quite right. I mean, one of the rather horrible things now is it's actually quite difficult to get a balloon ascent and actually get a temperature from it. Because what actually happens is they now get all fed into the big models and they're simulated and handled. And it's actually quite difficult to get back to raw data. It does exist, but it's not as readily available as you might think. So, yeah, there are huge worries in this. It has to be done, because if you want to use that data to update one of these models you do indeed want to get rid of all the nasty gravity waves and things you're not interested in. So, yes, there is a huge problem. And there are some interesting examples in history. Have you come across Hubert Lamm? He's a climatologist.
This probably ought to be off the record, but he wrote a wonderful paper once in which he re-interpreted a temperature record and invented an anomaly that didn't exist, and in order to count for this anomaly he blew up a non-existent island into volcanic eruption and created a dust pattern. It really is extraordinary.
Where is this paper?
It's in the Royal Society.
What year did he write it?
I can't remember now.
Okay. I'll have to look for it.
This refers back to the 1890s or there about. But he actually took this South American record and did something funny with it. I'm not quite sure what he did. But he actually got a temperature anomaly, which when I looked at the data, I couldn't believe it to be real.
Made an example of it.
Then he blew up an non-existent island, [???] Buva [?] Island, which we know now is just a single volcanic pile, there wasn't another island there. It was one of these navigation errors in the good old times. And so he blew this up in a non-existent volcanic eruption.
A little bit too much processing of the data.
Just a little bit of processing of raw data.
Okay. Well, in the philosophical transaction paper, talking about the calibration against the world standards, saying that it's regrettable that most of the world network is calibrated against it, since the individual station trends are not then independent.
Well, that's true. I mean, with all instruments, which sort of changed slightly. Then it depends on how well you maintain them whether or not you really do have a true unit. And if you certainly keep checking back to that, then the errors in doing that have to be taken into account. Most people think once you check them back against the center, then that's it. But it's not really. That's just one —
So you're saying the individual stations will vary over time? And by comparing back to them...
Sure. You've got some internal change, which gives you something about it, and then you've got the external change. So you've got to balance these against each other to see what's going on.
And if you just do the external checks against like the world standard, then you're sort of like —
Well, you're just accepting what they believe to be where there's a trend or not. If your instrument actually does have a real trend because you've been foolish and you've slightly displaced one of the pictures or whatever. Obviously internally.
So you could actually have a downward trend in the ozone over time that could be sort of eliminated —
Well, you never know, if it wasn't kept properly, if there was one there, or if it was badly maintained, then it would pass it's trend on to all the other people. Whereas other people at least have some independent data and should say, "Hey, wait a minute. This isn't quite the case." You must — people should — okay, it's nice to say all these things and to calibrate it, but you've got your own bit of work to do before you go there. To pay the price of going in and taking it, it seems to me.
I mean, I see there's this potential problem with calibrating against the world standard. But is there any way that, say if there was a problem with the world standard, and that's being propagated to all the other instruments, that that might be revealed somehow?
Oh, sure. It would be revealed from TOMs or other things, you know, coming. I mean, TOMs let's face it, isn't actually an absolute measurement at all. It's really a nasty cookage [?], but it works wonderfully in a relative sense. Because it's reasonably consistent, even if it's not right. Then, at least, it points to individual stations and says, "Hey, just a minute. You seem to be doing something that you didn't do before." What's coming is Gomas [?], the occultation one, which actually is an absolute measurement because the baseline is self-determined each time you do the measurement. So that will be rather fun to see how it compares with what we've been doing in the past.
How effective do you think the traveling lamps are for revealing errors in measurements?
Well, they still need improvement, but they're better than nothing, shall we say. They've improved a lot over the years. The original ones were pretty ropy.
Were there problems with the original ones?
It's just something, the basic geometry of the instrument, and the geometry of the standard lamp. You can't actually get the filament and the kerosene in place relative to the optics as you can with — You do your best, but there's no way in which you can actually relate where you put the filament, even on the other instrument. I mean, that's basically the problem. It's just simple geometry, unfortunately.
So it's primarily the relationship of the filament in the lamp to the objects?
Yes. It's how much you can fill the field underneath, and how much variation you're getting from the — You know, it's simply a question of uniformity of illumination.
So when you put the lamps on the Dobson's, then it doesn't — it does have a bracket which it goes in?
But even given that —
Even given that, it's still not good enough. Because you have to treat each instrument individually, it seems to me.
Because I guess the optics in the instrument are going to be a little bit different.
Sure. I mean, they're fine for telling you your gross errors, but they're pretty hopeless for — I don't know why you'd need them, because basically — well, I guess it's probably a sort of fault in the manual that it's suggested what observations you do. Once you have gone through the pain of actually going to the instrument and making some measurements, you might just as well stay there five more minutes. If you arrange the wave lengths properly you can actually make all the measurements you need to do your own calibrations without really wasting any time whatsoever. So there's really no excuse for people not to work up their own, just to compare it with what they get from the things anyhow. I know it always sort of seemed to me that when you have people in remote places, really it's their job to do the work, to make the most out of it.
So they take the extra five minutes to make the calibration measurements, and this can all be processed later?
Yes. That's right. Sure. I mean, it's a long and tedious business in the way these things go. I mean, you never sort of end up with an absolute answer, but you do get a reasonable answer for the year before. So you're obviously a year behind because you don't know what's going to happen the next time.
Well, I guess, that would segue into the next question, which was going to be the issue of why the British Antarctic Survey never reported their data to the World Meteorological Organization? Never got your data in the red books.
Well, actually we sent them a complete tape of our re-working. And then we were going to do another re-working and I suppose we got lazy. We just sort of felt that what's the point of sending off raw data like that when we know perfectly well that by the time we comb through it and made all the adjustments it's going to be changed again? I guess we have to sort of apologize in that sense. But it really did seem that...
There wasn't any point in reporting it.
Well, that's right. Other people weren't taking much trouble, but we wanted to. And in any case we were always going to be a year behind because the ship's got to come back; you've got to spend a year re-working through most of it. You have to realize, I guess, that we, like everyone else, were under tremendous pressure to do stuff. It was just one of the things we were doing there at the geometric stations, tide stations, earthquake stations, all the rest of it. So looking back, yeah, we worked under quite large pressure actually.
So as far as you were concerned, then the people that just report their raw data almost immediately.
I don't quite know what it's for, in a sense. You're not going to [???] errors in completion in one month's measurements. It's going to be a long and painful process and you want to know what's happened at each end of the series [correct phrase?].
So you wanted to take as much care as possible with your measurements?
Well, yeah. That's the kind way of putting it. The unkind way of putting it is to say we got a little lazy, I suppose. But the interest in ozone had died. I mean, there was all the fuss when Roland and Molina started, and then from about 1980 onwards, everyone was saying, "Oh, I think nothing is ever going to happen." I mean, the muddles now agree that nothing much is going to happen, two or three percent of the studies state, and so on and so forth.
Since you brought up the Roland and Molina stuff, you know, when Johnson suggested that nitrogen oxides released from the SSTs and Roland and Molina in '74 suggested that chlorine might reduce stratospheric ozone. When these papers came out in the mid-'70s, did you look for these trends, downward trends, in your Antarctic data?
No, not specifically. I guess, again, I mean, at that stage I wasn't interested in. I was in chemistry. That was enough work to get one occupied. And it was only later that one really — someone worked up the fact that — I mean, what was being said after all was really quite straight forward. That all of this was going to happen in the tropics and the last place to look for ozone depletion was in higher latitudes. And so on and so forth. And so one was aware this was going on. But it didn't really sort of trigger one to actually do anything, I'm afraid. I mean, since we had already published what we thought were trends through the '60s, and in '76 when we thought the things were reasonably well calibrated. And I guess it's a bit like the other side of it, that we also had our eye on the meteorology of Antarctica. And then when we published the ozone hole paper it was painfully obvious that winds, temperatures, whatever, hadn't changed over all those years. So the dynamics came along and it was like, "Yes, it could be dynamics." And the answer was, "Well, it might be, but it's going to be funny dynamics because it really hasn't yet produced anything." And so when the temperatures in the ozone hole got low, everyone was, "Oh, everything is getting cold." To which we replied, "Don't be silly; it's just they're not warming up anymore." That's probably what they meant. But it just seemed to me silly the way they were saying it. The change was indeed obvious. It was simply that that area no longer warmed up because the sun came back, and so on and so forth. And so we weren't completely blind to all these things. Then we had to say, "Gosh, this must be chlorine." And then there was this really nasty joke, which really was to irritate ICI, as much as anything else. Because they rang me up and said, "You can't publish that." I said, "We can publish what we like." It's not good science, but —
Yes, Imperial Chemical Industry. They saw a preview of the paper quite unnaturally.
Oh, okay. They reviewed the Nature paper?
They were shown a copy of the Nature paper before publication, in which they wrote back and sort of said they didn't think we should put the CFC part in, that it wasn't good science. And I said, "No, it wasn't good science."
But, "We're putting it in anyway."
"I'm afraid I'm going to put it in anyhow because I really do believe this is it." And I can't see anything else to change it. And what I'm always saying about putting the CFCs in is that, "You have changed the atmosphere, I'm afraid. And that's the only change at the moment I can see in Antarctica."
So the IC said that. I'll come to this again later. So you actually didn't look in your data at all, like at any point in the year, to try and see if there was any sort of reduction due to —
That's what I'm saying. If you read that paper, you'll realize what I said. And I still maintain that if you are expecting a trend because the ozone factor is being hit, then basically you look at the quietest time of the year, not the most active time of the year. You know, the climatology of Antarctica in October and November has always been vicious because you don't know when the timing of the thing comes along. So we did look in January, February, and March. That was in that '77 paper. So we did look at that. And I guess, again, we got lazy. We would have eventually gotten around to doing it in the next 10 years, but we hadn't reworked them. And this was half the problem. Jonathan came over and said, "The whole thing is looking low." And the answer was, "Yes, that's fine." "But please convince me that the instruments aren't drifting." And I sort of got angry. And I called Brian one weekend and said, "This is not good enough. We've got to sit down and convince ourselves that it's not the instruments, that there's actually a change."
When did Jonathan first come to you and say that there's something?
'74 I suppose. '73, '74.
When did he actually first come to you with the data and say, "There's something wrong here. It looks like there's a decrease."
I guess that was early '74, if my memory is correct. Or some such time. And as I said, what we did then —
Early '84, you mean?
I'm sorry. '84. Yes. Sure. No, it must have been '83.
Okay. So it was '83.
Yes. We decided we'd better send a new instrument, because that one had been there 15 years or so.
But sending in 123 had nothing to do with you thinking there was a downward trend or anything?
Well, that's right. Indeed. We just though that — as I said, what you've got to be really sure of is that it shows up in all the different measurements in the right sort of way. And the extraterrestrial constant hasn't changed. And because, as I said, we were waiting another five or six years before we actually did the whole homogenous game again and we sort of were caught in between. So I phoned Brian in one weekend and we actually sat down and we worked enough of the things to convince us that this was it. But I think, yeah, it was all very well plotting October. I mean, once it got to 30%, then it was obvious it was dithering. What really convinced me in the end was effectively — you have to realize that the ozone hole was awkward then because it appears at a time when you can't really measure. I mean, you're relying on moon observations to tell you what's happening in August, and you've got to wait for the sun to come up high enough. So lots of measurements are really —
Iffy, as it were. But what really convinced you are the values you find at the end of the polar night are different than the ones you entered in the first however many years of the record-20 years of the record, 18 years of the record. You came out of the polar night as badly as you went into it with. When you started plotting these in the right way, you discovered that, "Gosh, you're actually coming out of the polar night with values which are much lower than you went in with." And, of course, this is what you'd get, but you can't really measure.
So over the entire record of the — say Halley Bay data, when you went into the polar night and you came out, you had pretty much the same value?
Pretty much the same value. Yes.
And that's another thing that was different about the Ezob [?] solutions.
That's right. Indeed. That's why I'm looking at the October trend. Because the October trend is a mess. Because in the good old days you could actually lose the vortex in mid-October, so you'd have to look at each year. And if you really wanted to get a trend from it, you would have to do some sort of preliminary filtering and then just reject the years in which it was clear that — I mean, that sort of simplifies it to a state of absurdity. Basically you have a polar night in which you can't measure as [???] through ozone and moon observation. Not very good. In fact, if you come out of that polar night with a value at some indeterminate stage, somewhere between here and here, it will go up; and then it will fall down again and come back over to this side, and then you'd come back into the polar night like that. It's as simple as that. But this timing is horribly random. Now it goes into December. But in the early days you could have half of October, which is high or whatever; simply plotting the October main values is not really a very good way of looking at it.
Because the vortex breaks down at different points?
Sure. That's right. And from year to year. But what was pretty firm was that usually you found a little bit of trace at this level here, and then suddenly someone plotted these ones, and they end up falling off... So that's what really decided me that something had to be done.
So the values of the ozone, as you came out of the polar night —
Just kept getting lower and lower.
Lower and lower. From winter.
It would be nice to have the measurements that you could sort of see the thing doing that, at that stage. But it took a little while to sort of adjust to what was going on and to see how the thing goes.
You say in '83 John Shanklin first mentioned this to you. So what was your first response when he said that — I mean, did he tell you that the observers down there were mentioning that the values were dropping?
No, not particularly. Because I don't think they were mentioning that. I mean, it was just simply that he — plus the October monthly means [???], to which the first answer was, "Sure, but have we re-worked the extraterrestrial constant? I mean, are you sure there's something not happening instrumental?" And, at the same time, deep in the mythology of all this, he had actually encouraged the people down there to do something, which simply showed that he had forgotten how to tell them how to run an instrument, unfortunately. The Dobson spectrophotometer, the results depend upon which direction you point the instrument. You must point the instrument at the sun, or where the sun should be. If you then rotate the instrument, you'll get a nasty effect like that. And then that's something he wanted to publish, unfortunately. So it showed that he discovered this effect for himself. But it was very deep in Dobson literature. And the instructions are very clear that the instrument has to be pointed. So I was a little bit worried about this as well.
This was before he —
This was much the same time as we were pondering over this one.
So he brought this up about the fact of changing the orientation of the instrument and how that effects —
Well, the orientation of the instrument is already in the original Dobson thing. It was very carefully stated — and this was the sort of thing which can happen when you're a long way away —
But he —
So what does he say about all this?
Well, he mentions this and you telling him you need to check all these — and what specifically did you ask him to go back and check now, the first time he mentioned this? "Did you check the extraterrestrial constant? That needs to be re-checked?"
And what else did you suggest that needed to —
That's the first thing, I wanted to see whether the graphs, the standard tests, the waves tests and everything else, to make quite sure of the things that were going on.
And so then he went back and checked those things?
Well, no. In the usual way, we were busy, so that sort of delved on. And what really brought it to a head was that we had an advisory committee come in to see what we had been doing. And as I said, the only way I could say something sensible to the advisory committee was to actually re-work some of these things myself and convince myself.
When was this advisory committee?
Oh, gosh. That must have been October' 84.
October '84. John Shanklin mentioned the fact that there was an open house and —
Yes. There was an open house either in '83 or '84, and he wanted to put up the fact that there was this downward trend. And he said that he was told to put up a sign that there was no change in the ozone from — he showed me the placard that was put up. Does that ring a bell?
No. Ask Brian Gardner.
Okay. I asked him. He said he remembered it too. But I thought I'd just check with you.
They didn't ask me.
Okay. This was something that I guess was an open house.
Kind of an open house or something?
Yes. And I guess Jonathan wanted to actually put up the results showing there was a downward trend. And I guess the mix was put up there —
I honestly don't remember being asked about that. I think Brian must have sat on that.
So, yeah, a placard was put up saying that — So then there was this advisory committee meeting in October of '84. And what was going on with that? You said that if you were going to say something about this that you needed to know what additionally?
Well, I needed to be convinced that this was sort of worth drawing their attention to. And I guess I thought it might be interesting. But it was the dullest meeting I ever attended.
So what exactly did you say about this in the meeting?
I said, "There are unequivocal changes taking place in the Antarctic ozone and my guess is it's the CFCs." It just fell flat.
Like a lead balloon?
Well, you know, "You'd better publish it as quickly as you can," sort of thing.
Okay. So this was in October?
Yes, it was October.
So by that time you had been pretty convinced that there was something going on?
Okay. Jonathan gave me a copy of a paper that he wrote up November 2, 1984. Because he said, as he describes it, he still hadn't convinced you or Brian. So he wrote up this like three-page paper and he gave a copy to you, Brian, and there was another person, Reikroft. He gave copies to. Do you remember this paper?
Okay. What was your reaction when you received this paper? Were you already convinced that something was going on before you received that paper?
Well, if the timing is as you say, I honestly don't remember. Well, it depends. My memory would be that Brian and I were already busy writing what eventually appeared. If the timing —
I've got the thing Jonathan Shanklin wrote up. It's November 2nd or something. And he's actually got comments back from Reikroft on November 6th. And from talking to Brian, he said that you didn't start writing the Nature paper until after November 11 although November 11th was a Sunday that you had called him into the office?
I see, okay. Well, that seems reasonable yes. I believe that.
Okay. So when you got Shanklin's little three-page paper he wrote up, I mean, what was your reaction to that?
It wasn't particularly convincing. I mean, he had just plotted the monthly means. I felt one has to do better than that, I think, is the answer.
Okay. He plotted out the monthly means. And you thought because of the thing with the breakdown of the polar vortex and all of that, that it just wasn't good enough?
Yes, sure. I mean — well, you had to put it in context. Well, one of the sorts of convincing things about that we eventually wrote it up was that you couldn't find these silly things happening at other times of the year. You know, the point was to show in what sense the annual pattern had changed and that the instruments should be reliable because they're still giving moderately reasonable values after the vortex has gone away. Who knows?
So sometime in October or so you had been pretty convinced something was going on. You didn't know quite what. And then, November 11th, this Sunday, you had gone into the office and later in the day brought Brian Gardner in. Why did you go into the office this day? Did this just get under your skin and you needed to look at the data a little more?
No, I guess the advisory committee must have been the next week. I obviously got the date of the advisory committee wrong. I mean, it was preliminary to that.
It was just I had done it before that. I mean, looking back it wasn't an awfully happy time, I must admit.
Why is that?
I guess I was doing too many other things. I lost day-to-day touch with Brian and Jonathan I suppose. And I was hoping that they were getting on well together and still doing a sensible job on it. What was happening? Oh, I had this dreadful — this man — Well, it was when I acquired my interest in ozone chemistry, because I had this Ph.D., the modeling and so on and so forth. So I guess it was —
And you had been doing that like —
No. That was the first year.
'83 or '84?
Yes, it guess it started then. But the first year was more or less wasted I guess, in a sense, with me trying to teach him about the Antarctic and he tried to teach me a bit of ozone chemistry. And like all these things, it takes time to get the model up and running. And then you have to pull the model to bits to see what the hell its saying and whether what its saying is sensible at all.
Was he a student of Thrush?
I guess I'll ask you more about the Sunday night that you called Brian in. And when you called him in, what was the point of calling him in?
Well, because he had all the data sheets in his office. I mean, I didn't accumulate all these things in my office. I had to actually be able to put my hands on them and sit and look at them and sort of —
So what sort of exchange went on between you and Brian regarding the data?
Well, basically the first thing we did was quite simply look for all the direct sun observations, which are the most reliable. Work through these for the relevant month of October effectively and just convince ourselves that they made sense compared to what Jonathan had said. And do a quick test on what the possible instrumental change could be. I mean, this was in a sense almost a bonus I suppose. Because had you started these arguments a year or two earlier you would have been terribly confused because the effect was full. I mean, by being lazy and waiting for the thing to fall out at us, you didn't need any statistical tests. Your grandmother would have seen it. Provided you convinced yourself you were still running the instrument properly, something rather dramatic was happening.
So when you were looking at these direct sun observations, were any of those from any particular years that you eliminated for any reasons whatsoever?
No. Not really. They were all — I don't remember the exact dates, but we ended up by being convinced that, in fact, the —
So it was the direct sun observations that really convinced you?
Oh no. Clouds and things like that are always nasty [?] when they get much bigger.
So when you were looking at the direct sun observations you were looking at them from going into the polar night and coming out?
Okay. And you were seeing this effect where there was a big drop between entering and exiting?
Yes. That's right. It just sort of went down one heck of a rate when it came to —
Okay. And then in the Nature paper you said to interpret the difference, you're talking about this drop here, I guess, in exiting the polar night as opposed to entering. You said to interpret the difference as a seasonal instrumental effect would be inconsistent with the results — well, actually, no. I guess this is for the difference of the years. To interpret the difference as a seasonal instrumental effect would be inconsistent with the results of routine checks using standard lamps. Now, how would the standard lamps reveal that this was an inconsistency, the difference?
Well, you should see a seasonal variation in the standard lamps if it's instrumental.
Okay. So if this drop is due to a calibration problem, it's going to reveal it?
It should reveal itself in the standard —
In the standard lamps. And so you checked the standard lamp tests?
Yes. Standard lamps, the wedges, and the wave length settings effectively. I mean, idiots can — by re-setting wave lengths we managed to convince ourselves that all these things were effectively in order. Or at any rate, there was nothing in it as large as the difference we were seeing. Which wasn't surprising.
Now, this drop right as you're exiting the polar night, did you consider the possibility that that could be — I mean, it seems like the fact that you have it occurring in October, as you're exiting the polar night, and it's not —
That's the trouble. Let's draw this a bit better. The polar night is a bit in the middle. But there's a retched twilight period when you haven't got any measurements or you've got very little hope of making measurements. If you believe you need sunlight to destroy ozone, which at that time I did, then clearly if you could stop measuring there you ought to see the thing actually falling down. But we didn't have those observations where it showed anything of the sort. It just simply showed you picked up the sun on a bit of a curve, which was obviously falling away before your eyes.
Well, you've got this drop-off as you're coming out of the polar night, but it's not — this change in the record is not showing up any other time, other than October. Right?
Not actually. No.
Yes. And so, that in and of itself would seem to indicate that it's not an instrumental artifact. But again, I was talking with Brian, and he mentioned the possibility that the instruments were cleaned during the polar night, and all that. And so was it seriously looked at, the possibility that this drop right after the polar night might have been due to something that was —
That should show up in the test. I mean, when you clean it — you do a standard test before you clean it and a standard test after you clean it.
Okay. So sometime in October this would have shown up in the standard lamps then?
Well, I mean, we looked at the standard lamps before and after cleaning on all the other tests as well. I'm sure there are discontinuities that have to be added to the record. But they didn't seem to us big enough to need to quantify in face of what we were looking at effectively.
Okay. Now, you and Brian, this November 11th, that Sunday, pretty much convinced yourselves that there was something going on. Is that the night that you decided you needed to write this up for Nature? Or when did you decide that?
At the advisory meeting, more or less, I said that. I brought it up at the advisory meeting and laid down the law that something dramatic was happening and there was almost certainly CFCs. And their only response was, "You better write it."
And so what was the point —
We did, in fact, submit it on Christmas Eve. It took time to sort of think of the best way of doing it.
So it was about a month and a half from the advisory —
It must have been.
Because the advisory meeting was like what? A few days after that or something?
It must have been that week; otherwise I wouldn't have been getting so agitated about it.
So the plan for writing this up was not just to present the results? That, "Hey, we're showing a decrease." But also to try and explain it.
Well, it was my view. Yes. I mean, we could have just written it out. But I don't know that it would have made —
Quite the impact. And we're entitled to say that the dynamics, on the face of it, hadn't changed. I mean, you had to say something of the sort. Otherwise any sort of explanation would have been possible. It seemed to me we could really say that there didn't seem to be anything left but chemistry.
I'm kind of interested though, during this whole period, and in the Fall of '84, as to whether or not you had talked to anybody else about this? Because two of the people you thank at the end of the paper, Thrush and Mergatroy —
Were you just thanking them for discussions you had previously had about this sort of chemistry, or... ?
No, no. Bob had retired from the MET office. Bob was one of these lovely solid people whom you could bounce ideas off. Bob deserves his tribute, because when I convinced him I felt like I convinced everyone.
So you had talked to him about —
Oh yes. He was retired. But he was coming in as a consultant. So I saw him about once a month I suppose at that stage. He was at the meeting. He was actually sitting in on the meeting.
The advisory committee meeting?
Okay. Was that the first that he had heard about this? Or had you talked —
You had talked to him about this before?
Before, yes. As a matter of fact, I sort of cleared it with him before I went to the meeting. I said, "I'm going to say this."
When did you first talk to him about this?
Oh, goodness knows. I mean a month or two probably, I suspect.
So sometime in like maybe October or something?
Or in September even. Because he was sort of helping me. He was in on my discussions when we were deciding what we ought to do in the models. I mean, I don't know if you ever read that — Effectively we founded one of the models, that wouldn't represent an Antarctic summer unless you put quite a large vertical velocity to keep ozone there for you. And Bob had been concerned just before he retired with developing the method of his models for ozone. So we used it. So we talked about that mainly.
Was he doing 2-D modeling?
I guess they had everything from 1- to 3-D. I don't know if it was that stage, or 3-D models.
He was retired from the MET office, and he had been coming here like once a month or so to consult on various things?
So sometime in September or October you had talked to him about the —
Yes, I did. I said, "Jonathan is making noises, and we're going to need to look at it." Like all these things, you put it off and —
What did he say when you brought this up?
I guess the reason I presented it in the way I did was — the thing that sort of convinced him, rather than just saying a monthly value for October, was the fact that you could see this sudden change from, what should we say — goodness knows, everything up to '78, the values really just stuck within this coming into the thing here.
So you presented to him, as you presented it to me here, that's how you presented it to Mergatroy.
And that really, yes, I mean, I think convinced him more than just looking at a graph with a trend fitted to it.
A downward trend. Yes. Yes.
So we decided that when we published it we wouldn't just talk about that time of the year, but we'd show what was happening in February, at the same time, just as a contrast to show that things were going on.
Okay. What other discussions did you have with him regarding this? Were there further discussions?
Well, he wasn't convinced it was CFCs.
Okay. So when you talked to him in September/October, you thought it was the CFCs back then?
When we got to this stage, yes. When the figures got so huge. I mean, there didn't seem to be anything else that it could be. He concurred with this, with the statement we made. We didn't have enough space in the letter to spell out that the operating climatology was really so robust at this time of the year that there was no way dynamics could really be effecting anything. And he concurred with that. But that still didn't —
What did he think was going on?
I guess he agreed with us in his heart of hearts. But he felt he needed a little more evidence. He needed some chemical measurements.
When you presented this to Mergatroy, what evidence did you present that it wasn't the dynamics? What specifically did you say to him to convince him that it wasn't a change in the dynamics?
There was no sign of trends and stratospheric temperatures, winds, give or take the fact that there was some uncertainty in the timing, the vortex was there. The climatology is surprisingly robust. I mean, I don't know if you've seen some of the plots. Given the robustness of that through the winter, you would expect it.
I'm kind of surprised about this. The fact that you had all this data indicating that the dynamics hadn't changed, and once your Nature paper came out there were all these people putting forward these dynamical theories. Now what was your reaction to these dynamical theories?
Basically to laugh I'm afraid. To simply say, '''Wait until some chemistries have been measured and I think you'll find you're wrong. You're a better dynamacist than I am, but basically, I'm sorry, my money is on the chemistry."
So basically you were convinced that your data had pretty much constrained the problem to a chemistry problem?
Yes, yes. I mean most of the people who reviewed the paper — no names, no factual.
Well, I know the people, I think, who reviewed the paper.
But even Susan convinced me that she knew nothing about Antarctic meteorology at the time. And it was much the same with these other people. They're good dynamacists, but they really had no feel for what went on in Antarctica. So I guess we were arrogant to say, "Wait and see." And then other people came along and there were these knocks and all these other funny things going on.
They saw the proton events and —
You know, you can't — I mean, the one thing I regret about the paper is we didn't spell out in enough sorted detail for people what should have been obvious, that you can't change the ozone by 30%, and the only place you can do this is 70 millibars-50 to 70 millibars. You can't do it above and you can't do it below either. And it's just that bit of the atmosphere the balloons luckily go up to. So you can really be quite sure that really not very much has changed.
So you had these talks with Mergatroy. What about Thrush? I mean, he's thanked [?] too. Now Solniscus [?] was a student of Thrush?
That's right. Yes. He was the chemical advisor as it were. And he was somewhat equivocal about it, I must admit.
You had talked to him about this around the same time? Early Fall?
Yes. And, he was quite right. There was an awful lot more work to be done before anyone could really be pretty plain about it.
He was equivocal about the chemical explanation you presented?
Well, we didn't present the chemical explanation. We simply said that, "It must be the chlorine." And then the answer was, "Well, you can see a little bit how the chlorine interacts with lower temperatures and so on. And you obviously must reduce the nitrous oxide. "Well, that would have been all right if the thing had been 10% or 15% depletion. But it certainly wasn't good enough to explain the 30%. And so it probably got to 100% — I mean, it was perfectly plain that [inaudible]. I mean, I guess to some extent he had an interest in OCI and such other things, and was concerned that we shouldn't sort of say too many things without having good grounds for it. Don't— I mean, that's not for the record. I mean, he was the one who insisted we gave a copy to OCI so they could come in before we published it. Yes, I was a bit surprised. But he was telling us about relevance of [???] with chlorine nitrate and those types things.
So he had mentioned the fact that—
We were in the reference stage to Roland and Melina and chlorine nitrate and HTO reaction and so on and so forth. Which didn't help at all. And there was quite a bit of hand waving.
But he was convinced that — I mean, you had this downward trend?
Oh, yes. Yes.
But he was equivocal about whether or not the chemistry could explain it all.
Everyone was equivocal about it. I mean, you had to decide at that stage whether you were going to stand and make a song and dance about it, or whether you were just going to sit back and wait to see what happened.
I'll ask you this question; I've asked other people. Where did the name "ozone hole" come from? The "hole." Do you recall that coming up anywhere?
It first appeared in the Washington Times in '86. There is a hint in a paper written by someone who interviewed it all — in which I think he more or less says that Sherry Roland invented it and passed it on to Walter, or whatever his name was, the New York Times man. Whether that's true, I have no idea. But that's when it first appeared. Other people have blamed Susan for it, but I don't think she's the sort of person who would — I certainly didn't invent it.
I read something where you and two other people were sitting around and someone said —
No. Not that I'm aware.
Okay. Another pothole story.
The first time it appears in newsprint was... I guess it must have been October of '86 sometime.
Okay. So you submit the paper to Nature close to Christmas Eve or something. And it starts going through the review process. Well, let me ask you this before we go into that. Were there any other people that you talked to about this in the meantime? So you had talked to Thrush, Mergatroy, obviously this guy [inaudible]?
I mean, a copy of the paper also went to Harold Johnston. Because if you remember what was just happening was that WMO was producing their first stratospheric ozone report, and so we felt we had to sort of say something there. So it gets a brief mention. And I guess we thought that was enough, so we sit back and wait to see what TOMs people say. I mean, you know, we had thought that TOMs ought to be saying something, and it was a bit strange why they hadn't sort of said anything. I guess we didn't know the people well enough at that stage to sort of bang on the door and say, "Hey, what the hell's going on?"
Yes. Because I know John Shanklin has actually sent letters to people at TOMs, but it was the wrong people.
Yes. Everyone's busy and all the rest of it. I mean, I guess Johnston must have had the letter in January I suppose. ICO certainly saw it in January. I don't know about Mergatroy. I spoke to him, and he would have certainly said something to' anyone he met about these things. So the next sort of stage was Bob Watson. I could tell another little story. Bob sort of insisted I went along to meet...
This was Mergatroy.
Yes... To meet the MET Office Group, Adrian Tuck and his people, which Bob had been in charge of.
This was when? The late Fall?
I guess this was February or March, just before the paper came out, to be received with skepticism, I have to say. Two young men who now, I think, regret having sort of looked at me and said they didn't believe it. However, that was fine. And then I guess we just sat back and then October of the next year the American Embassy rang up and said, "Bob Watson's coming across with some satellite pictures which he'd like you to see." And then it just mushroomed.
Yes. Yes. Okay, so the paper gets submitted in the end of '84, and when did you get the reviewers comments?
Susan was fairly prompt. Ravi was rather dilatory.
So those were the only two reviewers? Right?
They were the only two I know of. Though it's not good to actually name them, I think, as you write this up.
Well, I mean, Susan Solomon's already been named in print as one of the reviewers. I mean, she doesn't seem to have a problem with it.
Oh, no. No.
I'm actually going to be interviewing Chandra anyway, so I'll ask him. But one of the things I actually wanted to ask you (and Susan said it was okay if you wanted to give it to me), is whether I could get a copy of her comments and the original paper.
To be quite honest, I'm not even sure I kept it. I've thrown a lot of things away.
Because I don't think Brian had it.
I know he saw it. Whether he took a copy, I don't know. You'll have to let me search. I have a horrible feeling — yeah, I guess I didn't really think in terms of history at that stage and it just sort of disappeared.
Yes. I mean, it would be interesting to have the comments.
Susan sort of drew attention to the Japanese and so on and so forth. Which Brian should have known about anyhow, but didn't actually tell me about when he went to Hawaii.
Oh, that he had met Chubachi. Yes. Actually, he met Chubachi in '85, the summer of '85.
Yes, in Hawaii. But he didn't actually tell me about that.
Oh, he didn't tell you about that. Okay.
Which shows how little we were talking to each other.
But he told you he met Roland there though, didn't he?
Well, that appeared later actually. That's another one of these improbable things.
Okay. So you got the comments from the reviewers at some point. What changes do you make in the paper in response to their comments? No substantive changes?
Just adding a few references or...?
No. We wrote back to Susan and pointed out that the Antarctic stratosphere didn't behave in quite the way she thought. She had this idea of sort of a steady streaming downward motion, etc., etc. And there was nothing which — we didn't feel like adding to the paper in light of the actual things she told us. And that was that. I mean, nothing. And Robby was just ridiculous. Robby said — well, I'll quote it. It was about as short as that. It started off by saying, "This is impossible." Then it said, "Well, if it's true, it's interesting, so it had better be published." And that was it for all practical purposes.
I guess I've read that quote somewhere.
It may have come out somewhere. I don't know.
Yes. Actually Arlin Krueger said that he — Chandra had mentioned something that he was reviewing a paper that was giving impossible results or something. He didn't say anything about who it was or what it was. But the amazing thing is that Arlin Krueger was at the same — well, actually before that had been going to the TOMs meanings, the ozone processing meetings, which actually showed that they were getting low values. And he never connected the two.
Oh. I didn't know that.
Yes. I can tell you more about that off the tape.
Actually, I first met Arlin in Tokyo in, it must have been '86. He put his arm around me and he said, "Joe, do you know what you've done?" And I said, "No." And he said, "You have enabled me to look at my data for the first time since then." I mean, he told me that he had been taken off the TOMs thing, and it was only when we published the paper that he was allowed to go back and start looking at it again. So, I don't know.
He actually wasn't part of the ozone processing team?
No. I try and keep out of the [???] argument.
Well, this is one of those things where the left hand doesn't know what the right had is doing.
Yes. It was the same with us in those days, I'm sure you've discovered to some extent.
Discovering it to a great extent that it was going on everywhere.
Sure. There's a question of sort of sorting it out in your mind what you think is significant and what you think is not. And what triggers you off, as you've discovered here, is something to which to most people probably wouldn't make very much sense. But it convinced me a lot more than just plotting a monthly meeting and drawing a line through it.
Yes. Okay, so the paper comes out in May 16, 1985. What sort of responses were you getting from anybody right after the paper came out? Or was it pretty much dead for a while?
It was dead until — like I said, the embassy called up and Bob came over with the reprocessed data effectively. And I guess we must have had a few people who rang up and told us it was impossible, etc., etc. Having put your money where your mouth is you just have to wear it and sit back. I mean, psychologically it was a rather queer time. You weren't quite sure what you should be doing-banging on the Prime Minister's door or — It was perfectly obvious no one was going to move very much until some confirming evidence came along.
But this was about five or six months before anything happened?
It was I-D modeling of the Antarctic summer. I mean, the irony of the situation is that — I sort of sat and thought about this, and basically I thought I-D modeling in the Antarctic Spring was a pointless exercise because you couldn't really get anywhere. The climatology wasn't reversed enough. But the Antarctic Summer was much more fun. You had this very rapid fall-off of ozone. For what seemed to me fairly obvious reasons, and still do, you get 24- hour photolysis and over the [???] and so on and so forth. The first runs of the model were quite categorical that you can't keep ozone layer in a static atmosphere. The only conceivable way the thing doesn't even run away even faster from you is the fact that you've got some downward motion bringing ozone down to compensate for things. So I don't know what Mike — I mean, there was really no relevance between what Mike was doing and what we were doing here. And it never even crossed my mind — I mean, I've even forgotten when he published? When was hit thesis finished? '85?
Yes. I think it was. Yes.
I mean, because I wrote — well, the irony is, I wrote that other silly paper with him and Brian and Bob Mergatroy. Which also follows the referees.
That paper you wrote?
Yes. That should have been published before the ozone hole paper. But because of the — well, the referees are not too far away. It got held up and held up. So it actually appeared afterwards. I think we wrote a lot of good sense in both papers actually.
I was going to ask you this too. When did you first hear about Chubachi? Is that like you said in Susan's Solomon's review where she had mentioned him?
Yes. Susan sort of drew my attention to it there. And then I met Chubachi in '86 when I went to Tokyo. But I forget where the meeting was now. Alan was there and Krueger was there, and Chubachi was there. I mean, the trouble with Chubachi’s results was that they were just isolated incidents. The COA didn't have a long enough ground record to convince people that something exceptional was going on. And Chubachi, of course, was completely-and correct me I suppose, his infatuation was the dynamical explanation for it. Which from the point of view of Showa Station, is in fact right. It's just that Showa Station in some years is in the vortex and sometimes it's out. So, yes, I mean, there is a lot of dynamics in one sense there. But it's not systematic dynamics. And I don't think he really had any idea of the sort of systematic change. I mean, he was just drawing attention to what he thought was enormous results. I think we can claim to be the first people to put it in its context, which something systematic was happening. But that was the trouble. As I say, I don't know, I suppose I got busy in the modeling and various other things. And the usual run of the bloody administration. I wasn't awfully in touch with other people before the paper was written. And then, of course, it all changed. Everyone sort of hailed around the world everywhere, "Do this, do that." And then you start to learn a lot and see what other people are talking about. I mean, given the number of accomplishments every year now, it's amazing to look back at how isolated everyone was in those days. Just totally immersed in their own little games.
Yes. But Chubachi's stuff — I assume you've seen the assortment of papers he's published. I mean, he seems to be focused more on the fact that the ozone is actually showing a drastic increase at a certain — like earlier than it had shown before, rather than focusing on the low value.
Yes. Sure. Yes.
I guess I'd like you to expand a little bit more on the idea that the Showa Station didn't have a really continuous record, even though it went back— I mean, the Station was operating since '66, but I guess I'd like you to speak to how you could actually interpret that there was like a long-term decrease in ozone at Showa, or how difficult it would have been to have done that, given the nature of their record.
Hmm. Not very easy. They were lucky they got money for the ozone suns. The ozone suns are actually nasty things. And if you wanted to show trends in suns, if you're not doing them every day you've got a huge problem on your hands. First of all, sorting out what the sampling thing is. I suppose the other thing one has to... I'm trying to remember now. I mean, life is so easy these days you just ring up your friends and you'll have charts pulled out on your desk like confetti. You know? I mean, one of the huge problems through the '60s and '70s was getting some idea of what the upper air conditions were. I mean, we had our own two stations, which were fine. But there were only about three other upper air stations in Antarctica. And drawing maps and understanding how the thing was behaving dynamically was really rather difficult. I guess given what was available; Chubachi really didn't have much hope of really looking at his data in a very sensible way. And then I can't remember what the Japanese were doing, if it was actually Dobson. I mean, he never referred to the Dobson.
They had a Dobson?
They had a Dobson there. But, I mean, they never seemed to refer to these things. And you didn't get the feeling that anyone thought that a measurement made in '66 would be comparable to one made in '83. I mean, this was the whole— I mean, that was one of the things which made me sort of determined about it, was that I knew that for 18 years this hadn't happened. I wasn't prepared to be sort of automatic about what had happened from '75 to '85, because Brian and Jonathan, effectively between them, were sort of sitting on that. And I was waiting to sort of see the worked-up fashion of it. But it was, you know, the very fact we could sort of put our hands on our hearts and say, "Here's 18 years of data." And this sort of thing didn't happen before. Which really sort of made the difference. And I think that's what sort of convinced Bob in a sense, given that we were prepared to put 2% or 3% on a measurement made in '57, and one made in '76, or whatever it was. It convinced him that we knew how to run a Dobson. As I say, other people have published also some spherous trends in that sort of time.
The Showa Station too was, as you said, sometimes in the vortex and sometimes out? Does that cause problems with the interpretation too?
Oh yeah, sure. I mean, I can't draw, but you'll recognize it no doubt. I mean, Halley is tucked away here, and Faraday is tucked away here, and Showa is somewhere around here. The typical decay of the vortex is the vortex sort of shifts off in this direction, ending up in the north, in the Atlantic or whatever. It's just occasionally it can't wander out over shore. So given the lack of upper air observations, it's not very easy to see. So you really have to almost go by the ozone measurements at the other stations and guess where the vortex has got to. So really you sort of — sometimes you may not see an ozone hole. The most recent 10 years it would have done — but not for very long. I mean, particularly in the end of October, November, the lower ozone can be, well, stuck over here, and the high ozone stuck over here. And I don't know. I mean, I never really sat down to try and interpret his measurements, I must admit. But I would have thought it would make it much more difficult.
Yes. But Halley Bay is pretty much sitting right in the center of the vortex. Right?
No. I mean, we can go and look at the TOMs data of the moon if you like. I mean, there's actually a long narrow mess sticking up towards South America. I'm drawing it very badly. With a very low ozone in here, and Halley would be sitting here. I mean, occasionally it sees the full depth of it, but not that often, in fact.
Okay. But it doesn't get really close to the edge of the vortex though?
It can. I mean, 1986 was a very nice case in point. This was one of the nice things that convinced us about the TOMs data. I've forgotten what October looked like now. October must have looked like this. Nice and flat out to the ozone hole, and then suddenly it sort of went whoosh, and then it slowly did that. And this was just the vortex edge coming over Halley. And then TOMs re-produced it with the Dobson beautifully. So we actually had a range of, I can't remember, 340 to 150 Dobson's, I think. Which tracked itself perfectly.
Right. And this was in '86?
That was '86, I think, if my memory's right. In October the thing is wobbling about like a rather soft balloon and it can be anywhere effectively. I mean, it very rarely gets back to Eastern Antarctica. But Halley — The first year we ever made measurements, you see, Halley showed the typical behavior in those days of having this huge sudden increase. And Faraday just sort of had a little minor bump. And what had happened was the collar of the vortex had actually come up and sat over Argentine Islands. And then all the sort of new ozone ratio was here. But Faraday never actually saw an increase that year. And that's unique in sort·of the whole 20-odd years records. We dizzied over this for a long while and we were sort of, "God, we must have been doing something wrong."
So what made you realize that you weren't doing something wrong?
Well, we had the upper air maps. And, as I say, the first sort of thought was this increase was so small compared to what was going on at Halley we were like, "Oh, what have we done?" And then actually we got the upper air maps, which for the IGY were quite good ones, and then we sort of suddenly realized, "Gosh, you know, the vortex came and sat over here. As everyone else was getting in the ozone, you were sitting in the middle of this vortex and seeing nothing."
How long did that vortex sit there?
About a month. I mean, from mid-October to mid-November. So we had these sorts of ideas in our mind. The other thing which was clearly in our mind which doesn't get very much mention was that when the Argon eruption came along, we show it very clearly that the volcanic dust didn't reach our stations until the vortex had been blown away. Dobson had already said the thing was isolated. We could put aside that. But that always seemed to me a wonderful tracer experiment, which clenched the matter almost beyond dialect. You don't move stuff about in the lower stratosphere until the vortex itself has been pushed off.
Okay. The positioning of the stations — actually going back now, back to the '50s for IGY, what determined the positioning of some of these stations? I mean, I'm sure some of it had to do with practical matters for transporting materials and things like that.
No, I don't think so particularly. I may be wrong. I mean, I wasn't senior enough at the time to get involved in these arguments. The Argentine Islands was quite straight forward. It was just a base which had for the last years been used for other purposes. We went through the exploring stage. And it was a base where you sort of sat there with not much to do then. And so there was room and we scoped it for putting in a station there. So that's why FIDS effectively put it in there. The Royal Society simply wanted to get as close to the pole as it could in British Antarctic Territory effectively. So it went to the head of the — as far as it could get to the head of the YDC, decided it couldn't quite get to the head of it, because the relief problems for admittance. You know, you might get in one year, but sometimes... I don't. Yes, sure, I mean, they wanted to get as close to the Royal zone as they could. Was it before the treaty? I can't remember very effectively on what was once British…
Yes. So what about South Georgia?
South Georgia was always awkward because South Georgia was technically a little bit of a dependency on its own. It was a separate dependency until the thing was reorganized in '63. So we weren't actually allowed to work at South Georgia in those days because it wasn't part of the — far enough as remitted [?]. And in any case, there wasn't enough money to run that number of stations for IGY. It was touch and go anyhow actually.
So after the IGY, was Halley Bay going to be closed down?
Yes. The Royal Society was effectively going to abandon it when the offer was made if they would transfer the equipment we would provide staff for it and re-fill it. And that was accepted.
So that was the primary reason for keeping Halley Bay there?
Well, no. I mean, it had two and a half years of good data.
You might as well extend the record and they're going to give us all the equipment.
It was just within the logistical capabilities of the survey as it were. I mean, technically, it was a long way...
Any special reasons why Halley Bay was maintained there?
No. I mean, I guess the Royal Society put pressure on them.
So again, going back to the question of the Dobson 123, when that was ordered. Again, just to be clear, that was ordered because you had been observing low ozone values at Halley and you wanted to check those. This was more for practical reasons, for pulling the instrument out and getting it overhauled.
Originally, yes. It arrived in time. And then it just seemed sensible that the two should sit there. I mean, normally they wouldn't have sat there side by side. One would have gone in and the other one would have come out.
So what was the reason for leaving them sitting there side by side?
Because I was worried about this 20% or 30% which was being talked about.
Because, let me see. The instrument, you received it in June, 1981. And it was actually installed in January of '82?
Oh, I have no idea.
Actually, I got that from John Shanklin. So I think that actually shows up in your Nature paper too. But it was installed in January of '82.
January of '82?
Yes. So it would have been about three years before the Nature paper came out.
It doesn't ring a bell at all. I'm thinking it was in '84, surely. Wasn't it? I mean, there must have been two instruments still there in ' 84.
I mean, from what Brian and Jonathan said, it went in in '82. And they don't recall there being an instrument in '84.
No, I would have had to have gone in in December, '83. I don't know. What does it say in the Nature paper. They passed what I said in the Nature paper.
It said that 123 had been in there since '82 or something like that. Since early '82. I mean, the reason that this came up is because there's a number of accounts out there that say that the reason 123 was sent down there was because you were getting the low measurements. And you wanted to have the instruments sit next to it for a while to make sure it was okay and all of this. But Jonathan and Brian don't agree with that.
I see. All right. Well, they're right. Um... yeah. Sure. They were in charge of all these day-to-day affairs.
Oh, okay. So there wasn't any instrument sent down there to check for —
Okay. And no instruments were ever sent?
I still hop back to the fact they kept showing me two instrument's records for one year to convince me.
Yes. Jonathan said — actually, I guess, Jonathan went down with the instrument then and the 123 and 31 actually sat next to each other taking measurements for like two weeks together. But you seem to recall a whole full year of measurements together?
Well, like you, I've probably read it somewhere else as an apocalyptical story and it's never crossed my mind to doubt it. But you can take Jonathan's and Brian's word for this. Because I don't have all the records.
Well, I got the shipping manifest and all of that. The one that was actually sent down.
Okay. I'd better look back at the Nature paper and see what we actually said at the time.
None of your instruments were ever sent to Boulder to be checked against the world standard?
No. They came back and went to the maker's specs and then to the UK MET office.
For check against their standard?
Okay. I mean, again, when they went to the MET office to have their calibration checked, that was just a rough check to see if —
Well, no. It was the best people could do in those days effectively. But the huge worry with these things is that you have to do a certain bit of dismantling to be able to pack the damn thing up so it could travel in a small ship.
Okay. So then you sent them out to the field and then the Langley plot method was used for long-term calibration and all that. Okay. And the Faraday Station data, when you were going over this, did that help you make up your mind at all about whether or not this was a real long-term —
And how? Because it seems like the Faraday data, that the —
Well, it's very equivocal. But it's what you would expect from something which is much near the edge of the vortex as it seemed to be. But yes, had it showed nothing I would have scratched my head. It doesn't show up quite so plainly as the Halley one does. But it's [inaudible].
Okay. But you would expect it to be that way though?
When Brian came in on November 11th, the Sunday. When you were looking at the direct sun observations, did you look at those for Halley and Faraday?
I honestly can't remember. Instinct tells me I would have looked at both. But I really can't remember. Does Brian? No. I think it was Halley which occupied most of our attention.
It's actually a question that only occurred to me after I had talked to Brian.
I've forgotten. I guess we must have done — probably not as systematically as the other ones.
I guess, again, I'd like to go back to the issue of the data reduction and how careful you have to be in reducing your data and knowing what's actually going on before you can actually give out reliable information. Because like I said, I've been out to Boulder and I've talked to a number of people now about the Dobson's and all that. And it seems, in a lot of ways, to be not a very easy instrument to get reliable information out of. And I've talked to people who have worked with it, and they seem —
I've forgotten. Haven't they automated the Boulder one?
Yes, they've automated it there.
No. It's an instrument which works whether you've got someone who perhaps has had trouble with it. I suppose one of the reasons why I wasn't immediately receptive of Jonathan's things. Because in the good old days— you've seen the thing and you probably realize that on top you have a smoked plate and this wonderful little clock. And you don't actually look at the dial readings at all when you're making the measurements. You just scribe on your wonderful smoked plate and then you sort of mark the things you want on it and you read them off. Jonathan got to the stage where he had a lot of young men just to sort of twiddle the dial and then take what they thought was the meaningful reading off. Which actually I thought was sort of retrograde. I mean, if you're going to do this, at least the way you play the game, or the way I play the game, I would hope other people still do, is you do go through this retched routine of the smoked plate. And you then get someone else to come along and to read it off to you. And to check that you actually transcribed the right dial readings onto the plate. Because that's one easy place where you can make mistakes. What else can you do? There are all sorts of silly things you can do. That's when we discovered about pointing the thing in the right direction and so on and so forth. Yes, I mean, you've got to be awake when you're using it. And how you would do a moon observation without the smoked plate I have no idea. Because moon observations has a huge variability. So the moon observations didn't really sort of come into the ozone paper, because anyone can make a mess of a focused moon observation, I'm afraid. And they have a terrible nongoshen [?] in error distribution. I mean, they're either right or they're frankly wrong. I don't know. I'm old and cynical I suppose. The young don't like taking control anymore. And you're quite right. Now you can automate things and so on and so forth. And then, in a sense, I suppose, that's why. There really ought to have been more sort of focused on the automation's which took part in this. But they didn't really, because people don't like taking the trouble you need to do to focus on observation. It's a skill you have to learn. A sort of muscular skill, to keep the sun where you want it.
So the direct sun observations is something that —
Well, no. As long as they're high and you're using the prism and the diffusion screen, that's fine. Because that's the beauty of the thing. It's then pretty robust. But if you've got a very low sun and you want to focus the image and so forth on the slit and you've got to keep it in between the judicial [?] marks you have to go through this business of scanning the photo receptor to find where the most sensitive part is and make sure you always use the same part of it and so on and so forth. I guess it's pretty tedious.
And that's something that takes a good deal of practice in order to do it properly?
Yes. Sure. Yes.
And then the Antarctic —
You don't gain very much. You gain ten days, I suppose; extra obviously. Because the sun rises so quickly. Well, no, we tried to do a CD focus as early as one can. And then pondered with the idea of just going de-focused ones, which you can actually do at a lower time, but the single wavelengths effectively tells you nothing. Single wavelengths, prior to the double wavelengths now. Because that's when we were still arguing about absorption convergence and that sort of thing. And again, in the sense, if you simply ran through all the way and if you always did ACD measurements, as they were. You had your own data and you could make your things consistent. Whereas, it wasn't until somewhere around then that Dobson and various people actually published the new determinations of the absorption convergence. So there was always some bits of arguments about whether that's wise to say we — for the purposes of the paper we start very rigidly to the '80s, and one says the sort of basic things we should-we really couldn't be wrong. If you had a mixture of — if you tried to pluck things coming out of the sunrise it makes your CD-AD. You'd have a huge story on your hand to convince people it was a homogeneous bit of data.
Even with the regular measurements, and you have to use a Langley plot method in order to calibrate the instrument. But that method, you're assuming that you have a constant amount of ozone over the instrument. Right? When you're doing the Langley plots?
In the main, yes; not on anyone day. You know perfectly well there's not going to be so on anyone day. And then that does fairly restrict it. I mean, it is quite difficult to do that sort of thing in an Antarctic spring, for example, when basically you're sitting on increasing. In the old days, you were sitting on increasing ozone once the rising started. You might ask Brian or Jonathan again. Most of the ones that are actually used, I guess, would be mid-summer ones or even a bit later. Although ozone is — remember I showed you the graph.
Not on the two steep descending, but otherwise you're in trouble, yes indeed. What you can actually do — I mean, it doesn't really matter from the point of doing the Langley plots. You can actually cheat and work under Cirrus, because it's wavelength neutral. You can actually use — you don't have to have a perfect cloud-free sky. You can have the right sort of clouds and it doesn't make any great difference. And I don't know if you've ever looked at the original thing Hamilton and I published. You can actually sit to the end of the year and add up all the observations and convince yourself that there's no spherous and nonsense going on. So it provided you a clear sky to run your Langley plots. Or actually systematically scattered through the year, you'd get a reasonable answer.
Okay. Let me ask you. I’ve generally been asking this question of people I talk to, again, something I've asked you before, about the distinction between models and data. And one of the things is the extent to which the data has to be modeled in order for you to get useful, reliable information from it. And I want to say something to the effect that there is no absolute distinction between data and models. It's more like a relative distinction. Because data is going to be more or less modeled — like satellite data, its going to more heavily modeled than the stuff for Dobson. But even with the Dobson, the ozone values that you get — I mean, those are more model dependent then the dial readings you get off the instrument. But even, it seems, to some extent those measurements, that data, the dial readings themselves, are model dependent in some ways. Because like you said, you get this raw data back and you calculate an ozone value and it comes out to some really weird number. So you go back and look at it and you find out, "Well, maybe someone had a transcription error here."
Oh, sure. Yes. They're distinguished if they're relatively huge, and they're not distinguished with the small unfortunately. But that doesn't really come into it. I mean, there is a whole lot of interest between your dial reading and what goes on. I mean, that's basically what all the routine tests are about. The dial corresponds to a proper weight reading, or the weight reading is — I mean, one of the earliest things in '57 was when we — it was '58, wasn't it? That the early wedges were made of gelatin and stuff that would deteriorate. But that turned up, and we could correct for that to some extent. Not very well, but at least you recognized it from what's going on. And that was the new rhodium, which is not quite the same. But nevertheless, you do have to sort of make quite sure that the dial hasn't slipped relative to the — I mean, that's sort of what these things are about. So, yes, there is an exchange of entrants. But I don't think it really takes you into what I call the muddling stage quite. I don't think it's a question of you have a muddle in your data, that somebody has to fit into this in a way that you like and can appreciate.
But what about the example of where you have the transcription error though? Where that doesn't fit into what your model of —
Well, that's not a sort of model. That's simply saying, "Gosh, there's a nasty discontinuity if you believe this." And then basically my model is quite simple. I believe the hole in the atmosphere to be continuous.
But isn't that a model though?
Oh yes, that's a model in some sense. It's a conceptual model. But not in a very sophisticated way. Argentina is a case in point where you can see Dobson give its change in a day in the early spring. Sure. You can't tell if it's a transcription error or not until you drag in the 100 millibar temperature to see if the discontinuity is in the direction you're expecting. Yes, you have a — sure. I mean, the only way you can reject data is by having some idea of what it looks like and how it ought to correlate with other things. And it's a dangerous game. One shouldn't take that too far.
So I am saying that there's models involved. And if you want to say, "Well, they're not sophisticated. It's not like I'm using a general circulation model or something."
But this is true about any measurement you make in your life. If you take an AC meter and you plug it into the mains and it tells you that it's 110 volts instead of 230, you think, "Oh, perhaps I've forgotten to search the right range."
Yes. Well, see, this is something, again, that comes up in political debates over this. Is again, there seems to be these people that want to say, "Well, the data is absolutely reliable and the models, we've got problems with them."
Well, but those are the sort of people who don't understand the science at all, it seems to me, unfortunately. If you want to say something new, you pay a price. You pay the price of being wrong perhaps. But it's up to you. You're, after all, supposedly in charge of these young men making all these things. It would be fine if you measured all these things yourself and then you might even sort of feel a bit stronger about it. But, no, I mean, human beings are human beings. We all do silly things. Show me a year's observatory records in the old days and I'm sure if I couldn't pick it up and point out to you ten mistakes within short of half an hour, I should be very surprised. I mean, you do it now. You pick up papers now and you see all sorts of idiotic things which are simply miss-typings or whatever.
Even with the data, when you've got these models, even though they may not be very sophisticated or what-not, it seems to me what's important is not the fact that they're not that sophisticated or what-not, it's a matter of whether or not the results are robust with respect to the models. So that they either depend in the proper way upon the modeling you're doing, such that you know the data or the raw readings or what-not are supposed to be related this way to the model. You know, we know that. Or it doesn't depend on the model at all. So in some way you have to have some model in order to model your data. If you can show that over a certain range or what-not.
Yes. But this can only come into sense with experience. I mean, you're sent down to Antarctica to measure ozone layers for the first time. You just have to believe what you're getting. You scratch your head when something looks a bit strange and— I mean, you're encouraged, at worst, to go and check and see that you're doing all the sensible things. And then gradually over the years you sort of build up a feel for this. And as I said, perhaps if you find something else that correlates with it, it gives you an extra check on it. It's a funny game, I agree with you. Science isn't objective. I mean, anyone who thinks that science is objective —
I mean, this is one thing that I'm actually interested in, is the extent that science is objective, regardless of the fact that —
It's objective it the sense that you're duty bound to tell other people exactly what you've done so they can go on and do it again. And of course when you're making routine observations from day-to-day you can never go back to the day before and make the measurement again. But nevertheless, you're duty bound to explain you've gone through the normal procedures and you've done this, that, and the other. And sure, if you start correcting 50% transcription errors, you'd get rather worried, wouldn't you?
I mean, there are usually two or three, or three or four a year perhaps. And if you're lucky they can be captured, I suppose. But I think people have forgotten with the days of satellites, in a sense, what the old days were like.
So say from experience and what-not you would expect to have a certain number of transcription errors or other sorts of errors. And if you're starting to get more errors of a certain sort, then you generally expect — then you think something has gone wrong somewhere.
Then you think people are being horribly careless. I don't know. I don't even know how Jonathan and Brian run the damn things nowadays. But in the old days, we had two physicists you switched between different subjects in different weeks. And part of the job for one week would be to check what the other chap had done the previous week. And, I mean, it all gets horribly tedious. And I'm sure those things aren't done as I would like them to be done still. That was part of the game. Because in those days an awful lot was done by hand scaling. I mean, you'd made the record, you'd read the description, you'd estimate the stuff. And, as I said, in order to make sure that you keep mistakes as small as possible, it was checked. And as you worked through them —
I mean, it seems like something like that, where you're switching off the measurements like that, does introduce a greater degree of objectivity into the situation. Because then the measurements aren't so dependent upon one person.
That's right. No, I mean, if you — I have seen sort of records from one man, for one year, doing things. But he can make a complete mess of it indeed. South lGY, I believe.
Actually, Bob Evins was showing me some records from Dobson stations. And actually some of them— the calibration problems when they have a certain person who's running the station, when that person leaves and a new person comes in, how it changes. It's like clearly correlated completely to his —
Yes. I was in McMurdoch in one occasion, I was sitting talking to someone and there was a bang on the door and the chap that I was talking to at that time said, "Excuse me, this is my replacement." And then he said to the chap, "Now, this is this, this is this, this is this." And then he said to me, "Well, I guess that's the end of my year. I'm off now." And then he said to the other chap, "I should be about for a few days if you want to ask any questions." That's how you hand over a scientific program, really.
What year was this?
Oh, around '71.
When things weren't automated and there were chart records and everything. But nevertheless, it just — I mean, this was one of the problems sort of at Halley, where originally the chap that used to stay there for about seven days a year, after changing over was really rather difficult. Which is why we had this sort of retched system in a sense you had two signs — one for being there for a year, and the new one to sort of bring it along and pass it on to the next year. Whether it worked, I'm never sure. It made it more difficult to change things. Because you always had one chap who had been there the year before. "We did it this way last year." "Sorry, we've changed our mind. We're going to do it this way."
Going back to the construction of the Dobson instrument. I mean, the construction of the instrument does involve a good deal of theoretical knowledge, understanding of the atmosphere, electromagnetics, things like that. Optics. That's all built into the instrument. And so in that sense, any of the data you get out of the instrument, wouldn't you say that it's dependent upon all of that stuff that's built into the instrument?
Oh, sure. You're relying on — but have you not seen the back checklist?
The back checklist?
The checklist of what they have to do to determine that things are in the right place and so on and so forth. And sure, the only way — I guess instruments have gone out from there which were not in a very good state. But it should appear if people do the right sort of things when they're running it — oh, sure. I'm not saying any of us were ever clever enough to take the thing to bits and put it back together again and have it produce perfect Dobson’s. But nevertheless, Dobson did actually give enough stuff in the IGY manual that if the instrument was to be kept going at all. I mean, if it was in a state where you couldn't have kept going, you had enough information to help you do that. I don't know. Yes... and then I suppose it's possible theoretically to send one somewhere and produce absolutely rubbish.
I was just thinking to the extent of which instruments actually embodied models in this sort of space.
Yes. The beauty of the Dobson in the strict sense is that nothing mattered. That you don't actually have to make any absolute measurements of elimination. Right? Which is what New Zealand just tried to do with the atomic thing and made some silly remarks in consequence that they thought they had some more information. An actual fact, you get very little extra information from the absolute intensities. Certainly Dobson's saw it very clearly. But it took them a long while to realize that you need two wavelengths passed to eliminate the aerosol. And then perhaps in this day and age— well, he knew about sulfur dioxide. Most stations don't have proper sulfur dioxide, unless you're actually given a replacement. It did take a really long while to latch on to the right way to use it. It's very sad he had all these observations he used to take himself, which I suppose go back to '31 when he introduced the new ones. I mean, to really make sense of his early readings and it simply was received as experimenting and changing things all the time. Even the erosion measurements were the longest one. I mean, people keep forgetting in the early years of that as a matter of fact are of a one wavelength pair. There is probably as much in the change of weather and smoke from various places.
Yes. I mean, you've worked with other instruments as well. And you were just saying that with the Dobson, it took a long while for them to actually get to the point where they could get reliable measurements out. It was not like he just built the first instrument and, voila! You know, getting perfect ozone measurements.
No. Let's be fair. Long before he built the first photoletic [?] instrument, he had these wonderful various things which were all worked out photogramatically [?]. If you look back at that, that's not actually a very good determination of ozone. It told us what the world of distribution was from Antarctica and so on and so forth. Because relative measurements are actually quite good enough for that. But if you want an instrument which is going to give you trends, that's a very different game. And the Dobson isn't really a very good instrument for that, with all due honesty. If you have to keep relying on the Langley plots.
So why exactly is it not a good instrument for that? For trends, if you have to rely on the Langley plots?
Well, if you're looking for small trends, it takes you a number of years to get out of the error bars. I refuse to believe anyone who thinks he can run an ozone in the long term better than 2% or 3%. In principal, you could adjust. But, my God, you'd need some dedicated people to really trust them. And then the typical conditions in which they operate. I mean, it's much nearer to 5% or 10%.
So you can't get trends that are below that?
No. And which I said many times when people were looking for trends, in fact, if all we can expect is a 3% change in the studies stakes, it's going to take 30 or 40 years to be sure of that.
So that was one of the reasons why you didn't expect to see any decrease in —
That. But also partly because, as I said, the story, as I understood it at any rate, it was really perfectly clear that the first place you see this is in the tropical upper stratosphere. They would have seen it in the tropical upper stratosphere.
Sort of like when someone sees —
And then Antarctic was right at the sort of end of the factory, as it were.
So, "Wait for someone else to see it in the tropics and then we'll look for it," type of thing?
No. I mean, we were in a sense looking, but not with intense interest there, because of all the problems having to wait for the days to come back and the collaborations and all the rest of it. You didn't really feel there was much —
Actually, I guess I need to ask a question about the Amundsen Scott data from the South Pole for around this time for October/November data for 1983. Supposedly what happened was — I have some specific questions about what possibly could have gone wrong. But what happened with the NASA people in 1984 was that they noticed the low ozone values in July — As I said, the NASA people, when they were looking at their July data, one of the first things they noticed was these low values.
When you say July, I don't know —
July of '84. July, 1984, when NASA first —
What were they looking at? You can't take data and —
Their TOMs-they were looking at their TOMs data from October, November, and December. And it took them about six months to process the data. I mean, actually I talked to Al Flagg who headed the ozone processing team. And from the time TOMs actually took a measurement until the time they could fully process it, they could have done it in six hours. But it took them six months. Because, as with Bass, they had other things to do. So when they first processed the data from October of '83, they noticed these low values. And they went off — they wanted to check against a ground station.
Yes, they would have flagged it at that stage. I mean, they were still —
Well, they were flagged, but when they noticed all these flags they went and looked at them and they were getting all the low values. So what they did immediately was try to check it against a ground station. And they were checking at the South Pole station. And they were noticing that — well, the South Pole was showing normal values. So they thought there was a problem with the instrument. So they want back, they looked at their instrument again, and sometime in the Fall of '84, they decided, "Well, there's a problem with the South Pole data. We've got real low values." Although they didn't know what to make of them at this point.
They were checking like with like.
They were what?
They were checking like with like. They were actually taking the over-pass values at the South Pole?
Okay. I mean, again, I don't carry all these things in my head. But one would need to look back and see how the vortex was behaving. You could well imagine years in which the vortex would have passed over the Pole quite early. And they would indeed, at the end of October, be seeing high values. That's just why I was curious as to —
That the Pole would be seeing high values?
Sure. It would see high values usually before Halley does, you see. In the Halley at least it's the end of the road, as it were, and the vortex is going away. So you've got more time to see the low values. I'd need to look back and see what' 84 was like and —
I mean, now, after the fact, Komer, in a paper that was published — I think it was Journal of Geophysical Research or Geophysical Research letters, one or the other, in '86, actually noted that the values for October, November, December, were erroneous. As he said, "erroneous and uncorrectable." And they didn't realize that until after the fact. So they had been sent to the WMO, and so that delayed NASA's processing and all of that. But the question is this, because he actually published-what he said there was that, and when I talked to him, is that he thought the measurements were taken instead of AD, ADA, which they were taken on the primed wavelengths instead. And I've asked Brian this and John this. I mean, is taking the measurements like that, could that have caused the sort of problem where they thought they were getting normal values, around 300? I mean, if the values were actually around 180 or something like that, as the TOMs data indicates, could taking the measurements on the prime wavelengths actually show that they were around 300 consistently? Or would you get something more erratic?
I'm not even sure I understand you. I don't know what you call the prime — I call the prime wavelengths a pair. I mean, the Dobson only measures, as I've explained, ratios of Lambda 1 to Lambda 1 dash, if you like, and Lamba 2 to Lamba 2 dash. You can't measure these separately. What you can do is that there's a separate lever on the thing which changes these to the long settings. Without that, you'd just produce rubbish. You wouldn't produce anything reasonable at all. Subject to correction, it never even occurred to me.
Now, that's the question. I've talked to Bob Grass as well. I mean, Bob Grass [???] to think that the lever was pushed in.
Well, if he thinks the lever was pushed in, I would be surprised that anything would ever be workable at all. Do you know what they were trying to take? I mean, were they trying to take —
I think they were doing AD —
Well, AD direction [???] cloud, which means a fairly high —
These are apparently all the observations for those months. Now, if the lever was in like that, would the —
Yes, it gives an answer. But, I mean, it's got nothing to do — anything interesting to do with ozone at all. The idea of C-1s is when you go from C to C dash, you're moving into a bit of the region which isn't affected by ozone, but which is merely determined by the Riley scattering and so on and so forth. So if you're working on the dash ones of these, you're beyond the Huggins bands, you're not really — it doesn't have anything to do with anything sensible. I mean, much more likely. I mean, like I said, it's likely that someone set the wavelengths wrong maybe.
How would they have set the wavelengths wrong? On the key levers?
Yes, on the key levers you have stops. You have two stops which you set to A and D. So you don't have to sit and fiddle every time; you just back the lever against it. If you set those wrongly and someone hasn't checked them — I mean, their temperature dependent, so you have to read the thermometer and then you set them to where you want to be. And the temperature is regulated. The temperature very well might change. So they may well sit there —
So someone could have taken —
If they were set wrongly, then of course you're not putting the right bit of the Huggins bands over the slit. So you can get any answer you like as to make high ozone and low ozone, depending on what you do.
Yes. So if someone took a measurement and got a really low value when they were expecting something high, could they have adjusted the key levers to get what they thought and just left them there?
Well, I wouldn't — I don't know.
I know that's not the proper way to do anything.
It would never occur to me to do that.
I mean, I'm just trying to figure out what potentially — I mean, I'm trying to talk to as many people that operate the instrument as possible to figure out what —
Well, this is the whole problem for our sins. I mean, Jonathan Nissage would probably know much more than I did. The only people I've ever been concerned with is to how they teach people to use it, where the MET office now is off the record. I'm so disgusted with the way they've sort of taught people. I used to go up myself and try and teach people what I thought was right. No, it doesn't really ring sense to me.
So if the lever had been in —
One of the things we used to do that other people don't do very often was actually to do the spectrum and find where these ratio things were and convince ourselves that the wavelengths were sitting in a right way that were sensible. But I don't know, it could well be-there were kind of silly things, like people didn't read correctly where — I mean, they thought they had to set the center of the stop instead of where the lever pushes against it. I don't know. One could only imagine. Or even in a sense I suppose that the damn things worked lose or were moved. I have no idea. I mean, I could conjecture all sorts of things.
So if the lever was actually pushed in when it should have been pulled out —
That, I think, would just produce nonsense.
I can't — I may be wrong. But you might, by luck, find yourself in the region where it made some sort of sense. But I really can't believe it.
I was trying to talk Bob Evins into actually trying to take some measurements like that and see what it actually produced.
Well, sure, to see what happens. I don't know. I mean, you're going a lot off the record, in a sense.
So it just doesn't ring a bell?
When one — it's really surprising how few times the American's have actually tried to use the Dobson at the South Pole and how often they seem not to be able to use it. Which I find a little surprising. I cannot visualize anything which does that at all. I mean, the more likely — the only thing which it can really do in these things is to displace the wedge relative to the dial. In which case you do get a systematic effect and it can move things up and down.
That would only happen if you actually got inside the instrument to do that?
Well, yeah, or if something slipped. Yes.
Or if something slipped.
They're on a supposedly non-slip —
So, I mean, if something like that happened, there would probably be some record of them having gone into the — I mean, you would —
Well, they ought to have wedge-tests. I mean, you do a wedge-test to find out what- you know, you've got a piece of glass with a known absorption which you shove into the thing and it should tell you whether or not you're wedge-style tables are right. So I wouldn't have —
I'm trying to get the original calibration records and everything for the instrument so I can actually figure out what might have happen. But everybody I've talked to since Komer says that that explanation just doesn't seem to ring.
It doesn't really seem very sensible to me. I would think frankly it sort of went off weight scale at times.
If they were doing it on —
If they were doing it on the other wavelengths with that. I need to think about it. I've long sense forgotten all the details of Dobson's over— what else can you do? The only other thing is shut off slits. If you shut off slits, again, you take everything off scale. No, I have no idea.
So there's nothing else you can do that would give you systematically higher measurements that you couldn't correct later?
No. I mean, if it's going to be systematic, it has to be something to do with the wedges. I mean, the only systematic — the better the instrument —
But that's something you could correct after the fact, if you knew what —
Yes, if you knew what was going on. Indeed. I don't know. Even setting the wavelengths wrong doesn't really give you that much of a systematic effect.
The key lever?
Yes, I mean, if you put those in the wrong place — even if you put them consistently in the wrong place, it's going to give a nonlinear effect, depending on the [inaudible] . You wouldn't see a systematic effect, no.
I have no idea.
Well, I guess that's about it.
The inquisition is over. All right. Months leading up to it, the opinion basically was that any CFC trend was just going to be horribly small. And as I say, I think it's still fair to say that basically in those days the discovering of who is trending your data was tantamount to see what the hell you were doing wrong. So that one's first reaction was, "Gosh, we better go through it with a fine-tooth comb what all the young men have doing and what was going on."
What was your reaction when you finally convinced yourself that this was a real effect in your data? I mean, you thought it was the CFC's.
Very excited for sure. Why didn't we see foresee this? What the hell is going on? It's going to be an exciting mixed party.
When you say excitement, was there anything like —
Well, sure. Yes, indeed.
Like, "My God, what's going on?"
Yes. It's moving so fast, how the hell is it going to stop, and where is it going to stop? Basically I suppose even then it wasn't very clear. But you felt it couldn't go much beyond two-thirds. No, no, it was an exciting time certainly. As I said, basically what should one be doing? Were any politicians early on or any position of power pays attention to you, until someone comes along and confirms it for you.
I mean, it really didn't take off until like the —
Oh, I gave two or three talks I suppose. Because I say [???] to know about it, and so on and so forth. But no, it wasn't until really the American's re-worked the stuff in October of '86, and then you could sort of see it.
So you had these nice pretty images of the hole?
That's right. You could actually see where it was. And you could sort of see why different stations might or might not see it. One of the sorts of problems was is it only Halley and Faraday seeing it? Or why the hell why? It doesn't make sense that one side of Antarctica sees it and the other side doesn't. It just doesn't really make sense, particularly dynamically. Why the hell should suddenly dynamics decide to play a funny game in half the vortex? It can't really be done. So one was not really sort of prepared to believe the diagnosis from that point.
Well, are there any other people that you may be discussed this stuff with prior to the publication, which we haven't mentioned? There was Thrush, Mergatroy —
That must be it, I suspect.
Did you talk about this at all with Reikroft? Because I knew, as I said, Shanklin —
Not really. Reikroft and I had a hate relationship. I mean, I just couldn't sort of talk to him and get a sense — I mean, I couldn't even keep my temper talking to him, frankly. He was a very curious man. I'm a curious man too. I mean, we all have our ways. But I just couldn't talk science to him. He didn't understand how difficult it is to take a year's Dobson measurements. I mean, this is what is very difficult to bring home to people. It requires dedication on the part of the young men. And you've got to encourage them along and so on and so forth. And it's no good at sort of saying in essence, you know, "You're just wasting your time doing these measurements." I mean, if you believe that, you stop them. He would have gotten along quite happily with ??? that we should stop them as being valueless. Perhaps we should have done — I mean, who can tell. You can go on making measurements forever. If it turns up trumps, it doesn't justify your...
John Shanklin gave me a copy of the paper that he gave to you and Brian and Reikroft. And from the comments that Reikroft gave back to Shanklin, it seemed like he thought Shanklin was going to publish the paper, which I don't think Shanklin was thinking about when he wrote it. But it seemed clear that Reikroft thought that it was publishable as it was. And what do you have to say to that? Because I know you thought there were problems to begin with.
I didn't see how one could publish it without really doing at least a bag of bones over the homogenization for the years I finished doing it. You know, frankly, if you wanted to publish a simply month's trend, once it occurred, you really had to back it up by pointing out what you had done to substantiate it. And then, I don't know, in the course of discussions with Bob and Brian, I suppose, this other way of looking at it sort of dawned on me and it seemed much more robust. I don't know really. It's funny in some ways, because we were all sort of more or less getting in each other's hair, and hating it, I supposed. Jonathan sort of kept saying things and I'd sort of say gently, "Yes, but please show me the results of all the monthly tests and so on so you can convince me that this is going on." And then he'd sort of disappear. And I wouldn't see him for a couple of weeks. And then he hadn't done it, and so on and so forth. So it was a funny time. Brian was sort of mixed in between. I mean, he was technically Jonathan's boss, as it were. And I was sort of somewhere above both of them. And so it got to be a somewhat awkward time.
So, was your attitude something to the effect of, you know, there's probably something here in the data, but you need to back it up more?
Yes, sure. Yes.
You need to show me that this isn't some artifact.
Sure. I mean, you know, it's perfectly obvious to say if you don't do things properly, then you can produce horrible artifacts in books and data.
What potential problems — I mean, in this paper, these two graphs where you're showing there's a downward trend, how could that have been produced in proper data processing? Or data analysis?
Not from that, but from neglect of the routine monthly tests and all these sort of things. And then the Langley plots. You know, a particular station running a Dobson will produce some discontinuities. One every three years, or more if people fiddle about too much. I mean, the Dobson is an instrument in the whole world which philosophy to me says, "Leave well alone. Keep doing your tests, but don't make modifications. Only make a modification if you're really convinced that something has really shifted or something too stupid has gone on." I mean, I'll see if I can dig out that Angel and Korshivo for you. I mean, they published Kaddock and Nowell, which went up by 40 Dobson units in ten years. That's how you produce a trend. I don't know how other people do it, but that's the sort of thing you can produce.
And what explained, again, why that was going up? Was that just something that was done with the instrument?
I think so. A wedge was deteriorating and they hadn't sort of caught it.
So with like the Halley Bay, with the —,
When — I mean, I guess in the early stages of the debate of whether we should write it up or what we should do, I would not have been too surprised to find that someone had done some of these earlier and it was instrumental. It wasn't within the bounds of possibility it seemed to me at the time.
Yes. Because you got this new instrument that was put in in January of '82. So, I mean, there could have been maybe something due to the new instrument that was causing the downward trend.
Yes. My guess would be that with the delay of moving things backwards and forwards that we hadn't really sat down to see a comparison of the two instruments at that stage.
That was January, '82?
I'm just baffled by this statement.
That was January of '82?
No. The instrument didn't actually stay there for a full year.
Yes. John and Shanklin said it was about two weeks or so.
I don't know. They have the things-I don't. And if I didn't, I'm baffled.
So there could have been a possible downward trend in the data caused from the new instrument going. I mean, could there?
Well, yeah. But it wouldn't —
It would be a level shift rather than a drift. So you can interpret. It depends on how you want to interpret the thing.
I mean, was that considered at this point? I mean, he's showing you this downward trend in the data. And is that something you would have expected with a new instrument going in like that?
But you never know.
No. I've forgotten all these details now. I guess we were just happy that the two things gave the right reading in the January of '82. As you said, there's so much gossip going on these days.
I can send an e-mail again and check. Because, I wanted to get —
I'll ring up Brian and ask him.
John was going to search around for the shipping manifest from — or look for some record when it actually came back from Halley. So I'll find out. But, I mean, he seemed to think it was only down there for two weeks, sitting side-by-side, before it was shipped back.
He might be certainly right.
Oh, well. Okay. Well, I guess that's about it for now then.