1
00:00:14,430 --> 00:00:19,560
Thank you for that kind introduction and it's great to be here. You all for coming this evening?
2
00:00:20,220 --> 00:00:25,740
I'm going to start by, as Eileen says, winding the clock back a bit and in fact,
3
00:00:26,220 --> 00:00:34,470
start by referring to what was an extremely what has proved to be an extremely influential popular science book,
4
00:00:34,470 --> 00:00:43,200
which was published in the second half of the 1980s by James Gleick, because it was this book that coined this iconic phrase,
5
00:00:43,200 --> 00:00:50,670
The Butterfly Effect, which I think has pretty much gone into everyday language and popular culture.
6
00:00:52,770 --> 00:00:57,420
In fact, the very first chapter of the book is called The Butterfly Effect.
7
00:00:57,900 --> 00:01:06,690
And in this chapter, Glick describes the work of MIT meteorologist Ed Lorenz, whose picture is up there,
8
00:01:07,200 --> 00:01:15,990
and a paper that he wrote back in 1963, which was pretty much ignored, I would say, for ten years or so.
9
00:01:16,380 --> 00:01:22,410
It was published in what most scientists think of as an obscure metrological journal.
10
00:01:23,040 --> 00:01:26,050
Turns out, actually it's the premier meteorological journal in the field.
11
00:01:26,100 --> 00:01:34,140
Anyway, there we go. But what Gleick describes, and I'm going to talk about this at some length in the talk,
12
00:01:34,950 --> 00:01:46,499
is how Lorenz came to discover three very simple equations which had no determinism or no randomness.
13
00:01:46,500 --> 00:01:52,200
Everything was precisely defined. The three equations coupled together in a way which I'll tell you.
14
00:01:52,920 --> 00:02:03,149
And it had the property that it was almost impossible to predict the future of these equations, in the sense that starting with just a tiny,
15
00:02:03,150 --> 00:02:09,960
tiny, tiny perturbation, let's say, to the initial state, the solution developed quite differently.
16
00:02:10,470 --> 00:02:14,370
So let me actually illustrate that right at the beginning with this animation.
17
00:02:14,370 --> 00:02:17,669
So we're going to look at the evolution in time of what actually are.
18
00:02:17,670 --> 00:02:25,290
Two, it's not apparent until about now that there are actually two different solutions of these equations.
19
00:02:25,740 --> 00:02:31,530
Starting as I say, from almost but not quite identical initial conditions.
20
00:02:32,040 --> 00:02:38,950
And after a while, although they track together the red and the blue for a while, they sort of correlate.
21
00:02:39,690 --> 00:02:48,060
And so knowing one of the solutions, if you like, tells you almost nothing about the second solution after a certain time.
22
00:02:50,190 --> 00:03:01,260
Now, Glik then went on to say that this model of Lorenz basically formed a, you know,
23
00:03:01,440 --> 00:03:09,059
a hard quantitative bedrock for his notion that he'd been that he Lawrence had
24
00:03:09,060 --> 00:03:16,300
been talking about for some years of how the flap of a butterfly wings could go,
25
00:03:16,350 --> 00:03:21,090
let's say, in Brazil, could cause a tornado, let's say, in Texas.
26
00:03:22,170 --> 00:03:27,749
Now, this is where the history actually becomes slightly confused.
27
00:03:27,750 --> 00:03:30,990
And one of the. So I've got two parts to my talk.
28
00:03:31,440 --> 00:03:34,860
One is to try to review what is pretty standard stuff,
29
00:03:34,860 --> 00:03:45,450
which is how Lorenz came to arrive at these equations and some of the amazing insights he had, incidentally, in linking these very simple,
30
00:03:45,960 --> 00:03:54,480
essentially equations that Isaac Newton would have understood very well with modern 20th century fractal geometry,
31
00:03:54,480 --> 00:04:04,620
a sort of concept that Newton would have found quite alien. So there's enormous insights that Lorenz had in deriving his 1963 model.
32
00:04:06,060 --> 00:04:14,400
But I also want to say, and this is something which I'm sure most most of you will not know is that actually Glick got it completely
33
00:04:14,400 --> 00:04:22,950
wrong about attributing this 63 model to the idea that Lorenz had in when you referred to the butterfly effect.
34
00:04:23,310 --> 00:04:26,790
So unfortunately, the butterfly effect has actually been misnamed.
35
00:04:26,880 --> 00:04:29,250
And I'm going to try to explain why that is.
36
00:04:31,240 --> 00:04:39,910
So the term the term well, the term the butterfly effect comes from Glick's book, but essentially refers to this paper or this talk,
37
00:04:39,910 --> 00:04:48,370
let's say, that Lorenz gave in 1972 at the American Association for the Advancement of Science and the title.
38
00:04:48,490 --> 00:04:55,150
It was a kind of a sort of semi it was essentially a public talk about predictability of weather.
39
00:04:56,110 --> 00:05:00,520
And that was the title. Does the flap of a butterfly wings in Brazil set off a tornado in Texas?
40
00:05:01,150 --> 00:05:06,250
So what I wanted to do in the sort of towards the end of the talk is actually
41
00:05:06,700 --> 00:05:13,480
to to say a bit more about precisely what Lorenz said in this 1972 meeting.
42
00:05:13,810 --> 00:05:21,910
And you will see that it's actually not referring to his 63 paper at all, but to something he published, in fact, in 1969.
43
00:05:21,910 --> 00:05:26,530
And that is much less well known, I would say, in the broad community.
44
00:05:27,160 --> 00:05:35,380
And in fact, this problem that Lorenz Lorenz wrote about in the 69 paper and which he talked about in the 72 meeting,
45
00:05:35,710 --> 00:05:43,360
turns out to be one of the great unsolved or closely linked one of the great unsolved problems in 21st century mathematics.
46
00:05:44,050 --> 00:05:51,370
So it actually gets to the heart of something which is actually much more radical than kind of chaos theory,
47
00:05:51,370 --> 00:05:55,090
although you might think chaos theory is fairly radical, but this is even yet more radical.
48
00:05:56,620 --> 00:05:58,350
So so that's the outline of the talk.
49
00:05:58,360 --> 00:06:05,410
You'll know some of it maybe, and I'll fill in some details, but towards the end I'll say stuff that I'm sure most of you don't know.
50
00:06:05,590 --> 00:06:08,980
And I hope to I'm not going to really change popular culture.
51
00:06:09,440 --> 00:06:16,890
It'll it will always be called the butterfly effect. But I'm going to call what I what what Lorenz really meant the real butterfly effect.
52
00:06:16,900 --> 00:06:23,799
So we'll come to that towards the end of the talk. Let me just say a little bit about the biography of Lorenz.
53
00:06:23,800 --> 00:06:32,650
So here he is, is as a young man, grew up in New Hampshire, very interested in astronomy as a kid,
54
00:06:33,820 --> 00:06:38,580
apparently had a passing interest in weather, but it wasn't a particularly sort of passion of his.
55
00:06:38,660 --> 00:06:43,240
In fact, probably weather was a nuisance because it obliterated the stars when he wanted to go out and look at them.
56
00:06:44,260 --> 00:06:48,910
But he claims, you know, in in later years, he had he had somewhat of an interest in weather.
57
00:06:49,810 --> 00:06:53,260
But he went to university, he went to study mathematics.
58
00:06:53,950 --> 00:06:59,169
And, in fact, he went to Harvard and went through his undergraduate degree and in fact,
59
00:06:59,170 --> 00:07:04,360
started a postgraduate degree in mathematics under the great George Birkhoff,
60
00:07:04,900 --> 00:07:09,850
where he was studying some topic in apparently remaining in geometry, nothing directly to do with chaos.
61
00:07:10,960 --> 00:07:15,460
And his plan was to carry on. That was going to be his Ph.D. in mathematics.
62
00:07:15,880 --> 00:07:20,590
But then the Second World War came and sort of upset the plan.
63
00:07:21,490 --> 00:07:31,330
And I think by chance, Lorenz discovered an advertisement that the CIA, the Army or the or the Air Force, I think it was the army at the time.
64
00:07:32,230 --> 00:07:38,920
They were looking for weather forecasters. Maybe it's the Air Force because it was two planes and so on.
65
00:07:40,330 --> 00:07:44,920
So he thought, well, okay, well, I've have a thought back to his childhood, a little bit of interest in weather.
66
00:07:44,920 --> 00:07:48,670
So maybe I'll that will be a good sort of career for me during the war.
67
00:07:48,700 --> 00:07:55,509
So he he joined up and, you know, he was sufficiently good that after he'd been through this course,
68
00:07:55,510 --> 00:08:00,129
they decided he was the right person to teach the next lot of weather forecasters to be weather forecasters.
69
00:08:00,130 --> 00:08:08,710
So he actually became a tutor, if you like, in weather forecasting for the for the Air Force.
70
00:08:09,670 --> 00:08:13,749
He eventually got sent out to Okinawa, I think, in the Pacific.
71
00:08:13,750 --> 00:08:19,420
So he actually was involved in some of the real action in the in the Pacific sector.
72
00:08:21,170 --> 00:08:28,549
Came back after the war and thought about, okay, do I restart my career in mathematics and decided,
73
00:08:28,550 --> 00:08:31,550
actually, you take quite an interest in a real interest in Weber.
74
00:08:32,060 --> 00:08:40,550
And so instead he enrolled at MIT just up the road to do a Ph.D. in meteorology.
75
00:08:42,010 --> 00:08:50,140
And the topic he was essentially given was trying to solve this equation.
76
00:08:50,410 --> 00:08:56,890
So it doesn't matter if you don't know this equation, but this is one of the archetypal equations in fluid mechanics.
77
00:08:57,310 --> 00:09:01,690
It's essentially Newton's second law of motion, you know, suit and circular motion force.
78
00:09:01,690 --> 00:09:05,709
It was mass times acceleration. This is well, it's actually the other way around.
79
00:09:05,710 --> 00:09:13,180
It sets knots times acceleration equals force, but written for a fluid, which could be the atmosphere or it could be the oceans,
80
00:09:13,180 --> 00:09:18,190
or it could be a laboratory fluid, a fluid that potentially has many, many, many scales of motion.
81
00:09:19,150 --> 00:09:22,960
And it is actually a remarkable thing that if you count, remember,
82
00:09:23,290 --> 00:09:27,280
the last time I counted the number of symbols, it was around 20 something, 22 or so.
83
00:09:27,520 --> 00:09:34,030
Mathematical symbols is always strikes me as a wonderful thing that with just 22 mathematical symbols,
84
00:09:34,030 --> 00:09:41,559
you can describe the dynamics of every scale of motion in the atmosphere from the very largest jet streams,
85
00:09:41,560 --> 00:09:48,170
you know, which extends thousands of kilometres in length in the upper atmosphere, you know,
86
00:09:48,250 --> 00:09:52,990
right down to clouds, right down to the little eddies coming out of my mouth as I speak.
87
00:09:53,050 --> 00:09:56,200
This is all described by this one set of equations.
88
00:09:56,200 --> 00:10:00,250
So this is Lorenz's job to look at. How do you know the problem is?
89
00:10:01,420 --> 00:10:07,740
So, okay, so this is this is this equation can be likened to very much a work of art.
90
00:10:07,750 --> 00:10:15,260
It is a work of art. But instead of likening it to a Renoir here, I'm going to liken it to a Russian doll.
91
00:10:16,980 --> 00:10:23,390
This is actually a very special Russian doll. Russian dolls unpack into small Russian dolls.
92
00:10:23,780 --> 00:10:30,439
This one unpacks into yet smaller Russian dolls and indeed yet smaller Russian dolls, and indeed yet smaller Russian dolls.
93
00:10:30,440 --> 00:10:33,500
And indeed, yet smaller Russian dolls. Dot, dot, dot.
94
00:10:36,300 --> 00:10:41,760
And in the same way, if you actually want to solve this equation, you have to, as it were,
95
00:10:41,760 --> 00:10:47,340
unpack it also into actually what turned out to be billions of individual equations.
96
00:10:47,760 --> 00:10:50,969
This is actually what makes weather forecasting one of the things that makes it so difficult,
97
00:10:50,970 --> 00:10:56,400
because you need to unpack it into the equations which describe the big jet stream,
98
00:10:56,400 --> 00:11:01,410
which describes low pressure systems, which describe clouds, which describe some cloud turbulence and all that stuff.
99
00:11:03,120 --> 00:11:13,420
So the way to solve these equations, which is what Lorenz did, is, is to look at truncating these equations.
100
00:11:13,440 --> 00:11:24,180
Oh, sorry. Before I do that, yes. Let's let's say that these Russian dolls can be likened to, you know, the worlds or the eddies in a turbulent fluid.
101
00:11:24,480 --> 00:11:33,030
And in fact, it's it's probably appropriate at this stage to refer to the little piece of doggerel that was written by one of the founding.
102
00:11:33,180 --> 00:11:39,900
Well, one of the real pioneers of turbulence theory and also, incidentally, a pioneer of of weather forecasting.
103
00:11:40,290 --> 00:11:44,760
Lewis Fry Richardson, sometime in the beginning of the century.
104
00:11:45,390 --> 00:11:53,160
Anyway, this famous little poem, big whirls of little whirls that feed on their velocity and little wells have lesser wells and so on to viscosity.
105
00:11:54,610 --> 00:12:01,979
Okay. What a great poet. But still, it makes the point that in a turbulent system, you have these many, many, many scales.
106
00:12:01,980 --> 00:12:07,590
So these scales, the big wells or the big Russian doll select little wells, the of the smaller Russian dolls and so on.
107
00:12:08,520 --> 00:12:13,979
And the fact that this notion of feeding on their velocity captures this idea that, as it were,
108
00:12:13,980 --> 00:12:19,350
the Russian dolls can sort of bash into each other and transfer energy from one one Russian doll to another.
109
00:12:19,860 --> 00:12:26,999
And this is manifest mathematically in the fact that the equation here this you is the fluid velocity,
110
00:12:27,000 --> 00:12:36,150
and it sort of multiplies itself and that makes the system nonlinear and not linearity is the thing that allows energy to move up and down the scales.
111
00:12:36,480 --> 00:12:44,809
And that's one of the complications of things. So Lorenz's PhD was actually about trying to whips was actually if I could go back,
112
00:12:44,810 --> 00:12:50,510
was trying to solve these equations in a sort of simplified way by getting
113
00:12:50,510 --> 00:12:55,669
rid of some of the scales and just treating maybe the system in a simplified,
114
00:12:55,670 --> 00:13:04,110
approximate way. And he came up with some what are called time stepping schemes and other types of what are called numerical schemes,
115
00:13:04,110 --> 00:13:10,709
which actually even to this day are still used. Anyway, the study was, I would say, moderately successful.
116
00:13:10,710 --> 00:13:13,830
It wasn't like I set the world on fire, but it was moderately successful.
117
00:13:14,340 --> 00:13:22,050
And he found that within a few years of his PhD, he'd been invited back to MIT as a proper member of faculty.
118
00:13:22,950 --> 00:13:28,170
But there was a slight rider on this appointment because he had to be in charge of a group
119
00:13:28,410 --> 00:13:33,470
that was whose kind of principal research activity was doing long range weather forecasting.
120
00:13:33,480 --> 00:13:37,020
So long range means, you know, a month or two ahead,
121
00:13:37,620 --> 00:13:42,630
whereas the people that were trying to kind of approach weather forecasting from the from
122
00:13:42,690 --> 00:13:45,960
the point of view of solving these navier-stokes equations were only thinking about,
123
00:13:46,430 --> 00:13:50,940
you know, maybe 12 hours ahead at the moment or maybe a day ahead at most, but just a very short time.
124
00:13:52,290 --> 00:13:54,989
So how would people even think about doing long range forecasting?
125
00:13:54,990 --> 00:13:59,040
So this was actually done by a completely different method to solving these equations.
126
00:13:59,040 --> 00:14:07,280
It was just done with the statistics. So the idea is you have a big pile of weather maps which let's say we do this today.
127
00:14:07,290 --> 00:14:13,140
We got a big part of weather maps from today going right back to whenever you like, 1800 for the sake of argument.
128
00:14:15,000 --> 00:14:22,650
So we're in what have we our may. So let's suppose our task is to forecast what the weather's going to be like a month from now.
129
00:14:22,660 --> 00:14:26,040
So the monthly forecast for June or July or something.
130
00:14:27,540 --> 00:14:35,670
So the idea that this group had and incidentally, he got a lot of this group got a lot of support from the Statistics Department at MIT.
131
00:14:35,820 --> 00:14:45,530
So this this this had some pretty high level support was you just go back in time and find a weather map that from the past that looks like today.
132
00:14:45,540 --> 00:14:52,470
So you go back and find, let's say 1961, May 1961, the weather looked very similar to May 2017.
133
00:14:53,490 --> 00:15:04,920
Then what you do, you predict for June, July 2017, what the weather was in May and June of whatever I said, 1961, it's kind of method of analogues.
134
00:15:06,150 --> 00:15:12,990
So Lawrence got put in charge of this team, and his first reaction was, Hmm, this didn't sound right to me.
135
00:15:13,020 --> 00:15:17,639
You know, what's the scientific basis for this? This idea of just so you can find an analogue?
136
00:15:17,640 --> 00:15:20,969
Because it seemed to suggest that the weather was somehow very periodic.
137
00:15:20,970 --> 00:15:30,600
You know, if it if it was safe, it was like it is in May 61, that necessarily what we see in the next couple of months will be what happened in 1961.
138
00:15:30,840 --> 00:15:34,680
And his kind of intuition was that the weather isn't periodic like this.
139
00:15:34,950 --> 00:15:43,110
It seems to be irregular. And apparently had long battles with people not only in this group, but again with the statistics, people that I met.
140
00:15:44,160 --> 00:15:49,620
And he said, you know, I've got to try and sort this out. I'm going to try to prove that this really this idea isn't going to work.
141
00:15:49,800 --> 00:15:56,490
How am I going to do it? So that's when he went back to his truncated navier-stokes equations.
142
00:15:57,030 --> 00:16:02,490
So maybe I can prove that this property of of periodicity just doesn't happen.
143
00:16:02,850 --> 00:16:09,120
If I take a semi realistic model and use I mean, computer technology was, was starting to in the fifties,
144
00:16:09,120 --> 00:16:17,040
computer technology was starting to arise so I could solve these equations on a computer and just show them that this won't work.
145
00:16:18,450 --> 00:16:22,580
So he set about, you know, taking actual weather type of equations.
146
00:16:22,590 --> 00:16:26,280
And for a long time, he worked with many Russian dolls of the eight.
147
00:16:26,310 --> 00:16:31,170
So actually, he worked for a long time with 12, you know, a 12 component.
148
00:16:32,760 --> 00:16:36,899
So truncating the equation down to 12 compares. But, you know, the computers just weren't up to the job.
149
00:16:36,900 --> 00:16:40,950
And he found this was really hard work trying to do anything.
150
00:16:40,950 --> 00:16:43,500
It just took too long to get anywhere.
151
00:16:44,160 --> 00:16:52,739
And he had a real breakthrough when he talked with a colleague of his from a nearby university who had been studying not so much weather,
152
00:16:52,740 --> 00:16:59,940
but what's called convection in a in a fluid where you heat a fluid from below and look at the circulations that develop.
153
00:17:00,900 --> 00:17:09,960
And this guy actually said that he had a seven component model that seemed to seem to have this property of non periodicity.
154
00:17:10,480 --> 00:17:16,410
Maybe Lorenz could look at that and Lorenz did in fact move on to that problem and
155
00:17:16,410 --> 00:17:21,720
he realised within this subset of seven equations there was actually a three.
156
00:17:22,110 --> 00:17:29,190
So he realised when he was looking at these non periodic solutions that four of these components almost went to zero.
157
00:17:29,200 --> 00:17:32,330
So we kind of had the intuition that there would be a three component.
158
00:17:32,340 --> 00:17:36,330
He had a three component subset that.
159
00:17:39,760 --> 00:17:47,790
That had the required property. And this is what led him to these iconic equations, to Lorenz 63 equations.
160
00:17:47,790 --> 00:17:52,180
So this is just a three Russian old truncation of a laboratory fluid.
161
00:17:52,930 --> 00:17:57,280
So you've got three variables X, Y and Z. So these are not space.
162
00:17:57,280 --> 00:18:00,970
These sort of describe a type of circulation. It doesn't really matter.
163
00:18:01,090 --> 00:18:07,059
This this equation has been truncated so much, it now sort of likes to lose contact a bit with the real world.
164
00:18:07,060 --> 00:18:13,930
So it's kind of an idealisation of the real world. So X, Y and Z, just variables thinking the variables.
165
00:18:13,930 --> 00:18:22,150
And in the three equations, the left hand side, the x by d t d y by d t d set by d to use the time rate of change of X, Y, and Z.
166
00:18:22,510 --> 00:18:28,400
And it's given by these terms on the right hand side. The numbers, ten and 28 are not.
167
00:18:28,820 --> 00:18:32,149
They don't have to be precisely those. These are the numbers that Lawrence used.
168
00:18:32,150 --> 00:18:35,540
But you can, you know, choose numbers nearby and you get the same behaviour.
169
00:18:36,840 --> 00:18:44,670
And you notice that in the right hand side you get these terms multiplied together, X and Z multiplied together, X and Y multiplied together.
170
00:18:45,000 --> 00:18:49,590
So it does retain that notion of nonlinearity, which as we come to, is all important.
171
00:18:51,260 --> 00:19:00,380
And when Lawrence integrated these equations. So now we're looking at the the X component variable as a function of time.
172
00:19:01,220 --> 00:19:07,760
You can see it's, it's, it doesn't have any it doesn't you know, it, it looks very irregular.
173
00:19:07,760 --> 00:19:15,200
There are periods when it seems to be up in one state, but then it kind of jumps down or periods when it's down here and then it jumps up.
174
00:19:16,660 --> 00:19:23,980
So this is exactly what he was looking for to try to kind of prove this counterexample to the statistical model.
175
00:19:25,240 --> 00:19:31,149
But the more Lorenz looked at this, the realised he didn't really understand what was going on,
176
00:19:31,150 --> 00:19:34,450
what on earth, you know, what is causing this behaviour.
177
00:19:35,970 --> 00:19:39,840
And he had the real insight into plotting it in a different way.
178
00:19:41,550 --> 00:19:44,129
So instead of taking one variable and plotting it again,
179
00:19:44,130 --> 00:19:50,010
this time you take all three variables and imagine a three dimensional space span by those three
180
00:19:50,010 --> 00:19:57,180
variables and then let time kind of represent be represented by the length along a trajectory.
181
00:19:57,180 --> 00:20:04,230
So this represents some fairly random starting condition in this three dimensional space of X, Y and Z.
182
00:20:04,290 --> 00:20:08,430
So you give it anything you like and then let it go with his computer.
183
00:20:08,970 --> 00:20:13,230
And what he found was that the trajectory start to fall.
184
00:20:15,540 --> 00:20:22,919
Sort of strange shape. Now, the first thing is you can see it's got this these two kind of wings, if you like.
185
00:20:22,920 --> 00:20:27,540
And sometimes people call this a it looks like a butterfly, but that's entirely coincidental.
186
00:20:27,840 --> 00:20:35,219
But it has these two kind of wings which which in a sense describe those that two that kind of regime behaviour,
187
00:20:35,220 --> 00:20:38,550
it tends to be the variable tends to be up here or down here.
188
00:20:38,580 --> 00:20:44,640
Well those if you like up here, up here is, is going around here and down here goes round here.
189
00:20:45,930 --> 00:20:47,729
So that was kind of that was an interesting thing.
190
00:20:47,730 --> 00:20:55,500
But then Laurence sort of said, well, what exactly is this thing that this this trajectory is kind of oscillating on?
191
00:20:55,500 --> 00:21:01,020
What is what is the geometric object behind this? And for a while he thought maybe it's some kind of surface.
192
00:21:01,020 --> 00:21:05,850
These two lobes are lying on the surface and somehow the surface is just sort of glued together.
193
00:21:06,510 --> 00:21:13,060
But then he realised that couldn't be the case because. Then the trajectories would have to kind of cross over each other.
194
00:21:13,100 --> 00:21:19,839
It just didn't work. And it was he agonised about this for by the way, I just want to talk about that.
195
00:21:19,840 --> 00:21:27,290
I've got a. A nice animation of the of how the actual state is not going.
196
00:21:29,470 --> 00:21:32,920
They just don't work. Yeah.
197
00:21:34,380 --> 00:21:36,450
It doesn't seem to work anymore. Okay. Not to worry.
198
00:21:37,170 --> 00:21:43,530
This would have been an animation showing a the state going round in a very irregular way around what is going to go.
199
00:21:44,550 --> 00:21:48,209
Okay. And you've got different perspectives. That's now the Y in the Z.
200
00:21:48,210 --> 00:21:51,290
I think you had the. The Z and the extraction.
201
00:21:51,290 --> 00:22:00,070
That's the name of the X and Y direction. Then it'll rotate round to the. So Lawrence kept asking, So what on earth is this geometry?
202
00:22:00,080 --> 00:22:04,550
What is it? What is it? Actually, what? What how can I describe this in a kind of mathematical way?
203
00:22:06,020 --> 00:22:09,140
And that's when you realise this has to be some kind of fractal.
204
00:22:09,830 --> 00:22:15,889
Now maybe these days we're kind of blasé about fractals, we see them everywhere, but just kind of caution went back to the 1960s.
205
00:22:15,890 --> 00:22:24,680
He's dealing with equations which as I say, Newton would have had no trouble understanding and suddenly how this fractal geometry kind of comes out.
206
00:22:24,690 --> 00:22:30,470
So what I want to do is spend a few minutes trying to get this insight into why why a fractal?
207
00:22:31,670 --> 00:22:36,260
And to do that, I'm actually going to use a slightly simpler dynamical system.
208
00:22:37,430 --> 00:22:48,649
It's just easier to explain. And this actually was was was derived, as you can see, 13 years later, by wrestler in a way,
209
00:22:48,650 --> 00:22:55,220
as a way of of kind of really trying to describe this phenomenon of chaos as as simply as possible.
210
00:22:55,550 --> 00:22:59,630
And again, it's a three component system of differential equation.
211
00:22:59,640 --> 00:23:04,250
So the DOT stands for D by d t, so divided you've actually divided by two.
212
00:23:04,250 --> 00:23:10,370
You've said but slightly different equations. But again, notice the non linearity where pointer is.
213
00:23:11,790 --> 00:23:18,200
Okay. So what I want to do is take you through the rustler a tractor and try to understand why that is fractal.
214
00:23:19,280 --> 00:23:24,200
So I'm going to use some nice pictures from a book by Abraham Henshaw in 1984.
215
00:23:25,280 --> 00:23:31,920
So first let me start with so what I want to introduce are three essential components for this type of chaos.
216
00:23:33,560 --> 00:23:40,310
And the first two components are illustrated here. And this is generic to all chaotic systems.
217
00:23:41,960 --> 00:23:46,570
So imagine you take a little area here.
218
00:23:46,580 --> 00:23:50,450
This is in this space of states, state space, X, Y or Z, whatever.
219
00:23:52,130 --> 00:23:59,720
And we're going to look at, let's say, two, two points on the edge here and let them evolve in time.
220
00:24:00,440 --> 00:24:10,040
Now, what this what what these lines are illustrating is the fact that one of the essential ingredients of chaos is the notion of instability,
221
00:24:10,400 --> 00:24:18,530
that the two points that are initially close start to diverge in general exponentially from each other.
222
00:24:19,560 --> 00:24:27,360
And that's kind of illustrated, as I say, here. So this this point goes off in this direction, whereas this one goes off in this direction.
223
00:24:27,360 --> 00:24:34,380
So this direction here is a kind of direction of instability where initially close trajectories are diverging apart.
224
00:24:34,410 --> 00:24:42,750
So that's one key ingredient. The second key ingredient is actually looking at this in the sort of transverse direction.
225
00:24:44,380 --> 00:24:47,920
So looking at points, let's say, which start on the other edge of the square,
226
00:24:48,520 --> 00:24:53,740
because the second feature of chaos is that these or this type of chaos at least,
227
00:24:54,100 --> 00:24:58,420
is that these trajectories in this direction are actually converging together.
228
00:24:59,680 --> 00:25:04,120
So they're getting closer. Now, the key point is that overall.
229
00:25:05,410 --> 00:25:08,410
Average over the whole system in all time.
230
00:25:08,830 --> 00:25:14,110
It's actually the converging direct directions win over the expanding directions.
231
00:25:15,090 --> 00:25:18,570
So that if I was to measure the area of this.
232
00:25:19,930 --> 00:25:26,640
So here's if I take this area at initial as imagine this is at some initial time and then that evolves into an area.
233
00:25:27,000 --> 00:25:29,910
Imagine looking at this object from the far side.
234
00:25:30,330 --> 00:25:37,950
Then there'd be an area here which was stretched out in this direction but compressed in the transverse direction.
235
00:25:38,340 --> 00:25:43,790
Then on average, the area will be smaller afterwards than it was before.
236
00:25:43,800 --> 00:25:51,110
So this overall shrinkage of area. So these are two key ingredients for chaos.
237
00:25:51,980 --> 00:25:59,030
So this just shows the same thing again, except they've just drawn or they've drawn this the transverse,
238
00:25:59,030 --> 00:26:03,230
if you like, direction as a, as a rather small, um, just much smaller.
239
00:26:03,440 --> 00:26:06,410
Okay. So that's fine. Now we have a problem.
240
00:26:06,530 --> 00:26:12,679
I mean, if, as I say, these if you just imagine exponentials of divergence, then this would just carry on.
241
00:26:12,680 --> 00:26:17,600
They would diverge as far apart as you like, go off to infinity, as it were.
242
00:26:17,990 --> 00:26:26,600
So this can't possibly explain the fact that we have a a kind of a geometry which seems to sit in a in a finite, compact region of state space.
243
00:26:27,200 --> 00:26:30,320
So how do we stop the trajectories going off to infinity?
244
00:26:31,310 --> 00:26:34,880
Well, this is where the third ingredient of nonlinearity comes in.
245
00:26:35,870 --> 00:26:40,999
And in the case of the Russell Attractor, it's rather simple. The nonlinearity just folds one of this.
246
00:26:41,000 --> 00:26:45,450
The surface over like this. So and so.
247
00:26:45,530 --> 00:26:53,390
Just take one hand and just fold it over. So let's have a look at that in a bit more detail.
248
00:26:53,630 --> 00:26:59,160
So let's let's take our. Let's talk a little surface.
249
00:26:59,700 --> 00:27:03,150
Here are two points, two trajectories.
250
00:27:04,290 --> 00:27:10,230
And they kind of evolve and the system becomes nonlinear and the surface folds over like this.
251
00:27:10,740 --> 00:27:17,280
So that one trajectory has gone to the top part and the other trajectory to the to the bottom part there.
252
00:27:17,280 --> 00:27:19,140
And it's kind of folded over in half.
253
00:27:21,230 --> 00:27:31,430
Now, if you remember what I said, if you try to let's try to kind of compress that down into the same cross-sectional area as we started with.
254
00:27:32,740 --> 00:27:41,110
Then what I said, if you if you recall, is that this is the combined area or the total area of this top part.
255
00:27:41,500 --> 00:27:44,590
Plus the bottom part is actually less.
256
00:27:44,920 --> 00:27:48,840
It's smaller because of the overall effect of this contracting this.
257
00:27:48,970 --> 00:27:56,140
This is the effect of dissipation, effectively. This total area is smaller than the the one I started with.
258
00:27:56,710 --> 00:28:05,650
So if I try to compress the whole thing into the same area, there must be a gap between these the top sheet, if you like, and the bottom sheet.
259
00:28:06,010 --> 00:28:10,060
This is a really crucially important point. There's a gap. There has to be a gap.
260
00:28:10,480 --> 00:28:14,320
They don't kind of completely because you've got some compression effectively.
261
00:28:14,530 --> 00:28:17,720
So. All right. So let's take this thing again.
262
00:28:18,670 --> 00:28:22,960
You can imagine this is just like here, but with the gap. But I've sort of blown this up a bit.
263
00:28:23,980 --> 00:28:29,290
And so this kind of comes round as two sheets and we'll go we'll go through the whole exercise again.
264
00:28:30,430 --> 00:28:34,960
Fold it over. And now you see what was two sheets have become force sheets.
265
00:28:35,830 --> 00:28:39,310
Now this gap.
266
00:28:40,410 --> 00:28:45,390
Has now kind of formed two small gaps, if you like, here and here.
267
00:28:45,690 --> 00:28:51,030
And a bigger gap has appeared in between them for the same reason that I said before.
268
00:28:52,400 --> 00:28:59,540
And now we do it again. We got now these four sheets we folded over and now we've got eight sheets with a big gap,
269
00:28:59,960 --> 00:29:05,990
two small gaps and four really small gaps, which was the original gap done, done twice over.
270
00:29:07,160 --> 00:29:13,010
Now, if we keep doing this over and over and over and over and over and over and over and over, what are we going to end up with?
271
00:29:14,620 --> 00:29:23,380
Okay. Well, one way to to try to understand that is to take a cross-section through this this structure here.
272
00:29:23,770 --> 00:29:30,759
After, you know, a very large number of of these types of iterations, this is sometimes actually called a Lorenz section.
273
00:29:30,760 --> 00:29:36,790
The whole thing is sometimes called a Poincaré section. And this kind of transverse through the trajectory is called a Lorenz section.
274
00:29:37,930 --> 00:29:44,889
So what is this? Well, this was this structure was discovered some years earlier, in fact,
275
00:29:44,890 --> 00:29:51,070
in the late 19th century, early 20th century by the great mathematician George Cantor.
276
00:29:51,880 --> 00:29:58,450
And his reason for for this he invented this set called the Cantor set was really nothing to do with chaos per say,
277
00:29:58,450 --> 00:30:02,390
but really to try to understand a bit more about the nature of numbers.
278
00:30:02,550 --> 00:30:10,570
And this is an interesting set is an interesting example of a of a set which has an uncountable infinite number of points.
279
00:30:10,570 --> 00:30:14,920
But but the points actually take up no volume at all.
280
00:30:14,920 --> 00:30:18,130
No volume at all in the on the on the line.
281
00:30:19,060 --> 00:30:29,680
So the Cantor set is is constructed by starting with just, let's say, the line between zero and one and throw away the middle third.
282
00:30:30,130 --> 00:30:34,690
So you got two pieces and then throw away the middle third of the remaining two pieces
283
00:30:35,080 --> 00:30:38,440
and then throw away the middle third of the remaining four pieces and continue,
284
00:30:38,440 --> 00:30:46,000
continue. And then just take the intersection of all of these iterates and you're left with this set.
285
00:30:46,360 --> 00:30:51,000
Now, so this big the big gap. Is is if you like.
286
00:30:51,300 --> 00:30:57,300
Is this part of the of the rustler a attractor or it's a gap in the rustler tractor.
287
00:30:57,660 --> 00:31:03,000
And then the next iterate are these two smaller gaps and so on and so forth.
288
00:31:03,030 --> 00:31:07,559
So we have this. So this is where fractals basically come from.
289
00:31:07,560 --> 00:31:11,940
The three ingredients, instability, dissipation, non linearity.
290
00:31:14,220 --> 00:31:22,740
And this is what Lawrence said in his 63 paper. We see that each surface is really a pair of surfaces, so that where they're they appear to merge.
291
00:31:24,630 --> 00:31:25,860
They're really four surfaces.
292
00:31:25,860 --> 00:31:32,159
Continuing this process for another circuit, we see they're eight processes and we finally conclude there's an infinite complex of surfaces,
293
00:31:32,160 --> 00:31:38,160
each extremely close to one another or other, the two merging surfaces, I think.
294
00:31:38,190 --> 00:31:41,850
Ian Stewart in his popular science book, Does God Play Dice?
295
00:31:42,270 --> 00:31:49,979
Absolutely Nails the reaction, I think of a lot of mathematicians when they read this paper and he said,
296
00:31:49,980 --> 00:31:54,780
when I read Lawrence's words, I get a prickling at the back of my neck and my hair stands on end.
297
00:31:54,810 --> 00:32:00,970
He knew 34 years ago. He knew. And when I look more closely, I'm even more impressed in a mere 12 pages.
298
00:32:00,990 --> 00:32:05,370
Lawrence anticipated several major ideas of nonlinear dynamics before it became fashionable.
299
00:32:06,930 --> 00:32:15,360
So if I had to say, you know, one thing that Lawrence for me absolutely characterises the genius of Lawrence,
300
00:32:15,990 --> 00:32:20,430
it's actually not so much the, you know, this divergence of trajectories,
301
00:32:20,730 --> 00:32:27,540
but the realisation of this fractal geometry, which underpins these sets of differential equations,
302
00:32:27,540 --> 00:32:31,290
which by all accounts should look very smooth and unexceptional.
303
00:32:31,470 --> 00:32:35,629
Unexceptional. So with that in mind,
304
00:32:35,630 --> 00:32:43,280
I am going to make the claim that in some sense Lawrence provides a bridge really between the classical
305
00:32:43,850 --> 00:32:51,930
physics and classical mathematics of Newton and some of the real key ideas in 20th century math.
306
00:32:51,950 --> 00:32:57,860
So we've already mentioned Cantor. This is Kurt Gödel and this is Alan Turing.
307
00:32:58,310 --> 00:33:10,400
And I'm sure people know that some. They really revolutionised our ideas in mathematics by bi by raising the notion that there are there
308
00:33:10,400 --> 00:33:18,620
are things which are true in mathematics but not provably true or in Turing's language propositions,
309
00:33:18,620 --> 00:33:24,620
which are kind of undecidable algorithmically, no matter how complex or your computation might be,
310
00:33:24,980 --> 00:33:29,090
you just cannot establish the truth of of certain propositions.
311
00:33:29,600 --> 00:33:37,549
And it turns out that many of the properties of these fractal geometries are quite isa morphic,
312
00:33:37,550 --> 00:33:46,010
quite similar to the types of propositions of good Alan Turing that are undecidable.
313
00:33:47,140 --> 00:33:50,920
In fact, the papers written on this showing this in a formal way.
314
00:33:51,850 --> 00:33:56,379
So there is this bridge between, as I say, the calculus of Newton, which, as I say,
315
00:33:56,380 --> 00:33:59,530
Newton would look at these equations, say, yes, I take that, I understand that.
316
00:34:00,880 --> 00:34:08,680
And but through this remarkable geometry that they generate, we have these links to some of the most profound things in in mathematics.
317
00:34:08,860 --> 00:34:14,319
Now, I've included Andrew Wiles here because there's lots of other interesting things you can ask.
318
00:34:14,320 --> 00:34:23,530
For example, suppose I want to have a type of arithmetic where I add numbers on on this or on this fractal set, on this cantor set,
319
00:34:23,530 --> 00:34:28,359
let's say on this transverse cantor set of multiply numbers such that the sum
320
00:34:28,360 --> 00:34:32,350
of the numbers or the product of the numbers also lies on the cantor set.
321
00:34:32,440 --> 00:34:41,710
Now, can we do arithmetic on the Cantor sets? It turns out you need a special type of number system to do that called padding numbers.
322
00:34:42,670 --> 00:34:50,230
So I think numbers are a very rigorous branch of pure mathematics that has a very close link to fractal geometry.
323
00:34:50,950 --> 00:34:58,300
And as I say, they they respect the sort of geometric constraints of fractals in doing algebra.
324
00:34:59,140 --> 00:35:05,020
And it turns out that a lot of the deep theory a number theorems from poetic numbers
325
00:35:05,020 --> 00:35:11,380
systems were very important in Andrew's celebrated proof of Fermat's Last Theorem.
326
00:35:13,350 --> 00:35:24,959
One of the things which really intrigues me as a physicist is whether there are links between these ideas and quantum physics.
327
00:35:24,960 --> 00:35:35,280
So people may recognise Erwin Schrödinger and Heisenberg, Paul Dirac and another very celebrated Oxford mathematician, Roger Penrose.
328
00:35:37,680 --> 00:35:45,540
Suddenly three of these people, Schrödinger, Dirac and Penrose, I'm not so sure about Heisenberg,
329
00:35:45,540 --> 00:35:49,630
but the three of them were very uncomfortable with quantum theory.
330
00:35:49,650 --> 00:35:56,100
It works and it does work to this day, but the foundations are very the why this keeps jumping.
331
00:35:56,550 --> 00:36:00,510
The foundations are very kind of bizarre and difficult to understand,
332
00:36:00,510 --> 00:36:09,630
and I think all three of them felt that there must be something deeper underpinning quantum theory, something much more deterministic.
333
00:36:10,170 --> 00:36:16,379
But Roger Penrose in his books like The Emperor's New Mind, which I'm sure some of you will have read very much,
334
00:36:16,380 --> 00:36:23,670
emphasised that if we if there is some deterministic underpinning to quantum theory, it has to have some notion of non compute ability.
335
00:36:23,670 --> 00:36:27,210
This undesired ability has to feature somewhere deeply.
336
00:36:27,510 --> 00:36:31,710
Just an ordinary type of deterministic system won't do it.
337
00:36:32,400 --> 00:36:35,790
So the reason I've included Roger here is because he has very much stressed the
338
00:36:35,790 --> 00:36:41,250
potential role of non compute ability in going deeper into fundamental physics.
339
00:36:41,640 --> 00:36:46,170
So again, perhaps these types of fractal objects play a role.
340
00:36:46,170 --> 00:36:54,270
I personally believe they do, but one has to say at the moment it's a matter for uh, for study and so on.
341
00:36:54,270 --> 00:36:59,910
So I'm not going to dwell upon this any more in the talk, but I'm happy to talk to people afterwards if they're interested.
342
00:37:01,200 --> 00:37:11,400
Chaos theory, for sure, has revolutionised many different branches of science in physics, in biology, in engineering, in economics and social science.
343
00:37:11,550 --> 00:37:17,430
It's hard to think of any area where it's left untouched. Unfortunately, it's sometimes also abused.
344
00:37:18,910 --> 00:37:22,270
So I want to give an example of an abuse of chaos theory.
345
00:37:22,270 --> 00:37:27,630
And this is actually something that's close to my heart because I have to deal with these sorts of people quite a lot.
346
00:37:27,680 --> 00:37:31,500
The climate, well, they don't like to be called deniers, climate sceptics.
347
00:37:32,830 --> 00:37:38,460
Now, one argument which is quite common actually goes like this You guys, okay,
348
00:37:38,470 --> 00:37:41,740
you might be able to forecast the weather tomorrow, but you're pretty hopeless.
349
00:37:42,160 --> 00:37:47,980
And even then, you sometimes get it wrong. And, you know, a couple of weeks ahead is pretty dodgy.
350
00:37:48,400 --> 00:37:52,120
A year ahead, no chance. So what wasn't on it?
351
00:37:52,130 --> 00:37:56,680
Why on earth should I believe what you say about the climate 100 years from now?
352
00:37:57,790 --> 00:38:04,119
So this is this was this was brought up not so long ago in a review of a book in The Telegraph.
353
00:38:04,120 --> 00:38:12,640
Daily Telegraph. The game's up for climate change believers, the theory of global warming as a gigantic weather forecast,
354
00:38:12,970 --> 00:38:17,440
and therefore it can have a century or more and therefore can have no value as a prediction.
355
00:38:17,830 --> 00:38:21,190
Okay. So I just want to debunk that very briefly and then we'll move on.
356
00:38:22,120 --> 00:38:26,770
Because basically that is to misunderstand the problem of climate change.
357
00:38:28,960 --> 00:38:36,280
What I've done here is to do a kind of climate change experiment in Lawrence world by adding an extra term to one of the equations.
358
00:38:37,480 --> 00:38:40,630
Now, just think of this term here. This is actually just going to be a number.
359
00:38:40,990 --> 00:38:47,350
But just think of this as a kind of a surrogate for doubling carbon dioxide concentrations in the atmosphere.
360
00:38:47,350 --> 00:38:52,120
So I'm a sort of applying an extra forcing term, if you like, to the equations.
361
00:38:52,510 --> 00:38:57,669
So how does that affect the Lorenz the Lorenz system?
362
00:38:57,670 --> 00:39:08,709
So here's again the equation X component as a function of time before the system would oscillate roughly between these two kind of regimes,
363
00:39:08,710 --> 00:39:11,980
the two lobes of the butterfly with equal sort of frequency.
364
00:39:11,980 --> 00:39:16,780
So it is likely to be up here in one lobe of the butterfly, as it would be to be down here.
365
00:39:17,680 --> 00:39:22,480
What's happened with this additional term is that I've increase the likelihood of it
366
00:39:22,510 --> 00:39:26,650
being in the upper lobe and decrease the likelihood of it being in this lower lobe.
367
00:39:27,370 --> 00:39:32,979
Now, the system is still chaotic because if I want to predict in detail at one of the trajectories,
368
00:39:32,980 --> 00:39:38,500
if I want to make a weather forecast, if you like, in this Lorenz world, it's still going to be predictable.
369
00:39:38,500 --> 00:39:42,340
It's still going to be subject to uncertainties in the initial conditions.
370
00:39:42,640 --> 00:39:47,980
But what is predictable is this kind of gross statistic that the probability.
371
00:39:49,010 --> 00:39:53,270
Of the state being up here has been affected in a very predictable way.
372
00:39:53,750 --> 00:40:00,530
In fact, we can you know, we can actually plot this on in the Lawrence attractor world in the state space.
373
00:40:00,950 --> 00:40:08,179
So here's the original Lawrence Attractor in the new system, these two kind of what we call centroid.
374
00:40:08,180 --> 00:40:10,790
So right in pretty much the same place as they were before.
375
00:40:11,090 --> 00:40:16,550
But now the system spends a lot more time zipping around here than it does zipping around here.
376
00:40:18,530 --> 00:40:23,060
So we could view in this perspective, we could view climate change as a problem in geometry.
377
00:40:24,100 --> 00:40:33,640
The question is how is the geometry of this attractor being affected by the addition of this term, which I've now set equal to ten?
378
00:40:35,470 --> 00:40:39,590
So it's important to understand this. So climate change, this is the problem of climate change.
379
00:40:39,610 --> 00:40:48,099
How does the whole climate attractor it, which basically means the statistics of weather change as we double carbon dioxide levels,
380
00:40:48,100 --> 00:40:53,890
which we surely will be sometime later this century or so over pre-industrial.
381
00:40:54,640 --> 00:40:58,210
How how how does that change the statistics of weather?
382
00:40:58,630 --> 00:41:02,230
And that's not the same problem at all as completely different class of problem to
383
00:41:02,230 --> 00:41:07,780
saying how does a particular trajectory evolve on its on the climate attractor.
384
00:41:08,770 --> 00:41:12,820
Over some from some initial condition. Okay.
385
00:41:13,060 --> 00:41:24,970
Uh, let me just now move on to the sort of part of the talk I wanted to, uh, I wanted to get to, which is what did Laurence actually mean?
386
00:41:27,430 --> 00:41:31,540
By the butterfly effect. And you'll see that, in fact, he wasn't talking about his 63 paper at all.
387
00:41:33,900 --> 00:41:37,610
He's talking about the fact that the weather is a multi scale system.
388
00:41:37,620 --> 00:41:48,299
So we have we have a you know, if you imagine a large low pressure system trundling across the Atlantic embedded in that low pressure system,
389
00:41:48,300 --> 00:41:54,030
there'll be, let's say, you know, thunderstorm clouds, big thunderstorm clouds where they match the some of the storm clouds,
390
00:41:54,210 --> 00:41:58,560
maybe 100 kilometres in scale, whereas the weather system itself is a thousand kilometres in scale.
391
00:42:00,780 --> 00:42:05,819
Embedded in that thunderstorm cloud, there are small sub clouds, turbulent eddies,
392
00:42:05,820 --> 00:42:10,139
and within those sub cloud, turbulent eddies there, yet smaller turbulence eddies.
393
00:42:10,140 --> 00:42:14,310
So we're looking at this this Russian doll hierarchy again.
394
00:42:16,520 --> 00:42:22,489
Now, Laurence said, okay, let's suppose our problem is to predict this weather pattern, the large scale weather pattern.
395
00:42:22,490 --> 00:42:23,630
That's what we're interested in.
396
00:42:25,210 --> 00:42:31,330
We might be able to measure, you know, the initial the starting conditions for this weather pattern very, very accurately.
397
00:42:32,710 --> 00:42:38,770
In fact, we can imagine almost perfectly for the sake of argument, measuring the initial conditions for that weather scale.
398
00:42:39,250 --> 00:42:44,770
But that's not going to give us indefinite predictability because sooner or later the fact that
399
00:42:44,770 --> 00:42:51,520
we haven't been able to measure the the cloud scales or the sub cloud scale motions perfectly,
400
00:42:51,520 --> 00:42:55,989
that's going to catch up with us because these areas are going to kind of propagate up non-linear,
401
00:42:55,990 --> 00:42:59,260
only from the small scale to the larger scale to the very large scale.
402
00:43:01,210 --> 00:43:06,550
If you want to actually read this, he's Lawrence himself wrote a popular book called The Essence of Chaos.
403
00:43:06,550 --> 00:43:14,410
And in the appendix, he actually describes what went into this talk in the triple OAS meeting in 1972,
404
00:43:15,130 --> 00:43:18,340
and he talks about errors in the course of scrapped structure.
405
00:43:18,350 --> 00:43:23,980
So that's a large scale weather pattern and then errors in these final structures, which are the clouds.
406
00:43:24,250 --> 00:43:29,620
And then in fact, the fact that errors in the final structure can start to induce errors in the course of structure.
407
00:43:31,170 --> 00:43:35,639
What I've decided to do for this talk is actually not to go through in detail with what he wrote,
408
00:43:35,640 --> 00:43:43,320
but return to my Russian doll example and try to describe what he what his thinking was with that.
409
00:43:44,810 --> 00:43:49,879
So let's imagine the Russian doll, the big Russian doll is this big low pressure system trundling across the Atlantic.
410
00:43:49,880 --> 00:43:53,600
And that's the thing we want to predict. We want to predict that as far ahead as we possibly can.
411
00:43:54,620 --> 00:44:01,280
And let's imagine we've got some observing, some instrument which has no error at all.
412
00:44:01,310 --> 00:44:10,110
I mean, infinitesimal, just quantum mechanics, just for the sake of argument, measures perfectly measures whatever its wants to do.
413
00:44:10,130 --> 00:44:17,810
Temperature, pressure and stuff like that. Now we have a whole load of these instruments and we're going to adopt them around the whole globe.
414
00:44:17,840 --> 00:44:24,980
We've got enough of them. We can dump them around the whole globe in a regular network where the distance between the two,
415
00:44:25,040 --> 00:44:28,970
between any two of these measuring systems is sufficiently,
416
00:44:29,840 --> 00:44:36,530
let's say, small, that it can it can resolve, you know, this this low pressure system really, really well.
417
00:44:38,470 --> 00:44:46,990
So that the initial conditions for this low for the scale, this low pressure system, the scale of this low pressure system are known almost perfectly.
418
00:44:48,220 --> 00:44:57,640
But the distance between these measuring objects is not sufficiently small that I can measure the initial conditions for all these smaller eddies,
419
00:44:57,640 --> 00:45:03,209
the smaller wells of Richardson. So I'm going to imagine that with that information.
420
00:45:03,210 --> 00:45:08,390
Let's say I can predict ahead five days. I mean, it's a reasonable sort of number.
421
00:45:09,740 --> 00:45:15,320
So what Lawrence said was, well, maybe we're not happy with that. How would we extend that predictions?
422
00:45:15,500 --> 00:45:17,390
Horizon We want to predict further.
423
00:45:18,020 --> 00:45:23,720
Well, what's limiting us in this case is not our ability to measure this scale, because we've measured this perfectly.
424
00:45:24,170 --> 00:45:28,250
What's limiting it is the fact that we don't have any information about this scale.
425
00:45:28,970 --> 00:45:34,760
So we'll stick a whole load more of these instruments down, filling the gaps between the ones we had.
426
00:45:35,330 --> 00:45:40,370
So now we have enough resolution that we can measure the initial conditions of this scale here.
427
00:45:40,940 --> 00:45:46,280
So the question is, how much is that going to bias? How much extra prediction skill is that going to give us?
428
00:45:47,360 --> 00:45:52,280
Now, Lawrence says the answer is not ten days, because these these things,
429
00:45:52,730 --> 00:45:56,960
the errors are likely to be growing faster for these smaller scale features.
430
00:45:57,620 --> 00:46:00,019
So, for example, you mentioned the cloud system, the error,
431
00:46:00,020 --> 00:46:05,660
the time it takes because the circulation patterns are more vigorous in a in a in an individual cloud.
432
00:46:06,110 --> 00:46:11,329
If you have a flown through a cloud in an aeroplane, you know, speaking of thunderstorm cloud, there's a very vigorous circulation.
433
00:46:11,330 --> 00:46:19,080
So air is a growing much more rapidly. So let's I'm kind of simplifying the for the experts.
434
00:46:19,090 --> 00:46:24,960
I'm simplifying things a little bit just to make the point. So but nevertheless, the idea is, is qualitatively correct.
435
00:46:24,990 --> 00:46:31,080
So let's imagine for this second, you only get two and a half days of extra prediction skill.
436
00:46:31,090 --> 00:46:39,000
So you've gone up now from five days to seven one half days. So Lauretta says, okay, that's fine, but let's, let's say we want to go further.
437
00:46:39,010 --> 00:46:45,930
So let's, let's have a whole load. More measurement systems are now fill in even the gaps in those gaps.
438
00:46:46,260 --> 00:46:52,470
So then now we can even measure this Russian doll initially perfectly and he says, okay,
439
00:46:52,980 --> 00:46:58,350
again, we'll only because these are growing yet faster, the errors are growing faster.
440
00:46:58,770 --> 00:47:03,240
We may only get an extra C I'm I'm halving the extra time each step.
441
00:47:03,720 --> 00:47:12,330
So now we only get an extra day and a quarter. So you can see this continual investment in imperfect measurement measurement systems,
442
00:47:13,200 --> 00:47:20,250
which is driving the initial error closer and closer to zero, is actually buying us less and less prediction time.
443
00:47:21,870 --> 00:47:27,839
And what Lawrence says. Let's take this to the limit and imagine we've got an infinite number of measuring
444
00:47:27,840 --> 00:47:32,490
systems which can basically take us down to infinitesimally small scales.
445
00:47:33,270 --> 00:47:43,710
He's saying that what this will bias in terms of prediction time is a series which if with these particular choices of doubling times,
446
00:47:44,100 --> 00:47:47,130
will actually converge to a finite number.
447
00:47:47,940 --> 00:47:52,650
Now, normally we have when we have series, we think of convergence as a good thing and divergence is a bad thing.
448
00:47:52,950 --> 00:48:02,040
But here, convergence is a bad thing because it's really limiting your prediction capability to some finite time,
449
00:48:02,040 --> 00:48:05,670
which in this simple calculation turns out to be ten days.
450
00:48:06,570 --> 00:48:10,500
So this has a remarkable. If you think about it, this is a remarkable thing.
451
00:48:10,500 --> 00:48:19,350
If it's correct, it says. No matter how small my initial error is, I will never be able to predict more than ten days ahead.
452
00:48:20,490 --> 00:48:29,370
Now, just think that's completely different to the 1963 paper, because the 63 paper is about just these three equations,
453
00:48:29,370 --> 00:48:36,030
which, although it's very difficult to predict them, you can always in principle predict as far as ahead as you like.
454
00:48:36,330 --> 00:48:41,740
Providing your initial error is sufficiently small. But here's a system where that's not true.
455
00:48:43,600 --> 00:48:51,010
And this is what it's about in his 69 paper, which appeared in a Swedish journal called Tell US.
456
00:48:51,100 --> 00:48:59,260
And if his 63 paper was, you know, is considered to be in an obscure journal, this is obscure squared, I'm afraid.
457
00:48:59,320 --> 00:49:02,610
I'm afraid so. Very, very not.
458
00:49:02,620 --> 00:49:10,509
Not really. Well read at all. But it's worth just if you don't mind just reading, I'll read out the the part of the abstract.
459
00:49:10,510 --> 00:49:15,190
So it is because it's really important this he claims it is he proposed that it is proposed
460
00:49:15,190 --> 00:49:19,630
that certain formally deterministic fluid systems which possess many scales of motion.
461
00:49:19,670 --> 00:49:26,140
That's the key point. Many scales of motion, not the three rationales, but a kind of essentially an infinity of Russian goals,
462
00:49:27,040 --> 00:49:30,370
are observationally indistinguishable for in deterministic systems.
463
00:49:30,730 --> 00:49:35,140
Specifically, that two states of the system, differing initially by a small observational error,
464
00:49:35,650 --> 00:49:42,129
will evolve into two states differing as greatly as randomly chosen states of the system with a within a finite time interval.
465
00:49:42,130 --> 00:49:50,260
In our calculation, it was ten days. Now, here's the real Ryder kicker, which cannot be lengthened by reducing the amplitude of the initial error.
466
00:49:51,400 --> 00:49:59,020
So this is very radical. So what I'm proposing to do here today is to illustrate so as to introduce you to two butterflies.
467
00:49:59,340 --> 00:50:04,450
Now it's we're never going to stop the butterfly effect being used to describe low order chaos.
468
00:50:04,900 --> 00:50:10,920
But I'm going to call this the common butterfly effect. So here's a common blue on the common butterfly.
469
00:50:11,090 --> 00:50:18,250
This is the one that everybody thinks of when they talk about the butterfly effect is about sensitive dependence on initial conditions.
470
00:50:18,820 --> 00:50:23,920
So that means it's certainly difficult to predict the future, but it's not impossible in principle.
471
00:50:23,920 --> 00:50:28,930
You can predict in something like a Lorenz 63 system, you can predict as far ahead as you like.
472
00:50:29,200 --> 00:50:36,450
Providing the initial error is small enough. On the other hand, I'm going to use this monarch butterfly.
473
00:50:36,460 --> 00:50:43,960
So this is a slight, a weak pun on the fact that this is monarch is a royal member of royalty and royal reality in Spanish or something is royal.
474
00:50:43,970 --> 00:50:47,050
So the real butterfly effect, the royal butterfly effect, if you like,
475
00:50:47,770 --> 00:50:52,419
is the 69 paper which basically says that there are finite predictability
476
00:50:52,420 --> 00:50:56,320
horizons which cannot be extended by reducing uncertainty in initial conditions.
477
00:50:58,400 --> 00:51:02,600
So we know that this is this has been certainly well verified.
478
00:51:02,600 --> 00:51:07,130
So the common butterfly effect is been is kind of absolutely part of standard science.
479
00:51:07,370 --> 00:51:12,500
How about this real butterfly effect? Is this part of a of a kind of rigorous science?
480
00:51:14,180 --> 00:51:18,829
Let's just put this into a slightly more mathematical phraseology,
481
00:51:18,830 --> 00:51:28,400
because what what we're saying here is if you take the initial conditions and change them as slightly as you like in a finite time,
482
00:51:28,430 --> 00:51:31,850
you'll you'll diverge, too, to find actually different solutions.
483
00:51:33,800 --> 00:51:42,680
Now in math, in mathematics, problems are referred to as well posed or ill posed according to certain properties that they have,
484
00:51:43,520 --> 00:51:46,010
apparently from a definition given originally by had them.
485
00:51:47,450 --> 00:51:54,140
So problems are well posed if solutions exist, they're unique and the solutions depend continuously on the data.
486
00:51:54,170 --> 00:51:55,790
This means initial data in our case.
487
00:51:56,180 --> 00:52:04,730
Now, this continuously means if you change the initial conditions very, very slightly, you're just going to change the solution slightly.
488
00:52:05,480 --> 00:52:14,840
And that's true in the 63 chaos, but it's not true in the 69 case, in the 69 real chaos, real butterfly effect.
489
00:52:15,920 --> 00:52:22,190
So if this effect is really true, it means that the navier-stokes initial value problem is not well posed in this sense.
490
00:52:22,850 --> 00:52:28,880
So now this has becomes a problem in mathematics. It's the navier-stokes problem for three dimensional, turbulent fluid.
491
00:52:29,900 --> 00:52:31,670
The initial value problem well posed.
492
00:52:33,170 --> 00:52:39,200
And this is why I say this is actually turns out to be one of the great unsolved problems in 21st century mathematics.
493
00:52:39,620 --> 00:52:49,010
So as I'm sure many of you know, the Clay Mathematics Institute put out these kind of key problems, unsolved problems in mathematics,
494
00:52:50,420 --> 00:52:55,580
which include famous things like this is the beginning of this millennium, the Riemann hypothesis and so on.
495
00:52:55,610 --> 00:52:57,590
And here's the the Navier-stokes one.
496
00:52:57,890 --> 00:53:04,219
It's actually not framed in terms of this continuity property, but rather whether solutions exist and are unique.
497
00:53:04,220 --> 00:53:09,290
And even that's not known. And we need to know that even before we can understand this property of continuity.
498
00:53:10,250 --> 00:53:15,680
So I think it's fascinating to me that what Lawrence speculated about in 69 and
499
00:53:15,680 --> 00:53:21,110
then this triple as talk in 72 actually refers to a very deep problem in maths.
500
00:53:21,770 --> 00:53:31,400
So if I got just a few minutes to finish, okay, so I want to kind of get back to two more practical issues because a question you might ask is,
501
00:53:31,640 --> 00:53:36,350
is it really the case then that we're limited in weather prediction to ten days
502
00:53:36,350 --> 00:53:40,880
or something and things aren't quite as simple as that in practice because.
503
00:53:43,210 --> 00:53:50,020
Ten days may well be sort of roughly an average timescale that we can make sort of detailed predictions about the weather.
504
00:53:50,380 --> 00:53:56,350
But what we rapidly find out is that there are many situations where we can make much longer range predictions,
505
00:53:56,350 --> 00:54:03,790
apparently with with quite good skill, but equally other situations where predictions do in fact go wrong within a couple of days.
506
00:54:05,200 --> 00:54:14,010
And that's very well illustrated going back to the 63 paper and actually looking at how errors, if you like, or uncertainty.
507
00:54:14,020 --> 00:54:19,360
So imagine this is an initial state with some kind of ball of uncertainty associated with it.
508
00:54:20,140 --> 00:54:28,660
This actually shows that there are some parts of the Lorenz attractor that are extremely stable, so small perturbations hardly grow at all.
509
00:54:28,930 --> 00:54:33,730
So this is telling you there are some parts of the system which are actually very predictable,
510
00:54:34,450 --> 00:54:40,450
other parts of the situation, other parts of the attractor, where you start to explore the state space down here,
511
00:54:40,720 --> 00:54:47,290
where you do start to see unpredictability and some initial conditions where the uncertainty just explodes.
512
00:54:49,720 --> 00:54:53,980
Now, those of you, most of you in the audience are not old enough.
513
00:54:53,980 --> 00:54:57,129
But I know there are a few people to remember this.
514
00:54:57,130 --> 00:55:06,590
But one of the most iconic weather events of the last maybe 300 years occurred almost 30 years ago, October 1987.
515
00:55:07,270 --> 00:55:14,620
And this poor guy whose name is Michael Fish, although he's I have to say, he makes a great after dinner living out of speech,
516
00:55:14,620 --> 00:55:20,859
living from this famously made probably the worst weather forecast in history by
517
00:55:20,860 --> 00:55:25,540
saying there's no chance of a hurricane hitting the UK in the next a day or so.
518
00:55:25,900 --> 00:55:39,860
And indeed, a hurricane did did hit. And this is actually a great example of this very intermittent phenomenon of explosive unpredictability.
519
00:55:39,890 --> 00:55:49,440
So what I'm going to do here is just animate two weather forecasts from two almost identical initial conditions.
520
00:55:49,460 --> 00:55:52,880
These are these are surface pressure maps, almost identical.
521
00:55:52,890 --> 00:56:00,280
Not quite identical. You can look in detail. You'll see differences. With just a couple of days before the storm hit.
522
00:56:00,400 --> 00:56:08,560
So just run this forward and they start off tracking each other pretty well.
523
00:56:08,830 --> 00:56:17,170
But then leading up to the fateful morning of the 16th, I think it was, you see that these are really quite different.
524
00:56:17,440 --> 00:56:23,350
This would be a very kind of benign day, and this is a really intense vortex.
525
00:56:24,310 --> 00:56:31,000
Poor Michael Fish was given essentially this solution by the Met Office, but for a flap of a butterfly's wing,
526
00:56:31,240 --> 00:56:35,910
he would have been given this one and he would then have been a national hero instead.
527
00:56:36,880 --> 00:56:41,500
Actually, you shouldn't feel too bad about this because the Met Office did really well out of this poor forecast.
528
00:56:41,500 --> 00:56:48,100
They said, Oh, well, our computers aren't big enough, you need to invest lots more money. And so I don't feel sorry for them.
529
00:56:48,790 --> 00:56:53,319
Okay. So the question is, how would a modern day Michael Fish deal with this situation?
530
00:56:53,320 --> 00:56:58,059
Because chaos is chaos. I mean, chaos in 1987 will still happen again.
531
00:56:58,060 --> 00:57:01,930
So this intermittent explosive unpredictability could still happen in 2017.
532
00:57:02,560 --> 00:57:06,010
So how would a modern day Michael Fish deal with this situation?
533
00:57:06,610 --> 00:57:11,440
Well, these days, with much bigger computers and in particular with massively parallel computers,
534
00:57:12,040 --> 00:57:18,550
we can take advantage of that type of architecture to run weather forecasts in what I call ensemble mode.
535
00:57:19,090 --> 00:57:25,180
So we actually we actually run here 50 different forecasts.
536
00:57:25,180 --> 00:57:28,180
This would be done typically every day at places like the Met Office.
537
00:57:28,660 --> 00:57:34,479
Some are actually everywhere around the globe. Now, you don't run a single weather forecast, you run an ensemble of them and very,
538
00:57:34,480 --> 00:57:42,280
very slightly the initial conditions and look to see ahead of time whether the forecasts are diverging and what you see in this case.
539
00:57:42,280 --> 00:57:45,669
So this is the 87 storm run retrospectively.
540
00:57:45,670 --> 00:57:50,950
It's run with a modern day ensemble system, modern day weather forecast model.
541
00:57:52,120 --> 00:58:03,969
But but on this on this 30 year old case and what this shows at about a two and a half day range is a phenomenal spread in in solutions.
542
00:58:03,970 --> 00:58:09,070
So these are all pressure maps. So you see, you know, you can get pretty much everything from, you know,
543
00:58:09,070 --> 00:58:16,720
absolutely balmy, calm days to to horrendous weather, very, very wide divergence.
544
00:58:17,290 --> 00:58:24,670
Okay. So what would you do with this information? How would you sort of how would you convey this information to the public?
545
00:58:25,360 --> 00:58:34,930
Well, one thing you can do is basically say, okay, how many of these members had hurricane force winds and count them?
546
00:58:36,010 --> 00:58:43,569
And if there were, say, 15 out of the 50 had hurricane force winds, you would say 59, 50, 30 out of 100.
547
00:58:43,570 --> 00:58:46,870
So you say is roughly a 30% chance of.
548
00:58:47,710 --> 00:58:50,850
Hurricane force winds, and that's actually pretty much what we see.
549
00:58:50,860 --> 00:58:59,260
So this is a a contour map showing the probability of hurricane force wind gusts on the 16th of October based on that ensemble forecast.
550
00:58:59,710 --> 00:59:03,730
And these are blue lines here. And the English Channel are up around 30, 40%.
551
00:59:05,720 --> 00:59:09,950
Now. So what do you do with that information? That's up to you.
552
00:59:10,220 --> 00:59:12,380
That's not up to the meteorologist. That's up to you.
553
00:59:12,710 --> 00:59:18,200
If you've just bought a brand new yacht and you're not terribly good at sailing and you're thinking of crossing the channel,
554
00:59:18,920 --> 00:59:27,140
maybe don't want to do it. If you just bought a brand new Lamborghini and it's sitting underneath a big oak tree, maybe you want to move it.
555
00:59:27,440 --> 00:59:33,470
30 or 40% is probably a big enough probability that that would be a prudent thing to do.
556
00:59:33,860 --> 00:59:40,630
But it raises a very interesting and a whole topic for another talk some time about making decisions under uncertainty.
557
00:59:40,640 --> 00:59:46,970
But the main point is, and indeed much better decisions can be made once you have a quantified quantify uncertainty.
558
00:59:47,240 --> 00:59:51,110
And that's precisely what this does. All right. So I think I'm finished.
559
00:59:51,110 --> 00:59:54,530
Sorry for going slightly over time. I'm going to be another shamed advertisement.
560
00:59:54,530 --> 00:59:57,919
If you're interested in reading any more about anything I've spoken about,
561
00:59:57,920 --> 01:00:03,530
I just want to highlight three papers which I can send you or you can get them on the on the Internet.
562
01:00:05,090 --> 01:00:13,579
Most of this talk is actually based on a paper that I wrote with colleague Andrea stirring from the physics department and Gregory Sagan,
563
01:00:13,580 --> 01:00:17,780
who's here in the Maths Institute, about this butterfly real butterfly effect.
564
01:00:17,780 --> 01:00:25,180
And we go into a lot more detail about the theory of the navier-stokes equations and things like that, but as well as the history behind the you know,
565
01:00:25,190 --> 01:00:34,190
what I spoke about, I also wrote a biographical memoir for at Lawrence, who died a few years ago, who was a former member of the Royal Society.
566
01:00:34,200 --> 01:00:38,440
So there's a lot of the biographical background to Lawrence in that which you can
567
01:00:38,520 --> 01:00:42,450
again get from the Royal Society and if you're interested in this area of this.
568
01:00:42,450 --> 01:00:47,629
So what I find rather fascinating linkage between Lawrence and the girdle incompleteness
569
01:00:47,630 --> 01:00:51,920
theorems and some of the ideas of Roger Penrose on non compute mobility in physics.
570
01:00:51,920 --> 01:00:59,330
That's in a paper in contemporary physics published by the Institute of Physics, based on a lecture I gave in the physics department a few years back.
571
01:00:59,990 --> 01:01:00,950
Thank you for your attention.