ABOUT THE SPEAKER
Gavin Schmidt - Climate scientist
What goes into a climate model? Gavin Schmidt looks at how we use past and present data to model potential futures.

Why you should listen
Gavin Schmidt is a climate scientist at Columbia University's Earth Institute and is Deputy Chief at the NASA Goddard Institute for Space Studies. He works on understanding past, present and future climate change, using ever-more refined models and data sets to explore how the planet's climate behaves over time.
 
Schmidt is also deeply committed to communicating science to the general public. As a contributing editor at RealClimate.org, he helps make sure general readers have access to the basics of climate science, and works to bring the newest data and models into the public discussion around one of the most pressing issues of our time. He has worked with the American Museum of Natural History and the New York Academy of Sciences on education and public outreach, and he is the author of Climate Change: Picturing the Science, with Josh Wolfe.
More profile about the speaker
Gavin Schmidt | Speaker | TED.com
TED2014

Gavin Schmidt: The emergent patterns of climate change

Filmed:
1,270,244 views

You can't understand climate change in pieces, says climate scientist Gavin Schmidt. It's the whole, or it's nothing. In this illuminating talk, he explains how he studies the big picture of climate change with mesmerizing models that illustrate the endlessly complex interactions of small-scale environmental events.
- Climate scientist
What goes into a climate model? Gavin Schmidt looks at how we use past and present data to model potential futures. Full bio

Double-click the English transcript below to play the video.

00:12
We live in a very complex environment:
0
864
2323
00:15
complexity and dynamism
1
3187
1904
00:17
and patterns of evidence
2
5091
2063
00:19
from satellite photographs, from videos.
3
7154
2885
00:22
You can even see it outside your window.
4
10039
3011
00:25
It's endlessly complex, but somehow familiar,
5
13050
3860
00:28
but the patterns kind of repeat,
6
16910
1960
00:30
but they never repeat exactly.
7
18870
2490
00:33
It's a huge challenge to understand.
8
21360
4127
00:37
The patterns that you see
9
25487
2132
00:39
are there at all of the different scales,
10
27619
3720
00:43
but you can't chop it into one little bit and say,
11
31339
2906
00:46
"Oh, well let me just make a smaller climate."
12
34245
2663
00:48
I can't use the normal products of reductionism
13
36908
4212
00:53
to get a smaller and smaller thing that I can study
14
41120
2722
00:55
in a laboratory and say, "Oh,
15
43842
2308
00:58
now that's something I now understand."
16
46150
2396
01:00
It's the whole or it's nothing.
17
48546
3367
01:03
The different scales that give you
18
51913
2552
01:06
these kinds of patterns
19
54465
2122
01:08
range over an enormous range of magnitude,
20
56587
3457
01:12
roughly 14 orders of magnitude,
21
60044
2416
01:14
from the small microscopic particles
22
62460
2491
01:16
that seed clouds
23
64951
2376
01:19
to the size of the planet itself,
24
67327
2560
01:21
from 10 to the minus six
25
69887
1276
01:23
to 10 to the eight,
26
71163
1077
01:24
14 orders of spatial magnitude.
27
72240
2292
01:26
In time, from milliseconds to millennia,
28
74532
3411
01:29
again around 14 orders of magnitude.
29
77943
3055
01:32
What does that mean?
30
80998
1387
01:34
Okay, well if you think about how
31
82385
1939
01:36
you can calculate these things,
32
84324
2660
01:38
you can take what you can see,
33
86984
1960
01:40
okay, I'm going to chop it up
34
88944
1026
01:41
into lots of little boxes,
35
89970
1379
01:43
and that's the result of physics, right?
36
91349
2355
01:45
And if I think about a weather model,
37
93704
1725
01:47
that spans about five orders of magnitude,
38
95429
2494
01:49
from the planet to a few kilometers,
39
97923
3127
01:53
and the time scale
40
101050
1538
01:54
from a few minutes to 10 days, maybe a month.
41
102588
4412
01:59
We're interested in more than that.
42
107000
1395
02:00
We're interested in the climate.
43
108395
1305
02:01
That's years, that's millennia,
44
109700
2141
02:03
and we need to go to even smaller scales.
45
111841
2573
02:06
The stuff that we can't resolve,
46
114414
1601
02:08
the sub-scale processes,
47
116015
1965
02:09
we need to approximate in some way.
48
117980
1980
02:11
That is a huge challenge.
49
119960
1762
02:13
Climate models in the 1990s
50
121722
2188
02:15
took an even smaller chunk of that,
51
123910
1970
02:17
only about three orders of magnitude.
52
125880
2018
02:19
Climate models in the 2010s,
53
127898
2095
02:21
kind of what we're working with now,
54
129993
1774
02:23
four orders of magnitude.
55
131767
2940
02:26
We have 14 to go,
56
134707
2303
02:29
and we're increasing our capability
57
137010
2200
02:31
of simulating those at about
58
139210
1870
02:33
one extra order of magnitude every decade.
59
141080
3546
02:36
One extra order of magnitude in space
60
144626
1895
02:38
is 10,000 times more calculations.
61
146521
3249
02:41
And we keep adding more things,
62
149770
2380
02:44
more questions to these different models.
63
152150
2374
02:46
So what does a climate model look like?
64
154524
2733
02:49
This is an old climate model, admittedly,
65
157257
2341
02:51
a punch card, a single line of Fortran code.
66
159598
4080
02:55
We no longer use punch cards.
67
163678
1978
02:57
We do still use Fortran.
68
165656
2241
02:59
New-fangled ideas like C
69
167897
1957
03:01
really haven't had a big impact
70
169854
3235
03:05
on the climate modeling community.
71
173089
2367
03:07
But how do we go about doing it?
72
175456
1400
03:08
How do we go from that complexity that you saw
73
176856
4624
03:13
to a line of code?
74
181480
2530
03:16
We do it one piece at a time.
75
184010
1573
03:17
This is a picture of sea ice
76
185583
1878
03:19
taken flying over the Arctic.
77
187461
2098
03:21
We can look at all of the different equations
78
189559
2038
03:23
that go into making the ice grow
79
191597
3112
03:26
or melt or change shape.
80
194709
2114
03:28
We can look at the fluxes.
81
196823
1131
03:29
We can look at the rate at which
82
197954
1952
03:31
snow turns to ice, and we can code that.
83
199906
2845
03:34
We can encapsulate that in code.
84
202751
2329
03:37
These models are around
85
205080
1226
03:38
a million lines of code at this point,
86
206306
2083
03:40
and growing by tens of thousands of lines of code
87
208389
3470
03:43
every year.
88
211859
1191
03:45
So you can look at that piece,
89
213050
1653
03:46
but you can look at the other pieces too.
90
214703
1922
03:48
What happens when you have clouds?
91
216625
1933
03:50
What happens when clouds form,
92
218558
2159
03:52
when they dissipate, when they rain out?
93
220717
1882
03:54
That's another piece.
94
222599
1742
03:56
What happens when we have radiation
95
224341
1846
03:58
coming from the sun, going through the atmosphere,
96
226187
2534
04:00
being absorbed and reflected?
97
228721
1926
04:02
We can code each of those
very small pieces as well.
98
230647
3979
04:06
There are other pieces:
99
234626
1416
04:08
the winds changing the ocean currents.
100
236042
3460
04:11
We can talk about the role of vegetation
101
239502
3770
04:15
in transporting water from the soils
102
243272
2329
04:17
back into the atmosphere.
103
245601
1969
04:19
And each of these different elements
104
247570
2914
04:22
we can encapsulate and put into a system.
105
250484
3624
04:26
Each of those pieces ends up adding to the whole.
106
254108
5148
04:31
And you get something like this.
107
259256
2297
04:33
You get a beautiful representation
108
261553
2848
04:36
of what's going on in the climate system,
109
264401
2622
04:39
where each and every one of those
110
267023
3389
04:42
emergent patterns that you can see,
111
270412
2782
04:45
the swirls in the Southern Ocean,
112
273194
2003
04:47
the tropical cyclone in the Gulf of Mexico,
113
275197
2756
04:49
and there's two more that are going to pop up
114
277953
1641
04:51
in the Pacific at any point now,
115
279594
2354
04:53
those rivers of atmospheric water,
116
281948
2713
04:56
all of those are emergent properties
117
284661
2857
04:59
that come from the interactions
118
287518
2124
05:01
of all of those small-scale processes I mentioned.
119
289642
3495
05:05
There's no code that says,
120
293137
1905
05:07
"Do a wiggle in the Southern Ocean."
121
295042
1857
05:08
There's no code that says, "Have two
122
296899
2668
05:11
tropical cyclones that spin around each other."
123
299567
2898
05:14
All of those things are emergent properties.
124
302465
3812
05:18
This is all very good. This is all great.
125
306277
2146
05:20
But what we really want to know
126
308423
1270
05:21
is what happens to these emergent properties
127
309693
1949
05:23
when we kick the system?
128
311642
1705
05:25
When something changes, what
happens to those properties?
129
313347
3533
05:28
And there's lots of different ways to kick the system.
130
316880
2989
05:31
There are wobbles in the Earth's orbit
131
319869
2033
05:33
over hundreds of thousands of years
132
321902
1879
05:35
that change the climate.
133
323781
2026
05:37
There are changes in the solar cycles,
134
325807
2136
05:39
every 11 years and longer, that change the climate.
135
327943
3105
05:43
Big volcanoes go off and change the climate.
136
331048
3574
05:46
Changes in biomass burning, in smoke,
137
334622
3238
05:49
in aerosol particles, all of those things
138
337860
1863
05:51
change the climate.
139
339723
1822
05:53
The ozone hole changed the climate.
140
341545
4059
05:57
Deforestation changes the climate
141
345604
2217
05:59
by changing the surface properties
142
347821
1926
06:01
and how water is evaporated
143
349747
1990
06:03
and moved around in the system.
144
351737
2466
06:06
Contrails change the climate
145
354203
2285
06:08
by creating clouds where there were none before,
146
356488
2867
06:11
and of course greenhouse gases change the system.
147
359355
4598
06:15
Each of these different kicks
148
363953
3021
06:18
provides us with a target
149
366974
2151
06:21
to evaluate whether we understand
150
369125
2835
06:23
something about this system.
151
371960
2161
06:26
So we can go to look at
152
374121
2392
06:28
what model skill is.
153
376513
2704
06:31
Now I use the word "skill" advisedly:
154
379217
2033
06:33
Models are not right or wrong; they're always wrong.
155
381250
2411
06:35
They're always approximations.
156
383661
1720
06:37
The question you have to ask
157
385381
1894
06:39
is whether a model tells you more information
158
387275
3079
06:42
than you would have had otherwise.
159
390354
1925
06:44
If it does, it's skillful.
160
392279
3381
06:47
This is the impact of the ozone hole
161
395660
2454
06:50
on sea level pressure, so
low pressure, high pressures,
162
398114
2860
06:52
around the southern oceans, around Antarctica.
163
400974
2595
06:55
This is observed data.
164
403569
1913
06:57
This is modeled data.
165
405482
2088
06:59
There's a good match
166
407570
1594
07:01
because we understand the physics
167
409164
1951
07:03
that controls the temperatures in the stratosphere
168
411115
3138
07:06
and what that does to the winds
169
414253
1746
07:07
around the southern oceans.
170
415999
2181
07:10
We can look at other examples.
171
418180
1519
07:11
The eruption of Mount Pinatubo in 1991
172
419699
2856
07:14
put an enormous amount of aerosols, small particles,
173
422555
2799
07:17
into the stratosphere.
174
425354
1587
07:18
That changed the radiation
balance of the whole planet.
175
426941
3147
07:22
There was less energy coming
in than there was before,
176
430088
2782
07:24
so that cooled the planet,
177
432870
1658
07:26
and those red lines and those green lines,
178
434528
2019
07:28
those are the differences between what we expected
179
436547
2565
07:31
and what actually happened.
180
439112
1688
07:32
The models are skillful,
181
440800
1783
07:34
not just in the global mean,
182
442583
1693
07:36
but also in the regional patterns.
183
444276
3044
07:39
I could go through a dozen more examples:
184
447320
2840
07:42
the skill associated with solar cycles,
185
450160
2850
07:45
changing the ozone in the stratosphere;
186
453010
2070
07:47
the skill associated with orbital changes
187
455080
2347
07:49
over 6,000 years.
188
457427
2056
07:51
We can look at that too, and the models are skillful.
189
459483
2398
07:53
The models are skillful in response to the ice sheets
190
461881
3094
07:56
20,000 years ago.
191
464975
1520
07:58
The models are skillful
192
466495
1671
08:00
when it comes to the 20th-century trends
193
468166
2904
08:03
over the decades.
194
471070
1515
08:04
Models are successful at modeling
195
472585
2282
08:06
lake outbursts into the North Atlantic
196
474867
2605
08:09
8,000 years ago.
197
477472
1765
08:11
And we can get a good match to the data.
198
479237
3090
08:15
Each of these different targets,
199
483463
2387
08:17
each of these different evaluations,
200
485850
2130
08:19
leads us to add more scope
201
487980
2391
08:22
to these models,
202
490371
1151
08:23
and leads us to more and more
203
491522
2744
08:26
complex situations that we can ask
204
494266
3988
08:30
more and more interesting questions,
205
498254
2569
08:32
like, how does dust from the Sahara,
206
500823
2710
08:35
that you can see in the orange,
207
503533
1734
08:37
interact with tropical cyclones in the Atlantic?
208
505267
3443
08:40
How do organic aerosols from biomass burning,
209
508710
3477
08:44
which you can see in the red dots,
210
512187
2723
08:46
intersect with clouds and rainfall patterns?
211
514910
2934
08:49
How does pollution, which you can see
212
517844
1787
08:51
in the white wisps of sulfate pollution in Europe,
213
519631
3899
08:55
how does that affect the
temperatures at the surface
214
523530
3335
08:58
and the sunlight that you get at the surface?
215
526865
3488
09:02
We can look at this across the world.
216
530353
3488
09:05
We can look at the pollution from China.
217
533841
3660
09:09
We can look at the impacts of storms
218
537501
3598
09:13
on sea salt particles in the atmosphere.
219
541099
3444
09:16
We can see the combination
220
544543
2561
09:19
of all of these different things
221
547104
2171
09:21
happening all at once,
222
549275
1468
09:22
and we can ask much more interesting questions.
223
550743
2407
09:25
How do air pollution and climate coexist?
224
553150
4624
09:29
Can we change things
225
557774
1509
09:31
that affect air pollution and
climate at the same time?
226
559283
2589
09:33
The answer is yes.
227
561872
2344
09:36
So this is a history of the 20th century.
228
564216
3044
09:39
The first one is the model.
229
567260
2243
09:41
The weather is a little bit different
230
569503
1407
09:42
to what actually happened.
231
570910
1289
09:44
The second one are the observations.
232
572199
2032
09:46
And we're going through the 1930s.
233
574231
2325
09:48
There's variability, there are things going on,
234
576556
2824
09:51
but it's all kind of in the noise.
235
579380
2182
09:53
As you get towards the 1970s,
236
581562
2862
09:56
things are going to start to change.
237
584424
2009
09:58
They're going to start to look more similar,
238
586433
2062
10:00
and by the time you get to the 2000s,
239
588495
2558
10:03
you're already seeing the
patterns of global warming,
240
591063
2642
10:05
both in the observations and in the model.
241
593705
2749
10:08
We know what happened over the 20th century.
242
596454
2127
10:10
Right? We know that it's gotten warmer.
243
598581
1760
10:12
We know where it's gotten warmer.
244
600341
1611
10:13
And if you ask the models why did that happen,
245
601952
2740
10:16
and you say, okay, well, yes,
246
604692
2125
10:18
basically it's because of the carbon dioxide
247
606817
1866
10:20
we put into the atmosphere.
248
608683
1979
10:22
We have a very good match
249
610662
1682
10:24
up until the present day.
250
612344
2627
10:26
But there's one key reason why we look at models,
251
614971
3420
10:30
and that's because of this phrase here.
252
618391
2221
10:32
Because if we had observations of the future,
253
620612
2495
10:35
we obviously would trust them more than models,
254
623107
3329
10:38
But unfortunately,
255
626436
1944
10:40
observations of the future
are not available at this time.
256
628380
5540
10:45
So when we go out into the
future, there's a difference.
257
633920
2705
10:48
The future is unknown, the future is uncertain,
258
636625
2562
10:51
and there are choices.
259
639187
2404
10:53
Here are the choices that we have.
260
641591
1833
10:55
We can do some work to mitigate
261
643424
2139
10:57
the emissions of carbon dioxide into the atmosphere.
262
645563
2795
11:00
That's the top one.
263
648358
1926
11:02
We can do more work
264
650284
1906
11:04
to really bring it down
265
652190
2176
11:06
so that by the end of the century,
266
654366
2218
11:08
it's not much more than there is now.
267
656584
2464
11:11
Or we can just leave it to fate
268
659048
3767
11:14
and continue on
269
662815
1493
11:16
with a business-as-usual type of attitude.
270
664308
3746
11:20
The differences between these choices
271
668054
3456
11:23
can't be answered by looking at models.
272
671510
4797
11:28
There's a great phrase
273
676307
1639
11:29
that Sherwood Rowland,
274
677946
1793
11:31
who won the Nobel Prize for the chemistry
275
679739
3864
11:35
that led to ozone depletion,
276
683603
2273
11:37
when he was accepting his Nobel Prize,
277
685876
2397
11:40
he asked this question:
278
688273
1379
11:41
"What is the use of having developed a science
279
689652
2311
11:43
well enough to make predictions if, in the end,
280
691963
3261
11:47
all we're willing to do is stand around
281
695224
2829
11:50
and wait for them to come true?"
282
698053
2707
11:52
The models are skillful,
283
700760
2737
11:55
but what we do with the
information from those models
284
703497
3318
11:58
is totally up to you.
285
706815
2171
12:00
Thank you.
286
708986
1938
12:02
(Applause)
287
710924
2916

▲Back to top

ABOUT THE SPEAKER
Gavin Schmidt - Climate scientist
What goes into a climate model? Gavin Schmidt looks at how we use past and present data to model potential futures.

Why you should listen
Gavin Schmidt is a climate scientist at Columbia University's Earth Institute and is Deputy Chief at the NASA Goddard Institute for Space Studies. He works on understanding past, present and future climate change, using ever-more refined models and data sets to explore how the planet's climate behaves over time.
 
Schmidt is also deeply committed to communicating science to the general public. As a contributing editor at RealClimate.org, he helps make sure general readers have access to the basics of climate science, and works to bring the newest data and models into the public discussion around one of the most pressing issues of our time. He has worked with the American Museum of Natural History and the New York Academy of Sciences on education and public outreach, and he is the author of Climate Change: Picturing the Science, with Josh Wolfe.
More profile about the speaker
Gavin Schmidt | Speaker | TED.com