1
00:00:00,896 --> 00:00:04,642
We know more about
other planets than our own,

2
00:00:05,446 --> 00:00:08,670
and today, I want to show you
a new type of robot

3
00:00:08,693 --> 00:00:12,733
designed to help us
better understand our own planet.

4
00:00:13,408 --> 00:00:15,020
It belongs to a category

5
00:00:15,044 --> 00:00:19,998
known in the oceanographic community
as an unmanned surface vehicle, or USV.

6
00:00:20,665 --> 00:00:22,595
And it uses no fuel.

7
00:00:23,047 --> 00:00:26,981
Instead, it relies
on wind power for propulsion.

8
00:00:27,005 --> 00:00:30,760
And yet, it can sail around the globe
for months at a time.

9
00:00:31,474 --> 00:00:34,352
So I want to share with you
why we built it,

10
00:00:34,376 --> 00:00:35,868
and what it means for you.

11
00:00:37,814 --> 00:00:42,398
A few years ago, I was on a sailboat
making its way across the Pacific,

12
00:00:42,422 --> 00:00:44,620
from San Francisco to Hawaii.

13
00:00:45,296 --> 00:00:48,650
I had just spent the past 10 years
working nonstop,

14
00:00:48,674 --> 00:00:51,790
developing video games
for hundreds of millions of users,

15
00:00:51,814 --> 00:00:55,335
and I wanted to take a step back
and look at the big picture

16
00:00:55,359 --> 00:00:57,296
and get some much-needed thinking time.

17
00:00:57,918 --> 00:00:59,463
I was the navigator on board,

18
00:01:00,034 --> 00:01:03,698
and one evening, after a long session
analyzing weather data

19
00:01:03,722 --> 00:01:05,351
and plotting our course,

20
00:01:05,375 --> 00:01:08,449
I came up on deck and saw
this beautiful sunset.

21
00:01:08,814 --> 00:01:10,247
And a thought occurred to me:

22
00:01:10,747 --> 00:01:14,101
How much do we really know
about our oceans?

23
00:01:15,252 --> 00:01:19,131
The Pacific was stretching all around me
as far as the eye could see,

24
00:01:19,155 --> 00:01:21,470
and the waves were
rocking our boat forcefully,

25
00:01:21,494 --> 00:01:24,700
a sort of constant reminder
of its untold power.

26
00:01:25,291 --> 00:01:28,523
How much do we really know
about our oceans?

27
00:01:29,235 --> 00:01:30,773
I decided to find out.

28
00:01:32,362 --> 00:01:35,471
What I quickly learned
is that we don't know very much.

29
00:01:35,495 --> 00:01:38,621
The first reason is just
how vast oceans are,

30
00:01:38,645 --> 00:01:40,986
covering 70 percent of the planet,

31
00:01:41,010 --> 00:01:44,555
and yet we know they drive
complex planetary systems

32
00:01:44,579 --> 00:01:45,759
like global weather,

33
00:01:45,783 --> 00:01:48,228
which affect all of us on a daily basis,

34
00:01:48,252 --> 00:01:49,870
sometimes dramatically.

35
00:01:50,562 --> 00:01:53,746
And yet, those activities
are mostly invisible to us.

36
00:01:54,944 --> 00:01:58,709
Ocean data is scarce by any standard.

37
00:01:59,268 --> 00:02:04,372
Back on land, I had grown used to
accessing lots of sensors --

38
00:02:04,396 --> 00:02:06,024
billions of them, actually.

39
00:02:06,576 --> 00:02:13,030
But at sea, in situ data
is scarce and expensive.

40
00:02:13,535 --> 00:02:17,530
Why? Because it relies on
a small number of ships and buoys.

41
00:02:17,919 --> 00:02:20,790
How small a number
was actually a great surprise.

42
00:02:21,329 --> 00:02:24,450
Our National Oceanic
and Atmospheric Administration,

43
00:02:24,474 --> 00:02:26,097
better known as NOAA,

44
00:02:26,121 --> 00:02:28,450
only has 16 ships,

45
00:02:28,969 --> 00:02:33,223
and there are less than
200 buoys offshore globally.

46
00:02:33,770 --> 00:02:35,441
It is easy to understand why:

47
00:02:35,465 --> 00:02:37,605
the oceans are an unforgiving place,

48
00:02:37,629 --> 00:02:40,963
and to collect in situ data,
you need a big ship,

49
00:02:40,987 --> 00:02:43,321
capable of carrying a vast amount of fuel

50
00:02:43,345 --> 00:02:45,163
and large crews,

51
00:02:45,187 --> 00:02:48,366
costing hundreds
of millions of dollars each,

52
00:02:48,390 --> 00:02:53,712
or, big buoys tethered to the ocean floor
with a four-mile-long cable

53
00:02:54,513 --> 00:02:57,508
and weighted down
by a set of train wheels,

54
00:02:57,532 --> 00:03:02,141
which is both dangerous to deploy
and expensive to maintain.

55
00:03:02,628 --> 00:03:04,554
What about satellites, you might ask?

56
00:03:05,028 --> 00:03:06,931
Well, satellites are fantastic,

57
00:03:06,955 --> 00:03:09,458
and they have taught us
so much about the big picture

58
00:03:09,482 --> 00:03:11,503
over the past few decades.

59
00:03:11,527 --> 00:03:13,903
However, the problem with satellites

60
00:03:13,927 --> 00:03:18,207
is they can only see through one micron
of the surface of the ocean.

61
00:03:19,124 --> 00:03:22,697
They have relatively poor
spatial and temporal resolution,

62
00:03:23,450 --> 00:03:27,724
and their signal needs to be corrected
for cloud cover and land effects

63
00:03:27,748 --> 00:03:28,956
and other factors.

64
00:03:30,094 --> 00:03:33,365
So what is going on in the oceans?

65
00:03:33,929 --> 00:03:35,561
And what are we trying to measure?

66
00:03:36,283 --> 00:03:38,869
And how could a robot be of any use?

67
00:03:39,361 --> 00:03:43,259
Let's zoom in on
a small cube in the ocean.

68
00:03:43,283 --> 00:03:46,591
One of the key things we want
to understand is the surface,

69
00:03:46,615 --> 00:03:48,654
because the surface,
if you think about it,

70
00:03:48,678 --> 00:03:51,977
is the nexus of all air-sea interaction.

71
00:03:52,001 --> 00:03:56,995
It is the interface through which
all energy and gases must flow.

72
00:03:57,552 --> 00:03:59,652
Our sun radiates energy,

73
00:03:59,676 --> 00:04:02,572
which is absorbed by oceans as heat

74
00:04:02,596 --> 00:04:05,452
and then partially released
into the atmosphere.

75
00:04:05,476 --> 00:04:09,644
Gases in our atmosphere like CO2
get dissolved into our oceans.

76
00:04:10,099 --> 00:04:13,728
Actually, about 30 percent
of all global CO2 gets absorbed.

77
00:04:14,603 --> 00:04:18,211
Plankton and microorganisms
release oxygen into the atmosphere,

78
00:04:18,235 --> 00:04:22,406
so much so that every other breath
you take comes from the ocean.

79
00:04:22,430 --> 00:04:25,462
Some of that heat generates evaporation,
which creates clouds

80
00:04:25,486 --> 00:04:27,780
and then eventually
leads to precipitation.

81
00:04:27,804 --> 00:04:30,057
And pressure gradients
create surface wind,

82
00:04:30,081 --> 00:04:32,878
which moves the moisture
through the atmosphere.

83
00:04:33,731 --> 00:04:37,564
Some of the heat radiates down
into the deep ocean

84
00:04:37,588 --> 00:04:39,716
and gets stored in different layers,

85
00:04:39,740 --> 00:04:43,045
the ocean acting as some kind
of planetary-scale boiler

86
00:04:43,069 --> 00:04:44,690
to store all that energy,

87
00:04:44,714 --> 00:04:48,451
which later might be released
in short-term events like hurricanes

88
00:04:48,475 --> 00:04:50,856
or long-term phenomena like El Niño.

89
00:04:51,329 --> 00:04:55,644
These layers can get mixed up
by vertical upwelling currents

90
00:04:55,668 --> 00:04:59,114
or horizontal currents,
which are key in transporting heat

91
00:04:59,138 --> 00:05:01,112
from the tropics to the poles.

92
00:05:01,623 --> 00:05:04,422
And of course, there is marine life,

93
00:05:04,446 --> 00:05:09,462
occupying the largest ecosystem
in volume on the planet,

94
00:05:09,486 --> 00:05:13,188
from microorganisms to fish
to marine mammals,

95
00:05:13,212 --> 00:05:16,220
like seals, dolphins and whales.

96
00:05:16,244 --> 00:05:19,821
But all of these
are mostly invisible to us.

97
00:05:21,821 --> 00:05:27,197
The challenge in studying
those ocean variables at scale

98
00:05:27,221 --> 00:05:29,395
is one of energy,

99
00:05:29,419 --> 00:05:33,542
the energy that it takes to deploy
sensors into the deep ocean.

100
00:05:34,597 --> 00:05:36,890
And of course, many solutions
have been tried --

101
00:05:36,914 --> 00:05:38,525
from wave-actuated devices

102
00:05:38,549 --> 00:05:40,090
to surface drifters

103
00:05:40,114 --> 00:05:42,836
to sun-powered electrical drives --

104
00:05:42,860 --> 00:05:44,972
each with their own compromises.

105
00:05:45,745 --> 00:05:48,915
Our team breakthrough came
from an unlikely source --

106
00:05:48,939 --> 00:05:53,288
the pursuit of the world speed record
in a wind-powered land yacht.

107
00:05:53,742 --> 00:05:56,043
It took 10 years of research
and development

108
00:05:56,067 --> 00:05:58,596
to come up with a novel wing concept

109
00:05:58,620 --> 00:06:02,428
that only uses three watts
of power to control

110
00:06:02,452 --> 00:06:05,540
and yet can propel a vehicle
all around the globe

111
00:06:05,564 --> 00:06:07,764
with seemingly unlimited autonomy.

112
00:06:08,262 --> 00:06:11,608
By adapting this wing concept
into a marine vehicle,

113
00:06:12,356 --> 00:06:14,662
we had the genesis of an ocean drone.

114
00:06:15,280 --> 00:06:17,614
Now, these are larger than they appear.

115
00:06:17,638 --> 00:06:22,214
They are about 15 feet high,
23 feet long, seven feet deep.

116
00:06:22,238 --> 00:06:24,291
Think of them as surface satellites.

117
00:06:24,315 --> 00:06:27,453
They're laden with an array
of science-grade sensors

118
00:06:27,477 --> 00:06:29,285
that measure all key variables,

119
00:06:29,309 --> 00:06:32,307
both oceanographic and atmospheric,

120
00:06:32,331 --> 00:06:36,688
and a live satellite link transmits
this high-resolution data

121
00:06:36,712 --> 00:06:38,754
back to shore in real time.

122
00:06:39,515 --> 00:06:42,393
Our team has been hard at work
over the past few years,

123
00:06:42,417 --> 00:06:45,925
conducting missions in some of
the toughest ocean conditions

124
00:06:45,949 --> 00:06:47,173
on the planet,

125
00:06:47,197 --> 00:06:49,662
from the Arctic to the tropical Pacific.

126
00:06:49,686 --> 00:06:52,410
We have sailed all the way
to the polar ice shelf.

127
00:06:52,434 --> 00:06:54,773
We have sailed into Atlantic hurricanes.

128
00:06:55,159 --> 00:06:57,117
We have rounded Cape Horn,

129
00:06:57,141 --> 00:07:00,380
and we have slalomed between
the oil rigs of the Gulf of Mexico.

130
00:07:00,404 --> 00:07:02,734
This is one tough robot.

131
00:07:03,629 --> 00:07:06,610
Let me share with you
recent work that we did

132
00:07:06,634 --> 00:07:08,508
around the Pribilof Islands.

133
00:07:08,532 --> 00:07:12,451
This is a small group of islands
deep in the cold Bering Sea

134
00:07:12,475 --> 00:07:14,637
between the US and Russia.

135
00:07:14,661 --> 00:07:18,111
Now, the Bering Sea is the home
of the walleye pollock,

136
00:07:18,135 --> 00:07:20,710
which is a whitefish
you might not recognize,

137
00:07:20,734 --> 00:07:25,421
but you might likely have tasted
if you enjoy fish sticks or surimi.

138
00:07:25,445 --> 00:07:29,067
Yes, surimi looks like crabmeat,
but it's actually pollock.

139
00:07:29,908 --> 00:07:32,918
And the pollock fishery
is the largest fishery in the nation,

140
00:07:32,942 --> 00:07:35,487
both in terms of value and volume --

141
00:07:35,511 --> 00:07:38,980
about 3.1 billion pounds
of fish caught every year.

142
00:07:39,695 --> 00:07:42,394
So over the past few years,
a fleet of ocean drones

143
00:07:42,418 --> 00:07:44,644
has been hard at work in the Bering Sea

144
00:07:44,668 --> 00:07:49,123
with the goal to help assess
the size of the pollock fish stock.

145
00:07:49,147 --> 00:07:52,918
This helps improve the quota system
that's used to manage the fishery

146
00:07:52,942 --> 00:07:55,110
and help prevent a collapse
of the fish stock

147
00:07:55,134 --> 00:07:58,078
and protects this fragile ecosystem.

148
00:07:58,538 --> 00:08:03,347
Now, the drones survey
the fishing ground using acoustics,

149
00:08:03,371 --> 00:08:04,895
i.e., a sonar.

150
00:08:04,919 --> 00:08:08,140
This sends a sound wave downwards,

151
00:08:08,164 --> 00:08:10,674
and then the reflection,
the echo from the sound wave

152
00:08:10,698 --> 00:08:12,456
from the seabed or schools of fish,

153
00:08:12,480 --> 00:08:15,390
gives us an idea of what's happening
below the surface.

154
00:08:15,849 --> 00:08:19,855
Our ocean drones are actually
pretty good at this repetitive task,

155
00:08:19,879 --> 00:08:23,784
so they have been gridding
the Bering Sea day in, day out.

156
00:08:23,808 --> 00:08:30,767
Now, the Pribilof Islands are also
the home of a large colony of fur seals.

157
00:08:31,175 --> 00:08:35,664
In the 1950s, there were about
two million individuals in that colony.

158
00:08:36,374 --> 00:08:40,134
Sadly, these days,
the population has rapidly declined.

159
00:08:40,158 --> 00:08:42,501
There's less than 50 percent
of that number left,

160
00:08:42,525 --> 00:08:45,033
and the population
continues to fall rapidly.

161
00:08:45,904 --> 00:08:48,007
So to understand why,

162
00:08:48,031 --> 00:08:51,520
our science partner at
the National Marine Mammal Laboratory

163
00:08:51,544 --> 00:08:54,881
has fitted a GPS tag
on some of the mother seals,

164
00:08:54,905 --> 00:08:56,382
glued to their furs.

165
00:08:56,406 --> 00:08:59,407
And this tag measures location and depth

166
00:08:59,431 --> 00:09:01,763
and also has a really cool little camera

167
00:09:01,787 --> 00:09:04,183
that's triggered by sudden acceleration.

168
00:09:04,207 --> 00:09:08,121
Here is a movie taken
by an artistically inclined seal,

169
00:09:08,145 --> 00:09:11,886
giving us unprecedented insight
into an underwater hunt

170
00:09:11,910 --> 00:09:13,724
deep in the Arctic,

171
00:09:13,748 --> 00:09:16,120
and the shot of this pollock prey

172
00:09:16,144 --> 00:09:18,910
just seconds before it gets devoured.

173
00:09:18,934 --> 00:09:22,720
Now, doing work in the Arctic
is very tough, even for a robot.

174
00:09:22,744 --> 00:09:25,334
They had to survive a snowstorm in August

175
00:09:25,358 --> 00:09:28,849
and interferences from bystanders --

176
00:09:28,873 --> 00:09:32,131
that little spotted seal enjoying a ride.

177
00:09:32,155 --> 00:09:34,716
(Laughter)

178
00:09:35,378 --> 00:09:41,665
Now, the seal tags have recorded
over 200,000 dives over the season,

179
00:09:42,395 --> 00:09:44,354
and upon a closer look,

180
00:09:44,378 --> 00:09:48,909
we get to see the individual seal tracks
and the repetitive dives.

181
00:09:49,413 --> 00:09:52,270
We are on our way to decode
what is really happening

182
00:09:52,294 --> 00:09:53,973
over that foraging ground,

183
00:09:53,997 --> 00:09:55,342
and it's quite beautiful.

184
00:09:56,362 --> 00:09:59,765
Once you superimpose the acoustic data
collected by the drones,

185
00:09:59,789 --> 00:10:01,913
a picture starts to emerge.

186
00:10:01,937 --> 00:10:06,309
As the seals leave the islands
and swim from left to right,

187
00:10:06,333 --> 00:10:10,746
they are observed to dive at a relatively
shallow depth of about 20 meters,

188
00:10:10,770 --> 00:10:14,859
which the drone identifies
is populated by small young pollock

189
00:10:14,883 --> 00:10:16,966
with low calorific content.

190
00:10:16,990 --> 00:10:21,114
The seals then swim much greater distance
and start to dive deeper

191
00:10:21,138 --> 00:10:25,810
to a place where the drone identifies
larger, more adult pollock,

192
00:10:25,834 --> 00:10:28,093
which are more nutritious as fish.

193
00:10:28,117 --> 00:10:32,239
Unfortunately, the calories expended
by the mother seals

194
00:10:32,263 --> 00:10:34,386
to swim this extra distance

195
00:10:34,410 --> 00:10:39,326
don't leave them with enough energy
to lactate their pups back on the island,

196
00:10:39,350 --> 00:10:41,384
leading to the population decline.

197
00:10:42,408 --> 00:10:47,755
Further, the drones identify that
the water temperature around the island

198
00:10:47,779 --> 00:10:49,464
has significantly warmed.

199
00:10:49,488 --> 00:10:53,788
It might be one of the driving forces
that's pushing the pollock north,

200
00:10:53,812 --> 00:10:56,267
and to spread in search of colder regions.

201
00:10:56,765 --> 00:10:58,550
So the data analysis is ongoing,

202
00:10:58,574 --> 00:11:01,953
but already we can see
that some of the pieces of the puzzle

203
00:11:01,977 --> 00:11:03,350
from the fur seal mystery

204
00:11:03,374 --> 00:11:04,975
are coming into focus.

205
00:11:06,690 --> 00:11:08,779
But if you look back at the big picture,

206
00:11:08,803 --> 00:11:10,304
we are mammals, too.

207
00:11:11,032 --> 00:11:14,856
And actually, the oceans provide
up to 20 kilos of fish per human per year.

208
00:11:14,880 --> 00:11:18,009
As we deplete our fish stocks,
what can we humans learn

209
00:11:18,033 --> 00:11:19,618
from the fur seal story?

210
00:11:20,719 --> 00:11:23,592
And beyond fish, the oceans
affect all of us daily

211
00:11:23,616 --> 00:11:25,490
as they drive global weather systems,

212
00:11:25,514 --> 00:11:28,275
which affect things like
global agricultural output

213
00:11:28,299 --> 00:11:31,928
or can lead to devastating destruction
of lives and property

214
00:11:31,952 --> 00:11:35,153
through hurricanes,
extreme heat and floods.

215
00:11:35,756 --> 00:11:39,536
Our oceans are pretty much
unexplored and undersampled,

216
00:11:39,560 --> 00:11:43,997
and today, we still know more
about other planets than our own.

217
00:11:44,021 --> 00:11:47,809
But if you divide this vast ocean
in six-by-six-degree squares,

218
00:11:47,833 --> 00:11:50,650
each about 400 miles long,

219
00:11:51,539 --> 00:11:53,693
you'd get about 1,000 such squares.

220
00:11:53,717 --> 00:11:56,013
So little by little,
working with our partners,

221
00:11:56,037 --> 00:12:00,135
we are deploying one ocean drone
in each of those boxes,

222
00:12:00,159 --> 00:12:02,581
the hope being that
achieving planetary coverage

223
00:12:02,605 --> 00:12:05,785
will give us better insights
into those planetary systems

224
00:12:05,809 --> 00:12:07,120
that affect humanity.

225
00:12:07,603 --> 00:12:11,215
We have been using robots to study
distant worlds in our solar system

226
00:12:11,239 --> 00:12:12,429
for a while now.

227
00:12:12,939 --> 00:12:15,570
Now it is time to quantify our own planet,

228
00:12:16,440 --> 00:12:20,279
because we cannot fix
what we cannot measure,

229
00:12:20,303 --> 00:12:23,195
and we cannot prepare
for what we don't know.

230
00:12:23,761 --> 00:12:24,921
Thank you.

231
00:12:24,945 --> 00:12:28,019
(Applause)