That means that I use radar to study glaciers and ice sheets.

And like most glaciologists right now,

I'm working on the problem of estimating

how much the ice is going to contribute to sea level rise in the future.

So today, I want to talk to you about

why it's so hard to put good numbers on sea level rise,

and why I believe that by changing the way we think about radar technology

and earth-science education,

we can get much better at it.

When most scientists talk about sea level rise,

they show a plot like this.

This is produced using ice sheet and climate models.

On the right, you can see the range of sea level

predicted by these models over the next 100 years.

For context, this is current sea level,

and this is the sea level

above which more than 4 million people could be vulnerable to displacement.

So in terms of planning,

the uncertainty in this plot is already large.

However, beyond that, this plot comes with the asterisk and the caveat,

"... unless the West Antarctic Ice Sheet collapses."

And in that case, we would be talking about dramatically higher numbers.

They'd literally be off the chart.

And the reason we should take that possibility seriously

is that we know from the geologic history of the Earth

that there were periods in its history

when sea level rose much more quickly than today.

And right now, we cannot rule out

the possibility of that happening in the future.

So why can't we say with confidence

whether or not a significant portion of a continent-scale ice sheet

will or will not collapse?

Well, in order to do that, we need models

that we know include all of the processes, conditions and physics

that would be involved in a collapse like that.

And that's hard to know,

because those processes and conditions are taking place

beneath kilometers of ice,

and satellites, like the one that produced this image,

are blind to observe them.

In fact, we have much more comprehensive observations of the surface of Mars

than we do of what's beneath the Antarctic ice sheet.

And this is even more challenging in that we need these observations

at a gigantic scale in both space and time.

In terms of space, this is a continent.

And in the same way that in North America,

the Rocky Mountains, Everglades and Great Lakes regions are very distinct,

so are the subsurface regions of Antarctica.

And in terms of time, we now know

that ice sheets not only evolve over the timescale of millennia and centuries,

but they're also changing over the scale of years and days.

So what we want is observations beneath kilometers of ice

at the scale of a continent,

and we want them all the time.

So how do we do this?

Well, we're not totally blind to the subsurface.

I said in the beginning that I was a radio glaciologist,

and the reason that that's a thing

is that airborne ice-penetrating radar is the main tool we have

to see inside of ice sheets.

So most of the data used by my group is collected by airplanes

like this World War II-era DC-3,

that actually fought in the Battle of the Bulge.

You can see the antennas underneath the wing.

These are used to transmit radar signals down into the ice.

And the echos that come back contain information

about what's happening inside and beneath the ice sheet.

While this is happening,

scientists and engineers are on the airplane

for eight hours at a stretch,

making sure that the radar's working.

And I think this is actually a misconception

about this type of fieldwork,

where people imagine scientists peering out the window,

contemplating the landscape, its geologic context

and the fate of the ice sheets.

We actually had a guy from the BBC's "Frozen Planet" on one of these flights.

And he spent, like, hours videotaping us turn knobs.

(Laughter)

And I was actually watching the series years later with my wife,

and a scene like this came up, and I commented on how beautiful it was.

And she said, "Weren't you on that flight?"

(Laughter)

I said, "Yeah, but I was looking at a computer screen."

(Laughter)

So when you think about this type of fieldwork,

don't think about images like this.

Think about images like this.

(Laughter)

This is a radargram, which is a vertical profile through the ice sheet,

kind of like a slice of cake.

The bright layer on the top is the surface of the ice sheet,

the bright layer on the bottom is the bedrock of the continent itself,

and the layers in between are kind of like tree rings,

in that they contain information about the history of the ice sheet.

And it's amazing that this works this well.

The ground-penetrating radars that are used

to investigate infrastructures of roads or detect land mines

struggle to get through a few meters of earth.

And here we're peering through three kilometers of ice.

And there are sophisticated, interesting, electromagnetic reasons for that,

but let's say for now that ice is basically the perfect target for radar,

and radar is basically the perfect tool to study ice sheets.

These are the flight lines

of most of the modern airborne radar-sounding profiles

collected over Antarctica.

This is the result of heroic efforts over decades

by teams from a variety of countries and international collaborations.

And when you put those together, you get an image like this,

which is what the continent of Antarctica would look like

without all the ice on top.

And you can really see the diversity of the continent in an image like this.

The red features are volcanoes or mountains;

the areas that are blue would be open ocean

if the ice sheet was removed.

This is that giant spatial scale.

However, all of this that took decades to produce

is just one snapshot of the subsurface.

It does not give us any indication of how the ice sheet is changing in time.

Now, we're working on that, because it turns out

that the very first radar observations of Antarctica were collected

using 35 millimeter optical film.

And there were thousands of reels of this film

in the archives of the museum of the Scott Polar Research Institute

at the University of Cambridge.

So last summer, I took a state-of-the-art film scanner

that was developed for digitizing Hollywood films and remastering them,

and two art historians,

and we went over to England, put on some gloves

and archived and digitized all of that film.

So that produced two million high-resolution images

that my group is now working on analyzing and processing

for comparing with contemporary conditions in the ice sheet.

And, actually, that scanner -- I found out about it

from an archivist at the Academy of Motion Picture Arts and Sciences.

So I'd like to thank the Academy --

(Laughter)

for making this possible.

(Laughter)

And as amazing as it is

that we can look at what was happening under the ice sheet 50 years ago,

this is still just one more snapshot.

It doesn't give us observations

of the variation at the annual or seasonal scale,

that we know matters.

There's some progress here, too.

There are these recent ground-based radar systems that stay in one spot.

So you take these radars and put them on the ice sheet

and you bury a cache of car batteries.

And you leave them out there for months or years at a time,

and they send a pulse down into the ice sheet

every so many minutes or hours.

So this gives you continuous observation in time --

but at one spot.

So if you compare that imaging to the 2-D pictures provided by the airplane,

this is just one vertical line.

And this is pretty much where we are as a field right now.

We can choose between good spatial coverage

with airborne radar sounding

and good temporal coverage in one spot with ground-based sounding.

But neither gives us what we really want:

both at the same time.

And if we're going to do that,

we're going to need totally new ways of observing the ice sheet.

And ideally, those should be extremely low-cost

so that we can take lots of measurements from lots of sensors.

Well, for existing radar systems,

the biggest driver of cost is the power required

to transmit the radar signal itself.

So it'd be great if we were able to use existing radio systems

or radio signals that are in the environment.

And fortunately, the entire field of radio astronomy

is built on the fact that there are bright radio signals in the sky.

And a really bright one is our sun.

So, actually, one of the most exciting things my group is doing right now

is trying to use the radio emissions from the sun as a type of radar signal.

This is one of our field tests at Big Sur.

That PVC pipe ziggurat is an antenna stand some undergrads in my lab built.

And the idea here is that we stay out at Big Sur,

and we watch the sunset in radio frequencies,

and we try and detect the reflection of the sun off the surface of the ocean.

Now, I know you're thinking, "There are no glaciers at Big Sur."

(Laughter)

And that's true.

(Laughter)

But it turns out that detecting the reflection of the sun

off the surface of the ocean

and detecting the reflection off the bottom of an ice sheet

are extremely geophysically similar.

And if this works,

we should be able to apply the same measurement principle in Antarctica.

And this is not as far-fetched as it seems.

The seismic industry has gone through a similar technique-development exercise,

where they were able to move from detonating dynamite as a source,

to using ambient seismic noise in the environment.

And defense radars use TV signals and radio signals all the time,

so they don't have to transmit a signal of radar

and give away their position.

So what I'm saying is, this might really work.

And if it does, we're going to need extremely low-cost sensors

so we can deploy networks of hundreds or thousands of these on an ice sheet

to do imaging.

And that's where the technological stars have really aligned to help us.

Those earlier radar systems I talked about

were developed by experienced engineers over the course of years

at national facilities

with expensive specialized equipment.

But the recent developments in software-defined radio,

rapid fabrication and the maker movement,

make it so that it's possible for a team of teenagers

working in my lab over the course of a handful of months

to build a prototype radar.

OK, they're not any teenagers, they're Stanford undergrads,

but the point holds --

(Laughter)

that these enabling technologies are letting us break down the barrier

between engineers who build instruments and scientists that use them.

And by teaching engineering students to think like earth scientists

and earth-science students who can think like engineers,

my lab is building an environment in which we can build custom radar sensors

for each problem at hand,

that are optimized for low cost and high performance

for that problem.

And that's going to totally change the way we observe ice sheets.

Look, the sea level problem and the role of the cryosphere in sea level rise

is extremely important

and will affect the entire world.

But that is not why I work on it.

I work on it for the opportunity to teach and mentor

extremely brilliant students,

because I deeply believe that teams of hypertalented,

hyperdriven, hyperpassionate young people

can solve most of the challenges facing the world,

and that providing the observations required to estimate sea level rise

is just one of the many such problems they can and will solve.

Thank you.

(Applause)