 Thank you, Valerie, and I wanted to thank the organizers of the NCAR Explorer series for allowing me to be here, and especially Lorena Luna who wasn't able to be here today but helped with this together, and all of you for coming out on a beautiful fall day to hear about a subject that's very close to my heart. So to begin with, aside from my sheets being interesting scientific phenomena in their own right, why do we here at NCAR care about them a lot? And the most obvious reason is because as ice sheets as they lose mass, that mass goes into the ocean and raises sea level. And the more sea level, the higher sea level rises, the greater the chances of flooding. Flooding is the most common and costly U.S. national disaster. It causes about $10 billion a year of damage on average. Some years when you have a big storm like Hurricane Harvey, significantly more than that. And the U.S. populations is increasingly concentrated in coastal counties with a lot of people and ecosystems and infrastructure near coastlines. And usually when you see pictures of flooding, it's associated with a major storm. Sea level rise, of course, raises the baseline level and makes it likely that a given storm will produce more damage because the storm surge starts at that higher baseline. But also increasingly we're seeing flood damage, which is more routine, which doesn't need a storm, but just happens during a high tide. So for example, this graph here is a time series going from 1950 to the present. And it shows the frequency of tidal floods above what's called a nuisance level, you know, enough to cause problems at various NOAA tide gauges along the coast. And you can see the frequency has increased a lot in the last 10 or 20 years, even with the relatively modest sea level rise that we've seen so far. And around the world, of course, there's many, many people exposed to harm from sea level rise. About 300 million people or close to 5% of the world's population lives within 5 meters of sea level. And the populations most at risk tend to be in Asia. If the climate were to warm 2 degrees Celsius over pre-industrial levels, we would expect to see about 5 meters of sea level rise, ultimately. And there's people in China have the largest number at risk. Outside of Asia, the United States has about 12 million people living within 5 meters of sea level. And if you happen to live in one of the small island nations in the Indian Ocean or the Pacific Ocean, sea level rise is already a real existential concern because many of these countries are in the Pacific, for example, are in coral atolls. And the entire atoll may be no more than 1 or 2 meters above sea level. So this, for example, is the most populated atoll in the Marshall Islands in the Pacific. And it sits almost entirely within 1 or 2 meters of sea level. And along with several other countries, including Kiribati, Tonga, Tuvalu, and others, would no longer be habitable if sea level were to rise, even a couple of meters above the levels we have now. If you go back to the late 1800s, which is when sea level measurements began, traditionally, sea level was measured by buoys and tide gauges, which are put in coastal locations. And the difficulty there is that sea level is rising at different rates in different places. And you have to somehow integrate that all over the world. But you do your best to combine these records. And you can pile a global sea level record going back to 1880. And you can see that sea level rise has been fairly steady over the last more than 100 years, with a total of about 20 centimeters of sea level rise since 1880. And here I want to pause for a moment. I'm a scientist, and I always think in terms of metric, because I find the other units to be too hard to keep track of. But in case you're less familiar with metric, I'll just take a moment and do some conversions. 30 centimeters is about one foot. So if you see a number like 3 millimeters per year, you can convert that to 30 centimeters in a century or about one foot per century. I'll be using units of meters a lot. A meter is a little over 3 feet. Kilometer is about 1⁄6 of a mile. And then when I talk about temperature, I'm going to talk about degrees Celsius. And to get a sense of how big one degree Celsius is, it's important to remember that one degree Celsius equals 1.8 degrees Fahrenheit. So if I give a number in Celsius, double it, take a little bit off the top, and that's what it would be in the degrees we're more used to thinking about. So since the Industrial Revolution, temperatures in the global average have risen by about one degree Celsius, or about 1.8 degrees Fahrenheit. But back to the sea level plot. Since 1993, there have been satellites using altimetry, measuring the distance between the satellite and the sea surface, which we can use to get a global picture of what sea level is doing. And satellites show that sea level is now rising about 3 millimeters a year, which is large in the 20th century average, reflecting some acceleration that's happened in the last quarter century or so. Now, the global average sea level rise has two main causes. You can either add mass to the ocean, or you can increase the volume of a given mass of ocean. So if you melt ice on land, that's adding mass to the ocean. That can be either in a form of ice sheets or smaller glaciers. Also, water expands as it heats. And since the ocean is heating, it's expanding and taking up more space. So that's responsible for about 1 third of current sea level rise with land ice melting accounting for about 2 thirds. But depending on where you are, the sea level rise rate can vary a lot. If your land is subsiding for natural reasons or because of pumping of groundwater or that sort of thing, you can have that enhances the effects of sea level rise, makes your relative sea level rise more. You can also have, in some places, slow land motions. For example, there's some places on earth that are still rebounding from being covered with ice during the previous ice age. And that tends to make it appear that sea level is falling. You can also have changes in ocean circulation. So for instance, the Gulf Stream has slowed down somewhat in the last few decades. And when the Gulf Stream slows down, that leads to more water piling up on the east coast of the US. So that's one reason that if you look at the numbers here, you can see the greatest increases in sea levels since 1950 on the east coast and the Gulf Coast with smaller increases on the left. So your local contributions can be, at least so far, as significant or larger than the global average contribution. And before I go farther, I wanted to talk about the different kinds of ice that I and my colleagues study. First, a glacier is just a mass of ice that's formed from snow. Snow that falls in the land doesn't melt in the summer. Over the years, it compacts into ice under its own weight and then starts flowing downhill under the force of gravity. And this is your prototypical mountain glacier in the French Alps called the Merida Glace. Now if you have a glacier that thickens and expands to the point that it's no longer topographically constrained, for example, by mountains, but in this case, the Vatnijökul ice cap of Iceland forms more of a dome shape. Then conventionally, if that's less than 50,000 kilometers squared, which is about the size of Switzerland, it's called an ice cap. And if it's larger than that, it's called an ice sheet. So an ice sheet is a large mass of glacier ice, which is not constrained by mountains or other topography. So the Earth today has two large ice sheets, one in Greenland and the larger ice sheet filling most of the continent of Antarctica. And on an ice sheet, the majority of the ice is typically grounded, meaning that if you dug down to the bed of the ice sheet, you would get to bedrock or soil. But especially in Antarctica, a significant fraction of the ice sheet is floating. So what happens is the ice starts flowing out to sea and at some point it's sufficiently thin that it floats. And once it floats, it no longer contributes to sea level directly where it to melt, like an ice cube melting in a glass of water. But ice shelves can exert back pressure on the glacier ice that's flowing into them. So if you remove an ice shelf and you remove that buttressing or pressure, ice shelves can contribute indirectly to sea level rise. So in Antarctica, the two biggest ice shelves are the Ross Ice Shelf here and the Ronny Filchner Ice Shelf here. Those are both about the size of France, so pretty good size. And the fastest flowing part of an ice sheet is called an ice stream. So for an ice sheet, fast would be anywhere between one or several kilometers per year, so not fast in a conventional sense, but that's a pretty fast speed for a glacier. And then the slower ice that would be in either side of the ice stream might be going somewhere between one and maybe a few hundred meters per year. And these are the so-called cyple coast ice streams, which flow into the Ross Ice Shelf. On this plot, here you can see the Greenland Ice Sheet. And here you can see the sea ice that Valerie mentioned, which is very reflective and surrounded by ocean north of Siberia and Alaska in this particular photo. The other kinds of ice sheets and ice shelves and so on are land ice. Sea ice is different, and then instead of forming from compacted snow, it forms from seawater directly on the surface of the ocean. And so it doesn't affect sea level. And also the dimensions are a lot different. Sea ice is typically not more than two or three meters thick, whereas an ice sheet can easily be two or three kilometers thick. These are just pictures of sea ice, first year ice, on the Ross Sea in Antarctica. And this is a big multi-year ice flow surrounded by first year ice, which is called pancake ice, because it forms these characteristic discs. But that's all I'll say about sea ice for today. This is a picture taken by one of my colleagues who was on doing field work in Greenland and took a picture out of the airplane of the largest and fastest outlet glacier, which is similar to an ice stream, in Greenland, which is called Jakobshap and Isbray, which at its nearest terminus, it fills up a fjord called the Lulisap fjord, which is only about five kilometers wide. But the ice, it goes several tens of kilometers upstream. And it flows at the rate of 10 kilometers a year, which is very fast. Around it is the slower moving part of the ice sheet, which might be going, oh, say, a couple hundred meters a year. And if you look over here, you're away from the ice sheet. You are on land where the ice hasn't reached, but there's summer snowfall, winter snowfall that then melts in the summer. And out here, this is the terminus of the ice sheet, where the main ice sheet ends and calves and forms icebergs. And then in the fjord, you can see this what's called a melange of icebergs and sea ice all mixed together. So that's just a picture to show lots of different kinds of ice all in one place. And I wanted to go back to the Merida Glass in the French Alps, partly because I happened to be there about a month ago. I was able to go to a workshop near Geneva and took a day to visit Chamonix, which is the town sitting below Montbelin, which has several big glaciers flowing off it down into the valley, including the Merida Glass. This is the longest glacier in the French Alps. And it's been accessible by a cog railway since 1908. And for that, people would ride up to the glacier by mule. And so you can see, sometime in the early 20th century, people taking the expedition out on the surface of the Merida Glass and all the crevasses coming down from the summit of Montbelin. And there was a guy named Edward Spell Terini who, in 1909, went up in a balloon so he could take this beautiful aerial photo of the Merida Glass. And then there was another photographer who went up by helicopter just a year or so ago and took a picture from the same viewpoint. And you can see, for example, that the sidewalls here, where there used to be ice, there isn't anymore. So that gives you some sense of the retreat there's been in the last century, about 100 meters of thinning compared to where the surface used to be. Which sounds like a lot, but when you're there, it's even more striking because you take the train up and then you get in a cable car which takes you down to where the surface of the glacier used to be a few decades ago. And then, but instead of seeing a glacier there, you see steps, lots and lots of makeshift metal steps. And as you walk down, there are signs at various points along the way telling you that you've reached the level of the glacier to a certain year. So when you get to where the glacier was in 1990, which is not so long ago, you look down, you've got a long way to go before you get to the terminus. And you walk a while longer and you get to the 2005 level and still quite a long way to go, quite a lot, at least a couple of hundred steps. And then you keep walking and then you get to 2015 and you think, well, that wasn't long ago at all, but you've still got about a hundred steps to go before you get to the, so the glacier is melting more than 10 meters a year at this point. And then you get to the bottom. And by the way, the reason most people do this is because there's an ice cape at the bottom which is periodically carved out so you can go inside and see this beautiful blue glacier ice. So you do that and they've also covered some of the ice with what looks like a big tarp to stop it from melting as quickly. And then you look back up and you've got 430 steps to climb just to get to where the level of the ice was a few decades ago, which for me was very striking and sad because this is one of the most iconic glaciers in the world and it's given pleasure to thousands and thousands of people and it's been there for at least 10,000 years and it's dying now. There will be no Merida Glass as we know it in a few more decades because it just is not, can't be in balance with the climate that we're in today. And if you look at how much ice it's lost in the last century, it turns out to be about a billion cubic meters or if a billion seems like a large number, you can imagine an ice cube that's one kilometer on each side and that's about how much ice has been lost from Merida Glass in that time. And if you were able to weigh that much ice, it would be about a billion tons or one gigaton in the units that are used. And it turns out that it takes a lot of ice to raise the sea level even a small amount. So for example, if you, to raise the sea level globally by one millimeter, you need about 360 gigatons of ice. So in other words, you would need to take all the ice that melted disappeared from Merida Glass in the last century and put that in the ocean every day for a year and that would give you one millimeter of sea level rise. And that happens to be about the contribution that ice sheets are making to sea level rise now about one century's worth of glacier melt every day, you know, day after day and increasing over time. So sort of going from the one glacier scale to larger scales, if you were to melt all the ice in the European Alps and add it to the ocean, you'd get about 0.3 millimeters of sea level rise. So, you know, not even a pencil point. If you were to melt all the ice in Alaska, which has considerably more ice, that would be about five centimeters of sea level equivalent. So SLE means sea level equivalent, which again is about 360 billion tons of ice. If you were to melt all the glaciers and ice caps in the world outside Greenland and Antarctica, you'd get just under half a meter of sea level rise. So while these are making a big contribution to sea level right now, because they're relatively melting so quickly, the contribution, the ultimate contribution is not as great as that from ice sheets because there's just not as much ice there. The Greenland ice sheet is about, you know, it's typically two or three kilometers thick and occupies most of the island of Greenland. And if you put all that in the ocean, that would raise sea level by about seven meters. But then about 90% of the ice is in Antarctica, about 60 meters sea level equivalent. So that's one reason people are most concerned about Greenland and Antarctica is just because there's so much ice there. Then I want to say a little about the physics and dynamics of glaciers. As I said, a glacier forms from snow that compacts into ice and then starts flowing downhill. And you can think of ice, although it appears very solid, you can think of it as a very viscous, slowly deforming fluid. So if the ice were happened to be frozen at the bed, at the base of the ice, then it would gradually deform under its own weight with the velocity increasing as you go up. So the maximum velocity of a glacier is always at the surface. And this deformational component is typically not too big. It might be something like 100 meters a year of deformation give or take. But if you want to get a glacier to slow, just to speed up, what you do is you add water at the bed. You could either do it in the form of a thin film of water that the glacier can slide on top of. Or if you get water into the soil or till and that becomes soft, the soil can deform and that allows the glacier to move forward at speeds of a kilometer a year or more. And so the big ice streams are all sliding in some way. And then glaciers, when they're in balance with the climate, they have an equal amount of mass gain and mass loss in a typical year. So the mass gain is almost entirely from snowfall. So you have snow falling on the glacier and not eventually turning into ice and some of it's blown around by the wind. But then you can lose that mass either by sublimation from the surface, which is where it goes directly from the solid to the gas phase. Or if it's warm enough in the summer, you'll have summer meltwater. And some of that will go back into the snowpack and refreeze, but some of it will run off to the ocean. So that's the way you can lose mass. And also, especially Antarctica, you can, the ice can thin as it reaches the ocean, but not melt at the surface. So you have a floating ice shelf here. And then what happens to a floating ice shelf is either you have some melt from the bottom if you have a warm ocean or you can cab off an iceberg. And once you cab off an iceberg, it eventually melts somewhere else. So the mass balance in an ice sheet is between snowfall in and some combination of melting and sublimation and calving out. And it's the difference that gives you the change in sea level. So, and relatively, these are big numbers and a relatively small imbalance can give you a big change in sea level. And how do we know that ice sheets are losing mass? We've only been able to measure the mass of ice sheets and ice mass changes since about the 1990s. And that's because of new satellites that were launched then. There's two main methods of measuring mass and mass changes in an ice sheet. One is satellite altimetry, which is where you bounce a radar or light signal off the surface and measure the time to come back. And in that way you can measure whether the surface is thinning or thickening. And so, for example, in Antarctica, this is called the Pine Island Glacier, that patch of red. And that's where the ice is significantly thinning as measured by altimeter. But one drawback of this method is that when the ice surface goes down, it could be because the ice is losing mass, but it could also be because the snow is compacting because it's more dense and an altimeter doesn't distinguish between those two things. But if you measure the ice sheets' gravity, which is what the GRACE satellites are designed to do, in this case you have two satellites that are orbiting the Earth one following the other. And it's possible to measure very small variations in distance between the two satellites, which tell you that the Earth's gravity field is changing in a certain location. And the changes in the Earth's gravitational field are associated with changing mass of ocean waters or, in this case, the mass of ice sitting below. So this is a map of ice lost from a Greenland ice sheet with the darker red colors showing where there's been the most thinning between the time the GRACE satellite was launched in 2002 and 2016. It was just ended its mission about a year ago, but there's another set of satellites that have been launched and are getting ready to go. The thing about gravity is you don't get... The picture you get is a little bit smeared out and diffuse, but it is a measure of mass as opposed to thickness. The Greenland ice sheet, as I said, it's about seven meters sea level equivalent. And the mass that comes in is balanced and roughly equal when the climate's in balance. You have about equal amounts of surface melting leading to runoff and iceberg calving. Since the 1990s, though, Greenland's been out of balance and you've had mass loss of averaging about 280 gigatons a year. This is Greenland velocities and you can see the blue colors are faster ice and the red colors are slower ice. And you can see how you have all these fast ice streams along the West Coast and Southeast and then here in the Northeast. And Greenland is relatively isolated from the ocean in the sense that most of its bed sits above sea level. This part here is where it's blue is below sea level because it's way down so much by the ice, but out in the margins, most of the bed that the ice sheet sits on is above sea level. And that makes it relatively less vulnerable to ocean warming than its counterpart, Antarctica. So Antarctica, as I said, has about 60 meters sea level equivalent. A lot of that ice is locked up in the cold, dry part of the ice sheet called East Antarctica, which is this part. And it's so high up and so cold that a lot of that is not vulnerable to any kind of melting that would happen in the foreseeable future. However, this part of the ice sheet called West Antarctica, if you look at its topography, all the blue spots are places where the ice sheet is grounded below sea level. So if you were to get, in some case, you have an ice shelf, like the raw shelf here. In other case, you have grounded ice that's sitting on the seabed. So if you were to get warm water up under the ice shelf, you have the potential to melt a lot of ice fairly quickly. And paleo climate or past climate record suggests that West Antarctica has ebbed and flowed a lot over the past couple of million years. And if you look, these are the fastest part of the Antarctic ice sheet is the two ice, big ice shelves here. And then here's Pine Island Glacier and Thwates Glacier, which you might have read about, which are two of the fastest thinning and changing glaciers. And they're both thinning, we think, because warm ocean water is getting up underneath them. Now if you go, so those are the two ice sheets we have now. If you go back in the past, there are of course more ice sheets. If you go back 20,000 years to a time called the last glacial maximum, you had large ice sheets over most of Canada coming down into what's now the Northern United States and also ice sheets covering a lot of Northern Europe. And there was so much ice then that sea level was about 120 meters lower than it is now. And so there were places that are sea covered now that where there are land bridges, for example, on the Bering Sea where it was possible at times to walk between Asia and North America. But in past warm climates, sea level has been a bit higher than it is now. The present period is known as interglacial because we don't have big ice sheets. And the last interglacial before this one called the last interglacial was about 125,000 years ago. And compared to pre-industrial temperatures, it was about one or two degrees warmer than those temperatures. So in other words, we're now about one degree warmer than we used to be, and that's getting to be at the low end of the last interglacial. At that time, carbon dioxide was not that high compared to what now it's about 400 parts per million, then it was about 280. But we think that sea level over several thousand years increased by somewhere between six and nine meters. And we think based on climate records and model simulations, this shows a simulation by one of my NCAR colleagues, Betty Otto Bluzner, that Greenland lost some ice but only a couple of meters worth of sea level. So you had to have an Antarctic contribution of at least five meters, which would be probably the majority of West Antarctica. And if you go back farther in time to the period that the waxing and waning rice ages started about two and a half million years ago, the period before that was called the Pliocene, which was two or three degrees warmer than the temperatures that have characterized the recent interglacial. And that was the last time carbon dioxide was as concentrated in the atmosphere as it is today, about 400 parts per million. And at that time, there's still a lot of uncertainty because the records are not as plentiful. But we think global sea level was somewhere between 10 and 30 meters higher than today with that temperature rise of two or three degrees. And even to get 10 meters, you need probably most of Greenland, most of West Antarctica. And to get 30 meters, you certainly need some of East Antarctica as well. So this is concerning because we potentially could have temperature increases at that size in the next several decades. However, they would take a long time to play out. The changes that happened in the Pliocene happened over at least several thousand years. But you can see that it may take a while, but you don't need a lot of warming to make sea level go up quite significantly. I wanna spend a little time on this plot just to talk about some of the interactions between sea level and climate. This plot is based on records from an ice core that was drilled in Antarctica. And in an ice core, you have bubbles that are trapped in the ice that preserve a record of the atmosphere that existed at the time that the snow fell and turned to ice. And so you can dig down to the bottom of the Antarctic ice sheet. And in this case, find ice that's more than 400,000 years old. And you find that there've been on roughly 100,000 year cycles, these changes in carbon dioxide concentration between about 200 and 300 parts per million. And global temperature changing very much in sync with the carbon dioxide, plus or minus about five degrees Celsius globally. And sea level going up or down by about 120 meters, depending on whether you have a big Northern Hemisphere ice sheet or not. And these changes that have happened in the past have been associated with changes in the Earth's orbit. The central idea being that depending on what the tilt of the Earth's axis is and what time of year the Earth is closest to the sun, you can have either cool or relatively warm summers in the Northern Hemisphere. And if you have a relatively cool summer, then the snow that falls during the winter may not melt during the summer and it has a chance to pile up and form an ice sheet. And we think that's what happened about at the end of the last interglacial, about 120,000 years ago. And then the ice sheet slowly built up over the next 100,000 years to the last glacier maximum. And then the orbital configuration lined up such that you had a high amount of melting in the Northern Hemisphere. And the ice sheet at that time may have somewhat overextended itself and been relatively unstable. And then you can melt a very large ice sheet in a fairly short time, getting up to several meters of sea level rise per century until 10,000 years ago when the big ice sheets in Canada were mostly gone. And then we've had a period of about 10,000 years called the Holocene, leading up to the Industrial Revolution when we had a climate that was fairly stable and benign. This, I should say, is actually not quite to scale. If we were to scale, the 400 would be something like this. And it's not possible to increase CO2 without having an effect on temperature. The relationships are somewhat complicated and CO2 and temperatures tend to feedback on each other. And that's what's different about the present times that we're way out of bounds of the natural changes in CO2 and CO2 is fundamentally what's driving everything else. If you go back about 3,000 years and look at sea level reconstructions, this is about 500 BC up to the present. You can see that there are some wobbles, but generally sea level was only varied by about five centimeters or so per century as compared to about 15 centimeters in the 20th century. So the rise we've seen in the last century was the largest in, we think, at least 3,000 years and probably during the whole Holocene. If you look just during the satellite era since 1993, this is a figure prepared by one of my colleagues at the university here in Boulder. This is global average sea level rise from satellites. And you can see some wobbles from year to year that are associated with things like volcanoes and whether there's an El Nino that changes the temperature of the Pacific. But overall, a steady increase with a little bit of acceleration in recent years, which is probably associated with the greater contribution from ice sheets than there was earlier. And the average here is about, as I've said, is about three millimeters per year. And of that three millimeters, you get about equal contributions from the Greenland Antarctic ice sheets from glaciers and ice caps and from thermal expansion of the ocean as it warms. And you can see the increasing contribution from Greenland Antarctica and that Greenland has a relatively greater contribution, although Antarctica has been apparently catching up in the last several years. Then if you look forward into the 21st century, you've likely heard of the Intergovernmental Panel on Climate Change or the IPCC, which every six or seven years puts out a massive report on the state of the science of climate change and climate change adaptation and mitigation. And their last report was in 2013. And they typically consider a range of emissions scenarios ranging from low greenhouse gas emissions if CO2 emissions were to sort of peak now all rapidly through the rest of the century. And high emissions, which is more of a business as usual scenario. And between the low and high emissions, you can see that by 2100, it makes a big difference to your temperature increase, where with high emissions, you have a projected global average temperature increase of around three or four or five degrees Celsius. Whereas with the low, you can top out around one or maybe one and a half. Similarly with sea level, which is highly correlated to temperature. Up through about 2050, the spread is relatively small. That even with the low emissions, we're probably committed to another 30 to 60 centimeters, one or two feet of sea level rise by 2100, even with these, with low emissions. But with higher emissions, we're looking at probably a range of half a meter to a meter with the separation coming in the later part of the century. But there was a sort of asterisk associated with this projection, which was that when the report came out, there were suggestions that the marine-based sectors, the part of the Antarctic ice sheet grounded below sea level, might be unstable to collapse. And if so, sea level could go up more than the likely range of half a meter to a meter during the century. And I wanna come back to that, but first just to add that again, global average sea level rise is not necessarily the same as what you see locally. I mentioned subsidence and glacial rebound, but also ice sheets are massive enough to have their own gravity. So if an ice sheet loses mass, it no longer tugs as hard on the ocean that's immediately around it. And so you could actually have potentially sea level fall around a melting ice sheet while sea level rises everywhere else. So if you're in North America, other things being equal, you would prefer to lose mass from Greenland if you had to rather than Antarctica. But these are, this is the relative contribution. This is still imposed on a overall sea level rise in most of the world. The Greenland ice sheet we think would not be viable if you had warming of more than a couple of degrees Celsius. You already have a lot of summer melting in Greenland and you had more summer melting. What eventually would happen is that you'd have, even if you took iceberg calving out of the picture, you would have more summer melting and runoff than you have incoming snowfall. And at that point, what's called the surface mass balance, which is the difference between the incoming snowfall and the mass loss would be negative and loss of the ice sheet would be inevitable if nothing were done about that. But it would take a long time, at least several hundred years and maybe two or 3000 years. This is a climate simulation showing the evolution of Greenland over several thousand years and starts off slow. But what eventually happens is that as you thin the ice sheet, the ice sheet is now, because it's lower in elevation, it feels a warmer atmosphere and the warmer the atmosphere it feels, the more it's melting. And so that can be a positive feedback that leads to more rapid decline. And once that gets going, it may not be possible to reverse. In Antarctica, there's not much summer surface melting at this time, but there's a lot more exposure to the ocean, which makes things more complicated. If you look at the ocean off the coast of Antarctica, the typical structure you see is you have fresh and cold water sitting on top of warmer and saltier water. Typically the warm water is at the top because warm water is lighter, but because in this case the warm water is salty, it actually sits below the colder water. And warm for an ice sheet means maybe one or two degrees about freezing, which is still pretty cold, but for an ice sheet, that's enough to melt maybe 10 or 20 meters of ice in a year. So you have, in many places, you have this warm water, relatively warm water, called circumpolar deep water, sitting just off the continental shelf. And if that water is able to get onto the continental shelf and under the ice shelf, you can do a lot of melting in a fairly short time. And then this is called a grounding line. That's where the ice is thin enough to become a float. And the way the bed is drawn is that it gets deeper as you go inland. And because of the dynamics of ice sheets, if you start melting and making the grounding line go back to here, the thicker ice is, the faster it flows. So you have this big arrow showing and representing increased flux from a thicker ice sheet. And you have an increased outflow of ice, that means you get more thinning and so on. So once this gets going, you can have retreat all the way back until the bed reverses slope and starts turning up again. And this is potentially something that once triggered can go on without needing any more warm water to trigger it. It's just gonna happen on its own. Another thing to concern Antarctica is although most of Antarctica's, the ice sheet is very high and cold. The ice shelves themselves are close to sea level and some of them have been warming in recent years. And this one in particular is called the Larson B, which was an ice shelf on the Antarctic peninsula. And in 2002, because of surface warming, you had a lot of ponding on the surface of the shelf, which you can see by the dark coloring here. And what we think happened is that the ponds led to crevasses that went all the way through the shelf and then the shelf shattered over the course of a few weeks in 2002. So there's no shelf there now. And the glaciers that were being buttressed by the shelf sped up and this was measured. And this is not one of the largest ice shelves, but if this were to happen to a larger ice shelf as it's been speculated, that could lead to a big increase in flow of grounded ice into the ocean. And that was the hypothesis of a paper that came out a couple of years ago, which is probably the most talked about paper on ice sheet modeling in the last few years by Rob DiCanto and David Pollard. And they have an ice sheet model and had found that their ice sheet model wasn't as sensitive as they thought it should be to the warming associated with past climates like the last interglacial. And they put in these new mechanisms, one called hydrofracture, which is what I just described with the melting ice shelf. And another thing called marine ice cliff instability, which is like the ice sheet instability that has to do with ice that ends abruptly with the shelf not there anymore. And in their model, they found really large rates of sea level rise, close to a meter from Antarctica alone in the next century and as many as 12 meters by 2500 with a big collapse of West Antarctica and some retreat of East Antarctica. But there's a lot of speculative things in this model and some people maybe have taken the model more at face value than is appropriate at this point because the model is relatively, for example, has relatively coarse resolution and it has treatments of physical processes that are not very certain. So at this point, this is still very speculative, but people like me and others who model ice sheets, what we really want to do is model these processes more realistically and get a sense of whether this sort of scenario is plausible. So that brings me to the climate model I work on, which is with many other people in NCAR and at universities around the US and the world called the Community Earth System Model or CESM. CESM has been around in some form for more than 20 years. Its predecessor was called a Community Climate System Model and traditionally in a climate model, you have components for the land surface, the atmosphere, sea ice and the ocean, which are connected by what's called a coupler that exchanges fluxes in various fields between the components. And up until about 2010, ice sheets at least were not considered as dynamic, not moving objects in a climate model. So Greenland, for example, was treated in the predecessor of CESM as just a big bright rock reflecting a lot of radiation. But then there were these observations in the 90s and 2000s showing that ice sheets actually can change on time scales of a few decades that climate models were supposed to be concerned with. And so there was a demand for a new generation of models which are often called Earth System models, which are like climate models, but typically include some other things that weren't traditionally included in climate models among those being ice sheets, but also many other processes having to do with biology and chemistry. And so in 2010, we had the first version of CESM and we had a simple ice sheet model in that that I worked on. And then just this past summer, we had the release of CESM too with a much more realistic ice sheet model that I and others have spent several years working on. And another thing we introduced is interactive coupling between ice sheets and the land and atmosphere. And what that means in this case is that the land model computes what happens to snow and how much snow melts and forms. And so the land model is what tells the ice sheet model how much it's thinning or thickening at the surface. And then the ice sheet dynamics model will respond in either thicken or thin. And with the interactive capability, now the ice sheet model tells the land model, okay, your surface is now lower than it was because I've thinned. So then the land model can incorporate the feedback that comes from having a thinner ice sheet with a warmer atmosphere. So that's a more realistic kind of coupling than we used to have. However, it's very hard, both scientifically and technically, to couple ice sheets to the ocean. And that's something that we have not implemented yet, but something we're working on and hope to have in maybe CESM-3. So they're parts of the simulation that look pretty good for ice sheets. This panel shows Greenland surface mass balance, the difference between accumulation and melting. And the red areas are areas of net melting and the blue areas are areas of net accumulation. And the left-hand panel is from a regional model called RACMO, which is run at very high resolution and is very well validated or checked against observations. And it's the closest thing we have to the true surface mass balance of Greenland. This is in the center is what CESM-2 does, the model you can check out off the web. And you can see that for the most part, Greenland's pretty good. We capture these ablation zones and we have maximum snowfall here in the southeast. But the snow in the south is too diffuse and we have too much snow getting into the central south part of Greenland. And that's probably mostly because we don't, the model's too coarse to represent the mountains here. And so we have too much snow leaking into the interior. So that's called a model bias that we're always working to reduce that sort of bias. And then one of my colleagues recently sent me this plot, which shows a new developmental version of CESM-2 with a feature called Variable Resolution, which means you can take the atmosphere model and give it finer resolution. The grid cells are smaller over a particular place you're interested in, in this case Greenland. And in that case, you can resolve these mountains better and you find that there's not nearly as much leakage in the center. So we're hoping that in future versions of CESM, this will be a regular operational thing and that'll improve our Greenland simulation. In Antarctica, we have for a global model, which is relatively coarse resolution of grid cells about 50 or 100 kilometers on the side, a very good snowfall simulation, although sort of similar to Greenland, the red areas are areas of high snowfall. And if you compare it to the RACMO version to CESM, you can see there's a little bit too much snowfall leaking in from the coast instead of being concentrated very close to the coast. But overall, the structures look pretty good. And in particular, that we have pretty good snowfall in temperatures here in the ice shelves. So we think that if we project a warming that leads to melting on the surface of the ice shelves, that the climate model may be able to depict that process fairly realistically. On the ice sheet dynamic side, we have something called a community ice sheet model or CESM, which represents ice sheet flow. And these are some CESM simulations I've been working on for Greenland, where the left hand panel shows velocity observations, specifically the surface velocity that you can measure from aircraft or satellites. And then the right hand side shows the model velocity. And here are the various ice streams and outlet glaciers on the west coast, which are resolved very well by the model. And the slow flowing ice in the center is well modeled. One thing we're missing though, is this feature called the Northeast Greenland Ice Stream. And that's much too weak in the model. And we think that's because we don't have a good representation of the basal water, the basal hydrology. And so we don't have enough water here. So one thing we'd like to do is have a better basal water model and capture this particular feature. Over in Antarctica, that's harder than Greenland. And one reason it's harder is because the shape of Antarctica in a particular, particular how thick and extensive these ice shelves are, is very sensitive to how much melt you have coming from the ocean underneath the shelf. And we don't yet have models that give us accurate pictures of what the melt is going to be underneath an individual ice shelf. So observations on the left, the areas that are red are the big ice shelves that are flowing fast. And then the red and yellow regions upstream are the ice streams flowing into the big ice shelves. And here is results from a recent model simulation. And you can see the model in this case is capturing the observed flow really nicely. You know, all these individual ice streams, for example, they depend on the topography. So if you have a good representation of the topography, you can capture a lot of the structure in a model. But this particular simulation depends on adjusting the melt rates to give you shelf, ice shelf extents that are similar to what's observed. If we didn't have those melt rates, we would get a simulation that didn't look as good. So a big future goal for Antarctica would be to have models that can give you these more realistic basal melt rates so we can kind of predict the whole system without having to adjust things along the way. And one of the big things that people will be working on who are interested in ice sheets at NCAR and many other labs and universities over the next couple of years is a project called ISMIP-6, which stands for the Ice Sheet Model Intercomparison Project for CMIP-6. And CMIP-6 is the sixth go-around of the Climate Model Intercomparison Project, which is a big international exercise to compare all sorts of climate models in all different ways. There's maybe 20 or so global climate models that get together every few years to compare their results and try to figure out why results are similar or different. And this is the first time that ice sheets have been part of that process. And so there are three main parts of ISMIP-6. The first thing we wanna do is take a standard global model CESM that doesn't necessarily have dynamic ice sheets in it, but just look at all the model results that are relevant for ice sheets, like how much snowfall and melting there is in Greenland or Antarctica. And then in addition to that, we wanna take the output from these various climate models and use that to give us our best estimate of how much additional melting there's gonna be over the next century or so, both on the surface and beneath ice shelves and ice sheets. And so the idea would be you get your best possible view of the snowfall melting, and then you apply that to something called a standalone ice sheet model, which is just an ice sheet model that's not connected to a climate model, but is responding to the forcing that comes from a climate model. And then finally, we're gonna do a round of what we call a coupled ice sheet climate experiments where you can have that interactive coupling I was talking about earlier, where the climate changes the ice sheets, the ice sheets change the climate in return, and the two can feedback on each other and we can look at how much they feedback. And so we will use CESM on its own for this and SISM or ice sheet model for this, and then the two in combination for this. So our group and our collaborators will be involved in all aspects of this project. And that's about all I wanna say on the technical side of things. I wanted to finish up by presenting a few broader questions and my best attempt to answer those questions given what we know, but also given the various uncertainties. So something if you're a coastal planner and say you wanna build infrastructure that will endure for the next 50 or 75 years, you wanna know how much sea level to plan for. And you may be very risk averse and so you may particularly be interested in what's the worst thing that could happen, what's the most sea level rise we could see. And that's what the IPCC reports always try to do is give a best estimate in an upper bound. But in the case of sea level, if Antarctica doesn't do anything abrupt and catastrophic, we think that sea level rise by 2100 is probably not much greater than a meter and maybe a bit less, which is certainly significant and certainly something that planners need to take into account and could displace people. But the issue there is that it could be worse and it's hard at this point to quantify what the probability is that it could be worse than that. So that's something that we very much like to have a better handle on. Another question is just, can we get reliable sea level predictions from climate models and our system models? For several decades we've been using climate models to predict changes in temperature and precipitation, but we haven't typically, you haven't typically been able to take, look at climate model results and say, okay, this is how much my local sea level is gonna go up. And that's something we're moving toward. And I think in terms of predicting changes in snowfall and melting, the climate models are now pretty good. But as I've said, the models are not yet good at modeling interactions between ice sheets and oceans. So that's a part that we want to improve on. And, but hopefully by the next generation of climate models we'll be able to do better with that. And then what everyone wants to know is what's gonna happen in Greenland and Antarctica in the long run. And based on what we know from models and recent events, but also from records of past climates like the last interglacial, we think that with global average warming around one and a half to two degrees Celsius. And at this point, we're gonna see how we could have warming of less than that relative to pre-industrial times. The long-term sea level rise is likely to be around five meters, which is a lot. However, with temperatures of one and a half or two degrees Celsius, that sea level rise is likely to occur pretty slowly. So it wouldn't happen all at once. And you would have, let's say, several centuries over which these changes start to happen that with new technologies or different policies you might be able to curtail the temperature increase that would stop the sea level from rising that much before you get the irreversible changes. However, if you have warming of, let's say, three degrees Celsius, a little over five Fahrenheit or more, it seems from past climates that the Greenland ice sheet is probably not viable that most of it would melt. And the West Antarctic ice sheet would probably collapse at some point. And maybe you'd lose significant parts of East Antarctic as well. And so you'd get sea level rise in the long run of at least 12 meters and maybe more like 20. And that would be pretty well catastrophic if it would occur on human timescales. Again, these changes will take place relatively slowly, but the higher the temperature, the faster they happen. And the more likely you are to reach a point where at some point the changes over practical purposes are irreversible and will just unfold inexorably over the next several centuries. And so I think for planning purposes, you'd like to know where the thresholds are that still give you time to reverse if, as you learn, as you get new information. And the temperatures I've mentioned are quite relevant to the Paris Climate Agreement, which you've likely heard about. This was signed at the end of in Paris, at the end of 2015. This map came out about a year ago and it shows all the countries that ratified and or signed. And two countries at the time, Syria and Nicaragua, that had not signed. And one country, this one, that had signaled its intention to leave the Paris Climate Agreement. And since then, Syria and Nicaragua have signed. So it's just this one now. And, well, interesting that the way the agreement was written is that you can't withdraw instantly. Once you give your intent, you have to wait three years and then there's another year. And so it turns out that the earliest date the US can effectively withdraw is November 4th, 2020. And it just happens that there's an election on November 3rd, 2020. And so we may not have seen the last of this yet. See what happens. But in the terms of the Paris Climate Agreement, in some ways this was unprecedented because just about every country in the world made a commitment to stop global climate change at a level that would not cause catastrophic harm. And so the agreement is to limit global warming, quote, well below two degrees Celsius compared to pre-industrial levels, and also pursue efforts to keep that increased to 1.5. And it was kind of surprised that the Paris Agreement came up with 1.5 because that was considered by many people to be too ambitious. But a lot of the small island states in other countries that were existentially threatened by sea level rise said, no, two degrees would be the end of our country. It needs to be less than that. So that was taken as at least a goal. However, the commitments countries have actually made are likely to allow warming of at least three degrees Celsius. So if we were to avoid some of the more catastrophic impacts with regard to ice sheets, but also other parts of the climate system, three degrees is probably not enough to prevent some things from happening that would make the planet look quite different than it does today. So, but I don't want to leave on a hopeless note because 1.5 or two, I think there's a big difference between 1.5 or two, at least for ice sheets versus three. And I think it's significant that this many countries have made that a commitment. And then the question is how quickly countries will, how aggressively countries will pursue keeping those commitments. So to summarize, sea levels, it's a very interesting time to be a scientist studying ice sheets in sea level because we're learning a lot, but there's still a lot of uncertainty, especially with regard to Antarctica. And to work on Earth's system models, it's an exciting time because the models are getting better at simulating all these things. But as I've said, the simulating interactions between ice sheets and oceans is still in its early stages. And finally, what humans do in the next several decades will be very, very consequential for ice sheets. So I would say the long-term fate of ice sheets has not been determined. And that what we do, what people do, in the next 20, 30 years will have a lot to say about what happens in the long term. And thank you very much for listening. Thank you.