 Good afternoon and welcome to the meeting of the committee for the committee. We are pleased to present a 2D program on solid earth science and sea level change. Sea level rise is one of the most critical problems facing society today. We need to understand how sea level will change in the future and how it will impact coastal communities and infrastructure. The solid earth plays an important role in unraveling the evolution of sea level over a range of spatio-temporal scales. This meeting will review the state of sea level science, discuss some of the interactions with the solid earth, and explore a number of the important questions that we will still need to answer. Today, we will begin with introductory talks on sea level and then follow up with discussions of glacial isostatic adjustment and solid earth deformation. Tomorrow, we will have talks on vertical crustal motions as constrained from geodesy and relative sea level, followed by a general discussion with all of our speakers. I have a few announcements before we begin. This session is being recorded and will be available on our website within a few days. In addition to questions from our speakers and committee members, we plan to take questions from the audience. As noted in the previous slides, simply click the Q&A button on the bottom of your screen. Type your question in the box and click send. Any questions you submit may be read aloud and included in our video recording. In the interest of time, we will skip committee introductions, bios of our committee members are located on the academy's websites, and a link can be found in the chat. Importantly, I want to thank all of our speakers and the participants for taking the time to join us today and tomorrow, and I very much look forward to hearing from each of them. I will now turn over to Steve Narm, who will introduce our first two speakers and moderate our first discussion. Steve. Thank you, Torsten. As Torsten mentioned, we look forward to some very informative and interesting talks this afternoon. First, I'm pleased to introduce Ben Hamilton and Chris Pycuch, our two introductory speakers. Ben will go first. Ben is a research scientist at the Jet Propulsion Laboratory, where he focuses on studies of sea level change using satellite measurements. He's also been the team leader of the NASA sea level change team since 2017. Ben, take it away. Thanks, Steve. Let me share my screen real quick. First of all, I want to thank the organizers for inviting me to speak today. With this introductory talk, I'm hoping to accomplish two things. One is to give a framiest societal context to what we're discussing here today, and then also just give an overview of the different drivers of sea level change that are going to be discussed in the talks that follow. So a lot of times we think of sea level as a future problem. I mean, you hear that a lot when we're talking about plans that 2100, we're going to have a certain amount of flooding and coastal locations. But I want to try to reframe that a little bit and recognize that sea level rise is an ongoing and current problem. So this is a really nice graphic that's been put out by NOAA, and it really highlights the issue that we have now versus what we saw decades ago. So with increasing relative sea level rise over the last century, we've had a decrease in the available freeboard. So freeboard is really the safety gap between where high tide typically occurs and where flooding conditions happen in these coastal communities. So coastal communities were established with this freeboard in mind. They knew where high tide typically is, and they knew that they could build their infrastructure safely above high tide so you wouldn't experience flooding on regular time periods. But with increasing sea level over the past century, we've seen a significant reduction in the available freeboard and that safety gap. So this image on the left here shows sea level in 1950 versus sea level in 2010, just as an example. And you see how much closer that baseline of sea level is to the infrastructure where people are living. So because of that, this increase we've seen, we've seen an increase in flooding as a result. So there's two things, and this is really going to be the focus of both my talk and also the talks that follow. So there's the increase in sea level, which is obviously an important factor in terms of reducing this freeboard. But it's also the land is sinking. So it's this interplay between land and ocean that's really of critical importance when you're talking about the threat of coastal flooding. And there's two types of coastal flooding that occur. So there's the flooding associated with large storms. So you have, for instance, a hurricane coming off the coast and you get storm surge associated with that. But then you also have what we call high tide or sunny day flooding. So we call this sunny day flooding because it really doesn't take a storm. It doesn't take rainfall to drive this type of flooding. It's your high tide, potentially a anonymously high tide, coupled with some other kind of variability. So to say the winds are blowing in the right direction on a given day, this all combines and can lead to flooding conditions. And this type of flooding, this high tide flooding has really increased dramatically over the past few decades. And in some of these coastal communities, I'm focusing here on the United States, we're reaching a tipping point in terms of the coastal flooding that we're seeing. So in that panel A, in the middle of your screen here, these are the thresholds. So this is the gap, that freeboard gap between mean higher or high water. So you're kind of typical high tide. And then when you start to enter in the flooding conditions, this is in meters. And for some of the areas around the United States, we're now less than a foot of available freeboard. So it really doesn't take much to push from high tide into flooding conditions. And this panel on the right is a time series compiled by NOAA. So it really just tracks a number of high tide flood days we see per year around the US coastlines. And you can see this dramatic increase that's happened just in the past couple of decades. So this high tide flooding, it's a current, ongoing, and worsening issue. Now if we think about planners and decision makers and what they need at the coast, because of nearing this tipping point, a whole range of processes contributing to sea level change become very important. So it's not just the long-term sea level rise associated with ice melt or thermal expansion. Those are certainly critical pieces here. But all this other variability that happens at different time and space scales becomes very important. So planners are very concerned with this full sweep of processes contributing to sea level change. And we can, so this figure on the right shows some of these processes. And I'll cover some of these in the talk today and others will be covered in the talks that follow. But we can attach these particular processes in these time scales. So on the left and blue, I have days to weeks. So those are kind of your storms. And then on the right, those are your century scales. So your longer time period processes like thermal expansion and ice melt. But we can reframe this in terms of the decisions that need to be made on these time scales. And we can even attach dollar costs to this. So these processes and these time scales can be directly linked to decisions that need to be made. So on the bottom left, there's obviously the emergency response as a hurricane comes off the coast. For instance, there's damage that's done. You have to respond to that repair infrastructure or all that stuff. So certainly that's a big cost. But now you have this mid-range. So with the increase in high-type flooding, coastal communities need to plan their annual budgets, how they're going to respond to this high-type flooding from year to year. And then on longer time scales, you have these flood protection projects. How are you going to continue to successfully live in some of these coastal communities knowing that flooding is going to occur? So again, we have this... It's not just a scientific problem. It's not something that we're just interested in for the sake of scientific curiosity. There really is real-world current impacts into what we're discussing here, both in my talk and the ones that follow. So the remainder of my talk, I'm going to organize it, kind of how I laid it out there. So I want to discuss the processes that contribute to sea-level change on the largest scale, so global sea-level change, and then the processes that contribute to sea-level change on more regional levels. And you're going to see a version, or maybe this figure in particular, a version of this figure quite frequently throughout the talks in this meeting. It really does summarize the different processes and the interplay between these processes and how they contribute to relative sea-level change. So I'm going to go ahead and visit this and try to frame some of my talk based on this particular figure. So I am going to organize this largely by spatial scale in terms of global and regional scales. But as part of that, I'm going to address the different time scales. So there is, as I mentioned on the previous slide, other time scales of variability. There's inter-annual to decadal variability, which plays a big role in what we're seeing along the coastlines, in addition to the centennial and longer time scale, which I'm going to discuss more in the talks that follow. And I'm going to focus largely here on the Modern Observation Network. As Steve mentioned, I do a lot of work with satellite observations, so that's going to be my primary focus here. And I'm going to try to relay how we use those measurements to help us understand processes driving sea-level change. All right, so as I said, I'm going to focus on these global sea-level change processes. And there's two primary reasons that sea-level is changing on a global scale. One is ice-driven or barrier-static sea-level change. So this is really just the transfer of water between land and ocean. There's a couple ways this happens. So one, land ice from the ice sheets or glaciers melts, and then that melt water goes into the ocean, increasing just the mass of water that's in the ocean. The other is there can be changes in terrestrial water storage associated with the change in the global water cycle, so the movement of water between land and ocean and kind of annual time scales. Ground water withdrawals, water out of the ground, and then put that in the ocean. Certainly that's going to impact global sea-level rise. Then also water impoundments. So if we trap water on land, again, that'll impact global sea-level rise. And then the other mechanism for driving changes in global sea-level is thermal expansion. So thermosterec sea-level change really is just the ocean absorbing heat, and then that water expands, increasing the volume. So these, I've just highlighted them with a box in red on the right here. Again, I'm going to use this figure for some of the topics here. But I first want to focus on how we actually measure sea-level change. So there's these different processes, but how do we measure sea-level change in total? How do we know global sea-level is changing? So for the last century, we've used tide gauges. So I think Chris has talked that follows he'll go quite a bit into tide gauges and how we use those measurements to infer different things about the processes driving sea-level change. But on the right here, I just show an example of the San Francisco tide gauge sitting there off the coast. These tide gauges are necessarily located on land. So they're measuring relative sea-level change that interplay between subsidence or vertical land motion and what's happening with the ocean. And then these, again, since they're located on land, they have to be located along coastlines or on islands. And you can see the distribution of tide gauges in the bottom panel there. So we have global coverage, although it's certainly biased. You don't have a large number of observations in the interior of oceans. You need an island, as I said. And a lot of these observations are biased towards the northern hemisphere where you have more developed countries and cities. So there are challenges associated with combining these records, and there's been a lot of efforts to combine these records and estimate global mean sea-level over the length of the tide gauge record. I've listed a few of those there. That's not really the topic of the difference between what we had during the 20th century prior to the satellite records and what we have now. So now, since 1992, we can use satellite altimeters to measure sea surface height in the ocean. So on the bottom left, that's a satellite that's going to launch in just a couple weeks, or I guess a week from Saturday, the Sentinel-6 Michael Freilich satellite. But the important thing with satellite altimetry is it provides near-global coverage. Basically, the coverage of the altimeter over the course of 10 days. So every 10 days, it repeats this ground track. And from this, we can take all of these measurements across this ground track. We can average those together, and we can get a pretty good assessment of global mean sea level since 1992. So just one quick slide on how satellite altimetry works. So it's essentially a rangefinder. So the satellite altimeter sends a radar pulse down to the surface of the ocean. That pulse interacts with the surface of the ocean, and then bounces back and returns to the satellite. So the altimeter measures how long it takes for that pulse to get from the surface of the ocean and back to the satellite, knowing very precisely where the satellite is. From that, you can then get an estimate of sea surface height. So that's a really simple explanation. I'm leaving out some detail there, but that's really the core of how these altimeters are making measurements. And on the right there, I have kind of the constellation, satellite altimeters that NASA has launched since 1992, starting with the Totex Poseidon, and then working up to the Sentinel-6 Michael Freilich satellite, which is going to launch again next week, next Saturday. So as I said, from these measurements, we can take these, we can average them all together, and you can get an estimate of global mean sea level. And that's what I'm showing here. So this is the time series of global mean sea level rise during the altimeter time period. And from this, you can see, there's a couple of features you can note. One is that it's fairly linear, right? So, I mean, you have this increase. It's kind of irrefutable that sea level has increased from 1992 to 2020 to present on these global scales. That rate of rise is about 3.4 millimeters per year, or to put it in terms of inches, about 1.4 inches per decade. There have been studies that have pulled out an acceleration of this. So the rate at the beginning of the altimeter record in the first half is smaller than the rate in the second half. So there is an acceleration that's occurring in part of this record. In the altimeter, we have very good measurements of global mean sea level. We know very accurately and precisely how much sea level is rising during this time period. And just to provide one slide of context. So in blue, here is an average of the tide gauges from a recent paper from Thomas Fragerica. He took available tide gauges. He made an assessment of global mean sea level over the 20th century, looked at the contributions of some of the processes I'll talk about in the subsequent slides, and compared that to the satellite data. So just one thing to note here, the rate that we see during the altimeter record from 1993 to 2020 is almost three times what we saw over the full length of the 20th century. So we've seen this increase in sea level rise during the satellite era compared to what we had previously. But one thing to note, too, is that the tide gauges and the altimetry data do match pretty well during the overlapping time period. All right, so from these altimeters and the tide gauges, we know that global sea level changing is rising. And we can start to drill down a little bit into the processes contributing to this change. So to start with this first factor, so I'm going to talk about mass driven or very static sea level change. So this comes from the ice sheet melt from glaciers and then also groundwater withdrawal or changes in terrestrial water storage. They go into the ocean. So I'm going to focus first on this. And due to additions to the observing network in the past couple of decades, we actually have really good information and improving understanding of these mass driven sea level change, the mass driven sea level changes that have occurred in the past couple of decades. So we have ground stations. We have in situ measurements we can use to understand what's happening on the ice sheets and in glaciers. We have laser altimeters. So ISAT and ISAT2 are examples of those. And then we have what we call grab emitters. So this is GRACE and GRACE follow-on. And that's really what I'm going to focus on here. I'm not going to touch as much on the other two, just in the interest of time. So the GRACE satellites, the gravity recovery climate experiment. So the GRACE satellite was launched in 2002 and was collecting observations up through 2016. And in 2018, the GRACE follow-on satellite was launched and is up there now collecting information. But these GRACE and GRACE follow-on satellites give us information about gravity changes on Earth. Okay. So in the bottom left, each GRACE and GRACE follow-on satellite is made up of a pair of individual satellites. And what the measurement of these two satellites make is really just the distance between the two satellites. And on the figure here on the right, I'm showing a little bit about how this works. So one satellite follows the other. As they fly over the Earth, they respond to changes in the mass, the gravity below the satellites on Earth. So as an example here, you can see the first satellite flying up to this mass, this mountain, and it gets pulled towards it. And that separation between the two satellites increases. As the second one then comes closer towards that mountain, it gets pulled, and as the other one flies over, it gets pulled back as you can see in that third panel. And then as they both pass, again, those two satellites, there's an interplay between them. By tracking that distance very carefully, we can infer the mass, the gravity below those satellites on the surface of the Earth. And as these continue to orbit over time, we can start to get an idea of the mass change that occurs. So one of the, one really key observation these gray satellites give us, is gracing grace ball and satellites give us, is changes in the mass, the ice mass on these ice sheets. So here I'm showing the Greenland ice sheet and the mass change during the grace record. We not only know that the Greenland ice sheet is losing mass, we also know where that, that mass is being lost from on the Greenland ice sheet. So again, this is a really key observation that these gray satellites give us. I'm not going to let this run through entirely, but you can kind of get a feel for what's happening here. And then we can make an equivalent figure for the Antarctic ice sheet. This is not a video just showing the final result here. Again, we've lost mass over the time period, the grace and grace ball and have been, been measuring. And again, you can see where mass is being lost from. And then also, we can look at individual glaciers around the, around the world and assess how they are also losing mass, how they're changing over this time period. So here on the right is a figure showing a few of these different regions with glaciers and how they've lost mass as well. So again, this is all valuable information we can get from grace and grace follow-up. Based on this, we can then assess exactly how much ocean mass is changing. So instead of looking at how much mass is being lost on the ice sheets, I'm now looking at the kind of inverse of that. I'm looking at how much mass is being gained in the ocean. So these are, this is the change in ocean mass during the grace time period from grace or from grace and grace follow-up. You can see the rate of change from 2002 to present is about 2.1 millimeters per year. Okay. So now shifting gears to the second factor. So the other way that global sea level changes is from thermostatic, a thermostatic sea level change. So when the ocean absorbs heat, the water expands, increasing global sea levels in the same way. And we have good observations of this as well over the past couple of decades. So as opposed to satellites, we're now relying on in situ observations. So the Argo profiling floats have been measuring temperature and salinity of the ocean from zero to 2,000 meters below the surface of the ocean again since roughly 2005. So the way these work is they spend most of their time below the surface of the ocean. They drift around the global ocean. There's about three to 4,000 of these in the ocean. I'll show that in the next slide. But they collect this information about the water column from 2,000 meters every 10 days. They cycle back up to the surface. They transmit that information and then again to send back down to depth making these measurements. So from these measurements, we can take those temperature and salinity observations and we can estimate thermostatic global mean sea level. A couple caveats here. The floats don't measure between 2,000 meters over this full record. We really can't get an assessment of what's happening below 2,000 meters. And then some parts of the ocean are on samples. You're relying on these floats to freely drift and to enter different parts of the ocean. Some areas do not get sampled. So we can assess global mean sea level. Given those caveats again, it's not quite global mean sea level in terms of the thermostatic contribution. And if we look at the Argo measurements we have. So this is just the Argo measurements that have been obtained in the past three days. So this doesn't reflect the full extent of the Argo profile and floats in the ocean. Only the ones that have reported data in the last three days. So again, they cycle every 10 days. There's many more below the surface of the ocean. But you can get an idea of the global coverage, the roughly global coverage we have from these profiling floats. All right. So from that we can then take this data again, take all those observations. We can average it and we can get an assessment of the thermostatic global mean sea level change, global sea level change. And it ends up being about 1.1 millimeter per year over the past couple of decades from 2005 onwards. All right. So based on these observations, we can do what we call the global sea level budget. So this is a very simple computation. It allows us to check these observations we're getting from these three systems, see how well they agree with each other. So very roughly the Argo mean sea level plus grace and grace following mean sea level should roughly equal what we see with the altimetry. So thermostatic plus barrier static should roughly give us global sea level. Again, understanding some of the caveats I mentioned previously. And it turns out that works pretty well. Okay. So this, I know there's a lot going on in this figure, but the point here is that the very static sea level change in green and the thermostatic sea level change in orange, you add those two together, you get the yellow line and it matches up reasonably well with the altimetry data. Okay. So from these observations, we have a pretty good understanding of why sea level is changing on global scales and what's driving those changes. All right. And I want to shift gears. And this is going to start to relate a little bit more to the talks that are going to follow talking more specifically about regional relative sea level change. This regional relative sea level change is really the what's important for coastal planners and decision makers. It's certainly important what's happening in these global scales. It gives us a good indication of the state of the entire climate system and how it's changing. But in terms of a planning or an impact perspective, you're really interested in what happens at the local or regional level. So what are the processes that drive departures from the global mean on the regional level? So I've drawn boxes around some of these. So anything that's in orange there is considered a regional driver or that orange yellow color. I'm going to focus on three of these in particular. So large scale climate variability causes changes. When ice melts, it goes into the ocean and it doesn't fill the ocean like a bathtub. There's a certain fingerprint or pattern associated with that ice melt. And then also vertical land motion. It's partially reflective of what's happening in the ocean, but an equally important part is what's happening on land. So with the ultimate tree data, I showed it we have this global cover so we can get a first look at how different sea level changes on a regional scale versus what it is on a global scale. So here on the right, we can take that data and compute regional sea level trends over the past couple of decades. And there's a couple of things to note. One, it's not flat everywhere. So it's not unsurprising given the processes driving regional sea level change, but we don't see uniform increase across the global ocean. We do see patterns and we see relatively high rates in certain parts of the ocean and reduce rates in other parts. So one particular example that's been a really key feature in the altimeter record is the very high rate of sea level rise in the western tropical Pacific approaching a centimeter per year and then the relatively low rates that we see in the eastern Pacific, including along the sea level rise during the altimeter time period. So this sea level rise, it's partially related to these longer term or longer time scale features. Natural variability on shorter time scales also plays a role in this. So there's all these different things that play, these processes that lead us to this map. It's really key to understand as we try to convey to decision makers how sea level might change in the future. And then just one different way of looking at this map, this is a very specific video. Just to give an idea of the roughness and the different features we see in the ocean. So this is the altimeter data gridded up, a monthly map just showing the change for month to month. One thing you'll see in just a second, popping up is the El Nino southern oscillation, a really big El Nino that occurred right here in the western tropical Pacific or when you fit the trends, it's certainly smooth some of this stuff out, but there's these different time scales and spatial scales of variability in the ocean that we have to be concerned with. All right, so just for one slide on interannual to decadal sea level variability, so there's large scale climate variability that drives changes in sea level on both global and regional scales. And I'll talk about how we see these signals on global scales in a second. But one in particular here, so in particular time periods in the satellite record December of 1997 in December of 2015 and both of these were marked by very strong El Ninos that occurred during the the altimeter record. And what we see during these time periods is that regional sea level can change by up to 25 centimeters greater than 25 centimeters on a year to year basis. So huge amounts of sea level change and what's noteworthy here is that coastal regions are also feeling this feeling this so it's not just an open ocean instance and off the west coast and the west coast in South America you see high rates of or high levels of sea level associated with these signals. So they can really drive changes on short time periods impact flooding in these locations again not associated with long-term climate change but it's this interannual variability which can really be a concern to decision makers. This is one particular signal but there are other natural climate signals and different ocean basins. I've just listed a few here so the North Atlantic Oscillation, the NAO, the Pacific Decadal Oscillation which is again centered here in the Pacific and impacts the northeast Pacific. The Indian Ocean Dipole which impacts the Indian Ocean but it's connected to Enso and in some different ways but there are these large scale dominant signals that impact sea level on these shorter time scales and again given that we're close to that tipping point and coastal communities these types of signals are actually critically important and to understand how they're going to combine with other sea level processes in the future. And just one slide here so we actually do see these signals show up and global means sea level so all that has been done here I showed you that nice linear relatively straight increase of global means sea level all I've done here is remove the trend so that blue line is global means sea level without the trend so you do see these variations about the trend and a lot of these are associated with large scale climate signals like El Nino. So I've highlighted a couple areas of circle the couple areas in green there so the 1992 1982 1983 there was a pretty big El Nino 1997 1998 again another El Nino in 2010 2011 there's a very strong La Nino so that's the opposite phase of from El Nino where you have decreasing sea level in the eastern Pacific and increasing sea level in the western Pacific and it turns out that these signals these climate signals drive precipitation patterns in the movement of water between land and ocean it can impact global means sea level as well as on these regional scales all right so just to touch on a couple more and these are topics that are going to be covered in much more detail in the subsequent talks so when ice melts or when we move water from land into ocean again it doesn't the ocean doesn't fill up like a bathtub right so there's these fingerprints that represent where that water actually goes into the ocean once it melts from these different sources so each source has a different associated fingerprint or pattern associated with it and you can see here I pulled out four particular locations so the Antarctic ice sheet in the upper left Greenland ice sheet in upper right changes in terrestrial water storage in the lower left and then changes in glaciers in the lower right and you can see where ice is being lost you actually have a drop in relative sea level so the gravitational pull of that particular area is reduced and actually water is shifted away from it on the other hand and far field from these sources you see an increase or a relatively large increase and relative sea level associated with this this ice mass loss so as we consider melting ice and the transfer of water from land into the ocean it's really important to consider how that water is going to move about the ocean and what areas what coastal communities will be most affected by mass loss from that particular source and again there's going to be talks that discuss this in more detail and then the last topic here I wanted to cover is vertical land motion so again I've made the point that we're concerned about the ocean clearly but we're also concerned about what's happening on land in terms of the movement of land relative to the ocean so vertical land motion is driven by large scale changes so what we call glacial isotatic adjustment and then also in smaller scales by effects like groundwater withdrawal so when you pump water out from a particular location then that area will respond by sinking or subsiding we do have observations of these in the past couple decades so one way that we observe them is through GPS so this is just showing the GPS coverage and the GPS stations maintained with the data collected by UNR the University of Nevada Reno on their website so you can see kind of the coverage of GPS across the United States and then from this we can actually then assess vertical land motion using what what's called the GPS imaging technique which I think Bill Hammett will speak a little bit more about tomorrow but you take those observations and you can create this map and showing exactly where we see subsidence versus uplift and can start to infer how coastal communities might be impacted by the subsidence so one noteworthy thing from this in particular if you look at the east coast of the United States you see broad scale subsidence approaching on the order of 2 to 5 millimeters per year so not a small contribution from subsidence if you look at the west coast conversely you do see some other features where some areas of uplift associated with some different signals but again with these observations we can get a pretty good idea of what's happening on these broad scales and then also GIA contributions to relative sea level change occur over a longer time period so these observations here are reflective of and part of GIA changes and these have to be modeled and available observations can be leveraged to improve those model estimates again that's going to be talked about in some of the subsequent talks all right one final point so for coastal planners these broad scale assessments of vertical land motion are certainly important but a lot of the adaptation efforts that are being done happen at a very local level so maybe you're going to raise a particular house protect a neighborhood raise a street issues like that so information about subsidence and vertical land motion at these higher resolutions these smaller spatial scales become critically important and one way we can get at that is through interferometric synthetic aperture radar in SAR analysis and Manu shows I tomorrow is going to be talking about this in more detail I just want to point to it a little bit that it's not just these large scales when we're talking about subsidence that matter the smaller scales are also critically important so this in SAR analysis allows us to get at those smaller spatial scales and make assessments closer to the street or community level that again are critically important to planners and decision makers. All right so just to summarize and point to where we're going with the rest of this meeting so I really just wanted to highlight today that relative sea level change it varies on a wide range of spatial and temporal scales so in terms of a scientific problem it's a really interesting problem how these different pieces fit together how these time scales and spatial scales and the processes that vary upon them how they interact and combine it's a really important question an interesting question scientifically but it's also really important to keep in mind that this matters from a societal perspective so we've reached these tipping points flooding as occurring it's ongoing it's worsening these different timescales in the interplay becomes really important from a decision making perspective the modern observation network is really valuable it's really increased our understanding dramatically improved our understanding of sea level change over the past couple decades one thing to note is that these records are still relatively important so continuity is key in this so the launch of Sentinel-6 Michael Freilich for instance that's going to continue this altimeter record that's critically important but also these comparisons to past sea level records are also very important you'll hear some of that in the remaining presentations it's not just about what we see during the altimeter record it's how the altimeter record is different or has changed over the past past few decades or even the talks that are going to follow are going to provide detail looks into some of the processes I covered they'll discuss this comparison between past and present observations and really start to highlight some of the key questions that scientists are grappling with that really impact our understanding of future sea level change and with that I will finish up thanks a lot Ben that was great we're going to take a few questions for Ben committee members please raise your virtual blue hand and state your name before you ask your question any questions from committee members well I have a question from the audience Ben so why don't we start with a question from the audience so the question is how sensitive a satellite altimeter is changes in the orbit how was that corrected and how frequently and they relate that to part of the global temperature record controversy involves the adequacy of similar corrections right so part of a big driver of the accuracy of our altimeter measurements is our understanding of the orbits and that's actually a lot of the improvements we've seen from Topex Poseidon on up through the current altimeters is associated with that improved understanding of the orbits trying to sorry I kind of forgot the flow of the question but I mean I think I would say that our uncertainty in the altimeter measurements is our understanding and our uncertainty is our understanding of the orbits and that's factored in when we try to make these assessments so again our the estimates of global mean sea level our accuracy in terms of that 3.1 millimeters per year is about 0.4 millimeter per year so again we know it pretty well part of that uncertainty comes from the orbits but yeah I think a simple answer is that we know the orbits well enough to make the inferences that that that I showed in my presentation yeah thanks Ben that was a great talk I'm super interesting I was curious in the regional maps that you were showing there's also a lot of high frequency spatial variability is that real or is that an artifact of the processing I was because I mean it looked like there were pretty large magnitude variations over maybe a hundred kilometers or even shorter wavelengths yep so so that is real variability when I showed the trend map so some of that variability that high frequency variability gets aliased in so I showed the trend map which you shouldn't trust too much in terms of the higher frequency those really small features but when I showed that animation and all the movement of those smaller scale features that's real variability certainly you have to keep in mind the gridding will do certain things to it but there are these what we call mesoscale variability in the ocean these eddies moving around which really play into the ocean so in the altenders are very useful and are understanding those scales too there are future satellites so the surface water and ocean topography satellite the swat satellite is focused more is going to be focused more directly on those smaller features on the mesoscale which will provide us even more information but you have short answer is those are that's real variability in the ocean great thanks alright thanks Ben folks I'll be moving on to questions later we need to move on to the next talk so I'd like to introduce Chris Pipewch Chris is a physical oceanographer at the Woods Hole oceanographic institution where he focuses on the physics and statistics of coastal and regional sea level variability and change so Chris why don't you take it away thanks Steve let me share my screen one second I'd like to share my point slide now again hi everyone I'd like to thank Steve and everyone else on the committee for giving me the opportunity to speak to you today I also want to thank Ben for his great introductory talk that segues nicely into what I'll be talking about so the theme of this meeting the next two days is sea level in the solider and in my talk I'm actually not going to talk too much about advances in solider geophysics I'm not a geophysicist I'm an expert in climate change so my time today is use my presentation as a motivation I want to motivate the talks that will come the four talks that you'll hear later today and tomorrow to show how with observations of sea level change coupled with knowledge of solid earth geophysics we can actually improve our understanding of ocean circulation and climate change so as Ben promised here's this figure again so it's a diagram from the report on oceans and cryosphere and again just to remind you what we're illustrating here is the myriad earth system processes that depending on spatial scale here is another version same figure this is a figure from Ben Horton and colleagues a couple of years ago that's trying to get across this same idea that depending on space scale and time scale there are any number of processes again involving an interplay of the climate system the ocean and solid earth geophysics that contribute to sea level change so on the left-hand side Horton and colleagues are identifying different processes again geophysical oceanographic and climate exactly those same processes that were indicated in my previous slide and they're indicating how much sea level change that given process can contribute to depending on spatial scale which is indicated by the colors to the left global and blue and local and regional to the in gray time scale is on the X axis on the logarithmic scale and the color shading indicates the rough order of magnitude of sea level change associated with that process so for example if you look at ocean dynamics which is the first regional process in gray and you look at the annual time scale so 10 to the 0 years you see that this process can cause on the order of decimeters to a meter of sea level change there's a lot of details here that really draw your attention to two main points if we sort of take our eyes and go horizontally at any given process we see that again depending on spatial scale and time scale any given process can be more or less relevant to sea level change that's one point the other point is if we take more of a vertical view and so go vertically and pick a particular time scale we see that sea level changes if we imagine we have observations of sea level changes often those changes don't represent the effect of a single cause rather they're the effect of multiple causes okay so again what we're seeing is this interplay of geophysics and oceanography and climate science all imprinted on sea level observations so here's my attempt to sort of simplify the previous two slides so this is my version drastically simplify for my purposes of what the previous two slides were getting across so again in this meeting we're going to largely focus on relative sea level and its relationship to solid earth geophysics what I'm going to drive home today is that again if we have observations of relative sea level change and we have good knowledge of solid earth geophysics and that informs our knowledge of ocean circulation and climate change so just to wrap up I know there were a few hiccups in interruption so just to make sure we're all tracking here here's the summary so far I mean the premise of this meeting is that relative sea level data are informative of solid earth geophysical processes things like glacial isaesthetic adjustment interior earth structure etc and my thesis here stated in the middle is that by combining relative sea level data and understanding of solid earth geophysics we can actually make progress towards understanding ocean circulation and climate change so today I'm going to summarize two things I'm going to look at past studies that use relative sea level data either to infer changes in ocean climate in the form of global mean sea level changes or ocean circulation and I'm going to pay particular attention to the extent to which solid earth geophysical knowledge is or is not incorporated so we'll start with global mean sea level again Ben provided me a really nice introduction here so I don't have to do as much explaining as I was planning on but suffice it to say global mean sea level or what I'll call GMSL for short it's an essential climate variable as Ben described global mean sea level changes can arise from one of two causes either veraesthetic changes that is you're adding or removing mass to the global oceans or thermostatic changes you're expanding or contracting ocean water by warming it or cooling it and so because GMSL changes arise from these two causes GMSL is an important index of land ice wastage the hydrological cycle and earth's energy imbalance so because GMSL is both practically and theoretically interesting it was a long history of estimates trying to infer changes in global mean sea level that date back to the 1940s actually and all these studies they use the network of global tide gauges so here I'm illustrating the global tide gauge network again global ocean here each one of these circles indicates the location of a tide gauge and the relative size of the circles indicates how long the record is so bigger circles go with longer records and the color shading is the rate of relative sea level change simply taking the available data at each site and fit a linear trend so the yellows are indicating more sea level rise and the blues are indicating sea level fall and so all of these studies going back to the very beginning the early studies by Gutenberg in 1941 for example all these studies recognize two things from the start one something that should be obvious looking at this plot is that relative sea level data from tide gauges are spatially and temporally inhomogeneous they're heterogeneous for example if we look at Western Europe or North America we see a lot of dots and a lot of big dots meaning that there are a lot of long tide gauge sea level records in these regions contrast that with looking at places like the South American coast and parts of the African coast where you see fewer dots and smaller dots and sometimes no dots okay so studies from the very beginning recognize that to estimate global mean sea level one has to cope with this spatiotemporal heterogeneity of the data the other thing that studies recognize right from the beginning again going back to Gutenberg in before is that clearly these relative sea level records are influenced by solid earth geophysical processes that are entirely unrelated to global mean sea level changes this is perhaps most clearly evidenced if you look at the deep blue colors around Scandinavia which are indicating rates of sea level fall on longtime scales associated with post-glacial rebound and the glacial isostatic adjustment process so it was recognized right away that these solid earth geophysical effects would have to be dealt with they are the signal to the global mean excuse me they are the noise of the global mean sea level signal so what I thought would be useful is to summarize past studies in terms of how they dealt with this geophysical signal so early studies again starting with Gutenberg in the 40s but going up to Eugenia Lisitsin in the 1970s and even some studies today the strategy is to avoid tide gauges from regions that are thought to be dominated by solid earth geophysical effects so for example those tide gauges from Scandinavia would not be used for example and this is how many authors proceeded for many decades until about the 1980s when folks like Vivian Gornitz and Ian Shennan started to not necessarily avoid data from regions that were affected by geophysical effects but rather paleoproxy records of long term geological rates and correct the tide gauge records the premise being if you estimate the long term geological rates and subtract them from the tide gauges then what's left over is more closely related to global mean sea level change starting around the 1990s you start seeing the availability of global glacial isostatic adjustment models so similar to those practices in the 80s you have folks like Vic Peltier and others that will use models of the glacial isostatic adjustment process again to correct those tide gauges under the assumption that the models accurately simulate the glacial isostatic adjustment process and that GIA is really the most important geophysical effect on the tide gauges at the turn of the century in the 2000s you have a couple important developments one is that folks like Jeremy Trafica and others started to include not only models of the GIA process but other solid earth geophysical effects for example the spatial patterns that Ben alluded to a few minutes ago related to the elastic fingerprints spatial patterns of sea level change related to contemporary ice melt this is really nicely demonstrated in the somewhat recent paper by Carling Hay in 2015 that's one development in the 2000s another is that by the 2000s we've started to collect sufficiently long records of crustal motion from global positioning system data GPS data and so the idea here is that removing land motion effect as measured by GPS from the tide gauges essentially taking a relative sea level measurement and putting it in the geocentric frame so early studies by Guy Wobelmann and others and more recently exemplified by Sonke Dangendorf in 2017 demonstrate this approach and as we move into the last decade what we've seen is a more and more comprehensive accounting of solid earth geophysical effects and more comprehensive uncertainty quantification so for example studies like Hay et al 2015 also studies more recently by Thomas Fredericks of include uncertainty in GIA models as well as those elastic fingerprints so here's I'm trying to on one plot summarize all the past literature on post industrial or industrial era global mean sea level trend so let me explain this plot a little bit each box corresponds to a different study the excuse me the vertical thickness of the bars indicates that the error bar of any given global mean sea level trend where we have rate on the on the y-axis and the x-axis is essentially the period over which that trend is computed and finally the color of the box indicates what year the estimate was published and long story short the history of global mean sea level estimates is typified by an increasing amount of solid earth geophysical knowledge being brought to bear and consequently these estimates grow more consistent and less uncertain such that in the beginning estimates were kind of all over the place and range anywhere from basically no rise to up to three millimeters per year of rise but over time over the last decade we've converged to estimates against the 20th century rate to be between one and two millimeters per year and in fact in the past five years we've really been converging to these lower numbers between one and one point five millimeters per year so we can have estimates in recent times not only that not only do we know rates of change over the 20th century now we're able to also independently attribute the causes so Ben made allusion to the sea level budget he was talking about it in the context of satellite and modern measurements we can also now do this over a longer time scale so here are some results from a recent study by Thomas Frederick essentially resolving what was known a couple of decades ago as monks enigma saying that we were unable to attribute the causes of past sea level rise well Frederick said here exemplifies how we can do that nowadays so there's a lot going on here let's look at panel a for example a couple of different curves here the blue curve and panel a is Frederick said at all's estimate of 20th century sea level change on an annual basis the black curve is essentially the sum of all the causes remember the thermostatic and barostatic contributions that Ben mentioned in fact those individual barostatic and thermostatic effects are given in red and orange here so again the point is that not only can we diagnose these changes we can also diagnose the causes not only that we can also identify acceleration in sea level rise for example as discussed earlier by Sonka Dungendorf last year and again more recently discussed by Thomas Frederick said if we look at panel C we're looking now at global mean sea level trends that's the rate of change again blue is the total change whereas black is the sum of the contributions and red and orange are the barostatic and thermostatic respectively I want you to look in panel C at the blue curve from 1960 to about 2010 what we see is sort of a linear increase from about the 60s to the modern and again these are trends so it's an increase in the trend or an acceleration in global mean sea level okay so we can identify these higher order features of the past sea level record so long story short if you come away with anything here is that we have a really good understanding of 20th century global mean sea level change and its causes so so what's next well I would suggest we start looking at a longer geological time scale specifically looking at the late Holocene or the common era why do I highlight that that time theory will a couple reasons one is it performed it provides a an informative sort of longer term geological context within which to put modern rates of sea level change the other is a bit pragmatic there's a relative abundance of proxy sea level records with things like salt marsh sediments micro atolls et cetera so here are a few of the only available estimates of common era global mean sea level change the earlier estimate by Iceland grisken and others is shown in yellow is actually based on it is a temperature based estimate actually the orange curve from Andy Kemp from 2011 this is based on salt marsh sediment from North Carolina essentially in that study the authors removed an estimate of GIA from a model and essentially assumed that what was left over was global mean sea level arguably the most comprehensive and authoritative estimate we have of common era global mean sea level is the blue curve from Bob Kopp and colleagues and what this represents is a probabilistic simulation of a global database of high resolution common era proxies and what's really interesting about the Kopp at all curve is that of course if we look far to the right to the last century we see the uptick of the modern acceleration in rates of global mean sea level but what's really really interesting in the broader common era context is that there are all these bumps and wiggles in the pre-industrial common era that is there is prominent centennial and millennial scale variability that we actually don't know the causes of so I'm identifying this as a potential area to try to study more what are the causes of those centennial to millennial scale fluctuations. Kopp et al show that there is a relationship to surface temperature but at this point it's unclear to what extent these represent thermostatic or barostatic changes or to what extent they're driven by greenhouse gases or solar variability or volcanic activity for example there are also other areas where solid or geophysical knowledge can be brought to bear to make better and more robust these estimates for example one caveat of the Kopp et al study is this issue of identifiability they were unable to unambiguously identify the long-term rate of pre-industrial change and they had to assume that global mean sea level rate between zero and 1800 common era was zero and they acknowledged that that assumption so with better geophysical knowledge we can perhaps lose that assumption and have better knowledge of how sea level changed on millennial time scales in the pre-industrial era and finally we want to improve the representation of solid or geophysics in these models so the way a lot of probabilistic algorithms work is that you assimilate sort of sparse data and you share that information across space and time according to specified covariance functions so I'm giving an example here of two covariance functions related to the GIA process at New York City where I'm showing the star here so on the left is the covariance function used in the Kopp et al study and you see this sort of yellow bullseye right around the New York region you can enter saying measurement at New York how far a field does my probabilistic algorithm allow me to share that information and so in the Kopp study it's a relatively local neighborhood of New York City okay so compare that assumed structure to the structure on the right so this is the essentially the era covariance structure of the GIA process at New York City with all other sites from a suite of GIA models where we're varying the lithosphere thickness and the mantle viscosity and the ice history okay and we see a very different structure here for example we see far field regions of strong covariance for example if we look at Northern Canada or even over just north of Scandinavia we also see anti covariance so these structures are very different they're important because these structures are determining how information is shared across space and time in these models so we want to move to more realistic representation of geophysical models so point here is identifying some areas where climate scientists and solid earth geophysicists could work more closely to advance our knowledge of common era global sea level so about halfway through now let me switch gears to sea level and ocean circulation I'm going to sort of shamelessly focus just in the North Atlantic here so large scale North Atlantic ocean circulation plays an important role in climate variability and abrupt change carbon cycling ecosystem productivity and a bunch of other fields so suffice it to say there's a lot of interest in knowing how North Atlantic ocean circulation has changed over time but there's a problem and that is that measuring the circulation is both difficult and expensive so past changes in climate relevant features of the circulation things like the overturning circulation they're not very well constrained and they represent a key uncertainty related to climate change just so everyone's on the same page I know not everyone is an oceanographer here I'm sort of sketching a little cartoon of the North Atlantic circulation so all these arrows are indicating different different circulations I'm just noticing right now that I've indicated that Greenland has three E's apologies for that anyway the red arrows indicate sort of warm surface currents and what we see if we follow those red arrows on the bottom we see this arcing clockwise motion which traces out the subtropical gyre which involves things like the Gulf Stream and the North Atlantic current and is predominantly driven by the surface winds but we do see that some of those red arrows travel further north bringing warm salty water to the northern North Atlantic where they encounter harsh conditions they densify they cool and they sink to depth and return a equator word and that's indicated by the yellow arrows and the blue arrows so that's again a cartoon picture of this complicated circulation what do our observations look like well here's a schematic of the modern day Atlantic circulation observing system so to the left is a map and all these different colored lines show the nominal locations of different contemporary observing arrays this is summarized from a paper by Eleanor Fraka Williams last year on the right I'm showing some of the data from those from those sites and so we have time on the x-axis on the bottom and transports of ocean circulation transport of volumetric rate indicated by spare drips on the various y-axis this is I should emphasize these data are a remarkable achievement in the context of ocean observing however if we're interested in long-term change these records are just not long enough so what do we do where does sea level come in well the basic fluid dynamic equations of oceanic motion suggest that changes in large-scale features of the circulation things like the vertical overturning or the horizontal gyre as I just mentioned it turns out that that changes in these circulation features are in fact manifested in and coupled to changes in sea level along the western boundary so if we're interested in the north Atlantic ocean our western boundary here is the east coast of North America I'm showing some kind of cartoony equations here that try to illustrate this so this Psi VO this is the transport of the vertical overturning circulation you can demonstrate based on simple conservation principles this should be directly related to zeta w which is sea level on the western boundary this other cartoony equation Psi Hg this is the transport of the horizontal gyre again through basic physics you can show that this should be related to the the latitudinal structure d by dy or y is latitude of sea level along the coast divided by f where f is the Coriolis parameter a function of latitude so again the idea here is that through basic physics these large scale features are coupled to what's happening at the coast none of this is new I don't want to claim any novelty here this has been recognized for almost a century and in fact there are studies dating back to the 1930s that try to exploit these physics to use tide gauge records to say something about ocean circulation and particularly the Gulfstream again going back to the early studies by Montgomery in the 1930s and in fact not just the Gulfstream a lot of these papers focus on the Gulfstream at its origin at the Florida Straits otherwise known as the Florida Current I'm trying to illustrate that region here in these maps so on the left we have the western north Atlantic and the color shading here is surface ocean currents so blues are more sort of slower more sluggish currents whereas yellows are more rapid swift currents and what you see here is that you have a whole sea of blue except starting around Cuba and southern Florida and the Bahamas you see the emergence of this bright yellow snaking band again right below Florida is the origin of the Gulfstream so I'm zooming in there with that red box and showing the region to the right again this is a zoom in of southern Florida and Cuba and the Bahamas all of the black and red lines across the ocean here are the locations of historical measurements I should clarify also that the color shading now indicates ocean depth not speed so I'm trying to exemplify that this is a very narrow shallow region the Gulfstream passes through here the historical direct ocean transport data are sparse but you'll notice in both figures here I have all these gray squares and gray circles these indicate the locations of long tide gauge records so people have exploited the fact that the Gulfstream here is so trapped by the topography so constrained by ocean depth that we can use coastal sea level measurements and so again most of those studies I cited in the previous slide do that one thing I should clarify remember my earliest slides where I was showing either this map from the SROC or this figure from Ben Horton's figure at different time scales different processes are responsible for sea level change so with that knowledge in mind physical oceanographers have made assumptions when they've analyzed these tide gauge records specifically they assume that at the high frequencies anything you're measuring by a tide gauge is related to oceanographic processes whereas at the low frequencies this is all solid or geophysics so they've done one of two things they've typically removed the trend from the tide gauge record again assuming that that's largely due to things like isostatic adjustment however that causes two problems one is that it implicitly assumes that once you remove the trend you no longer have solid or geophysical effects in your tide gauge record which is not strictly speaking true it also precludes you from looking at possible ocean circulation changes on the longest time scales because you just removed the trend but again this sort of procedure is really typical of most studies here are two examples again so what we're doing is we're looking at tide gauge records in the southern Florida or Bahamas region versus transport so on the top right this is results from an earlier paper by Iceland and he is using sea level data from a few tide gauges along the southeast and note that sea level was on the y-axis but he's inverted the axis so that positive values are to the bottom and negative values are to the top this is because sea levels inversely correlated with transport and he's interpreting these sea level changes during the 20s and 30s in terms of changes in the Gulfstream so if you look at this time series in first that there was a strong drop in the Gulfstream transport at around 1931 to 1933 so that's the inference he draws from the tide gauges but there were no concomitant transport measurements to corroborate that hypothesis on the bottom left is a more recent paper by Malin colleagues from mid-80s this is from the Stax experiment kind of a busy plot but bear with me here so we've got about a year and a half of data from the early 1980s data both of transport measurements transport shown in volumetric rate on the left y-axis and sea level on the right y-axis and so the thicker lines are the transport measurements thinner and dashed lines are the sea level measurements and I should clarify the sea level measurements are the sea level difference sea level in the Bahamas minus sea level on the Florida coast this is what you expect from geostrophic balance that the fundamental momentum balance in the ocean point being is that all these various curves are correlated they're correlated and this suggests that you might be able to use coastal tide gauge records as a low cost system for ocean circulation but again all these studies are looking at the high frequencies just this past summer I published a study where I try to tackle this problem head on in other words I try to tackle the problem of trends so I used all those tide gauges I indicated in the previous map slide along with available transport measurements to try to estimate Florida current transport changes try to estimate Gulfstream transport changes over the past 110 years and here here's the result so this is my estimate of transport again time on the x-axis transport in a volumetric rate on the y-axis the thick line is the median estimate the shading and dashes are different measurements of uncertainty I won't go too much into the details and the solid thin lines are two members of the ensemble so ensemble based estimate of past transport change and the key point here is again remember I haven't removed the trends from the tide gauges and I'm identifying a likely decline in the Florida current over the past century and this is consistent with independent studies based on other proxies that the larger scale overturning circulation has weakened over that same time now crucial to my estimate was that again I'm looking at the trends in sea level and trying to distinguish what component of those trends is due to solid earth geophysics things like tectomics or glacial isostatic adjustment and how much is due to ocean dynamics and the relevant quantity here is the trend in the sea level difference across the Florida you can remember I'm looking at the difference sea level in the Bahamas minus sea level in Florida this is the relevant quantity for inferring circulation so I'm showing here in this bar graph different estimates of that sea level difference trend again the difference across the straits and the trend in time anything that's labeled as modeled is something that I've estimated in my study and observed is an independent data set that I've held in reserve and not included in my probabilistic estimate and what I want to show here so again rate is on the bottom the horizontal thickness is a measure of uncertainty and the vertical thickness has no significance it's just so you can see it so if we start at the top at the modeled relative sea level trend again this is the difference across Florida straits we see that it overlaps zero there's not much of a significant change in the sea level difference over the study periods however that sort of zero trend actually is the canceling out of two different processes one there's a positive trend across the straits that is sea level rising higher in the Bahamas due to what I'll call isostatic processes this is where your solid geophysical effects would come in that this is the blue box on the second row and there's a negative trend in ocean dynamic sea level this is on the fifth row down in orange so it's this canceling out and in fact if you compare to independent data they corroborate this so if we look at the second third and fourth rows again I have my modeled isostatic rate I'm comparing that to the rate of sea level rise invert from GPS data and also the the rate of long-term change inferred from proxy records from mangroves and salt marshes and admittedly there's some uncertainty here but the qualitative story that there is that there is a positive trend is consistent likewise if we look at the bottom two rows again the orange is my modeled trend in dynamic sea level the yellow is what you get from satellite altimetry which although it's over a shorter period gives the same sense the same sign of trends so again the the important point here is I've tried it for circulation changes on the longest time scales so this is encouraging that encourages us to do a few things one is again my record only goes back to the early 1900s due to the availability of tide changes we should go back further in time a few ways we can do this one is to recover old archival records such as exemplified by chef and talker another is to use a high-resolution paleo proxies as I hinted to a moment ago for example from salt marshes or coral micro-tolls and this would allow us to test an earlier hypothesis based on Delta 018s of oxygen isotopes that the Florida current strengthens coming out of the little ice age spent a lot of time talking about the Gulf Stream but there's nothing special about that in the sense that we can do this kind of study elsewhere and so I would encourage using long tide gauge records elsewhere to try to estimate other ocean currents Kristen Richter provides a nice example in the case of the Atlantic inflows of the Nordic seas over the past 50 or 60 years the caveat in that study is that she and her colleague did remove that linear trend so I think that this issue should be revisited in that region but trying to deal with that linear trend issue and of course there are long tide gauges elsewhere what I think is particularly salient in the case of a meeting on solid Earth geophysics are these long records in places like Japan and New Zealand which potentially can tell us about past changes in Pacific ocean circulation of course one has to contend with things like tectonics which presents an interesting challenge and finally it'd be good to explore more generally these growing global databases of high-resolution proxy records so I've gone over time a little bit and I apologize for that sorry for the initial hiccups but I'll quickly summarize so again I hope I've convinced you of the following one is that relative sea level records have motivated fruitful collaborations between solid Earth geophysicists physical oceanographers and climate scientists and the history over the past 80 years of global mean sea level study shows us that as we improve our understanding of solid Earth geophysics and incorporate that in the GMSL studies we improve and make more consistent our estimates of past global mean sea level change and I'm suggesting that those global mean sea level changes are a little bit different from what we've seen in the past we've seen in the past where there hasn't been as much collaboration and progress towards estimating past circulation changes based on sea level records so it's an encouragement to move forward there so finally to sort of point to the four talks that you'll see in a little bit what do we need to make progress on these processes and problems that are identified with the moon circulation and climate geophysical models and others we need to improve the spatial and temporal resolution and scale that we can resolve we also need to improve our understanding of interior Earth structure and ice history which we'll hear more about this afternoon from Pippa and Jackie also really important part of the reason why GIA has featured so prominently in the history of sea level studies is that by and large it's been the one solid Earth geophysical study for which we have models of other processes and feedbacks between them and finally as we have more models more data we need to also develop in tandem probabilistic frameworks that can assimilate in a mathematically coherent fashion all these disparate data streams and finally in addition to the models of course we need to grow the observations in space and time things like Ben has hinted to and we'll hear more about tomorrow we need to grow in lengthen the GPS network we need to start exploiting and of course keep growing these global paleoceanographic proxy archives so with that I'll take any questions and again apologies for the hiccups initially thanks Chris that was great so we have time for a few questions Torsten thanks Chris great presentation asking a little bit the question that also came up in Q&A that we'll probably revisit the business about fingerprinting right you and the previous speaker alluded to this with an understanding for how different ice melt should be expressed and we know that the ocean system is very much you know part of that and so I wonder if you could comment on the scales to which they interfere and to the scales to which they decouple I mean how much of the fingerprinting can we do given the uncertainties in the ocean dynamics yeah great question so again I'll caveat this with this is an oceanographers and I would defer to my geophysics colleagues to give more technical input but I can speak to a couple aspects so you asked about interference and coupling I'll speak more to the interference so I'll translate the question of to what extent so I made the comment that a lot of ocean dynamics studies assume that once you remove the trend from a sea level record you've essentially removed all the solid earth geophysical effects and what's left is largely an ocean circulation in climate that's just not true and there have been a number of studies specifically for example along the east coast I'm thinking of things papers by Jim Davis and Nadia Vinogrova also earlier work by Thomas Frederick said that it showed that for example recent accelerations along the U.S. east coast particularly south of Cape address so this is along Florida, Georgia and the Carolinas much of that acceleration in recent decades has an important contribution from melting of land ice and also Thomas Frederick said his paper says looked at longer times over the past 50 or 60 years and it showed that sort of ice and water mass flux matters a lot for these decadal fluctuations so certainly strictly speaking it's not true to assume that decadal variability in a tide gauges just due to ocean dynamics so in that sense again we understand very well if you're given an amount of ice melt or water mass redistribution knowledge is fairly good about how the fingerprints will respond so we know the fingerprints are very effective there is having a good history of ice melt say of the Greenland ice sheet or of mountain glaciers things like that so we have a lot of that knowledge there's been a lot of progress in sort of recording and estimating past changes in land ice and again we know the fingerprints fairly well and their uncertainties so it's just a matter of having a little bit more communication between the geophysicists who have that knowledge and the oceanographers and again there are a few papers that do this but we need more of them for your talk let's I want to bring Ben into the discussion here and have the two of you kind of address questions together we can take questions from the committee if you raise your virtual hand we can take questions from the audience through the Q&A feature so anybody have a question they want to bring up or I can bring up one of the questions from the audience let me go to one of the audience questions and I'll paraphrase here but they are asking and this could be either for Chris or Ben you can decide which one should handle it but they're asking if long-term changes and atmospheric pressure have been considered in studies of sea level change go ahead Chris so there was the PhD thesis a few years ago so I wrote my PhD on this this is certainly considered it's an effect that is especially important at short time scales and high latitude so this is the so-called inverted barometer effect so by and large when you have changes in barometric pressure they will affect sea level change but it's largely passive that is it doesn't couple the changes in ocean circulation and as I said these effects tend to be more important at high frequencies and high latitudes but as we start talking about these multi-decadal changes over the global ocean they tend to be less important there are isolated incidents that of course one needs to be aware of but it's more of a quantitative thing I would argue on the whole doesn't make much of a qualitative difference and I should clarify that especially we're talking about global mean sea level since ocean water is largely incompressible atmospheric pressure doesn't have any effect on global mean sea level alright thanks I got another question from the audience which is kind of a big picture question which I think is important to address here where do the biggest uncertainties lie and what new observations are most important to understand regional sea level change yeah so I mean I think Chris and I might have slightly different answers or different perspectives so I'll go first but I mean in all these processes I mean there's a lot of science still going on a lot of uncertainty on all of these so I mean in terms of one of the biggest drivers of rapid sea level change at the coast I mean certainly what's happening with the ice sheets and what could trigger a rapid change in the contribution from those ice sheets I mean that's certainly a big driver on these longer time scales but also something as simple as knowing what's happening at the coast right so I talked a lot about the altimeters those altimeters don't give us observations right up to the coast so there's this mismatch a little bit between our open ocean observations and what we see with tide gauges obviously we're really interested in what's happening directly at the coast that's where the impacts are happening that's where you start to have some of these questions of vertical and motion what's happening there versus what's happening in the ocean so there are certainly observations that are needed I think we would all say for different things but I mean improved new observations of the ice sheets I think is important for rapid increases in sea level in the future and then more observations in that coastal environment I think are very critical too so we can get that information closer to the coast yeah I would second some of the things that Ben said of course there's the issue of the ice sheets and that's sort of often the elephant of the room especially as we go up to longer time scales the issue of the coast really really important again these tide gauges that I spoke of they give really good information about spatial scales along the coast but they are really blind to its offshore scales which if you combine that with the issue Ben mentioned about altimeters the current generation of altimeters being somewhat degraded for various reasons within 10 or 20 kilometers of the coast there really is this sort of zone near the coast that we're really sort of blind to and don't have good knowledge of and that goes for the ocean physics too there's a lot of developments in physical oceanography right now that are trying to understand a little bit better I'm maybe guilty of oversimplifying in that slide I had with the cartoon equations there's a lot of really subtle questions involved in terms of how change in the large scale circulation is transmitted to to the coast so there's a lot of questions there and the last thing I'll add and this is basically just to plug the talks that will come up is vertical land motion like I said you know if I could have redone one of my slides you could also trace the history say of ocean dynamic studies or global sea level studies in terms of the assumptions that are involved in terms of land motion geophysical effects again you know Ben Ben showed that nice figure of the available GPS data but even with that really dense GPS network one has to make you know often those GPS stations aren't co-located with a tide gauge for example so one has to make assumptions when sort of mapping from the GPS data to a tide gauge in terms of determining you know how do we correct a tide gauge for vertical land motion effects and there are various ways of doing that various assumptions folks make but I would argue that our knowledge would be surprised and encouraged by the talks that come is that right now the knowledge of say the magnitude spatial scales of time scales of vertical land motion are relatively unknown in this context and we see this more and more as for example Ben showed some figures from California there's the recent work presented by Brett Buzango looking at the Norfolk region using INSAR and GPS there's all this very large magnitude short spatial scale behavior that's critically important for sea level that's only just now coming to light that's a crucial uncertainty thanks Cindy do you have a question yes I do I think I well let me just give a bit of background I'm hoping that my home in New Orleans isn't part of our experimental sea level guide or reference sites but no engaging in looking at the coastal environment in this near shore it's non-trivial in particularly in areas of active sedimentation because they're added signals and how are we going to work in this near shore area and how far along are we in cooperating or collaborating and capitalizing on the large number of industry wells in that are actually pinned very deeply and stably and how is the community is there any progress or any motion in the community trying to capitalize on the existing infrastructure in the near shore off shore areas may I take a first stab at that I don't know broadly the answer to your question I do know there's specific efforts to improve our understanding on those spatial scales in the coastal region so for instance NASA has a program called delta X an observation program which tries to combine airborne spaceporn and in situ observations in the exact region you're talking about so try to really saturate a particular area with observations making some of the connections you're referring to and seeing what improved understanding we can get out and that definitely helps in these areas that are threatened again I mean that's I think what you refer to as a big area for coastal communities you obviously have one specific case or location I think a similar thing could come up in other locations too but yeah I don't know Chris if you have anything any specific examples or general thoughts on this yeah I mean I'll sort of I guess echo you I mean I don't have the answer and I think that you identify a really really crucial important that we need to make improvements not because the answer doesn't exist but because that's sort of beyond my expertise but again I can speak to a few things that I'm aware of that are sort of more generally sort of moving towards that knowledge I mean one is the proliferation of citizen science you know a lot of coastal towns again Norfolk being being one example where the sunny day nuisance high tide flooding events are becoming more and more frequent you know everyone's noticing it and there's efforts to use citizen science to better map the geography of coastal flooding so again you know if you look at the tide gauge data that I've been showing you know a city might have one tide gauge you know and the geography of coastal flooding is really complex and you're not going to alias all that spatial structure if you have a single tide gauge in a location so you know crowdsourcing you know the citizen science is one way another you know I'm thinking of folks down at Georgia Tech for example and in other institutions that are trying to develop low cost coastal sensors that are deployable on mass you know you wouldn't be looking for the same sort of stability and accuracy of say a NOAA tide gauge but you would want to have you know these sensors that could be deployed around a coastal community or a city that would be able to better to map you know when it floods where does it flood and why because again you really can't get at that kind of sort of granular information with the tide gauge network so people like Manady Lorenzo and Kim Cobb at Georgia Tech are spearheading efforts like that say in cities like Savannah and efforts like that are ongoing in other cities too so those are those are to the efforts that I'm aware of again they don't speak specifically to your question in geographic region but those are the things that I know of that more generally get at the issue you're speaking of hopefully that's informative oh I was I was looking to the general question I just kind of made a joke about my specifics Maya do you have a question yes I just have a quick question for Ben you were talking about the importance of regional signals and you showed this amazing plot from Southern California where you have an Anaheim going up by I think it's seven millimeters per year and then Lompop going down rapidly is that just tectonic effect or what's more is it water cable what's causing those kind of drastic changes so I can actually point to my new precise talk tomorrow I think that plot was actually from his work so he'll go over that in detail I don't want to step on his toes but I think just a very very general answer to your question there's a lot of different processes contributing to this local these local effects it's sometimes surprising how local the vertical land motion signals can be which again that just points to the importance of the answer analysis that he'll talk to you have good answers to that question Torsten do you have another question I guess mainly for Chris I wonder I mean it's clear that the problem of estimating local relative sea level is complex in terms of its uncertainties and I guess I wonder you know if I'm a policy planner and I want to have information say for the next 50 years right so I can see how the long-term uncertainty obviously has to do with what we do with our carbon emissions right what's going to happen but sort of you know in maybe on a slightly shorter time scale if I were then to ask well what are the uncertainties are there data products out there that would provide that given like a local estimate of uncertainty for a certain time scale and can we turn our understanding of the dynamics around and ask for specific application what are the additional data sets that we needed that we would need to collect right can we turn it around and say alright for this question really it's the relative time gages for this question it's the 50 kilometers scale geodetic uplift that would make the most impact in terms of doing a better job and mitigating these changes yeah no great question and I think in your question you're sort of moving towards the answer I mean the answer is I mean it depends it depends on spatial scale and time scale as you're hinting to and you know eluding back to those earlier slides that I showed at the beginning of my talk and then also Ben showed in his time the answer is going to depend on where you are when you are and what scales you're talking about but we're roughly speaking if you're if you're talking about not that the deep long-term say 50 100 years if you're talking more to the next decade or two decades you know that that is that is a that is a time scale a time horizon where sort of this natural variability of the ocean atmosphere things that Ben mentioned natural climate modes for example things like north Atlantic oscillation or and so I mean you can you can frame it in that language you can also frame it equivalently sort of different side of the same coin of looking at you know variations in ocean circulation in which for all intents and purposes you know might be sort of random and sort of stochastic going into the future I mean it really is that those those couple processes they are quantifying and so and things like that right yeah so so right so so so speaking very broadly that's going to be on a large scale one of the most important processes so that being said our our you know predictions are only as good our predictions of sea level for instance on those time scales are only as good as our predictions say of ENSO or future changes in ocean circulation and that's getting to an uncertain place however while I might not be able to say 10 years from now whether or not we're going to be in a positive or negative ENSO phase you know you can start thinking of these bulk statistics that you know that are informative to someone like a planner for example Phil Thompson at the University of Hawaii has come up with a really innovative flooding tool that gives that sort of incorporates these longer term projections with tides and then storms and decadal variability all these things together to project forward year by year what are the odds that any given town or city will experience X number of days of flooding and what's helpful there is that you know while the uncertainty is it can be large for any given year you can start imagining computing things like okay I'm not concerned with 2041 per se but tell me but I'm considering 2040 to 2050 and what are the odds that one year in there we're going to see you know 20 or 30 flooding days I mean that that sort of knowledge that more bulk aggregated knowledge is really really informative and we're moving towards being able to make statements like that again it's it's partly a function of ocean dynamics tides do come in storms do come in long term sea level change comes in but we can start making those general statistical statements but again if we're making a particular prediction it's only going to be as good as our predictions of those component parts thanks all right well thanks everybody thanks Ben and Chris those were excellent introductory talks the focus of this workshop so sorry we didn't get to everybody's questions but we're going to take a little break now into for about 30 minutes at two o'clock eastern time we'll come back with more talks and Mark will introduce our next two speakers so everybody go grab a coffee and we'll see you back here at two o'clock eastern time all right well welcome back everybody our next session this afternoon is going to focus on solid earth deformation and glacial isostatic adjustment we will have two speakers in this session Pippa Whitehouse and Jackie Osterman and Pippa will speak first Pippa is an associate professor in the geography department at Durham University in the United Kingdom her expertise lies in modeling the processes of glacial isostatic adjustment which consider feedbacks between ice dynamics sea level change and solid earth deformation and her talk today will focus on neglected processes the role of solid earth and controlling ice sheet contributions to sea level change Pippa I'll turn it over to you fantastic thank you let's just see I think I've shared my screen can someone confirm you've seen the right slides yes it looks great fantastic thanks Mark thanks for the introduction thanks for the invitation to speak and thanks to Ben and Chris for mentioning the elephant in the room which is the ice sheets I've bent the rules a little bit here I'm going to talk about solid earth and sea level change I'm going to bring ice sheets into that equation and in particular I'm going to think about how solid earth change influences ice sheet dynamics which obviously then has a direct influence on sea level change unlike perhaps all the other talks I don't have a picture of the sea in the background here I've got a picture of the ice sheet the largest ice sheet Antarctica here is one of the GPS instruments installed by Polnet the NSF funded project which is measuring earth deformation in response to ice sheet change and not only does that tell us about ice sheet change but it tells us about the rheology of the solid earth it tells us how the earth behaves when you force it with a surface load such as ice change or sea level change and we really need to understand the rheology and these feedbacks between ice sheets dynamics and solid earth deformation I sort of flagged up there that the large ice sheets exert primary control on long term sea level change just changing something I can see a bit better and I've included here the figure which the earlier speakers showed as well the contributions to sea level change and Jackie asked him and queried whether I should talk about neglected processes in my title but actually the process I'm going to talk about is not quite on here so the process is around the role of the solid earth in controlling ice sheet change and so what I'd actually like to do is I'd like to take this little GIA symbol over here and I'd like to add another symbol which says GIA underneath the ice sheets and that's really what I'm going to talk about how the land deformation under here influences the dynamics of the ice sheet and how that feeds into sea level change if I was going to add something I'd say our understanding is sort of low to medium on this and the strength of this process the sort of details but the time scales are actually medium it's something that can happen relatively quickly before I dive into the detail I decided to include this take-home message which really captures the essence of what I'm going to be talking about for the next 25 minutes and that is that we understand a lot of the theory behind what I'm going to talk about but we don't understand the detail we don't understand the strength of the feedbacks between ice sheet dynamics and solid earth deformation the thing that we do know at the moment is that it really depends on the earth rheology and Jackie will talk more about this about how we're mapping out 3D earth rheology and one of the reasons we're doing this is to try and understand the rheology beneath the current remaining ice sheets to get the strength of these feedbacks correct this is the last of my introductory slides and I just thought it would be important to touch on why it's important to understand these processes the first thing to say is that ice sheet change models of ice sheet change have actually included a component of subsidence or rebound in them for a number of decades so it's been included in some of the original ice sheet models that were developed to model continental scale ice sheet change but until recently we didn't really think about whether that representation of solid earth deformation was accurate or not we included it but we hadn't really tested whether the parameters that we'd chosen were correct and we hadn't really thought about other details in there perhaps from a solid earth community that actually needed to be included as we're modelling the ice sheets and the reason it's important to get these processes right is because they're included in models that are being used to predict future change so we had subsidence and rebound included in models for a while but we're including more complex versions of those processes now and we need to get those details right those models are calibrated often on their ability to reproduce observations and I've mentioned past change here and actually this can include relatively contemporary observations so we have good observations contemporary ice sheet change and some of these models are being tuned on the basis of whether they reproduce those observations on a decadal scale we also more traditionally tune ice sheet models on their ability to reproduce much longer term change perhaps over sort of the last deglacial cycle if we are tuning these models we need to make sure that we've represented the processes right in the models and we also need to make sure that we've interpreted the data correctly that we are then using to tune the models and if we don't do that then we can end up with biased models and I'll just bring you back to the first point up there that these models are being used to predict future change if we over tune the models with the incorrect processes to some data that we don't fully understand then we're going to end up with biased projections so hopefully that's a little little piece of motivation for why we're digging into some of the details here the structure of my talk I'm going to have three sections here I'm going to talk about some key concepts since in fact three key concepts I'll then talk about two areas of recent advance and then three areas where I think there are open questions that still need to be tackled so the first of these concepts much of this talk is about how the solid earth influences the ice sheets but I'm going to start the other way around and I'm going to talk about the ice sheets influence the solid earth some of you be familiar with this process of glacial isostatic adjustment what I've done here is divided it up into two time scales when we have for example the melting of an ice sheet there's something that happens straight away we refer to this as the elastic response as the ice sheet shrinks the land underneath rebounds upwards and of course we're taking melt water and putting into the ocean and the sea level changes so the global volume of the ocean increases there's also as part of this sort of instantaneous process of changing the shape of the geoid and Ben referred to this in his talk at the start that the gravitational attraction of an ice sheet is so large that it deforms the sea surface as the ice sheet shrinks the mass of the ice sheet is smaller and locally the height of the sea surface will fall so although the volume of the ocean has increased in this instantaneous idealized representation of an ice sheet melting close to the ice sheets the sea level is falling and these are the sort of processes that are represented when we talk about a sea level fingerprint and I'll come into that in my next slide but this is a snapshot of why we see these patterns the other time scale that I'd like to talk about is the viscous time scale and that's really what happens after that instant movement of ice into the ocean and response to the solid earth and this is really about the time decaying deformation of the earth it's about this ongoing rebound that happens this rebound where the ice sheet was there's deformation also under the sea floor due to changes in loading across the surface of the earth this affects the height of the sea surface in this idealized scenario I'm assuming there's no change in volume of the ocean but there will actually be changes in relative sea level because of this relaxation process so these are the two processes that I'm going to talk about and I'll come back to these a little bit a lot of the research to date has assumed an elastic response to ice sheet change I'm going to talk about some areas where we have low mantle viscosity the important thing about low mantle viscosity is that what that means is that the earth can respond faster than average in response to a surface mass change and so I would argue that in some areas where you've got ice sheet change we need to be considering where there's a viscous response to that ice sheet change on relatively short time scales on the order of decades that was the first key concept the second one I will mention here is about fingerprinting and this has been mentioned a couple of times and this is a classic sea level fingerprint of the pattern of sea level change that you would expect if we had melt from Greenland and West Antarctica totaling one millimeter per year and you see this far field sea level change which is greater than the mean value so it's greater than one millimeter per year and this near field sea level change we actually see a drop in relative sea level and for the purposes of this talk it's this feature of the drop in sea level in the near field of the ice sheets which actually plays a very important role in controlling ice sheet dynamics this is very much a a signal of contemporary change what's been hinted at in the other talks is actually there's if we're thinking about relative sea level change around the world there are other processes going on one of them being the ongoing response to past ice sheet change so we have this background GIA signal and this is the instantaneous signal that we'd expect with ongoing ice sheet change the reality is a little bit more complex and we saw a range of fingerprints earlier from terrestrial water storage perhaps other sources of ice sheet change so this pattern can be adapted basically depending on where we think we know there is ice sheet change or indeed water storage change something else to bear in mind is that concept that I mentioned about low viscosity regions in areas where we think the mantle is relatively low viscosity such as beneath Iceland perhaps Alaska, Patagonia and the Antarctic peninsula we think that instantaneous ice sheet change so I'm talking over a couple of years five to ten years of ice sheet change this actually trigger not just an elastic response but also a viscous response in which case these fingerprints get a little bit more complex and they evolve through time and there's a couple of papers here which I've mentioned which go into the detail of what's going on there this is very much a contemporary idea of what's going on nowadays we can map out ocean change we saw in Chris's talk there's the beautiful sort of altimetry observations that's going on that we have if we go back into the past we're much more restricted in the information we have about the pattern of sea level change which you could argue can be used to infer the pattern of ice sheet change we're very much restricted to relative sea level records around the coastlines and as you can see from this detailed map of sea level change it's really important to get sea level records that cover the whole of this area of change in particular if we're thinking about the polar ice sheets then getting a good north south distribution of records of sea level change captures these sort of gradients that are in these predictions so it's actually quite difficult to reconstruct past sea level change from fingerprints because of the restrictions of where we actually get the data there's an important caveat here as we saw from the satellite altimetry data the contemporary pattern of sea level change doesn't look anything like this because we have primarily the steric component of change that's going on and that's something which I think is actually needs more research if we look into past sea level data and think about what are the potential steric biases in the data are there differences in tidal range or ocean circulation that are being recorded in those sea level records that we're not accounting for when we're just thinking about that global GIO process if we're thinking about the solid earth then a couple of areas to think about is whether as came up in one of the questions whether sediment loading has an impact on what's actually recorded in a paleo relative sea level record and then on even longer time scale where the dynamic topography feeds in so these are all other factors to bear in mind if we're trying to tune models using the data here so that's a summary of how ice sheets impact sea level the last two slides the last of my key concepts is going to move to the ice sheets and again this is a figure from the recent special reports on oceans and cryosphere and changing climate and it takes a snapshot through a marine ice sheet so the third concept I'm going to talk about is the marine ice sheet instability Antarctica is a marine grounded ice sheet which means that it flows from the continent directly into the ocean and an important thing to get your head around is this concept of a grounding line and that's essentially the point where the ice stops being grounded touching down on the bed and it starts to float now the dynamics of an ice sheet depend on the thickness of ice at the grounding line essentially the flux of ice from into the ocean depends on this thickness and this is demonstrated here by the fact that as the ice sheet retreats to a deeper position the thickness of ice increases and you can see that these arrows increase so as the ice sheet retreats if the bed is sloping downwards the flux of ice increases and we get into a runaway scenario and that's the marine ice sheet instability and so these are these scenarios which predict the potential for very rapid collapse of the ice sheets if they're perturbed into this unstable situation the but it points there so on this bed we get unstable ice loss and something to bear in mind here is that models often assume a constant sea level there is a little arrow here which talks about rebound and I'll come on to the effect of that in a moment but quite often we think about the sea surface being constant if we're driving ice sheet models we're perhaps using a U-static sea level curve of what we think the ocean did in the past and the details are a little bit more complex which plays to our advantage if we're thinking about trying to find ways of slowing down ice sheet loss. So those are the three key concepts I wanted to introduce and I'm just going to talk about two areas that have recent advances and we're really getting into the cutting edge of the processes that are being thought about in recent research the first of these is the stabilizing effect of GIA and I hinted at that in the previous slide the idea that the bed will uplift as the ice sheet thins. What I've taken here is this image on the left and just there's a little bit more complexity here but I've emphasised the idea that the bed rebounds over a large spatial scale as the ice is losing mass. Another thing that happens is the sea surface falls. This comes back to the fingerprint and the idea that close to an ice sheet you have relative sea level fall even though in the far field we have sea level rise we actually get a reduction in water depth at the grounding line as the ice sheet retreats. So this is a negative feedback and it helps to stabilise the grounding line position. It might slow retreat, it might stop retreat, it could potentially cause re-advance. Those are the details that we don't understand yet but the details largely depend on earth rheology. It depends on how quickly the land responds to the ice loss and it depends on the spatial pattern of that rebound as to whether it can influence the water depth at the grounding line sufficiently to slow down the retreat. That's the theory over the last seven or eight years these have been implemented into models. This was the first and still the neatest representation of this process worked by Natalia Gomez. In the top image she has not implemented any GIA feedbacks and in the bottom one we do have this rebound process of the solid earth and the slight decrease in the height of the sea surface and you can see going from dark blue to light blue in the upper picture we have runaway retreats and in the lower picture that rebound is stabilised. Thinking about longer time scale processes again shortly after this process was identified Basta Burr run this simulation through multiple glacial cycles including these feedbacks that's the coupled model that we're talking about here and it's interesting that these discrepancies between the coupled model the solid line and the standard model the dashed line actually occurred during glacial maxima so this is the volume of the Antarctic ice sheet down here and he predicted that he couldn't actually model as large an ice sheet if we included these feedbacks so not only do these feedbacks damp the rate of grounding line retreat this shows that they also damp the rate of growth of an ice sheet it's really sort of a way of putting the brakes on the dynamics of the ice sheets and it can potentially reduce the interglacial variability of ice sheet volume this is pretty much looking at sort of past change here and and so something that this sort of vetted very quickly into is the idea of how this might influence future ice sheet change and this is where we started to think more carefully about mantle viscosity model predictions show that the GIA process can reduce rates of future ice sheet loss is a couple of papers from 2015 now where the authors ran projections into the future where ice sheet dynamics was coupled to GIA processes and they demonstrated that if we include both the feedbacks but especially the feedbacks where the mantle viscosity is very low and that rebound kicks in very quickly then we can limit the rates and maximum amount of ice loss in the future these are very much proof of concept asking the question of can this process help the next step was to move into thinking about what the details here what is the mantle viscosity which is low enough that it's actually going to be able to prevent future ice loss in a large scale way this is an article a figure from one of David Pollard's papers couple of years later this is quite an extreme scenario this is an RCP 8.5 scenario the dashed line here is the CO2 forcing and you can see that the CO2 forcing is so strong that we completely lose the ice sheets we've got East Antarctica and a certain portion of East Antarctica with a sea level rise shown on the left here within about a thousand years what happens from that point onwards when the CO2 returns to much lower levels is that for this pink line the ice sheet starts to regrow and this pink line is the one with the lowest mantle viscosity it's where this rebound kicks in and helps to stabilise the remaining ice sheet at the other end as you might expect the blue line at the top here that's the one with the strongest mantle viscosity so that's the one where the rebound process is not as quick and it doesn't provide this same damping effect on ice sheet dynamics so this starts to tell us that we need to understand the mantle viscosity to know which of these trajectories we want these models applied a single mantle underneath the whole of the ice sheet but actually we know the one thing we do know about mantle viscosity and Antarctica is under West Antarctica it's much weaker and under Eastern Antarctica it's much stronger so we need to home in on the details a little bit just earlier this year there was a study around Pine Island Glacier which is one area where we do know the mantle viscosity is very very low because we see the rebound happening so quickly in the GPS measurements and this study demonstrated that if you have the lowest mantle viscosity it's this upper line here then you reduce the amount of sea level rise that you predict on sort of a decade or centennial time scale so going through these this range of predictions the lower one here is one where there's no feedbacks and we get the maximum amount of ice loss when we include feedbacks particularly with a weak earth model in the Pine Island region then we reduce the rate of ice loss so we're getting there we're starting to to nail down the details of what these trajectories might look like if we know the mantle viscosity the ultimate step here is to use one of these coupled models and implement spatial variations in mantle viscosity across the whole of Antarctica and I'll show you in a couple of slides time that we don't fully understand the details of the mantle viscosity but we do know that if we include spatial variations in mantle viscosity it makes a difference this is the difference between two coupled models one with a simple earth structure and one with a detailed earth structure and we can see that the sea level relative sea level predictions make a difference so including those the 3D Earth Structure makes a difference to the sea level predictions this is for 15,000 years ago so it's essentially water depth around the grounding line at that time and we need to get that right to get the ice sheet dynamics correct those are two areas where we've seen relatively recent advances so hopefully you're now getting an idea that GIA helps stabilize ice sheets and if we have a weak enough mantle viscosity actually that stabilisation is much much stronger so I'll just go into the last couple of slides where I've posed a few open questions the first one is what the details of that relationship between earth deformation and ice sheet dynamics the right hand figure here you've actually seen before this was reproduced from the Pollard paper but it had this very strong RCPA 0.5 forcing the left figure here is less strong forcing it's a doubling of CO2 and it looks at the trajectory of the Antarctic ice sheet over a couple of thousand years the black line is one with no feedbacks so we get this runaway ice loss situation the red one includes feedbacks but it uses a strong earth model the blue one uses one of these weak earth models so we can see that on the side here this is global mean sea level implications we prevent the collapse of West Antarctica there's nothing we can do about it we can't change mantle viscosity but it will be prevented if the mantle viscosity is lower under West Antarctica versus if it is stronger I think what we really need to think about here is try and understand what trajectory we're on does the rate of climate forcing matter does the rate of the ice loss matter in terms of the way that the rebound is triggered in response to the ice sheet change the next step here these are from 1D models the first step in the structure is to map out the viscosity there are two ways to do this that I'm going to summarise here the first one is to think about the idea of an understanding of surface mass change and a measurement of how the earth responded to that surface mass change if we have those two things we understand the reology so if we know how the earth is forced we know how it responded that's a measurement of reology and this is an experiment which has been going on in the Antarctic Peninsula recently due to the rapid ice loss that is going on but we have this network of GPS instruments the pink dots here on the left which are surrounding this area of ice loss this is altimetric measurements of ice sheet loss since 2002 following the breakup of the Larson B ice shelf and on the right here I show the GPS measurement from one of these sites from the Palmer Station and we can see the rapid ice loss commenced in 2002 and we see this drastic change in the rate of rebound of one of these instruments so this is work from Grace Neal to PhD she modelled the elastic response to ice sheet change that's the red line here and it gets nowhere near the rebound that we were seeing in the GPS instruments she then searched through a range of viscosity values and identified the one which gives us the best fit for the change in rates and it's around 10 to 18 pascal seconds which is several orders of magnitude to global average mental viscosities so we're starting to be able to nail down what is the viscosity at this location due to this natural experiment that's going on and this is something which has been picked up by a large number of authors since using this GPS network the second way just so we don't have these instruments everywhere this sort of natural experiment going on everywhere so we need to map out mental viscosity in a much more independent way and I know Jackie's going to talk about this in a lot more detail but essentially the approach that we can be used is to map out seismic velocity perturbations which is what is shown in this figure here which help us understand mental temperature distributions and hence viscosity now there are many assumptions as you go from seismic velocity perturbations to mental viscosity and so this is basically all I'm going to say is there are probably more questions than answers at the moment here and so this is not something that we can take and map out absolute values of mental viscosity and start to implement models there's a lot more research needs to be done to understand how we do that and I will just mention the added complication if we go to lab experiments in biology we see that actually the viscosity of the mantle probably depends on the stress that is applied and if you think about it the stress that's applied is the ice sheet change and that changes through time so there's a question of whether the viscosity actually evolves through time as well as being spatially available and the final question I'll just mention here is thinking about the past to predict the future and we often tune them to see if they can represent what's happened in the past on the left here there's an image of a model reproduction of the retreats of the West Antarctic ice sheet and what's actually happening here is the ice sheet retreats it retreats back to this pinning point it retains an ice shelf and then we have this rebound process what happens is not at the grounding line but this high point here touches down on the base of the ice shelf and the ice shelf re-grounds which is quite a threshold process with an ice sheet changes the whole stress regime and the ice sheet starts to re-advance back to this pinning point so you can see there's a lot more subtlety here which comes down to evolution of those floating ice shelves is absolutely key and understanding the bed topography underneath Antarctica so if we're trying to run ice sheet models and reproduce the past these are the sort of details we need to be thinking about representing in those models just very quickly I will make a quick mention of longer time scales this is not to do with rebound in response to ice sheets so much but the idea that topography has changed on a million year time scale so this is a recent article from Guy Paxman where he used reconstructions of the pythymetry 34 million years ago compared to the present day pythymetry and asked the question of what happens if you try and force an ice sheet model on those two different topographies and you actually get very different answers so the sensitivity of the ice sheet to climate forcing depends on the topography that's underneath the ice sheet this is my second to last slide we were asked to think about ways to tackle some of these questions the first point I've put here is to think about other ways we can measure the rebound of the uplift or deformation of the solid earth as it's forced by surface mass change I agree with the earlier speakers that we need observations of surface deformation I'd love to throw this one out I'd like to know what earth deformation is underneath the ice sheets and underneath the oceans if we can map that out in a much larger scale and whether there are ways of trying to get it past deformation in a smarter way we use relative sea level records but the resolution the accuracy there is relatively coarse are there smarter ways to understand past surface deformation if we're thinking about modelings then there's two areas which I think the field will continue to head towards one is data inversion so the use of data to tune models and the other is coupled modelling I've got my notes in front of me here all I've put here is that if you get ice sheet modelers together with GI modelers that's a really good idea and if you can bring in people that understand data and the uncertainties that's even better so that's my last point here is to really we're moving towards a better treatment of uncertainties and that's in the data but also in the models and as well as understanding those uncertainties I'm sure this will come up in questions we need to be moving towards the idea that we can't just find the right answer but we need to be thinking about these probabilistic approaches to understanding all these processes so this is my summary slide hopefully you've seen that actually sea level change regulates ice sheet dynamics which in turn obviously affects sea level change and these details depend on earth rheology we need to account for feedbacks in models and also when we're interpreting data that doesn't mean that all models need to be fully coupled or fully 3D but we need to understand what's in the models what we need to correct for and potentially what we could parameterise in the relatively simple way and do a good enough job and then the last area which I touched on map our earth rheology Jackie will talk more about this and this really has implications for a wide number of areas I've talked about ice sheet dynamics and I just touched on this one here current reconstructions of global ice sheet change over the last 20,000 years have largely been built assuming a 1D earth it's a massive sort of inversion and data tuning exercise to build these models of what the ice sheets did we need to revisit some of this and think about do these still hold if we actually build them on top of some sort of 3D earth so I'll just show here just to close my take home message that to be able to predict future sea level change we need to understand the strength of the feedbacks between ice sheets and solid earth deformation and we know that this depends on variations in earth rheology thank you thank you very much for that excellent talk we're right at 255 so what I think we'll do is we'll go straight on to Jackie for her talk and then Pippa when we come back and the end we'll have questions straight for you and then for both of you to get it thanks Matt so next I'd like to introduce Jackie Osterman Jackie is an assistant professor at earth and environmental sciences at Columbia University and works to understand how sea level and ice sheets have changed over the past hundreds and thousands and millions of years her work uses large scale numerical simulations to quantify the magnitude and rate of sea level change in response to warming temperatures and to unravel the interactions between solid earth processes and the paleo climate record of earth geologic past her talk today is going to be entitled using paleo sea level records internal structure and decide for drivers of sea level change so Jackie thanks a lot for joining us today and please take it away thank you so much Mark let me pull up my slides does that look good yeah that looks perfect okay fantastic okay excellent well I also want to start by just thinking the organizers of this has been really great to see already the talks and engage on this topic and also thank all the participants because I know zoom fatigue is real so really appreciate everyone being here today it's fantastic to speak after Pippa and after everyone else which already lays the groundwork for a lot of the things that I'm going to be talking about I also have a token field field picture on my front slide here to Pippa where we were in the Bahamas last year to map out some paleo shorelines so I will move a little bit further away from the ice sheet compared to Pippa's talk and a little bit deeper into geologic time we've heard a lot about GIA we've covered the process so I'm glad that I don't have to go over these details again and instead I'm going to jump right in here we'll just look at sea level change over while sea level change today or over the haul scene and how it's driven by GIA and Pippa referred to this as the background GIA so I'll snuck that into the title of this slide here so this is a prediction that shows how much sea level is changing today only as a result of the ongoing adjustment to the last deglaciation and what stands out here of course are these areas in red areas formerly covered by ice sheets where sea level is falling as a result of postglacial rebound and areas in blue here where sea level is rising as a result of both the eustatic changes so ice mass loss over this time period but more importantly so for today and for this rate of change here areas where we're on the peripheral bulge of the former ice sheets that are subsiding over with the interglacial time periods there's also a lot of structure here in the FA field compared to one of the earlier slides that were shown in some of the introductory talks here we're looking at sea level change millimeters per year, sub-millimeter per year which is really this kind of small scale features in the FA field that are hard to detect and I think we'll hear more about this tomorrow as well and definitely on the same order of magnitude that has a lot of the other sea level changes that are occurring but we see also patterns here where that are sort of following the coastlines this kind of halo around the coastline as a result of the oceans gaining and losing mass over the glacial cycle and therefore we have excess load or reduced load and I think about these kind of outlines around the continents as sort of the peripheral bulge of the ocean load instead of the peripheral bulge of the ice sheets and so as a result we see sort of small scale this is just a meter here small scale sea level changes over the Holocene so we understand and this overall you know I'm starting off with kind of the general things that I think we understand quite well before I go into the aspects that I think we don't understand quite well and propose some paths forward of how we can understand them better so I think we have a good understanding of this first order pattern in fact you know some of these observations were the first to tell us about Earth's viscosity so Haskell in 1935 used uplifted shorelines and finoscandia to impose and suggest that Earth's viscosity is 10 to 21 Pascal seconds which is really a measure that pretty much holds today and similar work of course has been done since here is a sort of an emergence curve from Hudson Bay where sea level is falling the crucial aspect of this work here that really allows us to tease out viscosity more clearly than in other locations is that while the on the right here the magnitude of how high sea level was at an earlier time depends on how thick the ice sheet was back in time which is a measure we don't know as well what is not dependent on the ice sheet thickness is that this sort of exponential shape so the decay time of this exponential shape is pretty much only sensitive to the viscosity and therefore we can use these round curves to constraint put constraints on the viscosity there's been a lot of work and mentioned this as well but there's been a lot of work using sea level histories from around the globe and trying to understand Earth's viscosity as well as ice sheet variability over the deglaciation and highlighting one study here and others exist where they used this is from Kurt Landbeck and they used a variety of locations that are shown here as black and red markers there are sea level records mostly Holocene but also during the deglaciation and they inverted these in an iterative procedure to find the optimal parameters for lithospheric thickness upper mantle viscosity lower mantle viscosity and then pose a resulting global mean sea level history what they find is they get constraints on the lithospheric thickness that range there are two scenarios the red and the blue but they find this what's shown here is misfit on the y-axis and essentially these values that have lowest misfit are the ones that are most likely in best fit the data they obtain a sort of likely range for lithospheric thickness these are all one dimensional models in which the viscosity in Earth's structure only varies with depth they find that the upper mantle has a viscosity of about 10 to 20 and that the lower mantle viscosity has sort of two solutions one around 10 to 21 and the other around 10 to 23 there has been a history in the community of the debate of what the viscosity is in the lower mantle also using not just sea level observations but also geodetic observations J2.0 for example here in this specific study what the authors favorite was this higher for a couple of reasons but most notably that it required a smaller Arctic ice sheet smaller ice volume at the last glacial maximum which as you just heard from Pippa the sort of updates to the reconstruction of the ice sheet volume over the deglaciation from an Arctic sort of come in lower and lower so as in there's less and less change over the glacial cycle actually and this is sort of more in line with this higher viscosity here and we get a deglacial global mean sea level history that starts with sea level at around 130 meters during the last glacial maximum and then a deglaciation with sort of punctuated ice melt that is where in particular the past rates of sea level rise occur during Mount Water Pulse 1a Mount Water Pulse 1b for example and those have been used and studied as trying to understand speed limits on ice retreat which is of course important if we think about the future so there are details here in these reconstructions that will differ a little bit from study to study but I think this is the overall there is an overall understanding here and this is just to show this as well and again the sort of ice sheet evolution over the last half a million years with major ice sheets in North America in Fennoscandia the exact distribution of them is not known very well but we have a pretty good I would say first order understanding of this ice sheet variability you see of course here during the interglacials like this one the last deglacial where the sea is actually rebounding and you see how during glacials land bridges form and are disconnected during the deglaciation there's a lot of connections to archaeology for example as well so this is sort of you know this is the biggest picture that I think will have a pretty good sense of now we're going to move to thinking about okay some of the details that we probably that I want to argue that we don't know as well on to how we can understand them better so one of those you know I say details but it's really it's pretty pretty significant difference is actually the the exact geometry of the past ice sheets and just to demonstrate this here I'm showing you three different learned at ice sheet reconstructions at 18,000 years ago and they are fit to different observations some of them are more sort of ice physics based than others but they are quite significant differences here and this is you know I would say one of the very important areas of ongoing research really trying to understand these differences there's a lot more data that can be used in assimilated and I'll get to that when we talk about future directions along those lines of sort of outstanding questions for the deglaciation there is also a the question still of what's referred to as the missing ice problem at the last glacier maximum which means that at the if we look at the if we look at sea level records from the last glacier maximum we require an ice volume an ice equivalent volume that is larger than we can really reasonably distribute among the possible ice sheets so there's this missing ice there's some emerging some suggestions of how this could be reconciled but it's certainly not a solved problem similarly thinking about how much what the contribution is from different ice sheets to meltwater post 1a and other meltwater events there is work being done in this direction and which is really interesting but these are definitely still that's the level of detail that is really important if we want to understand ice sheet variability at the moment the other big component in the GIM modeling is of course also lateral variability and viscosity and Pip already gave you a good overview of this I'm showing you here just to give you a sense of sort of the order of magnitude variability that we're expecting you see lithospheric thickness here in the top left and then variability and viscosity at a couple of incremental depths so viscosity likely varies by an order of magnitude we can what is used here rely on seismic tomography and rheological laws again Pip talked about this there are uncertainties in this here we're using some geodynamic constraints to better map shear wave speed into temperature and then into temperature and then we have certainly uncertainties remain both in the scaling relationships as well as just the tomography itself and one aspect to mention here is that during calculations that account for this full variability and viscosity is a lot more computationally expensive and the reason for that is that we only vary whether the structure is regularly symmetric we can actually solve parts of the equations semi analytically this has to be fully numerical in finite volume, finite element pick your method of choice but it has to be fully numerical which makes these calculations the actual sea level calculation associated with this earth structure a lot more computationally expensive and as a result a lot harder to explore uncertainty and as a result that makes it a lot harder to explore uncertainties so where does it leave us so what are sort of the data model misfits that we still have for the deglaciation and I'm just plotting out a couple of sites here you see different locations here over the deglaciation you might have noticed that time for me sometimes goes in one direction the other direction so you always have to keep an eye out for that so here it goes towards the left with the present being on the left and I am showing you sea level data as sea level index points which mark sea level marine limiting which mean that sea level has to be higher than that and terrestrial limiting which means that sea level has to be lower than that from a series of locations and the locations are kind of group locations that are shown here on this plot the prediction that I'm showing here is the 6GVM5 sea level prediction and this is a great sea level model and sea level prediction that's widely used what's great about it is that it's actually very easily available and it's been shared widely which is one of the reasons why it's so widely used it fits some of the sea level observations so it's a model that assumes a radially symmetric viscosity so it fits quite well in some locations here possibly up here and there are other locations where it doesn't or it fits in Barbados very well the Barbados data also went into constructing this ice model and then it doesn't fit so well in other locations I pick these somewhat randomly just not to show that this model is not a good model but just to show any ice model that you were to choose would do better in some locations and not so well in other locations and it's a result of the uncertainty that we have in the ice reconstruction and the uncertainty that we have in the viscosity structure so how can we move forward here how can we come up with a model with a GIA model that actually fits these observations well and that does allow at the same time for variability in Earth's viscosity I'm going to kind of propose one suggestion here of an approach and just before I go into that here's sort of the setup of the problem we have our GIA model what goes in is the ice and ocean loading most notably the ice sheet reconstruction the ocean loading is calculated from the ice reconstruction and the earth reology that goes that's input into a GIA model the output is our solid earth deformation it's relative sea level change and other quantities we can we make a prediction we can compare that to sea level observations and then adjust the ice and earth model accordingly so the crux here is of course what does it mean to adjust the ice and the earth model accordingly I'm going to show you one approach of how to do this now well we're just going to be adjusting the earth model and we're going to lend the methodology from seismology essentially what we want to solve is an in risk problem and when I started working on this I thought about this in terms of seismic tomography and since this is a solid earth crowd I thought this would be a good analogy to set up this approach as you all know seismic tomography we have an earthquake we have seismic stations waves travel through earth's interior and we measure the seismic speed at the seismic station there are sensitivity kernels in the earth's interior that connect where the earthquake happens and where the seismic station records the seismic waves and that means these are the areas that whatever the seismic waves that travels through the mantle is sensitive to a fast or slow structure so we have these sensitivity kernels here which are sort of the darkened areas and if we have a structure that is a feature in the mantle that is slow the wave will arrive later than expected if it's fast it will arrive faster than expected and we can take that knowledge that mismatch between the observation what we expected and what actually happened to update the internal structure and here we go that gives us our seismic tomography we can do the same thing with sea level observations so and this is work that's been pioneered by David Alatar and and his student Ophelia Crawford Crawford so what he's calculated here is a sensitivity kernel for the viscoelastic deformation problem so if you were to imagine that the ice load is just a point load which of course in reality it isn't just for analogy sense and our rebound observation is a point observation which it is then we can calculate the sensitivity kernel for the physics of this problem and we find a sensitivity that's actually quite similar to the sort of banana donut seismic sensitivity kernels we can calculate the sensitivity kernels in a 3D earth model very cheaply by using adjoint equations which have been used in seismology and geodynamics and now are also derived for the GIA loading problem so what does the sensitivity kernel look like not for an idealized point load and point station but actually for a real world extensive ice sheet as well as kind of a sea level record and that's shown here so we're looking at sensitivity kernels that have a specific ice you know the results of a specific ice history we're looking at a station so we have you know the Laurent Hyde and the fan of scanning ice sheet as I showed you this before and we have a station here that's located in Bonaparte Gauls we're looking at the sensitivity kernel at two depths here at a thousand kilometer depth and you see that there are areas essentially you see the areas that are red and blue are the areas that this specific observation so the sea level record at this side is sensitive to and you see it's sensitive to structure right underneath where the record is where the sea level record is and it's sensitive to the area ice loading again similar to seismic tomography we can use the knowledge here where if we have if we have a positive viscosity kernel here in blue we know that if we increase viscosity in that region we will increase local sea level where else if we have a negative viscosity kernel say in red here we know that increasing viscosity will lead to a decrease in the actual sea level prediction and this is fantastic because it gives us a means of updating and inferring 3D earth structure based on a data model misfit and I'm going to show you an example of how this can work with a synthetic test so the synthetic test here is that we're going to use the same database that I showed you earlier from this land back paper so this is sort of the geographic range of where we have sea level observations and we're going to set up a a known 3D viscosity structure of the earth and we want to then we then produce data with this known 3D viscosity structure and then we use the data to back out and infer back what the viscosity variability was and compare how close we got to the known structure so this is all synthetic no real data in here so we're putting this is the viscosity model that we assume to be correct this is based on the S20RTS seismic model we use a fixed ice geometry we're not inverting for ice geometry here at all the data in the far field so details in the ice geometry probably don't matter as much but if we go to a global scale of course we don't you know there's uncertainty in this which we're connecting here in this in this simplest case scenario and we run forward to projections of sea level at these different locations and produce our synthetic data we assume the data have no uncertainty okay so perfect data again easiest setup and then we start with no variations in viscosity and use our that inversion scheme those sensitivity kernels to update what the actual earth structure what the earth structure looks like and then compared to the true I'll show you the results for that in a second this is an iterative approach because we get updates to our viscosity structure then again do forward predictions compared to opposite compared to our synthetic observations update the viscosity structure and it's if we use far field data it's pretty well behaved we use near field data it gets a little bit trickier and we can also you know test do we actually match our observations after those iterations and that's shown here so this is our synthetic data is shown in black the initial one the earth model so what we started with is in is the dash blue line and then in the red is the inverted results so we see okay our inversion met matches the observation this is great but do we actually map out the correct our structure right so I'm showing you here on the left and I'm going to go through this kind of depth by depth what the viscosity structure is that we get out and you know ideally if this is a perfect inversion it looks exactly like this in reality you will already expect that it can't it can never look like this because our data really is mostly located in these regions so we don't really have any information in big parts of the earth but let's just look at the regions where we do have data okay so this is our inverted the result of our inversion the inverted viscosity structure we see sort of higher viscosity here and around here lower viscosity here which does you know our features that do to first order match up with the with the true with the true viscosity structure we also see the sort of higher viscosity sorry lower viscosity here which is a little bit present but not quite as striking if we go to 500 kilometers depth we see quite nicely that we are mapping you know the subduction zone here we're mapping these high viscosity slabs that are in our in the true model and we continue to map the slow viscosity zone here we're starting to pick up this high viscosity which is not really present at this slice but if we go a little bit deeper we're seeing sort of some bleeding between the depth slices where it's starting to pick up this high viscosity here again continue these features and one slice deeper you know we see some of this variability in the Pacific I would argue that we're sort of actually matching and are able to map some of the first order features in the viscosity structure we can actually quantify this and look at the correlation between what we know is the true viscosity and our inverted viscosity the left shows you the correlation coefficient between the two red is we're doing really well we're doing really well in the structure blue means it's actually anti-correlated so we're getting the opposite of what we should get so blue is bad red is good you see overall in the in the whole glow we don't in some regions we do better in other regions we don't do well at all however if we only look at regions where we actually update the viscosity so we actually feeding information into our inversion we are left with these regions that all have really high correlation so we are actually able to recover some of the 3D structure through this tomography viscosity tomography here's the dream my dream is to move towards such tomographic imaging of Earth's 3D viscosity I think the key part here is that it's towards a tomographic image there are a couple of very important caveats that make this very difficult but I think still possible and I'm going to end with showing you are reflecting on a few of these complications so one is of course that we need more data and more data exists so this is a compilation of the Holocene C level data that are available there have been great community efforts from the Halsey working group to compile and standardize the Holocene C level data so we can start working with them there are obvious spatial gaps that you just thought in the inversion we don't have data we don't have any information so the more data we have over this time period the better this imaging of the viscosity will work and I should also say that the time period we are most limited to the Holocene where not that much ice change happens the glacial records will be better those are much harder to get to because most of them are submerged today we can also start incorporating present day constraints GPS gravity and there was already a lot of talk about that today ice sheet reconstructions need to be improved I talked about some of the outstanding questions here in these inversions we can explore tradeoffs between inverting for ice or inverting for the Earth's viscosity but there is I think a lot of data that aren't actually assimilated in these and this goes more into PIPA's direction on the cryospheric side but I think there is a lot of data that we can leverage to actually distinguish between which of these ice sheet reconstructions is right so beyond sea level data and present day constraints there are fluvial and pro-gacial lakes all around here that have tilted shorelines that we can use river routing across here, across here, across here that tell us something about the GIA signal could use eschers to constrain the sub-glacial hydrology of the warranted ice sheets people have also used constraints from atmospheric circulation so it's kind of an eclectic set of constraints but bringing those together and not just in with a handmade ice model but with a physics-based ice model would really I think push this field forward significantly and I should say there are people who are doing this and that's great there is of course also still a lot to be done on this assimilation techniques I had described a gradient-based optimization using adjoins there are other approaches out there and it would be great to see them the drawback there are a lot of advantages to the approach that I described to you drawbacks are that we find a global we find a local minimum that means we find one possible solution we don't actually get a good sense of uncertainty this is a problem not unique to this community but to the seismic community to the geodynamic community and approaches like looking at, there are approaches out there and actually working together across the Solar Earth Sciences I think we can learn a lot from each other here I also would advocate for better benchmarking people have done benchmarking for radially symmetric GA models and Wouter van der Waal has led some of those efforts but I think on the 3D GIA models they have not been benchmarked and I think they should be benchmarked and it's boring work but it's really important and of course on the rheological parameters the better our constraints on the viscosity structure from other disciplines is the better the starting model in such an inversion the better our end result I may dream about joint inversions between doing a seismic inversion a geodynamic inversion and a GIA inversion together which might be possible at some point and I think would be really amazing and I think this is my last point here which is also on the rheological properties and Peppa mentioned this but everything I told you was assuming a Maxwell rheology a single dashpot and a spring but it's very likely that the Earth has transient behavior for example burgers or sort of an extended burgers material which we know from laboratory experiments at high stress high strain location especially close to the ice sheet we might move into dislocation creep that means stress dependent deformation so the rheology of a Maxwell body is something that we also need to reconsider and there's work being done in this direction but I think there will be a lot more in the next decade and oh no this is the last point and I'm going to just end on this very quickly but I think and Peppa and Chris mentioned this I think one really exciting aspect of looking at the deglaciation is that the records are becoming better and better and resolving smaller scale variability and here's a record that shows you sort of centennial variability in sea level which you know might not be a GIA process but might be an oceanographic process or might be related to sedimentation terrestrial water storage variability so a lot of the processes that we will hear about tomorrow are a lot I think I think the community is starting to be able to pick those up and separate them out from GIA which I think is another avenue that I'm certainly very excited about so to summarize and finish up here we understand the general deglaciate evolution and earth's internal structure and this spittles on decades of work from a lot of you know fantastic scientists improvements that start now allow us to map 3D variability and earth's viscosity better understand some of the important detail but very important isodynamic features and interpret signals associated with ocean dynamics or I should say really just non GIA processes a lot of work I think is still needed to actually get there on the data assimilation technique side including more physics based ice model Pippa said it right getting GIA models together with ice sheet modelers and observational data people that's the dream team getting diverse assimilating diverse types of data being creative about that getting better coverage and then connecting to some of these other disciplines that have you know complementary knowledge that we need and I didn't really motivate my work at the beginning because it was motivated through all the other talks but if we have this better understanding of course it affects present-day sea level changes it affects our knowledge about the feedback of solid earth and ice sheets as Pippa elaborated on and it also affects our study of sea level during past warm periods periods of rapid change for example in that water events that are really important to understand ice dynamics and sea level change into the future and I'm going to stop here and I'm really happy to take any questions. Great, thanks Jack that was another great talk I think what I'd like to do is let's switch to both Jackie and Pippin will take questions for both together and just go into a general discussion of both talks at this point and I think to start off I think Thorsten has a question so maybe we can start with him. Thanks Mark, thanks to both of you for great talks and I'm I have a billion questions and I'm really very excited that three-dimensional variations in viscosity have now become something that people outside the geodynamics community are excited about because we've been at this forever and we kind of had a hard time getting at it but when you bring in the sea level and you realize the feedbacks with the cryosphere it becomes much more relevant obviously for society and so there are a number of ways to get at the different spatial temporal scales of viscosity from geodesy, from geoid and from sea level exchange and I guess I have a number of questions that are to do with that maybe sort of specifically starting out and then someone else can pick up on this for Jackie so you showed these really amazing androine based inversions for relational viscosity variations which is I think exactly the right approach but I was very surprised that you had deep sensitivity for viscosity and these were lost right and so my understanding is when you look at GIA alone then the length scale of the ice load gives you a pretty good idea about how deep you're sensing and then you do sort of a fairly localized average. Now GIA with the sea level equation then brings in the geoid response so I presume it's this geoid response and the flexure of the lithosphere that Pippa really nicely illustrated that gives you a bit of a deeper sensitivity but when the geonemesis tried to get at lateral viscosity variations from the geoid or dynamic topography it's always like because those are mainly sensitive to density rather than viscosity and so the rates are much better so I wonder what is it when you go to sea level that then gives you sensitivity that also appears to be sort of less localized. Yeah I would say two things come to mind here one is that what you just described we think about and the rule of thumb is the sensitivity often the depth sensitivity of the ice sheet is kind of the size of the ice sheet roughly right so for the Laurent height people say I think a thousand kilometers the depth sensitivity but this applies to if you look at viscosity in the rebounding area but we are not looking at I mean we're not only looking at area at sea level in that in the same location where we also have that low change so if you were to map out the sensitivity kernel for a site for a sea level record in Hudson Bay that's exactly what you would get you'd get sort of a sensitivity kernel that extends to about a thousand kilometers but since we're measuring kind of further away we are getting this broader sensitivity the other part is that it's not just the ice load that changes over the glacial cycle but it's also the ocean load and the ocean load is a much larger spatial feature and contributes to kind of a greater depth sensitivity that is sort of my question does ocean loading right that gives you a georic kind of feeling thing that is but why does that better than you know looking at say the effect of internal density distributions is it because the forcing is different because when we do internal density variations and internal viscosity variations the geoid doesn't much care right but your sensitivity analysis says that when you look at these surface loads it seems to be more sensitive to lateral viscosity variations yes sorry is your question whether it's the trade-off with density? No the question is where does the sensitivity come from compared to other geoid forward modeling studies that mantle dynamics have done yeah I mean I would say it's the combination of both of those I mean you're seeing if you look at the if you think about the figure that I showed you with the point load of your ice sheet and the point load or the point measurement of your sea level record getting a kernel that is sort of banana donut shaped and so that just gives you I mean there's trade-offs right if you can accommodate all of that in the upper mantle but you are getting sensitivity in the lower mantle from that as well I think next we have a question from Maya I actually have a couple questions but first of all thank you to you both that was such such fabulous informative and clear talks thank you I'll start with Pippa and both my questions are very simple compared to Thorsten's they're really just for my own clarification so Pippa you talked about how this difference in viscosity is going to affect whether or not the glacier retreats because of how it responds and I saw I know this is buried somewhere in your plots that you were looking at sort of maybe a thousand year time scale but could you expand a little bit on like exactly how much difference it makes in terms of time and how quickly it is responding to these differences in viscosity yeah I think actually what you've asked there is what we're grappling with at the moment the some people are looking at very short time scales and often they're running models which can run at higher spatial resolution and so you're sort of resolving the processes the detail of the rebound if you I think it's important to say if you run ice sheet models at different spatial resolutions you'll get different answers some of those ones which have run for very long time scales are fundamentally slightly different have different sensitivity to the forcing you applied to the climate forcing essentially so that's just something to make you aware of the what we think is happening is somewhere like Pine Island in Antarctica is that we can trigger a response to ice loss within a decade which is a couple of orders magnitude greater than just the elastic rebound which we traditionally assumed was happening and we know that because the GPS are recording rates which are a couple of orders a couple of times greater two or three times greater than the elastic response we're building that into models that have run over decadal and centennial scales but we're very much looking at specific processes there we're asking the question of what can the earth do other models will be thinking about ice shelf processes ocean processes in there as well and I think it's fair to say that there's lots of different groups tackling these different forces and some of those longer time scale ones necessarily try and combine all of those but maybe with less detail I think there's not surprising that we're getting different answers for long term projections at the moment thank you as one of the things I really loved about your talk is you so clearly illustrated how complex it is and how dependent these different feedbacks are on each other and so I guess that completely makes sense given that overall theme in your talk so thank you and I'll just switch can I ask one more question of Jackie go ahead with your seismic tomography models you know one of the things that always makes me nervous about seismic tomography models is their blobbiness particularly in the shallow mantle and particularly under the ocean basins where where we have so few measurements I mean obviously you can measure it other ways with bances but how confident are you in those shallow models and does getting the shallow viscosity right does that disproportionately impact your results versus the deeper viscosity yeah so I mean in the if we use a viscosity derived from seismic tomography really at the mercy of the resolution of the tomography right just through the process of converting to viscosity we can't really add any more resolution to it that using the sea level observations as independent constraints on the viscosity might be able to help with that but I think the resolution that we're going to get and we saw this a little bit in the synthetic case the resolution that we're going to get is also going to be very blobby or very smooth out right it's really the first order features that we possibly can start to resolve the question of whether that is a concern or something that is particularly important I guess it depends on where you're looking if you're looking in Antarctica and you're looking at a location where you have your ice sheet melt and you have your station right next to it you're going to be very sensitive what happens in the upper mantle there if you're looking at a site that's further and that also applies say looking along the U.S. east coast the upper mantle is really important the further you go away from the ice load the less important that sort of oceanic upper mantle structure is and mapping out these sensitivity kernels can actually help us answer that exact question because they're telling us which part of the of the mantle we're seeing that observation is seeing so I would say it depends on what you're looking at maybe actually just to segue off of that Jackie in terms of the spatial variations there's a question from the audience whether you can comment on the sharp spatial variation in lithospheric thickness in Africa 50 to 300 kilometers what's the evidence for it and what might have caused it yeah so we're seeing some of the strong gradients in lithospheric thickness once we move on to kind of a continental craton so you know this the models that I showed I derived from seismic tomography and they sort of map out the craton in those regions which just have a much higher you know thicker core you know thicker lithosphere then the the surrounding regions so that drives those kind of relatively stark spatial gradients great thanks we have a question from Jessica Warren thanks Jackie and Pepper for a great talk so it was really interesting and my talk kind of follows up on what Maya was asking about which is I'm curious with upper mantle rheology you know if you for example put in dislocation creep which I think the seismic data suggests in many areas is dominating what does that do and maybe the other question is why the preference for diffusion creep is it it's computationally easier and just in terms of like moving towards a closer rheology for what we think is there can you know can we go in that direction of putting dislocation creep in I can start on that but I'm also happy to then hand it off to Pepper because I think particularly in Antarctica people have looking at that more than in other places in general I think that the choice of using Maxwell is just that you know start with a simplest model that can explain the data there are definitely approach you know it's not actually that I mean that more computationally expensive to run different those different realities that aren't that are still one-dimensional so it's possible and people have are doing it you're sort of adding you know more free parameters to the inversion but I think it's it's necessary and ideally they're not fully free parameters but there are parameters that are actually constrained from by the you know the laboratory experiments but yeah the initial choice for Maxwell is just for starting with the simplest model and people have looked at this in particular in Antarctica so I don't know if you want to add to that yeah I it would be good to just flag up some work so I I run a 3D GIA model in Antarctica working with Vata van der Waal where we do include dislocation and diffusion creep as Jackie says it's it's computationally more expensive a problem that we hit there is how to parameterise things like water content and grain size we the model we have at the moment we just choose a uniform value for water content of the mantle underneath Antarctica partly because we don't know any better we find very different estimates of viscosity depending on that water content which obviously has knock on implications for the GIA model an area to flag up I'm not super familiar with it but I'm aware that Magneto lyrics can tell us a little bit about water content in the mantle and and luckily people in that field are looking to Antarctica and potentially introducing some observations there which will help constrain that bit of the problem but yeah absolutely thinking about temporal variations in mantle viscosity so the dependence of viscosity and stress is something which is sort of high up on the radar in Antarctica because we're able to get these areas of rapid ice sheet change and see the rebound and we can't quite model that rebound with a Maxwell rheology at the moment thanks I but I'm going to cut them off I'll just say that I was happy to hear water invoked as well because that seems like one of the easier parameters to allow you to have viscosity variations literally as well I think I'm going to go we have a question now from Cindy Cindy Evinger yeah I'm whoops I don't know sorry something odd just happened I was going to ask both of you how we I mean one of our roles on the committee is to find ways look for intersections or ways to support the community and it seems that particularly with both of your talks that you have strong advice to where new observations should be collected where new stations should be installed and so is there adequate dialogue between the observation list particularly say Antarctica where there's a specialist community and I'm just saying I've heard a whole series of talks on Antarctica related to research of colleagues here and I and I'm not what just seems like there's potentially disconnects and wondering how we could help with the community share the modeling results to better inform data acquisition so I'll pick that up because you mentioned Antarctica there Cindy I think you've you've almost flipped around the answer that I was going to give by saying can we share them those model predictions and that's that's actually an interesting direction and there are there are studies more around the northern hemisphere ice sheets asking the question of where could we get data that would help our modeling approaches that's not really been done so much for Antarctica I think the general approach is if we can get any data which is grateful because it's logistically so challenging but I think that's an interesting direction to be more targeted in identifying specific observations that would be helpful for the modeling would be interesting and Antarctica is an interesting location it's obviously very challenging to get to and carry out field work but it's actually also an area where perhaps there's the most collaboration in the world because of that challenge so starting those conversations is actually you often get quite a long way there I can add from the sort of far field perspective there is some work that as specifically what are the areas that are most sensitive to the parameters that we want to constrain and then go there and and look at observations but of course you're constrained by not every location has the data that we're looking for not every location is accessible it's not there's a reason why most of around Africa doesn't have any of the Holocene data in this database it's a little bit of a difficult but there is definitely that conversation and I would say a little bit less so on the where should we go to collect data but a lot more what's happening more and more over the last decade and hopefully more and more in the next decade it's sharing just starting to have that conversation between the observational science and the data modeling so the Holocene what I mentioned just collecting compiling the data in a way that is meaningful and useful for the modeling community and vice versa producing model predictions and making them available that then are also useful for observational scientists I think that's the stage we're probably at at the moment and it's picking up which is really great to see. I think we have time for one last question Jeff I think you had a question. Yeah there actually my question was almost exactly the same as Cindy's so I'll just say thanks to Pippa and Jackie for great talks really appreciated it. Great actually I'm going to use my I'm going to do it right here for just a quick second to ask you guys both one of you can maybe explain something to me that I've always wondered about so we often in like intro talk just about you know if you have an ice shelf and it breaks off it doesn't change sea level right because it's already floating but thinking about these fingerprints I'm curious how does that map is that fingerprint talking only about grounded ice that's melting I would assume though that that gravitational effect would actually also pertain to floating ice directly adjacent to an ice sheet is that true I really don't know the answer I'm very curious what it is Shall I have a go at that one Jackie? So my way of thinking about ice shelves is that they're already in the ocean so obviously as you said if they break off they don't change the sea level so I'm now struggling with the concept of the gravitational attraction of the ice shelf and I have a hunch that it doesn't affect it because it's displacing ocean water which would I would say yeah I'm struggling with this answer you're absolutely right that when they break off they don't change the sea level I don't think they are massive enough to have the same sort of effect as a change in the volume of the grounded ice sheet behind if you think about them then transporting into the ocean it's just mass moving around it's not mass changing it melts into the ocean essentially and what the models do is they do account for the self gravitation of the ocean itself so I think I may be getting back to my first statement is that if you count them as being part of the ocean then it is accounted for by these models of considering the distribution of mass throughout the system so even though an isostatic canal mountain range would deflect the gravitational attraction due to the floating that wouldn't be the case for an ice shelf I think so I think that's right it's the same mass and the mass is in the end what matters for gravity I would assume that it doesn't have a change on gravity and since you mentioned mountain glaciers if the ice melts and you allow the mountain to rebound there's actually also no gravitational effect but because there's a delay in the rebound you'll have that initial change in gravity that then is slowly being made up and I would also add which isn't exactly what you asked but the ice shelf is an important betracing role so as soon as you lose the ice shelf everything else can go great well thank you both for both great talks and stimulating discussion I'm going to throw it back now to Steve Nair who's going to do a quick wrap up for the day right thanks Mark well these talks have all been great and it's a lot to summarize I'm going to take a stab at here and I apologize if I missed some points but I think what we heard today is that first satellite measurements are providing a lot of new information about sea level change including sea surface height and ice sheet height and ice sheet mass changes that's telling us that sea level is rising and the rate of that rise is accelerating and that this is being caused by a combination of melting ice and expansion of the oceans both of which are being driven by global warming there's a lot of regional variations in sea level rise due to the figure prints of the ice sheet mass loss on sea level and variations where the oceans absorb the excess heat from global warming so I think we also heard that the big uncertainty for projecting future sea level change the elephant in the room someone said is determining how quickly the large ice sheets will melt studies of relative sea level data such as tide gauge data that are important for studying both the solid earth geophysical effects as well as sea level change and understanding isostatic adjustment, tectonics sediment compaction, subsidence etc so by combining this relative sea level data with knowledge of solid earth geophysics we can make progress towards not only understanding the solid earth but also ocean circulation and ultimately climate change changes in the ocean currents and circulation can have major impacts on sea level change along the coast for example changes in the Gulf Stream will have a big impact on how much sea level change we see along the east coast of the US we also discussed that we don't fully understand the feedbacks between ice sheet dynamics and solid earth deformation but we know that 3D variations in earth reality matter and it's going to be important for determining how solid earth deformation impacts ice sheet dynamics as well as the regional variability on sea level change due to the ice sheets and the fingerprints they make on the oceans and sea level the impact on ice sheet dynamics will depend on how quickly the solid earth deforms due to the ice mass loss and what the spatial pattern of this deformation is and so mapping mantle viscosity is critical for understanding these interactions and their feedbacks we also learned how paleo sea level records are being used to understand the earth's internal structure and the drivers of ice contributions to sea level change distribution of past ice sheet loads is a major air source here and is important for modeling glacial ecostatic adjustment and other effects approaches from seismic tomography have been used to study the earth's internal structure and so viscosity tomography is a new tool for doing this somewhere starting to map 3D variations in earth viscosity using these techniques however we also need to consider other changes in sea level not related to GIA such as ocean dynamics and so it's going to be important for scientists from many different fields to work together to solve these problems of understanding how the solid of this contributing to ice sheet dynamics and I would say in closing personally I think sea level science is very much an interdisciplinary problem and so getting oceanographers and geophysicists and geodesists and others working together to understand these problems it's going to be critical to move forward so I'll stop there and just say on behalf of the committee I want to thank all of our speakers today they were all great talks I'd also like to thank everybody in the audience for joining us today and the questions that we got so I also want to know that tomorrow we're going to be continuing this meeting with a few talks on vertical crystal motion on how that contributes to sea level change those will begin at noon eastern time tomorrow so thanks everybody and we'll see you tomorrow