 So, thank you all for joining us today. Welcome to the course webinar on Subduction Zone Initiatives. The course committee has asked three presenters to provide information about new NSF research coordination work that they're in. Harold Tobin is going to start from the University of Washington. He'll be speaking on a research coordination network for the SE4D initiative that followed by Torsten Becker from the University of Texas at Austin. He'll be discussing a modeling laboratory for Subduction Zone Science. And we'll finish it off with Tobias Fischer from the University of New Mexico, who will be presenting on a community network for volcanic eruption response congress. Each of the speakers will have about 15 minutes for their presentations, followed by three to five minutes for questions for each presentation. And this would be a really good time for clarifying questions for the specific presenter. Joining all of those three presentations, we'll have about 30 to 45 minutes for questions for all the speakers, Harold, Torsten, and Tobias. We'll ask for questions from the course committee members first. We ask that all of you remain muted while the speakers are presenting. During the question period, please use Zoom's raised hand feature if it works, and we'll unmute you in order. If it doesn't work, please feel free to chat to me, Deb Glickson, or Eric Edkin. Either one of us will be available and monitoring the webinar. With that, why don't we start with Harold? So Harold, if you would start sharing your screen. Harold, you're muted. Let's unmute you. How's that? Can you hear me now? Yep. Okay. And can you guys see my slides? Yes, we can. Okay, great. Well, good morning, everybody, or good afternoon, priests, coasters. I hope everybody enjoyed their holiday. First of all, thanks so much for the opportunity to present this information on this kind of developing, nascent planning program for what's called SC4D, or Subduction Zones in Four Dimensions, if you will, or perhaps through space and time. I'm going to be talking about the overall SC4D project as it stands right now and giving kind of the overview, as well as the more detailed description of the research coordination network that's been known as kind of the SC4D umbrella, or just simply SC4D research coordination network that was funded by NSF. So I'm kind of speaking on behalf of the steering committee that was set up for this coordination network. You can see all the names of the steering committee members on the slide there. It represents a number of researchers from across the country, from across a number of different disciplinary areas related to subduction zone science. We have a lot of people from academia, but also from USGS and NOAA involved in the steering committee. So it's a pretty broad group of people. And I'm really, I'm the chair of the steering committee and just the spokesperson for this group. Let's see how, there we go, making the slides go. So as you probably are aware, the SC4D initiative really grew out of a subduction zone observatory idea or push that started with some, some events that were at Iris and Unavco meetings going back six or seven years and kind of culminated in the organization of this very large workshop in late 2016 in Boise, Idaho called the subduction zone observatory workshop. Something like 250 scientists from around the world came together at that meeting and really looked at, you know, in the context of exciting new developments in our field of science and also in the context of the results that have come out of big decadal scale programs like Earthscope and Geoprisms and others, margins before that. What were the exciting things in science that would, that would lead on from that. So Jeff McGuire and Terry Plank led a writing team that a number of us were on to produce this vision document that was really the outcome of that meeting. And I think, let's see, you probably can't see my mouse if I slide it around, but if I use this little fancy laser pointer function, maybe you can see that. So, you know, the document that was written is called the SZ4D initiative, this phrase emerged in that meeting, understanding the processes that underlie subduction zone hazards in four dimensions, a vision document. We just tend to refer to it as the vision document. You can download a copy if you guys haven't already taken a look at that at our website. Some of the key things I think that emerged from the Boise workshop was a focus on saying there are many, many, of course, different processes at many different scales in time and space that go on at subduction zones. But one of the themes that kind of unifies them and unified the consensus of that very large meeting to the extent that a consensus can form at such a big meeting is a focus on this idea of fundamental basic science processes but that underlie the hazards that subduction zones present. So a close link to the ultimate societal relevance but still a focus on sort of NSF-style, NASA-style discovery science. Another key theme that emerged was the idea that one thing we're doing now and in the past decade or so that is really a sea change is looking at time-evolving phenomena and the power of observational science over time. And that ranges from sort of very short time scales of, you know, minutes to hours through months, decades, years, and even geological time. That there was opportunity because it's, of course, a global question and problem for a great deal of international coordination and that any kind of field program and or observational science had to be very intimately tied with modeling and experimental collaboratories along the way. So we'll talk a little bit about that. So, you know, these phrases come straight out of the document targeted experiments to make next big leaps in understanding these processes and comprehensive suites and measurements taken in the same place over the same time or maybe it should say same suite of places over useful time scales. And so, you know, I think one of the things that came through loud and clear was and this also came out of discussions with NSF was that a futurist subduction zone kind of focused research program had to be just that had to be focused and targeted subduction zones are such a large chunk of solid or science that a kind of something that was simply a kind of covering the entire waterfront of all subduction zone science was not really what could emerge as a focused and decadal scale program. There's still lots of room for subduction zone science and all sorts of core NSF programs, but that this had to really represent something with a thematic focus that was very clear. So, at the same time as this SD-4D vision document was emerging, a number of other reports, of course, were taking place. This erupt report from the National Academies is underlies a lot of what Tobias Fisher will talk to you about in a few minutes. The USGS was pushing forward on their own very substantial kind of focus subduction zone science report at about the same time. NASA's new decadal plan highlights the number of subduction zone science programs or potentials with new space missions like the NISAR radar satellite, new GRACE follow on and other other programs. And then, of course, at NSF, the end cycle of both Earth Scope and Geoprisms kind of presented the group with an opportunity to really think about what might come next and how to reimagine what's exciting for the next phase of big subduction zone science. So, this is taken straight from the table of contents of the vision document, big science questions, motivating the subduction zone initiative, and you can see what they are. They're thematically focused on a few of the key things that are large enough questions to really be worth tackling over a decadal scale or beyond and yet also tied back to those geohazard problems, like when and where do large earthquakes happen, what the processes are. Similarly, how is magma production connected to volcanic eruption? How do you go from magnetism to eruptive processes and what are the time scales and processes that do that? Variations in subduction inputs affecting both of those things. So looking at the sort of the spatial scales of variability and how do the surface processes that represent hazards from landslides and other things link back to the subduction problem and the crystal deformation. So, the key components in interdisciplinary science program, in infrastructure and instrumentation program, I'll talk more about in a second, in the modeling and laboratory and experiment collaboratories. So the goals, these are just some sort of a quick time series examples down here. One is the deformation using synthetic aperture radar data, seguam volcano over a kind of a decadal scale leading up to eruption. The other is just a several day time scale of foreshocks leading up to the Tohoku earthquake in Japan in 2011, but looking at different ways in which we are now able to capture and model these emergent phenomena and use those for really fundamental insight into the processes that underlie these things. So again, a focus on collecting 4D data sets from real time to geological time. So in the vision document, there's an implementation set of recommendations that were developed, I would say in a fairly close consultation with NSF on sort of the feasibility of moving forward. And so there's a timeline there with integration, SC4D Working Group and Steering Committee, that's essentially the umbrella RCN that NSF kind of provided us with the RCN process or vehicle for funding this planning process that has been developed in order to tackle these interdisciplinary science program questions that are kind of shown in here with the vision document starting in 2017. The NSF process led us to funding of the actually the Steering Committee in late 2018. So we're not quite on the timeline that's here, but we're not far off of it either. Torsten Becker will talk about the community modeling collaboratory that was identified as a key sort of, you know, top level component of the whole SC4D initiative process. And then the questions of how we will plan towards large-scale infrastructure, which is one of the biggest questions. There's a lot of recognition that infrastructure, meaning instrumentation in the field and potentially, you know, large expense ambitious programs are exciting to the community and potentially one of the key unifying features in the way it has been for programs like Earthscope, but are a very big challenge to undertake. So those three elements, the science program, the modeling collaboratory and the infrastructure program, all recognized as important to this. So what NSF did then was fund us for these RCNs, which are, and I want to really emphasize this, this is the planning process to turn the vision document into a potential implementation plan. The vision document just lays out the science objectives and the kind of really big scale let's reach for the stars kind of goals, but turning that into a blueprint for how this might actually take place as a new program or new initiative, sort of capital I initiative at NSF, NASA and other agencies is the task of the research coordination network. It's very much a task that is just still in its relatively early stages. And so many of the questions about what it will actually look like are what the RCNs are designed to address. So the planning that's funded are these three things you're hearing about today. NSF committed a significant amount of money to these planning activities over basically these three years through 2021. And we were very pleased to see support from five different programs in both EAR and OCE, suggesting a broad buy-in to at least the concept that planning for this SD4D initiative and looking at what it might entail and seeing whether that fit NSF's future goals made a lot of sense. One piece of implementation that's worth bringing up and you've probably heard about this is this recent news that the MSRI program and the director's office at NSF funded this seafloor geodesy instrument pool as a first step towards some of the things we're interested in in the SD4D program at about five and a half million dollars just in the last few months that was announced. Okay so the umbrella RCN has these fundamental goals, you know turn the vision into an implementation plan for an actual initiative as I said. It's really important to note again that the scale and the scope of it is still an open question and it's a complex one. We've got we've had a lot of discussions with our program managers at NSF that have suggested you know the implementation in particular for something like infrastructure could range from you know something quite modest through the tens of millions of dollar kind of scale and then potentially one could dream all the way up to the MREFC scale that you know that was something like what Earthscope did or big ship projects do at the hundreds of millions of dollar scale but we don't know and we don't know what the lay of the land will be as we move towards this. It's our task to make the argument. Will it be an ongoing research funding program? Some of the ones that exist now. What about data management systems, collaboratories and so on? So our task then is to build the community and we're doing that with sort of two nested levels of groups that I'll talk about in a second. The first is to basically engage everyone with interest in it in these thematic interest groups surrounding the key big questions that I laid out a few minutes ago and then specific detailed working groups that will be much smaller groups of people that we're staffing right now with a dozen to 20 people on each one that have the detailed task of actually designing the program taking all that input from the thematic interest groups and writing the essentially white paper or collection of white papers that would represent a kind of an implementation plan a proposal if you will although not a standard style proposal with a capital P to NSF and so working with the partners in fact to design that viable and compelling program and make the case. So just some of the examples of the types of things that were talked about and these are all directly taken out of the vision document you know one idea would be comprehensively choosing a few key areas around the world where mega thrust systems exist where you know that there is either potential for very large earthquakes or there's hypotheses that these are places that the subduction thrust is somehow different from places that are locked and have earthquake and synomagenic potential and instrumenting them in a detailed way onshore and offshore to understand the locking the processes that are going on inter-seismically and potentially to actually capture say a magnitude eight scale or above earthquake in the act if you chose enough places around the world then statistically one might be able to do that over the span of let's say 10 years so an integrated network you know seismic ocean bottom seismometers as well as land stations GPS or GNSS including offshore which is what the A for acoustic GNSS is as well as onshore ocean bottom pressure sensors active source seismic reflection and refraction imaging high resolution bathymetry basically all the geophysical tools one could bring to bear on this as well as the kind of geo history meaning the structural geology through both onshore and offshore studies open questions how broad a region and how many regions might be needed to capture those kind of events on a large scale and where in the world are the right places to go those are very much part of the planning process not preordained in any way the motivation that I'll do very quickly through is just that you know recent events like the ekk chili earthquake that was in 2014 showed run-ups in seismicity foreshock sequences before the main shock earthquake which occurs here over the span of two weeks and GPS evidence for for displacements in the period leading up to the earthquake as well suggesting slow slip events that are precursory to to the main earthquake Tohoku had some interesting similar features we'll talk about volcanoes in just a second and come back to this one but new time series evidence that are really driving the frontier of science on understanding how these mega thrust fault systems work and here's just an example from Mavramadis said all of GPS displacements over the span of a number of years actually prior to the 2011 Tohoku earthquake in northern Japan suggesting that that you know pre-earthquake precursory phenomenon were going on that had not previously been detected open questions of course about how common persistent that is and what that's actually telling revealing to us about fault structures offshore and this is the Nankai trough of Japan so southwestern Japan now seafloor geodesy has really revolutionized our understanding of the locking of the plate boundary all of these are offshore GPS stations and all the small gray arrows are the onshore stations and they show as we knew already that the subduction thrust is fundamentally a very well locked system similar for example to Cascadia but that that locking is patchy and has a lot of complexity to it in both time and space so the inversion of the locking of the fault offshore just shows in these color scales here as kind of not uniform as was previously thought but patchy across the whole subduction thrust system and in fact very recent results still in pre-print form are suggesting that slow slip events are actually taking place up in the lock zone offshore similar to but very distinct in location from the down dip ETS or slow slip events we've all heard about before similarly to that for the volcano side of the question the group has been discussing and the vision document suggests an arc scale volcano observatory in one or more places chosen carefully again to represent key key arc processes this happens to be South America this is Alaska illusions over in here and looking at common instrumentation on a number of volcanoes gas sensing and some of the other things that I just mentioned for mega thrusts and look at eruptive processes on a decadal scale with these common instrumentation networks motivated by all sorts of evidence of time series of magma scent and recharge processes prior to eruption that we are seeing now and I will talk about them in much detail except that chemistry and seismicity have been known for a while as pre-eruptive events petrology of crystal clocks and understanding diffusive processes in these are for example in olivine crystals showing water loss that can be used to time magma ascent down to time scales of minutes to weeks as opposed to many years and some of the other more traditional dating methods the geology is an important component of the whole thing and so moving quickly through again just recognizing that there are a number of ways to use surface geology as proxies for different down-dip locations this is Christy Rose beautiful compilation of a number of field sites that represent different places down the seduction interface and I'll just kind of summarize that by saying that there's this understanding that or this emerging view I think that it's a holistic approach is needed for understanding how faults actually lock up why slow slip exists why rapid slip exists and what controls the boundaries between them looking at a holistic approach to mineralogy, lithology, rock fabric stresses and friction this is blue schist, neclogite, brittle and ductile materials intermixed right around the lower boundary of the of the seismogenic zone so that really is a link between the geology field imaging of deep subduction processes or deeper megathrust processes and then using experiments to understand these friction and stress questions and how fabric relates to those these are the frontiers for that kind of work and surface process is similarly just a very brief vignette example these are not meant to be remotely comprehensive for looking at you know this is a some work that's going on here at University of Washington using synthetic ground motion simulations for a Cascadia earthquake and then examining the coastal region for evidence of paleo landslides and looking at evidence that whether or not landslides were triggered by ground motion during the magnitude nine earthquake inferred for 1700 AD and recent work suggests that there may be a peak in landslide activity tied with some really interesting new lidar methods for dating surfaces to the 1700 earthquake okay so and what are we doing now I'll try to wrap up in just a couple minutes here sorry I'm going slightly long we have a website that I encourage you all to take a look at at sc4d.org it's evolving quickly we're still adding a lot of material to it but we have links I'll show you there to the thematic interest groups that we're using to build our community and we have hundreds of people who've signed up for that in recent weeks for the interest groups and we are in the process of using self nominations and some of our targets to invite people for those working groups we've had a number of recent events we have upcoming town hall at AGU next Thursday a week from Thursday a special session at AGU and we've been promoting and kind of getting the word out about the sc4d planning process at all these different meetings if you scroll down the web page you see these interest group links faulting in earthquake cycles is one interest group magmatic drivers of eruption landscapes and seascapes is the phrase that our group has used for surface processes and subduction zones and then the international partners working group enlisted as well all three of these were building the direct working teams to start writing materials we have an international partners group it's sort of nascent group meeting informally at AGU and then we are going to build off of that over the course of the next next half year or so we want to get our domestic kind of large-scale wireframe of the plans in a row before we go too far down the international partners direction just because of location being an important key question and that image that was used for that international partners is just this one suggesting look if you want to do big scale infrastructure implementation there are a number of places around the world where you might think about both volcano and megathrust instrumentation and the onshore offshore components of all of that in order to comprehensively study the subduction zone it's hard to imagine that you could do the whole world with the kind of density of the instrumentation that would really address these questions so talking about which targeted experiments and which parts of the global scope make sense for a U.S. program which parts are being done well already like in Japan perhaps New Zealand which places are more and less just sort of practically feasible in terms of politics data access and other things so there's a lot of questions there that still remain both for U.S. subduction systems as well as other ones like South America down in this region Mariana's and so on okay so over the next two years the timeline that we're looking at is basically right now we are populating these working groups to get specific and tractable plans started we have this international partners group meeting at AGU in this town hall meeting in January we're planning the first working group meeting we'll bring them together about 60 to 70 people in Albuquerque for three days to kind of give the marching orders to the working groups and start that process of that really happening then over the course of the next year all through 2020 the working groups will both seek input from the community on priorities potential experiments themes and focus sites via a number of vehicles online forum surveys town halls etc meanwhile they will also meet online regularly and then eventually in person next year again to rough in that developing implementation planning documents so lots of decisions to be made lots of process to take place but that's the kind of process we're looking at by 2021 we'll have another all hands meeting with the steering committee in the working groups together to work on the final kind of white papers and implementation blueprints and perhaps we will be able to target some science workshops on the biggest scale questions although that will require some additional funding because that couldn't be built into the scale and scope of the RCN funding and finally by fall 2021 finalize those implementation plan documents so just to wrap it up you know the key challenges I think in thinking about a coordinated SC-40 program are this need even as we plan for flexibility because we don't know what scale and scope will actually emerge as really feasible we need to sell maybe multiple plans you know depending on kind of the lay of the land as time moves forward here we need to use opportunities as they arise so for example the mid-scale research infrastructure call that that C4G Odyssey was a response to this year meaning not just creating something out of whole cloth but using some of the opportunities that already exist and be kind of flexible towards that and the need for focus I'll sort of end with that that we're not simply here to duplicate core program priorities this can't be all subductions on science for all people if we want to make it I think a successful and focused program that's been a key theme both from the meeting in Boise but also from our discussions within SF along the way okay well with that I think this pretty much just re-creates re-creates some of those those statements I just made this certainly aspires to be one of the next big things in post-earth scope post-geoprism's world but in no way is it already an NSF program this is a planning process only and really we have an opportunity to create something new that builds off the best parts of existing programs but is not the sequel or continuation of any one of those that it must be in our minds amphibious earth and ocean science and the scale remains to be determined but the community drives the whole thing okay that's it for me I should probably maybe stop my sharing thank you so much Harold that was great let me start with asking if any committee members have maybe maybe we'll just take one question now and then we'll come back after are there any questions right now can I ask a question there we go hi Carolina yes absolutely yes hi Harold how are you good good hi yeah and hi everybody yeah so one of the things that I'd like to understand better and you said it that you don't see this you see this as the next or you the community I guess see this as the next big thing but not just a continuation of Geoprisms and Earthscope right but I'm not entirely clear on how it's not an extension and growth and you know I'd realize that it's a you know rather that a continental scale has got the interface but still you know how is it not just that yeah I mean I guess one statement to make that's probably obvious is that those programs have been very successful and they've led a lot of the science that leads up to this so of course to some extent it's an outgrowth of all of those things we're informed by them and there are only so many models for great great science programs I suppose out there but I think the key element here is is this decision to make a focus on on time series science and a renewed emphasis on this kind of observational experiments that are very targeted so you know Earthscope took on very very large scale instrumentation and was focused heavily on that and then used that of course to get a lot of measurements that were that were revolutionary in many ways in our science geoprisms had by contrast no you know infrastructure program at all right it's only been a science program SCAC is yet a third model where you know it's it's building a kind of a center that actually in a sense runs its own science program with many different individual p.i.'s involved in it I think our challenge is to take kind of the best of all of those and create a program that is that is targeted that is just the right balance of infrastructure and science program and goes after some pretty focused questions and tries to pick the best geographic locations to address those focused questions rather than simply just trying to cover a large geographic region with as much instrumentation as possible for example so yeah it builds from those but I think the objective really is to create a small set of key hypotheses and then drive time series observations married to modeling and experiments that will answer those questions in a reasonable length of time thank you thank you both so much why don't we move on to Torsten if you can share your screen Torsten all right okay thank you so much thanks a lot Deb and thanks to the committee for giving me a chance to talk about the modeling Collaboratory for Subduction Zone Science and thanks to Harold for introducing the activities of SC40 and providing a lot of the background for related efforts in the community I'm speaking here as the PI of a research coordination network that is to discuss how the modeling Collaboratory might come about and I'm speaking on behalf of our steering committee where I where we've listed here the we have listed our great team just like Harold's we have folks here from the USGS from academic institutions and we're trying to to span a range of different disciplines from marine geophysics to landslides from volcanoes to earthquakes and I'm also speaking here on behalf of our program manager Gabe Lotto who's been instrumental in getting a lot of these things done and Gabe is actually someone who's working on the interactions between earthquakes and tsunamis so we're really happy to have him here now a little bit following up on Carolina's question what motivates thinking about a modeling Collaboratory are some of the lessons that we've learned from initiatives such as the ones that Harold just mentioned and I think the solid earth community meaning geologists geophysicists patrologists geochemists could really benefit from an integrative approach that has an understanding of the physical processes as its core as we collect this data we should proceed in a way that allows us to embed those constraints in an understanding of the mechanics of how things work and getting such a mechanical or physical model will in turn help us to design better experiments to actually answer the scientific hypotheses that we're after and so to illustrate some of the amazing data sets and some of the amazing constraints that we've been able to arrive at things to investment such as earthscope I'd like to show some of the science examples that are really motivating our joint desire to move ahead here and arrive in a new understanding what's happening in the subduction system what you're seeing here is one estimate of the genetically determined vertical rates over North America where you can see that along Cascadia we have subsidence on decadal scales overall even after corrections for GA and to some extent hydrological features this subsidence could be related to being late in the viscoelastic cycle as we think we are for Cascadia where you can show that you can actually have a little bit of subsidence due to viscoelastic effects just because of an earthquake cycle however for North America also based on earthscope results we know that there are positive poency anomalies negative sorry negative poency anomalies along Cascadia in the mantle in the asthenosphere where convective flow pulls things down in a way that is of course associated with present day subduction but also goes a little bit further south so we have one component of change in topography that is due to long-term convective flow another component that might be related to the earthquake cycle if we're trying to use the verticals to say something about earthquakes and Cascadia we have to correct for the different distributions across the scales and we need good physical models to do so in a confident way Harold already mentioned the amazing observation that Tohoku Oki actually had the overriding plate sort of moving toward the subduction zones over time scales of years to a decade and we are now beginning to see here unpublished work from the Berkeley group that such year-long variations in what could be the locking or the loading at a plate boundary also observed for Cascadia and that region that appears to change the locking for Cascadia lies between the locked and the term as well so with these very interesting observations of transients they could possibly be related to preparatory processes one meeting that really brought this home to me was at the National Academy's standing committee for seismology and genomics where we met in the spring of this year to discuss new opportunities for the study of tectonic precursors and here is a modified slide from Laura Wallace where she explained to us at this meeting how the Kakura earthquake that happened in 2016 changed stress on the Higurangi margin and that stress change was apparently associated with the change in the recurrence time of slow slip and at the time the prime minister of New Zealand came to GNS and the institute was faced with real-world challenges to and capture these observed transients these changes in the system in a way that could be parsed into statements about seismic hazard and what they arrived at was something like a change in an increase in likelihood of a rupture on Higurangi by 25% over three months or something like that but really what what is becoming increasingly clear is that we have all these intriguing observations about changes in the state of a megathrust just using the megathrust as an example but we don't really have the physical models that can capture things across the scale and that can also be used to do something that can be really helpful that can be specifically helpful for hazardous sets so the modeling collaboratory is supposed to help with some of these challenges and the idea is that this collaboratory is geared toward understanding subduction zones across the scales including megathrust and volcanic hazards so we're trying to establish a framework and a set of models that can be used to understand the core physics but to also then take data streams such as from seafloor observatories as available in Japan as might be available eventually for Cascadia and in the next time we detect a transient like we saw for Tohoku okay we are hoping to have a model that can assimilate those data that we can convert the changes in the system to probabilities now Harold already mentioned the observations for changes in volcanic systems and to treat both as a challenge but they're also commonalities such as fluid transport and we're hoping that the tools that will be needed to generate those models will be of general utility for a range of earth scientists so the idea then for the modeling collaboratory for subduction is to establish not just a set of tools but a set of models and a community that can integrate the data from different disciplines into a process based understanding of subduction zones so we very much would like to engage with a number of observationalists from geology from geophysics just listing a few fields here and from rock mechanics from seismology from geodyssey to arrive at this collaboration across the earth sciences engaging with engineering computer science and to do this in a fully open approach supporting diversity having a shared set of models having the approach the models and the data fully documented and doing things in a reproducible way so modeling framework is what we're after but this modeling framework is not just a bunch of computer codes it's a collection of constraints it's a collection of data that can tell us about the state of a certain volcano of a certain subduction zone but it's much more than that it's also a way for the community to interact and work together and for us to train the students that are needed to do these to work on these interdisciplinary problems and so the scope of the modeling collaboratory it's supposed to be a center at some place in the US but it's clearly has a global scope for one if we want to understand subduction in general we need to look at different realization that the earth provides us in terms of the age of the incoming seafloor the stage at which we are in the seismic cycle or in the volcanic cycle and we're also so there's a mandate to integrate as many sites as many constraints as possible but we're also faced with the challenge where for example as Harold explained Japan has a very data rich environment has something like a hardware observatory going already there's fairly good instrumentation for example in South America but in Cascada we're just beginning to build something along the lines of what might you know what we already have in other places on earth and so the establishment of a modeling collaboratory can help us out from a US perspective here in terms of engaging with the international community while we're building up our hardware investment ourselves and so the idea is that a modeling collaboratory will interact with a number of global observatories and here are just examples of the different sites that we have and by building these models they can then be deployed on different sites really support a community for open science and research and training support workshops and support summer schools and support tutorials hackathons and things like that and so how are we going to do this here's a here's a sort of illustration as to what the modeling collaboratory might do in terms of providing Lego blocks Lego blocks of tools that will be not just you know numerical models but that will also be curated datasets that can be assembled in a way to be realistic for a certain subduction zone or they can be used in isolation to understand processes and we want to have robust tools that can be applied to say run a forward model of the Nankai trough using real-time data streams or that can be assembled to understand how fluid transport of the general subduction zone can explain the heterological signatures of the surface the development is supposed to happen at a center but also in a distributed way with grants to PIs in the community to work on different approaches for some problems we know how to solve them numerically others we have no clue so it's important to have a diversity of approaches the modeling collaboratory will be very much engaged with benchmarking and validation of these codes we have to make sure that our codes will be useful for data assimilation will be useful for use in actual hazard assessment which will be first for the solar earth community really I mean we're always worrying about hazard but when the rubber hits the roads the geonemesis have shied away typically from using the models a good reason for it but we have to get serious about it and we want to make sure to proceed in a way that we can do uncertainty quantification and that we can do optimal experimental design right once we have a model then we can turn it around and we can say where should we do those geological and geodetic observations to nail down the stress state on the interface in a way that's useful for dynamic rupture computation something like that and and we also want to do this in a way to really help the solar earth community to do more in terms of accessing leading edge computer hardware we're a little bit behind here when we're comparing ourselves for example to the climate people and so we the modeling collaboratory will not do this in isolation of course we are we'll partner up with other community initiatives outside sc 40 and within sc 40 we are so far as in terms of an RCN funded by NSF but we have a lot of interest from USGS and also from NASA on the high computing high performance computing side and so we will I know if you flip the Lego model around there's sort of they sort of the detailed under valley that we'll sort out but we want to establish this set of tools to do it in a way that we can do regional data assimilation right but that we can also work on the fundamental physics we can understand we want to engage with people who are asking detailed questions about an abstract setting what is the friction law doing in general but we also want to engage with the operators who are running doughnut and observing some change in the system how can we build an integrated framework high performance computing will be very important particular for access to data local computing data and modeling at the same place workflows and access to databases and we've already partnered in terms of the RCN with a number of different initiatives perhaps you know some of those you know other ones you might not know cheese is an initiative on the European side trying to take codes of PIs in the solid earth sciences and then scaling them up so that they can run on the on the biggest machine so there are some exciting opportunities here there are some challenges but I think over all the opportunities are really there we're trying to do things in a different way and I just like to note that the fundamental physics here that we're talking about the tools that we're building will be of use outside subduction zones as well right if you have a subduction zone model you just reverse the boundary condition you have a rifting model so we're not just limited here but it's the approach that I think is is is is quite nice so what what has happened so far is that NSF has funded us for to discuss the the way a modeling collaborator might come about we have a website and we have a Twitter feed which I encourage you to visit and our our aims of the of the modeling collaboratory RCN are really to to push things forward and to come up with a proposal for the establishment of a modeling collaboratory over now the remainder of the next year and we seek to do a better job in identifying the knowledge and implementation gaps right we're we're dealing with a multi-scale multi-physics problem where parts of the physics are just are not known yet this does not mean that one shouldn't build a model because once you have a physical model in the America model you can be much more specific as to asking which parts of the system which what aspects of the unknowns are important and which ones you might be able to average over and things like that and and part of that the process of understanding the knowledge gaps is to think about the interactions that we have to consider right um we have the volcano system and we have the earthquake system and in between we have these different spatial scales and these interactions for example we have inactions between the solid earth and the water layer we have interaction between the elastic and the and the fluid part of the earth the slow term the long term creeping part of the earth we have interactions between surface processes and the deep mantle and so on and um we're trying to understand well which of these interactions do we understand where do we know how to average and where do we need to do more to understand even you know which interactions to take into account of which ones to neglect we want to evaluate the state of the art in the megathrust volcano modeling communities and they're actually quite different levels in terms of the physical models and we want to understand well what are the specifics of the mcs I gave you some examples this is not a done deal but we're still trying to figure out what's happening our process to figure that out is through a series of workshops and these workshops are centered on some of the key sites in subduction zones where we have the interactions where we have the multi physics processes that's the fluid transport is one of the recurring themes matters for conduit perhaps in the three phase kind of setup but also for the interface in the deep mantle and with other interactions between say the short-term dynamics and the long-term dynamics like large-scale geodynamic models and telling about pressure and temperature in the wedge which will then affect the loading of the megathrust and so on so our workshops are trying to actually deal with the signs but then also take those signs problems and parse them into recommendations for the mcs what has happened so far is we had a fluid transport modeling workshop which had a number of folks from academia and also industry and with the idea being that this is the common one of the common aspects between earthquake and volcanoes it was a great workshop with which happened in minnesota and key takeaways included that we need a better understanding still of the processes a lot of the physics not sorted and perhaps not surprisingly a modeling collaborator might help resolve those challenges and might distill the ways that the science should proceed like independently of whatever model we want to build for volcanoes or earthquakes it's important to interface scientists and we all know that we need to learn each other's language to communicate but more than that I think it's important for folks to engage and actually learn about each other's tools in the framework of tutorials similar to what side has been doing I think the modeling collaborator with it can help with that and the community for fluid transport was really very keen on model validation reflecting that some of the approaches aren't clear yet and we want to understand how can we run benchmarks and how can we define problems to test our codes and the cost of cinema your training this was really important and the modeling collaborator again is supposed to have that sort of aspect to it now there's a report you can about the fluid workshop which has just been completed you can download that along all of the presentations and we're sharing all of the documents including our internal discussion stuff online on our webpage where I provided you the address earlier and more recently we had our second workshop the second workshop happened in Oregon on the megathrust system for that workshop we are in the process of writing the report but you can already download all of the presentations this was an exciting workshop I was sitting there writing down ideas for new student projects five a day but at the same time we had a really great discussion as to how the modeling collaborator might have and one of the things we did is we identified some of the intermediate next couple of year model specifics that would really help this the signs we identified that a viscoelastic model with fluid transport would be great we thought a global geodynamic model three dimensions getting at the pressure temperature conditions with two phase flow and production zones was something that's doable what might also be sort of in a five to 10 year times that we doable was would be a flexible communique community code including rupture and earthquake cycle and cycle processes and it was thought that a hierarchy of models going from conceptual to simple to complex would be really important right we don't want to build a single black box model for say cascadia quite the contrary we want everything to be fully transparent so that folks can look under the hood and we also want things to be flexible in the sense that if you don't like the thermal model you can switch it out if you don't think that the seismological constraints are good enough to say anything while you can still use the tools for fundamental tests so it's really important to have this hierarchical approach to be flexible and have to individual Lego pieces be robust and you know we want to explore multiple realizations of models right so those are our two workshops which already happened what is coming up is our volcano workshop the location is still to be determined and it's supposed to happen in July of next year so we can go progress there we will also start a series of webinars focused on the way that people collaborate where we're trying to bring together geologists geophysicists geodynamicists seismologists and have sort of a fireside chat to see what has worked what hasn't worked in terms of multi disciplinary research also trying to cover some of the aspects of high performance computing and some of the nitty gritty stuff that the workshops are not covering as the workshops are focused in the science bombs so we are on track the way it was envisioned in the sc 40 report and we're hoping that we'll move toward a proposal submission on the modeling collaboratory however that might work by the end of 2020 and so that's it from my side I guess we have time for quick questions and we'll return to everything later Yep thank you so much Torsten let's see let me start and ask if there are any committee members that have questions no is there anyone else that might have one or two clarifying questions for Torsten you might have to unmute yourself if you'd like to ask a question or send a quick note to me or Eric okay hearing none why don't we move on to Tobias we're going to get your slides up in just a second hold on just one minute Tobias can you see your slides I can see my slides yes all right why don't you start us off okay and just say we'll switch the slides for you okay great so good morning or good afternoon I'm Tobias Fischer University of New Mexico I am the PI of the community network for volcanic eruption response converse RCN and I'd like to start out by thanking the course subduction zone initiative committee for the opportunity to report today about this RCN and also thanks to Harold and Torsten for their nice and eloquent introduction to the kind of topic of RCNs and how how these came about so like the other RCNs the RCNs the converse RCN grew out of the reports that were already mentioned and it became apparent that perhaps something is needed to kind of organize the community around figuring out why volcanoes erupt and what can be gained scientifically from monitoring or documenting eruptions and precursors to these eruptions next slide here's an old plot that shows you a very few data points on seismicity and the flux of sulfur dioxide from a volcano and you can see that over the span of a few months we have a few data points whenever people were able to go out to make measurements and we have some data from seismometers that show us the length or the duration of long period events without much information on their source or their location but even with that very limited data set we can see that there are some patterns emerging prior to volcanic eruptions so for example the duration of long period events goes up prior to an eruption and the gas flux perhaps goes down but we don't really have very much data if you go to the next slide we can see now that for example the carbon to sulfur ratios on the left side this is from the Arika volcano 2004, 2008, 2012 we have one data point every about four years when people went to collect some samples and make some measurements but if you look on the right diagram bottom right that represents data that was collected in real time with an instrument on top of the volcano that continuously measures the carbon to sulfur ratio so you see that the data density really goes up and we see a pattern of increased carbon to sulfur ratio prior to this eruption at V Arika in fact the ratios go down right before the eruption and then the data stop because the instrument's destroyed so if you go to the next slide so we realize now this is actually from Karl 2013 we can combine gas measurements carbon sulfur ratios for example with seismicity that allow us to locate the earthquakes and we can see that there are clear precursors to the eruptions here at Etna so we see carbon sulfur ratios kind of tagging along with between five and 10 there's an earthquake location you know nine to 16 kilometers steps in 2005 then there's another earthquake that was located at six to seven kilometers 10 to 15 kilometers and then the carbon sulfur ratio really shoots up about you know a few days prior to the eruption of Etna in 2006 when you combine that with crystal clocks data that tell us about magma injection we can get information from that to tell us something about magma injection early or yeah early in 2005 and then again other magma eject injection events in mid 2006 so these slides just illustrate the power of combining different disciplines different techniques during the run-up of eruptions and that power of the combination of these techniques allows us then to understand better understand what is happening in the next slide we see that this is the eruption report so erupt report and that report really had the suggestion in it to make such community building efforts like the converse RCN to strengthen volcano science and the ability to collect critical data at an eruption in order to maximize scientific return and overcome observational bias so the converse RCN is an effort to do exactly that to build the community and collect data from a US volcano prior to the eruption and during the eruption to maximize scientific return next slide and we have a an RCN funded for two years and it's about $300,000 most of it actually in fact almost all of it goes for workshops there is no overhead in terms of institutional overhead because it's all participant support costs there is no salaries in here it's all to bring people together for these workshops and we have a steering committee and and then we have these disciplinary leaders importantly in the steering committee we have rotating USGS and NASA folks as representatives in the disciplinary leaders we're covering pretty much all the disciplines that are involved in volcano science so we have seismology geodesy infrasound petrology, geochemistry experiments gas remote sensing water modeling eruption dynamics and TEFRA UAVs sample curation and public communication next slide so what has happened so far is we really started this RCN with a steering committee kickoff meeting in Albuquerque late November 2018 we had a panel session at the AGU 2019 with representatives from USGS, NASA and academia where we talked about some of the issues that have arisen and will likely arise in the future when we try to respond to volcanic eruptions and we had a steering committee teleconference in February 21 February 2019 so the results from these teleconference from this teleconference and from this kickoff meetings on the website next slide please and the outcomes of these initial discussions were the following that are listed here strengths and volcano science and the ability to collect critical data and samples at a US eruption to maximize scientific return establish a framework of protocols and infrastructure for data collection integration this could also include some sort of a facility and formulate a collaboration strategy between the USGS other national institutions or federal institutions and academic institutions next slide and so here's just the timeline of what we have done and where we're going so far so the RCN was funded late 18 next slide and so we had these this kickoff workshop in Albuquerque we had the AGU panel session the steering committee teleconference next slide and we have members of the community that had submitted some proposals to NSF so far they were not successful next slide and in the volcano cider there was a converse modeling workshop now we're in the state of the disciplinary workshops that are happening currently more about that in a minute in March next year March 2020 we'll have a cross disciplinary workshop in Albuquerque next slide and in probably September or so we'll do a volcano response dry run next slide and during this entire time the report writing is continuing or is happening and we hope to publish the report before or shortly after the funding ends next slide so the disciplinary disciplinary workshops are ongoing we had a workshop with strong USGS participation about gas and remote sensing of gas at CVO in September Tefra eruption dynamics workshop is happening this fall at AGU a joint Unafco geodesy workshop happened already in Portland infrasound workshop just happened in Fairbanks and Petrology or Chemistry Experimental Petrology sample workshop happened at Pre-GSA in September seismology workshop is planned for February 2020 and UAS workshops or UAS essentially is participating in these different workshops contributing experience about drones and applications so all these workshops are happening and they're about two or three days and involve 10 to 15 maybe sometimes up to 20 people next slide and we have a Google life report and the leaders from these workshops are encouraged to fill in the results of the discussions and address these questions so the first one what are one or two science questions that can only be answered with transient data observed during run up to and during eruptions how do we identify and manage precursory signals and then during later stages of the report writing we will look at what are the resources resource requirements for data and sample sharing and dissemination what are mechanisms for funding rapid response to U.S. volcanoes and what are the processes for obtaining permits on the ground next slide we are also thinking about facilities and we are in communication with the natural hazards reconnaissance facility that is run out of the engineering program at NSF this facility actually has some equipment that could be useful for volcano response such as drones some cameras some seismic stations and I've been in touch with Laurie Peek who is the PI and Joseph Wortman who is the director of the natural hazards reconnaissance facility Rapid and we will meet up at AGU to discuss this further so this facility has some equipment however we feel that likely we need additional or different equipment and that there should be some sort of facility for the converse program or converse RCN next slide so out of these workshops and initial conversations we think that converse is going to be really successful in building this scientific community and bridging the correspondence and collaboration between the academic volcano community and the USGS and NASA folks who are also obviously interested in volcano processes and we hope that the converse idea could remain an entity beyond the two-year NSF grant thank you thank you so much Tobias so let's start with again the committee members are there any one or two questions just for Tobias Doug do you have a question yeah just have a quick one and it's maybe unusual when I'm just curious how do you measure success in an initiative like this as compared to what a baseline might be yeah so measuring success in such an initiative would entail being able to instrument volcanoes and get people organized prior to and during an eruption to collect these critical data so for example some volcanoes erupted recently Bokoslav is an example and many of the precursory signals were not effectively recorded and so a lot of information was lost about the run-up to this eruption as well as the actual eruption so success would be to collect data prior to the eruption and get some time-series information of the type that I showed as examples leading up to the eruption in real time and then hopefully understanding what is happening and being able to better forecast what is happening at when the eruption is happening and what size it'll be and so forth okay and you would be able to represent that this would not have occurred or this kind of effort would not occur without this kind of facilitation yeah I think so because for example we have new instrumentation now that we could deploy on these volcanoes that would allow us to collect much better data for example NASA has very sophisticated remote sensing techniques that you know they could possibly use at certain volcanoes and then collect data that the USGS by themselves would not be able to collect so the academic community can significantly contribute to the volcano monitoring efforts that are currently ongoing okay thank you great are there any other committee members who might have a question for Divya I'd like to ask a quick question right so I might have missed this because I have to step out for a second but how do you envision international sort of partnerships here with data because I mean certainly you know there are lots of observed volcano observatories out there from you know certainly all along Central America South America Europe and you know how do you envision this might contribute to the to the effort of collecting data analyzing what's a precursor to interruption and so on yeah that's a good question and this RCN currently is focused on US volcanoes only however we envision that our data that we collect will go into databases that will inform the international volcano community so if we collect data let's say at Augustine and you know we find some interesting precursory activity we would envision to put that data into a database for example it could be the Volvo database where folks can then access that data and use our precursory information or precursory data to better understand data that they are collecting at their volcanoes wherever in Central America South America you know Asia to better understand volcanic processes in general because what we're observing are precursory signals that are not just typical for one volcano but they have occurred at many different volcanoes does that answer your question partly but we can revisit later okay maybe are there any more questions for Tobias and if not wait one two three if not what I think I'll do is open up the questions for any of our speakers and I think Bruce has the first question for that yeah thank you Deb and thank you to the three of you it's been really very very enlightening I have really what are two comments they may lead into questions and there is a proviso that both of them arise I think from the context of course we're at a very early stage in this process but the first comment is simply at the moment the interface between parts major building blocks so is the 4D isn't necessarily that that clear and what I mean is for example you could you can hear 20 minutes talking about the modeling collaboratory without any reference to the vital need for new observational information and you can also hear about the new observational initiatives but the link to the modeling isn't isn't there and I think in any environment like this a webinar it's pretty critical to to demonstrate that these these are not absolutely separate initiatives that don't rely on each other but there is a dependency there uh I'm sorry so I'm not quite sure I understand the comment speaking for the modeling collaboratory the integration of data from the different disciplines is a very important part here and the alignment of the computational geoscience with the observational one is is crucial for this and it's actually a very it's a new model that's being suggested here where you can think of subduction zones as just one exemplar of a new way of dealing with business right if you think about sage gage and Pascal deployments for example how many times have we done deployments and then we couldn't answer or address the hypotheses that we meant to answer right and what is suggested here is to complement the data collection efforts with the modeling effort that's very much geared toward optimal experimental design right what's suggested here is an incorporation of data streams in terms of a real-time model of the state of a volcano or megathrust so I'm not quite sure where the data didn't quite make it in there and in terms of the relationships to sc 40 sc 40 I think as a wider umbrella thing will be crucial for us for the for the United States to help advance subduction zone signs perhaps with an investment in our local subduction zones and it will be outstanding if we can incorporate those data streams in the future but the modeling collaboratory can start with this effort right now right with the existing global data so I wonder if you could clarify what what what your concern is here yeah no exactly I mean and you've come right to it with those last remarks there's nothing in what you presented to suggest that there's any need for new data at this point in time you could argue simply that there could be terrific development in modeling approaches and in code without necessarily any new data and so a cynic might say right let's take the modeling collaboratory and treat that as simply an entity a standalone entity why does it need to be in a why does it need an umbrella and I'm not arguing that's not I'm doing that as devil's advocate that's not my own perspective but I'm saying you guys if you want to preserve this large entity have to make it very very clear that you need new data I think it's much easier for them to say they need new models if I could chime in yeah I we take that on board very much but I think that that's exactly what is going on and maybe we need to make sure we convey it even even more forcefully one of them I was attended both of the MCS workshops and one of the most remarkable things is the degree to which the modelers spend almost all their time talking about the need for observations and demonstrating to us the ways in which well constructed models and modeling schemes drive clarity in what observations are most useful to make and pinpointing exactly what kinds of new observations are needed from the the infrastructure and field experiments it's very clear that we can't blanket the entire world with geophysical and geochemical instruments and geological studies so we need to pick and choose and target very well and we've we've heard quite a bit actually already out of the MCS about exactly how models can drive targeting of those experiments to get the most bang for the buck as if you will so I think that and Torsten included some of the data sets already that really are driving the modeling community and as you said Bruce it's absolutely true that the models can drive not only the field experiments but also the laboratory experiments yeah and I'm really only saying how keep that word new in front every time somebody says data because otherwise you can say there are some wonderful data it sets out there you know which which are underutilized at the at the present time and that's not my perspective or I'm on your side in this my other comment however is the timing of our report is not ideal in terms of the timing of the development of SC4D you know we nice of our report was two or three years hence in many ways it isn't and one of the things that comes out at the moment and this is inevitable is the open-endedness of SC4D at the moment and by that I mean when you see a map of seduction you show a map of seduction zones around the world what strikes people is the diversity in terms of the level of existing infrastructure the existing state of knowledge the ease of difficulty of assembling decadal scale new data sets so it's a hard sell at this stage it'll be much better in one or two years time it's harder for a group like ours to totally endorse everything that's going on board at this very very early stage and that's just a comment it's not a criticism yeah I certainly appreciate that and we wish the same in a way I mean essentially what I laid out was that the next year I think the next 12 months or so is the critical time for kind of tuning in geographic focus and kind of intellectual question focus that can be tractable and addressed in the kind of program that we're designing those global maps are certainly not maps of what we expect to do in the sense that all of those things are very unlikely to be on the table for say major new infrastructure for the intellectual questions though I think that we can recognize that there's a lot of the global authority being worked on well there are key areas where key questions exist and maybe a more a more concise answer a more concrete answer to Carolina's earlier question about how this differs from geoprisms in Earthscope too could be stated something like that you know Earthscope what Earthscope did very very well was this big data collection but maybe I think it would be fair to say and probably people wouldn't disagree that it was basically a geophysical experiment and didn't incorporate a very deeply integrated you know sort of geochemical geological modeling and other things into the Earthscope program itself geoprisms the other side of the coin had virtually no you know sort of sustained long-term data collection of its own although it ran a number of campaign field programs but didn't have an infrastructure component yet what it did do extremely well or has done extremely well is integrating these various disciplinary areas into really truly multidisciplinary science for both subduction zone and rifted margin problems and so taking those two key aspects and making sure a new program actually draws on the best of both is critical and goes I think directly to your question about the integration yeah I you know yes thank you that's very useful and if I actually could have a concrete example of how geoprisms did something that has changed how we think about either a subduction zone or a rifted margin I think it would be useful for me to understand something concrete but I see Bruce's point and I think you've both answered this that I think you know as it's perhaps it's just a matter of semantics and words you know the you could see these as completely separate entities even though you know the ideas as they would work together right there's been the modeling collaboratory and the observational side let's put it that way you could see them as separate entities on their own you know funder on their own that someone enterprising might integrate at some point right what it would be nice to see I think it's that one necessitates the other you know that you cannot do one without the other I think that's what I would like to see I mean I totally see that the need but I'm just saying it would be nice to convey that one cannot exist without the other in an effective way yeah and I I mean you're I I guess I agree with everything that was commented on here and it's it's it's interesting and we're dealing with with a number of different things right as you know Harold said Earth's scope was was great but had several shortcomings for one as you said it was disciplinary quite focused and for me as a geodonemesis I I you know I felt that you know we really in the end didn't yet bring it all together to understand continental dynamics that much better than before right we have all this stuff that we know now sits under the plate but what does it do we don't really still don't know and so when I first attended what was then an SEO meeting thinking about the follow-up right I raised my hand and I was like please this time let's integrate something from the bottom up that deals with process right and so I think SC40 as a as an initiative would very much benefit from something like the modeling collaboratory to have an interaction with the people who are doing the observations right and I think that's where that's where the added value comes in the modelers can tell the geodesists where to measure things to get a better locking estimate right and on the other hand working with the observation lists tells the modelers which sort of codes to build to be able to map the data and also working with the folks who are interested in hazard drives the modelers forward to do better in terms of uncertainty quantification robustness of models and so on however it is true that the modeling collaboratory itself can sort of stand on its own and is a little bit more of a philosophy how we should maybe push science forward by integrating modeling with data collection and subduction zones is like I mentioned one example where this could really be beneficial right and so I think if you're like want to narrow it down to subduction zone signs which in itself is quite broad but if you want to narrow it down I think for us as you know scientists working in the U.S. engaging now with a modeling collaboratory would be great because we can already start the conversation within national community that's further ahead in terms of these observatories and then SC40 once it comes online right we can sort of go on that and be more of a partner internationally when it comes to the observations right so there's different ways to to think about it and it's not really sort of my place to comment on this further but it's true that the modeling collaboratory has a lot of you know independent value but there's really benefits of doing it with SC40 and to have that as an example to to do a big scale infrastructure investment for the first time in a way that's that is taking the modeling seriously to actually deliver on saying something about the scientific hypotheses we want to test and to deliver something in terms of hazard to cattle scale physics-based hazard assessment for volcanoes and earthquakes right and so we've we've tried bits and pieces of that before right and and there are lessons not just from earth script but also from you know skek for example and and I think it's time to do something a little bit new here and we're all seeing the challenges that are involved in that right how how do we do that what's the timeline you know which agencies do we engage how do we how do we do how do we do that but I think there is there's a huge opportunity here and there's there's a lot of excitement and I think we have a we have a mandate to do something about figuring out these precursors just picking one example just very briefly to cap that off you know I think the value of in keeping the and the modeling Collaboratory inside a framework like SC40 is we have all this talent and firepower and potential for leveraging computational power that could happily go off and work on very important and interesting subduction zone problems along with lots of other things but if we keep it focused on the same questions that the observational side is looking at then we just we concentrate all of that into the into the same problems I think we just have a much more powerful community great thank you I'm gonna jump in for one second because we have at least two other questions from committee members Tim I think you're next and then we'll take Donna Hey this is a question for I guess any of the three of you could answer but maybe Harold is best qualified to answer has to do with question six that was in Deb Glickson's list of pre questions so most subduction zones are outside the U.S. and we can make a big contribution to our scientists by getting data that's already be being collected in those places by other national agencies more widely distributed so I'd like to get I'd like to see a little bit more effort on getting these international partners involved from the get go I am concerned that right now they're not intermittently involved that they're kind of an after I've heard I've heard it talked about as an afterthought and I think it would be more fruitful to get them involved in the planning process early on so that they feel like they are real partners we can't instrument the whole world so maybe we should start by focusing on places where international partners are willing to do contribute and do some of the heavy lifting so that's a point any of the three of you could address perhaps yeah thanks Tim I mean that's a that's a comment that's really important and in a topic I should say that's really important to us and it's true that there's been a certain amount of trying to to sort of get a good grasp on what the scope of what's being attempted here is important in order to really drive what the key questions are with some of those international partners but having said all of that then it's really clear that as you look around the world first of all that that this is not intended on the on the either the the sort of megathrust side or actually the ARC volcano side beyond the converse program to be us to be us only or us focused or be just Cascadia and Aleutians or something like that that's certainly not what's envisioned by the steering committee for the umbrella what is envisioned is leveraging exactly that off of you know we can sort of divide the world of international partners up into places that are already doing a lot with very interesting subduction zone systems where we just want to directly partner with them and kind of leverage off data that's being collected ones that are at various stages of building up infrastructure or observations in really interesting places from the point of view of the key problems that are being addressed just a concrete example that might be Chile or a couple of the different Central American nations you know from Mexico through to Costa Rica and so on where there's the real possibility of a direct partnership that will involve U.S. infrastructure partnering with things that are already happening in those places and then they're international partners from the form of other sort of what's called a deep resources countries that are heavily engaged in the research without being sort of field sites of their own like let's say Germany for example and you know country like Japan fits into more than one of those categories simultaneously so you know we are now organizing a meeting with a bunch of key players from about eight or nine different different countries Mexico Chile Japan New Zealand well the Singapore group and Australian groups that relate to Indonesia up through Myanmar field sites and so on there are no mothers Germany UK I didn't mention but there's also this U.S. NERC partnership U.S. UK partnership proposal to do a Mediterranean Subduction Zone Observatory or MED SZO in the Hellenic Arc that is already ongoing and has a whole bunch of U.S. collaborators along with a big group from from UK several different EU nations including Greece and Italy but also Germany and France and so so there are a number of steps going on that haven't really sort of shown up in the public view of our website or activities yet but that I think will start emerging over just the next few months and we're inviting some of those key international partners to this January kickoff meeting of the working groups so there's some structural reasons why we wanted to get our domestic team all under the same page and starting to pull in the same direction but it's pretty clear that if we want to do a solid and effective science program it's going to involve a number of locations around the world and we'll have to involve that kind of data leveraging Yeah, I'll just chime in quickly here I completely agree this has to be an international effort and I see the modeling collaborator is serving the the global community really in a way that is engaging with similar efforts and that is hopefully providing value real value for the global community so our RCN has travel support for international participants build in anticipating this and we had for example folks from UNAM give keynotes that our megathrust workshop was really quite inspiring we had great participation from Europe and Japan and as Hal mentioned you know there's different ways to to engage I spend personally a lot of time talking with Takana Hori at Gemstech who's part of a very similar Japanese modeling effort they are trying to move their computer hardware beyond excess scale what have you and he's the leader on that part of this Japanese project to integrate data streams from their seafloor observatories into physical models very much what we're trying to do and I met with folks from NID running the network and asking me so what should we do in terms of complimenting our data streams with physical models the Japanese are facing this challenge and they have already made huge investments and we're hoping to really benefit from each other so far you know we've played the role of being each other's cheerleaders I'm going to attend an earthquake engineering conference in Kobe next year just to to be a cheerleader and help the Japanese and I hope that these investments in terms of personal ties will pay off I mentioned the European cheese initiatives where the Europeans are actually sort of again a little bit ahead of high-performance computing in the solid earth communities in some extent when it comes to training and supporting the workflows you know we've signed a memorandum understanding with those folks there's EPOS the European Playbonder Observatory System that we're hoping to tie and so so very much right we it should be global and as you all know there are challenges right the communities operate and different have different cultures of operating access to data is you know is treated differently the degree of openness is unclear sometimes and so there are challenges right and we are aware of those and their ways of engaging for example I was part of a skek earthquake research institute Tokyo NID summer school program a while back very much with the idea of exporting some of the culture to data sharing to different communities and get engaging and building long-term personal context so that the students have partaken our workshops in a couple of years down the line will be faculty and then grow into leadership positions so yeah I mean this is I think we have important choices here to make as to what we do with our U.S. science and how we maintain leadership but of course the science is global and we need to understand the difference of direction zones and even if if our investment is saying Cascadia or Alaska we need to bring in these other communities we need to bring in these other data and integrate them for sure Tobias did you want to respond? Yeah just I agree with everything that's been said that's a really good response to that question from the volcano side of course the RCN is a U.S. volcano focused and we have of course a lot of good collaborations with the international community of volcanologists and we could immediately implement our processes internationally if you know the observatories and the other collaborators would be okay with that but that's kind of the point of this RCN we want to put in place protocols and ways to monitor volcanoes and collect key data and then you know we could implement that anywhere around the world in collaboration with local observatories and local scientists Can I add to that so that so what Tobias just said in fact is the larger picture part of the the Magnetic Drivers of Eruption subgroup of the Umbrella RCN so the Converse RCN that Tobias told you about has a specific mandate for the U.S. volcano sort of response but the larger SC4D has a much broader portfolio and so that group that subgroup is being led up by Diana Roman and Terry Plank and they are talking about this you know common set of instrumentation on a number of volcanoes to be selected and they're talking a lot about Central American volcanoes just for example as well as South American as well as you know Alaska Aleutians and other places so that international collaboration is also moving well along within the volcano side of SC4D. Great, thank you. Donna finally let's get you unmuted. Am I unmuted? Okay. Yep, now you are. This is unanswerable but I wanted to just explore more about the overarching questions of SC4D. You said it won't be all things subduction so there's a hazard focus but it's on the fundamental processes that underlie hazards which is kind of everything right? I mean you showed a beautiful picture of a Lucius eclogite if you look at the presentations in your upcoming AGU session you know there's not real obvious connection to hazards sometimes so how would you grapple with this that that you could pretty much say oh it's on a subduction zone so it's on hazards just what kind of discussion or how will you how will you decide what is included? Yeah I think though like you said that's a great question that that gets to the heart of the fundamental philosophical question of how to do all of this. I think one part of it is just make you know inculcating a culture within the SC4D community of recognizing that despite the fact that we can define things pretty broadly we still need to be cognizant of and draw those links back to what the what the process that one is studying is what it has to do with the fundamental hazard questions. I mean an example of that is both I and Torsten showed essentially you know aspects of the the mantle dynamics and let's say viscous or viscoelastic plastic response to either precursory or post earthquake events and looked at you know how that much larger scale response drives the loading the stress loading of the the part of the plate boundary that actually is is locked and going to participate in rupture. That doesn't mean that you know I think that one that SC4D will prioritize all types of the numerical simulation of of you know the flow at the largest scale of subduction systems and slabs and so on. It's going to have to parse out those questions and really decide which things are important to target and focus on. So that's one example where the aspects of that kind of geodynamic scale work that would be prioritized will still have to link back. We could make similar kind of statements about processes driving driving volcanic systems to use one of the old phrases from the margins and geoprism stays the subduction factory was a thing that generated lots of really interesting questions and really interesting science from the point of view of everything that goes into the subduction zone driving what comes out the volcanoes. Not all aspects of that to me fit into the SC4D mandate parts of it do and parts of it would be seen as broader important but different scientific questions surrounding the subduction system. The earthquake community within SC4D I think is really focused on emphasizing especially from the point of view of the experiments that can be done and the models that can be run things that actually can tackle tractable questions about you know the drivers of rupture for example and how loading leads up to to a rupture event and then what happens dynamically in response to a set of existing sort of quasi static conditions at the beginning of an earthquake. So there's still lots of science to do that's very broad around that zone but it's got to be tied in. That's that's I think that's my best answer to that and I'd be happy to hear the others chime in on it. I guess I I can comment on that and you know I I think the the the goal of providing the physical models for Canaan earthquake hazard assessment is an exciting one is one that can help us hone in on our science but just taking the megathrust workshop we just had as an example one of the most exciting talks was by by John Platt who told us about micro structure and geological constraints on the megathrust and I think it's it's a true statement that if you're trying to figure out how a certain megathrust operates you you need to understand processes that go on on one million year timescale you know in terms of how does the accretionary wedge respond to the setting what is the interaction with long-term mass transport of their writing played landscape evolution but also what what are the tectonics doing how is the back arc spreading how are different trenches advancing or retreating all of those on timescales of a million year will affect the general state the megathrust operates and then from there you can go down to shorter timescales now the modeling collaborator for seduction will probably not be concerned with questions like did plate tectonics operate for a billion years ago but the tools we will hopefully help and contribute to can be used for research on whatever sort of genomic problem you want to to apply to in this planet or other planets so I think it helps helps to focus but it's really quite quite broad and so the you know we're trying to walk that balance and um you know asking just again as an example of the megathrust how is the megathrust loaded right that gives you a way to focus and I think that means you've got to understand trench retreat and you you know and you have to understand the shorter timescales as well yeah but it's go ahead first it is not just geophysics right I like to emphasize this and geology petroleum mechanics it's trying to bring it all in there and trying to you know give not just say we need more data right we always need more data but to give more specific answers like maybe we don't need to understand this particular aspect of you know rate dependent friction maybe we need to understand a different component maybe we don't need better seismic data everywhere but maybe we need highly detailed you know surveys in a certain location and we're hoping to understand sort of how the different scales talk to each other right and sort of if you build an amic rupture mall what is the information in an act of suicide mix that you can use about roughness and structure and and and what can we do and what can we can we not do you know okay great do any of the committee members any other committee members have any other questions not hearing any are there any other questions from anyone who's on the line and you'll just have to unmute yourself and introduce yourself please so i'm going to do one last call for questions oh donna do you have another question i do i i just wanted to ask more about cyber infrastructure and whether it would be something at an sc4d level coordinating the different other aspects or interacting with i don't know iris yunafco or how would you deal with it i know it's a huge issue let's see maybe all of us could tackle parts that torsten do you want to talk first about for big-scale computing yeah so um from the modeling collaborator perspective i i think there are many opportunities to engage with different community initiatives and so i for example if you take sage gage right um i think something like a modeling collaboratory approach would very much be useful for sage gage operations in general right the next time you want to collect um you know data somewhere you know you can you can figure out what by running forward models and by producing synthetic data as to what you should where you should be observing what to test whatever you want to test right in general and i think we can this we can do this better now because the genomic codes and capabilities have advanced sufficiently so we i think personally should get really serious about sort of a third leg here in terms of computational geoscience helping us out in terms of guiding observational efforts for any sort of sage gage activities and the infrastructure needs that uh you know are similar right we are already doing that people are already doing this with global wave propagation computations to complement observations so data storage and things like that there's commonality and really a lot of value in engaging now there's other efforts supported by nsf for example earth cube right earth cube has invested a lot of money in terms of developing infrastructure that we hope to engage with in terms of the modeling collaborative when it comes to reproducibility for example right so nowadays sort of getting the workflows out there completely for complete transparency that's a that's a challenge for individual pis and earth cube can help with that now in terms of the hardware access the modeling collaborative you know would should make it easier for people to access how performance computing to say take our lego pieces put them together load in a structural model for cascadia and then run your particular test case that you're interested in on the biggest computers we are hoping to to facilitate that and we can't do that alone but the modeling collaborative will have to engage with the super computing center that has expertise in doing that such as just tack not to just because it's sitting here on top of tack but such as tack to help with that right and there's precedent for that and I think there's ways of engaging with whatever the future cg will do right where you know cg has done a lot of great stuff and is a little bit in a sort of in a reorganization stage itself right and so there's there's ways how you can see different entities within solid earth to to usefully work together and so I guess my answer is you know earthcube is really important for example for the workflows sage gage in terms of the ways of having data and computations at the same place and I think the modeling collaborator can help out sage gage and then cg you know it really depends on what the future cg will do but there is a need to interface more where the super computing centers right because the super computing centers have expertise there in terms of having access to data storage and things like that I can just add that from the point of view of the kind of the broad sc40 picture well for you know some things are easier and some are harder so for the the kinds of systems that are now getting well established and aren't perfect but getting better and better like everything that iris and you navco does and so on there's not much appetite for reinventing reinventing wheels that's much better for partnering with with whatever entity comes out of the sage gauge iris you navco world in the next couple of years and for using the systems that are there and certainly a total commitment to open data access and the idea that these are community experiments now the harder notes to crack of course for the geological data the laboratory data you know the the kind of interoperability of models and all these other things so that's the discussion that still is ongoing is at an early stage I think it's fair to say but some of the exciting things coming out of the earth cube world in terms of using data mining techniques to and other things to get at how to bring geological data on to and geochemical data on to more into more archivable database systems are what's under discussion we don't have a lot of concrete things to say about that side of data for sd4d specifically yet just because I think of the stage that we're at it's in the discussion I guess I had a few more things come into came to mind one is of course big data and machine learning right where machine learning is just sort of another tool that will possibly be usefully applied here so we will complement physically deterministic mechanisms with machine learning whenever useful right and and we have another interesting workshop on that at c osg but in terms of what we're all going to do with the with the cyber infrastructure I think it's we've learned a lot of lessons from the past decades of the community trying to do that and and for the modeling collaboratory it's really science first and it's the science that has to drive the code development and we want to make sure that we have very close integration we want to work with computer scientists and engineers but the science has to be the driver and we want to make sure we're not in a situation where we're building products for you know and we don't know our customers we want to have the science be the driver and another thing that we've we've learned is from past efforts is that if we're trying to couple things right we're always we've been trying to couple things for 20 years but what was really clear for example at the surface process lithospheric deformation workshop that NSF sponsored where CSDMS another community organization which we haven't talked about and CIG co-sponsored is that if you want to do coupling right you need to do it at a very low level you need to take a code take this other code take of all the bits and pieces and then integrate them at the low level and that is hard but that is the only way to really do it if you want to respect you know the the proper physics and the time scales and heroin and and so we need to to do a lot of sort of nitty gritty stuff we can't just put it all together bits and pieces but we want to do it in a way that's driven by the science and that makes sense so very much engagement with these initiatives but keep the signs be the driver and have the scientists guide and do the code development that is I think really important yeah I just want to add to that and of course from the volcano science side data is you know very very important and ideally we want that data to feed into the models to help us really forecast what's happening at these eruptions and for deformation and seismic data we have yunafgo and iris but for example for gas data or for images from eruptions for tephra size distribution dynamic information from eruptions that are ongoing we currently don't have any place to put those data and get them accessible for people in a modeling collaborative to really go in there and use that data and you know try their models with that data that is collected in real time so we're collecting these high frequency data many different types of data but we don't have places for these data to go and those discussions will be held at the workshop in in march that will happen here to really figure out what what to do with these things thank you all so much for the answers to those questions I think we might have time for one more question so if anybody would like to ask one please unmute and jump in then instead what I'd like to do is on behalf of the committee I'd like to sincerely thank Harold, Torsten and Tobias for taking the time to provide this information about the RCNs to the course committee it's been incredibly helpful and we really do appreciate the time you've taken for everyone my pleasure oh thank you Torsten for everyone on the line there will be a video of this webinar available probably in the next couple of weeks and it will be available I believe on our course website if you have any particular questions please feel free to email me and that's dglickson at nas.edu dgliscson at nas.edu and I wanted to thank you all and that's it have a great Monday