 Hi, everyone. Welcome. I guess I have to put this on something. Oh, this is the biggest crowd that's attended a systems meeting. So it's a pleasure to have you all here. It's a nice topic. So I think that's probably why that we've got a larger crowd. It is an exciting time for systems, as we'll see in a moment when we look at the plans for systems 3.0. But I'd like to take the first bit of this and introduce you to some of the things that we've done in this last year. And in the years before, sort of a look back. And Greg will come up as the incoming director and tell you about systems 3.0. So for those of you who have not been here before, can I have a show of hands who this is their first meeting? OK, good. So this slide's for you. So I have dozens of different slides that talk about the various things that underpin our science. But I chose this one because it provides a pretty overview, but it also engages in some of the systems 1.0 and 2.0 goals. And as Michael just alluded to, we're moving into more coupling into the human domain. But I think everyone should understand at least that these are the goals that form systems 1.0 and 2.0, such as talking about the different transport processes and their interactions with their environment. And of these processes, which of them are self-organizing and what are the pattern formation? It talks about material fluxes and surface evolution and deposits. So there's a geological side that has driven some of our activities. And then there's this really big coupling that has always been with us, but we're now taking it to the new level, which is coupling between the physical, the ecological, and the human processes. And even the outside domains that influence all of that, which are the Earth's interior and Earth's atmospheric dynamics. We covered last year was climate. So that was to try to get a handle on those kind of interactions. For many of you, you obviously haven't seen this, but this is how systems 2.0 has been functioning. We have a number of things that we need to get done and we need to maintain, such as making sure our community is active, engaged, and all of these other activities. So let me just quickly walk you through a few and point out what I consider the salient point for each. So this is our governance structure. We're basically a community of communities and we have working groups, we have focused research groups. And they're quite large in size in hundreds of people who are a part of each of them. But then there are these individual projects and that's what I want to concentrate on. So for every dollar that we get at the integration facility to buy NSF to sort of manage this whole thing and develop some of the middleware and software. NSF has handed out $13 for every dollar to individuals and teams for their affiliated systems projects. So that's a force multiplier of 13.1. And when you add the other agencies, systems has served a large community. Its impact is much larger than the initial grant cooperative agreement with CSDMS. So this is a slide that I've also not shown before because I never knew this. So this is the population by country. And we knew we had 69 countries. I knew what our growth rate was about 150 persons a year. And I've been showing where the people are choosing to populate themselves in these various working groups. But the country side of things is sort of new for me and I'd like to share it with you. So two thirds of our membership comes from the US. And this is both government agency, scientists, as well as academic people and those from industry. And then there are the countries and I don't think the countries that are next highest up are should be of any surprise. If I was surprised at any of this, I have some German colleagues in the audience is that there weren't a lot of Germans who choose to be members of CSDMS. So we have to change that somehow given their population. So I'd like to welcome four new co-chairs. The first one is Kim DeMutzert who's co-chairing the ecosystem dynamics focus research group. Where's Kim? There we go. And Myra Zellner, the human dimensions focus research group. Hi Myra and Mary Hill who can't be here right today but she's coming a little later on in another day or two shows. She will be here who's gonna co-chair hydrology the focus research group. And then Scott Peckham. Scott here. Scott's not here either. Who's co-chairing with Tom, the Cyber Numerics Working Group. And since this last year, we've had a number of models that have come into the system. And when I'm listening here in the domains is these are sort of the basic domains. Some of these are both terrestrial and coastal. Some of them are coastal and marine, et cetera. Some of them are hydrological and terrestrial. Those are the models that we have entered into our repository. And we have not quite as large, but we have a lot of heavy code tools that are associated with some of these models or can help other models. And then we have this compliant side of things. And the compliance side is for those models that have been refactored, altered, wrapped, whatever way you wanna word it so that they can be in a plug and play modeling world that we've been putting together. So I wanna thank all the new contributions. These are the models that came in this last year. And for those of you who have not seen sort of the world of modeling that we've been in and I don't wanna go through it all. Some of them are Lagrangian. Some of them are direct American simulations. Some of them are social or local or time marching, et cetera. So systems has always been a community of communities. And in this community, there is a wide variety of models and languages that they offer. For each model that's submitted to our repository, the integration facility does a few things on behalf of that submission. We can start and usually start for all of our models, we've done this citation indices for both model overview papers and model application publications. So you can get credit for writing models. This is something that I've been pushing academic community to recognize. So it's not just scientific papers but it's actually the models themselves that should be credited. Our model metadata with its digital object identifier for each stable model, we now provide this. And by doing this, this has really helped. And we've been working with the various publishers to make sure that when people publish a paper with a model in it, people can reproduce that model. You can reproduce it in such a way that we can get back to what is called normal science because in the past, without the code, you would have a model paper that could never be reproduced. So we thought that was a mistake and we've worked hard to try to sort that out. So obviously all the code that comes into repository is open source and our version control we use is given. Quick word on our soon to be retiring supercomputers. Janice has already retired but we brought in these supercomputers in three stages. First there was beach, that was the largest on our university for a short time and then there was Janice and then now there's Summit. And beach has, we've been keeping beach afloat much longer than it shelf life. So that'll be the next one that we replace. We replaced Janice with Summit. But if you look over this last 10 years of supercomputers, you'll notice that not only are the teraflops, computational speeds that we can attain increased over time, but the entire architecture has really changed. This is always a challenge for modelers because modelers like a stable hardware platform from which they can write code towards. So if the hardware keeps changing, we have to adopt and that's why we put on these clinics for helping people use HPCCs. And now you can see your interconnect bandwidth is so large in our new systems. It's not just clock speed. It's not just number of cores. In fact, you can see the number of cores have gone down yet the teraflops isn't coming. We put on a number of in our educational repositories. We work with the community to take what you have developed and what we developed internally and make sure that we can sort of spread the word, the domains, whether it be creating new student labs that you can download and run on your behalf and in your classes or short courses or lectures or whatever the method is. We have our own YouTube channel. So please, if you've not experienced our educational EKT side of our web domain, please go there, have a look around and use it. We put on a number of special issues. I think this one that we're attending right now deserves a special issue. So you may want to talk amongst yourselves and whether you think that the social surface dynamics community can get together and put a special issue out in an appropriate journey. In our EKT that Irina Overeem is trying to develop this quantitative toolbox, she wants to tag the concepts that come in that you may submit in terms of concepts, discipline, domains, level and model difficulty progression because you need to know that when you're giving us first year student course and they have a lab and you're using one of our things, you know that this is something that you can use or it may be more appropriate for graduate students. So please work with us on trying to make this system happen. It'll help all of us in our academic teaching pursuits. So one of the things that we've done is we've looked at how models have been written and how model coupling systems have sort of combined allowed one model to talk to another model. And there is this thing that's been around for a long time called IRF. IRF has initialized, run or advance, finalize. But a couple of the systems have that get values, set values. When you do that, you can have your models wrapped in such a way that they can enter the plug and play world. And I've shown this slide many times and there's this blue arrow that basically indicates that once you've done one, you've done another, you can then all all these models where there's an appropriateness of either subject matter or need to couple them. But I never tell you about the blue arrow. The blue arrow is actually quite a complicated arrow. And so I thought on my last time up here, I would tell you a little bit about the blue arrow. The blue arrow is quite complicated. It is really the guts of the middle aware that has been developed over the last 10 years by CSDMS. And if you bear with me for a moment, I wanna walk you through some of this. So when we get a model written in any of our open source languages, C, C++, Python, Java, Fortran, any of the Fortran. This model needs to be BMI. It needs to be put into this IRF get set system. So we've developed ways to help all of you do that. If you're a model developer, we've put a builder in place, we put templates in place, and we have a tester that allows us to test whether it really meets the standards. And then you end up with a BMI model. So once you have a BMI model in our system, then it needs to be babelized. So a systems babelizer allows a BMI model to speak Python. And it does that through some pretty heavy sophisticated code. It's babel, the language neutral compiler that spits it out in one model into any of the other languages. And in our case, we've chosen one language. So no matter what the languages are coming in, we spit out one language called Python. And we make that BMI model a Python component. And together with the BMI metadata, we can allow it to get into the Python modeling tool called PyMT. And then that will allow it to get into the web modeling tool. And if you're just a user, not a model developer, the only thing you may see is this last box called WMT Client, which is your browser code. But there's a lot of magic that goes on here. In fact, there are these are new concepts, as Eric will show you tomorrow, and when he provides his, with Greg, the keynote on BMI, you can actually take a model and in PyMT, change the model and its whole structure without actually touching any of the code. So you actually don't need a couple with another model, you just need to get it into this PyMT to make things happen, and you'll show you a little bit about that. So it's a whole new concept. And we have a bakery repository that allows you to get access to our recipes, pre-built binaries for all of your components, tools, and models. And of course, again, something we can do and of course, again, something we don't often talk about, you can't have one model talk to another model if they actually can't agree on what discharges or what units discharge, or even what discharge means and there is called. So we have standard names and these standard names have a method that's being picked up now by other groups. And it's really important that this becomes part of how the semantics and the whole one model can talk to another. We've also worked on, and there's a clinic on this, on getting Dakota, our DOE uncertainty tool, to get it to be much more useful. And so we've been PMIing various portions of Dakota so that they can be used in this plug-and-play world. And we've been pushing ahead on the benchmarking and our first off is supporting this island, this benchmarking tool that we've been developing for permafrost exercises and work. So some of the history back in 2002, 15 years ago, there was about 60 people that had came to a meeting here and you can see some of the young faces that were there then. I still had an old face back then. And the core concepts that we moved forward with was developed from this first meeting and elected from that meeting was a bunch of representatives to go to the agencies in DC, NSF, NOAA, USGS, O&R, all these various agencies. And they were to sell the contents of that first meeting. They sent the agency people sent us back to then write the implementation plan and a few years later we were funded. So if you look carefully up in a tree, you will see Brad Murray, who's gonna become our new chair, freshly elected. You also see Pat Weiberg and Brad as one of these leaders. This group has been giving their blessed wit and tears to systems for a very long time. So I would like to thank Pat for her great contributions and I would like to welcome Brad for the task of keeping this organization on track. So I'm not gonna walk you through this history but this is a history where you can see we developed the integration facilities, our first HPCC, the plans, the repositories, the various tools started coming out, new HPCCs came online and we walked our way through these various developments to get us to where we are today. And hopefully this will lead us to a submission in the near future for systems 3.0 and new visions and goals for various working groups and focus research groups. And since this is my last time, I have to say the following. This group that you're looking at here is some of the most talented, kind and generous people that I've had the pleasure to work with around the world and I owe them everything and so please join me in thanking them for their contributions to where we are today. And if you need to see any of them or you have any needs, please see one to support you. Okay, so welcome everybody, it's great to see you all. I wanna tell you a little bit about plans for the next phase of CSDMS, CSDMS 3.0. Before I launch into what that's about, let me make a couple of observations. One is that CSDMS is still growing. This is sort of remarkable to me. Jai mentioned that the membership has grown by about 150 people per year. That's shown in this chart here extending back to 2008 and if you had asked me 10 years ago to predict, I would have said, well, after a few years, this will level off. We've found our community, people are those who are interested have joined and yet it's a straight line pointing upward. Then it says something about the level of excitement and enthusiasm among you, the community. A second point has to do with the diversity of this community. So if you look at a distribution of membership in our various working groups and focus groups, right away you recognize that just from the titles of the groups that this is a remarkably diverse community in terms of its interests. So, for instance, among you all are people interested in long-term tectonics and geodynamics, the building of mountain ranges. There are people interested in ecosystems, fisheries, vegetation. There are people interested, of course, in human dynamics and human systems. Now, the National Science Foundation is very interested in the concept that they call convergence, which they define as the merging of ideas, approaches and technologies from widely diverse fields of knowledge to stimulate innovation and discovery. This is one of the sort of 10 big ideas that NSF has recently put out. And I think that this community actually has a credible claim to be achieving convergence in its diversity and its eagerness to work together and share ideas. Okay, so a year ago today at this meeting last year, we asked you all those who were in attendance then to get together in a group of a series of breakout group discussions and to help shape a vision for what systems 3.0 might look like. And in the years since then, the integration facility has sat down, collected the notes from those conversations, worked with the working and focus group chairs to try to craft those ideas into a vision of systems 3.0. And that has now been translated into a draft proposal that we hope will be submitted to NSF sometime in the coming months to support the integration facility for a systems 3.0. So I wanna tell you a little bit about what the outcome of those conversations has been. First of all, if we try to sort of synthesize a high level view of the science that comes out of this, we've tried to encapsulate it in the phrase you see here with the white arrow, model data synthesis in Earth surface science and applications. So it's a little bit of a mouthful, but those words are carefully chosen. First, model data synthesis. This reflects a view that came out of those discussions last year and that one of the key opportunities before us lies in bringing our computational models together with the rapidly expanding number, quality and resolution of data sets that pertain to the Earth's surface and its dynamics. I'll say more about that in a minute. The second part here, science and applications reflects the view that emerged from these groups that we can and should be doing more as a community to embrace applications, to bring systems models and systems expertise to bear in the applied world where the state of the art models are not always state of the art. So those two things together gives us our theme. Another way of casting this is that we're seeking to move toward a more predictive science of the Earth's surface where predictive here means not just forecasting the sort of sedimentary weather tomorrow, but it means prediction in the broadest sense of a good scientific theory is one that predicts the observations. So let me say a few words about model data synthesis and data opportunities. Last year at this meeting, there was a poster presentation that talked about a project called Arctic DEM. So this is one of a few examples I wanna share with you. Arctic DEM is an audacious project designed to create high resolution digital elevation models of the Arctic. High resolution meaning a few meters per grid cell, the Arctic meaning literally the entire Arctic landscape. That's what makes this an audacious project and they're doing it with satellite photogrammetry. Now this is an example of a dataset that means that we are now on the verge of having for this part of the world high resolution images of the landscape that can be repeated. They can be regenerated as long as the satellites keep flying. That means that we can do change detection. And this is a place where change is happening fast as we know. So whether your interest is in coastline retreat, or thawing permafrost, or shifting river networks, we're on the cusp of being able to measure those changes in a way that we never could before. Another example comes from LiDAR. So LiDAR has been with us for, I guess two decades or so now as a technology, but of course as you know, its applications are rapidly growing. Its coverage is expanding fast. Some states in the US have LiDAR coverages. A lot of countries now are acquiring LiDAR for things like coastlines and river quarters. So the coverage is increasing fast. And it too is repeatable. And it too raises interesting opportunities and change detection. So you're looking at images now from some of Marielle Perignon's work. If you look at the sort of bluish figure, that's showing an image, a difference image between two LiDAR data sets collected five years and one big flood apart from one another. This is an image showing in detail the patterns of sedimentation along a flood plain that resulted from a large flood. You know, when I was a grad student, we never had a prayer of having anything like this. To collect a data set like this would have been a major investment in time and money. So this too is an opportunity in change detection and it means increasingly as modelers, we're not gonna have the excuse to say, well, my model looks a little bit like a river, right? It's gonna have to be how much like this particular river and its dynamics over this period of time does your model capture? So it's gonna data like this are gonna push us to the limits of predictability possible. Another example from the coastal realm, this comes from Chris Sherwood. Chris showed some of these data last year. And you're looking at a digital elevation model of a stretch of the Cape Cod National Seashore that was obtained using drones and cameras on poles with structure from motion photogrammetry. I'm sure some of you are interested in using structure from motion now. It's a relatively new technology that makes high resolution surveying of topography fast and cheap, much faster and much cheaper than it used to be. Again, change detection becomes a real and affordable possibility. Again, pushing our models to say, not my model looks like a coastline, but my model resembles this change in this stretch of coastline in these ways and it fails in these other ways. The world of asymmetry is also part of this data revolution whether it's from submersible drones and improving quality of sonar technology. We're getting better and better images of the submarine world as well. Here's an image of the Baltimore Canyon trough at a resolution of five or 10 meters per grid cell. So all of this makes change detection and modeling of change detection and testing of models much more feasible and powerful than it was just five years ago. Now, not all of us deal with, many of us deal with processes that happen in scales of our lifetime and some of us extrapolate them into the geological record, right? The geological time scales. We think about mountain building, we think about sedimentary archives and things like that. But there too, we have a data revolution, not so much in our ability to live for 10 million years and measure things over that time. That would be cool. But in the form of experiments. So we have with us today at this meeting, representatives from the sediment experimentalist network. So welcome to you all. These are some of the experiments that the SEND group has produced. And what's interesting here is not so much that experiments are happening. Experiments in sedimentology and geomorphology and related fields have been undertaken for a long time. But what's new is that we can now actually measure their dynamics very precisely. So in sort of visualizing and teasing our visual imagination, it's also captured digitally. We can capture stratigraphy evolving, landscapes evolving in a controlled laboratory setting. And we have no excuse not to be testing our models in these controlled environments. So a wealth of data that we didn't have when systems was first launched and that's a huge opportunity. So in a way, the challenge moving forward then is to embrace this and other opportunities while also sustaining the good things that the integration facility has created and that are valued to the community. And doing that in uncertain budgetary times, right? So doing it if we're lucky on a flat budget. So that's the challenge that the integration facility is gonna be grappling with over these next months. Now, talking about some of the ideas that we have on track for systems 3.0 that again, came out of your vision last year, it's useful to think about the CSDMS integration facility activities in terms of three pillars, right? These are sort of three core areas that systems works on. And they reflect three things that many of us do in our science, scholarship, engineering, and so on. First of all, we collaborate with one another that's partly why we're here. We share resources, whether it's model codes or data sets or whatnot, and that's the community piece. Those of us who work with computational models, we do models, we run models, we write models, we debug models, and we need computers to run them on. That's the computing piece. And then finally, many of us are educators and all of us are learners. And that's true whether we're just starting out, we're just learning how to do basic programming, we're learning numerical computing and so on, or whether we're interested in learning a new specialized technique, whether that's the latest advances in computational fluid mechanics or it's agent-based modeling of human systems. We're all learners. That's the education piece. So I wanna talk about some ideas for systems 3.0 in each of those three pillars and I'll start in reverse order with education. I won't go through each of these items in detail. The key is that the light text is things that systems is doing now and will continue doing. The heavy text is sort of new ideas. One of the messages that came out loud and clear from last year's breakout group discussions was that clinics at these meetings are great, but there's only so much you can learn in two hours. So there is an appetite for more in-depth training opportunities and one idea that emerged from at least one breakout group was the concept of summer schools, some kind of fairly intensive training opportunity to bring people up to speed with best practices in scientific computing and applications. So we hope to launch that in systems 3.0. These will be probably something like 10 day summer courses. Let me say a few words about computing resources. There's a lot to say about computing more than we really have time for now, but let me highlight three things. First of all, the data challenges mean that we need better ways to bring models and data together, whether it's ingesting data set into a model or it's comparing model outputs with data to try and figure out how well a model did. That calls for tools to help that process along and so we hope to be able to create something like a basic model interface for data sets and to enhance GIS capability. So if you need to project a data set or project a model output, so you can compare them with one another. Let me say a word about beach. So Jay talked a little bit about super computing resources. What's unique about beach, of course, is that it is dedicated to this community. It's specifically for CSDMS users and it has a sort of a low barrier to entry. It's a starter system. As Jay mentioned, the beach facility is aging and it needs to be replaced. And so we hope to partner with research computing here at CU Boulder to create essentially a chunk of a modular system that they maintain that will probably have fewer cores, at least initially than beach does, but those cores will be a lot faster. Connections will be a lot faster and the memory will be a lot bigger. So we hope to see improved performance with that. And then finally, let me mention the Python modeling tool that Jay also mentioned. This now exists as a beta product and Eric is gonna show you a demo of that tomorrow morning. We hope to turn it into a full-featured product that you can use in doing your science and we'll talk about why it's, we'll show you some examples of what makes this such a cool and useful tool tomorrow morning. Finally, the community end of things. Again, lots of ideas came out of last year's discussions and we've tried to synthesize them. One thing that became clear is that many people in our community are interested in cold regions, whether it's glaciology or it's thawing permafrost or it's Arctic coastal retreat. And CSDMS has never sort of formally recognized that area of activity. So we hope to do so by forming a cryosphere focus group to complement our existing suite of focus groups. Let me also say a few words on the idea of science teams. As Jay mentioned, the working in focus groups are big. The smallest group has something like 70 odd members in it. Now the biggest has over 700, right? So that's a big job for the group chairs to manage. And yet we want to provide, we're hearing from many of you that you want to become more closely involved. You want to contribute in some fashion. So one of the ideas that came directly out of one of the breakout groups last year was look, what about the idea of forming teams within the groups? These are people who volunteer their time to engage in one or more projects on behalf of the broader community each project year. And those projects could be anything from wrapping a model to writing a group proposal to doing a multi-authored review paper and many other kinds of things. Proof of concept study and what have you. So that provides a venue to sort of recognize your contributions and provide a way for broader participation while honoring that formally and also providing a mechanism by which we the community can do projects that directly benefit both your science and the broader base of infrastructure that CSTMS supports. So stay tuned in the coming months for more information about the process for applying to join a group or nominating your colleagues to join. So that's a quick sneak preview of some of the things that we have on tap again. That's a direct outgrowth of the discussions that were held last year. If you have further thoughts, we would greatly welcome them. Just pull an integration facility person aside during this meeting or send us an email at systems at colorado.edu. So back to Jay. So I won't even put this on, maybe I should, I don't know. So here we are at this meeting. It's called the dynamic duo. And I think it's gonna be an exciting time, but let me tell you a little bit about how we got here in the first place. So, you know, we had a meeting here last year. It was the linking earth system dynamics and social system that both work. It was also co-sponsored that, I guess we're having some mic problems. Anyways, this small group forged some very interesting ways forward. And they came up with a list of ideas that I think are worth pursuing by our community, you know, from the natural hazard side of things to migration to how various, we deal with various low-frequency high energy events. And that white paper went around. And so I think all of you will have seen some copy of that as members. There was a follow-up meeting just months later in Kyoto that broadened that group, which was sort of less of an international group of folks. And this Kyoto meeting was part of the a future earth initiative and that future earth initiative under the auspice of the analysis integration modeling of the earth system aims project. And so that's that a group of people you see there in that met in Kyoto. And they really pushed, you know, what are the barriers for that the social scientists are facing when they are developing their models. Right now they're locked into many of them agent-based models or integrated assessment models. And they really have a hard time interacting with the kind of modeling that we all are doing. So that meeting really brought out some of the finer points. And then there was a follow-up meeting at Potsdam as Michael said, integrated modeling of the social environmental systems. This was a real blueprint forward way of trying to get that community to make advances. And now there's this meeting. So we will have an opportunity to meet later today. In fact, after a few more keynote speakers to really finally get sort of the service dynamics community with populated with social scientists who are gonna be leading this effort in reporting to try to get your thoughts into these developing white papers. Now the developing white papers, I don't know for many of you folks if you've all experienced writing them, they almost always lead to some agency turning on money. And so whether this money is at NSF or whether it's in the international community, I think it's really important that you get the opportunity to put your thoughts down. And no matter how young you are in your career, your thoughts matter. So please attend this breakout session and to get your thoughts recorded. You will notice that we have clinics every day. And if you forgot which ones you've signed up for, you can look at the back of your tag and see which ones you signed up for. We have limitations on some of our room sizes. So please try to attend the clinics you suggested you were going to attend. We have plenary keynote speakers. Every year that we have these plenary keynote speakers, they've been given advice on how to make their presentation so that together there is a coherent way forward. And you can find all of these presentations online. So I really urge you to attend these keynote talks that will include some award-winning student talks. And then we have poster sessions on today and tomorrow. And we will vote, you will all vote, as my father used to say, vote early, vote often. Vote for the best poster so we can hand out the award at our banquet. And so the groups that we'll meet, the focus research groups and breakout groups, they each have their own individual attendance structure and again, we hope you attend them and make sure that they can start developing the visions for systems 3.0. And finally, my last slide is to let you know that we have three outstanding winners today for this meeting. Mike Ellis are winning the program director award for 2017, Julia Moriarty winning the student model award and outstanding thesis and the competition this last year was fierce with the numbers of people that we've had to go through the various publications and theses and Julia won. So I'll hand it off to her for her achievements. And then finally, our lifetime achievement award, Bob Anderson, I don't think there could be a better winner of our lifetime achievement award than Bob. And I'm so pleased that he's winning this award while I'm still director. So thank you, Bob, for this achievement. So what you see in all of this will be celebrated at the banquet and please congratulate them when you see them. And now I will introduce 30 seconds left. Our next speaker is going to be Marko Janssen who's going to be talking about two modeling cultures. So please welcome Marko Janssen.