The Open Modeling Foundation: a Global Community for Standards-Based Modeling of Human and Natural Systems
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
| Title |
| |
| Title of Series | ||
| Number of Parts | 5 | |
| Author | ||
| Contributors | ||
| License | CC Attribution 4.0 International: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
| Identifiers | 10.5446/62444 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
|
2
00:00
TransportRapid transitSpare partRolling (metalworking)MaterialTrainTextileBuick CenturyHot workingSizingAlcohol proofDepth soundingSeparation processComputer animationPanel painting
08:56
ToolToolMultiple birthHot workingBuick CenturyShip of the lineSpare partToolFood storageBook designHose couplingComputer animation
17:51
Multiple birthCylinder blockFord FocusEngine-generatorWorkshopPaperFood packagingToolFirearmSpare partRutschungComputer animation
26:47
Multiple birthCanberra <Flugzeug>WorkshopEngine-generatorSpare partShip of the lineHot workingSpaceflightFinger protocolAutomobile platformCylinder headTrainPhotographic processingMail (armour)Spring (device)WorkshopCommitteeHose couplingComputer animation
35:42
CommitteeWorkshopPaperFinger protocolAutomobile platformFullingSpare partStagecoachCommitteeMoving walkwayHose couplingWorkshopComputer animation
Transcript: English(auto-generated)
00:00
Thanks for inviting me to talk about the Open Modeling Foundation, the OMF. And so what I thought I'd do is kind of give you a background of why the OMF was formed in the first place and sort of how this happened, give you a background in it, and then bring
00:21
you up to date on where we are now. It's very recent and we've only just started our work this year. I want to start out. My training is in both social and natural sciences, in anthropology and earth sciences and within anthropology, long-term social change in archaeology.
00:43
And so I'm interested in the long-term interaction of people and the environment. Let me click here, there we go. And so I want to start out by pointing out that we live in a coupled natural human system. Sometimes if you're in a big city, you think that it's all human and everything you see
01:04
is constructed by people and we've pretty much created an artificial world. And of course, if you're out in the wilderness, you say there's a lot of world that has no impact to people. But neither one of those is true. In even the most densely populated built-up city, natural processes still occur, buildings
01:26
weather just like rocks or waterfalls for the sky and rolls downhill, even if it's conducted in streets and sewers as opposed to streams and rivers. And in the most distant part of the world, even in places like Antarctica, we have evidence
01:44
of particulates from human activity, the effect of people on the atmosphere, in all of our natural preserves, our wildlife areas. These are largely managed by people, you know, what animals survive and which do not.
02:01
So we live in a world that's very much a human and natural world. This picture is, I think, a good illustration in that you see fields cultivated by people, plants cultivated by people, but those plants are growing on their own using ecosystem services like oxygen and CO2 and things like that.
02:23
Animals activities also produce more of this CO2 and methane and things like that and other gases. We contribute organic matter to the surface, but so does other things and pedogenesis that soil formation still happens. So there's this tightly connected world that we live in.
02:41
And it's not just any part of the world, it's the whole planet. You know, as a social scientist, I think of only a few short millennia ago, as a historical scientist, people were normal in terms of how they lived their lives. They were a normal, large mammal, large primate, a very successful one.
03:03
We lived in groups like this of 25 to 50 people successfully collecting and obtaining resources both to eat and to make things that we needed in our lives. And we're so successful that people spread all over the world.
03:20
And so this, but we were still normal, right? And in terms of large mammals, but we're not normal anymore. Now there are an enormous number of people. We're closing in on 8 billion people on the planet. That's a lot of any animal and especially a very large mammal like ourselves, a really,
03:44
really large number. And instead of living in normal social groups, and people have always been highly social mammals, and part of our ability to transform the world comes from our ability to cooperate and coordinator activities and social groups, right?
04:01
But we live in really unusual social groups. We live in these enormous hives of millions of people we call cities and urban areas. And so the only parallel in the animal world to the way that people live today with over half the population living in these hives are social insects, ants and termites, I mean,
04:25
that's really weird. They're not even mammals, right? Not even vertebrates. And this makes us truly, truly unusual and in fact, some of our largest cities, places like August and Mexico City and Tokyo are larger than the hives and nests of some of
04:45
the most prolific social insects like termites. So this is a really, really weird world, social world that we have created as part of this coupled natural human system. And on top of it, on top of it, are currently today, our digital media and our ability to
05:07
move really rapidly. I was just in Europe a week ago, and I'm back here in the US, and I'm talking to you over digital media. This has connected everybody socially, culturally, economically, and of course, as we know from
05:23
the recent pandemic, epidemiologically, in a planetary scale network of cross-cutting relationships and interactions. As of several years ago, the number of mobile phone subscriptions exceeded the number of
05:42
adults on the planet. And we're closing in on everybody on the planet being within one degree of separation of access to a smartphone. That means that anybody on the planet, any of you, me or anybody else, can talk to anybody
06:01
else in real time, 24-7. And through smartphones, we have access to this enormous cumulative amount of knowledge and information of the entire human species, and that includes everything from the most profound and deep and serious scientific knowledge to trivial sound bites and misinformation.
06:28
So again, we live in a really unprecedented social world, and this is a coupled social and natural world. So it's not just social, we're affecting the entire planet.
06:43
As a graduate student in anthropology, one of the things that was drilled into me is that we're mammals like all other mammals, right? All animals are unique. But in fact, humans have become really odd, and so the things that we do, our ability to manipulate the planet, the size of our social groups, the amount of interaction is
07:05
unprecedented for any organism on the planet in its whole history so far. And this has consequences in terms of our ability to get along in the world. A few centuries ago, and again, as a historical scientist, I mean, just a few short centuries
07:22
ago I can say this, people lived in communities. You would have lived in a small village, and anybody living in that could understand, observe, and be able to get along socially with everybody in the village. You knew everybody.
07:40
You knew all of their roles, and good and bad, how to get along in this. And you had a pretty good sense of if you did things, whether you were clearing fields or hunting animals or building something, you had a good sense of how this would affect
08:00
your world today and in the future. We don't know, we can't do that anymore. I don't live in the largest city in the world by any stretch of the imagination, but there are almost 5 million people in the Phoenix metro area. There's no way that I have any contact with or understanding socially, economically of even a tiny fraction of that number of people.
08:23
Likewise, all of us have impacts on the world that extend around the world. All the things that you guys are sitting and using right now, just ignore the high technology stuff that we're using, the clothes you wear, the things you have on your desk, all of those. If you were to track down the materials that were made from, where they come from in the
08:45
world, you would find out that you have a footprint over the entire globe. So we don't really have any sense, we can't have any clear sense of how we impact the world today. And so in this world, that's dominated by these rapidly changing, telecoupled human
09:05
and biophysical processes. How do we try to make a world that we want to live in, that we can survive in, but one that is worth living in for today and in the future, for our children and our grandchildren?
09:24
And we have to face the reality that we can't use the same kind of conceptual abilities that have served us so well socially and ethologically over the past couple of million years. We live in a totally different world than the ones that we've successfully managed and
09:42
been able to manage in the past. So somehow we need to be able to expand our vision, our ability to both understand and manage, in some sense, the world in which we live from just those local villages and gardens that are our history and where we've been successful to an entire planet.
10:07
Because we live in this complicated, globally connected socio-ecological system, if we want to have a successful and sustainable future, we have to think about planetary management. And quite honestly, that's something that no one's ever done before.
10:25
If we feel worried that we're not doing a good job of managing the planet, we should at least be realistic and say, look, we've only been trying to do this for a very short period of time. No one's ever done it, much less done it successfully.
10:41
So of course we're stumbling around with this. Because of the complexity of the world that we live in, we really can't just depend on our powerful intuition to be able to understand and organize and drive the societal system
11:01
in which we live, especially the interaction between people and the environment. There's still this feeling that if enough good meeting people get together and talk about it, we can figure out how to manage the planet. And I would say that that's beyond our native capacity. But people have always used technology, going back a couple of million years, we've always
11:23
used technology to expand our ability to get along in the world. This is part of the human niche, we're a technological animal. And we have created increasingly sophisticated information technology from the first clay
11:40
tablets through printed books and now with digital technology that allow us to store data and information outside of our brains to manipulate that information and be able to understand complex interactions in ways that we could not do otherwise. And so I think that if we will have any hope of being able to manage an entire planetary
12:01
socio-ecological system, we have to face the reality that we have to utilize the technologies we have to do this. We can't just do it on our own, no matter how well-meaning and how smart we are. And of course, we are doing this. So we've begun developing these digital tools, mathematical and computational models beginning
12:24
half a century ago, to better represent and study the complex biogeophysical processes that drive the world. And they've been very successful. We would have no idea about our society's impact on climate, on oceans, on land, especially
12:47
at global scales without these kinds of tools that would be impossible. So they've given us really, really important insights, not only on how the socio-ecological system that we live in works today, but it lets us begin to look into the future, at
13:04
least potential futures, and get a better sense of what they might hold in ways that are just impossible otherwise. So these have been very, very important, what I've talked about in terms of first-generation models, but they're beginning to face limitations.
13:21
So we've been building these and using these now for half a century. Importantly, because they're really important for representing interactions between different aspects of the geophysical and biological world and social world, they should be critical
13:43
for looking at these interactions between human systems and natural systems. But currently, most of these models do not do this. They're not set up to represent these kind of complex interactions between people and the environment.
14:00
And in fact, some of the most successful and best developed modeling of the world focuses primarily on natural processes, atmospheric physics, ocean currents and thermal clients,
14:21
biogeochemical cycling, ice sheet movements, things like this. Those are really important, really important. But there's this big emphasis on modeling natural processes over social processes, and certainly, the funding and the amount of effort dedicated to modeling natural processes
14:42
is much, much orders of magnitude greater than the resources dedicated to understanding and modeling societal processes, even though the modeling that's been done shows that human society is having such an enormous impact on the world. And then when we do model human systems, they tend to be modeled completely separately from
15:06
biophysical processes, as if the biophysical world is unimportant, it doesn't exist equally. Once the models of biophysical processes ignore people, all of the large global earth system
15:21
models that are really important in IPCC reports and trying to help us foresee the different effects of what would happen in the future under different human scenarios, right, don't include people, they don't include people, which is a real issue. And then we can't look at the feedbacks between natural and social processes.
15:42
Additionally, a lot of these models, the most successful ones, the most widely used ones, have been people who have been building them now for decades. And they've become very large, very complex, these big code bases, especially primarily in Fortran, that have accumulated over a really long period of time.
16:02
And this makes them difficult to modify, difficult to understand. And many of the most successful are system models, I would maintain that probably no single person understands how those models work, any of those models work. There are very talented, very dedicated people working on them.
16:20
I work closely with people at the National Center for Atmospheric Research that manage one of the best community or system model, and they do want to work on it, but each software engineer and team works on some chunk of it. They don't know how all of it works together because it's millions of lines of code. This of course makes it very difficult to modify, to use.
16:45
You can download the community or system model, you can get it, it's open source, but could you use it? Probably not. You would need a supercomputer, you would need specialized knowledge. And so this, in and of itself, restricts accessibility to this kind of technology.
17:05
Additionally, most of the kind of models used to try and understand the interaction between people in the environment today, most of the best known ones, have been developed in well-funded research institutions and universities in the global north by limited teams of specialists.
17:25
And in most cases, they're not available to anybody outside of those teams. You can get CESM, that model, but you can't get most of them. And they're sort of like black boxes.
17:40
So you have just to trust that these people have done the right thing. You don't know what the algorithms are and policymakers don't know what they are. How can you trust these models? This has led to problems in the past and in the present where people accuse scientists of fudging the data and making things up because these are secrets.
18:03
So this is another series of kind of models that are out there. The same situation makes them relatively inflexible. A lot of them are designed to answer questions that were important a decade or more ago simply because it takes a lot of effort and time to be able to modify any of these to answer
18:22
questions. So it makes it difficult to study these kind of socio-natural systems and also multiple scales. CESM, it runs globally and that's great, but it's fairly worthless for trying to understand the detailed environmental dynamics of the Phoenix metropolitan area or Berlin or Paris
18:44
or anyplace else like this. And so they can't scale up and down, you have to pick a scale, you have to pick a set of questions and we're stuck with that. So these are some of the limits that are there. How do you deal with this? So beginning in 2015, a number of workshops were organized around the world, organized
19:05
by several different organizations. Ames was the real lead in this. This is analysis, integration, and modeling of the earth system. It's an arm of Future Earth. It also involved CompsysNet, which I direct, which is a network and code library for socio-ecological
19:25
sciences. CSDMS is a similar scientific network and code library for the earth sciences. And then there was a number of other groups. So we organized a whole series of workshops in different places in the world to talk about
19:42
the future of modeling. So what's the next generation of modeling social and ecological systems if this is a really important tool. This is just a quick overview. You guys are recording this and I can save the slides so I won't spend time on it. But the point is that we had a lot of these workshops in many different places to try
20:02
and bring in the global community of people doing modeling, developing modeling, policymakers, to talk about what is needed for the future. These are some of the questions that we discussed. So how do we integrate modeling of human and biophysical systems? How do we put these things together, which they're not done now?
20:22
How do we make modeling more flexible so that you can scale up and down to answer different kinds of questions and challenges as they arise? From the global to the very local. How can we move beyond these limited research teams in the global north to access more diverse
20:43
conceptual approaches to modeling? There's a lot of really smart people in the world in many different places outside of where these models are created. And certainly to some extent, some of these organizations bring in scientists from elsewhere,
21:00
a few, to participate. But how can we do a better job of getting a much more diverse range of intellectual contribution to thinking about how to model human and earth system dynamics? Along with the flexibility, how do we set up a kind of the ability to rapidly answer
21:23
questions we have not thought about and future challenges? I mean, the questions that we're facing now, there's going to be new ones in the future, new challenges in the future as we move ahead. And then going the other direction in terms from creating these, if this kind of modeling is such an important tool for understanding the complex world in which we live and being
21:45
able to manage it in a better, more sustainable way, how can we make sure this kind of technology is available worldwide? How can we democratize this beyond the way it is now, where it's very restrictive?
22:02
And in order to do all these things, we need to create some kind of a scientific agenda for integrative modeling. How do we do that? And so this is part of the discussion we had and building this new capacity in not just next generation modeling, but next generation modelers, how do we build human capacity for
22:23
data science and modeling to be able to do this at multiple scales and not throw away what we've already done, but go beyond these legacy systems? So there was a lot of wide ranging discussion in these, and there are a lot of things came out of it, but there was one of the things kind of a consensus that the future of next
22:46
generation modeling is a shift from these big piles of Fortran that we have now to some kind of a more distributed computational environment.
23:01
Think of it as an ecosystem, potentially interoperable models, if you're familiar with something like the R ecosystem or the Python ecosystem, where you have many different kinds of models that are created, but created in such a way that to answer different questions, you can put them together in different ways to answer different questions at different scales.
23:22
This would allow modeling to be more flexible and adaptable, respond to changing challenges, and would make it easier and actually encourage contribution from a more diverse scientific community around the world.
23:42
If more people could contribute to this ecosystem in a way, then more people would have a stake in it and be able to use it to make this technology more accessible and usable to a wider global audience. So this is a future that we thought was great, but how would we actually make this happen?
24:02
Well, you need to have code that's available and discoverable by other people. It needs to be understandable and reproducible. Those of you in the FAIR community would recognize that we're talking about the kinds of criteria that are under the FAIR acronym and then some.
24:24
We probably need to do a better job of leveraging new technologies like containerization, packaging models with their dependencies to make them both more reusable and usable in a wider variety of platforms, including supercomputing platforms.
24:41
And of course, we need some kind of standardized global APIs to allow us to actually link models together. If you want to have an ecosystem, potentially interoperable models, there need to be some common APIs that would allow this to happen.
25:00
Technology is an important part of this, but to have all this happen, this involves also common protocols, common ontologies, and we have to do this across multiple scientific domains. So it's not just good enough that social science does this, but we need to have this across ecological sciences, geophysical sciences.
25:23
And so while technology is important for realizing this kind of ambition, it's not sufficient. This is really a social challenge, the science as a social practice. How do we change science to enable a new generation of modeling that we need for a sustainable
25:44
planet? And the idea is one of the underlying things is we need to shift the way people think about doing modeling, which is normally today, there's some questions, you build models to answer those questions, you write a paper on it, and you're done.
26:04
We need to think about designing models produced by others and not just ourselves. And so we need to have some kind of widely adopted community standards for how models are deployed and documented and integrated with this kind of a vision to actually happen.
26:27
It doesn't exist today. And so these are the things that we would need. We need to have standards, software standards for best practices and adopted by multiple
26:40
scientific communities. We need to communicate information about standards, administer them so that people who are looking at code will know whether or not a particular model meets certain standards or not. We need to create professional incentives because this is extra work to actually create,
27:03
to do your models rather than just one off to answer your question, to write models so that you can use by others. This is extra work, better documentation. So how do you know, everybody's busy, everybody has a lot of things they have to do. How do we provide incentives that enables scientists to get professional rewards for
27:25
best practices, and then provide education and training in this, and all of this requires some kind of a coordinated international effort. So this is the basis for creating the Open Modeling Foundation as an initiative, as this
27:41
kind of meta-organization to try and fill this role currently. So after these workshops, a number of things happened and one of them was, it was kind of serendipitous, there was a meeting of the International Environmental Modeling and Software
28:00
Society in Fort Collins, Colorado at Colorado State University in 2018 that was attended by representatives from numerous large modeling organizations from around the world. And so several of us got our heads together and contacted people from those organizations
28:22
and said, look, we know some people at Colorado State, we can get a room, but you'd be willing to stay on after the conference for a day or two and talk about the idea of creating some kind of an international consortium for modeling standards and best practices. And they agreed.
28:40
So we ended up with six or eight different organizations represented and got together in Fort Collins. I gave a little spiel about why this might be a good idea, I think it would take some time to convince everybody. After 15 minutes, everybody said, that's great, let's do it. So we started talking about how we could create an international organization for standards
29:05
and best practices in computational modeling. And after those days, the goal was then to try to seek funding, organize workshops, bring in the community. And in 2019, we got funding from the Alfred Sloan Foundation to do this kind of planning.
29:23
And so we held some strategic planning workshops, we were going to hold them in interesting places around the world, to bring in the community in 2020. And of course, as you know, that kind of thing didn't happen. And we ended up doing them all on Zoom, though they worked out.
29:42
We got a lot of people involved, we contacted people from over 100 organizations, we managed to get representatives from almost 70 of them to participate in these planning workshops, representatives from many different kinds of organizations, and from most parts of the
30:03
world, not every part, but a lot of parts of the world. So we started trying to put this together. Because we had to do it all online, we created a collaboration platform based on GitHub,
30:20
in order to capture the discussion and ideas out of those workshops, so that each workshop could build on the work of the previous workshop. So we kind of had to throw that together quickly when all of our live workshops were canceled in 2020. But it actually worked out pretty well.
30:41
And you can go to the Open Modeling Foundation site, and you can dig into the GitHub part, and you can see all the discussions in the discussion forum in the issue tracker, all the responses and ideas and things like that, and the documents that were produced.
31:00
We did some other things. We did a pilot project of model coupling. We did some papers, and a parallel project put together by Ames on trying to create large scale behavioral models to couple them into our system models is ongoing.
31:21
So after two years, actually almost three years of planning and discussion, community involvement, we drafted a charter, kind of a constitution, a charter for the organization, and scheduled a founding meeting, as at the end of 2021, online and in Washington DC at
31:44
the issues facility there. And it was attended by representatives from 50 organizations around the world, 47 of which agreed to become founding members. And we used our GitHub platform to actually go through the charter line by line, have
32:02
amendments, have votes, and final approval of the charter and a statement of value. So you can see here. So I'll read the first one, you can read the other one, but this is part of our mission that the Open Modeling Foundation is an international open science community to enable next generation
32:23
modeling of human natural systems. And importantly, it's an alliance of modeling organization to coordinate and administer common community standards and best practices and provides the resources to do that.
32:41
So how's it organized? Well, it's not a professional society of individuals. The idea is it's a consortium or a collaboration of organizations, formally constituted organizations that represent modeling scientists or modeling around the world.
33:05
And each member organization has one representative, one voting representative to the members council. And that's kind of the ultimate governing body of this consortium that approves all changes to the charter, all activities, all standards, and things like that.
33:24
However, then there's also actually working groups that put all this stuff together that create standards that help identify ways, come up with ways to identify standard based models to provide professional incentives to do education, to develop cyber infrastructure
33:43
to support this. Then we also have a working group on early career scholars to try and encourage next generation scientists to participate in this. So whereas the members council are representatives of organizations, individuals, whether they
34:02
are from a member, a represented member organization or not, can join working groups. So any individual can join a working group. So one thing that we've built into the charter is each working group must have multiple chairs. These are volunteer positions.
34:21
And so we want to spread the work around. And also we want to try and diversify the leadership. This is a way to try and diversify the leadership of the Open Modeling Foundation. So they have to have at least two or more chairs. And then to kind of manage it all, we have an executive committee. There's elected members from the members council, the chairs of the working groups, non-voting
34:45
members from cyber infrastructure, early career scholars and non-voting director, me in this case, elected position. And we have regular meetings to try and deal with the day-to-day business of the organization.
35:01
We had our first members council meeting, annual meeting, this spring and hosted in Leipzig, Germany and online. So we're finally getting started on that. And this is like we're trying to set up kind of a mailing list for people who can't participate but want to keep up to date.
35:20
And so this is just a quick visual of the range of organizations that are currently members of the Open Modeling Foundation. And we're actively seeking additional member organizations. And so if your organizations are interested in participating in developing and administering these standards, I would invite you to join the O&F.
35:43
So what have we done so far? We created an executive committee. So we had to elect a director, elect members council representatives. We had elections in 2022. And then those people nominated chairs of working groups that then had to be voted on
36:00
and approved by the members council. So they became part of the executive committee. And you can go into the site and see all the current members of the executive committee, the working chair heads, and they represented this. Then we started seeking funding for startup. We haven't been successful with National Science Foundation, but the Sloan Foundation has been
36:23
incredibly supportive and interested in what we're trying to do. They gave us a substantial grant to get us started in the first two years. We can't pay working group chairs, but we can give them time, the idea. So we're going to pay graduate students to help them out.
36:41
And then we've got travel support for members council members, working groups to attend meetings and travel support for early career scholars to try and present research and activities on standard-based modeling. And then a project coordinator.
37:01
We were tasked with several things to get things started in 2022. One was to set up a task force to create a code of conduct to go along with the value statement that's done and is out for approval to the members council now. We're supposed to communicate to the larger modeling community, the OMF. So we did a couple of papers, one in PNAS and one in environmental modeling and software.
37:25
This came out last fall. And we're in the process now of changing, transforming this GitHub collaboration platform into a full featured science gateway. So to move from kind of the planning stage to turn it into a much more informational
37:43
and useful platform for the organization. And we're working with the science gateways community Institute on that. And so here you can see what it looks like currently. Hopefully it will be looking better in the coming year as we begin to try and implement some of the suggestions from SGCI.
38:03
Even though we're just barely getting started, there is broad interest. I looked with Google Scholar and this is just a quick one to show the people that have so far visited the OMF science gateway. We've had visits from every continent on the planet. Some continents better represented than others, but people, so it's not just a one country
38:26
thing. This is truly an international effort. I'm really pleased about that. And so what we're trying to do now is improve the gateway, want to expand membership. We're planning on doing a special workshop at the summer meeting in 2024 of the international
38:43
environmental modeling and software society. All the working groups are getting started. The standards working group has already had a couple of meetings. We're meeting with the certification working group, Chris Erdmann's on that I think next week and education and outreach group are planning a speaker's forum and note the newsletter
39:01
and the early career scholars are working on a high profile paper on best practices. So things are getting started. I'm really happy about the level of enthusiasm and activity in the big challenge of any organization like this or any other is getting people to be involved and be active.
39:20
You know, it can't just be a couple of people you need a broad involvement and so far that's happening. So that's very encouraging on this. And so of course we don't, this is what we don't want and I'll end it here and see if you guys have any questions. Thank you. Thank you.