 Good morning. Good afternoon and good evening. Welcome to the IBM Research Horizons events. My name is Anna Topol. I'm the Chief Technology Officer and a Distinguished Engineer at IBM Research. It is truly my pleasure to be a host for you today for the virtual event that actually will be going on for the next three days. It is my pleasure to comment that we are having speakers and panelists from all over the world, but this session specifically is being streamed from our IBM Research Lab in Zurich, specifically in the Russelikam. So our horizon event is focused on trying to help you gain insights of what IBM Research is doing around the topic of future of climate. But we also want to have a discussion and thought leadership provoking discussions with our panelists and speakers who are joining us from outside of IBM. We really want this session to be interactive for you. So please ask questions on the right side of the video. You see the box for a chat. So kindly type your questions throughout the day. And without any further ado, I will introduce our first speaker for today, Salomon Osafa. Dr. Salomon Osafa is a Vice President of IBM Research. He has research strategy, development, execution and partnerships around what we call the Impact Science Organization, which covers future of work, future of climate and future of health. Furthermore, Dr. Salomon Osafa is also ahead of the Research Lab in Kenya and South Africa. And today he will take time to really help us understand what drives the research agenda of his team and what are some various types of technologies, specifically the structure of technologies that the team is using to address the challenge and climate impact. So Salomon, back to you. Thank you. Thank you, Anna. If we could show the slides. So great to be here. This is going to be a very exciting three day event with, you know, keynotes and panel discussions and also some deep dives. And of course, as many of you know, to tackle a topic such as future of climate partnership is going to be very important. And I'm also, you know, aware that many of you know the topic very well. A good reference is actually the UN international or intergovernmental panel climate change report that just came out a month ago, which highlights, you know, the importance of this topic. And there's a couple of points that the report makes, right? One is climate changes here. We're all witnessing it from, you know, the perspective of the different extreme events that are happening that we're seeing, you know, more frequently. It's also very clear that, you know, human activities are at the midst of it, right? So they're the major cause for global warming. And of course, the third point that the report makes is it's a massive problem. It's a global pressing challenge. Now the silver lining in all of this is the fact that, you know, technology can really help here, especially with the advent of some of the latest technologies. If we bring that together for the next decade, there is a real possibility that through partnerships we can make, you know, a big advance to tackle climate change. Next. And from IBM research perspective, we do believe that we could be, you know, very helpful in terms of tackling climate change. And that's really for three main reasons. One is we're a global organization. As you can see on the map, we have labs across the world in almost all the continents. We have research labs actively developing some of the latest technologies and working on problems like climate change. The second is our team of 3,000 scientists is actually composed from different disciplines all the way from AI to material science to mathematics and, you know, semiconductor industry and in many different disciplines coming together to tackle these problems. And that's the type of skill that's required to really make a dent on this problem. And then, of course, the third one is the fact that we have over 75 years of experience in developing technologies. We have been at the center of the IT revolution. And, you know, recently we have been at the forefront of advancing some of the latest technologies such as artificial intelligence and quantum computing. So we believe that we can partner with other industry leaders and other enterprises and governmental entities and many other partnerships, which will really allow us to be at the center of making progress when it comes to developing technologies for climate change. If you go to the next slide. So many of you are, again, aware that, you know, enterprises are really accelerating their action when it comes to climate change. And there's a few drivers here. The first one is, of course, investor pressure, right? A lot of investors are actually taking voting action and making sure that company directors are being accountable in terms of their climate change. Their climate change commitments. And, you know, shareholders are really holding companies accountable. So there's a lot of, a lot of activities going in that space. The second one is consumers. Consumers themselves are asking companies to be more environmentally conscious, right? They're being very proactive. They're being very conscious about the products that they're getting from, you know, different companies. And this is actually something that we see across generations, right? Across generations, people are becoming more and more inquisitive about companies and their commitments and what they're doing towards climate change mitigation and adaptation. And then the third one is, of course, a lot of policy shifts that are happening. Governments making commitments, for example, the European deal or the Biden plan for clean energy and also China and others really making commitments in this space, which is also shifting the landscape. So there's a lot of, you know, these points that are accelerating actions by enterprises. If you go to the next page. Now, from IBM perspective, we have been committed for many decades regarding environmental action. We have been publishing our environmental reports for over three decades and disclosing some of the reductions committing and, you know, reductions and being very transparent about it. But this year has been very big and major for us as well because of some of the big commitments that has been made by our CEO. And the biggest commitment that we have made this year is the fact that we will achieve net zero by 2030. And our goal to achieve net zero is actually also by really increasing our commitment to procure more renewable energy and by reducing our emissions. So not just, you know, purchasing offsets, but really we're clear that we want to use more renewable energy across our operations and we want to reduce our carbon footprint and really make commitment regarding that. So by 2030, we will have 90% of our operation being run by renewable energy. Beyond our own commitments also, we are participating in forming many partnerships, right? One of the big ones where we're central to the formation was the MIT climate and sustainability consortium. And it's a very unique one which actually brings some of the biggest companies in the world together for just to mention a few names. Apple, Cargill, Dow, Pepsi, Boeing and many others bringing them together so that we can actually jointly identify some of the biggest problems that we can solve and deploy the latest talent from various universities and really make significant progress. So it's one of the exciting kind of consortiums that we have joined, the partnerships that we have formed and we're in the process of forming many more. Because like I say, this is a global problem, it will require partnership to make significant advances. Next. So IBM has, you know, a unique lens in terms of the problem and the technologies that are needed because we work with some of the biggest enterprises across different sectors, right? And that gives us visibility on some of the business drivers and some of the pain points, right? For example, if you take oil and gas, you know, they are interested in making, you know, driving action as it relates to regulatory pressure and some of the investments and the change that's happening in terms of renewable energy, right? So how do they become more transparent? How do they reduce their carbon footprint emissions and also diversify their investments, right? If you look at energy and utilities, they are very much interested in energy transition, right? And the new electrification efforts that are happening and how do they quickly manage a shift to support that effort. But they're also interested in managing risk to their infrastructure, to their assets, right? And informing their operations based on some technologies that are able to forecast extreme events, right? If you look at the financial services industry, they're definitely interested in identifying and quantifying risks that are related to the investment portfolio. And they're also interested in exploring new opportunities when it comes to carbon reduction and climate resiliency. And then last example would be retail supply chains, which, again, there's a lot of mitigation efforts that need to happen because supply chains emit a lot of, you know, carbon. And also importantly, supply chains need to be more and more resilient to the extreme events that are happening more frequently and they're becoming more severe, right? So if we look across all these different industries, all these enterprises, there's a few patterns from technology perspective that we believe are emerging. The first one is that enterprises need carbon accounting and reduction technologies so that they're able to do better reporting and they're able to drive their own business KPIs. The second one is that there's more and more of a need for technologies that enable modeling and prediction of climate risk and impact so that it informs operations and informs resiliency. And there's also more of a need for data and AI technologies that are able to ingest and correlate a lot of the data that's coming out, you know, from different sources and make it useful for decision makers. And the last part is when we do that, it's also important to make sure that these technologies are able to plug into already existing process flows or workflows, right? It's not just inventing the wheel but make it much easier to integrate these technologies into already existing tools and workflows. So those are the patterns that we have identified across different enterprises. If you go to the next page. So it's with this in mind that we have structured agenda for the Horizon event in the next couple of days. We're doing quite a lot when it comes to developing technologies related to mitigation and adaptation. And we're doing this not just, you know, in our own silos, we're developing these technologies through different partnerships. So throughout these three days, you will be able to see some of the progress that we're making, for example, in materials discovery for carbon capture, enabling carbon performance and integrating that into applications, as well as some of the platforms that are being developed to enable geospatial analytics and decision making and how we're also enhancing climates prediction or weather forecasting using AI and ultimately once again integrating this into applications. And, you know, all of this is being enabled by the discovery technologies such as AI and quantum that's being developed within IBM Research. We believe that these technologies will be at the heart of accelerating the development of the tools that I mentioned and accelerating even the scientific discovery that will be required in the next decade so that we can make some significant progress. So with that, let me introduce our speaker for the keynote. Greg Downing is the director of sustainability at Cargill and he leads the development and execution of Cargill's climate change strategy. And he's doing that by partnering with the business units across Cargill to develop reduction strategies across all of Cargill's global activities. And Greg holds a PhD from Columbia University in New York and previously he has worked with Target on leading energy and greenhouse management programs. And also he has worked with a lot of ClinTech startups. So with that brief introduction, Greg, welcome. Thanks Solomon and thanks to IBM for inviting us. We're really excited to be able to participate today. We've been working with IBM quite a bit lately both within the MIT consortium but also just directly with IBM and have been really impressed with the level of knowledge that they bring to the topic and really the passion that they have around it. So we really see IBM as an important partner and an important player in this space and really driving action on climate change. And kind of also in that spirit of partnership, I'm excited to have Wayne Honeypot from the SWEI-Haus Institute with us today as well. Wayne is going to talk a little bit about some of the work we've been doing together on understanding some of the economics behind a little bit of what I'm going to talk about. So I'm going to talk about what Cargill is up to primarily focused on the agricultural side of what we're working on. So just a little bit of background on Cargill, I think most people have heard of Cargill but a lot of people have no idea what Cargill actually does. We are one of the largest privately held companies in the world. We work in many countries in the world and certainly sell to most countries in the world. And really we are kind of in the middle of the food and agricultural supply chain. We work a lot with farmers who we consider customers of ours and then we work with downstream customers like food service customers, CPG, consumer baggage goods companies, retailers, and really help to get food from where it's produced to where it's consumed and needed. And at the heart of what we do is really our company's purposes which is to nourish the world and to do that in a safe, responsible and sustainable way. And obviously climate is a key piece of all of those things. And so we are actively working on, very actively working on figuring out how agriculture can be a part of the solution to climate. And we also know that there are other things beyond the climate like land and water, people that really agriculture is a key component of. And as we are building out programs and really trying to accelerate what we're doing, we really center it on how do we make sure that investments are made in farmers so that they can take the actions that we all need them to take in order for agriculture to have the biggest impact that it can have. Certainly we do a lot in our operations around energy and trying to reduce our footprint there. We've got a large transportation footprint as well that we're doing a lot in that space as well. I'm going to focus today on the agricultural part of what we're doing, which is really the biggest part of our footprint and the biggest part of our efforts. And our climate program, which is based, we have greenhouse gas reduction goals both for our operations and for our supply chain that have been approved by the science-based targets initiative. So they're in line with the Paris Climate Agreement. And on the agricultural side of things that's really centered on trying to drive regenerative agricultural practices in all of the geographies and supply chains that we work in. And really regenerative ag has a lot of benefits. It has climate benefits both in terms of reducing emissions but also pulling carbon out of the atmosphere and storing it in the soil. But the same practices also have really big, really good water benefits in terms of reducing the amount of water needed for irrigation, reducing the amount of fertilizers used and therefore improving water quality. And then, done right, regenerative ag can have a really positive impact on farmer livelihoods both in terms of potentially increased crop productivity but also in terms of resilience to drought, to floods and other extreme weather events, which will only become more common or extreme as the world continues to warm. So that's kind of why we're focused on regenerative agriculture and sort of what we're focused on within that. Really, there's a handful of practices that we're trying to drive on the farm. So, you know, one is to keep the soil covered. After harvest is a possible plant cover crops that help hold the soil in place over the winter in the northern hemisphere. Other opportunities to reduce tillage to help not turn the soil up and help keep the carbon in the soil. Increasing crop diversity also helps with soil diversity and that helps enable the soil to store more carbon. Similarly, keeping living roots in the soil is really important. Then, where livestock can be integrated into real crop production systems, that can be a really effective regenerative ag practice. It can be a challenge to do logistically, but when and where we can do it, it can be actually one of the most impactful practices that we have to help try regenerative ag. So that's kind of generally, and then at Cargill, sort of how we're approaching kind of the principles that we have in mind as we're building out our climate program. You know, we know we need a portfolio of programs. We don't. There's no silver bullet. We need to do a variety of different types of activities. We need to drive a variety of different activities until we're taking a number of different approaches to see what works, see what doesn't work and figure out what ultimately we can scale to the level that we need to scale. We're really focused on outcomes. As I mentioned, we have a science-based target initiative approved goal, and so our scope three target is really is at the heart of our regenerative ag. We also have context-based water goals and land related, land use related goals. And so as we're designing programs, we really have those outcomes in mind. And as much as possible, we're trying to design programs that address most, if not all of those outcomes that we're trying to address. And that's one of the great things about regenerative ag is that a lot of the practices address multiple issues. And one of those, actually, the third bullet is being farmer-centric. You know, there can be really good financial benefits to farmers on the regenerative ag. And we know that anything that we do that we want to scale has to work for the farmer. There's no way we will drive the chains that we need to drive if it doesn't work for the farmer. And so that's always top of mind for us as we're building programs. I just walked through a couple of examples of some of the things we're working on. So the Soil and Water Outcomes Fund is a program in Iowa and in the United States. It is a market-based program where farmers participate. They generate both carbon and water related benefits that are then registered in a marketplace and anyone on the band side can buy them. We help support the initial startup of the Soil and Water Outcomes Fund and we're one of the off-takers of the carbon benefits from there. But one of the really interesting pieces of this, and this goes back to the stacked benefits side of things, is that there are big water quality benefits to the practices that farmers are adopting, particularly less nitrogen and phosphorus runoff. And there are municipalities in this region that have regulatory requirements to reduce the amount of nutrient loading in their water. And so there's financial value there. So the municipalities are buying the water benefits. We're buying the carbon benefits and so the farmer is seeing a nice revenue stream because of the dual benefits that are coming out of the projects that they're doing. So this is a model that we're really excited about and are hopeful that it can actually scale. Another example in Europe is a wheat carbon reduction program in this one. We're actually generating farmers to help them generate carbon credits, certified carbon credits, which are then put into the voluntary carbon markets. This one is specifically on carbon, so there are water and farmer livelihood benefits, which help us support our goals. But for the farmers, it's primarily the carbon benefit that's really driving the value for them. And then just one other example just for a little bit more geographic context. In Mexico, we're working with corn farmers. And this is an example where we're paying farmers directly for specific practice adoption. And so that's another model that we found to be successful. How long can and where at scale will the funding come to continue to pay farmers for those practices? They probably have to be some way to price that into the market, but at least initially we're able to work directly with farmers. And oftentimes we'll have programs like this where we work with customers, where we share the cost of paying the farmers to adopt practices and are kind of experimenting with what those incentives need to look like, what the dollar value of those needs to be to really try to change. So those are just a few examples of things we're doing in various parts of the world. The biggest challenge we see is really how do you scale. So these types of projects, it might work on 10,000 acres, it might work on 100,000 acres, but how do you start to get to a million acres to 10 million acres to 100 million acres? And that's where the challenge really comes in. And there's sort of a number of different areas of challenges. So one is just how do you measure the impact that you're having? The best way is through soil sampling, but that's really expensive and time consuming. We're pretty excited about remote sensing, being able to do a lot of the work, and particularly remote sensing combined with modeling with a little bit of ground truth and from soil sampling. So we think, and this is where I think companies like IBM can really play an important role, we think there's a role for technology on the measurement side of things that end up pretty bullish on that. Assuring quality is another area which is obviously related to measuring, but making sure that the accounting methodologies are in place and from a sort of a corporate inventory perspective, those methodologies are currently evolving and being developed. We know there needs to be verification of improvements that are made, but that can be an extensive and burdensome process. So how do you remove some of the overhead of verification? And then permanence is always going to be an issue with soil carbon. The carbon that's being stored in the soil that can be pretty quickly reversed through floods or through killing or any number of things. And so how do you either discount the accounting for permanence or how do you put in monitoring in place to make sure that reversal is not happening? And then the third area is really around driving adoption. So how do we get farmers to adopt practices? There are some educational and technical things that we think we can help farmers with there. And then there's sort of financial and de-risking pieces. I think farmers are the ones who have to take action, but if you look at the value chain, they're sort of the least able to take on the risk in the financial burden. So how do we set up the system that help make it financially and not make the risk tolerable for farmers? And that last one, the financial piece, I'm not going to hand it over to Wayne from the Soil Health Institute. We've been working with them on trying to figure out some of the financial implications of adoption of soil health practices, of regenerative ag practices. And we actually think it's in the long term a net positive for farmers from a financial point of view. But Wayne, I'll talk more about that and kind of some of the data that we're starting to pull together for that. So that's kind of quickly what we've been working on. I'm going to hand it over to Wayne now. And then after Wayne, thanks for opening up for a little bit for Q&A. So Wayne's the president and CEO of Soil Health Institute. They're one of our key partners and we really like working with them. They're very thoughtful, very practical in this space. So I'm excited to have Wayne here and I will hand it off to you now. Great. Thank you very much. I'm trying to get it rolling here. Hello, everyone. Thank you very much for the opportunity to visit with you. And thanks very much, Greg, for teaming this up really well. Economics is a really important subject. Obviously, when it comes to for farmers and ranchers, cause they are business persons at the beginning of the day and the end of the day. And we know, as Greg alluded to a lot about these soil health practices and systems are very good for the environment. You can increase carbon sequestration and reduce greenhouse gas emissions. You can improve water quality because in many of those practices, you're reducing nutrient leaching. You're also reducing nutrient runoff. But it's really going to become really comes down to understanding what the farmers need when they're deciding whether or not to adopt these practices. And economics is of course one of those key questions that they have. We know that a lot of these practices can be really beneficial for increasing available water holding capacity and therefore drought resilience, enhancing nutrient availability. But still, it really does have to come down to dollars and cents for them because they are those business people that have to make a living on the land too. So this project was to set up to provide farmers with some of the economics information they need when they're deciding whether or not to adopt these soil health management systems. We interviewed farmers on 100 farms across nine states in the Midwest US where some 70% of the corn and 67% of the soybean is raised. The farmers we selected have been using a soil health management system for at least five years. Many of them had been doing it considerably longer than that. In total across those 100 farms, they represented 194,000 acres. They were using no tillage on about 85% of their cropland and cover crops on about 53% of it. You can see we covered quite a wide range in climate and mean annual precipitation ranging from 20 to 55 inches per year or 51 to 140 centimeters per year. And mean annual temperature ranging from 43 to 61 degrees Fahrenheit. And what you don't see here is there's also a considerable range in soils across the central part of the US where the project was conducted. So these farmers were interviewed about their management practices, their impacts on yield, other experiences. And from that our agricultural economists developed these partial budgets comparing a soil health management system. That's what the SHMS stands for there on the slide. The soil health management system compared to the conventional system. And we did not include US Department of Agriculture payments because we know that those payments are temporary. And also because we wanted to see how well these practices stood up on their own economically. So I obviously don't have a lot of time here to go through all of the results, the details of the partial budget analysis. And so the best I can do is just kind of give you just a brief summary. The average farm size where we interviewed the farmers was 1,940 acres. So these were not backyard garden plots. These were very commercial operations. For those of you that don't use acres, that's about 776 hectares. And our results, we found that the soil health management system has reduced the average production costs by about $24 per acre in the US for corn. $17 US per acre for soybean. And the soil health management system increased net farm income for some 85% of farmers that were growing corn and 88% growing soybean. The soil health management system also increased net farm income by about $52 per acre for corn and 45% for soybean. But there are a number of additional benefits that they are also reported. 93% reported that it increased access to their fields. And we think a lot of this is because the soil health practice really helped improve the drainage properties of their soils, allowing them earlier access to those soils. They also reported that the soil health management system increased resilience to extreme weather. And this is very consistent with some of our recent equations and algorithms that we've developed for predicting the relationship between with sand, soil and clay and soil carbon. When we increase carbon, we can increase that available water holding capacity. And so it's not only these co-benefits of additional carbon storage in the soils and could be an additional motivator for farmers to adopt these practices and systems that do store more carbon. And then lastly, we concluded that because these adoption rates are relatively low, just in these nine states, only 41% of the land is in no tillage and 6% using cover crops. And we really think that many others can benefit economically from adopting these soil practices and systems. And lastly, I just wanted to briefly share with you our outreach program on this project. We do not do anything at the Soil Health Institute for academic reasons. We really want to have impact for all of our work, whether it's research or trainings or economics or evaluating environmental impacts or whatever it might be. This particular project, we developed state-based fact sheets and held state-based webinars so that farmers can have the information that's most relevant for them. And in order to do that, we reached out to over 500 state-based agricultural organizations, invited them, asked them to invite all the farmers on their mailing list. We ended up with over 1,900 that registered, 250 of those were farmers that are managing some 386,000 acres. And since we had these opportunities this past spring and early summer, we've had over 5,000 views of the webinar so far. So with that, I will turn it back over to you all. And thank you very much again for the opportunity. I'll turn it back to you, I guess, Anna. Thank you. We are like two minutes before the end of the session, so I will kindly ask you to go back later to the chat and answer the questions. I'll just pull the first one. Let's see how quickly we can go through them. The first one was related to technology, specifically asking, you mentioned, Greg, that IBM is helping with some measurement technologies. But could you comment more in view of maybe even one of the use cases, what specific technologies are differentiating and helping here? Yeah, I think if you look at the remote sensing side of things, we're starting to get between the imagery and the right models and algorithms. There's some ability to look at practice change so you can see tillage remotely. You can see cover crops remotely and even now sort of type of cover crops. And so I think, you know, continued improvement in the imagery space and then the models behind what the carbon impact of the practices that imagery can see I think is a rapidly evolving and I think really important space. I think we'll continue to evolve over the next couple of years. Great. Thank you. I will just grab one more. And again, there's few more, so please go back to the channel. The next question was stated as a challenge. The participants stated that AGRF report on agricultural states productivity in sub Sahara Africa region increased by 4.3% over the past decade, which is twice the world average, but 75% due to the crop area expansion. So how can we improve for our sustainability with those numbers? Yeah, and that's the huge question we face right going forward. We think we're going to need maybe twice as much food as we're currently producing by 20 to be based both on population growth and increased wealth of populations. And so increasing productivity of active lands, it has to be a primary focus of what we're doing. And I think that's one of the encouraging things about some of the regenerative ag work is it does seem like there are product productivity benefits to them. And so I think, you know, just being thoughtful about what one is driving technology improvements from a productivity perspective, you know, from from Genetics and practices, but then to how do you how do we drive practices that really help increase that yield on currently current working lands so that we don't have to expand into new land. All right, thank you. Thank you both of you. Thank you both Wayne and Greg for participating again. Please continue answering the questions if you can. And we will now move to the next part of our presentation. Thank you both. When Solomon presented he commented on the need for the accelerated discovery and so the next presentation will specifically focus on accelerate discovery for carbon capture materials. And I'm delighted to introduce our next speakers, Dr. Matthias Stainer and Dr. Stacy Gifford both co lead the IBM Research Global Future of Climate Materials Discovery Initiative, and they will comment on how their team really uses and leverages the various disruptive technology AI and others to discover new materials for the carbon capture or storage to build partnerships and to scale that technology. They will be joined by Rodrigo Newman, a research scientist and IBM Research in Brazil, and by Geeta Moll, a staff member and IBM researcher in the UK lab. So team, I'm turning to you. As you heard from Solomon, materials discovery has historically been a challenging problem that requires significant resource investment and decades to accomplish. Just think of the scale and effort it takes to develop and deploy novel high performance materials or new drugs. Climate change has its own set of challenges that are rooted in an urgent need for accelerated materials discovery and the deadline is only getting closer. Even if we stop emitting carbon dioxide and other greenhouse gases today, the planet will continue to warm for at least another decade because of the carbon dioxide that is already there. To avoid the worst outcomes of a changing climate, we need to capture carbon both from point sources such as power and cement plants as well as directly from the air and combine this with a global program to put it back underground where it can be stored safely for millennia. Many types of carbon capture and storage technologies exist today, but their application has been limited to a handful of solvents which are expensive to operate due to high regeneration energy costs and rapid degradation. New materials with improved cost and performance are needed to drive global wide scale adoption. Over the last year and a half, IBM has been developing tools for researchers in industry and academia that will accelerate discovery of new materials for carbon capture and better ways to store it permanently. These tools form an end to end computational pipeline, starting with tools that can machine read scientific publications and organize the information into knowledge graphs. It follows the prediction of new materials with AI generative models and performance validation with autonomous labs. All this research illustrates how computers can speed up the discovery of new materials, learning from the past developments and designing new and better materials for future applications. Carbon capture is an urgent problem at global scale. IBM is ready to help. Forming the starting point of materials discovery end to end, the knowledge hub can interact with the human, answering questions or compiling tables of facts on request. As another example, we have developed a discovery tool that simulates the flow of fluids such as supercritical carbon dioxide through the tiny pores of underground rocks using realistic models made from geological core samples. Geasty mail from our UK lab in Darsbury will now begin the materials discovery journey with a demonstration of the IBM Knowledge Hub for Carbon Capture Materials and Methods. Thanks, Rishi and Matthias for the introduction. I will be talking to you about our Knowledge Hub to accelerate the material discovery process for efficient carbon capture. To accelerate the discovery process, we need to understand new domains rapidly as required knowledge is often distributed and cross domain. As shown here, we comprehend by consuming knowledge from various information sources related to a field. One of the critical bottlenecks in any given domain for accelerated discovery is that users are drowning in data. Domain comprehension is achieved by quickly making links and inferences about the concepts in the field. With the right tools, domain-specific knowledge graphs can assist the process by especially highlighting non-obvious connections. For example, in the knowledge graph here, the end-usage reminded that there is a relationship between gas absorption and ionic liquid, which might not be apparent otherwise. Our toolkit collectively called Knowledge Hub aims to assist the discovery process by automatically generating knowledge about carbon capture domain using complex data, be it unstructured scientific content or tabular data which is more structured. The tool helps users understand complex data by performing domain-specific information extractions. Underneath, the tool uses knowledge graphs to store this extracted information while preserving any relationships among data. This enables the rapid development of conceptual understanding of a given domain. For example, the tool can help a material scientist efficiently identify classes of materials, for example, gas absorption or ionic materials, or materials that have high potency in carbon absorption, for example, metal-based materials. Now, you will see our tool in action. Let us assume that a material scientist wants to know about all the materials that can support carbon absorption. The scientist can ask a natural language query from the system to the following event, materials about absorption. The system then converts the query to match the underlying information of our knowledge graph and returns the appropriate data. It does so by automatically interpreting the user question by mapping the question to the underlying data schema. The above question is identified as gas absorption as our domain models are all about carbon dioxide capture. As you can see, the tool can automatically identify the kinds of data visualizations enabled by the underlying data. In this case, we have high potential views. A tabular data view that can provide information about the materials that can support absorption with associated type information and molecular structures. A frequency view that can show the temporal variance of results. For example, what materials are frequently researched by other scientists looking for information about carbon absorption? A neighborhood view that can show the bird's eye view of materials related to absorption. For example, in our case, it is shown that ionic liquids are related to gas absorption. We conducted user studies and design thinking sessions with scientists while we developed this tool. They identified the ability to quickly obtain neighborhood information as a critical enabler in the discovery process as these links may not be evident at the point of the initial query. Using this neighborhood graph, the scientists can further explore the domain information. For example, the scientists may be interested in knowing more about ionic liquids or materials that have high carbon potency such as metal. The user can also find the sources of the information that resulted in the generator knowledge by exploring the source view. As you can see, the findings are grounded based on three publications. The tool can also provide clues to the scientists wherein this information can be located in the source. Once satisfied with the search, the scientists can navigate to the externally hosted source for the complete information. The ability to do so is critical for organizations as the tool can be configured to consume specific data only. For example, sources with specific provenance guarantees. In addition, this tool can be used to bootstrap discovery tasks in any domain. For example, a tool can be used in finance by a market analyst or could be used in drug discovery by a chemist who wants to know how two molecules relate to each other. Now, Rodrigo Norman from our research lab in Rio de Janeiro, Brazil will show you how we can investigate carbon dioxide underground storage with our scientific discovery and simulation tools. Today, I'll be discussing the geological storage of carbon dioxide, which prevents it from being released into the atmosphere. We can store CO2 by injecting it into abandoned oil and gas reservoirs, selling aquifers and other rock formations, or use it in industrial processes that have a net negative carbon footprint, meaning that they store more CO2 than they release. There are several microscopic mechanisms which we can exploit to trap CO2 inside a geological formation. In structural trapping, CO2 is injected below an impergable capped rock layer that prevents it from floating up to the surface and escaping into the atmosphere. Residual trapping occurs when CO2 manages to permeate the rock, but it remains trapped as isolated bubbles. The third mechanism is solubility trapping, and is driven by the fact that when CO2 dissolves in water, it becomes denser than pure water and therefore sinks below the surface. Finally, mineral trapping occurs when CO2 is dissolved in water and reacts with other minerals to form carbonates, which can store CO2 indefinitely. Mineral trapping is the most stable geological storage mechanism for CO2. Our contribution comes in the form of a cloud-based prototype that lets you simulate the interactions between the relevant materials and accelerate the discovery of new CO2 storage strategies. We simulate rocks and other materials down to the level of individual pores, because that is the scale in which the fundamental physical and chemical phenomena take place. We start by taking a rock sample, like this one, and performing an x-ray microtomography to generate a three-dimensional image of the pore space. This shows all the pores, channels and pockets within the rock on a microscopic scale. We then upload this microtomography image to our platform, alongside image properties such as size and resolution. In this case, we are using a carbonate rock sample, which is 100 by 100 by 400 voxels in size, and each voxel corresponds to a four micron cube. Once we load this image onto our platform, we can visually inspect the rock microstructure here represented in grayscale levels. Darker grays represent the pore space, while lighter grays represent a solid rock. You can also use the distribution of grayscale levels to help you pick the algorithm that we will interpret what is pore and what is rock. This is called image segmentation. Once the image is fully segmented into pore, shown in black, and rock, shown in white, we launch an algorithm that extracts a capillary network representation of the pore space. Our tool shows the pore structure, color coded by the capillary diameters, so the purples are narrower channels and the yellows are wider channels. The tool also shows relevant information like the distribution of capillary diameters across a rock, which can be useful for understanding whether this rock is a good candidate for CO2 storage. This connected network of capillaries is the fundamental geometric representation on top of which we deploy the physics simulations. The physics simulation takes as input the properties of the fluid and the driving forces that are used to impose flow. In this case, we are simulating CO2 subject to a pressure equivalent to hundreds of atmospheres. Once the simulations are completed, we can visualize and compare the pressure fields, the flow rates, and flow speeds at each point inside the rock. This allows us to find the most favorable points to inject carbon force sequestration. We have extended this methodology to enable the simulation of two fluids, such as water being pushed by supercritical CO2, as you see in the video. We are also incorporating the effects of chemical reactions that can alter the pore structure to simulate the mineral carbonation process. With that, we will be able to simulate the full range of physical and chemical processes that can enhance CO2 geological storage and accelerate the screening of mineralization additives. This simulation ability will allow users to optimize the CO2 storage efficiency. If you want to test the tool with your own rock data, please reach out to us. Thank you, Matias, Stacy, Giet and Rodrigo. Let's for a few minutes spend time now answering the questions. The first one that came from in the chat was specifically asking, are there efficient ways to boost trapped knowledge in the new domain? Okay, actually, I can take that question now. We are actually working on a couple of potential techniques. I think George is the person who asked the question. There are multiple other ways to do this answer, but we are specifically focused on transfer learning. This idea is that once we have some understanding about a particular domain for the related fields, we can bootstrap this new model generation with less amount of data and especially around annotations. And we are also working on a new technique. Our initial results are looking promising where we can actually use a small amount of labeled data from a given domain to generalize that. So that industries don't really have to spend a lot of time in annotating this data. In scientific and in manufacturing domains, I think one of the big problems is that if you really want good high quality data, you have to spend a lot of time annotating the goal standard. So we are trying to figure out a way to reduce that effort. Our current indications say that it's going to give us good results, but you know, watch out for this forum. We might be able to provide you some good answers in the quarter. I really appreciate it. Another question that just came is how limiting is the geological opportunity globally for injection? So I'm happy to take this one, Anna. Thanks for this question. It's an excellent question because the numbers are indeed mind boggling if you're looking at the amount of CO2 that actually needs to be captured and stored. And it turns out that the geologists, familiar with the matter, they actually analyze the pore space like they are looking through the geological formations and characterize the space at the micro scale. Like at that scale that Rodrigo showed to you in a rock where we have this capillary pore space and turns out that like when we look at it globally that the geologists are telling us that there's in essence enough pore space available to sequester all the CO2 that mankind would like to remove from the atmosphere. So this is a very important finding because that means there is an opportunity to scale, which means if we are getting better in the science and the technology of CO2 capture and sequestration there is a pathway in which we can actually get there. We can capture the CO2 more efficiently with better materials and processes and we can also inject it and store and mineralize it for the long term more safely and less costly. So that's why geological storage and the geological storage potential of the planet are very important. Thank you, Matthias. Thank you, team. Unfortunately, our time is over here. Please kindly stay online and answer additional questions as they come. And now I'll move to the next session. Thank you all. Our next session is focused on carbon performance and it's my delight to introduce the next speaker, Shantanu Gudbole. Dr. Shantanu Gudbole is the team lead and senior manager in the future of climate organization and he leads specifically future of climate and supply chain department in the IBM Research India Lab. He will review with us the concepts of carbon performance and of course the related aspects of climate resiliency or carbon accountability. Again, thank you, Shantanu for joining us. The microphone is yours. Good morning. My name is Shantanu Gudbole and I'm a senior manager and team lead for IBM Research's future of climate initiative. Today, I will talk to you about enterprise carbon performance. Climate change is turning out to have a double whammy effect on enterprises. Enterprises cause climate change primarily due to their greenhouse gas emissions, more generally called just carbon emissions, but they're also affected by climate change when severe extreme weather events happen and global temperature rises. The two main technological strategies for combating climate change are mitigation and adaptation. In our talk today, we will focus on mitigation. You will hear more about adaptation tomorrow. If you look at carbon responsibility and the aim that organizations are setting themselves to decarbonize and reach net zero emissions in say 10 years time or 20 years time based on the complexity of the organization, this has become a movement all over the world. There is increasing consumer, regulatory and investor pressure to act on enterprises now. When I say enterprises, I'm covering built infrastructure, assets and the supply chains that these enterprises are part of and responsible for. Look at some of these announcements from the EU, which is for example, phasing out palm oil from transport fuels. Look at the investment world where big companies like BlackRock are taking voting action against companies which are not doing enough to cut down their emissions and combat climate change. Many companies are also being put on notice and being asked to pull up their socks when it comes to decarbonization. Very big enterprises like Walmart have started Project Gigaton, which aims to eliminate one gigaton of greenhouse gas emissions from their supply chains by 2013. What are some of the main questions that these enterprises across industries are worrying about when it comes to decarbonization? First, they're asking, how do I accurately report my emissions? How do I incentivize my vendors, partners and suppliers to repair their own emissions? And how do I, for example, bridge the carbon level gap between my suppliers and my consumers? Let us break carbon emissions down in terms of the organization's carbon footprint. This is something which is very well understood, but it is worth spending a few seconds on. An organization's carbon footprint can be broken down into what are called three scopes. This is as per the greenhouse gas protocols definition that we follow and we implement throughout the rest of our work. Scope one is the direct emissions that a enterprise is responsible for by assets, vehicles, you actually own by the enterprise. For example, these are fugitive emissions, these are mobile combustion emissions and stationary emissions. Next is scope three. These are indirect emissions of the enterprise, which come from the purchase and use of electricity, steam for heating and cooling purposes, whether in your office buildings or in your factories and process plants, etc. Finally, scope three are the indirect emissions that a company is responsible for, but which occur outside your organization's boundary. These are typically supply chain emissions from your vendors and partners and suppliers, which could be upstream or downstream from you. Scope one and scope two are the two kinds of emissions which are reliably measured and can be inferred and calculated based on primary data, which is typically available in building management systems, ERP systems, etc. Scope three is a very hard beast because all of these emissions are occurring outside your organizational boundary and you typically don't have control over the other players in the supply chain, whether for competitive reasons or otherwise. Nonetheless, there are a few subcategories of scope three emissions where some reliability and some kind of estimation of emissions can be accurately done. At the end of the day, the scope one, scope two and scope three emissions are reported to organizations like the CDP or the carbon disclosure project and pretty much 90% plus of enterprises have been doing this reporting now over the last few years and this is only going to accelerate over time. So how do we go from carbon accounting to carbon performance, which is what we are really here to talk about today. We can look at an enterprise or we can look at the business processes that the enterprise engages in. Let us look at this three layer stack diagram from top to bottom. At the top, you have the enterprise scale, enterprise reporting and these goal settings of carbon emissions, right? Reaching net zero by say 2030 and things like that. One level below are the organizational units where the reporting roll-ups happen and the finance and execution and the strategies are set in those organizational units. At the bottom, you have the base business processes, which is your built infrastructure, your assets. These could be your factories, your machines, your equipments and finally the supply chains that you operate in. These could be transport networks, logistics networks. These could be the flow of inventory and spare parts across the supply chain all the way to end consumer depending on the industry that you may be in. If you look at the same diagram bottom up, you can see that you can start with a process centric view and aggregate it upwards towards your enterprise scope one, scope two and scope three emissions, whether for reporting purposes or to take decisions around carbon optimization, carbon reduction and carbon performance. So as over the next few years, business performance will help companies grow, change in size and shape and footprint. So also accounting and carbon reporting go hand in hand with that, which is what we call carbon performance to say that it's not just important to accurately account for your carbon emissions, but also you need to optimize it, reduce it and eventually be carbon performance in all your applications, your supply chains, your data centers, etc. So the cycle of carbon emissions reporting and performance is that first you gather the data, you collect it from your primary data systems wherever you can, or you estimate it using averages, but this needs to be done transparently. You report this, you benchmark it against your others in the industry, and then you try to optimize it by running programs for optimizing your building energy consumption or purchasing more renewable energy, etc. And then you go back to the loop of gather, report, benchmark and optimize. Let us see an example of this three layer diagram where you can start at an enterprise level or at a business process level and either go up to down or down to up. In this case, let us start at the bottom and let us look at one specific category of an enterprise's emissions, which is for example, say a vehicle free fleet that it owns. So this could be a logistics company or otherwise. So in a vehicle fleet, the primary greenhouse gas emissions will come from the burning of fuel which is used and these vehicles are then used to transport people, equipment, finished goods, etc. So the primary data in this case is going to be the amount of fuel consumed, the distance traveled in miles depending on the data availability, etc. Now, if this is there at a fleet level and you aggregate it upwards, you can then start identifying hotspots and see whether it is maybe the kind of fuel used and this of course varies from country to country. You can, as you aggregate and do it at the enterprise level, you can immediately start identifying these hotspots and those will become your targets for investment where you may choose to purchase cleaner fuel and things like that. At the end, all of it rolls up at the top layer at your enterprise scale. Maybe there's a dashboard or there's a way where you are kind of, this will be your scope one emissions. In some cases, it could be scope three emissions, which is your vehicle fleet and how is that contributing to your overall organizational footprint and do you need to do something about it and all those kind of things. So with this example pattern, this can actually be used in other kind of asset classes. So for example, the retail industry or the supermarket industry, for example in the US has hundreds or maybe thousands of stores and there are refrigerators in the stores which are responsible for fugitive emissions. So the usage of refrigerants, etc. leaks some very potent greenhouse gases into the atmosphere, which are a few hundred to a few thousand times more potent than carbon dioxide in their global warming potential as it is called. So this example of having this enterprise scale view, operational view and business process view with this example of vehicle fleets, you can extend it to say refrigerants. You can extend it to build infrastructure. You can extend it to energy consumption, electricity consumption across say your, how many of our office buildings you have, etc. So this drives insight and action across the organization and you can take this view scope one at a time, scope two at a time, scope three at a time or you can slice and dice it through your asset classes as per this example. Next, let us look at the technical strategy that we have been developing from a carbon performance perspective. Keep the previous example in mind and we'll build up some more interesting technical strategies around that. So we have a few building blocks at the bottom of this slide, which are our carbon accounting APIs. This is table stakes and this is an implementation of the greenhouse gas protocol, which as I mentioned, pretty much is the reporting standard that 90 plus percent of companies over the world use. So the carbon accounting APIs that we have our own custom implementation of contains scope one direct emissions, scope two indirect emissions, scope three indirect emissions, but not all scope three. We limit to the travel and transport kind of categories. We obviously use reference data from IPCC, IEA, EPA, etc. These kind of regulatory government trusted sources. So these carbon accounting APIs are not just a simple calculator. They have a lot of data quality checks and AI built into it, which can figure out missing data and improve the data quality and make the carbon accounting more accurate. And then that becomes a building block which we then feed into the next layer, which is the AI driven data augmentation. Obviously, along with the carbon accounting APIs, another important building block for us is geospatial and weather data, which again we have a large depository and an engine which can ingest it at scale and make it scalably available for various applications. The next layer on top is as I was talking about the data quality enhancement. And in some cases, when we do, you know, GHG downscaling using spatio temporal data, say, you know, remote sensing satellite data, etc. There's a lot of AI driven data augmentation that we do in this second layer. Next comes a differentiated technology layer where we have three parts to it. But today I will talk specifically about emission hotspot identification, which is really anomaly detection problem. And I will talk about building model ensembles for greenhouse gas emissions across space and time resolutions. Let's look at these examples one by one. So if you remember the vehicle fleet example, you can either take similar data or maybe these you can take data about say your various buildings and factories and processing plants that you own as a company, and they all have energy consumption. So if you see the bottom left of this slide, you really have a lot of red dots. Each of them is one building and its energy consumption. This space is actually a multi dimensional space because you are, but just for illustrative purposes, I'm showing a three dimensional view. We have additional data obviously like weather data, geolocation data, maybe in the case of refrigerant assets, it is leakage rate, age, the refrigerant used, you know, the kind of heating and cooling that is going on, etc. And then from there we use some AI algorithms, whether it is MDSS or other anomaly detection algorithms, which will find out anomalous assets, which in this feature space of energy consumption, weather data, codified with some of these energy mixes and the energy intensities of, you know, whether it's a renewable energy source or whether it's say a coal plant, which is supplying energy, all of that then leads to an outlier detection problem or an anomaly detection problem, which then helps you classify these assets. In this case, as I said, we are talking maybe about, you know, refrigerants or building energy consumptions into subgroups, which can help us identify, start identifying anomalies. We then use explainable AI to say that there are some factors which are causing some assets or some asset types or subgroups to behave significantly different from the others. And this using explainability will then get us to specific actionable insights saying that in a particular location X, you may have a building which is significantly responsible for the carbon footprint of your organization, because it is using maybe, you know, coal power plant based energy, or maybe there are assets which are due for repair and they're not performing well, they're leaky. And in some cases, you could be at a regulatory risk because you may have to report this and get those assets fixed within sometimes a period of 30 days and so on. So, using this technology around data quality and hotspot identification, you can identify at the level of assets where you should focus on and what you should concentrate on with respect to your carbon footprint at your scope one and scope two across your organization. Next, let us come to the example of GHG downscaling and model ensembles. Now we know that there are many, many satellites which have been launched for, you know, monitoring from the sky various aspects of what's the activity happening on earth. Once such class of satellites are these GHG sensing satellites, specifically for methane and carbon dioxide, there are already satellites which are covering the entire earth at various spatial and temporal resolutions. And of course, we know that the energy and agriculture industries account for nearly half of global emissions. So obviously enterprises in these spaces would really like scalable methods to account for and figure out the carbon footprints. And then again, similar to the previous slide, find out where are the anomalies, where are the emission hotspots, where do they need to focus investment and attention on etc. So at the top, you see some diagrams around, you know, these level two and level three satellites, satellite information, which we use, which we ingest, and we are turning all of them into carbon emission maps in some sense, which again can be ingested by geospatial platforms, including, you know, what IBM has. At the bottom, you then see that if you look, for example, in the agriculture and food supply chain industry, if you kind of plot the yield of, say, a particular agricultural commodity versus the income, right, you see that there are interesting effects happening in terms of your net revenue, your income from yield, your carbon footprint, and companies need to choose operating points by kind of understanding all of these curves. And this is some very interesting new work that we have undertaken with some large clients sticking to the energy and the agricultural industries. We know that sometimes, obviously the satellite bit remote sensing based approaches are going to be scalable because of course they can cover the entire earth, but they are not going to be as accurate as using primary data and some very specific calculators for example, what is available in the agricultural world. So what we are working on is building an ensemble of static and dynamic approaches for multi-scale greenhouse gas estimation and reporting it as carbon layers in geospatial platforms. Let me tell you what I mean. Across the agricultural sector, there are static approaches using these large alliances of food supply chain companies which have built very accurate calculators around what is happening down to the level of a farm or a plantation across many, many types of crops, all the major crops have been covered. And this requires a lot of data in terms of the irrigation practices, the tilling practices, the fertilizers used, the schedules of fertilization, what was the weather at all those points in time, etc. Now these static approaches require a lot of data, but they are also the industry standard which are used today. But we are also working with some of these dynamic approaches which do one better by actually going to the level of weather and soil related things that change on pretty much a daily basis. The static approaches are kind of calculators which are fed in once a season or once a year, whereas the dynamic approaches really track everything in a very dynamic manner in a continuous manner. In the dynamic approaches also you have these remote sensing based techniques that I talked about previously. What we are building is that across the spectrum of data availability and accuracy, how do you build an ensemble of all of these models to choose operating points which will help you accurately and confidently report your carbon emissions. So as you can see, we are collecting all of this technical strategy along these building blocks, adding a layer of AI-driven data augmentation and building this differentiated technology where I talked about two of them, which is this emission hotspot identification and model GHG model ensembles. The third one will be talked about by my colleague Kedar Kulkarni in the next talk. And then all of this enables a set of use cases, some of which I talked about, like enterprise carbon accounting and reduction, carbon optimized operations, carbon aware inventory optimization and asset carbon performance. So some of these carbon performance applications as we say include supply chain data centers, your cloud data center and IT operations and you will hear about them in some of the subsequent talks. With that, I thank you and I am available for questions and answers and it has been a great pleasure talking to you. Thank you again. Thank you, Shantanu. We have a question here live for one question, time for one question, but please later go back to the Slack. So we had on the Slack discussions, right? How good are we in really accurately measuring carbon footprint, right? And that was a comment, you know, are we ignoring energy use to deal with the increase of entropy regarding the ambition and so forth. And you also mentioned accuracy and estimate problems for scope three emissions, right? Could you just elaborate a little more because that clearly seems to be an important topic. Yeah, thanks, Shantanu. And thanks for whoever asked the question. I think this is a very important topic. So obviously entropy and those kind of topics are, there's a complex interplay between carbon emissions, energy usage and so on. But I think the overall scope three issue is a very big issue. Obviously, that is where the bulk of the emissions of an enterprise lies, right? In some cases, it's even more than 80, 90% of an enterprise is emissions, which are scope three. But the problem of scope three is exactly the fact that as per the definition, it is not in your organization boundary. So it's not in your control. So how will you get accurate reliable data from, you know, your upstream and downstream partners in your supply chain or value chain to kind of, you know, do accurate measurement because the act of measurement as in scope one and scope two is quite, you know, it's quite, I won't say straightforward, but it's a known art. Whereas with scope three, when you don't have data, you tend to rely on, you know, national estimates, averages, and you know, industry averages and so on. And in some cases, like for example, I think it's well known that scope three has 15 categories. Many of those categories have up to, you know, 40, 50% uncertainty. Even with whatever, you know, some of the large enterprises report with respect to, you know, their carbon footprint, you know, to some of these regulatory and, you know, these CDP kind of organizations and so on. So with that, I think it's, I mean, you can do it because, you know, there's not much else, but that's really the research and the scientific problem in terms of, you know, bringing the method to the madness in scope three. And that's definitely an active topic that we are, you know, thinking and kind of researching more. Yeah. So I mean, that's the main problem, you know, the lack of primary data, reliable data, which, you know, you and your organization and your CFO and your CEO will stand behind and certify, right. So that's the issue. Thank you, Shantana. I really appreciate it. I hope that participants are enjoying the session so far. The next in our agenda is a well deserved break. So please grab a glass of water, stretch your legs, and we will see you at the top of the hour to go to the next session that now will focus on applications of the carbon performance work. Again, see you back in a few minutes at the top of the hour. Thank you. Welcome back to the IBM Research Horizons event. Let's get right back into it. Our next session will be focusing on carbon performance implications. And the first session will actually be on supply chain. It is my pleasure to introduce Kedarkul Parani, who will present the session. He's a senior research and a manager in the IBM Research Lab in India. He also leads a future of climate team of IBM researchers in Africa, Kenya. Again, with the focus on supply chain sustainability and resiliency. Please remember to ask questions on the right side of your screen in the chat box so that when Kedark finishes his presentation, he can answer your questions to you. Hi everyone. My name is Kedarkul Parani. I am a senior researcher and the manager of the future of climate team at IBM Research Kenya. The topic of my talk today is carbon performance applications supply chain. Like you saw in one of the earlier sessions, businesses disclose their organizational footprint as scope one, scope two, or scope three emissions. Of these three categories, scope three, which represents emissions due to upstream or downstream business partners in an organization's supply chain is often the largest source of emissions. It can contribute up to about 80% of the overall organizational emissions. And thus, this category can represent the most significant opportunities for emission reduction. However, it must also be mentioned that due to lack of visibility in far off upstream or downstream tiers, scope three emissions are also often very tough to estimate. In spite of that, quite a few carbon responsible companies have taken this challenge up recently. And this trend is increasingly visible across several industry segments. Here are a few examples. Unilever in the consumer goods segment is investing a million euros in a new climate and nature fund. It has also pledged to reach net zero by 2039. Whereas BASF in the chemical segment and L'Oreal in the personal care segment have both committed to carbon label all of their products. IBM's asset management and supply chain offerings are unique in that they have the ability to help customers on their journey to reduce scope three or supply chain emissions in the context of several use cases. The carbon performance framework described in an earlier session with its modeling and calculation capabilities and access to rich emission databases provides APIs that allow computation of emissions under diverse scenarios. These APIs lend themselves readily to infuse carbon calculation and reduction for a wide variety of use cases as outlined here in the middle column. To help customers realize emission savings or reduction, we will focus on the following two repeatable solution scenarios across all use cases. The first one is as is carbon savings. In this case, we quantify how much emissions we are already saving as is without explicitly considering carbon. And the second one is explicit carbon optimization. Here we account for the economic impact of explicitly considering carbon emissions in our decision making. In particular, today we will describe two of these use cases related to the supply chain, order fulfillment and inventory optimization. The question we wish to answer in the context of the two use cases, order fulfillment and inventory optimization is the following. How to minimize carbon emissions such that the impact on the economic costs is minimal. In particular, how smoothly can we navigate the emissions versus costs trade off. Like we shall soon see in detail in the context of order fulfillment, we are interested in figuring out which stores or distribution centers must be chosen to fulfill an incoming order in a carbon aware manner. While in the context of inventory optimization, we would like to figure out how much and when to replenish inventory in a carbon aware manner. Let's start with the first use case, order fulfillment. Order fulfillment concerns optimizing decisions about the last mile delivery of online retail orders in an omnichannel setting. Like the schematic shows here, an online retail order might come in from some destination and the order management system has to decide in real time which fulfillment centers in the retail clients network and which carrier modes must fulfill this order. Current order fulfillment practices typically employ large scale integer linear optimization approaches to minimize the upfront costs which include the shipping labor and processing costs. As well as the predictive item node costs such as stock out markdown avoidance costs etc. However, they do not explicitly account for the carbon emissions generated due to transportation and distribution involved in dispatching order items from chosen nodes through the chosen carrier modes. While these transportation and distribution emissions are directly attributable to the carrier company, for example FedEx, UPS etc. as their scope 1 emissions, they are also indirectly attributable to the retail client as part of their indirect supply chain emissions or scope 3 emissions. Thus, carbon performance order fulfillment would optimally balance the economic costs like it does currently along with these scope 3 transportation and distribution carbon emissions. This slide summarizes the results for the first of the two repeatable solution scenarios as is carbon savings meaning how much emissions does order fulfillment optimization already save while minimizing only the economic costs in comparison with the baseline. In this case, the baseline is a rule based heuristic that recommends optimal decisions only on the basis of the distance between the fulfillment centers and the ordering destinations. Here's a screenshot of the demo created with IBM's Sterling Fulfillment Optimizer that summarizes the as is carbon savings for a major sports retailer over a period of 45 days in 2019. Looking at the summary column in the middle, about 122 tons of carbon dioxide were saved as is, which is equivalent to about 730 trees saved and about 182 NYC2 Paris strips. The graph on the left shows which of the transportation modes registered a major decrease in emissions due to the current fulfillment optimization approaches. As can be seen, a big chunk of emissions was saved due to reduction in air transport shown in green and that ended up increasing the ground transport in purple. However, since air transport is about five times as carbon intensive as ground transport, the reduction in emissions due to avoiding air transport far exceeds the increase in emissions due to increase in ground transport. Thus leading to net net emission reduction. The graphs on the right show a plot of the daily emissions over 45 days for the two approaches, the rigorous fulfillment optimizer in purple and the rule based baseline in green. As can be seen, the purple line is consistently on or below the green line indicating that the current fulfillment optimization approach saves carbon emissions as is on a day to day basis. This slide summarizes our motivation and approach behind the second of the two repeatable solutions scenarios in the context of order fulfillment. Explicit carbon optimization. Like we saw in the previous slide, optimizing the economic costs yields a net net reduction in emissions on a daily basis. However, since orders are fulfilled optimally as they stream in real time, an important question to be addressed is the following. Does economic cost optimization also lead to emission reduction on an order by order basis consistently? Take a look at the graph on the top left that plots the total number of daily orders over a period of 40 days for another major sports retailer. All orders in this graph were fulfilled minimizing the economic costs alone. However, when compared with the baseline, about 75% of the orders shown in gray showed a reduction in emissions to along with economic costs. And for the remaining 25% shown in red, while there was a reduction in costs, there was also an increase in emissions. This is largely due to the phenomenon of package consolidation, where ordered items are consolidated into a smaller number of packages to lower the shipping costs. However, when these packages are sent from further fulfillment centers, they invariably lead to an increase in the emissions as well. The carbon performance fulfillment optimization would have the ability to balance the economic costs against carbon emissions for this class of orders through multi-objective order fulfillment optimization as shown in the picture on the top right. Let's now move on to the next problem class, inventory optimization. Inventory optimization largely concerns optimizing when and how much to replenish inventory such that the incurred costs are as small as possible. As shown in a representative schematic on the left, the inventory position in red decreases as it caters to the demand, and decisions related to when an order must be placed, which is the ROP or the reorder point, and how much to order, which is the ROQ or the reorder quantity, both of these need to be determined optimally. Another question that must be answered is, how often should the ROP and ROQ be updated? Now, most inventory optimization approaches minimize the upfront costs that include replenishment and holding costs and some form of predictive stock out costs, while ensuring that the service level constraints related to the percentage of the satisfied demand are satisfied. However, they do not explicitly account for carbon emissions generated due to transportation and distribution. Like we saw in the case of order fulfillment earlier, these emissions are directly attributable to the carrier company as its direct or scope 1 emissions. However, they are indirectly attributable to the inventory owner as well as indirect or scope 3 supply chain emissions. This slide summarizes the formulation for the first of the two repeatable solution scenarios in the context of inventory optimization, as is carbon savings. Meaning, how much emissions does inventory optimization already save while minimizing only the economic costs in comparison with the baseline? In order to do that, we need to closely examine the following two formulations. The current cost only optimization as shown on the left, and the proposed emission only optimization as shown on the right. The cost only optimization, parameterized by cost usage and lead time parameters, recommends inventory decisions such that costs alone are optimized while ensuring that the service level constraints are satisfied. Similarly, the proposed emissions only optimization recommends inventory decisions such that the ordering, holding and stock out emissions are minimized subject to the same service level constraints. However, along with the cost and usage parameters, it is also parameterized by the following. The transportation mode used for delivery of inventory from supplier locations to the warehouses, the travel distance, the warehouse and inventory item, weight or volume specifications. Taking a closer look at the components of the proposed emission only optimization, ordering and stock out emissions are essentially due to transportation. Regular delivery in the case of ordering emissions and expedited delivery if applicable in the case of stock out. On the other hand, holding emissions are total warehouse emissions allocated to inventory items and depend on the warehouse location and item specifications. Once these components are constructed, evaluating them for the current solution and the baseline provides the necessary information to compute the as is savings as shown in the equation at the bottom. In the last slide, while evaluating the as is carbon savings, we also walked through the emission only optimization formulation. The cost only and emission only optimization formulations form the basis for explicit carbon optimization that balances the tradeoff between costs and emissions. In this slide, I will walk you through some early results obtained using representative industry data. On the top left, we have the numbers for as is savings and on the top right, we have the numbers for explicit carbon optimization. As can be observed in the case of as is savings, the current cost only approach saves both dollar costs and emissions when compared with the baseline. Whereas in the case of explicit carbon optimization, emission only optimization significantly reduces emissions by about 17% but leads to a slight increase in costs by about 3%. The picture on the bottom left summarizes the Pareto or the tradeoff between costs and emissions for different carbon constraints and in general suggests that it should be possible to decrease emissions significantly with a slight increase in the overall costs. With that, I would like to close my talk and will be happy to answer questions if there are any. Thank you. Thank you, Kedar. I just want to one more time encourage participants in asking questions. And just in case you have a challenge, maybe the web browser is not working or something. If you are getting error, please just refresh the web browser. All right. So, Kedar, the first question that came was related to what is the ballpark commission savings we get on a daily basis due to the as is savings. In order fulfillment optimization, could you comment on that? Sure. So based on our preliminary results, we see that on a day to day basis you get an as is emission savings of the order of 10 to 12%. And that's that's pretty significant. Thank you. I have another question. This is more of the intuition. You know, one would think that for the case of the order fulfillment, the smallest distance between the fulfillment center and the destination will give you the best emissions, right? Cost savings. Is that always the true? I just just wanted to say does intuition always apply here or do you find this one that's not the case? That's a very interesting question. Actually, I mean, if you're if you're essentially transporting a single package, then yes, most likely that will almost always be true. However, you know, I mean, depending on whether a particular fulfillment center, when an order comes into a system, right into the system, you need to figure out which of the fulfillment centers can even begin to, you know, start satisfying this order. They may not have every item in the inventory, in which case we need to sort of split the orders like you get, you know, a couple of items from one node and couple of items from the other node, right? So which may like be from the cost perspective, that may not be a great idea. But maybe because these are close enough from the emissions perspective, they are pretty good. However, the other, you know, end of the spectrum is when, you know, you may have a fairly far off fulfillment center, which has all of the items in inventory. And hence it can, you know, sort of send all of these items in the same package. And hence it will also save on the shipping cost. But because it's at a far off distance, it will not be not so good in terms of emissions. So there is this tradeoff between, you know, closer fulfillment centers that may not have all of the inventory items versus far off fulfillment centers. Which may have all of the items in the inventory, but just because they are far off, the emissions are going to be higher. So the issue really is to be able to navigate this tradeoff, you know, going from one end of the spectrum to the other. Where exactly do we want to, you know, stay at? Yeah, that's great. So my following that question, there's another questions. Do your calculations include the opportunity cost of miss sales when out of stock? The person is assuming that an expedited shipment to recover from a stock out to business may choose air transportation if necessary. Absolutely. We do that. In fact, that's a very good point. A lot of times, you know, stock outs, again, depending on the as depending on the kind of goods that we're talking about may incur significant costs, right? Particularly, there might be certain items that are so critical that, you know, if they we just don't have it, then the entire plant might go down, right? So it's necessary to have some of those on standby. And yes, I mean, if you just don't have it, you just need to make sure that you have an expedited delivery, in which case, you will end up using a very fast model. Transport, typically air transport. And like I said in the presentation, air transport is extremely carbon intensive. And yes, we do account for that in our analysis. But Shantanu and you commented how difficult it is to get the information. And so we had the question, can we provide stepwise information of carbon emissions and CO2 saving and then cost reduced throughout the entire supply chain process? Can you comment how easy, how hard that is? Yeah, that's a, if I may say so, I mean, the in terms of like the supply chain has been around for quite some time. So the holy grail of, you know, the supply chain area is, you know, the so-called coordinated supply chain where each and every node in the supply chain graph sort of talks to each other in such an ideal scenario. Yes, you know, you can always have whatever information you want. However, in reality, that's not how things work. And hence it depends a lot on the kind of cooperation that different nodes in the supply chain have with each other. A lot of times, you know, any node in the supply chain will typically have visibility to, you know, maybe tier one above tier one down at the most tier two above tier two down. And to that extent, you know, the scope three calculations will be limited. So yeah, I mean, ideally, if you have information for all of the nodes in the supply chain, then of course you can, it should be possible to generate accurate estimates on scope three emissions. If we don't have that, then we'll have to, you know, use approximations and, you know, in the process be transparent about the resulting accuracy. Thank you again. Thank you for discussing this topic with us. And now we'll move to the next session. We just discussed the applications of carbon performance session. That was the supply chain. Now we're moving to applications of carbon performance related to hybrid cloud. And continuing with that topic, Tamar Ilyam will be focused on this hybrid cloud discussion. Tamar is an IBM fellow working with the IBM Research at TJ Watson Research Center in New York, New York Heights, New York. And she leads the worldwide technology strategy in the area of cloud native environments and DevOps. For those of you who don't know the IBM nomenclature, IBM fellow is the highest technical position in IBM Research. So it's really my pleasure to introduce Tamar and Tamar. The microphone is yours. Hi everyone. I'm going to give this presentation together with Ali Miki from Bivouac BNP Pariba Transformation Hub. Let's start by talking about what is a carbon performance hybrid cloud. Let's define it. The way we define it is the ability to measure, quantify and ultimately reduce per tenant carbon footprint at every layer of the cloud stack. In and across data centers. Why now? Why is it important for us to pay attention to this problem right now? Well, I believe we're at an inflection point. The demand is growing for internet services at an exponential scale. In addition, we're seeing the emergence of energy demanding workloads such as AI. With AI, first of all, the energy which AI training jobs consume doubles every three to four months. And some training jobs take the equivalent of five car carbon footprint throughout their lifetime. We're also at the end of Dennard scaling and what it means is that we cannot keep up. We cannot expect that we can get from the same cheap better performance with reduced energy. So we need to find new innovative ways to reduce the energy that is consumed by the computation. If we look at market trends, data center carbon footprint is the new battleground for cloud providers according to IDC. CIOs lead sustainability projects in more than 50% of companies, which basically means that a lot of companies, many companies choose to start with IT in terms of the sustainability strategy to reduce carbon footprint starting with IT. Saving carbon footprint is a key motivation to move to cloud. And once customers move to cloud or at least plan to move to cloud, they want to know how much they're going to be saving in migrating to cloud, how much carbon footprint. And once they're in the cloud, what's the carbon footprint of their workload in the cloud? On the positive side, ICT is also the largest purchaser of renewable energies from across all the sectors. All right, so let's move to a little bit of a terminology. How do we calculate what the carbon footprint, which is associated with a data center? We basically need to multiply three terms. One is the IT equipment energy, which is simply the energy that is going into powering the servers, the routers that participate in the actual computation. The second term power usage effectiveness is the overhead, which is the energy that is going to everything else in the data center, including cooling, lighting and power conversions and distribution. And the last component is the carbon intensity, which looks at the origin of the energy, whether it comes from a renewable energy generation or from more dirty sources such as coal. Obviously, the intensity of coal is way, way more larger than the carbon intensity of renewable energy sources such as wind and solar. So any strategy to address the data center carbon footprint has to actually go and address each one of these components and try to reduce them. All right, so any optimization requires quantification. We cannot optimize anything. We cannot work on reducing the carbon footprint without quantifying it first. And the method of quantification will depend on the goal that you're trying to achieve. If you're trying to optimize or if you're just trying to report, make policy, set goals or make purchasing decisions. One thing is clear. Today, none of the cloud providers are providing a transparent method for calculating the carbon footprint of the tenants, the workloads of the customers that are running on the cloud. This conundrum did not go unnoticed. This is an article from Nature that talks about hiding greenhouse gas emission in the cloud. And basically what it's saying is that the cloud provider is responsible for either making the data available to their customers or calculating for the customers, what's the carbon footprint of their workloads. Our point of view is that energy must be considered as a first-class resource that can be traced, monitored, as well as accounted and optimized for. This is why we decided to work on the cloud tenant carbon accounting. Our method factors in cloud characteristics, such as multiple tenants, multiple services, and resources that are shared across tenants. We calculate the energy and the carbon footprint for all of these elements. The engine is programmable and dynamic and extensible, and it complies with the GHG protocol guidance document. Before I pass it to Ali, let me talk a little bit about the principles of the GHG protocol. It can be summarized in one sentence, and the sentences were all on the same boat. This is how I'm thinking about it. We're all in the same boat. There is energy that is going into the data center. According to the GHG protocol, that energy must be completely split across all the tenants. That includes the overhead. That includes all of the devices. That includes the equipment which is used in order to manage the workload, such as billing equipment, scheduling, monitoring. That includes reserve capacity. The method also must factor in what is a fixed emission and what is a dynamic emission. Let me explain a little bit what this means. Fixed emission is based on the allocation, so it's based on size. It includes the reserve capacity, for example. Dynamic emission is based on the actual use, so it depends on the load that the tenant is getting from the users. The method must be specialized based on the characteristics of the service. The method to allocate or attribute a carbon footprint or energy to tenants of a transaction service is going to be different than the method to do that for a storage service or a data service or for a service that provisions VMs in an IS environment. Our method is general enough to cover all of these elements and to do that in a way that is fair and complete. Right now, the implementation only supports bare metal and VMs, including the case where VMs of different tenants are running on the same server. In the future, we're going to extend it to other services. The code, the engine is available on our cloud innovation lab. As I said, our first customer is the Bivouac team in BNP Paribas. With that, let me pass it to Ali, who will talk with you about how they're consuming the APIs and what they're doing with them. Thank you, Tamar. Hello, everyone. Bivouac is a BNP Paribas transformation hub. We have four main pillars that are project extension, individual upskilling, collective performance and real-life and digital base camp. Some main goal is to be able to create, develop and deliver new services and new business models. Bivouac has been committed to a responsible digital approach and obtained the numeric responsible level. As shown here, this implies a lot of different tracks. One of them means to be able to provide reliable data about our impact and define best practices to use our resource on the cloud and build applications that run in the cloud. So, by obtaining this label, we started to structure our approach. And here are some first action we took, like sharing experience and tool with various players within the group, BNPPs, entities or functions like ITG. We have a lot of experience inside BNP, a lot of expertise and a lot of different initiatives. So we need to learn from all the existing experience that we have within BNP Paribas. We also organized a digital king week to raise awareness among all Bivouac residents. So it took the form of conference gathering more than 400 participants and daily challenge, very simple actions that we could take that allowed to learn good digital practices and reduce our collective footprint. We are also developing a proof of concept called Blue Tracker to give a score of the environmental impact of the product developed and run on the cloud. This core could be integrated in a wider impact score that could be considered when planning a project as we do with budget or a return on investment. So the proof of concept, Green Tracker, is about providing insights on cloud resource usage and environmental footprints to be able to give advice to product teams on best eco-responsible practice during the different steps of the lifecycle of the product. So what we offer today is part of the Green Tracker proof of concept. We provide monitoring metrics on the Kubernetes clusters that we initiated for the product, from development environment to production environment. So IBM's team in GCCAT help us use CCD KPIs to gather this data and expose them to the teams. Teams can still access detailed dashboard directly on the IBM cloud console if they need to have more details, but by making them accessible to not only technical team members. We want to further practices such as high sizing of the resource we deploy in the cloud and be able to integrate them directly in the backlog of the product. We also provide the product impact in terms of electricity and consumption of electricity consumption and carbon footprints. So we spent some time trying to figure out how we could provide those estimates. And that's when we made aware that IBM research could provide such a tool. So IBM provided us an API that we can use to get the environmental metrics we need. For now we focus on Kubernetes consumptions and we would like to extend this metrics to any type of resource we concern in the cloud. All this metrics allow us to give a green score to our product ranging from A to F and that is calculated relatively to similar products. So the next step for a green tracker is to first evaluate the relevance of our approach by testing green tracker with team within Pivac. Then we want to go from a proof of concept phase to the MVP phase by proposing a complete user job. We of course want to integrate more advice and more relevant data. And for that we need to continue our collaboration with IBM research. Thank you. Before we conclude I'd like to touch on the subject of optimization. One of the main goals why we want to quantify the carbon footprint is so that we can optimize and work to reduce it. We're working on a dynamic set of controllers that basically operate at each layer of the stack in order to minimize each one of the components that go into the calculation of the carbon footprint. The energy for IT, the overhead power usage effectiveness and make the best use of carbon intensity or renewable energy. In the first layer we look at VM placements and power management at the same time. So where are we placing the VMs on what servers and how can we power down machines that are not being used. In the second layer we're focusing on our hybrid cloud platform OpenShift and there we're putting optimization that has to do with scheduling the containers, placing them, and also right sizing the containers, determining the right size, the optimal size for each one of the containers in a dynamic fashion. And in the third layer we're doing dynamic dispatching based on the available renewable energy. That's actually a fascinating area. As you know renewable energy is dynamic. It comes and goes together with the sun and wind. And what we're employing here is prediction to predict how much energy we're going to get in different times of the day. And then based on that dispatch jobs such that essentially you're running more jobs when you have more renewable energy. So the cloud sort of expands and contracts based on the amount of renewable energy that you have at each moment. This is just like scratching the surface. But if you're interested to get more of a deep dive, please contact me and I'll be happy to have a follow-up discussions with you. Thank you. Thank you Tamar and Eli for your presentation. Let's get to questions. Tamar, I saw you seen the first question. You want to go directly and answer it? Yes. So the question is, can you hear me first of all? Yes. Yes, we can. The question is whether developers can reduce the question from Jen Elwood. Whether developers can reduce the carbon footprint of the code. So that's a very interesting question. I think that there are opportunities to reduce the carbon footprint at every stage of the software lifecycle, starting with development. You look at developers, obviously using practices such as being aware of what algorithm you use, the efficiency of the algorithm. What lighter is your use and techniques such as caching the data, for example, in order to avoid transfer of data, which is associated with a high carbon footprint. Go to packaging. Packaging the code as containers or as microservices. Why? Because they're more lightweight and then more containers can be condensed on fewer machines. You reduce the carbon footprint. Going into deployment. Are you putting the data close to the compute? If you do that, you reduce again the data transfer carbon footprint. And last, I think the biggest opportunity here, the new emerging programming models on the cloud. Look at serverless computing. When it comes to serverless computing, you're only running what you need, when you need it. And because of economy of scale, where all of the workloads of all the clients are on the same platform, you're pushing the complexity down to the cloud platform to manage it for you. And because they're dealing with economy of scale, they can do much more in order to optimize the carbon footprint of all tenants together. So a long answer to a short, interesting question. No, that's great. Let's ask the second one. Is cloud good or bad for the environment and why? Yeah, so I think I just answered the second one, which is I think that cloud is good for the environment, for all these reasons. There are multiple reasons. I mentioned the serverless and the emerging programming models in the cloud that are focused on running what you need, when you need it only, saving all this idle power. Because you know, idle power of a machine that is not running, not doing job can be up to 40% or 50% than the maximum of that server. That's huge. So saving on that, all of the automation that comes with the cloud, the auto scaling, all of that. There is a huge potential to do even more, which is the type of controllers that I described with scheduling and so on. So again, rather than focus on the developer to do that work, shift the complexity to the cloud provider to build in optimization and to leverage economy of scale. Another question was you started already quantifying some of the numbers, because the question is really how much can be saved using those techniques? Do you have some assessments of what you've seen so far? Yes, I believe according to, so it depends, you know, it depends on the use case, when it comes to dispatching based on renewable energy, the gain could be huge. Let's say if you are very flexible in when you can, you want to run your job and you're waiting until you have the maximum renewable energy versus the minimum renewable energy, the saving could be used because maximum renewable energy means almost zero carbon footprint. So the saving can be used when it comes to scheduling dispatching out of that, I would assess that, you know, 30 to 40%. Again, it really depends on the context on where you're starting from and what techniques are you using, powering down machines and so on. There was another question, do we have any client example of where we would have implemented green truck and ensured the calculation of carbon footprint optimization metrics? Yes, so we're working right now with BNP Paribas, and that's my colleague Ali. And they're making use of these APIs, it's a work in progress. It was very late for Ali, so he didn't join to answer questions. He's in Europe, and we're early on, but the idea for them, and that's why it was so important for me to show that end-to-end use case, is that they're consuming the APIs that we provide, the cloud provider provides in order to actually detect hotspots from their side. For example, cases where there is a container or an application that consumes a lot of energy, but serving very few users. That's a hotspot, basically, and that's exactly what they intend to do with the technology. Tamar, thank you again. Really appreciate you sharing with us your know-how and your expertise. We'll move to the next session. Thank you. And our next session is an exciting panel discussion, and our moderator for this session will be Bruce Almagry. He is a principal research staff member at IBM Watson Research Center in New York, and he co-leads the material discovery program for the future of climate initiative. Bruce, turning microphone to you. Thank you, Anna. I'd like to start by introducing our three panelists. Eric Falco Silva is research manager in the SINTEF industry with responsibility for solvent amine-based CO2 capture. He holds a PhD in the field of CO2 capture from the Norwegian University of Science and Technology. His technical background is in physical chemistry and chemical engineering. Eric has held positions as senior research scientist at SINTEF, program manager at GAS, treating R&D at Shell, and senior CO2 scientist at TOE. Richard Andercheck is a senior vice president at STANTEC, a top-tier architectural, engineering, and environmental design consulting firm. Rich leads environmental growth initiatives with a focus on innovation, incorporating fourth industrial revolution concepts. Rich and his colleagues have been developing and delivering integrated solutions to mitigate greenhouse gas emission and adapt to a warming world. His academic training is in mechanical engineering and his professional career as an engineer and consultant in the environmental sector spans nearly 40 years. Brian Yatko is the chief engineer for sustainability and future mobility at Boeing, leading a team focused on electrification, energy storage, aircraft design, and autonomy. Brian also sits on the board of directors for Boeing's Electric and Autonomous Aircraft Joint Venture, WISC. Brian has a PhD in aeronautics and astronautics from MIT and a Bachelor of Science in Aerospace Engineering from Penn State University. Welcome, Eric. Rich, Brian. So let's begin. Eric, your company, Sintef, works with clients to help them develop carbon capture and mitigation strategies. Tell us something about what you do and what your clients are asking for. Yeah, so Sintef is a fairly broad research institute and we do lots of things related to hydrogen and other energy technologies, but CCS is also a big topic and we have a broad range of clients these days. And you notice that there's a lot of globally, there's a very large interest in addressing these issues now. And so we have, I would maybe put, divide our clients into three classes. There's the technology vendors, all the companies that hope to provide the technology elements to contribute to these solutions, the CO2 capture units, the transport systems that ships to transport CO2, and the storage of it in geological formations. So you have the vendors, then there's, I would call them the informed end users, so companies that have CO2 emissions but also want to have substantial knowledge about the technology themselves and be involved in it. And that's typically like traditional oil and gas energy majors, or maybe just energy majors is what they would call themselves these days. And finally there's lots of, we have a lot of interactions in Europe. So you see lots of cement industry, steel industry, all types of industry with large CO2 emissions, and they simply have a problem. And they want a solution. And sometimes they will go directly to some technology vendors or sometimes they will approach entity like Sintef and please help us understand our problem and what is our best way out of our situation. What's the most economical way to get our CO2 emissions done? Thanks Eric. Rich at Stantec, you work with diverse international clients, environmental engineering. Can you explain what that is all about? Sure, thanks Bruce. And it's a pleasure to be here today with everyone. We should have some charts. I do, but there we go, there we go. Thank you. Thanks. So I guess we'll start with just our promise whenever we work on projects and sites is that we always design with community in mind. And that community can be the human community can also be the environmental community. So that's just our promise of how we approach projects. If you can go on to the next slide. And we have different businesses in the company. And the issue of mitigation, GHG emissions mitigation needs to be thought of in a comprehensive approach. So our disciplines beyond environmental including architecture, other sciences and engineering thinks about how we can mitigate GHG emissions and all the work that we do in buildings and the energy and resources sector. Our environmental services team supports all of the other sectors looking at permitting, compliance and nature based solutions. And we also support our infrastructure business. This includes transportation with zero emission vehicles, community development and our water teams that deal with all aspects of the hydrologic cycle. Next slide. We are a global firm as you mentioned, Bruce, and our client based varies around the world. We have offices from in the southern hemisphere in New Zealand and Australia in Africa and South America in the northern hemisphere of China, India, Middle East and Europe, UK, United States and Canada and then the Caribbean. And we offer these mitigation services for the full breadth of our clients that include municipalities, governments, private industry those traded on the stock market. And we deliver the greenhouse gas emissions services for a broad range of those clients depending upon where we are. The last slide, please. And I'll just give you one example of some of the work we do on the environmental side. So we work alongside of our colleagues with citing transmission routes, if you will. And in this case, it's showing the construction and placement of a tower in a difficult terrain. So we work with them on citing them and then during construction making sure that the environment is protected, species are maintained and then post construction monitoring restoration activities. So that's a little bit about work that we do at StanTech. Thanks, Rich. And now we have another set of charts. Brian, the airline industry is essential to our way of life, but it faces unique challenges as the world moves toward net zero. How does Boeing plan to make this move? Yeah, thanks, Bruce. And thanks for having me here. If you can go to the next slide, please. Yeah, I guess I'll start by just giving a bit of context, which is every year, four and a half billion people travel on airplanes around the world, four and a half billion passenger and plane mints, and about $7 trillion of goods are transported in their transportation system. And as we look forward and try to think about how to decarbonize that system while maintaining the same or increasing levels of benefits to humanity that come from those kinds of connections and economic activities, we have laid out a four part strategy, which I'll kind of walk through here and then talk about some vignettes on the future and how we're seeing this scenario play out and where in particular technology will be used to solve some of the most interesting problems. So just to give you a sense of context, right now the global air transportation system contributes about 2% of manmade CO2 emissions, but it does so at altitudes for which the impacts may be larger than 2% that's represented in just the carbon accounting. And so as we look at how to affect those emissions, we kind of think about that in terms of four primary pillars. The first is fleet renewal. So right now the fleet is full of aircraft that have varying levels of technology. If we were able to bring the fleet up to the same level of technology that currently exists and is in certified aircraft in the latest generation of products from companies like Boeing and Airbus, we would be able to save about one third of the total fuel in the system. So what's interesting is there are technologies that exist to save roughly one third of the total system fuel. The second pillar is operational efficiency. And so this relates to the technologies that allow us to fly aircraft more efficiently in the airspace. So we account that about 12% of total fuel burn could be saved if we were able to fly more direct routings, more optimal speed and altitude. And so this is a combination of having more automated systems but also having kind of more coordination within the system. The third and perhaps largest pillar is actually a transition to renewable energy both off the airplane and then what's ultimately used on the airplanes. And so here we're thinking about how to increase the use of what in our industry is called sustainable aviation fuels, which are fuels that are derived from existing sources of carbon within the environment, non-fossil-based fuels. So things like biofuels, electrofuels, synthetic fuels, synthetic hydrocarbon fuels, etc. And then the fourth pillar relates to some of the more advanced technology where I spend a lot of time, things like electric aircraft, autonomous aircraft, aircraft that use new propulsion systems and new configurations and ultimately new fuels. If you go to the next slide, I just want to briefly kind of show sort of a vignette of what the future of air transportation might look like in airports a few decades from now. So you see there is an aircraft that likely has a new configuration here. We're showing what we call the transonic thrust brace wing, which is an aircraft configuration that allows for increase in fuel efficiency at the aircraft level for a variety of technological reasons. But what's interesting is both the notion of how the fuel pathways that get to that airplane will work. So I mentioned the sort of biobase fuels which are represented in this, but also some of the autonomous technologies that will find their way both onto the airport infrastructure itself. So airports themselves are actually relatively, you know, have emissions themselves, carbon emissions themselves. And so the ability to have electric autonomous ground vehicles, small electric and autonomous aircraft that are used for passenger delivery around the airport environment, those are things that we think that we'll see in the generations ahead. And then maybe the last slide. This shows a vignette of the future of not a traditional airport of the type that we might be operating in today. But the kinds of environments that may be in urban scenes in the future. And so one thing that we're seeing in our domain is that the ability of electric flight to create aircraft that can be tailored to fly in or near urban environments in highly precise ways to allow for operating within infrastructure that can be in relatively dense urban environments. And so we have a joint venture called WISC, which is building one of the aircraft, actually the little yellow airplane that you see in the background there. But one of the things that may be interesting about the decades ahead is the incredible amount of new activity and new flying machines that will enter into these markets due to the enablement of autonomy and electrification. So I'll stop there and really looking forward to talking with you all about these. Thanks, Brian. That's a fabulous view of the future from all of you. These are really essential steps. I'm wondering a question for all of you as you've been discussing your current activities and goals. We know that many companies are promising net zero by 2050. It's only 29 years away. Where do you think we'll actually be in 10 years? And what needs to change to make this possible in 29 years? Eric, I see you have a hand up if you want to take a stab at this. Yeah, someone did slightly scary calculation last year. Because of COVID, CO2 emissions went down to some extent last year. And if I remember the calculation roughly, you would need to have an additional COVID effect every single year for the next 10 years. Not necessarily by those means, of course, but that kind of gives a hint of the scale. And also having been in the oil and gas industry, typically from your starter project, like imagine you want in a completely new place in North America or Europe or New Orleans in the world, you want to build a new CCS project from scratch. You do have to do all the feed studies, pre-feeds, go into construction phase until you're finally into operation. And that's easily seven, eight, 10 years. So for us to achieve these by 2030, it will require an effort on the monumental scale, I think. And starting now. Starting today, yes. And we don't want to shut the world down again. Right, right, right. In Eric's right, the magnitude of the effort needed to get to net zero by 2050 is huge. And remember that we're along the pathway and this is really a rate problem, which we know are the most complex problems we face. And the rate of change that we need to implement to electrify our systems and get more of the energy that we consume and use from renewables needs to accelerate rapidly. The National Academies has done a nice study that was published earlier this year that really lays out a nice pathway for at least the U.S. to decarbonize our economy. It requires huge investment by the community, by us as society, and a focus on our efforts to really transition away from combusting fossil fuels into other renewable resources. Yeah, major. So what do you see as the major roadblocks along the way? We think of scale, and it's hard to imagine gigatons of CO2, but then it's hard to imagine carbon capture places all over the world. What are the roadblocks? Are they technological? Are they science? Are they government? Yeah, Eric. Yeah, I would say that it's all about incentives. The private market moves when there's sufficient incentives in place. And for something like CCS, incentives have not been there. If the incentives, and for example in some countries they are talking about significantly higher CO2 taxes, and at some point this would simply become an attractive option as opposed to paying taxes. So if one society, if and when society is willing to commit to a certain level of saying you have to do this or pay this much money, then things could potentially move much more quickly. So we know how to do it. And CCS does have some, you have some specific issues with geological stories, so it's not everywhere in the world where it's as easy, but there's plenty of places where you could do it. So I do think it's more the resources and being willing, someone being willing to allocate the needed resources. The other thing I'll mention is that, you know, if you look at the level of effort that just was talked about this morning and the degree to which people are thinking about reducing their carbon footprint, society seems to be moving in that direction. And when we talk about this at the global scale, I think there's a lot to be learned from our European colleagues and those in the UK who have really been working on this topic a couple of decades ahead of most of us in North America. So I think there's an opportunity to apply lessons learned and see what we can do and bring across the pond, if you will, to expedite our transition here in North America. So we haven't mentioned the workforce. Well, the workforce that we need to make this change in the next generation of engineers and scientists, will they be available? Will they be trained? Will they be ready for the task? It is one generation, 29 years from now. I'll give my views. I think the academic training is in place and that those that are coming out of university are well trained to tackle these problems. I think also that the focus on using multi-disciplines to help get to solutions is beneficial. What's not clear to me, though, is the number of people who will be needed to really support this transition. Again, the magnitude of the effort required is large. I believe that there's hope in digitalization to help expedite activities that maybe can be backfilled from a human labor perspective. So our younger generations are very focused on this topic of climate change and extremely motivated to help support this transition to a more sustainable world. I'm very optimistic about this point. I think that these are incredibly meaningful problems and I think that these kind of problems are very motivating for people to work on. They're incredibly cross-disciplinary. So I think as it relates to education, especially in the aerospace domain, everything is a constrained optimization problem and the more knowledge that you have about the other domains, the better solution ultimately that we'll get to. I think not only are these meaningful problems, but they're interesting intellectual problems. I don't suspect that we'll have trouble motivating people towards these goals. I do think that this is a long-term... This is basically the challenge of this generation and I think that we'll see education systems evolve towards ultimately some of the needs that we may need as they become exposed in various industries. I'm not pessimistic about that at all. I'm extremely optimistic that the systems will work as they're designed. I'll support this. I also think actually the existing workforce is to a large extent suited for these challenges. It's a lot of these technologies that we're talking about, both with renewables and CCS and hydrogen and so on. It's engineering skills and computational skills and so on that are not too different from what are applied in industries today. So I think there's a future workforce and I think even there are plenty of people in the current workforce that will be happy to make some transitions. That's good to hear. So you mentioned a few market sectors. You mentioned aviation. We saw this beautiful shot of a stand-tech putting in a transmission tower in a formidable place. What market sectors do you think will have the hardest time improving carbon performance? Well, I guess I'll... I think it's fair to say that if you sort of categorize things as, quote, hard to abate, I would view aviation in that way. And now I would say that... And so I spend a lot of my time grappling with the idea of, okay, well, what does that mean? Why is it hard? And so I think anywhere where the physics are very difficult and already extremely optimized in order to perform the task at hand, I think if you look back at the dawn of the jet age, fuel efficiency has improved almost 80% since the first commercial jet. So there has been just... Because of the economic forces in the industry, there has been this incredible many, many decade-long journey of developing technologies that improve fuel efficiency as the primary competitive differentiation of the products. So I think anywhere that you kind of have this existing pressure to be extremely fuel efficient and the physics dictate that it's sort of extremely difficult to change some primary elements of the system, I think you'll kind of end up in these scenarios. Yeah, aviation does look difficult. Eric? Yeah, if I can comment. So I agree that aviation is one of the most difficult ones to abate, but I think there's other industries that are equally optimized, but I think the distinction is that it's hard to put the battery on the airplane. It is more constrained in terms of going long distances and relying on the energies carrying with it. So I think it's some inherent features there too. And also putting a CO2 capture plant on the airplane is also not an option. Yeah, and I guess I would add too that some of the industries that use the other greenhouse gases that are non-carbon dioxide-based, methane and nitrous oxide and hydrochloro-fluorocarbons they're more problematic I think to minimize than say carbon dioxide. So there's probably a way to deal with that by reducing CO2 in the atmosphere, but nevertheless those industries that use those or emit those compounds are going to have a little bit more of a difficult time. The other one I would mention and only because I saw something in the news the other day was that I think and maybe one of the Nordic countries, maybe Eric you remember, but just the other day there was making steel without any source of fossil fuels and that was the first of its kind. And in China our teams are doing a study around the sustainable steel industry. So those energy-intense industries are one that's going to be more difficult, but again there's positive progress being made and lots of effort being put everywhere around the world to help resolve these problems. So the title of this panel is Carbon Performance as a Business Imperative. And that implies necessary for business. Why is this necessary for business? How do you view this? Is it necessary for business? Companies that you work with, are they compelled? What is pushing this? Yeah Eric. Yeah, not naming specific companies, but up to a certain point I think companies were in a state of if we have to pay taxes or if regulation changed we should know enough about these technologies to be ready. So it's all about when are the taxes so high that it's going to hurt or when are there going to be specific government rulings against our industry. And in Europe you are looking at we have this European trading scheme for CO2 emissions and I think if I remember correctly it's hovering around 60 euros per ton CO2. And that is reaching levels where you really start doing your math at least. So I'll also add that in many ways our workforce is interested in how we as a company are doing around this topic. And it's important for us to make sure that we're minimizing our emissions and making commitments that allows us to attract and retain the best and brightest in the industry to help work for us. So that's another driving force. And it's a very positive thing for us to be able to make carbon neutrality commitments by the end of 2022 as an example and being one of the drivers for taking action. I second the point there on the workforce and also just to mention another dimension is consumers are demanding it. I mean I think there's a great public realization of the impacts that we're all having on the environment from various industries. And so I think that not just workforce and taxes and some of the other things but I think our customers and consumers will ultimately expect that companies have a plan in place to mitigate and ultimately eliminate their impact on the environment. Yeah I think we're seeing that too. So now maybe a harder question. When I think of businesses I think of profit and growth which makes me wonder if net zero by 2050 is economically feasible. So will the public pay more for a product that has a carbon neutral life cycle? Or for example can any one company afford to be the first to net zero if that means their product is more expensive than their competitors. Age old product. I guess I would say Bruce that I'm not sure that getting to net zero is inherently more costly and that if done well and planned and executed properly then achieving net zero may be more cost effective. There's more energy coming into the system all the time from renewable resources and there are ways to become more sustainable and more efficient within a company's own performance that ultimately reduces costs. So I don't know that I make the concluding link that getting to net zero is inherently more expensive. That's a good point of view. Eric, I saw your hand up. I'm used to the idea of carbon capture which of course could cost a lot. Eric, what do you think? Yeah, carbon capture is inherent. It's like waste disposal. You pay for it. So and I will, here I will disagree. I mean, there's no doubt that the cost of renewable power has come down dramatically and that's really good development and it's impressive. But I mean, the reality is that the world so far, if you look the last 100 years, has not been moving in the right direction. So that technology by itself will sort of make this cheap enough to achieve this goal without any societal effort. I'm very skeptical of that idea. And frankly, I think it's an idea we should be careful about because the risk is that our technologies would not come down in price quickly enough and the airway are not having met those targets in 2050. Yeah, so do you think we need government regulation? So all companies are protected by the cost in the same way? I see no country that... Well, the US did, for some years, the US CO2 emissions did go down because of transition from coal to natural gas. So things like that can help. But ultimately, for example, in Europe, there has been heavy subsidies of renewable energy and at some point those subsidies have become smaller. But still, I do think strong government intervention is going to be part of the story. I'll just add though that we seem to be willing to pay for the effects of a warming world in the context of flooding. Look at Hurricane Ida. My community was ravaged by wildfire. I'm here in California with tons of wildfires and luckily my home was spared. But my neighbors are lost without property yet we're willing to incur those costs and accept it without thinking about what if we were able to refocus and work toward mitigating GHG emissions and taking the money we can't take it from. But if we allow it to continue to worsen, the impacts are just going to worsen as well. More communities will be ravaged and it will become a more difficult place for us to live here on Earth. Yeah, that's a good point. And Brian, I know you are bowing our concern about cost too. Will all of these new technologies and time transition, will they increase travel costs? Do you think is there an economic hit for this change as well? Well, one of the things that I think is interesting in terms of those pathways that I laid out is with respect to sustainable fuels, we know and have available technologies that can produce sustainable fuels but they're not yet economical to scale. So one of the things that's going on in the industry is a real drive towards scaling sustainable fuels such that they contribute to a meaningful amount of fuel that's put on board the airplane. And in our industry there has been relatively significant action both in the United States with positive incentives, things like tax credits for developing and using sustainable fuels and in Europe mandates for developing and using sustainable fuels such that it looks like the sustainable fuels industry will have to roughly 100x by 2030. So that's a significant, if you think about scaling an industrial process, in order to meet the commitments that are in place, there's 100x that has to go into that industry. And what's interesting is it looks like the economic conditions may line up such that those investments will happen and deliver on that promise. And then from there the industry has to 10x another times further by 2050 to meet 100%. So I think that's kind of interesting on the fuel side. And then with respect to technology, everything in aviation is long lead. It's about 10 years for any aircraft development program from the time you start to the time that you finish and aircraft in the industry once they're certified last about 30 years. So airplanes that are produced today in 2020 will fly their last flight sometime in the 2040s or 2050. And so if you think about long lead technology and the ability to kind of get something into the fleet in 2030 such that it can make an impact over those next three decades, there's a huge amount of R&D costs that goes into that, but that's largely kind of part of the business models of the aerospace industry. Thanks. So Eric, a question for you. Maybe these questions are getting harder and harder. So we read in the IPCC report that direct air capture we needed to remove many gigatons of CO2 from the air. What do you see as the prospects for direct air capture and what will make it feasible on a global scale? Yeah, so I mean the math of the IPCC is very simple math that if we're going to reach these stay within these warming targets then at some point we're overheating in terms of CO2 emissions so the simple logical consequence of that is that we will have to to remove CO2 from the atmosphere. And they just say that's a simple mass balance kind of thing, it's very simple analysis. And then you start to think about that and what will this cost and I will say one thing, one of the important point here because that is getting a lot of attention and that is by nature always going to be more expensive than capturing CO2 from point sources. That's simple thermodynamics and engineering because it's so much more dilute in the atmosphere. So to start putting lots of, to do large scale back in the world at the same time as you're doing lots still lots of large point source emissions doesn't really make sense. So large scale deployment of that would only start to make sense when the world have spent a lot of resources in getting its emissions close to zero. So this is some decades into the future in my view. And that requires energy, all CCS requires energy and that even more so that's also inescapable and this needs to be green energy but this makes sense, it needs to be renewable. So you need large volumes of renewable energy. You can do some simple estimates on this and the scale of it is quite daunting. So the short story is because we're not doing enough today we are giving a very large challenge to future generations. That is what it is. Yeah, I think just what was it, just last week the largest direct air capture system was just started in Iceland if I remember right and 4,000 tons metric tons per year is what it's after using geothermal as the energy source. Yeah. So Rich can you give us an idea about the engineering or environmental challenges that might arise as we transition to more and more renewable energy? Yeah, one of the ones that I called a few of my colleagues here over the last few days just to get some further insights as to some of the issues that they're running into and a couple of my colleagues, Don Urbanbeck and Kurt Bierlein indicated that some of the issues are associated with the distribution of power from these renewable resources and the importance of maintaining the frequency at a very close tolerance across the grid and how these intermittent sources can be problematic. So that's a significant engineering problem that many are working on but as the amount of renewables increases over the coming decade that problem is going to continue to be an issue. One of the ways that we've been involved with and trying to help resolve that intermittent power generation is through pump storage projects and they provide an excellent opportunity, a great battery if you will, of being able to stabilize the delivery of the electrons across the grid in a very quick way. Yeah, interesting. I see we have an extra five minutes for our panel, so I'll continue. Brian, you gave a wonderful view of a future of aviation. You've talked about more carbon neutral fuel and you mentioned autonomous planes. You showed this vision of cities with little airports, almost portable airports inside of them. How likely is it we'll have personal travel in say electric no-pilot vehicles? Is that something that your joint venture with WISC is aiming to do and will that be, it sounds like a major game changer? Yeah, well, we are actually flying electric autonomous airplanes today. So we have a prototype airplane which seats about two people and we fly it very regularly out of California and New Zealand. And what's interesting now is we're working towards certification of a product version of that airplane which we expect has the ability to be certified and actually perform commercial operations within this decade. So I do think that certainly within our lifetimes we will see people flying commercially on electric airplanes and autonomous airplanes. And what's really interesting about electric airplanes, just putting aside the sustainability aspects for a second is for the first time in aviation industry we have this relatively kind of transmission agnostic propulsion system which allows us to configure aircraft that can do these really sort of interesting things. So you'll notice if you've sort of delved into the space at all, you'll notice that at the high end of the market airplanes all generally look the same, tube and wing kind of configuration. What's happening at the low end of the market is this great sort of experimentation of what airplanes look like and how they fly and how quiet they can be and how much stuff they carry and how far they go. So there's this incredible innovation that's happening at the low end of the market which is driven by electrification and those aircraft will operate safely and in high density because of autonomy. And so I do think that that's something that's going to happen in our lifetime. It's what we're doing in our joint venture at WISC and it's an incredibly exciting time to be working in that field in the aerospace industry right now. That sounds like it. I wonder, we just have a few minutes. I wonder if each of you could give an example of some selected technology that really inspires you because it's novel, evolutionary, very different. Something in your field, something that your customers have asked for, something that really gets you interested and can get the younger generation interested. At the risk of being a relative broken record, I suppose, there are two general purpose technologies that are happening in aviation right now. General purpose meaning they're kind of infusing everything that's happening in the industry and that's electrification and autonomy. And so electrification in terms of battery electric systems won't make their way to the higher end of the market immediately because of the physics involved. You're not going to be flying on a battery electric 737 within our lifetimes. But what's happening is this interesting activity at kind of the low end of the market and autonomy is truly scale agnostic. So as we work out how to automate safely more functions of the system, that will happen at both the low ends of the market and the high ends of the market. And there are just incredibly interesting technology challenges within the autonomy domain to make them really robust and safe for operating in the aviation industry. Yeah, I think I'm ready for that. That'd be great. The one I'll add is that Boeing's definitely, you know, certainly involved in this are we're focused on Earth observing systems and the data flows that come from them and how they can help characterize the natural environment. There are more and more satellites going up every month, every year and the spatial resolution and the data returns is tighter and tighter and the quality of the sensors is improving. So that's a technology that we're ingesting and applying to problem solving on both the natural and built environments. And lastly, I'll just say I'd like to just acknowledge a couple of other colleagues, Nicole Flanagan and Carrie Saban, our sustainability officer, who also provided insight and support and helping bring our understanding of mitigating GHG emissions to the marketplace. Thanks, Rich. Eric, any thoughts? Yeah, I'll just sort of a bit of a cop out here. So I started in the field of CCS 20 years ago and it's not a novel technology per se, but through these 20 years, a number of companies, SYNTF, other universities and so on, we've been working on it and there's been hiccups, it's a new technology. There was some issues you didn't realize before you started piloting. You go back to your drawing board, make adjustments, and finally we're getting the plans built. So it's just your mind of sometimes technologies is, you take it step by step and you finally get there to the full-scale commercial deployment. Yeah, yeah, yeah. So in our last minute, a quick yay or nays, is carbon capture, adaptation, climate change mitigation, is that? The growth industry of the future. There's one. There's one. I look at it as a trillion-dollar industry and I'm sure you all agree with that. Thank you so much. It's very thrust to do well as a designer. It certainly is. No choice. Thanks everybody. Thank you so much. Thanks. Anna, back to you. Eric, Rich, Brian and Bruce, thank you so much for sharing with us. Not just your experience, but your thoughts. We all disagree together. That's really marvelous and thank you again for joining us. For pleasure. Thank you. All right. With that, our IBM Research Day 1 of the Horizons event that focus on mitigations of climate change just through the close. I do want to remind you that your feedback is extremely important to us. So at the closure of the session, you will receive a service, so please kindly share with us your feedback. The event itself will go on tomorrow and on Thursday, each of the day, starting at 9.30 in the morning. Tomorrow specifically, Hendrik Hammond will open the session and the keynote is by Ibrahim Acheidong. He is the very important participant in our session as the UN Assistant Secretary General and Director General of the Africa Risk Capacity Group. He will talk about this group and what does it do for the client risk and management in Africa. We'll also have presentation on analytics for the client impact from our IBM Research Team in the UK, US and the Weather Forecasting Science Team. We'll later have a session on artificial intelligence for better climate predictions. That's from our team in IBM Research Brazil. We will be moving then to the climate aware applications and the man forecasting and that's from our research in IBM India. Then we will have a session on climate aware applications with focus on decarbonizing the electric grid. And then we will close the session again with the panel of exciting pod leaders in this area. So thank you again for joining us today and we are really looking forward to seeing you tomorrow. Thank you.