About the talk
Financial institutions are playing an increasing role in the low-carbon transition by taking steps to accurately estimate, price, and disclose future climate risk. By quantifying their exposure to climate risks, financial institutions can more effectively allocate investments, avoid ‘stranded’ assets, and track adherence to Paris Agreement goals and shareholder commitments. However, it remains difficult for these institutions to assess climate related risks across a portfolio of assets and across different benchmark warming scenarios.
We will cover large scale data transformation approaches as part of an end-to-end framework for quantifying annual, asset-level climate risk over multiple climate hazards including wildfires, inland flooding, and heat waves using simulations from global climate models participating in the Coupled Model Intercomparison Project Phase 6 (CMIP6).
We quantify forward looking climate risk from 2020 to 2100 under multiple climate scenarios such as high-emissions (SSP5-8.5) and medium-emissions (SSP2-4.5) warming scenarios. We will also showcase intermediate steps to make the climate simulations and spatiotemporal data interpretable and actionable.
Gopal Erinjippurath serves as CTO and Head of Product at Sust Global, a venture focused on geospatial analytics for climate adaptation. Most recently, he led the Analytics Engineering team at Planet Labs, an integrated aerospace and data analytics company that operates history’s largest commercial fleet of earth observation satellites. Planet Analytics serves a range of customers from city planning teams in governments and the World Bank to defense and intelligence functions across the world. He is known for agile engineering execution from concept to scalable high quality products. He has been an invited speaker at global industry conferences like Google Cloud Next and leading technical conferences in the machine learning space such as ICML and CVPR. Previously Gopal managed teams working on industry-leading analytics products from early concept demonstrations to multiple customers at Captricity (Acquired by NASDAQ:SSNC) and Harvesting Inc, where he advised the CEO. Previously he led the algorithm engineering development of Dolby’s (NYSE: DLB) first imaging display product, the Emmy Award winning Dolby Professional Reference Monitor. Gopal holds an MS in Electrical Engineering from University of Southern California and completed the Ignite Program, connecting technologists with new commercial ventures, at the Stanford Graduate School of Business.View the profile
Aquacal is presenting his talk fast and efficient climate risk estimation of global asset portfolios. Take it away. Alright, thank you, Sarah. I am. The midst of figuring out the stool. And I think, You folks should be able to see my screen now. Okay, Sarah can just say you can you can see it. Yes, we can see your screen. Okay, thank you. I just want to make sure I'm not, I'm not talking and I know that I'm super excited to be here and machine learning conference again, in 2018 when the world was a much different place and it was at
this location. So I'm still in California and I wanted to share with you and I updated the title of the talk to be scenario analysis. That's relevant to this crowd and Analysis using Google,. So, I'm a little bit about myself and I spend the rest of my career working a multimedia products. And some interesting people lives in Lodi. Oh, and a video codecs. And in the last five years of looking at it, as well as And in my most recent role to a prior to SAS Global, I was the head of
Athletics and insights of Planet Labs. As a few if you know, the largest completion of observation and my last ones around. It's just how do you use Suicide Squad, stealing formation signals on the world and the being able to sell that it scared us actionable Insight. So will our goal is transforming idea, Frontier climate science into actionable intelligence and the next 15-20 minutes. I'm an intelligence, which y'all give you a landscape overview of our time attack. And from that.
Models, I know how that impacts would do, you know, and then get to what the exhaust call acid level juice Wasteland lyrics. So, impolite to address it after. I had the moving in the context of dido machine, and the components of understanding rest. We are all Avera flight, you know, how they Baldwin climate at, create something specific interest in terms of closets, in terms of exposure. And a few minutes comes to bigger organizations in defining policy, a big part of that at translate sued, you know, the
context of Planet. So how do you plan for the rest? And then there's the element of like dealing with coping and adapting. To the changing climate is more in the context of Disaster Response in the context of carbon capture. In the context of carbon sequestration is our technology and the new eagle. The focus at tasks. Level has been largely on understanding and planning for recipes of financial institutions. City governments on the capability to be able to look at long-term forecasting, affect the climate
and understand how that impacts their tangible assets on the ground. So when you look into changing climate, like dealing with nausea, three, kinds of three times of slide of Guillain-Barre and datasets. So on one side, you have these catalogues of exposure. So being back like new and has Adidas at around, what's the temperature of the Cyclones, tropical storms tires, he pays. And then strings of active. Has it's happening all over the world as we speak, but you could be active fires could be active slides. And then
landscape seems to connect is it is getting the historic theater using that for inference and forward-looking climate and In the context of time in intelligence, all these three are equally important in understanding of the storage, and the near real-time data that helps. You better interpret, the forward-looking projections, and looking the projections head to info on like a coincidence that ever occurred in the past. So that's roughly like the 3D two components. And
the word scenario is a lot and like I feel awesome. You can be a little confusing. So if you can't enter the naive definition of a scenario, still like mostly to sequence of events and Is life. Scenarios are just like a grouping of his dump ships made on the future. So when we talked about that in the context of climate climate scenarios, they like a group of possible social and economic Barclays between now and the toast of the century. So is a fairly complex based in New York and you talkin climb that you're talkin about. I
need to bleed a global phenomena. Are you talking about a long time Horizon? Evolution of Flight of the variables and in order to accurately project into the future, you need to take into account the social interactions. The economic policy has been a slight, the initials for 6 that go into the into what important modifies the other subsystems. So that's likely the cryosphere, the landmasses the ocean. Yeah, and the atmosphere. So that's kind of what
is called the time. It's an hour is. So if you see a lot of isolation and thighs need to be in to our understanding of example of a few years ago and still traumatized, but the reality is, if you look at the historic perspective and use you of age was calibrating. What do your thing today on the last few years? Compared to what happened in history and gives you the trends into the future. In the context of climate scenarios, one of the land and climate
modeling one. Is the ancient all metal panel on climate, change the ipcc. And in the last five years, is a comprehensive framework by which the climate projections and five and modeling across many different research, Labs, all over the world come together on the common framework, and that's the female sex. So, the sea music stands for like the Top Model, India comparison project, and it kind of defiance different social and economic scenarios that allows you to create some bundle together projection, which are similar scenarios scenarios
that you hear about a lot in the Future. Instrumental two words, interpreting forward looking for Jackson. One of them is the SP one has called The Green Road. So be as a society, make the right at, take the right steps to as a tactician and mitigation. That's where we end up at 6. And I'm on a flight extreme hazard in the least amount of global warming between now and the toast of the century, but it's still not extreme baby. Doing the getting to the Green Road and be currently projecting us house because it's like the greatest in the bus Keys. We aren't doing enough on
climate and climate adaptation and that's a business-as-usual scenario, which projection uploads of Vine, v watts of additional energy. Some systems and that's what's called a 65. So that's happening. The world has tremendous, things are all these research institutions, all over the world that are contributing to the scene of six train book of projection and significant amount of research in almost all continents, pitches that you encouraging and a big pot of this is Theresa child into something actionable. Do understand. What's the climate risk? The assets that
you on? What's the time of this blasted said the organization you work for owns. What's the climate risk to the regions of the state and city taxes? So that's kind of like the Broadway such large said that I had in the back of the year. And to the close of the century blackest kind of like the projection based on the historic data from the specific model. So you can say that this is inferences can be tested. And what you see in different colors is inference that is protected under these different
socioeconomic Pathways of the ESPYs and she is like, you know, in the business as usual case, some significant warming up to 5°, and if you, if you go to the more conservative Green Road Park Cycles are the two 2° which is what I mentioned earlier. So this could be broadly like the realm of climate and tell them to just living in history. And we understand just from warming, I will be calibrating two projections from a model. You can calibrate do like the mole, more recent data that you collect, and then you can help you calibrate better and characterize the performance of
these models over the long-term. So I just want to talk to you in a little bit in terms of all me and temperature instead of abstract. So when I get to a couple of studies, which could be interesting example, so, you know, the fundamental variables like temperature, precipitation and extreme precipitation in humidity. In then project out of the stream has the most sophisticated physical hazard moms. And one of the most significant physical hazards that emerge from climate of wildfires in the
direct impact into the human settlement heat resistant and the Pacific Northwest in particular as being like a, an extreme case, you've all heard of like the wildfires in California or the last couple of years and this year in particular who she like in September. Experienced in recent history and this kind of like a great summary of Kingsland. So this a graph from from NASA shows the emulator fire detections across the Pacific Northwest and for every year in the century and you can see that in 2020
you had like the human actions to be higher than any other year is higher than just any other year, but it's also significantly higher. It's almost twice what we saw in 2018. That's a so far end of September and we knew enough to where they were still fires in the Southern California and increase. Temperature. Increase brightness in specific parts of this region and has a simulation from the domestics. The usual scenario of what you see across the Pacific Northwest Eye in terms of a monthly fire detection. Using those models on sampling, the
outputs from a mini model, such a part of the scene of 6 across the specific business as usual scenario and then mapping that into Bulgaria crossword, self and day by. You can see what the the projections look like between between over the last five years, which allows you to calibrate against. So that the mood is also satellite from NASA that want to fire. The Texans is much higher than what a scene at 6 business as usual. So now your job is not just in the Pacific Northwest.
Russia. And that's and that you can see through data, because and this is from the oldest mod, 14, even transfer, station. You can project out physically has its like wildfires into the future. You can also a map that to observed phenomena, with the recent recent. Is he holiday calibrate and that allows you to have more accurate projections into the future. So, now, you can see the fires in the Pacific Northwest. You don't see the fires in Alaska. Southern California rasina consistent over the last few years. And if you look at the time in Marvel's, they predict something very similar between
now and through the centuries as I possibly can. Do you want a One standard deviation you getting like 3.3% gone. In that specific and very high Allegiant by contrast. The region like a laska you see no damage fire today in from the models, but over the course of the next decades, you see an increasing amount of barn area and that's a function of the permafrost in Alaska that exposes Moreland, which makes it more exposed to dry and warming scenarios which breed
projections for Popeyes in the future. So that's kind of what you see like in Southern California fires right now, like the fire for brand. Is it expected to be something similar, but in an area that you don't see that many fires in the last several years and then you begin to see the fires you seeing an increased rapidly over the next. Next to eat a jacket. So that's kind of interpretation of the forward-looking climate prediction long time Horizon Global phenomena to all of these is not the same when it
comes to this sort of projection. Another key study mention, is that a flooding, you know, just like pie, pie's, gardenias are so huge. A disturbance to the human system of the human settlement in different parts of Africa, with a need for infrastructure development in the rural areas and a growing population. So you just kind of like to eat Max flooding exposure. That's not from the same business models over the business-as-usual scenario, for the next 30 Years, which indicates know, there's some reasons which are definitely having significant exposure over the next 30 Years and the
enemies of progress, against the Mets. There's no guarantee that they'll be a flood in the specific year or a specific month, but you can Uber a goes off, make a longer time. When do you can predict around? What the what, what level of exposure could be? And that's the snapping. I like the restaurant part of Africa is home to many metal mine site, and around a metal mining activity, is one of the biggest sources of conflict between mining companies and the local population because it connects into what is streams. Mining requires a lot of water at Candy's connection to their local
ecology. It connects into the economic stimulus as a fabulous like even take the remaining companies and the local population and you in mine sites across other active, inactive and Reclamation sites. And across each one of those lizards total number of the remaining. And the reason for the mining companies flooding is like a huge destruction because normally they ain't been there is flooding. And it also creates plugging in the mine site, which also creates in our discussions about music on mine, any longer. You got to wait for it to dry
or get the butter out in some other way. The other aspect of this more easier in some cases has more significant. Is it from the the flooding and over topping from Trading? Stocks have a favorite of toxic material. You also getting rid of I mean you have flooding or extreme precipitation chance going to flood an Uber and then that's another huge in my math class until these sites up high risk when there is flooding. And if you could do specific locations, one of the other one is with
over the next three decades, that gives you a sense of both of these are very significant flooding that has destruction over the next couple of decades. And there's a very high probability that there is flooding in the next 20 years around this Christmas. Location of the possibility of the water getting contaminated, Uber talking to Tina things. In what I tried to highlight those two examples, is the fact that need to be looking projections, which indicator of what the future looks like a cross different scenarios and
across regions and then comes to city planning on when it comes to an individual and organization or like a region. Trying to understand the risks. They have to find a cheap flights are much smaller resolutions, not Global exposure, my home, or my community. I to buy Fireside Cycles, that's kind of like the beginning of like are orange dress plus tax, which one of them and then also impacts the business decision, making a tangible asset, or I just walked to and
go from a global to a celebrity match, higher position position with a higher degree of confidence that it is going to be in extreme weather at your acid-notation and then And you look at the next few years. There's a little better. So today, because you have a better understanding to Historic or an early warning and what the power to keep Rejection. It mean, these words together. You need to be able to update brain the state-of-the-art time and models
the most active, and leading-edge climate data streams, as well, as like the event catalogs, and help analysis at the local level, and the ability to create metrics, that are injectable notice to machine learning for science, folks, part 2 professionals. And then Package that together and do a simple interactive tools that all over the world can use and make meaningful relationship and that's kind of focused on at Tuscaloosa with these areas are open. Casting. That's definitely something that you know, I would love to talk to you
more about and this is the most important and pressing challenges in our society today and visiting the ramifications of climate change as we speak of with the active fires and floods in different parts of the world Community. Indeed. It can be a huge difference. So I would definitely feel free to get in touch. Say hello. You're always hiring. I'll be with the punch bowl controls that they looking for a guy in the next few months. And if you always interested in talking, so
I just want to know, for like this opportunity and for the chance to share an introduction to 5 and bottles and climate scenarios, and I can make a little intelligence a possibility. Thank you, Google. How climate change affects our future and understanding the risks associated with climate change. Should definitely be at the Forefront of everyone's mind, including businesses.
Buy this talk
Buy this video
Our other topics
With ConferenceCast.tv, you get access to our library of the world's best conference talks.