Matthew Eby, Founder and Executive Director, First Street Foundation
In conversation with: Kristen Bellstrom, FORTUNE

Brainstorm Design 2022 - Climate Risk At Your Fingertips 

Conferences

with the effects of climate change increasingly evident. American homeowners can find it difficult to properly assess the risk to their properties and lives from wildfires and floods. The First Street Foundation with its team of client scientists, Computer modelers and researchers can help. Last week, this Brooklyn-based nonprofit released a wildfire risk model showing that almost 80 million us residences and buildings and commercial businesses Face some risk of wildfire damage over the next 30 years. Um, and it also comes on the back of their flood risk model, which showed that actually millions of homeowners are at risk from flooding is not actually disclosed by, disclosed by federal maps. Um, first reads founder who's going to join us in a moment, is going to help explain how they designed their sophisticated model and why they're making their predictive data available to everyone. So please join me in welcoming Matthew eby, founder and executive director of First Street Foundation in conversation with brainstorm design. Co chair christian nelson. Welcome Matthew, Thanks for having me. Um, so earlier today I was checking the news uh, and I saw that the National Oceanic and atmospheric administration has announced that it is expecting an above normal atlantic hurricane season this year. If that happens, this would be the seventh consecutive year. Uh, 2021 was a record year for wildfires. Uh, 7.7 million acres burned in the us. Uh, we're coming up on the 2022 season for, for both flooding and fires, I assume. Um, so I think, you know, these stats make it clear, uh, how important the mission of your organization is. So tell us a little bit about these tools that you're building. Yeah, absolutely. So we, at First Street Foundation originally started with flooding. Uh, and we start with flooding because it's the costliest natural disaster. So like you're saying the stats showing year on year, these things are going up into the right and from Canada. So we would say that the hockey stick, but unfortunately, the tools to turn these things, these physical climate risks into, uh, something that your average american can understand just don't exist. So the federal government produces flood rate maps that allow you to understand if you're in a flood zone or you're not in a flood zone, but they don't have anything to do with the property specific risk. They don't include certain elements like precipitation in their modeling. And most of all, they don't include climate change. So it's all a historical back looks. So what we've done is we've created not only the, uh, models, So working with over 80 climatologists, hydrologists, um, all the, as I say, uh, to produce the model, but then we've taken that and applied it to every home so that we can then turn it into statistics that your average american can understand. So we can tell you what's at risk for your property. How do you protect your property? And so the design behind that is really the, the UX and ui to turn these really complex topics into something that's really simple. Um and we have a slide that shows, I guess the difference between the fema models and your flood models and and that is because some of the things you were just talking about, like them not taking into account rainwater and things like that. Yeah, that's right. So when we first launched in june of 2021 of the maps that we produced was a comparison with the fema special flood hazard area. So we showed homes that were at risk from flooding, that 1% annualized risk. So that special flood hazard area and then what showed in our model for the same level of risk. And what we saw was there was about six million homes that had that level of risk that weren't in the special flood hazard areas from fema. Um now I think both of the tools, the flood tool and the fire tools also have future risk built in. Right. So my question about that, just from a design perspective is, you know, climate change is changing very quickly. It's happening more quickly than scientists have predicted. How do you look that far in the future with any degree of accuracy? Yeah. So there's ever evolving climate science as you're mentioning and we're getting smarter and smarter about these things and the global climate ensembles and the models that are leveraged for it. So what we do is we continuously upgrade our models with the latest science. So We look 30 years out, which in climate terms is not very long, but for us, you know, that's the term of the mortgage. So that's very important when you're thinking about buying and selling a home or renting a home and where you want to live. So we're able to update year on year with the latest science and in continuing to evolve in that nature. Um, but we're always able to give that shorter term window, which doesn't have too much about divergence in the emission. So when you look at emission scenarios over a 30 year period, most of the, what are called representative concentration pathways are very kind of similar. And then when you move further out to the kind of 2100 is where you see the big well, are we going to be on a one degree, two degree three degree four degree pathway. Um, so as you talked about with the flood tool, you had some data to start with, right, you were starting with the fema data or at least something to build on um, with the fire tool was that the case, are you starting completely from scratch And it's so like what kind of data sources are you, are you looking for that one? So with wildfire, the biggest source of data that you need off the hop is what are called fuels, which are things that burn, which is generally the forest. So there's something called the wildland urban interface, which is the wUI, it's a fun fun thing to say. Um and the WUI is where in the forest are the certain types of trees? How high are the trees? What's the condition of them? What's the type so that you can understand when there is a wildfire, How likely are they to to burn to combust? Uh So that data is available from the Forest Service, but there's not really a model out there that you're able to leverage. So we actually used wildfire software uh called Elm fire to then take all these different data sources like um the Forest Service um, dataset around the, the wildland urban interface. Use that with um previous known wildfire combustion sources and then use a 10 year time series from Noah, the same group that you're talking about at the beginning to then say, given all these weather patterns, given where wildfires can start and the conditions of the forest, how would they spread? And we did that 100 million times, um, to then have a, what's called a monte Carlo simulation to get two probabilities of things happening. Um, did you find anything with the fire map that surprised you or surprised other people? I mean, me personally, I'm not, there's a lot of climatologists on the team and experts on the team. I had no idea that florida had so much wildfire risk. And our, our chief data officers from florida, he said of course there is. Uh but there's not huge fires. You always think of the west coast when you think of fires and you think of large burning acres for days, weeks months of of long burning fires that were trying to put out and suppress. But within florida there's smaller fires that are very close to properties. So the number of properties that have that level of risk is actually very similar to California. But what we saw within the introduction of climate when we look at what's going to change within the forests and within the ignition points and those other elements that go into risk, we actually see that California has a six X increase in the homes that have that severe level of risk where texas and florida which have about the similar amount of risk today. Double so double is still bad but it's just not the same as six X. Yeah. Um I want to check in to see if we have any questions for Matthew yet. I have several more if not. Um Okay I don't oh yes sorry. Lights over their heart please. As a west coast person. I'm wondering about the six X stat. So where where do you find our places in the US that looks favorable when you look at these dimensions? Sure. So Rhode island is the place to be. Uh So uh Rhode island has what we use in our model, we use suppression which means people go out and put the fires out and uh fires that don't make it over 100 acres are aren't in our model. And Rhode island has zero fires that have made it over 100 acres because they're very small, close proximity to water and they have a great fire suppression team. So we actually show zero risk in Rhode island, which is kind of counterintuitive, there's obviously risk, but the way that we model risk, it doesn't show. So that actually in the east Coast, for the most part has a bunch of areas that are that are like that. Um can I ask the same question about flooding? Where, where are we safe from flooding? Flooding is a bit different because it rains everywhere and so you tend to have wrist, so when you're further away from the ocean, you tend to be in a better spot. But then that's not necessarily true for places like west Virginia that have a tremendous amount of precipitation risk. What's called alluvial risk or riverine risk, which is called alluvial risk. So flooding, there are some better places than other, but there's no, there's no Rhode island. Unfortunately Rhode island has a lot of blood red. Okay, we have another question over here. Yeah, I'm curious. I love this. Um and I love the premise that it will empower better consumer choice. Also think of the broader stakeholders and I'm curious how you think about it being utilized and even weaponized by other stakeholders, like say, insurers. So what we do is we try and democratize the data as we say. So we have great partnerships with realtor dot com or Redfin. So anyone who's kind of buying or selling real estate will find it when they go onto these portals or if you just go to risk factor dot com, type in your address, you get that information. Um, Insurance companies Tend to have a different view and a different model than us because they only care about 12 months. You sign up for a 12-month policy. They know what that risk will be for 12 months and then they address the premium the next 12 months just to compensate for that 12-month area. So they use a different model and a different technique to us. Uh, so our data actually doesn't inform insurance companies in the way that you might think. Um, but what it does do is it empowers consumers to understand what they can do about something. So there's a lot of people that have, for instance, homeowners insurance typically covers wildfire. But if you're on the west Coast, as we were just talking about some places are very severe risk. You actually getting a letter from your insurance company saying unfortunately we can't have wildfire, be part of your coverage, you'll have to get a separate policy and then that policy costs a lot. Uh, so what we're hoping to do is try and find solutions around what you can do for your property to minimize that risk. So then it'll actually be included in the pricing of your insurance where you can be certified. So they're great. There's a great nonprofit I BHP that's working on that exact thing. Um, I'm curious if you know beyond insurers, did you get any pushback from homeowners? Like I feel like this, this data not being available and then suddenly you're trying to sell your house and ah, here's a new level of transparency that perhaps you had not hoped for. We, we've had a lot of conversations internally about that, especially when we were very excited to land originally the partnership with Realtor dot com and now subsequently many other portals. Um, it's a difficult conversation because do you disclose this, which may impact someone today, but then the person who's buying that home doesn't understand the risk either. And then at what point does that happen where you distribute the data allow people to have this democratization of this type of data so that you have an informed country. And so we agreed internally that that was the right thing to do and to, to put this data out there, we have a full support team to answer questions. We aren't saying don't buy this home that's beside the ocean because we all want to live beside the ocean and vacation there and it's beautiful. Just know you should elevate your home have the right insurance to know what's going to happen over the next 30 years. I know that when you go to sell this home, someone's gonna be looking for the next period past that, which means your risk is not just your whole period, it's the next hold period for someone as well. Um, so tough ethical choices um, that we have to make. Um, but we believe that it's the right thing to do to give everyone this information and allow for uh, you know, a free flow of information. Yeah. It's unbelievable that some of this hasn't been available in the past. Um, I'm sure there are many risks to look at, but have you guys thought about what you might model next? Yeah, we just launched this other one. But yeah, no, it's okay. Uh, so wildfire just launched last monday. Uh, and we have our, our heat product, our heat factor coming on august 16th. So, uh, an interesting one there. If you take your hottest week generally in a year, the hottest seven days. What you'll see 30 years from now is that tends to be 30 days. So if you have seven really hot days right now this year that are in the city probably about 100 you're gonna end up with 30 of those, uh, in 30 years. So about a month. Uh, so not great news, but that'll be the next, um, foray that we have and then past that we're looking at Air quality, which includes smoke and things called particular matter 2.5 and ozone, uh, and wind and drought. So we have a bunch of models under development with the same kind of playbook that we run in the past. Okay, You guys have some work to do. Um, Thank you so much, Matthew. Thank you for having me.