Extreme weather risk is changing—are businesses ready?

Brian: Before we get into the results of the survey, what do you do for FM?

Angelika: I'm leading a global team of scientists, based out of our science and technology centers in Singapore, Luxembourg, and Norwood, who are looking into anything hazard related to natural perils. So that's, earthquakes and tropical cyclones, wildfires, heat, of course a lot of flooding.

It's a very mixed group of people, many nationalities, very, very smart minds. So it's been really fun in untangling the natural hazards risks, for FM and for our clients and getting our clients into the changing risk, in some of those fields.

Brian: So tell us a little bit about this natural hazard survey. What was FM hoping to learn by doing the survey?

Angelika: What we want to keep learning is what are the key concerns of our clients? What have they experienced? How can we ultimately support, preparing our clients for climate risks?

But also on the broker side, the survey was looking into what's our brokers from outside in view that clients may miss. Or vice versa. Do actually clients feel there's more to it than brokers may perceive?

Brian: What was your big takeaway? Was there a headline for you here?

Angelika: I guess my biggest surprise was that business disruption, because either the infrastructure that feeding into some of our clients' assets or impacted communities, that seems to be the number one concern of our clients.

Looking at the broker results, this is a concern they would classify as maybe a mid-level concern. Certainly something I've been surprised you would've thought, maybe direct property damage is the key concern of clients and it is the key concern from brokers. But there's definitely a strong difference in perception on how significant the risk is.
Brian: How would you have answered that question? What the biggest natural hazard risk is right now?

Angelika: It really is flooding. Because, the extreme rain events are increasing most and  no matter what you look for, if it's overland thunderstorm activity or if it's tropical cyclones that hold much more water in their systems or convective systems in the mid- latitudes, we've seen last year the big DANA storm in Valencia, which was a huge impact event.

So these type of events we see are increasing and I think it's just probably the hardest peril to grasp for people. Because you stand next to a small creek and of course when rain comes down, a small creek can turn into a five, six meter wave and, and wall of water. And people don't comprehend.

It is hard to comprehend how a little creek can actually turn into a wall of water. So it's always hard to prepare when the water is coming quick. So people also need to react quickly to where in the region really the rainfall comes down. So we have seen in particular in Spain last year, how important sufficient warning and preparation is for the people on the ground to know what to do actually when a warning comes out. But also that actually even a warning is released and the right infrastructure measures are done. Is there enough flood defense in place? Do we have emergency services to jump into action? It's just a very, very complex risk that is changing quickly and that makes it such a big risk.

Brian: What does it look like when that creek turns into a wall of water for businesses that are in the way of that?

Angelika: Well, it depends how, how well prepared the business is, to start with. Because there's a lot you can do, to at least limit the losses. But if it intrudes into your building, and it gets into your power system, that will wipe out your power system and turn it down. If you have chemicals not stored appropriately, you may deal with very polluted water. You may have machines that need to be replaced, which are very tailored, fabricated, and may have very long delivery times. So you may face really, at least weeks, if not months, in a worst case, you have to shut down your facility due to downtime, because you don't get back to business.

Brian: So what can people do to protect against floods?

Angelika: Well for once, first you really need to understand what is actually the flood risk you're prone to. Where could water come from to your facility?

Are you in a mountainous area and there's a large runoff, or is there creek beside you that may swell to very, very high levels? Or is a river system nearby where some dam may break water can intrude over topping and impact the facility.

So depending the nature of where the water comes from and how much of the water, there are ways to secure any low level areas of your building like loading docks. You can build flood walls or flood panels in front of your doors. There are shutters for windows. There's certainly ways to secure any hazardous materials, gas tanks, chemicals. There's easy tricks just to raise up anything that is particularly vulnerable to flood intrusion, any machinery, put them on a bit of a small pedestal, so isn't impacted firsthand by water intruding into the facility.

It's often small things. It doesn't necessarily need to be big costly investments that are very, very impactful to help.

Brian: Why is extreme weather important for an insurance company to be investing resources in?

Angelika: Well, first we actually have losses due to extreme weather. Extreme weather such as, windstorms or thunderstorms cause flood events, but also strong winds. So in order to understand which clients need to actually be better prepared for which perils, we just really support our clients in avoiding downtime. And that's really one of the key goals at FM, what we want to achieve that our clients keep in business.

So we want to understand who is exposed to very severe winds or severe flooding, get the right prevention measures in place so they don't experience losses during these events and, and keep continuing their services or their production.

Brian: What makes you excited about this work?

Angelika: I think I always had a passion for natural hazards, volcanoes, weather. I think they had this moment in my studies, when I realized there are those jobs out there in the insurance industry where you do really cutting edge science, but bring it into a very impactful use-case. And that's certainly what motivates me. That we still have this innovative nature of our day-to-day work. But it's not just writing a paper and that was it, but actually turning it into something that our clients or maybe even society, benefits off. And that brings an extreme strong purpose to me for sure.

Brian: How does your work get to the clients?

Angelika: Mostly this work is getting to our clients through our engineers on the ground. So we provide our information on the risk exposure, so field engineers and account engineers can take the information to the client to have meaningful conversations and actually building a partnership with the client to establish what is the right prevention that needs to be in place in order to keep the business as good, as up and running as possible, even in case of an event occurring.

Brian: What's a time that happened recently that seemed most meaningful to you?

Angelika: We just developed, within the team, a tool that allows very high resolution, flood assessment for single locations. We've done quite a number of analyses in the aftermath of Hurricane Helene, where we looked into what has actually been causing loss events? Is this something that's gonna happen again? How likely is it, was it like a bit of a freak event or is this something clients will need to prepare for? And as a result, many actions really have been rising from the clients. 

Brian: I've heard you talk before about how AI and machine learning will help modeling. How do you think that these new technologies are gonna help change how these perils are modeled?

Angelika: I think it's super exciting. First of all, 'cause it, it is such a huge tech capability field that starts to open up to scientists to unveil new patterns, new analysis methodologies, but also bridge some of the computational limitations we used to have. We come from numerical modeling. So numerical modeling means really dynamic, very complex climate models, weather models, that use explicit calculations of each physical process in the atmosphere. So you just can imagine how much calculation points are necessary to simulate single events or predict certain climate patterns. So there's a certain element of just making more data, more analysis, methodologies available.

Let me give you an example where we are currently working on very tangibly in applying AI methodologies. 

So we are trying right now to understand, based on historic data, where we have thunderstorm activity? What conditions have to be present in order to have thunderstorm clouds forming, but then what's the change in conditions in order to have these thunderstorms producing hail? or extreme rain? or very, very strong winds? or a combination of the above? And I think that's where it’s very, getting really exciting. You can actually use data to identify even different nature of the same meteorological phenomena, but untangle the peril impact and ultimately the hazard, our clients are dealing with.

Observational data has always its limitation. First the person needs to be there to observe something to start with. But also, as now in the modern era, of course we have much more technology for people reporting back observations. 20 or even worse, 30, 40 years ago, there was by far not the sensibility to observe weather or report back. So there's a huge temporal bias in terms of how much observational data we have. And somewhat, you gotta bridge this knowledge gap. And AI really helps us when we train on events, we know there have been occurred. And where do we have observation gaps where the same type of single peril impact or combined parallel impact has been happening in the past?

It's amazing and it's really for me it's just so cool to work with my team. You have this, just incredibly clever people who think up these concepts and build out these codes. And, I’ve got to call out to one of my team members who's just so excited every time he finds something and it comes up and “Angelika, I have dopamine. I cracked this, I found this, and it works.” And it's, it's so exciting to see how we really reveal new findings.

Brian: So obviously a hurricane is a very large storm system, but a tornado or a hailstorm is a smaller discrete event that might be more difficult to model and AI will help us bridge that gap. Is that right?

Brian: I understand we have a new partnership with Princeton. Can you tell us a little bit more about that?

Angelika: Yeah, sure. We're super excited about this partnership with Professor Gabriele Villarini. So we have two projects up and running right now. One project is looking into how is riverine flooding changing in the future, just due to changing rainfall rates, but also how are rainfall rates ultimately changing the risk, and, and the water heights in the river systems.

So we just stood up a second project looking into tropical cyclone induced rainfall and what is driving the very extreme nature of such events, which are often not purely due to the tropical systems on its own, but actually, are often seeing interactions with other, more extratropical systems. Understanding the dynamics and the physics behind these, low probability, but high impact flood events, will really help us to get our grinds into some of the biggest floodings we've seen in the past years, such as Hurricane Helene, Hurricane Katrina, another example are for sure the Beijing floods we had a couple of years ago.

And it's these events we are really worried about and we want to see what do we need to understand in order to be able to model such events much better. And, the partnership with Princeton University is really absolutely crucial to build the base data in order to improve these understandings.

Brian: What do you think climate research is gonna look like in 2030, 2035? How is it gonna evolve and differ?

Angelika: Well, we, there's certainly the rise of AI, which will have a very big impact. We have seen in last year and this year, a release of a couple of AI weather models, being it from Microsoft, from Google, from Nvidia.

And that will be super interesting to see. There's very little validation done so far, how they compare to the more classical weather and climate models. but as these models also evolve, as AI is just on such a steep increase, it's exciting, for me, as a scientist, this is fantastic to see and explore, and also explore the opportunities for us how to use it to do better risk assessment.

Brian: There's obviously been in California wildfires, in Florida, there have been hurricanes, there's concerns about areas of the world that might become uninsurable because of climate change. How can insurers, not just FM, but the market more broadly, respond to climate change in a way to make sure that the market will remain insurable?

Angelika: It's one of the big questions we have as an industry. I think for us it's much easier to keep relevant as an insurance and keep insuring our clients because we can work together on limiting the risk, pending what the hazard exposure is.

It's much, much harder in the residential market, because you won't have the same insight in terms of individual risk prevention, measures being taken. So you can't really differentiate the premium as much, as much more trust relationship, and that makes it sometimes very hard to, to keep insuring.Brian: What are some pieces of advice that you would have for a business that is concerned about these risks and is wondering how they should address them?

Angelika: I would say the three elements to this: first, understand tomorrow's risk. Don't get stuck on how your risk has been in the past. Second, mitigate to the new normal. Don't be surprised. If you know there are changes, adapt to it. And now as I already said, adapt. Adapt to the changes, right? If you are expanding your operations, consider what is the environment you think is sustainable for the future to build a facility.

Brian: Thank you very much for being with us in Angelika.

Angelika: It’s been fun. Thank you so much, Brian.