Can we use technology to predict the weather?

Transcription:

Transcripts are generated using a combination of speech recognition software and human transcribers, and may contain errors. Please check the corresponding audio for the authoritative record.

Patti Harman (00:06):

Hello and welcome to this edition of the Dig In Podcast. I’m your host Patti Harman, Editor-in-Chief of Digital Insurance. The National Oceanic and Atmospheric Administration predicts an above average hurricane season, which could mean 17 to 25 named storms with eight to 13 of them expected to become hurricanes with wind speeds of 74 miles per hour or more. And forecasters also anticipate that four to seven of these storms could become major hurricanes. And we’re talking like a category three, four, or five with wind speeds in excess of 111 miles per hour. But it only takes one storm making landfall in a heavily populated area to cause millions in damages. So I’ve had some experience with hurricanes in Florida, South Carolina, Maryland, and New York, and I’ve covered damage resulting from some really ferocious storms. And I can almost guarantee that if I go on vacation toward the end of the summer, there will be at least one hurricane that makes landfill landfall somewhere along the East coast. I think I have like a 95% average for some reason. It’s really kind of funny. But here to talk about weather, hurricane forecasting and the technology available is someone far more qualified to discuss this topic than I am. Howard Altschule is the CEO and Certified Consulting Meteorologist for Forensic Weather consultants, one of the largest weather expert firms in the country. He works with carriers to confirm past weather data, to verify hail storms, high winds, storm surges and more. So thank you so much for joining us today, Howard

Howard Altschule (01:56):

Patti, thank you. It’s a pleasure to be here.

Patti Harman (01:58):

So let’s start with an easy question, which is why did you decide to become a meteorologist and what did that involve?

Howard Altschule (02:07):

I’ve wanted to be a meteorologist and have had this fascination with weather since as long as I can remember. So you’re talking what, seven, eight years old, nine years old, 10 years old. I often tell people, like right now I’m looking at autograph pictures of Al Roker and Willard Scott from probably 20 years ago, I was a weather watcher for Al Roker when he had hair, and that just gives you an idea of how long ago that was. Back in the 1980s, I was a weather watcher. I was the guy that had that weather station in his backyard measuring the rainfall amounts and the temperature, calling the TV station and talking to Al or Nick Gregory or WOR Radio in New York City, and then listening for my name on the radio. After that, I just had my nose peeled to the window, my eyes outside anytime snow would fall, anytime there was a blizzard and certainly thunderstorms were my big passion and still are probably my favorite thing.

(03:10):

So I’d be up at all hours of the night watching thunderstorms, giving forecasts to the point where I earned the nickname Hurricane Howie back in junior high school. And then in high school did the weather on the local cable TV station. So I think that’s where this was all born from. And I ended up following my passion and going to college at the State University of New York at Albany, which is in upstate New York, and that’s only maybe like five miles down the road from where our corporate headquarters is right now for Forensic Weather Consultants. And I obtained my degree, a bachelor of science degree in meteorology or atmospheric science back in 1995. So quite a long time ago. And I’ve been in the business since then and had many different weather jobs, or not many, just a few. Some pretty glamorous where I was forecaster for the Mets and the Yankees, movie producers and things like that. And I also had about eight years on TV at NBC as an on-camera meteorologist. So I mean, the passion has been there up until today and it continues to be there both professionally and also as a hobby, and it always will be a hobby no matter what, even on the day I retire.

Patti Harman (04:29):

So how did all of that training help you with the weather and the storm tracking that you’re doing now?

Howard Altschule (04:36):

Well, SUNY Albany is one of the best universities in the country for meteorology. And even back then it was one of the top. So we had excellent professors and excellent classes and coursework, and one of my first jobs out of college was at a company down in Southeastern New York where I did a lot of forecasting as I mentioned, but they also said they needed some help in the forensic division, and this was the mid-1990s. I didn’t know anything about forensics back then, but they said that they would train me and I’m like, okay, I’ll help out brand new out of college my first year. So I started learning about forensics there and started learning about what they did, and I was trained minimally back then about what to do. And after that, I left that company, went to work for the TV station and the TV station back then as a number four guy on the totem pole, didn’t pay a very good salary.

(05:35):

And I was getting up at 3:30 in the morning every day to go get ready for the morning shows. And I was like, it’s going to be something better. I need to supplement my income. So I started my business, Forensic Weather Consultants, in 1999, 25 years ago this year, this September. And I started out working on one case every four months where I literally opened up the yellow pages, typed up a letter to all the attorneys in Albany, got one case every four months, reinvested in mailings and advertising. And now we’re probably about seven or 800 cases a year that we’re working on here with my team of eight other forensic full-time meteorologists.

Patti Harman (06:19):

Wow, that is just a fascinating background. I didn’t realize all of that about your company. Thank you so much for sharing that. How has the technology that you use to track weather changed over the years? And I guess what everybody would like to know is has it increased the degree of accuracy for predictions at this point?

Howard Altschule (06:40):

Yeah, no, it definitely has. When I was talking about doing the weather on TV back in high school and other meteorologists or weather people like Willard Scott, back in the day, we used to put magnet-like cloud figures and Thumbtack raindrops on a corkboard showing where the weather’s going to be. So over the decades, technology has really advanced and it continues to just explode with new technology and advancements with artificial intelligence and machine learning and predictive modeling. But one of the biggest tools we have that goes towards severe weather tracking, and not only forecasting, but also going back and looking at hail claims and stuff like that, but also for hurricanes is Doppler radar. There’s Doppler radar imagery that ingests all the information. We have satellite imagery that we never had anything like before. So satellite images go into the computer models, especially ones that forecast hurricanes.

(07:41):

There’s weather sensors or weather data on airplanes, commercial aircraft now. So like Southwest Airlines or United, some of those, or many of those have weather sensors that transmit data to NOAA. We have weather balloons that go off. You have UAV drones that are being flown. You have sensors in the Atlantic Ocean and the oceans that are just kind of floating there as buoys and all of that data, I mean, there’s weather sensors in parking meters in some places and on electric vehicles. So all of that data goes into computer models, new high resolution computer models that people are generating, and that goes toward forecasting hurricane forecasts primarily, and the most important because that could have the biggest devastation and the most significant impact on life and property over a very large scale. So those are just some examples of how technology has really, really improved over the past, not only several decades, but I mean especially over the past three or four years.

Patti Harman (08:48):

Very true. We’ve seen at Digital Insurance, where we cover that intersection between technology and insurance, and it’s just been absolutely mind boggling to me how much technology has changed so many different facets of the industry. So I mentioned earlier that NOAA is predicting an above average Atlantic hurricane season. What kinds of weather developments do you and other meteorologist track then to see where and when these storms might originate?

Howard Altschule (09:19):

Yeah, so using climatology and past weather data from the past 15, 20, 25 years, we can track things like ocean temperatures or El Nino or La Nina, which is a change in the global circulation patterns. And in the Pacific Ocean, one means above normal water temperatures. The other when that happens is above or below normal water temperatures, and depending on which one occurs, changes where thunderstorms develop, where the winds are stronger and basically changes the whole global jet stream and circulation pattern. So when you’re in a La Nina pattern, that has a tendency to mean a much more active hurricane season for the Atlantic, the Caribbean, and the Gulf of Mexico. So all of that technology and being able to track the water temperatures and now these computer models and using artificial intelligence going back year by year by year, saying, okay, when the water temperature was the way it is now in 2024, what happened in that same type of scenario?

(10:29):

And we can gauge how busy or active we expect the hurricane season to be based on that or vice versa. Now for this year, all of the correlations are saying that we’re going to be in the La Nina pattern, which usually almost always means above normal or well above normal hurricane or tropical storm activity. But what we have now going this year is near record ocean temperatures in the Atlantic and the Gulf of Mexico. So now when you have an atmosphere that favors development of tropical storms or hurricanes combined with way above normal water temperatures, ocean temperatures, which is the fuel for tropical cyclones, all the indications are that, yeah, we’re going to have a lot more tropical storms and hurricanes than usual, and they could be quite fierce. And just the way history has shown the steering patterns suggests a higher threat of maybe tropical cyclone or a hurricane hitting the East Coast of the U.S. or Florida or the Gulf Coast or kind of the Caribbean Islands.

Patti Harman (11:37):

Wow. Yeah, just even with the storms that hit Florida a couple of weeks ago, it was amazing how much rain came down in just such a short period of time. How are you using technology to monitor the weather before and after an event like a hurricane or a sleet or ice event? I know you can confirm specific types of weather events, but how do you actually track and confirm that they either did or did not occur?

Howard Altschule (12:06):

Yeah, so that is why we are so busy at Forensic weather Consultants, as I mentioned, we’re 10 times busier than we’ve ever been, Patti. And I think that’s because of the outreach I’ve been doing, speaking at conferences, going to a lot of insurance carrier headquarters or regional offices or doing webinars about forensic meteorology, and they’re seeing the value in what we do. They’re seeing, wow, we only thought you could look at the airport wind speed or an automated report, which has problems quite often. We didn’t know that this type of data was available. So when I give these presentations, and in New York City just yesterday, I’m giving one of these, I showed all these different types of weather records, all these different types of stations that most claims adjusters or attorneys or engineers aren’t even aware of. And I showed how we use them. Like doppler radar, one of the biggest things, what we do is we plot a specific property or loss location, right on the radar images.

(13:13):

Those are processed every one to five minutes, and we can track precipitation. We can see if it rained at the 7-Eleven, 10 miles away from the airport, but maybe not at the airport itself. We can see if it hailed at your house, but maybe not a mile away. And I think history has been such that a lot of adjusters and engineers, they can only use what’s being reported by NOAA or what the winds were at the airport. And quite frankly, the weather changes dramatically over short distances. So we’re able to answer those questions. Now, technology with Doppler radar has really advanced in the past 10 years or so. We have things called dual pole radar products, which not only does it show you what the weather’s like and where it’s raining during the weather on the TV news at night, but we can now see whether there’s hail in the thunderstorm or whether it’s just rain or whether it’s a mix or whether it looks like it could be really large. And we corroborate that with storm reports on the ground and other sources of information. So radar can detect tornado debris, so we can actually confirm sometimes if there’s a tornado occurring by the radar picking up debris, 15,000 feet of the air. It’s fascinating stuff.

Patti Harman (14:36):

Wow. It really is. I’m based in Maryland and we had, I think 10 tornadoes come across the state a couple of weeks ago, and it was just really interesting to kind of watch the weather forecasters talking about this is how we confirmed that it hit in these areas and all of that sort of thing. So it’s just really so interesting. So we’ve seen the frequency and the severity of storm events kind of grow over the years, the hundred-year and we hear about the thousand-year flood events, what do insurers need to consider when they’re writing coverage in these areas. And I’m wondering, can you really assume that a hundred-year flood really won’t take place for another one hundred years? I’m thinking locally here in Ellicott City, I think within a two or three year span, they had two, 100-year flood events. So as a carrier, how did they even begin to anticipate what kind of coverage or how severe some of these events could be?

Howard Altschule (15:41):

I think I worked on one of those flood claims in Ellicott City, actually, because I remember when they occurred maybe 10 years ago or something like that.

Patti Harman (15:48):

Yes.

Howard Altschule (15:49):

Yes, it’s a small world, right? It’s interesting because the weather is definitely changing. There can be a lot of people arguing that it’s global warming or climate change, and it is to some extent. There’s no doubt about that. And I think we all agree the weather is changing and getting worse, and I think there’s new climate models, there’s new updated information and better sources of data that’s being compiled so that insurers can see what kind of risk there is. For instance, we have an automated hail swath program that pretty much every hour tells us what the Doppler radar was showing hail size to be anywhere in the country and Canada. So you can go back and use that data every day. And we have probably eight, nine years of data every single day. And you can see what areas have had hail and where the radars have shown there to be hail.

(16:47):

So they could base their risk on, alright, what areas had minimal hail, what areas might’ve seen more hail than they’ve ever had before or rainfall. And in terms of those one-hundred year storms, because it’s warming up, that helps with evaporation and more moisture in the atmosphere and bigger storms, more powerful storms. So those hundred-year storms are happening more frequently. So I think using the more recent data, whether maybe not a 20-year historical data set, but maybe that in conjunction with maybe a five-year, five or 10-year historical data set might capture some of the more frequent recurrence intervals of that. So I think there’s a lot of information out there, and carriers I think really need to consult with forensic meteorologists and climate modelers to understand, all right, what data’s available, how can we model this? And account for the changing weather.

Patti Harman (17:50):

That really helps them in terms of their underwriting accuracy then and kind of figuring out what could be. So that’s great. Thank you. So what factors can affect the trajectory of a storm? I know we’re always watching when storms are coming in and you hear the forecast and then all of a sudden it’ll go out to sea or further inland unexpectedly. What kinds of factors can affect a storm then?

Howard Altschule (18:17):

Yeah, one of the biggest problems, Patti, is that there’s very limited data in the oceans to take measurements, to take measurements of temperatures or air pressure. That’s the information that gets ingested into the NOAA computer models that run and forecast where we think a storm will go. So I know there’s a lot of research and new grants trying to put more things out there in the ocean to gather that, and that’s what the hurricane hunter aircraft do as well. So things that affect the steering patterns of a storm, that’s a big one, but also where the exact high pressure center is over the Atlantic Ocean can make a huge difference. For instance, and let’s just say a computer model doesn’t pick up a cold front that’s coming through the Great Lakes and maybe it’s 12 hours faster than the computer models are thinking, what happens then is it can break down what we call this blocking high pressure and allow for a storm maybe to recurve earlier than what everybody thought or vice versa.

(19:25):

Maybe that cold front or trough as we call it in meteorological terms, is weaker than the computer models think, and that blocking high pressure over the Atlantic will as a result be stronger and that will tend to keep the storms moving further west and not recurve out to sea. So those are just a couple of things that we always watch, and it tends to be a big issue, especially with those cold fronts coming through. I always watch and I’m like, well, I’m going to watch because that looks like it’s slowing down and if that slows down, the Bermuda High is going to be stronger and keep this on a further west track. And that was the perfect example with Irma, right? It kept on shifting further west, and I saw it doing that, and we updated our forecast for our big telecommunications company clients. And other thing is atmospheric dust, right?

(20:19):

The Sahara dust layer coming off the African coast. So that tends to inhibit tropical development. So if there’s a lot of dust being picked up by satellite imagery, that would tend to suppress hurricane formation and tropical storm formation. If the models don’t pick up on that and they don’t forecast where it’s going to be in five days, that could have a big implication on how a storm strengthens or where it’s going to go. So those are just a couple of reasons why it’s so difficult to forecast hurricanes sometimes and really why people need to err on the side of caution. If you’re in that cone, you really, really need to pay attention. And if you’re in an evacuation area, you know what? Heed those warnings, evacuate, go to a safe place because once you are convinced that storm surge is going to hit, it’s going to be too late in a lot of situations.

Patti Harman (21:11):

Very true. It’s fascinating to me how many different types of, I guess, factors within the atmosphere, even things halfway around the world can affect what happens here in the U.S. It surprises me every time and I think, oh yeah, I should be watching that. And sure enough, in a couple of days it can really have a major impact. So do you think that technology is going to continue to make it easier to track or predict where storms will go or will there always be a certain element of surprise even with more data and more technology, those sorts of things?

Howard Altschule (21:52):

I don’t think we’re ever going to be smarter than Mother Nature. There’s a lot of research going on, a lot of money going into modeling, forecasting, computer models and AI and machine learning. I think we’ll get better and I think it will get better. I don’t think it’s ever going to be perfect or nearly perfect. We like to say a three or four- day forecast just in general is usually a good forecast just for the weather. But once you get out to six or seven days, or there’s a company that I think has a 45-day forecast that has a public very, very, very well-known forecasting company. And that’s a joke that’s just based on a climatology and models. And if they get it right, they’re lucky. But a 3, 4, 5-day forecast, we should start to get better improvements, especially as more money goes into research and we get more of these sensors which go into computer models and more high resolution computer models.

(22:54):

And the big thing, like you mentioned earlier, AI and machine learning going back and looking at, okay, it’s hot and humid out here today, what happened with the same exact situation in the past however many times. And then the computer model can look at that in half an hour and then give us some scenarios about what most likely will occur that should get better as AI continues to improve and we get a better understanding on that type of thing. So it’s really fascinating and I think that’s where the industry’s going. But to your point, I don’t think it’s ever going to be perfect.

Patti Harman (23:30):

Okay. I didn’t think so. But then what do I know?

Howard Altschule (23:35):

I can’t say it will be or else I might be out of a job one day. Right.

Patti Harman (23:40):

No, I don’t think AI will ever be smarter than you are with all of this.

Howard Altschule (23:44):

No. Well, thank you. That’s a nice testament. The check’s in the mail, Patti. Thank you.

Patti Harman (23:49):

So we could probably talk about weather and its implications for hours, but is there anything that I haven’t asked you that you think is really important for our listeners to know about weather or tracking storms or preparing for worst case scenarios?

Howard Altschule (24:05):

I think it’s really, like I mentioned a few minutes ago, it’s really important to pay attention to that National Hurricane Center cone of uncertainty. And there’s a lot of webpages out there, a lot of information out there. There’s a lot of armchair weather forecasters out there. And I guess that just popped into my mind for everyone listening, Joe Blow, Mid-Atlantic Weatherman has his Facebook page or whatever it might be, or Twitter, and they’re making their own forecast. A lot of those people have no experience or background education in meteorology. So I strongly advise people to just trust your sources, trust NOAA, trust reliable meteorologists, people that don’t sensationalize everything, people that don’t put up a picture of a 15-day computer model showing a hurricane hitting Miami because it’s likely to change and the general public doesn’t know that. So they get very scared and they get these people who are just doing it for clicks, doing it for notoriety and maybe money if that’s their full-time business.

(25:14):

So insurance carriers I think really need to invest in a good meteorologist or a meteorology company that specializes in forecasting and knows what they’re doing. I mean, we do that, Patti, we have arguably the largest telecommunications company in the world as our client, and we do hurricane forecasting and storm surge forecasting for them for the whole United States and the Caribbean and Hawaii and the West coast. And we’ve never gotten it wrong. All of their assets have been protected. They have prepared, like we told them to, so storm surge never affected their infrastructure. So really, if it’s important or if you’re an insurance company and you want to triage resources appropriately, a lot of people are turning to private companies like us for triage forecasting, even private corporations or insurance carriers. They can reach out to people like us, forensic weather consultants or whomever sign a contract, and they’ll get very frequent updates about what the hurricane forecast looks like, what we think is going to occur in the tropics, what we think is not going to occur, and maybe why we disagree with the National Hurricane Center.

(26:27):

They’re limited to what they can put out in public, we’re not. So we will give a lot of scenarios and we can educate and inform people to make a decision and prepare properly so they’re not caught off guard. So that’s really I think something that’s very important that your listeners always consider no plan A or scenario A, scenario B and scenario C, so that you can triage your resources, be prepared and not get caught off guard, whatever that means for them. I think that’s also where I think where the trend is going, more information from reliable sources helps these people make a reliable decision in terms of where to send adjusters or their trucks and things like that.

Patti Harman (27:11):

Very true. As a journalist, I totally agree about the importance of having reliable sources.

Howard Altschule (27:16):

A hundred percent, yes. I knew you’d like that.

Patti Harman (27:21):

Well, thank you so much for joining us today, Howard. This was just really fascinating.

Howard Altschule (27:26):

It’s my pleasure. Always a pleasure to speak with you and see you, Patti, and write articles with you. So thanks so much for inviting me, and hopefully we can do this again on some other topic. And maybe at the end of hurricane season, we’ll talk about ways for adjusters or insurance carriers or attorneys or whoever to identify what caused the damage to a specific property now that there’s potentially going to be five different storms in the past couple years, and it’s important to know what damage occurred and when. So maybe that’s a topic for another time.

Patti Harman (28:00):

Oh, yes. I will definitely keep that in mind.

Howard Altschule (28:04):

Well, thanks for having me, Patty.

Patti Harman (28:05):

Thank you. And thank you for listening to the Dig in podcast. I produced this episode with audio production by Adnan Khan. A special thanks this week to Howard Atschule of Forensic Weather Consultants for joining us. Please rate us, review us, and subscribe to our content at dig in.com/subscribe From Digital Insurance, I’m Patti Harman, and thank you for listening.