MBA 2020 Tech All-Star Howard Botts: Meet the Wizard of ‘Weather Forensics’
Howard Botts could be in Los Angeles and tell you within 15 minutes of a tornado touching down in the Midwest, if any, properties sustained damage and the severity of the damage.
He’s not psychic; he’s scientific.
Executive Leader of CoreLogic’s Science and Analytics Center of Excellence in Irvine, Calif., Botts is a pioneer in computer mapping and geographic information systems technology, with broad expertise in natural hazard modeling.
He’s been recognized as a 2020 MBA Tech All-Star thanks to the impact his intelligent computer models have had on housing finance, insurance and industry professionals’ understanding of location-based risk, down to a specific house, and how that risk will change over time.
Botts earned his Ph.D. in Geography from the University of Wisconsin-Madison and founded Proxix Solutions, a geospatial information provider, in 2003. His company built very granular hazard risk maps, using three-meter grid cells (about 10’x10’) up to 10 meters (30’x30’), showing property boundaries overlaid with data about weather, terrain and more.
“With these maps, risk could be assigned to an exact location,” Botts explains, “and we could get an extremely accurate view of risk to that property from wildfires, floods, hurricane driven storm surge and about 40 other hazards.”
CoreLogic acquired Proxix in 2007 and Botts became Chief Scientist, leading a team that focuses on how natural hazard risks equate to risk in the real estate environment, affecting everything from home prices to likelihood of foreclosure or default after a natural disaster event. His team works with lenders and servicers to understand portfolio risk. For instance, if a lender has homes in California, where only 12% of homeowners have earthquake insurance, they’ll want to understand their exposure and get reinsurance if needed.
The team also works with servicers after major events, such as Hurricane Harvey, Irma and Maria, to show them which properties are likely to be flooded and the degree of severity. “That lets the servicer respond right away to just the affected households,” Botts says.
How GIS Models Work
GIS models combine unique types of data about a location, depending on the type of risk being assessed. For example, a wildfire risk model starts with a satellite image of vegetation around the property to tell if a hillside is covered in Douglas fir, mixed Douglas fir/aspen or chaparral — each of which offers varying fuel load characteristics.
Other data layers are then added, such as a digital terrain (topo) map to determine slope and aspect (facing direction of the slope) because fires spread faster on steeper slopes, and south-facing slopes are dryer and more likely to ignite. Other layers of geo-spatial information that can then be added include historic frequency of wildfires, weather, wind patterns, road network and urban-wildland interface.
Each separate geospatial layer is weighted and combined to produce a highly granular wildfire risk map with a structure or property-level risk score.
It’s ground-breaking work that saves money over manual site inspections and gives insurers and banks a more complete understanding of their risk exposure. Botts earned his first two patents in 2013. Since then, he has secured three additional patents for risk models involving flash floods, sewer backup and basement flooding and landslides.
Currently 15 of the top 20 insurers are using these models for pricing, risk mitigation, presentation to regulators and for reinsurance placement. The CoreLogic California Earthquake Model developed by Botts’ team is the only CAT (Catastrophic risk) model approved by the California Earthquake Authority to estimate losses and price policies.
Real-Time ‘Weather Forensics’
The team uses data from NOAA weather radars across the United States–familiar to most Americans by its use for TV weather news. These radars send a pulse of electromagnetic energy into the atmosphere to detect precipitation, calculate its motion, intensity and determine the type (rain, snow, hail, etc.) to identify property damaging events such as large hail or intense rain events likely to cause flash flooding.
Weather radar data allow the team to see what’s happening in the atmosphere and on the ground, so they know where there are rotating winds, which may become tornadoes on the ground. “That tells us exactly what the footprint of that tornado is within 15 minutes of the event,” Botts says. “We know in real time which properties have 100% damage, down to 10% damage.”
Then, they use machine learning and AI technology to text-mine and imagery-mine social media data for real-time on the ground validation. “So as people post information and photos, we can validate our findings.”
Since 2014, Botts has presented at more than 50 conferences across the country on a variety of topics, including hazard database development, understanding portfolio natural hazard risk, market potential models, and geographically based market analysis. These conferences include the Mortgage Bankers Association Risk Management Conference, EPIQ and Single-Family Rental.
Although CoreLogic is a B2B company, Botts is keen on ensuring consumers have information on natural disasters and risk. With this in mind, he assisted with development of the CoreLogic Hazard HQ website – a publicly accessible risk information resource center that offers individuals, media and companies a high-level analyses and up-to-date data insights on the immediate risks that natural catastrophes pose to properties across the country.
“The site was designed to drive visibility of natural catastrophe risk as the economic impact of natural catastrophes increases, and to act as a home base for all insights pertaining to these risks,” Botts says. He also shares his knowledge with a broader audience through multiple annual reports, including the Storm Surge Report and the Wildfire Risk Report.
One important new focus is determining impacts to homes as the earth warms and sea levels rise.
“Our clients are asking ‘What’s it going to look like 10 years, 20 years, 30 years from now?’” Botts notes. “Climate change is going to affect everything from home prices to likelihood of foreclosure or default after a natural disaster event in many areas. Already in some areas of Florida we see a decline in home prices for areas that get tidal flooding during king tide events, where you get super-high tides which will flood streets and make whole areas relatively inaccessible for a short period. We know that’s only going to continue.”
“We’re trying to understand what the future holds because when you insure, you’re insuring for years at a time,” Botts adds. “When you get a mortgage, you may be looking 30 years into the future. So, understanding what that future looks like is important for the housing finance industry and homeowners.”