In the title song from Rodgers and Hammerstein’s musical “Oklahoma!”, the lyrics say “Oklahoma, where the wind comes sweeping down the plain. Where the waving wheat can surely smell sweet when the wind comes right behind the rain.” That is, unless it’s a supercell and then that wind could be destroying the wheat and everything else in its path. Welcome to the Historic Tornado Alley.
From the 1950’s to 1990’s an area that centered over south-central Oklahoma and northeast Texas became known as Tornado Alley because it was the area in the United States where most of the tornadoes took place. Flashforward 30 years and that area is moving east by 400-500 miles and is now centered over eastern Missouri and Arkansas, western Tennessee and Kentucky and northern Mississippi and Alabama. Climate change is pushing the dry line, the area where the moist humid air meets the cooler dry air that can cause severe thunderstorms and tornadoes, east.
Meteorologists and the National Weather Service feel that climate change factors such as rising temperatures, warmer winters, higher humidity and altered wind shear patterns are behind the movement of the area with the greatest development of tornadoes out of the plains and into the southeastern US. Not only are those factors changing the area of where more tornadoes form, but it is expanding tornado season from spring and summer to year-round. The shift east to areas that have not had years of research and preparation to protect their citizens from tornado outbreaks means that tornadoes are more dangerous and costly. The southeastern part of the United States has fewer storm shelters than the historic tornado alley of the plains and has more residents who live in mobile homes, both of which increase the chance of injuries and fatalities. Additionally, the area is prone to more nighttime tornadoes, which result in 2.5 times as many fatalities because of the warm moist air coming from the Gulf that continues after dark.
The need to accurately predict the likelihood of tornadoes forming to give residents early warnings to reduce the loss of life and injury is critical. The weather service has gathered data for years of tornado activity in Tornado Alley. Using that data and AI with predictive modeling, they can analyze incoming weather data quickly to identify subtle patterns that might be a tornado forming. AI and convolutional neural networks can be used to analyze doppler radar images to identify rotational and hook echoes that are key in predicting if a storm can form or is forming a tornado. AI can also create predictive modeling that identifies combinations of atmospheric variables like temperature, humidity and wind shear that can lead to the creation of tornadoes to monitor those factors and identify the areas that are at risk more closely. Connecting predictive modeling and analysis to emergency warning systems can help the weather service alert the media, emergency personnel and the residents of the risk that a tornado has formed or is likely to form to give them earlier warning to take shelter and be prepared to respond.
Insurers and local emergency and government personnel can use AI predictive modeling to identify critical pockets of the population who are at risk and the infrastructure for that area that, if damaged or destroyed by a tornado, would have a significant impact on the area, and rescue and recovery operations to be prepared to respond. Utilizing this data also helps residents have regular access to accurate information about storm formation, storm damage and rescue and recovery operations and when they can expect help from emergency personnel and insurers’ storm teams or FEMA in their area. Insurers use the information to help identify and coordinate their storm teams and claim operations to get help to their impacted insured customers in the area quickly. The quicker insurers can complete their inspection and get funds for their insured customers, the quicker the area can have the resources needed to start rebuilding and get back to their day-to-day life.
For large disaster areas that are not yet safe for emergency and non-emergency personnel to enter, the use of AI to analyze satellite imagery and aerial drone footage can help identify damaged buildings and infrastructure to prioritize rescue and recovery operations. With new deep-learning modeling that uses high-resolution satellite images, AI can classify different levels of destruction, including the types of damage. They can tell if the building has collapsed or if there is a partial collapse. It recognizes the geographic features in various locations to identify their impact on disasters and recovery. Emergency personnel and insurers can know the day after how many buildings are damaged and the extent of that damage. AI can analyze the images to provide faster and more accurate damage assessments to help with the allocation of resources more effectively. Insurers can use this data to help them identify impacted insured customers to start their loss estimates to get loss and extra expense funds into their hands to provide food and shelter and help with their recovery.
Post-disaster data helps with future pre-disaster storm preparations by identifying areas where the response was poor to improve rescue and recovery response, while recognizing what was done well. Emergency personnel and local governments can use geospatial data ahead of storms to help identify bottlenecks for recovery and rescue efforts and what infrastructure risks are in the area that will be impacted by a tornado. The use of real-time data of available resources vs. the needs and damage severity can help to optimize the use of emergency personnel and supplies and equipment and re-route traffic for recovery and rescue. AI messaging platforms and chatbots can help facilitate updates to residents and rescuers of the progress to get help to the disaster areas and provide all clear when the emergency is over. With AI, emergency personnel can proactively monitor and adjust rescue and recovery operations. AI using location-based data can help communities and insurers see how quickly people return to different areas to identify where funds and resources are needed now to start recovery and what areas will take longer for people to return and resources to be needed.
As part of pre-disaster preparation, AI helps to check data to look at changes in population trends and property values as well as environmental and weather changes to help analyze the level of risk facing insurers and governments to tornado damage in certain areas. It can help them identify and evaluate the projected loss of life, property, and economic and infrastructure disruption. It can also use that data to help with the allocation of assets and resources as the population and wealth shift from area to area while assessing the frequency and severity of disasters in that area. It also helps builders and cities to adjust building codes to rebuild homes that have tornado-safe rooms included.
The use of data and AI helps project the likelihood of a disaster to protect communities in Tornado Alley when the wind comes sweeping down the plain or the delta as the Alley moves east.
About the Author: Brenda McDermott, CPCU, CLP, SCLA, CIIP, SCLA, ARM, AIDA, AIC is a workers’ compensation claims specialist in The Hartford’s Major Case Unit. She is a past International Rookie, Claims Professional of the Year, Risk Management Professional of the Year and International CWC Speak-Off winner. She was the 2022 Region V Insurance Professional of the Year. She has been a long-time member of IAIP and served in multiple offices at the local, state, and regional levels. A Past Region V RVP she is currently serving as the Region 5 Marketing Director and Assistant to the RVP. She is Co-Chair of the International Marketing and Today’s Insurance Professional Committee. She is an MAL in Region 5 from Missouri.