IDSC is pleased to welcome Michael Scott Fischer, IDSC Core Faculty Member and Earth Systems Science Assistant Professor in the Department of Atmospheric Sciences at the Rosenstiel School of Marine Atmospheric, and Earth Science.

Story by Alan Gomez

Michael Fischer was too young to remember Hurricane Andrew tearing apart his family’s home south of Miami in 1992, but he remembers the destruction the Category 5 behemoth left in its wake. He remembers the chainsaws and hammers pounding away as the region recovered from the historic storm. And he remembers wondering how wind and rain could cause so much damage. That curiosity quickly morphed into Fischer’s passion, leading him on a lifelong quest to better understand hurricanes, why they form, and how they intensify so quickly.

In his younger years, that meant eagerly waiting until 50 minutes past the hour to watch the tropical forecast and charting incoming storms on a tracking map taped to the door of the family laundry room. Now, Fischer is armed with a Ph.D. in atmospheric science and has joined the University of Miami as a joint appointment between the Frost Institute for Data Science and Computing (IDSC) and the Rosenstiel School of Marine, Atmospheric, and Earth Science.

“As a kid, it was a curiosity, but now it’s totally evolved into my research,” he said. “My goal is that by improving our foundational understanding of what drives these systems and the strength of them, that the research that comes from that can provide tools that can generate better, more accurate forecasts, and that process can help save lives.”

Prior to joining UM, Fischer spent years at the National Hurricane Center and NOAA’s Hurricane Research Division, where he focused on understanding the phenomenon of hurricanes that intensify rapidly with little warning. Ben Kirtman, the Deputy Director of IDSC and Director of the Cooperative Institute for Marine & Atmospheric Studies (CIMAS) at the Rosenstiel School, said he’s excited to see what happens when Fischer has full access to the expertise and computational power at UM.

“His work in understanding how the storm structure and the large-scale environment affect the potential for rapid intensification is truly groundbreaking.”

“His work in understanding how the storm structure and the large-scale environment affect the potential for rapid intensification is truly groundbreaking,” Kirtman said. “As Professor Fischer brings the data-driven approach from machine learning with IDSC data science experts to the problem, I fully expect new breakthroughs in our ability to predict rapid intensification.”

Fischer has approached the challenge of understanding hurricane intensification in several ways.

While at the National Hurricane Center* in Miami, researchers realized he had a knack for computer programming and asked Fischer to develop a way to help forecasters get a more detailed look at storms in real-time. At the time, researchers had access to only flat, two-dimensional snapshots of each oncoming hurricane, much like seeing individual images of an MRI scan. Fischer was able to take predictions from state-of-the-art hurricane models and convert them into an interactive, three-dimensional visualization that forecasters could explore with their mouse, turning and twisting and zooming in and out of each storm.

“These packages were initially used more so in a research setting,” Fischer said. “This project aimed at being the first time the hurricane specialists were really getting a 3-dimensional visualization into their workflow.”

Fischer has also looked to the past to understand the hurricanes of today. To do so, he focused on Hurricane Irma, a 2017 storm that was best known for the evacuation nightmare it caused as people drove east, west, and north to escape the storm charging up from the south. But it was Irma’s two periods of rapid intensification (including a surge where it grew from a Category 3 to a Category 5 in 24 hours) that made Fischer want to analyze other storms that experienced similar intensifications. The problem was finding the data.

Fischer said that radar data is invaluable, but it was all stored in “clunky binary files” that would take hours to process. He did just that to produce a widely cited research paper published in Monthly Weather Review[1] in 2020 analyzing Irma’s rapid intensification. But that agonizing process convinced Fischer that there must be a better way, so he wrote a new program to extract, organize, and simplify the data to be more accessible and understandable for researchers and forecasters. And then, he decided to make that database public, allowing any researcher to use it.

“My goal is not to have my research out there,” he said. “My goal is to improve our understanding of tropical cyclones, and to do that, the best way is to make this data readily available to everyone in a very user-friendly format so we can help achieve that goal as quickly as possible.”

Now that Fischer has started his joint appointment with IDSC and the Rosenstiel School, he hopes to advance his research in two key areas:

• One will expand on work he’s been doing with researchers at Colorado State University to train computers to better identify the data gathered from each storm. Currently, researchers manually sort through each data point to make sure it’s valid (a process that can take weeks) or use an automated but conservative method that often throws out good data that is confused for background noise.

• Fischer’s other main project will be creating a 3D analysis of a hurricane when only minimal data is available. All tropical cyclones that target the United States are closely monitored by satellites, airplanes, and ground-based systems, but cyclones in other regions of the world are usually only tracked by whichever satellites happen to be passing overhead.

Fischer believes he can train a computer to combine those satellite observations with historical data to produce an accurate 3D representation of each storm. That would not only help people in remote areas prepare for oncoming storms, but continue his quest to finally figure out how hurricanes do what they do. And he hopes his new position at UM will help him get there.

“My formal training was in atmospheric science and I dabbled in machine learning and data science stuff I learned along the way,” Fischer said. “But this is what they do. Becoming well versed in the emerging capabilities of AI by working with the folks at IDSC and the computational capabilities there will not only benefit my goals of trying to improve intensity forecasts but open additional avenues to improve our understanding and prediction of the atmosphere.”

---

1  Fischer, M.S., R.F. Rogers, and P.D. Reasor. The rapid intensification and eyewall replacement cycles of Hurricane Irma (2017). Monthly Weather Review, 148(3):981-1004, https://doi.org/10.1175/MWR-D-19-0185.1 2020  |  Ref. 3763