Space Science and Engineering Center University of Wisconsin-Madison

With the 2022 Atlantic hurricane season approaching in June, forecasters at the NOAA National Hurricane Center are preparing to track and predict if, and where, storms will make landfall. They rely on a variety of tools to issue advanced warnings that aid emergency planners who may issue evacuation orders to the public. The goal is to save lives and property with these advanced warnings.

This year, a well-established, satellite-based method for determining tropical cyclone intensity, known as the Advanced Dvorak Technique, is receiving a significant upgrade within NOAA operations in time for the 2022 Atlantic season. Developed by researchers at the University of Wisconsin-Madison Cooperative Institute for Meteorological Satellite Studies, the ADT has benefitted over the decades from user feedback, especially from forecasters who rely on it to deliver more rapid and accurate analysis of storm development.

“For us the goal is to provide a tool so forecasters can do their jobs better,” says CIMSS Researcher Tim Olander. “We have always tried to support them by incorporating their recommendations and ideas for improving our algorithms.”

The original Dvorak Technique was first introduced in the early 1970s by NOAA Researcher Vernon Dvorak as a method to estimate hurricane intensity by analyzing visible and infrared satellite images of a storm. It was a labor-intensive process requiring manual interpretation and comparisons of images over time to detect common patterns, such as the intensity of the convection, cohesiveness of cloud bands and the development of the hurricane center, or “eye.” These estimates could take up to 20 minutes to construct by hand and proved to be less accurate when the hurricane lacked a well-defined eye.

Since then, advances in satellite meteorology have revolutionized the field, enabling forecasters to see hurricanes as they travel across the Atlantic Ocean or other tropical bodies of water. Weather satellites like NOAA’s GOES-16 capture visible and infrared imagery of a storm every 30 seconds. The increased frequency of the images, combined with modern computing and analysis techniques, has enhanced the accuracy of the Dvorak Technique.

CIMSS researchers took advantage of emerging satellite capabilities and developed the ADT in the late 1990s, building on Dvorak’s earlier methods. Use of the ADT has expanded beyond the NHC and it is now used to analyze tropical cyclones around the world, including those form in waters around Australia, Japan, India and Hawaii.

“The ADT itself is not a prediction tool, but it does help describe the current state of tropical cyclone intensity which provides the critical starting point for both forecaster and model-based predictions of future track and intensity,” says Senior Scientist Christopher Velden who leads the CIMSS Tropical Cyclone Group.

Hurricane forecasters gather in situ measurements from aircraft flying directly over a hurricane. By dropping probes, known as dropwindsondes, into a storm, scientists are able to collect real-time information like location, pressure, temperature and humidity. Combining these data with satellite imagery and the ADT gives forecasters an in-depth view of storm structure. Importantly, the ADT provides an indication of how a storm might strengthen – especially those approaching populated coastal areas.

Storms that develop outside of the Atlantic Ocean present a challenge for hurricane-hunting aircraft due to their limited range and inability to observe storms for extended periods of time. The ADT; however, provides an alternative way to continuously assess the state of a tropical cyclone wherever it exists on Earth. Because landfall preparations and evacuations are costly and disruptive, accurate forecasts, aided by the ADT can have huge implications for emergency planners who must decide whether to issue an order and for residents who must follow it.

Looking ahead, the CIMSS Tropical Cyclone Group will continue to improve satellite-based applications like the ADT to increase accuracy and reliability.

“The use of artificial intelligence methods to analyze large datasets is exploding in the research world, and our group is no exception,” says Velden. “We are already testing a new version of an AI-enhanced ADT that may someday replace the current version in operation. More accurate hurricane analyses contribute to improved warnings that lead to better public preparations for an approaching storm, ultimately reducing the negative impacts on coastal populations.”

This work is supported by NOAA