TROPHY Framework for Tropical Cyclone Tracking

Scientists in EVS introduced TROPHY, a new tropical cyclone (TC) tracking framework that is a useful tool for understanding and improving the future projections of TCs.

Tropical cyclones (TCs) are the largest drivers of losses among natural hazards, bringing wind gusts, high waves, storm surges, and heavy rainfall. Studies and simulations have found the intensity of TCs is increasing over the North Atlantic. Realistically and efficiently detecting and tracking TCs is critical for improving weather forecasting and Earth system analysis to better issue early warnings.

Knowing the location and the movement of the TC’s eye (a region of mostly calm weather at the center) precisely and in a timely manner is crucial. Errors in forecasting tracks could have negative impacts on the areas under watches and warnings.

Dr. Lin Yan and Jiali Wang from Argonne’s Environmental Science Division (EVS), in collaboration with Argonne’s Mathematics and Computer Science Division (MCS), Ohio State University, and the University of Utah, introduced a new TC tracking framework, called TROPHY (topologically robust physics-informed tracking framework). Its main goal is to integrate physical knowledge (such as duration and impact radius) of TCs to drastically improve tracking efficiency for large-scale climate datasets.

What TROPHY Can Do

Example of topological-based visualization. The left image shows wind vector fields identified as a cyclone; the right image is not a cyclone.

TROPHY is different than many TC trackers. It tracks the eyes instead of the impact areas. Other trackers use multiple variables at different altitudes (e.g., minimum sea level pressure or maximum rotation) as the basis to identify a TC candidate. TROPHY only requires near-surface wind speeds and directions.

Additionally, TROPHY can capture TC characteristics, such as frequency and intensity, that are comparable to and sometimes even better than a tracking algorithm that requires more data inputs. TROPHY can even detect TCs at least partially, during their beginning and end periods, even if their wind speed is low.

A More Efficient TC Tracker

Several elements of TROPHY make it a more efficient way to track TCs compared with another topology-based tracking framework. For instance, TROPHY can filter out features that cannot be considered as a TC when it comes to physical properties such as duration and intensity.

TROPHY calculates the topological properties of a detected TC using the radius of its impact region. A physics-informed calculation strategy builds a bridge between TCs and vector field topology. This makes TROPHY meaningful in understanding real-world scenarios.

Outcomes

TROPHY provides a useful tool for understanding and improving the future projection of TCs. Efficient tracking can help assess impacted areas and provide risk assessments for public and critical infrastructures.

The paper on the TROPHY framework was accepted to the 2023 IEEE Visualization Conference. This top conference for the visualization community only has a 20-30% acceptance rate. The paper is also being published in the top journal of the visualization community, a special issue of the IEEE Transactions on Visualization and Computer Graphics (TVCG).

Annually detected tropical cyclones by observations (left), a well-validated tracking algorithm called TempestExtremes (middle), and TROPHY (right) for the year 2004, as an example. Tracks are colored by maximum wind speed.