A team of researchers from Indian Institute of Technology Madras (IIT Madras), IIT Hyderabad and Potsdam Institute of Climate Impact Research (PIK), Germany have developed a novel approach to predict merging tropical cyclones as tropical cyclones ravage coastlines across southern Asia, Australia, and the Americas.

This approach, according to a statement from IIT Madras, is based on the interdisciplinary methodology of complex networks in an article titled ‘Study of Interaction and Complete Merging of Binary Cyclones Using Complex Networks’ published in Chaos: An Interdisciplinary Journal of Nonlinear Science.

As one cyclone starts affecting the other and vice-versa (this interaction is referred to as “Fujiwhara interaction” by meteorologists), there can be number of possible outcomes — their trajectories and/or strengths may suddenly change, or they can merge to form a much stronger cyclone.

Such binary interaction of cyclones has neither been completely understood nor fully incorporated in weather prediction models, and therefore leads to erroneous forecasts. Such inaccurate forecasts increase the threat to life and property due to unpreparedness caused by misinformation and the lack of early warning.

According to Prof RI Sujith, Department of Aerospace Engineering, Indian Institute of Technology Madras, analyzing cyclone interactions using the novel framework pioneered in this study has the potential to improve the accuracy of the early warning signals provided by meteorological organizations to the government so that they can take pre-emptive and early action to reduce the impact of such disasters.

A complex network encodes the pattern of interaction of a complex system, and can be directly applied to study the Fujiwhara interaction between two cyclonic vortices, said a statement released by the Chennai based institute.

Indicators derived from this methodology were found to clearly distinguish the different stages of mutual interaction between two cyclones and provide an early indication of cyclone merger, often better than conventionally used indicators such as the separation distance between two cyclones.

Dr. Somnath De, the lead author of this study, said “We believe that this network-based approach can be used to study binary cyclone interactions from observational or model-based relative vorticity data to obtain better insights on the possibility of cyclone merger. It paves the path to analyze such highly unusual/rare events in which sudden alteration of cyclone tracks or re-strengthening occurs, on a case-by-case basis, and facilitate improved prediction of cyclone tracks and the fate of such interactions”.

Dr. Vishnu R. Unni from IIT Hyderabad added, “Data-driven methods for prediction of extreme weather events have a unique advantage since they allow one to identify critical patterns in the evolution of such weather events that are elusive to traditional methods.”