Title

Early Warning Signals for Impending Hazards: A Data Driven Approach

Bio

Dr. Gopalakrishnan completed his PhD from Indian Institute of Technology Madras in 2016. His PhD thesis focussed on developing robust early warning signals to predict impending transitions in a bi- stable oscillator. His areas of interest in research are Complex System Theory, Early Warning Signals, Data Driven Modelling & Analysis, Time Series Analysis, Artificial Intelligence etc. He has published research articles in journals such as Nature Scientific Reports, Journal of Fluid Mechanics, Physical Review-E, Chaos, IEEE Transactions, Expert Systems with Applications etc. His article on early warning systems was cited by the office of the British Prime Minister in a policy document.

Abstract

Early Warning Signals for Impending Hazards: A Data Driven Approach Gopalakrishnan E A, Center of Computational Engineering and Networking, Amrita Vishwa Vidyapeetham

Geosystems can be modelled as complex systems as they have many interacting subsystems such as atmosphere, biosphere, hydrosphere so on and so forth. These systems interact with each other in a nonlinear fashion and their interaction may lead to geohazards. Thus, the geohazards can be viewed as catastrophic transitions that happen in a complex system due to self-organisation or emergence. Such tipping points which result in undesirable consequences need to be predicted. Universal measures which precede the catastrophic transitions can be employed to predict the impending hazards. We can calculate the early warning measures from the enormous data that we obtain through the sensor networks employed. Further, we can derive complex networks from the time series data that we obtain. Data driven methods such as Dynamic Mode Decomposition can be employed to analyse the dynamics of geosystems. We can also employ AI/ML techniques to develop robust early warning measures to detect the impending geohazards.