Skip to content
  • Home
  • Thesis Topics
  • Modelling Tropical Cyclones for Disaster Risk Management with Deep Learning Approaches

Research Thesis Topic

Modelling Tropical Cyclones for Disaster Risk Management with Deep Learning Approaches

Topic ID:

Thesis Topic/Title:
Modelling Tropical Cyclones for Disaster Risk Management with Deep Learning Approaches


Tropical cyclones are natural disasters whose tracks or intensity are challenging to predict. Physical and dynamical models are gold standards in such predictions, but a lack of reliable forecast models can have a devastating impact on people’s lives and infrastructure. Data-driven approaches can help in the design of cyclone forecasting models to complement physical methods and help in disaster risk management.

This project aims to adopt deep learning artificial intelligence methods to develop reliable forecast models, to predict the formation, tracks and intensity of tropical cyclones. Deep learning refers to machine learning (or computational) algorithms that can identify antecedent (past) features in covariate data to model the future property of a target variable (e.g. a cyclone path, intensity, etc). Deep learning has been extensively used in climate and weather forecasting due to its capability to model nonlinear and relatively complex relationship of many atmospheric variables. The project will develop data-driven approaches to forecast tropical cyclones, provide a higher accuracy and reliability advantages, adequate lead times and more certain characteristic features that can lead to user confidence and improved disaster mitigation. Such methods can lead to innovations in cyclone modelling and thus, save lives and minimise infrastructural damage.

The project will build deep learning models with high resolution weather prediction data, European Centre for Medium-Range Weather Forecasts (ECMWF), National Centre for Environmental Prediction (NCEP), Global Forecast System (GFS), UKMET (UKMON), Hurricane Weather Research and Forecasting (HWRF), Coupled Ocean–Atmosphere Mesoscale Prediction System-Tropical Cyclone (COAMPS-TC) and other statistical-dynamical methods. Students will validate model data, arrange cyclone data into categories, basins of formation, recurving and straight-moving, intensity at model initialisation for analysis. The project will build models for 6 hourly interval predictions, hybrid forecasts using deep learning and copula-statistical methods for probabilistic conditional forecasts upon given variables to account for uncertainties in tracks or intensity and sensitivity to atmospheric conditions. The study will use artificial intelligence to make positive outcomes for natural disaster risk management.

The project is strongly centred on disaster risk mitigation and will suit students with a background in meteorology, atmopsheric sciences, physics, mathematics, computing or climate science. The project is suitable for MSC but is also scalable to a PhD level. It provides opportunity to publish in high quality Q1 journals. The student will be part of Advanced Data Analytics Research Group under Professor Ravinesh Deo. See

Principal Supervisor

Associate Supervisors

Research Affiliations
  • School of Agricultural and Environmental Sciences

Field of Research
  • Artificial Intelligence and Image Processing
  • Atmospheric Sciences

Available Academic Programs
  • Doctor of Philosophy (DPHD)
  • Master of Research (MRES)

Application Open Date

Application Close Date

USQ Scholarship Applications

Other Scholarship Funding Details
Australia Awards - discuss with Professor Ravinesh Deo for support to apply

Pre-approved for Ethics
Not Applicable

Admission Requirements

Please review the admission requirements for the academic program associated with this Thesis Topic

Some programming and statistics skills are assumed. Student will learn Python, r or MATLAB programming, data analysis and AI.

The project will advance student’s careers and acquire programming skills with publication opportunity in Q1 journals.

Contact Professor Ravinesh Deo or see

Back to List