**5. Conclusions**

Historically, multi-hazards associated with TCs (destructive winds, storm surges, torrential rain and related flash-flooding) had significant impacts on population and coastal infrastructure of Australia and island countries of the Indian and Pacific oceans. Improved forecasting of TC seasonal activity is an important part of the Climate Risk Early Warning System (CREWS) for improving resilience of the society to potentially destructive impacts of TCs. Currently, a statistical model-based prediction of TC activity in the coming season is used for operational seasonal forecasting in the Australian region and the South Pacific Ocean by the Australian Bureau of Meteorology, the National Institute for Water and Atmospheric Research (NIWA) in New Zealand and the Guy Carpenter Asia-Pacific Climate Impact Centre (GCACIC) at the City University of Hong Kong.

In this chapter, a possibility of improving the accuracy of TC seasonal forecasting using advanced statistical model-based approach (e.g. support vector regression methodology) was demonstrated. Moreover, it was shown that dynamical (physics-based) climate models have potential for skilful seasonal TC forecasting. Transition from a statistical to a dynamical prediction system will ultimately provide more valuable and applicable climate information about TC seasonal variability, which can inform decision making, responses and adaptation in Australia and Pacific and Indian Ocean island countries.
