Prospects for Improving the Operational Seasonal Prediction of Tropical Cyclone Activity in the Southern Hemisphere - Monitoring and Prediction of Tropical Cyclones in the Indian Ocean and Climate Change - page 130

The forecast of the number of TCs was also derived based on the developed
model with the values of October-November-December 5VAR index in 2010.
The resulting predicted number of TCs in the Australian region in 2010-11 is
13.45, with the standard error 1.11. The obtained results of our pilot study for
possible application of statistical-dynamical model-based approach encourage
us to continue this investigation using POAMA outputs which we aim to conduct
in our subsequent study.
4. Discussion and Conclusions
Interannual variability in the intensity and distribution of TCs is large, and
presently greater than any trends that are ascribable to climate change.
Historically TCs have had major impacts on agriculture, water supplies, safety
and economic well-being of Australia and island countries of the South Pacific
and South Indian Oceans. Better managing the year to year variability in
cyclones is a practical means for decreasing current and future vulnerability to
TCs. In addition, understanding the drivers of variability provides greater
confidence in future predictions and projections. This is particularly important
as the current understanding of TCs and seasonal conditions is mainly drawn
from historical data and past covariability with drivers such as the ENSO,
which are less valid in a changing climate. One issue that has emerged is a
problem in predicting TC occurrence based on historical relationships, with
predictors such as the SOI and SSTs now frequently lying outside of the range
of past variability which was demonstrated by over-predicting TC activity in
the Australian region for 2010-11 TC season.
Currently, a statistical model-based prediction of TC activity in the coming
season is used by the NCC for operational seasonal forecasting in the Australian
region and the Pacific. Statistical models are also used by other agencies, for
example the National Institute for Water and Atmospheric Research (NIWA)
in New Zealand and the GCACIC, Hong Kong. In this study, we demonstrated
a possibility of improving the accuracy of seasonal forecasts in the Australian
region using statistical model-based approach. However, we also found that
statistical models cannot produce skilful forecast of TC activity for other regions
of the Southern Hemisphere.
We also made a first step towards investigating a possibility to apply
statistical-dynamical model-based approach for TC seasonal prediction. The
statistical models are based on historical climate data. Consequently, it leads
to shortcomings because the statistical models cannot account for aspects of
climate variability and change that are not represented in the historical record.
This is an increasing problem as climate change brings new and unforeseen
climate conditions. Dynamical (physics-based) climate models do not have
this short-coming and are consequently better at incorporating the effects of a
changing climate, whatever its character or cause. Therefore, the transition
from a statistical to a dynamical prediction system will ultimately provide more