A0688
Title: Periodogram regression, a semi-parametric mixed effects approach for modelling non-stationary tropical cyclone frequency
Authors: Lyuyuan Zhang - University of Melbourne (Australia)
Guoqi Qian - The University of Melbourne (Australia)
Sourav Das - Curtin University (Australia) [presenting]
Abstract: Tropical cyclones (TC) are significant indicators of evolving climate dynamics. Two primary responses of interest are the cyclone frequency and intensity. We propose a novel integrated modelling framework for simultaneous modelling of TC frequency across several meteorological regions in Australasia. We take a two-stage semi-parametric approach where large scale environmental variation is modelled using generalized linear models and stochastic spatio-temporal variation is estimated using spectral analysis of time series. This framework offers flexibility in modelling and leads to a confluence of several disciplines in statistical methodology including, hierarchical modelling, generalized additive mixed-effects modelling and periodogram estimation, concluding with a linear prediction for integer-valued time series with forecast uncertainties.