Frequency investigation of rainfall in the sugarcane growing region of North Queensland, Australia

Abstract

Understanding the underlying patterns of atmospheric systems is an important step in understanding climate variability. The potential impact of climate variability on agricultural industries is an important industry issue. Greater understanding of climate variability on both global and local scales is required. This local scale study analyses the frequency of rainfall in the sugarcane growing region of North Queensland, Australia. Spectral analysis of time series signals allows for the identification of periodic and quasi-periodic frequency components in the temporal structure of climate signals and how these have varied over time. This study used wavelet decomposition to deconstruct the rainfall records for 3 locations in North Queensland: Macknade, Kalamia and Tully to identify the underlying frequency spectrum. Seasonal accumulations of the weather station rainfall records from 1889, 1913 and 1926 to 2008 were computed, and wavelet analysis suggest that North Queensland rainfall at these three locations are influenced by a number of underlying cycles, covering both short and long time periods. A separate analysis of each season across all locations showed similarities between autumn and winter rainfall variability, and summer and spring rainfall. It was shown that the rainfall variability in each season is influenced by both ENSO length periodicities, as well as longer term decadal and interdecadal oscillations. The analysis did not show any changes to how the rainfall signals may be varying, however a quantitative approach may detect such changes. The identification of the periodic cycles influencing the local climate is a useful step towards understanding and controlling local changes in climate variability.