Characterization of daily rainfall variability in Hong Kong: A nonlinear dynamic perspective

Modelling and forecasting of rainfall behaviour are of essential importance in various hydrological and meteorological studies, which depends largely on properly understanding the underlying rainfall dynamics. In this paper, 20‐years (i.e., 2000–2019) measurements of daily rainfall data at 6 rain gauge stations in Hong Kong were analysed in the framework of nonlinear dynamic analysis. Phase space reconstruction (PSR), recurrence plot (RP) and recurrence quantification analysis (RQA) were carried out to identify and characterize the chaotic nature of rainfall dynamics. The results of this study underscore the complex processes and variability that is associated with Hong Kong's rainfall data. The existence of chaos in rainfall process was well recognized across various stations. The rainfall dynamics were further investigated with respect to its inter‐annual and monthly variability. The dependence of rainfall dynamics on different temporal scales was likewise examined, which indicates that the rainfall dynamics at finer resolution are less complex than those at coarser resolution, thus the consequent rainfall variability is less pronounced. This paper extended the concept of the chaos theory and nonlinear dynamic analysis to the daily rainfall data recorded at six weather stations in Hong Kong, with a view to characterize the rainfall dynamics and variability. The existence of chaos in rainfall process was clearly identified. The rainfall dynamics were investigated in detail using phase space reconstruction, recurrence plot and recurrence quantification analysis. Station‐to‐station variability of rainfall dynamics was noticeable. The dependence of rainfall dynamics on monthly variation and different temporal scales were likewise examined.