Identifying Synoptic Patterns Associated With Rainfall Regimes over Sri Lanka

During recent decades, Sri Lanka has experienced significant damage associated with frequent extreme rainfall events. In this study, a multivariate HMM (hidden Markov model) was applied to identify a synoptic pattern over the 20 stations having 55 years (i.e. 1961-2016) of daily rainfall records which can be grouped into three different regions in the context of rainfall patterns over Sri Lanka. The main rainy season for the each region was selected to explore the synoptic patterns informed by the hidden states within a multivariate framework. We choose the optimum number of hidden states ranging from 3 to 7 by maximizing the log-likelihood for each station. The hidden states are clearly obtained from the Viterbi algorithm on daily basis. Furthermore, the distribution of hidden states can effectively represent the temporal evolution of synoptic pattern associated with rainfall pattern.

Keywords: Multivariate, Hidden Markov Model, Extreme rainfall, Sri Lanka, Hidden state

Acknowledgement: This work is supported by the Korea Agency for Infrastructure Technology Advancement (KAIA) grant funded by the Ministry of Land, Infrastructure and Transport (Grant 18AWMP-B083066-05).