NON-LINEAR TREND IN TIME SERIES OF HEAVY RAINFALL IN VIET NAM

Abstract

Viet Nam frequently experiences the most heavy-rainfall-associated severe flood events. Duo to global warming, extreme heavy rainfall tends to occurs more often and induce significant disaster. Therefre, it is important to study the trend of heavy rainfall in Viet Nam. However, past studies of heavy rainfall trend are mostly based on linear regression or nonparameter method such as Sen method and Mann-Kandall test. However, these methods work with assumption that the time series are considered stationary. This assumption is not true for hydrometeorological variables, which are mostly nonstationary and nonlinear.
This study aim to analysis the heavy rainfall trend in Viet Nam using three methods: Linear regression, Sen method and singular spectral analysis. The purpose of using singular spectral analysis to capture the nonlinear trend of heavy rainfall. The results show that, in general, heavy rainfall exhibits increasing trend in Northwest, Northeast, North and South Central and Central Highlands while it displays decreasing trend in Red River Delta and Southern Plain. However, the heavy rainfall shows large fluctuations in different time period.