Bayesian Precipitation Regime Shift Detection In South-east Asia

Vasily Kokorev, J. Ettema, Peter Siegmund

Research output: Contribution to conferencePosterOther research output


The changing climate affects precipitation regimes worldwide, altering not only means but also extremes and variability. South-East Asia in recent years has suffered from an increased number of droughts caused by a delayed start of the wet season and longer dry spells. Accurately assessing recent changes in precipitation regimes is difficult due to high variability that can mask trends and large-scale regime shifts. Precipitation changes are often not gradual but abrupt, and detecting such rapid regime shifts is necessary for understanding current climate and predicting future changes.
In this study, we use the Bayesian regime shift detection method to analyze historical precipitation changes in South-East Asia. The used method has an advantage of being able to detect shifts not only in mean values but in all of the distribution parameters. Use of Bayesian statistic provides probabilistic estimates of regime shift occurrence.
Our results demonstrate that precipitation regime shift occurs in the Indochina region and the Philippines in the mid-seventies. The observed regime shift is in line with the shift of the Pacific Decadal Oscillation of 1975/76. We demonstrate that significant changes in probabilities of a dry day occurred after regime shift. We argue that recently observed droughts in the South-East Asia region is a result of a new regime and cannot be explained by natural variability alone.
Original languageEnglish
Publication statusPublished - 5 Feb 2018
EventAOGS-EGU Joint Conference: New Dimensions for Natural Hazards in Asia - Tagaytay, Philippines
Duration: 4 Feb 20186 Feb 2018


ConferenceAOGS-EGU Joint Conference

Cite this