Prediction of Meiyu rainfall in Taiwan by multi-lead physical-empirical models
Journal
Climate Dynamics
Journal Volume
44
Journal Issue
11��22��
Pages
3033-3042
Date Issued
2015
Author(s)
Abstract
Taiwan is located at the dividing point of the tropical and subtropical monsoons over East Asia. Taiwan has double rainy seasons, the Meiyu in May–June and the Typhoon rains in August–September. To predict the amount of Meiyu rainfall is of profound importance to disaster preparedness and water resource management. The seasonal forecast of May–June Meiyu rainfall has been a challenge to current dynamical models and the factors controlling Taiwan Meiyu variability has eluded climate scientists for decades. Here we investigate the physical processes that are possibly important for leading to significant fluctuation of the Taiwan Meiyu rainfall. Based on this understanding, we develop a physical–empirical model to predict Taiwan Meiyu rainfall at a lead time of 0- (end of April), 1-, and 2-month, respectively. Three physically consequential and complementary predictors are used: (1) a contrasting sea surface temperature (SST) tendency in the Indo-Pacific warm pool, (2) the tripolar SST tendency in North Atlantic that is associated with North Atlantic Oscillation, and (3) a surface warming tendency in northeast Asia. These precursors foreshadow an enhanced Philippine Sea anticyclonic anomalies and the anomalous cyclone near the southeastern China in the ensuing summer, which together favor increasing Taiwan Meiyu rainfall. Note that the identified precursors at various lead-times represent essentially the same physical processes, suggesting the robustness of the predictors. The physical empirical model made by these predictors is capable of capturing the Taiwan rainfall variability with a significant cross-validated temporal correlation coefficient skill of 0.75, 0.64, and 0.61 for 1979–2012 at the 0-, 1-, and 2-month lead time, respectively. The physical–empirical model concept used here can be extended to summer monsoon rainfall prediction over the Southeast Asia and other regions.
SDGs
Type
journal article