The Study of Wind Field ERA-20C in Monsoon Domains for Rainfall Predictor in Indonesia (Java, Sumatra, and Borneo)

Abstract

In recent years, various research institutions have developed diverse global data reanalysis projects. This provides an opportunity to gain long-term of meteorological data for local scale. This study aims to select the potential predictor of wind fields u and v of the ERA-20C dataset, a reanalysis dataset, at 850 mb from seven domains or windows of Asian, Maritime Continent, Australian, and Western North Pacific monsoon related physically to rainfall anomaly patterns in Indonesia. The vector wind velocity scalar was obtained by using a Helmholtz decomposition to separate the total circulation v = (u,v) into the divergent component/velocity potential (χ) or Phi and rotational component/stream function (ψ) or Psi for obtaining the scalar variable of vector wind velocity. The method applied Singular value decomposition (SVD) to identify pairs of spatial patterns (expansion coefficients) between the predictors of Phi and Psi in seven domains, with rainfall data from 48 stations in Java, Sumatra, and Borneo Islands from 1981 to 2010. The results revealed that spatial patterns correlations of Java Islands were the highest in the Maritime Continent monsoon domain (80o−150o E and 15oS−15o N), while Sumatra and Borneo Island were in the Western North Pacific monsoon domain (100o–130o E and 5o–15o N) with predictor Psi. The lowest correlation for Java, Sumatra, and Borneo was the Australian monsoon domain (110o E–130o E and 5o S–15o S) with predictor Phi. In general, spatial pattern correl-ations of Java Island were higher than others, agreeing with monsoonal rainfall type dominantly in the region.