Singular Spectrum Analysis to Identify Excessive Rainfall
Sisti Nadia Amalia *
State University of Medan, Medan, Indonesia.
Sahat Saragih
State University of Medan, Medan, Indonesia.
Zul Amry
State University of Medan, Medan, Indonesia.
*Author to whom correspondence should be addressed.
Abstract
Indonesia is known for its excessive rainfall. Rainfall trends in an area have different characteristics. Differences in latitude, apparent motion of the sun, geographical position, topography, and the interaction of many forms of air circulation all contribute to this. Rainfall time series is essential for engineering planning, particularly for water infrastructure like irrigation, dams, urban drainage, ports, and wharves. Although meteorological technologies provide short-term rainfall predictions, long-term rainfall prediction is difficult and fraught with uncertainty. Unpredictability and seasonality can cause complex behavior in rainfall time series. This research utilizes the Singular Spectrum Analysis approach to extract trends; seasonality, cyclists, and noise can all be identified with potentially high accuracy.
Keywords: Singular spectrum analysis, excessive rainfall