Aplikasi Change Point Analysis (CPA) pada Data Curah Hujan Harian

Abstract

The highly change on the rainfall average can be analyzed by using the Change Point Analysis (CPA). At any change point, CPA analyzes it in detail according to the confidence level dan confidence interval which is obtained by bootstrap methods. CPA was done on daily rainfall data of wet months in October to March 1996, 1999, 2002, 2007, and 2008 at Pondok Betung Jakarta station as well as in 2007 and 2008 at Darmaga Bogor station. The results show that CPA can detect the average changes when rainfall is high at Pondok Betung Jakarta and Darmaga Bogor station in January and February. Besides to analyze the daily average change of rainfall using CPA which is based on historical data, forecasting using ARIMA model was also conducted and the pattern forecasted rainfall was also analyzed using CPA. However, ARIMA model (0,1,1) did not precisely forecast the daily rainfall so the pattern of change was not detected well.