Rainfall analysis using frequency analysis approach and seasonal time series in Nusawungu District

Abstract

High rainfall can cause floods that damage the environment. To deal with potential extreme rainfall, an estimation of the frequency and intensity of rainfall at certain time intervals in the future is required. Analysis of the frequency of rainfall is also important in calculating the risk of flooding in an area and developing appropriate risk mitigation and management strategies. The purpose of this study was to analyze the frequency of rainfall in the Nusawungu sub-district using the Pearson Log Type III distribution method and forecasting using the moving average method. Daily maximum rainfall data for 20 years were obtained from the Google Earth Engine with the CHIRPS satellite in the Nusawungu sub-district and analyzed with the Smirnov-Kolmogorov distribution fit test and the Pearson Log Type III distribution method to estimate the frequency and intensity of rainfall in plan years 2, 5, 10, 25, 50, and 100. The results showed that the Pearson Log Type III distribution was suitable for analyzing the frequency of rainfall in the Nusawungu sub-district and the highest rainfall intensity occurred in the 100-year plan with a value of 207,22 mm/day. The moving average method is used to predict monthly rainfall in the following year, but the error test using the Mean Absolute Percentage Error (MAPE) shows that the forecasting model still needs to improve its accuracy. Analysis of the frequency of rainfall and forecasting can be an important reference in the planning and risk management of floods in the Nusawungu sub-district area.