Statistical analysis of precipitation variations and its forecasting in Southeast Asia using remote sensing images

Abstract

The Climate Hazard Group InfraRed Precipitation with Stations (CHIRPS) dataset was examined for its variability and performance in explaining precipitation variations, forecasting, and drought monitoring in Southeast Asia (SEA) for the period of 1981–2020. By using time-series analysis, the Standardized Precipitation Index (SPI), and the Autoregressive Integrated Moving Average (ARIMA) model this study established a data-driven approach for estimating the future trends of precipitation. The ARIMA model is based on the Box Jenkins approach, which removes seasonality and keeps the data stationary while forecasting future patterns. Depending on the series, ARIMA model annual estimates can be read as a blend of recent observations and long-term historical trend. Methods for determining 95 percent confidence intervals for several SEA countries and simulating future annual and seasonal precipitation were developed. The results illustrates that Bangladesh and Sri Lanka were chosen as the countries with the greatest inaccuracies. On an annual basis, Afghanistan has the lowest Mean Absolute Error (MAE) values at 33.285 mm, while Pakistan has the highest at 35.149 mm. It was predicted that these two countries would receive more precipitation in the future as compared to previous years.