Rain, Rain, Go Away: Stochastic Model in Predicting the Future Rainfall

ASM Sc. J., 20(2), 2025

Published on August 18, 2025

https://doi.org/10.32802/asmscj.2025.1581

Author: Shamshimah Samsuddin, Nordiana Rosdi, Noor Hidayah Zainal, Nurul Ain Zailan, Nurul Faqihah Zamri and Rose Mayla Omar

Abstract

This study investigates the use of the Markov Chain (MC) method to forecast future rainfall in Tanah Merah, Kelantan, by observing daily rainfall data and categorising it into five distinct states, denoted as S={1,2,3,4,5}, each representing different levels of rainfall intensity. A structured examination of rainfall transitions between states is made possible by the research's discrete definitions of the state space and the temporal set. The study reflects the dynamic character of weather patterns by capturing the possibility of changes in rainfall amounts through the creation of a Transition Probability Matrix (TPM) for each month. In addition to forecasting rainfall, the study computes the limiting distribution of the TPMs to create risk matrices for every state. These risk matrices, which are based on recent and past rainfall data, offer a probabilistic evaluation of future flood hazards. The monthly risk matrices provide important insights into flood prediction and disaster preparedness by showing how the likelihood of rainfall in each state can affect the chance of flooding in succeeding months. The study illustrates the potential of the MC technique in enhancing flood risk management in the region and improving rainfall forecasts by utilising the concept of long-run behaviour.

Keywords: Flood, Markov chain, Rainfall