Analysis of drought and flood occurrence using markov chain

Abstract

Flood and drought are among the most common natural disasters affecting the world. In this paper, Markov model has been used to analyse and predict flood and drought occurrences in Birnin Kebbi, Nigeria. The Standardized Precipitation index(SPI) was used to classify the annual rainfall of Birnin Kebbi into three states (flood, normal and drought). After some successful iterations of the model, the model stabilized to equilibrium probabilities, revealing that in the long-run 20% of the years in Birnin Kebbi will experience flood, 60% will experience normal rainfall and 20% will experience drought. It was also observed that, a drought year cannot be followed by a flood year and the probability of a drought year to be followed by a normal year is high while the probability of a normal year to be followed by a drought year and a drought year to be followed by another drought year is extremely small. Results from this research is an important information to the government and people of Kebbi state for better understanding of rainfall dynamics in their locality