A MARKOV CHAIN ANALYSIS OF DAILY RAINFALL OCCURRENCE AT WESTERN ORISSA OF INDIA
Abstract
This paper makes an attempt to investigate the pattern of occurrence of rainfall such as sequences of wet and dry spells, expected length of such spells, expected length of weather cycle etc. in the western part of Orissa state of India by fitting a 2-state Markov chain probability model to the collected daily rainfall data for a period of 29 years. The key assumption behind reliability of this model is that the occurrence of a wet or a dry day is dependent on the weather condition of the previous day.