Weekly rainfall analysis using the Markov chain model in the Dharmapuri region of Tamil Nadu

Abstract

During a rainy season, dry and wet spells tend to persist and can be represented using a Markov process. Knowing the succession of dry and wet periods is necessary to plant crops and carry out agricultural operations. This study aimed to analyze the probability of dry/wet spell rainfall using the Markov chain model in the Dharmapuri district of Tamil Nadu, India. In estimating the chance of dry and wet spells, the model used rainfall of below 20 mm in a week as a dry calendar week and rainfall of 20 mm or more as a wet calendar week from the years 1980 to 2019. From the 1st through the 32nd Standard Meteorological Week (SMW), a continuous dry week probability was 75-100%. The likelihood of a dry week trailed by another dry week was more up to the 32nd standard week, while the chance of a dry week followed by a wet week was more up to the 31st standard week, ranging from 75 to 100%. During the 37th to 45th weeks, the conditional likelihood of a rainy week followed by another rainy week ranged from 43.8 to 68%. According to a review of consecutive dry and wet spells, two consecutive dry weeks had a 55 to 97.5% chance of occurring within the first 32 weeks of the year. In the first 32nd week of the year, the chance of three successive dry weeks ranged from 32.6 to 92.6%. Consecutive dry weeks suggest the need for additional irrigation and proper moisture management practices. In contrast, consecutive wet calendar weeks indicate an abundance of extra water available for rainwater collection and the necessity for proper soil erosion control measures.