Markov Chain Model of weekly rainfall probability and dry and wet spells for agricultural planning in Coimbatore in western zone of Tamil Nadu

Abstract

The historical rainfall data for the period 1981-2011 of Coimbatore in western zone of Tamil Nadu were analyzed using Markov Chain Probability Model to describe the long-term behavior of wet or dry weather spells. The model considered less than 20 mm rainfall in a week as a dry week and 20 mm or more as a wet week to estimate the dry and wet spell probability. The probability of occurrence of dry week was very st th high from 1 to 38 Standard Meteorological Week (SMW) ranging from 75-100%. The probability of occurrence of dry week preceded by another dry week was also th high up to the 40 standard week and that of dry week preceded by a wet week was th high up to 35 week and varied from 75-100%. The conditional probability of wet week preceded by another wet week was comparatively high and varied from 43.8 to th th 92.3% during 39 to 48 week. The analysis of consecutive dry and wet spells revealed that there are 50 to 100% chances that two consecutive dry weeks will occur within the first 38 weeks of the year. Similarly the probability of occurrence of three consecutive dry weeks was also very high (53 to 100%) in the first 38 weeks of the year. The successive dry weeks indicate the need for supplemental irrigation and suitable moisture conservation practices whereas, occurrence of successive wet weeks gives an idea of excessive runoff water availability for rainwater harvesting and to take up suitable measures to control soil erosion. The analysis also showed the mean, earliest and delayed weeks of onset and withdrawal of rainy season in th nd th th th nd Coimbatore as 28 , 22 , 39 and 47 , 44 , 52 , respectively. Hence, it is suggested that th sowing of crops may be started in 27 week. High rainfall, more wet weeks and th delayed withdrawal of rainy season with high probability (73.4-95.3%) from 49 to nd 52 week could be utilized successfully for rabi crops. Since dry probabilities are rd th high during kharif (23 to 39 standard week), short duration crops and other less water requiring crops may be grown which have high return value.