Wet and Dry Spells

Back to Agrimet Division's Research Activities

Analysis of distribution of wet and dry spells using Markov Chain models has been completed for some selected stations of all the 35 meteorological subdivision. Based on these, a drought index has been developed and drought prone areas were demarcated. Runs of dry and wet spells have also been computed for some States.

In this study a Markov Chain Model has been applied to obtain the probabilities of wet spells over districts of five sub-divisions. they are Punjab, East Madhya Pradesh, Plains of West Uttar Pradesh, Gangetic West Bengal and Kerala. For this purpose daily rainfall data for the period from 1901-1990 for all the available stations in the sub-divisions have been considered.

There are many weather sensitive operations which require occurrence of certain amount of rainfall and its continuation for a critical period with a given probability. The probability analysis has an advantage over the study of arithmetic means and helps water resource management in different parts of the country.

A week is defined as wet when the amountof rainfall for that week is more than specified threshold value based on C.V. of rainfall for the monsoon season and at least 50% of the total number of the stations should report rainfall of 1mm ormore for 3 days in aweek. A week which is not "Wet" week is assumed in the study as "Dry"week.