Analysis of dry and wet weeks of rainfall by using Markov Chain - A case study at Jorhat (Assam), India

Abstract

Knowledge of the weekly dry and wet spell rainfall analysis is an important aspect for successful agricultural planning as well as water conservation techniques. The rainfall data during the period 1996-2018 from Jorhat station has been analyzed and the average, maximum, minimum, standard deviation and co-efficient of variation for each 52 Standard Meteorological Weeks (SMWs) were calculated.  The initial and conditional probabilities of dry and wet weeks were investigated by employing Markov Chain probability model. It was observed that, during the first 12 SMWs, the chance of two and three consecutive dry weeks was ranged from 40 to 100% and from 15 to 96%, respectively, while the chance of two and three consecutive wet weeks was ranged from 0 to 20% and 0 to 16%, respectively. The 20^th  SMW (14^th to 20^th May) is the earliest and the 23^rd SMW (4^th to 10^th June) is the most delayed week of the onset of rainy season. Also, the 40th SMW (1^st to 7^th Oct) is the earliest and the 47^th SMW (19^th to 25^th Nov) is the latest of withdrawal of southwest monsoon. The length of the rainy season is 161 days (21^st May to 28^th Oct).  The 41^st to 43^rd SMWs were vulnerable to 50% probability for dryness.  The monthly effective rainfall (ER) of the station was calculated and it was observed that, the total annual ER accounts only 47% of the average annual rainfall of this region.