Please use this identifier to cite or link to this item:
Title: Higher order Markov chain models for monsoon rainfall over West Bengal, India
Authors: Dastidar, Avik Ghosh
Ghosh, Deepanwita
Dasgupta, S
De, U K
Keywords: Markov chain model;Bayesian information criterion;Rainfall probability;Steady state probability;Mean recurrence time
Issue Date: Feb-2010
Publisher: CSIR
PACS No.:; 02.50.Ga
Abstract:  Two-state Markov chain models of different orders have been used to simulate the pattern of rainfall during the monsoon season (June-September) over Gangetic West Bengal (India). The analysis is based on the relevant data for 31-year period (1970-2000) for four major meteorological stations in the region. The determination of the proper order of the Markov chain that best describes the rainfall pattern is an interesting problem and Bayesian information criterion has been used for the purpose. Bayesian information criterion (BIC) reveals that third order Markov chain model best describes the rainfall pattern in general except for one station. This is verified and it is found that third/fourth order chain simulates the observed data more closely than the chains of other orders using the classical goodness of fit test. The time independent behaviour of the chain has also been studied with the help of steady state probabilities. The theoretical and observed values of the mean recurrence time have been found to be in close agreement.
Page(s): 39-44
ISSN: 0975-105X (Online); 0367-8393 (Print)
Appears in Collections:IJRSP Vol.39(1) [February 2010]

Files in This Item:
File Description SizeFormat 
IJRSP 39(1) 39-44.pdf140.98 kBAdobe PDFView/Open

Items in NOPR are protected by copyright, with all rights reserved, unless otherwise indicated.