Stochastic analysis of rainfall and its application in appropriate planning and management for Eastern India agriculture

Abstract

Rainfall analysis is essential for agricultural crop planning and water resources management, especially under water scarcity conditions. The 36 years (1977–2013) of rainfall data for Kharagpur were analyzed for characterization of different seasonal events. The Weibull's formula predicted the probability of mean onset on 23rd standard meteorological weeks (SMW) (3rd–9th June) and withdrawal on 43rd SMW (21st–27th October). There was 80–83% probability of a wet week [P(W)] occurring within 25th–35th SMW. The first order Markov chain process shows the conditional probability of one wet week preceded by another wet week [P(W/W)] varied between 0 and 86%; whereas, dry week preceded by another dry week [P(D/D)] varied in the range of 70 to 100%. The stochastic analysis of successive wet or dry weeks formulates the adaptation strategies to avoid the possible effect of wet or dry spell during cropping seasons. The wet spell analysis suggests rainwater harvesting to control soil erosion and maximization of water use efficiency. The probability of getting consecutive wet [P(W/W)] and dry days [P(D/D)] were varied in the range of 40%–70% and 50%–90%, respectively. The probability of getting different magnitude of rainfall (10 to 40 mm) during the monsoon weeks (25th–39th week) were found to be more than 50% probability level, which suggest for harvesting of excess runoff water for future supplemental irrigation.