Stochastic modelling of monthly rainfall series at Pantnagar (Uttarakhand)
Loading...
Date
2013-07
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
G.B. Pant University of Agriculture and Technology, Pantnagar - 263145 (Uttarakhand)
Abstract
The monthly rainfall data of 1981 to 2010 of Pantnagar were used for development of stochastic models. While the data for the year 2011 and 2012 were used for forcasting. To bring stationarity in the historical data series square root transformation was used. Fourier analysis was performed for determination of periodic component. The base period and number of significant harmonics was determined by autocorrelogram analysis and analysis of variance test respectively. Periodic component was removed by deducting from trend free series to obtain stochastic series. Stochastic series was standardised and normalised. The stochastic series was modeled by autoregressive (AR) family of Models. The Box-Pierce Portmanteau level was used to fit the test. The performance of model in regeneration and prediction of monthly rainfall value were evaluated quantitatively and qualitatively by comparison of historical and selected model correlograms and goodness of fit test such as mean forecast test, mean absolute error, root means square error. Results showed that properties propagated in the regenerated and predicted data series are similar to those of the historical series. In general, there is a good agreement between autocorrelation functions of historical and regenerated time series.
Description
Keywords
null