Partial Duration Trend Analysis for Yearly Rainfall and its Sequential Distribution: a Conceptual Model

Abstract

In this study a new conceptual model for partial duration series analysis was applied to determine the likely rainfall of the monsoon season and the pattern of rainfall events distribution. Example case study was carried out with rainfall data of semiarid area, Agra, India, experiencing extremes of yearly rainfall; ranging between high (1200 mm) to low (300 mm). The coefficient values for the exponential decay function (K) were computed for all decays of partial dura- tion rainfall serises. Using the range of values of K and the yearly rainfall amounts further optimization produced the best results on the basis of the minimum Chi square value. The optimum K value was 0.92. Using this K, yearly rainfall values were predicted and compared with rainfall received for the validation period (1988-2002), not included in determination of optimum value of K. All the predicted yearly rainfall values were bit higher than observed ones. Further, for matching the predicted values with observed rain fall values, correction equations were also developed. For distribution of rainfall events in the year, successive rainfall event that could have preceded at any time were con- trasted with those of the past years with corresponding total rainfall values. As the rainy season progresses, the matching rainfall distribution of the corresponding rainfall year becomes the rainfall distribution pattern for year under the forecast; enabling selection of crops and cropping patterns before the onset of rainy season and carrying out strategic planning for successful agriculture under the changing scenario of rainfall pattern.