Analysis of Trends and the Persistence Structure in the Daily Rainfall Occurrences in Indiana

  1. AGU Hydrology Section
  1. M. L. Kavvas and
  2. J. W. Delleur

Published Online: 19 MAR 2013

DOI: 10.1029/SP004p0153

Precipitation Analysis for Hydrologic Modeling

Precipitation Analysis for Hydrologic Modeling

How to Cite

Kavvas, M. L. and Delleur, J. W. (1975) Analysis of Trends and the Persistence Structure in the Daily Rainfall Occurrences in Indiana, in Precipitation Analysis for Hydrologic Modeling (ed AGU Hydrology Section), American Geophysical Union, Washington, D. C.. doi: 10.1029/SP004p0153

Author Information

  1. School of Civil Engineering, Purdue University West Lafayette, Indiana 47907

Publication History

  1. Published Online: 19 MAR 2013
  2. Published Print: 1 JAN 1975

ISBN Information

Print ISBN: 9781118656037

Online ISBN: 9781118668993

SEARCH

Keywords:

  • Covariance and spectrum;
  • Daily rainfalls;
  • Intensity function;
  • Stochastic model;
  • Variance-time function

Summary

The long term prediction of the rainfall occurrences can be made by means of stochastic models. For a reliable prediction the stochastic model must account for the cyclicities, the long-term trends and the covariance structure of the rainfall counts. It is inadequate to test the goodness of fit of a stochastic model only in terms of the marginal probability distributions of the stochastic process of rainfall counts. Instead of hypothesizing a model and then testing its goodness of fit, first the covariance structure of the rainfall occurrence counts can be identified in terms of Barlett's [1963] counts spectrum and Cox and Smith' [1953] variance-time function of counts. Then based on this explicit persistence structure and on the theoretical behavior of the various point stochastic models, the model which can best fit the statistical behavior of the rainfall data can be selected for prediction purposes. The cyclicities and the long-term trends can be identified by the rate of occurrence function, the intensity function of Cox and Lewis [1966], the counts spectrum, and the variance-time function of counts. The application of these statistical functions to the daily rainfall counts in various stations in Indiana is given. The results of the statistical analysis agree with the physical meteorologic facts and the previous statistical work in Indiana.