Abstract

This chapter presents a relatively new time series analysis tool called wavelet analysis. The limited availability of fine time step rainfall data for use in long-term continuous modeling could be alleviated by synthetically generating rainfall data of fine time step increments from temporally coarse rainfall data. This study explores the spectral behaviour of rainfall of various temporal resolutions for dominant periodicities and presents a simple method of generating credible rainfall data that combines the approaches of stochastic modeling with a disaggregation goal. In other words, the goal is to get better temporal resolution rainfall data from existing data. Large- and small-scale periodic components are identified in the daily, hourly, and tipping bucket time-between-tips data for the City of Edmonton.

A method of generating the desired rainfall data series is explained using the spectral behaviour of the rainfall data available. The desired rainfall data series would retain the record length of the most temporally coarse data (typically the longest period of record) and would have the desired fine time steps. The large-scale periodic components were extracted from three years of daily rainfall data and the small-scale components were extracted from one year of tipping bucket time-between-tips rainfall data for use in generating the synthetic rainfall data. The total rainfall volumes of the generated data compared well with observed values but tended to produce lower rainfall intensities and longer rainfall durations per event.

This paper is only available in PDF Format:

  View full text PDF