This research is for finding the suitable the synthetic data generating model for daily rainfall in the region of Middle Brantas River Basin in the East Java Province. There are 7 models being considered, 4 models for single-site generation, (1) the two-part, model group, (2) the transition probability matrix model group, (3) the resampling model group, and (4) the time series model group, and 3 models for multisite generation, (5) the conditional, model group, (6) the extension of single site Markov chain model group, and (7) the random cascade model group. All of the time-series produced by the daily rainfall synthetic data generation are then tested statistically. The results show that statistically the differences between the historical time series and the synthetically time series are not too significant. It turn out that the multisite model have produced better synthetic time series compared to those which have been produced by the single-site models.Keywords: generating model, synthetic data, daily rainfall.
Copyrights © 2013