Determining appropriate probability distribution models for rainfall observations at Davao International Airport synoptic station /
Francisco L. Laurilla, Jr.
- 2003
- 82 leaves
Thesis (BS Applied Mathematics) -- University of the Philippines Mindanao, 2003
Dependable rainfall is important for the estimation of the irrigation diversion requirement. The common approach that engineers have been doing to estimate dependable rainfall is to assume normality in the variables being analyzed, thus making use of the Normal distribution to approximate values of rainfall at the exceedance of 80%. Normal distribution, however, is not necessarily representative of all rainfall data. And to give alternative to the use of this model, this study was proposed to fit other distribution models that significantly describe the probability distribution of rainfall data specifically the monthly rainfall totals, daily rainfall totals and daily maximum intensifies recorded at Davao International Airport Sypnoptic Station in Sasa, Davao City. The parameters of the fitted distribution models were estimated using method of moments (MOM) and method of maximum likelihood (MML). Chi-square (CS) and Kolmogorov-smirnov (KS) tests were employed in determining the goodness-of-fit of the fitted models. According to the evaluation of Cs test, Beta (MML), Gamma (MOM and MML), Gamma (3P) (MML), Log Gamma (MML), Gumbel for maxima (MML), Weibull for minima (MOM and MML) and Weibull for minima (3P) (MML) were found to be the best-fitting models for monthly rainfall totals. Aside from these models, KS also found Rayleigh (MOM) as best fitting model for this data. On the other hand, Gamma (MOM) and Weibull (MOM) were found to be the best-fitting models for daily rainfall totals according to CS test. For maximum intensifies, Gamma (MOM) was found to be best-fitting model according to CS test. Furthermore, it was shown that Normal distribution did not fit any rainfall data utilized in this study.