Fitting models to daily rainfall data of Los Baños

Date

4-2009

Degree

Bachelor of Science in Applied Physics

College

College of Arts and Sciences (CAS)

Adviser/Committee Chair

Alvin Karlo G. Tapia

Abstract

The daily occurrence of rain in Los Banos was modeled by Markov chain. Zero-, first- and second-order Markov chains were computed. Several types of functions were used to fit the Markov chains. Fourier series with four harmonics was chosen to he the function that best fits the curves in terms of the regression coefficient R' of the Markov chain. The proper order of Markov chain using the Akaike Information Criterion was found to be of order 1. First-order Markov chain was modelled for different time periods. For 25 years time interval, the fitted curves indicate that there is no climate change during the second half of the year and there was a little increase in the probabilities of rain following rain. For 10 years time interval, the fitted curves indicate that there is no climate change for the whole year. The fitted model was compared to the actual daily rainfall data of 2008 and showed that the amount of rainfall follows the pattern of the fitted curve. The rainfall amounts were modelled by several types of distributions by comparing the percentiles of the rainfall amounts of each model with the percentiles of rainfall amounts of 1959-2007. The analyses showed that gamma distribution with value of zero best described the rainfall amounts. The comparison of the distribution with the actual rainfall amounts through analyzing the length of the peaks showed that the model was adequate to model rainfall in Los Banos.

Language

English

Location

UPLB Main Library Special Collections Section (USCS)

Call Number

Thesis

Document Type

Thesis

This document is currently not available here.

Share

COinS