| dc.description.abstract | Extreme rainfall can give a negative affect for human activity. Negative effect of production failure in agriculture and plantations can be anticipated by extreme rainfall estimation that may occur in the future. The aims of this study is to compare the extreme rainfall estimation using the modified Champernowne distribution and generalized Pareto distribution (GPD), and determine the optimum weights of combined extereme value estimation. Modified Champernowne distribution patern has convergence to the GPD that the distribution function consists of three parameters that describe the data center, the diversity, and characteristics of the end of the distribution. While GPD consists of two parameters that describe the diversity and characteristics of the end of the distribution (heavy tail/tail light). The data used are daily rainfall data in Darmaga Bogor Station during the period 1985-June 2011 were obtained from the Indonesian Agency for Meteorology Climatology and Geophysics. GPD estimation tends to over estimate while modified Champernowne distribution estimation tends to under estimate. The combine of the estimated value by weighted produces more accurate extreme value estimation. Forecasting results for 1, 2, and 3 month ahead show that the best prediction based on modified Champernowne distribution, while forecasting results for 6 and 9 month ahead is very well predicted by GPD and combined value. | en |
