KOKOH

Planning and design of hydraulics works requires hydrology analysis with rainfall data as the main reference. Rainfall has space and time characteristics. Therefore the rainfall data from local gauge are needed. The recorded rainfall data have not always been complete, therefore a technique for providing data is needed to make optimal interpretation based on historical data. This study utilizes data generation technique to estimate the rainfall characteristics at the area of interest. The rainfall data to be analyzed is the recorded data during five year period in four rainfall gauges in Madiun river basin. The rainfall characteristics is analyzed by descriptive statistic method for studying the parameters comprise the average, variances, standard deviation, skewness, kurtosis and coefficient of autocorrelation. To investigate the characteristics of rainfall in the future, a synthetic data is generated base on standard random numbers, then transformed them into exponential or gamma distribution which in agreement to historical data characteristics. The similarity of means and variances between historical and synthetic data is analyzed by test of hypothesis. The result of analysis shows that the average depth of rainfall is between 2 to 4 mm/hr, with standard deviation between 4.64 to 5.25 mm/hr. Most of rainfall is in the range of 0 to 8 mm/hr, therefore it has high skewness and kurtosis. Average rainfall duration for all gauges is 2 hours, and it is identified that the rainfall movement is from upstream to downstream. The data distribution is of exponential or gamma, the goodness of fit test with ?= 0.05 shows that ?2calculated<?2table. The result of hypothesis test for similarity of means shows that Zcalculated