Development of a Frequency Analysis Tool of Environmental Extreme Events |
---|
학술지명 Int'l Symposium on River and Lake Environment
저자 이상욱,강신욱
발표일 2014-08-25
|
In many cases pollution data sets are skewed so that the symmetric normal distribution is not a suitable model for estimating quantiles, proportions, or means. The lognormal distribution is frequently used for a frequency analysis of skewed datas in environmental engineering. The Weibull, gamma, and beta distributions are sometimes used to model environmental pollution data. The K-Water Frequency Analysis Tool (K-FAT) was developed to perform for performing environmental pollution, rainfall, and streamflow frequency analyses. The K-FAT operates through the inclusion of 14 probability distribution functions. Cunnane(1989) investigated the mostly utilized probability distribution functions in the world. The parameter estimation of the probability distribution functions were conducted in three ways: Method of Moments, Maximum Likelihood Method and L-Moments MethodThe goodness-of-fit tests include Chi-Squared, Kolmogorov-Smirnov, Cramer von Mises, Probability Plot Correlation Coefficient (PPCC), modified Anderson-Darling tests. To determine the best fitting probability distribution type, the methods AIC or Akaike Information Criterion (Akaike, 1973), and BIC or Bayesian Information Criterion (Schwarz, 1978) were utilized. These methods were included in the system to help users in deciding on which probability distribution function is best for the environmental and hydrologic data time series. The main program and the sub-program utilizes the Compaq Visual Fortran version 6.6 with 18,530 lines of codes. The GUI of the K-FAT consists of 3 modules: a pre-processing module, which analyzes and generates data from the Databases; a main module, which performs the frequency analysis system; and a post-processing module, which displays the results of the analyses in the form of figures, tables, graphs, plots and texts. This system can contribute to further researches regarding the continuous advancement of environmental and hydrologic statistics. |