• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.170-170
• /
• 2020

Due to the development of technologies, complex computation of huge data set is possible with a prevalent personal computer. Therefore, machine learning methods have been widely applied in the hydrologic field such as regression-based regional frequency analysis (RFA). The main purpose of this study is to compare two frameworks of RFA based on the artificial neural network (ANN) models: quantile regression technique (QRT-ANN) and parameter regression technique (PRT-ANN). As an output layer of the ANN model, the QRT-ANN predicts quantiles for various return periods whereas the PRT-ANN provides prediction of three parameters for the generalized extreme value distribution. Rainfall gauging sites where record length is more than 20 years were selected and their annual maximum rainfalls and various hydro-meteorological variables were used as an input layer of the ANN model. While employing the ANN model, 70% and 30% of gauging sites were used as training set and testing set, respectively. For each technique, ANN model structure such as number of hidden layers and nodes was determined by a leave-one-out validation with calculating root mean square error (RMSE). To assess the performances of two frameworks, RMSEs of quantile predicted by the QRT-ANN are compared to those of the PRT-ANN.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.57 no.5
• /
• pp.139-152
• /
• 2015

Along with climate change, it is reported that the scale and frequency of extreme climate events show unstable tendency of increase. Thus, to comprehend the change characteristics of precipitation data, it is needed to consider non-stationary. The main objectives of this study were to estimate future design floods for Wonpyeongcheon watershed based on RCP (Representative Concentration Pathways) scenario. Wonpyeongcheon located in the Keum River watershed was selected as the study area. Historical precipitation data of the past 35 years (1976~2010) were collected from the Jeonju meteorological station. Future precipitation data based on RCP4.5 were also obtained for the period of 2011~2100. Systematic bias between observed and simulated data were corrected using the quantile mapping (QM) method. The parameters for the bias-correction were estimated by non-parametric method. A non-stationary frequency analysis was conducted with moving average method which derives change characteristics of generalized extreme value (GEV) distribution parameters. Design floods for different durations and frequencies were estimated using rational formula. As the result, the GEV parameters (location and scale) showed an upward tendency indicating the increase of quantity and fluctuation of an extreme precipitation in the future. The probable rainfall and design flood based on non-stationarity showed higher values than those of stationarity assumption by 1.2%~54.9% and 3.6%~54.9%, respectively, thus empathizing the necessity of non-stationary frequency analysis. The study findings are expected to be used as a basis to analyze the impacts of climate change and to reconsider the future design criteria of Wonpyeongcheon watershed.

• The Sea:JOURNAL OF THE KOREAN SOCIETY OF OCEANOGRAPHY
• /
• v.16 no.4
• /
• pp.223-237
• /
• 2011

The intermittent freshwater discharge has an critical influence upon the biophysical environments and the ecosystems of the Yeongsan Estuary where the estuary dam altered the continuous mixing of saltwater and freshwater. Though freshwater discharge is controlled by human, the extreme events are mainly driven by the heavy rainfall in the river basin, and provide various impacts, depending on its magnitude and frequency. This research aims to evaluate the magnitude and frequency of extreme freshwater discharges, and to establish the magnitude-frequency relationships between basin-wide rainfall and freshwater inflow. Daily discharge and daily basin-averaged rainfall from Jan 1, 1997 to Aug 31, 2010 were used to determine the relations between discharge and rainfall. Consecutive daily discharges were grouped into independent events using well-defined event-separation algorithm. Partial duration series were extracted to obtain the proper probability distribution function for extreme discharges and corresponding rainfall events. Extreme discharge events over the threshold 133,656,000 $m^3$ count up to 46 for 13.7y years, following the Weibull distribution with k=1.4. The 3-day accumulated rain-falls which occurred one day before peak discharges (1day-before-3day -sum rainfall), are determined as a control variable for discharge, because their magnitude is best correlated with that of the extreme discharge events. The minimum value of the corresponding 1day-before-3day-sum rainfall, 50.98mm is initially set to a threshold for the selection of discharge-inducing rainfall cases. The number of 1day-before-3day-sum rainfall groups after selection, however, exceeds that of the extreme discharge events. The canonical discriminant analysis indicates that water level over target level (-1.35 m EL.) can be useful to divide the 1day-before-3day-sum rainfall groups into discharge-induced and non-discharge ones. It also shows that the newly-set threshold, 104mm, can just separate these two cases without errors. The magnitude-frequency relationships between rainfall and discharge are established with the newly-selected lday-before-3day-sum rainfalls: $D=1.111{\times}10^8+1.677{\times}10^6{\overline{r_{3day}}$, (${\overline{r_{3day}}{\geqq}104$, $R^2=0.459$), $T_d=1.326T^{0.683}_{r3}$, $T_d=0.117{\exp}[0.0155{\overline{r_{3day}}]$, where D is the quantity of discharge, ${\overline{r_{3day}}$ the 1day-before-3day-sum rainfall, $T_{r3}$ and $T_d$, are respectively return periods of 1day-before-3day-sum rainfall and freshwater discharge. These relations provide the framework to evaluate the effect of freshwater discharge on estuarine flow structure, water quality, responses of ecosystems from the perspective of magnitude and frequency.

• Journal of Distribution Science
• /
• v.15 no.8
• /
• pp.47-54
• /
• 2017

Purpose - According to agency theory, managers have incentives to adjust firm revenues to meet earnings expectations or delay bad news disclosure because of performance-based compensation and their reputation in the market. When the bad news accumulates, stock prices fail to reflect all available information. Thus, market prices of stocks are higher than their intrinsic value. After all, bad news crosses the tipping point, it comes out all at once. That results in stock crashes. Auditors can decrease stock crash risk by reducing agency costs through their informational role. Especially, stock price crash risk is expected to be lower for firms adopting high-quality audits. We focus on distribution and service industry to examine the relation between audit quality and stock price crash risk. Industry specialization and auditor size are used as proxies for auditor quality. Research design, data and methodology - Our sample contains distribution and service industry firms listed in KOSPI and KOSDAQ during a period of 2004-2011. We use a logistic regression to test whether auditor quality influences crash risk. Auditor quality was measured by industry specialist auditor and Big4 / non-Big4 dichotomy. Following the approach in prior researches, we use firm-specific weekly returns to measure crash risk. Firms experiencing at least one stock price crash in a specific week during year are classified as the high risk group. Results - The result of analyzing 429 companies in distribution and service industry is summarized as follows: Above all, it is shown that higher audit quality has a significant negative(-) effect on the crash risk. Crash risk is alleviated for firms audited by industry specialist auditors and Big 4 audit firms. Therefore, our results show that hypotheses are supported. Conclusions - This study is very meaningful as the first study which investigated the effects of high audit quality on stock price crash risk. We provide evidence that high-quality auditors reduce stock price crash risk. Our finding implies that the risk of extreme losses can be reduced through screening of high-quality auditors. Therefore investors and regulators may utilize our findings in their investment and rule making decisions.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.27 no.2
• /
• pp.23-37
• /
• 1985

This studies were carried out to get characteristics of frequency distribution, probable flood flows according to the return periods, and the correlation between return periods and those length of records affect the Risk of failure in the annual maximum series of the main river systems in Korea. Especially, Risk analysis according to the levels were emphasized in relation to the design frequency factors for the different watersheds. Twelve watersheds along Han, Geum, Nak Dong, Yeong San and Seom Jin river basin were selected as studying basins. The results were analyzed and summarized as follows. 1. Type 1 extremal distribution was newly confirmed as a good fitted distribution at selected watersheds along Geum and Yeong San river basin. Three parameter lognormal Seom Jin river basin. Consequently, characteristics of frequency distribution for the extreme value series could be changed in connection with the watershed location even the same river system judging from the results so far obtained by author. 2. Evaluation of parameters for Type 1 extremal and three parameter lognormal distribution based on the method of moment by using an electronic computer. 3. Formulas for the probable flood flows were derived for the three parameter lognormal and Type 1 extremal distribution. 4. Equations for the risk to failure could be simplified as $\frac{n}{N+n}$ and $\frac{n}{T}$ under the condition of non-parametric method and the longer return period than the life of project, respectively. 5. Formulas for the return periods in relation to frequency factors were derived by the least square method for the three parameter lognormal and Type 1 extremal distribution. 6. The more the length of records, the lesser the risk of failure, and it was appeared that the risk of failure was increasing in propotion to the length of return periods even same length of records. 7. Empirical formulas for design frequency factors were derived from under the condition of the return periods identify with the life of Hydraulic structure in relation to the risk level. 8. Design frequency factor was appeared to be increased in propotion to the return periods while it is in inverse proportion to the levels of the risk of failure. 9. Derivation of design flood including the risk of failure could be accomplished by using of emprical formulas for the design frequency factor for each watershed.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.25 no.4
• /
• pp.191-199
• /
• 2013

It is estimated and analyzed that the design wave height and the confidence interval (hereafter CI) according to the return period using the fourteen-year wave data obtained at Pusan New Port. The functions used in the extreme value analysis are the Gumbel function, the Weibull function, and the Kernel function. The CI of the estimated wave heights was predicted using one of the Monte-Carlo simulation methods, the Bootstrap method. The analysis results of the estimated CI of the design wave height indicate that over 150 years of data is necessary in order to satisfy an approximately ${\pm}$10% CI. Also, estimating the number of practically possible data to be around 25~50, the allowable error was found to be approximately ${\pm}$16~22% for Type I PDF and ${\pm}$18~24% for Type III PDF. Whereas, the Kernel distribution method, a typical non-parametric method, shows that the CI of the method is below 40% in comparison with the CI of the other methods and the estimated design wave height is 1.2~1.6 m lower than that of the other methods.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.17 no.1
• /
• pp.69-74
• /
• 2015

To analyze quantitatively environmental characteristics of cultivation area of citrus, Satsuma mandarin (Citrus unshiu Marc.), we made digital maps of environmental elements such as topography and climate. Elevation, degree of slope, and slope aspect were selected as elements of topological environment, and the annual mean air temperature, annual total precipitation, mean air temperature on January, extreme value of daily minimum air temperature, and the number of days below $-5^{\circ}C$ were selected as elements of climatic environments. The grid values of 8 environmental elements were extracted by shape of citrus farm area and analyzed distribution patterns. We can determine 3 agroclimatic criteria for growing Satsuma mandarin as over $14.5^{\circ}C$ of annual mean air temperature, over $-10.0^{\circ}C$ of extreme value of daily minimum air temperature, and less 5 days of below $-5^{\circ}C$ of daily minimum air temperature.

• Journal of Korea Water Resources Association
• /
• v.52 no.12
• /
• pp.1001-1009
• /
• 2019

Due to recent increasing heavy snow in Korea, the damage caused by heavy snow is also increasing. In Korea, there are many efforts including establishing disaster prevention measures to reduce the damage throughout the country, but it is difficult to establish the design criteria due to the characteristics of heavy snow. In this study, snowfall frequency analysis was performed to estimate design snow depths using observed snow depth data at Jinju, Changwon and Hapcheon stations. The conventional frequency analysis is sometime limted to apply to the snow depth data containing zero values which produce unrealistc estimates of distributon parameters. To overcome this problem, this study employed mixed distributions based on Lognormal, Generalized Pareto (GP), Generalized Extreme Value (GEV), Gamma, Gumbel and Weibull distribution. The results show that the mixed distributions produced smaller design snow depths than single distributions, which indicated that the mixed distributions are applicable and practical to estimate design snow depths.

• The Korean Journal of Applied Statistics
• /
• v.29 no.5
• /
• pp.807-822
• /
• 2016

The composite lognormal-GPD models (LN-GPD) enjoys both merits from log-normality for the body of distribution and GPD for the thick tailedness of the observation. However, in the estimation perspective, LN-GPD model performs poorly due to numerical instability. Therefore, a two-stage procedure, that estimates threshold first then estimates other parameters later, is a natural method to consider. This paper considers five nonparametric threshold estimation methods widely used in extreme value theory and compares their performance in LN-GPD parameter estimation. A simulation study reveals that simultaneous maximum likelihood estimation performs good in threshold estimation, but very poor in tail index estimation. However, the nonparametric method performs good in tail index estimation, but introduced bias in threshold estimation. Our method is illustrated to the service time of an Israel bank call center and shows that the LN-GPD model fits better than LN or GPD model alone.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.32 no.1
• /
• pp.26-38
• /
• 2020

A statistical test was carried out on the IID (Independently and Identically Distributed) assumption of the AM (Annual Maxima) data used to estimate the design wave height. The test was divided into independence (randomness) test and homogeneity test, and each test was conducted on AM data of 210 and 310 stations in coastal and inner coastal grids in typhoon and non-typhoon (monsoon) conditions. As a result of the independence test, the rejection ratios of the test are in the range of 1.8~5.3% and 1.4~6.0% for the non-typhoon and typhoon data sets, respectively. On the other hand, in the distribution difference test of typhoon data and nontyphoon data, the same distribution hypothesis was found to be rejected in the range of 47~79% according to the test method for both coastal grid and inner coastal grid. Therefore, in estimating design wave height by extreme value analysis, the estimation process by dividing the typhoon and non-typhoon data is appropriate.