Gwangseob Kim

Monthly dam inflow forecasts using weather forecasting information and neuro-fuzzy technique

Abstract The purpose of this study is to evaluate the applicability of monthly weather forecasting information to the improvement of monthly dam inflow forecasts. The ANFIS (Adaptive Neuro-Fuzzy Inference System) is used to predict the optimal dam inflow, since it has the advantage of tuning the fuzzy inference system with a learning algorithm. A subtractive clustering algorithm is adopted to enhance the performance of the ANFIS model, which has a disadvantage in that the number of control rules increases rapidly as the number of fuzzy variables increases. To incorporate weather forecasting information into the ANFIS model, this study proposes a method for converting qualitative information into quantitative data. The ANFIS model for monthly dam inflow forecasts was tested in cases with and without weather forecasting information. It can be seen that the model performances obtained with the use of both past observed data and future weather forecasting information are much better than those using past observed data only.

Excitation functions of the proton induced nuclear reactions on natZn up to 40 MeV

We have measured the production cross-sections of the residual radionuclides for proton-induced reactions on natural tin by using a stacked-foil activation technique in the energy range from threshold energy to 40 MeV at the MC-50 cyclotron of the Korea Institute of Radiological and Medical Sciences. The results were compared with the earlier reported experimental data and theoretical calculations based on the TALYS and the ALICE-IPPE codes. The present results are in general good agreement with the available literature data and calculated results by using the computer codes TALYS and ALICE-IPPE. The thick target integral yields were also deduced from the measured excitation functions of the produced radionuclides.

Experimental determination of proton-induced cross-sections on natural zirconium

We measured cross-sections for the formation of (86g,87m,87g)Y, (88,89g)Zr, and (90,92m)Nb radionuclides for proton-induced reactions on natural zirconium by using a conventional stacked-foil activation technique in the energy range between 1 and 40MeV at the MC-50 cyclotron of the Korea Institute of Radiological and Medical Science. We compared the measured data with the available literature data and the theoretical calculation by the model codes TALYS and ALICE-IPPE. We also deduced the integral yields for thick targets from the measured cross-sections of the produced radionuclides. The optimum production possibility of the medically important (89g)Zr and (86)Y radionuclides were discussed elaborately.

Evaluation of Drought Indices using the Drought Records

본 연구에서는 가뭄지수들의 적합도와 가뭄의 변동특성을 파악하기 위하여 보고서 등 각종 문헌과 신문기사를 통해 수집한 실제 가뭄발생 기록을 정량화하고 1973년부터 2009년까지 전국 69개 기상관측소의 강수 및 기온 자료를 이용하여 산정한 기후변수와 각종 가뭄지수를 행정구역 별로 산정하여 비교 분석함으로서 기상학적 가뭄지수의 실제 가뭄에 대한 표현정도를 평가하였다. 평가를 위하여 ROC 공간상의 각 지수의 위치정보를 검정통계량으로 이용하였다. 분석결과 전반적으로 비교적 단기가뭄의 평가에 적합한 SPI3, PDSI, PN, Deciles이 실제 가뭄발생 기록과 가장 높은 상관성을 보였다. 행정구역에 따른 산포정도는 전반적으로 비교적 낮은 지역적 편차를 보이며, 기온과 증발량은 상대적으로 높은 지역적 편차를 나타냈다. 본 연구를 통해 우리나라 전 지역 가뭄의 시 공간적인 가뭄의 변화도를 파악하고 실제 가뭄에 대한 가뭄지수의 반영 정도를 파악하고, 더불어 ROC 분석을 통한 통계적 검증방법을 이용하여 가뭄지수의 적합도를 분석할 수 있는 가능성을 제시하였다. 【In this study, the suitability of drought indices was analyzed using the quantified drought records from official reports, newspapers and drought indices estimated using precipitation and air temperature data of 69 weather stations from 1973 to 2009. Test statistics of the suitability of meteorological drought indices were evaluated using the ROC analysis. Results demonstrated that PN shows the best relationships with drought records. SPI3 and Deciles Distribution Ratio also show good relationships with drought records and their variability according to the administrative divisions is relatively small. Results of the analysis of the spatial and temporal variability of drought and the accuracy of the drought indices can be used to evaluate the accuracy of drought indices in drought monitoring and prediction, and to select the best index in drought management.】

Investigations of the natTi(p,x)43,44m,44g,46,47,48Sc,48V nuclear processes up to 40 MeV

Independent and cumulative production cross-sections for the (nat)Ti(p,x)(48)V, (43,44m,44g,46,47,48)Sc nuclear processes are reported here, for the energy region of 4-38MeV by using a stacked-foil activation technique. Measured data were critically compared with the earlier reported values, and also with the theoretical data from the TALYS and ALICE-IPPE codes. The measured (nat)Ti(p,x)(48)V reaction is important for charged particle beam monitoring purposes, whereas the (43,44,47)Sc radionuclide have various practical applications in nuclear medicine.

Parameter Estimation of the Neyman-Scott Rectangular Pulse Model Using a Differential Evolution Method

The NSRPM(Neyman-Scott Rectangular Pulse Model), a stochastic rainfall generation model using clustered point process theory, is widely used in hydrology since it is able to reflect the cluster characteristics of rainfall events but the RPM(Rectangular Pulse Model) is not able to handle that. DFP(Davidon-Flectcher-Powell) and GA(Genetic Algorithm) are generally used to estimate 5 model parameters of the NSRPM. However DFP is sensitive to initial value and has tendency to get not global solution but local minimization. GA has an advantage of optimizing without Hessian matrix, but it needs relatively long computing time. In this study, a DE(Differential Evolution) method applied in model parameter estimation in order to overcome those drawbacks. The rainfall model parameters during the summer season (June-August) at Seoul, Pusan, Daegu, Mokpo, Gangrung, which have more than 30 year long hourly precipitation data, were estimated using hourly rainfall data from 1961 to 2011. Results were compared to those using DFP and GA. The performance of the DE method was evaluated using test statistics and demonstrated better performance in preserving physical and statistical properties of observed rainfall data.

The Impact of Climate Change on the Trends of Precipitation Effectiveness Ratio and Runoff Data in South Korea

To analyze the regional impact of air temperature increase and precipitation variation on water resources, the variability of precipitation-effectiveness (P-E) ratio which is estimated using precipitation and air temperature data of 59 weather stations operated by the Korean Meteorological Administration (KMA) during 1973~2009 was analyzed. Also runoff data resulting from the Precipitation-Runoff Modelling System (PRMS) modelling were analyzed during 1966~2007. The overall spatio-temporal variability of P-E ratio and runoff data in South Korea is corresponding to the variability of precipitation amount. However some region shows that the P-E ratio decreases even though the trend of precipitation amount increases which may be caused by the air temperature increase. Runoff trend is similar to that of P-E ratio. Precipitation and P-E ratio have decreased all seasons except summer season and it means the reduction of available water resources during those seasons. These variability should be reflected in the spring, fall, and winter water supply strategy.

Development of Distributed Rainfall-Runoff Model Using Multi-Directional Flow Allocation and Real-Time Updating Algorithm (I) - Theory -

In this study, a distributed rainfall-runoff model is developed using a multi-directional flow allocation algorithm and the real-time runoff updating algorithm. The developed model consists of relatively simple governing equations of hydrologic processes in order to apply developed algorithms and to enhance the efficiency of computational time which is drawback of distributed model application. The variability of topographic characteristics and flow direction according to various spatial resolution were analyzed using DEM(Digital Elevation Model) data. As a preliminary process using fine resolution DEM data, a multi-directional flow allocation algorithm was developed to maintain detail flow information in distributed rainfall-runoff simulation which has strong advantage in computation efficiency and accuracy. Also, a real-time updating algorithm was developed to update current watershed condition. The developed model is able to hold the information of actual behavior of runoff process in low resolution simulation. Therefore it is expected the improvement of forecasting accuracy and computational efficiency.

Development of a Short-term Rainfall Forecasting Model Using Weather Radar Data

The size and frequency of the natural disaster related to the severe storms are increased for recent decades in all over the globe. The damage from natural disasters such as typhoon, storm and local severe rainfall is very serious in Korea since they are concentrated on summer season. These phenomena will be more frequent in the future because of the impact of climate change related to increment of concentration and the global warming. To reduce the damage from severe storms, a short-range precipitation forecasting model using a weather radar was developed. The study was conducted as following four tasks: conversion three-dimensional radar data to two-dimensional CAPPI(Constant Altitude Plan Position Indicator) efficiently, prediction of motion direction and velocity of a weather system, estimation of two-dimensional rainfall using operational calibration. Results demonstrated that two-dimensional estimation using weather radar is useful to analyze the spatial characteristics of local storms. If the precipitation forecasting system is linked to the flood prediction system, it should contribute the flood management and the mitigation of flood damages.

The Behavior of the DNPAOs at the Anaerobic-Anoxic Process According to the Change of the Influent NO3⁻-N Loading

This study was conducted to investigate the ratios of phosphorus release to COD uptake, phosphorus release to nitrate removaland phosphorus uptake to phosphorus release by denitrifying phosphate accumulating organisms in the system composed of an anaerobic reactor and an anoxic reactor as the change of the influent nitrate loading. The influent concentrations of COD, NH4+−N and PO4−P were 300, 20 and 6 mg/L, respectively, whereas the influent nitrate concentrations were varied at 10, 15, 20 and 30 mg/L. As aresult, the ratio of phosphorus release to carbon uptake was increased in accordance with nitrate supply. The denitrifying phosphate accumulating organisms simultaneously took up phosphate and removed nitrate from the anoxic reactor. With above 20 mg/L of sufficient NO3−−N supply, phosphate was taken up excessively by the DNPAOs in anoxic condition, resulting in the ratio of phosphorus uptake to nitrate removal to be 1.97. The large amount of both uptake and release of phosphorus occurred above 20 mg/L of nitrate supply, achieving the ratio of phosphorus uptake to phosphorus release to be 1.05. Conversely, if nitrate as the electron acceptor was limited, the ratio was found to be low.