During the period from August 18 to August 19, 1972, a heavy rainfall was observed in Kyeonggi district. The total amount of rainfall over that period exceeded 450㎜. Some synoptic features of the heavy rainfall were studied by the use of synoptic data. The notable feature was the synoptic situation which built up a deep convectively unstable layer over Korea peninsula. The observed low level jet stream seems to be formed by the heavy rainfall activity.

전국 9개 고정조사지에서 1968년이래 13개년간 조사한 솔나방의 10월의 밀도를 종합검토한바 다음과 같은 결과를 얻었다. 1. 13개년간에 솔나방의 10월밀도가 특히 높았던 년도는 '70년과 75년의 2회였으며 '76년이후는 단속 저밀도를 유지하고 있다. 2. 밀도가 현저하게 감소하였던 년도는 72년도와 76년도로서 8월중의 강우량 특히 24시간내 최대강우량이 밀접한 관계를 가지고 있었다. 3. 9개 조사지의 솔나방밀도증감은 대체적으로 유사한 경향을 보였으며 이것은 전국적인 강우의 대체적인 동시성에 기인한다고 본다. 4 8월중 일부 최다강우량이 100mm을 넘을 경우 솔나방밀도가 감소하는 확률은 였다.

In this study, the hydraulic gradient was calculated using the groundwater level and rainfall observed in the Hyogyo-ri area for a year, and the change in the hydraulic gradient according to the rainfall was analyzed. It was found that the groundwater level increased as the rainfall increased in all groundwater wells in the research site, and the groundwater level rise decreased as the altitude of the groundwater well increased. The hydraulic gradient in the research site ranged from 0.016 to 0.048, decreasing during rainfall and increasing after the end of the rainfall. As the rainfall increased, the groundwater level rise in the low-altitude area was more than the high-altitude area, and the hydraulic gradient decreased due to the difference in groundwater level rise according to the altitude. Through this study, it was found that the influence of rainfall is dominant for the fluctuation of the hydraulic gradient in the unconfined aquifer.

In this study, not only the increase in rainfall for a short period of time but also the increase in rainfall for a longer duration is frequently occurring according to climate change. Changes in rainfall patterns due to climate change are increasing damage to steep slopes. The Ministry of Public Administration and Security has been operating the criteria for evacuation of residents in steep slopes since 2015. However, the damage to steep slopes due to torrential rains in 2020 has been increasing. In this study, rainfall data from areas affected by steep slopes from 1999 to 2020 were collected and compared with the existing criteria(2015) for evacuation of residents at steep slopes of the Ministry of Public Administration and Security, and the status of the issuance of resident evacuation forecast was compared. Through this study, the rainfall criteria for each region were calculated and presented by reflecting the rainfall characteristics of the steep slope destruction area due to climate change, and it is believed that it can be used as a standard rainfall to reduce human casualties in the steep slope area in the future.

In this paper, among the W-S-R relationship methods proposed by Lee, et. al., (2020) to produce rain-based rain information in real time, we tried to produce actual rain information by applying machine learning techniques to take into account the effects of wiper operation. To this end, rain sensor proposed the Graded Descent and Threshold Map method for pre-processing the cumulative value of the difference before and after wiper operation by utilizing four sensitive channels for optical sensors developed by Kim Byung-sik (2016) and using rain sensor data produced by five rain conditions in indoor artificial rainfall experiments. This method is the method of producing rainfall information by calculating the average value of the Threshold according to the rainfall conditions and channels, creating a Threshold Map corresponding to the grid 4 (channel) x 5 (thinking of rainfall information) and applying Optima Rainfall Intensity among the big data processing techniques. For the verification of these proposed findings, the application was evaluated by comparing the rainfall observations with the methods presented by Lee, et. al., (2020).

In this study, the prediction technology of Hydrological Quantitative Precipitation Forecast (HQPF) was improved by optimizing the weather predictors used as input data for machine learning. Results comparison was conducted using bias and Root Mean Square Error (RMSE), which are predictive accuracy verification indicators, based on the heavy rain case on August 21, 2021. By comparing the rainfall simulated using the improved HQPF and the observed accumulated rainfall, it was revealed that all HQPFs (conventional HQPF and improved HQPF 1 and HQPF 2) showed a decrease in rainfall as the lead time increased for the entire grid region. Hence, the difference from the observed rainfall increased. In the accumulated rainfall evaluation due to the reduction of input factors, compared to the existing HQPF, improved HQPF 1 and 2 predicted a larger accumulated rainfall. Furthermore, HQPF 2 used the lowest number of input factors and simulated more accumulated rainfall than that projected by conventional HQPF and HQPF 1. By improving the performance of conventional machine learning despite using lesser variables, the preprocessing period and model execution time can be reduced, thereby contributing to model optimization. As an additional advanced method of HQPF 1 and 2 mentioned above, a simulated analysis of the Local ENsemble prediction System (LENS) ensemble member and low pressure, one of the observed meteorological factors, was analyzed. Based on the results of this study, if we select for the positively performing ensemble members based on the heavy rain characteristics of Korea or apply additional weights differently for each ensemble member, the prediction accuracy is expected to increase.

In meteorological data, various studies are being conducted to improve the prediction performance of rainfall with irregular patterns, unlike temperature and solar radiation with certain patterns. Especially in the case of the short-term forecast model for Dong-Nae Forecasts provided by the Korea Meteorological Administration (KMA), forecast data are provided at 6-hour intervals, and there is a limit to analyzing the impact of disasters. In this study, Hydrological Quantitative Precipitation Forecast (HQPF) information was generated by applying the machine learning method to Local ENsemble prediction system (LENS), Radar-AWS Rainrates (RAR), AWS and ASOS observation data and Dong-Nae Forecast provided by the KMA. Through the preprocessing process, the temporal and spatial resolutions of all the data were converted to the same resolution, and the predictor of machine learning was derived through the factor analysis of the predictor. Considering the processing speed and expandability, the XGBoost method of machine learning was applied, and the Probability Matching (PM) method was applied to improve the prediction accuracy of heavy rainfall. As a result of evaluating the HQPF performance produced for 14 heavy rainfall events that occurred in 2020, it was found that the predicted performance of HQPF was improved quantitatively and qualitatively.

In this study, the impact of cumulus parameterization usage in Weather Research and Forecasting (WRF) model on reproducing summer precipitation in South Korea is evaluated. Two sensitivity experiments are set up with using cumulus parameterization (ON experiment) and without using cumulus parameterization, which is called Convection Permitting Model (OFF experiment). For the both ON and OFF experiments, the horizontal grid resolution is 2.5km, and initial and lateral boundary conditions are derived from ERA5 reanalysis data. Overall, both of the two experiments can capture the spatial distribution of 2014 summer mean and extreme precipitation but show dry biases in the southern region of Korean Peninsula. Occurrence percentage analyses for different precipitation intensity reveal that OFF experiments show better performance than ON experiment for extreme precipitation. In the case of heavy rainfall over Gyeongnam region for 25 August 2014, OFF experiment shows similar characteristic of rainfall to the observations, although it simulates earlier precipitation peak. On the other hand, ON experiment underestimates the amount of precipitation. Also, vertical distribution of equivalent potential temperature and strong southerly wind which play an important role in developing heavy rainfall on 25 August 2014 are better simulated in OFF experiment.

Due to recent heavy rain events, there are increasing demands for adapting infrastructure design, including drainage facilities in urban basins. Therefore, a clear definition of urban rainfall must be provided; however, currently, such a definition is unavailable. In this study, urban rainfall is defined as a rainfall event that has the potential to cause water-related disasters such as floods and landslides in urban areas. Moreover, based on design rainfall, these disasters are defined as those that causes excess design flooding due to certain rainfall events. These heavy rain scenarios require that the design of various urban rainfall facilities consider design rainfall in the target years of their life cycle, for disaster prevention. The average frequency of heavy rain in each region, inland and coastal areas, was analyzed through a frequency analysis of the highest annual rainfall in the past year. The potential change in future rainfall intensity changes the service level of the infrastructure related to hand-to-hand construction; therefore, the target year and design rainfall considering the climate change premium were presented. Finally, the change in dimensional safety according to the RCP8.5 climate change scenario was predicted.

In this study, we analyzed the characteristics of climate variability in summer rainfall during Changma over three sub-sector regions (Middle, Southern, Jeju) in South Korea for the new climatological period of 1991- 2020 using observation data from 60 ASOS stations. There was a significant interannual variability in rainfall, wet days, and rainfall intensity but the long-term trend of rainfall was not significant over the three sectors in South Korea. Comparing the new climatology (P2: 1991-2020) with the old one (P1: 1981-2010), it was found that the precipitation during Changma in new climatology (P2) was enhanced in Middle sector but reduced in Southern and Jeju sectors. In P2, wet days increased only a few stations in the Middle sector but the rainfall intensity was strengthened over the 50% stations including Middle sector, south and west coast of the Southern sector. Wet days above 25, 50, 75, 95%ile rainfall during Changma in Southern and Jeju sectors all decreased in P2. Climatological change from P1 to P2 showed a large variability not only in temporal frame but also in the spatial distribution in South Korea.

The determination of soil parameters is important in predicting the simulated surface runoff using either a distributed or a lumped rainfall-runoff model. Soil characteristics can be collected using remote sensing techniques and represented as a digital map. There is no universal agreement with respect to the determination of a representative parameter from a gridded digital map. Two representative methods, i.e., arithmetic and predominant, are introduced and applied to both FLO-2D and HEC-HMS to improve the model’s accuracy. Both methods are implemented in the Yongdam catchment, and the results show that the former seems to be more accurate than the latter in the test site. This is attributed to the high conductivity of the dominant soil class, which is A type.

The determination of soil characteristics is important in the simulation of rainfall runoff using a distributed FLO-2D model in catchment analysis. Digital maps acquired using remote sensing techniques have been widely used in modern hydrology. However, the determination of a representative parameter with spatial scaling mismatch is difficult. In this investigation, the FLO-2D rainfall-runoff model is utilized in the Yongdam catchment to test sensitivity based on three different methods (mosaic, arithmetic, and predominant) that describe soil surface characteristics in real systems. The results show that the mosaic method is costly, but provides a reasonably realistic description and exhibits superior performance compared to other methods in terms of both the amount and time to peak flow.

The long-term variability of summer heavy rainfall in the Seoul metropolitan area was investigated in this study for the period of 1970-2018. The study period was divided into two phases; first phase from 1970 to 1994 and the second phase was 1995 to 2018. Long-term variability of summer heavy rainfall was examined using the change-point analysis method. Annual mean heavy rainfall amounts showed increasing trends in Seoul and Incheon during summer monsoon season (June to August). Results revealed that the changes in frequency and amount of heavy rainfall were observed larger in the months of July and August as compare to June during the second phase. The upper-level trough was prominently developed at the west of the Seoul metropolitan area and the core region of the upper-level jet was shifted to the east of the area during second phase. The western North Pacific subtropical high was expanded westward and moisture flux flowed along the southwesterly wind, resultant increasing moisture supply. The temperature and humidity tended to increase recently at the lower and mid-levels.

The purpose of this study was to evaluate the method of estimating the areal precipitation reflecting the altitude of the mountainous terrain on Jeju Island by comparing and analyzing the areal precipitation using the Thiessen polygon method and the isohyetal method in mountainous streams. In terms of constructing the Thiessen polygon network, rainfall errors occurred in 94.5% and 45.8% of the Thiessen area ratio of the Jeju and Ara stations, respectively. This resulted in large areal precipitation and errors using the isohyetal method at altitudes below 600 m in the target watershed. In contrast, there were small errors in the highlands. Rainfall errors occurred in 18.91% of the Thiessen area ratio of Eorimok, 2.41% of Witseoreum, and 2.84% of Azalea Field because of the altitudinal influence of stations located in the highlands at altitudes above 600 m. Based on the areal precipitation estimation using the Thiessen polygon method, it was considered to be partially applicable to streams on Jeju Island depending on the altitude. However, the method is not suitable for mountainous streams such as the streams on Jeju Island because errors occur with altitude. Therefore, the isohyetal method is considered to be more suitable as it considers the locations of the rainfall stations and the orographic effect and because there are no errors with altitude.

기존의 Markov Chain 모형으로 일강우량 모의시에 강우의 발생여부를 모의하고 강우일의 강우량은 Monte Carlo 시뮬레이션을 통해 일강우 분포 특성에 맞는 분포형에서 랜덤으로 강우량을 추정하는 것이 일반적이다. 이때 강우 지속기간에 따른 강도 및 강우의 시간별 분포 등의 강우 사상의 특성을 반영할 수 없다는 한계가 있다. 본 연구에서는 이를 개선하기 위해 강우 사상을 1일 지속강우, 2일 지속강우, 3일 지속강우, 4일이상 지속강우로 구분하여 강우의 지속기간에 따라 강우량을 추정하였다. 즉 강우 사상의 강우 지속일별로 총강우량의 분포형을 비매개변수 추정이 가능한 핵 밀도추정(Kernel Density Estimation, KDE)를 적용하여 각각 추정하였고, 강우가 지속될 경우에 지속일별로 해당하는 분포형에서 강우량을 구하였다. 각 강우사상에 대해 추정된 총 강우량은 k-최근접 이웃 알고리즘(k-Nearest Neighbor algorithm, KNN)을 통해 관측 강우자료에서 가장 유사한 강우량을 가지는 강우사상의 강우량 일분포 형태에 따라 각 일강우량으로 분배하였다. 본 연구는 기존의 강우량 추정 방법의 한계점을 개선하 고자 하였으며, 연구 결과는 미래 강우에 대한 예측에도 활용될 수 있으며 수자원 설계에 있어서 기초자료로 활용될 수 있을 것으로 기대된다.

본 연구에서는 기존의 정량적인 강수량 정보를 제공하는 방식에서 벗어나 호우발생에 따른 생활환경의 변화에 끼치는 영향을 고려한 호우영향예 보서비스의 필요성을 기반으로 호우위험영향도 평가가 가능한 호우재해 위험영향 매트릭스를 개발하고, 이를 통해 호우위험영향을 평가하는 방법을 제시하였다. 사당동 일대를 대상으로 실제 발생 호우사상(2011년 7월 27일)을 적용하였으며, 호우에 의한 침수로 영향을 받는 대상별(사람, 교통, 시설) 호우위험영향평가를 수행하였다. 이를 위해 1 km 격자기반으로 호우위험정도(Impact Level)를 산정하고, 침수심 결과를 조합하여 격자 기반의 잠재호우위험영향(Potential Risk Impact)을 산정하였다. 여기에 강우발생가능성 Likelihood와의 조합을 통해 호우영향예보가 가능한 호우위험영향(Heavy Rainfall Risk Impact) 값을 산정하여 사당동 지역의 호우영향정도를 격자기반으로 4개의 등급으로 분석, 제시하였다.

한국에서 현재 사용되고 있는 홍수예보모형은 집중형 강우-유출모형을 적용하여 유역의 유출을 계산하고 하도 및 저수지 추적모형 등을 활용하여 하천의 수위를 예측한다. 집중형 모형은 유역을 동질의 배수구역으로 가정한다. 따라서 유역내의 다양한 공간적 특성을 고려하지 못한다는 단점이 있다. 또한, 사용되는 강우자료도 지점강우를 활용하기 때문에 공간적인 분포를 자세히 고려하지 못한다는 한계가 있다. 따라서 홍수예보모형에 분포형 모형을 적용하기 위한 연구가 다양하게 진행되고 있다. 본 연구에서는 GRM모형을 한국 홍수예보시스템에 적용하기 위해 모형의 다양한 해상도에 따른 유역유출의 결과의 차이를 분석하여 최적의 해상도를 결정하고자 한다. 모형의 격자가 너무 조밀한 경우 계산시간이 과다하게 되어 홍 수예보모형에 적용하기에는 적합하지 않다. 너무 성길 경우에도 분포형 모형을 적용하여 공간적인 분포를 파악하고자 하는 목적에 맞지 않게 된다. 본 연구의 결과로 유역유출 예측의 정확성을 만족시키고 홍수예보에 적합한 계산속도가 나올 수 있는 최적 해상도를 제시하였다. 유출량 예측의 정 확도는 Nash-Sutcliffe model efficiency coefficient (NSE) 값의 비교를 통해 분석하였다. 본 연구에서 도출된 최적해상도 산정 결과는 분포형 유 역유출모형을 홍수예보모형에 적용하기 위한 기초자료로 활용될 것이다.

Two main sources of data, meteorological data and land surface characteristics, are essential to effectively run a distributed rainfall-runoff model. The specification and averaging of the land surface characteristics in a suitable way is crucial to obtaining accurate runoff output. Recent advances in remote sensing techniques are often being used to derive better representations of these land surface characteristics. Due to the mismatch in scale between digital land cover maps and numerical grid sizes, issues related to upscaling or downscaling occur regularly. A specific method is typically selected to average and represent the land surface characteristics. This paper examines the amount of flooding by applying the FLO-2D routing model, where vegetation heterogeneity is manipulated using the Manning’s roughness coefficient. Three different upscaling methods, arithmetic, dominant, and aggregation, were tested. To investigate further, the rainfall-runoff model with FLO-2D was facilitated in Yongdam catchment and heavy rainfall events during wet season were selected. The results show aggregation method provides better results, in terms of the amount of peak flow and the relative time taken to achieve it. These rwsults suggest that the aggregation method, which is a reasonably realistic description of area-averaged vegetation nature and characteristics, is more likely to occur in reality.

For the purposes of enhancing usability of Numerical Weather Prediction (NWP), the quantitative precipitation prediction scheme by machine learning has been proposed. In this study, heavy rainfall was corrected for by utilizing rainfall predictors from LENS and Radar from 2017 to 2018, as well as machine learning tools LightGBM and XGBoost. The results were analyzed using Mean Absolute Error (MAE), Normalized Peak Error (NPE), and Peak Timing Error (PTE) for rainfall corrected through machine learning. Machine learning results (i.e. using LightGBM and XGBoost) showed improvements in the overall correction of rainfall and maximum rainfall compared to LENS. For example, the MAE of case 5 was found to be 24.252 using LENS, 11.564 using LightGBM, and 11.693 using XGBoost, showing excellent error improvement in machine learning results. This rainfall correction technique can provide hydrologically meaningful rainfall information such as predictions of flooding. Future research on the interpretation of various hydrologic processes using machine learning is necessary.

This study investigates the performance of four Bayesian methods, Random Walk Metropolis (RWM), Hit-And-Run Metropolis (HARM), Adaptive Mixture Metropolis (AMM), and Population Monte Carlo (PMC), for estimating the parameters and uncertainties of probability rainfall distribution, and the results are compared with those of conventional parameter estimation methods; namely, the Method Of Moment (MOM), Maximum Likelihood Method (MLM), and Probability Weighted Method (PWM). As a result, Bayesian methods yield similar or slightly better results in parameter estimations compared with conventional methods. In particular, PMC can reduce parameter uncertainty greatly compared with RWM, HARM, and AMM methods although the Bayesian methods produce similar results in parameter estimations. Overall, the Bayesian methods produce better accuracy for scale parameters compared with the conventional methods and this characteristic improves the accuracy of probability rainfall. Therefore, Bayesian methods can be effective tools for estimating the parameters and uncertainties of probability rainfall distribution in hydrological practices, flood risk assessment, and decision-making support.