• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2021년도 학술발표회
• /
• pp.373-373
• /
• 2021

Rainfall forecasting is an important issue that is applied in many areas, such as agriculture, flood warning, and water resources management. In this context, this study proposed a statistical and machine learning-based forecasting model for monthly rainfall. The Bayesian Gaussian process was chosen to optimize the hyperparameters of the Stacked Long Short-term memory (SLSTM) model. The proposed SLSTM model was applied for predicting monthly precipitation of Seoul station, South Korea. Data were retrieved from the Korea Meteorological Administration (KMA) in the period between 1960 and 2019. Four schemes were examined in this study: (i) prediction with only rainfall; (ii) with deseasonalized rainfall; (iii) with rainfall and minimum temperature; (iv) with deseasonalized rainfall and minimum temperature. The error of predicted rainfall based on the root mean squared error (RMSE), 16-17 mm, is relatively small compared with the average monthly rainfall at Seoul station is 117mm. The results showed scheme (iv) gives the best prediction result. Therefore, this approach is more straightforward than the hydrological and hydraulic models, which request much more input data. The result indicated that a deep learning network could be applied successfully in the hydrology field. Overall, the proposed method is promising, given a good solution for rainfall prediction.

• Water Engineering Research
• /
• 제5권2호
• /
• pp.81-99
• /
• 2004

Long term streamflow regime under virtual climate change scenario was examined. Rainfall forecast simulation of the Canadian Global Coupled Model (CGCM2) of the Canadian Climate Center for modeling and analysis for the IPCC SRES B2 scenario was used for analysis. The B2 scenario envisions slower population growth (10.4 billion by 2010) with a more rapidly evolving economy and more emphasis on environmental protection. The relatively large scale of GCM hinders the accurate computation of the important streamflow characteristics such as the peak flow rate and lag time, etc. The GCM rainfall with more than 100km scale was downscaled to 2km-scale using the space-time stochastic random cascade model. The HEC-HMS was used for distributed hydrologic model which can take the grid rainfall as input data. The result illustrates that the annual variation of the total runoff and the peak flow can be much greater than rainfall variation, which means actual impact of rainfall variation for the available water resources can be much greater than the extent of the rainfall variation.

• 한국농공학회지
• /
• 제17권3호
• /
• pp.3848-3859
• /
• 1975

The author attempted to find most suitable formulas for probable rainfall intensities with analysis and consideration for characteristics of rainfall intensities according to the short and long period return periods at Cheong-Joo district. Above mentioned formulas induced by this study can be contributed to the credibility of runoff estimation for urban sewerage system, drainage works in small catchment area and embankment works in the rivers. The results of this study are summarized as follows: 1 Calculation values by Gumbel-Chow method were selected as a mean values for the calculation of probable rainfall intensities according to return periods in the short period. 2. Calculations for probable rainfall intensities for long period are based upon to the result by Iwai's method. Talbot type, {{{{I= {a} over {t+b} }}}} is confirmed as a most suitable formula for probable rainfall intensities among calculation methods in the short periods at Cheong-Joo district. 4. Specific coefficient method, I24=RN24${\beta}$N was selected as a means of calculation for suitable formulas of probable rainfall intensities according to return periods in case of long period. 5. Runoff estimation with high credibility by rational formula can be anticipated by establishment for the most suitable probable rainfall intensities at Cheong-Joo district.

• 한국수자원학회논문집
• /
• 제37권7호
• /
• pp.553-567
• /
• 2004

본 연구에서는 SPI의 문제점을 보완하는 측면에서 수정 SPI(MSPI)를 제안하고, 이를 서울지점의 가물분석에 적용하여 비교하였다. MSPI는 강수량을 이동평균하기 전에 정규화 하는 과정을 추가하는 경우로 SPI가 강수의 절대량을 이월하는 반면 MSPI는 강수의 상대량을 이월하는 형태를 가지게 된다. 서울지점 월강수량자료는 1777년-1996년까지의 것으로 각각 전체자료 및 1900년을 전후로 한 장기건조기의 전반부 및 후반부를 따로 분석하여 비교될 수 있도록 하였다. 그 결과를 정리하면 다음과 같다. (1) MSPI가 SPI에 비해 보다 현실적인 결과를 준다. 이는 1900년도 전후로의 장기건조기가 포함된 기록과 포함되지 않은 기록을 분석한 결과의 비교에서 확인되었다. (2) MSPI는 SPI에 비해 극단적인 강수사상에 덜 민감한 결과를 준다. 장기건조기의 전반부 및 후반부에 대해 MSPI를 적용한 결과를 비교하면 가뭄의 발생가능성은 보통 가뭄의 경우는 대체로 비슷하나 아주 심한 가뭄의 경우는 약간 줄어들었음을 파악할 수 있다(재현기간 약 18년에서 16년으로). 그러나 가뭄 발생 시 그 지속기간은 전반부에 비해 후반부가 길게 나타나고 있다(아주 심한 가뭄의 경우 약 2개월에서 2.5개월로). 이러한 결과는 상대적으로 극단적인 결과로 나타내는 SPI(아주 심한 가뭄의 경우 재현기간은 약 25년에서 10년으로, 지속기간은 1.5개월에서 3.5개월로)에 비해 보다 현실적이라고 판단할 수 있다.

• 물과 미래
• /
• 제5권2호
• /
• pp.30-43
• /
• 1972

The paper describes on the hydrologic analysis of point rainfall data of the three major areas, such as in Seoul, Pusan and Taegu. Scheme of the paper is analyzed five research cases. Contents of the analysis are carried out five kinds of transformed variables for determination of rainfall distribution types and two kinds of reliability tests on unusual(extraordinary) values each rainfall durations:short durations, long durations, long durations, monthly and yearly. Rainfall depth probability had been computed methods of hydrologic amounts analysis namely logarithmic transformations or Gumbel-Chow method and so on, but in this paper it is calculated log xi, n-square root transformations by using normal distribution function and normalization of rainfall distributions is examined graphical tests and $X^2-test$(chi-square test). Furthermore, rainfall depth probability is calculated taking into account the safty factors of project life of hydraulic structures. We think it is advanced contents that considering priceless experiences, the life of structures, conditions and more problems of planning engineers and designers, proposed rainfall amounts(proposed values) are presented charts or figures.

• 한국수자원학회논문집
• /
• 제45권5호
• /
• pp.431-444
• /
• 2012

본 연구에서는 강우 지속기간이 짧은 경우와 긴 경우에 대한 연최대치 계열과 연최대치 독립 호우사상 계열의 차이를 살펴보았다. 연최대치 독립호우사상은 1961년부터 2010년까지의 서울지점 시강우 자료에 다양한 IETD와 절단값을 적용하여 추출하였다. 결정된 연최대치 독립호우사상 계열은 기존의 연최대치 계열과 비교하였다. 그 결과를 요약하면 다음과 같다. 첫째, IETD와 절단값의 영향은 예상할 수 있는 수준을 벗어나지 않았다. 예를 들어, IETD의 증가에 따라 짧은 지속기간 및 긴 지속기간을 갖는 독립 호우사상의 발생 빈도는 유사한 감소비율을 갖으나, 절단값의 증가에 따라서는 특히 지속기간이 긴 호우사상의 발생빈도가 급격히 감소하는 것으로 나타났다. 둘째, 추출된 독립 호우사상의 평균 강우강도는 강우 지속기간에 관계없이 거의 일정한 것으로 나타났다. 이는 연 최대치 강우강도가 주로 지속기간이 긴 호우사상에서 결정될 가능성이 크다는 것을 의미한다. 마지막으로, 강우지속기간이 짧은 경우, 연 최대치 계열과 연 최대치 독립 호우사상 계열의 차이가 매우 큰 것으로 확인되었다. 그러나 이러한 차이는 강우 지속기간이 증가함에 따라 현저히 줄어들었다. 이러한 결과는 특히 매우 짧은 집중시간을 고려하는 작은 유역의 경우, 전통적인 자료 분석방법을 통해 결정된 설계강우가 실제 발생 가능성이 없는 비현실적인 경우가 될 가능성이 매우 큼을 의미한다.

• 상하수도학회지
• /
• 제28권6호
• /
• pp.635-645
• /
• 2014

The changes of rainfall pattern and impervious covers have increased disaster risks in urbanized areas. Impervious covers such as roads and building roofs have been dramatically increased. So, it is falling the ability safety of flood defense equipments to exist. Runoff coefficient means ratio of runoff by whole rainfall which is able to directly contribute at surface runoff during rainfall event. The application of accurate runoff coefficients is very important in sewer pipelines design. This study has been performed to estimate runoff characteristics change which are applicable to the process of sewer pipelines design or various public facilities design. It has used the SHER model, a long-term runoff model, to analyze the impact of a rising impervious covers on runoff coefficient change. It thus analyzed the long-term runoff to analyze rainfall basins extraction. Consequently, it was found that impervious surfaces could be a important factor for urban flood control. We could suggest the application of accurate runoff coefficients in accordance to the land Impervious covers. The average increase rates of runoff coefficients increased 0.011 for 1% increase of impervious covers. By having the application of the results, we could improve plans for facilities design.

• 한국수자원학회논문집
• /
• 제49권1호
• /
• pp.41-49
• /
• 2016

본 연구에서는 일강수량의 누적 패턴을 고려하여 새로운 가뭄지수(RADI)를 개발하였다. RADI는 누적 강수량의 장기간의 평균과 특정 기간에 관측된 누적 강수량을 비교하여 일단위로 가뭄지수를 간단히 산정할 수 있도록 개발되었다. 우리나라의 대표적인 가뭄기간 동안의 RADI의 시 공간적 변동성과 재현 주기를 살펴봄으로써 장 단기 가뭄의 감시에 대한 RADI의 적용성을 평가하였다. SPI의 경우, 장 단기 가뭄을 표현하기 위해서는 여러 가지 지속기간을 가지는 SPI를 사용해야 하지만, RADI는 하나의 지수로 단기 및 장기가뭄을 표현하고 모니터링 할 수 있다. RADI의 전국 평균값을 살펴본 결과, 우리나라에서 20년 주기로 가뭄이 발생하고 있다는 사실을 확인하였다. 또한 RADI의 발생확률 및 통계 분석을 통해 5단계의 가뭄 등급을 제안하였으며, 이는 가뭄 예보와 대응을 위한 기준으로 활용될 수 있을 것으로 판단된다.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2004년도 학술발표회
• /
• pp.1290-1294
• /
• 2004

In this study, new stochastic point rainfall models which can consider the correlation structure between rainfall intensity and duration are developed. In order to consider the negative and positive correlation simultaneously, the Gumbels type-II bivariate distribution is applied, and for the cluster structure of rainfall events, the Neyman-Scott cluster point process is selected. In the theoretical point of view, it is shown that the models considering the dependent structure between rainfall intensity and duration have slightly heavier tail autocorrelation functions than the corresponding independent mode]s. Results from generating long time rainfall events show that the dependent models better reproduce historical rainfall time series than the corresponding independent models in the sense of autocorrelation structures, zero rainfall probabilities and extreme rainfall events.

• 한국안전학회지
• /
• 제37권6호
• /
• pp.148-157
• /
• 2022

Every year, particularly during the monsoon rainy season, landslides at the Chuncheon province of South Korea cause tremendous damage to lives, properties, and infrastructures. More so, the high rainfall intensity and long rainfall days that occurred in 2020 have increased the water content in the soil, thereby increasing the chances of landslide occurrences. Besides this, the rainfall thresholds and characteristics responsible for the initiation of landslides in this region have not been properly identified. Therefore, this paper addresses the rainfall thresholds responsible for the initiation of landslides at Chuncheon from a regional perspective. Using data obtained from rainfall measurements taken from 2002 to 2011, we identify a threshold relationship between rainfall intensity and rainfall duration for the initiation of landslides. In addition, we identify the relationship between the rainfall intensity using a 3-day, 7-day, and 10-day antecedent rainfall observation. Specifically, we estimate the rainfall data at 8 sites where debris flow occurred in 2011 by kriging. Following this, the estimated data are used to construct the relationship between the intensity (I), duration (D), and frequency (F) of rainfall. The results of the intensity-duration-frequency (IDF) analysis show that landslides will occur under a rainfall frequency below a 2-year return period at two areas in Chuncheon. These results will be effectively used to design structures that can prevent the occurrence of landslides in the future.