• Proceedings of the Korea Water Resources Association Conference
• /
• 2018.05a
• /
• pp.147-147
• /
• 2018

In this article, we use an open source software library: TensorFlow, developed for the purposes of conducting very complex machine learning and deep neural network applications. However, the system is general enough to be applicable in a wide variety of other domains as well. The proposed model based on a deep neural network model, LSTM (Long Short-Term Memory) to predict the river water level at Okcheon Station of the Guem River without utilization of rainfall - forecast information. For LSTM modeling, the input data is hourly water level data for 15 years from 2002 to 2016 at 4 stations includes 3 upstream stations (Sutong, Hotan, and Songcheon) and the forecasting-target station (Okcheon). The data are subdivided into three purposes: a training data set, a testing data set and a validation data set. The model was formulated to predict Okcheon Station water level for many cases from 3 hours to 12 hours of lead time. Although the model does not require many input data such as climate, geography, land-use for rainfall-runoff simulation, the prediction is very stable and reliable up to 9 hours of lead time with the Nash - Sutcliffe efficiency (NSE) is higher than 0.90 and the root mean square error (RMSE) is lower than 12cm. The result indicated that the method is able to produce the river water level time series and be applicable to the practical flood forecasting instead of hydrologic modeling approaches.

• Journal of Korea Water Resources Association
• /
• v.39 no.10 s.171
• /
• pp.833-844
• /
• 2006

The objective of this study is to evaluate the snowmelt effects on the hydrological components, especially on the runoff, by using the soil water assessment tool(SWAT) which is a continuous semi-distributed long term rainfall-runoff model. The model was applied to the basin located in the upstream of the Chungju Dam. Some parameters in the snowmelt algorithm were estimated for the Chungju basin in order to reflect the snowmelt effects. The snowmelt effects were assessed by comparing the simulated runoff with the observed runoff data at the outlet of the basin. It was found out that the simulated runoff with considering the snowmelt component matches more satisfactorily to the observed one than without considering snowmelt effect. The simulation results revealed that the snowmelt effects were noticeable on March and April. Similar results were obtained at other two upstream gauging points. The effect of the elevation bands which distribute temperature and precipitation with elevation was analyzed. This study also showed that the snowmelt effect significantly affects the temporal distribution as well as quantity of the hydrological components. The simulated runoff was very sensitive to the change of temperature near the threshold temperature which the snowmelt can occur. However, the reason was not accounted for this paper, Therefore, further analyses related to this feature are needed.

• Journal of Korean Society on Water Environment
• /
• v.24 no.3
• /
• pp.345-353
• /
• 2008

In order to design storage-based non-point source management facilities, the aspect of statistical features of the entire precipitation time series should be considered since non-point source pollutions are delivered by continuous rainfall runoffs. The 3-parameter mixed exponential probability density function instead of traditional single-parameter exponential probability density function is applied to represent the probabilistic features of long-term precipitation time series and model urban stormwater runoff. Finally, probability density functions of water quality control basin overflow are derived under two extreme intial conditions. The 31-year continuous precipitation time series recorded in Busan are analyzed to show that the 3-parameter mixed exponential probability density function gives better resolution.

• Ecology and Resilient Infrastructure
• /
• v.3 no.1
• /
• pp.1-7
• /
• 2016

River restoration has recently progressed in consideration of ecological functions along with flood controls and conservation. For river restorations that consider ecological health and diversity, it is important to contemplate the recovery of hydraulic and hydrologic connectivity in isolated spaces by longitudinal structures. In this study, as a first step for the provision of hydraulic and hydrologic data, which is necessary for the ecological connection analysis in isolated spaces, we developed a one-dimensional numerical model for rainfall runoff and channel routing and applied it to the Cheongmi watershed. The developed numerical model can simulate hydraulic and hydrologic analysis at the same time using the rainfall data. Numerical results were compared with observed data and other numerical results. As a result, a very reasonable agreement is observed. The results of this study will be improved so that the long-term hydrologic and hydraulic analysis is possible to predict ecological change.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.33 no.5
• /
• pp.1861-1870
• /
• 2013

A climate change-driven increased hydrological variability has been widely acknowledged over the past decades. In this regards, rainfall simulation techniques are being applied in many countries to consider the increased variability. This study proposed a Homogeneous Hidden Markov Chain(HMM) designed to recognize rather complex patterns of rainfall with discrete hidden states and underlying distribution characteristics via mixture probability density function. The proposed approach was applied to Seoul and Jeonju station to verify model's performance. Statistical moments(e.g. mean, variance, skewness and kurtosis) derived by daily and seasonal rainfall were compared with observation. It was found that the proposed HMM showed better performance in terms of reproducing underlying distribution characteristics. Especially, the HMM was much better than the existing Markov Chain model in reproducing extremes. In this regard, the proposed HMM could be used to evaluate a long-term runoff and design flood as inputs.

• Journal of Korea Water Resources Association
• /
• v.30 no.4
• /
• pp.379-387
• /
• 1997

This study involves field monitoring of hydrlolgic characteristics of paddy fields under common irrigation practice, statistical analysis of maximum retention storage, determination of CNs for antecedent moisture conditions. Curve numbers were estimated from observed rainfall-runoff relationship of two years data. The estimated CN for AMC-II was 78, and the CNs for AMC-I and II were 63 and 88, respectively. A water balance model was used to find the effect of ridge height changes and initial ponding depth in paddy fields on runoff. The probability distribution of initial ponding depth was also investigated. The initial ponding depth follows normal probability distribution. Initial ponding depth corresponding 10%, 50%, and 90% probability were considered to be equivalent to AMC-I, AMC-II, and AMC-III, respectively. Long-term runoff data from paddy fields were simulated by a water balance model using recorded climate data, ridge height and estimated initial ponding depth derived from probability distribution. The estimated CNs using simulated runoff were 70, 79, and 89 for CN-I, CN-II, and CN-III, respectively.

• Journal of Korean Society on Water Environment
• /
• v.35 no.1
• /
• pp.35-42
• /
• 2019

Urbanization has led to extreme changes in land use on urban watersheds. Most cities are becoming residential, commercial and industrial areas, making infiltration and storage of rainfall less favorable. The demand for LID (Low Impact Development) technology is increasing in order to mitigate this water cycle distortion and return to existing hydrological conditions. The LID technique is effective in reducing runoff by permeating the urban impervious area. However, considering the limit of the installation area and the financial requirement of the installation, there is not much research on the linkage of each LID component technology for optimum efficiency according to the appropriate scale. In this study, the effects of the LID facilities applied to the target site were simulated using the SWMM model, suggesting the optimal linkage method considering interconnectivity, and applying the effects as an existing installation of individual facilities. The water balance at the time of application of the LID technology, short-term and long-term rainfall event were compared. Also, the individual application and the linkage application were compared with each other. If each component technology has sufficient processing size, then linkage application is more effective than individual application.

• Journal of Korea Water Resources Association
• /
• v.55 no.5
• /
• pp.333-343
• /
• 2022

For water resource management, the design flood is calculated using the flood frequency analysis technique and the rainfall runoff model. The method by design flood frequency analysis calculates the stochastic design flood by directly analyzing the actual discharge data and is theoretically evaluated as the most accurate method. Actual discharge data frequency analysis of the measured flow was limited due to data limitations in the existing flood flow analysis. In this study, design flood frequency analysis was performed using the measured flow data stably secured through the water level-discharge relationship curve formula. For the frequency analysis of design flood, the parameters were calculated by applying the bayesian inference, and the uncertainty of flood volume by frequency was quantified. It was confirmed that the result of calculating the design flood was close to that calculated by the rainfall-runoff model by applying long-term rainfall data. It is judged that hydrological analysis can be done from various perspectives by using long-term actual flow data through hydrological survey.

• Journal of Korea Water Resources Association
• /
• v.54 no.spc1
• /
• pp.1107-1118
• /
• 2021

Climate change brought on by global warming increased the frequency of flood and drought on the Korean Peninsula, along with the casualties and physical damage resulting therefrom. Preparation and response to these water disasters requires national-level planning for water resource management. In addition, watershed-level management of water resources requires flow duration curves (FDC) derived from continuous data based on long-term observations. Traditionally, in water resource studies, physical rainfall-runoff models are widely used to generate duration curves. However, a number of recent studies explored the use of data-based deep learning techniques for runoff prediction. Physical models produce hydraulically and hydrologically reliable results. However, these models require a high level of understanding and may also take longer to operate. On the other hand, data-based deep-learning techniques offer the benefit if less input data requirement and shorter operation time. However, the relationship between input and output data is processed in a black box, making it impossible to consider hydraulic and hydrological characteristics. This study chose one from each category. For the physical model, this study calculated long-term data without missing data using parameter calibration of the Soil Water Assessment Tool (SWAT), a physical model tested for its applicability in Korea and other countries. The data was used as training data for the Long Short-Term Memory (LSTM) data-based deep learning technique. An anlysis of the time-series data fond that, during the calibration period (2017-18), the Nash-Sutcliffe Efficiency (NSE) and the determinanation coefficient for fit comparison were high at 0.04 and 0.03, respectively, indicating that the SWAT results are superior to the LSTM results. In addition, the annual time-series data from the models were sorted in the descending order, and the resulting flow duration curves were compared with the duration curves based on the observed flow, and the NSE for the SWAT and the LSTM models were 0.95 and 0.91, respectively, and the determination coefficients were 0.96 and 0.92, respectively. The findings indicate that both models yield good performance. Even though the LSTM requires improved simulation accuracy in the low flow sections, the LSTM appears to be widely applicable to calculating flow duration curves for large basins that require longer time for model development and operation due to vast data input, and non-measured basins with insufficient input data.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.29 no.1B
• /
• pp.11-22
• /
• 2009

Long-term rainfall-runoff modeling is a key element in the Earth's hydrological cycle, and associated with many different aspects such as dam design, drought management, river management flow, reservoir management for water supply, water right permission or coordinate, water quality prediction. In this regard, hydrologists have used the hydrologic models for design criteria, water resources assessment, planning and management as a main tool. Most of rainfall-runoff studies, however, were not carefully performed in terms of considering reservoir effects. In particular, the downstream where is severely affected by reservoir was poorly dealt in modeling rainfall-runoff process. Moreover, the effects can considerably affect overall the rainfallrunoff process. An objective of this study, thus, is to evaluate the impact of reservoir operation on rainfall-runoff process. The proposed approach is applied to Anseong watershed, where is in a mixed rural/urban setting of the area and in Korea, and has been experienced by flood damage due to heavy rainfall. It has been greatly paid attention to the agricultural reservoirs in terms of flood protection in Korea. To further investigate the reservoir effects, a comprehensive assessment for the results are discussed. Results of simulations that included reservoir in the model showed the effect of storage appeared in spring and autumn when rainfall was not concentrated. In periods of heavy rainfall, however, downstream runoff increased in simulations that do not consider reservoir factor. Flow duration curve showed that changes in streamflow depending upon the presence or absence of reservoir factor were particularly noticeable in ninety-five day flow and low flow.