• Proceedings of the Korea Water Resources Association Conference
• /
• 2002.05a
• /
• pp.43-50
• /
• 2002

Accurate quantitative forecasting of rainfall for basins with a short response time is essential to predict streamflow and flash floods. Previously, neural networks were used to develop a Quantitative Precipitation Forecasting (QPF) model that highly improved forecasting skill at specific locations in Pennsylvania, using both Numerical Weather Prediction (NWP) output and rainfall and radiosonde data. The objective of this study was to improve an existing artificial neural network model and incorporate the evolving structure and frequency of intense weather systems in the mid-Atlantic region of the United States for improved flood forecasting. Besides using radiosonde and rainfall data, the model also used the satellite-derived characteristics of storm systems such as tropical cyclones, mesoscale convective complex systems and convective cloud clusters as input. The convective classification and tracking system (CCATS) was used to identify and quantify storm properties such as life time, area, eccentricity, and track. As in standard expert prediction systems, the fundamental structure of the neural network model was learned from the hydroclimatology of the relationships between weather system, rainfall production and streamflow response in the study area. The new Quantitative Flood Forecasting (QFF) model was applied to predict streamflow peaks with lead-times of 18 and 24 hours over a five year period in 4 watersheds on the leeward side of the Appalachian mountains in the mid-Atlantic region. Threat scores consistently above .6 and close to 0.8 ∼ 0.9 were obtained fur 18 hour lead-time forecasts, and skill scores of at least 4％ and up to 6％ were attained for the 24 hour lead-time forecasts. This work demonstrates that multisensor data cast into an expert information system such as neural networks, if built upon scientific understanding of regional hydrometeorology, can lead to significant gains in the forecast skill of extreme rainfall and associated floods. In particular, this study validates our hypothesis that accurate and extended flood forecast lead-times can be attained by taking into consideration the synoptic evolution of atmospheric conditions extracted from the analysis of large-area remotely sensed imagery While physically-based numerical weather prediction and river routing models cannot accurately depict complex natural non-linear processes, and thus have difficulty in simulating extreme events such as heavy rainfall and floods, data-driven approaches should be viewed as a strong alternative in operational hydrology. This is especially more pertinent at a time when the diversity of sensors in satellites and ground-based operational weather monitoring systems provide large volumes of data on a real-time basis.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.40 no.3
• /
• pp.303-314
• /
• 2020

The hydrological characteristics of watersheds and hydraulic systems of urban and river floods are highly nonlinear and contain uncertain variables. Therefore, the predicted time series of rainfall-runoff data in flood analysis is not suitable for existing neural networks. To overcome the challenge of prediction, a NARX (Nonlinear Autoregressive Exogenous Model), which is a kind of recurrent dynamic neural network that maximizes the learning ability of a neural network, was applied to forecast a flood in real-time. At the same time, NARX has the characteristics of a time-delay neural network. In this study, a hydrological model was constructed for the Taehwa river basin, and the NARX time-delay parameter was adjusted 10 to 120 minutes. As a result, we found that precise prediction is possible as the time-delay parameter was increased by confirming that the NSE increased from 0.530 to 0.988 and the RMSE decreased from 379.9 ㎥/s to 16.1 ㎥/s. The machine learning technique with NARX will contribute to the accurate prediction of flow rate with an unexpected extreme flood condition.

• Journal of Korean Society of Environmental Engineers
• /
• v.36 no.5
• /
• pp.367-377
• /
• 2014

Paldang lake which is the most important water resource in Korea is classified as a stream type reservoir and water quality of Paldang lake can be significantly influenced by external pollutant source. So this study was aimed to determine focused control BOD and TP sources of each unit watershed upper Paldang lake through analysis of pollutant source distribution and pollutant runoff characteristics. Generated load, discharge load, delivery load and each load density of 11 unit watersheds upper Paldang lake were calculated using data of water quality and flow rate from pollutant sources and 74 small streams. As a result of generated load, discharge load and delivery load of BOD and TP from pollutant sources, the most BOD generated load was taken by livestock with 66% of total BOD discharge load and domestic had the most BOD discharge load, 42.7%. The ratio of delivery load of livestock and domestic was 36.4% and 34.3%, respectively. Livestock occupied high ratio of TP generated load, discharge load and delivery load with 82.5%, 44.4% and 46.7%, respectively. Gyeongan watershed which had high population density showed the highest BOD delivery load density of $14.6kg/km^2/d$ and the highest TP delivery load density with $1.23kg/km^2/d$ was analyzed in Cheongmi watershed including the biggest number of livestock. From these results, management of domestic sewer and livestock excrement was determined as a focused control pollutant source. And intensive management about domestic sewer in Gyeongan stream and livestock excrement in Cheongmi stream is required for water quality improvement of Paldang lake.

• Journal of Korean Society of Environmental Engineers
• /
• v.30 no.8
• /
• pp.760-767
• /
• 2008

A watershed model was constructed using HSPF(Hydrological Simulation Program - Fortran) for predicting flow and suspended solid in the Imha dam watershed. The whole watershed was divided into 33 sub-watersheds in the watershed model, which was calibrated for flow using measured data from 2001 to 2007. The accuracy of watershed model prediction was evaluated using statistical coefficients of R$_{eff}$(Nash-Sutcliffe), R$^2$(Correlation coefficient) and graphical comparison. Then, the model was calibrated for suspended solid using field data measured during 3 major rainfall events in July 2006, and then validated against data obtained in 2 rainfall events from July to August in 2007. Overall, the model showed good agreements with the field measurements for flow and suspended solid. The watershed model constructed in this study can provide flow and suspended solid entering the Imha reservoir and will be utilized for turbid water management in linkage with reservoir water quality models.

• Journal of the Korean Society of Hazard Mitigation
• /
• v.10 no.4
• /
• pp.95-104
• /
• 2010

The storage function model is one of the most commonly used models for flood forecasting and warning system in Korea. This paper studies the physical significance of the storage function model by comparing it with kinematic wave model. The results showed universal applicability of the storage function model to Korean basins. Through a comparison of the basic equations for the models, the storage function model parameters, K, P and $T_l$, are shown to be related with the kinematic wave model parameters, k and p. The analysis showed that P and p are identical and K and $T_l$ can be related to k, basin area, and coefficients of Hack's law. To apply the storage function model throughout the southern part of Korean peninsular, regional parameter relationships for K and $T_l$ were developed for watershed area using data from 17 watersheds and 101 flood events. These relationships combine the kinematic wave parameters with topographic information using Hack's Law.

• Water for future
• /
• v.26 no.1
• /
• pp.115-124
• /
• 1993

This study is to determine the critical duration of design rainfall and to utilize it for the design of detention pond with pump station. To examine the effect of the duration and temporal distribution of the design rainfall, Huff's quartile method is used for the 9 cases of durations ranging from 20 to 240 minutes with 10 years return period, and the ILLUDAS model is used for runoff analysis. The storage ration which is the ratio of maximum storage amounts to total runoff volume, is introduced to determine the critical duration of design rainfall. The duration which maximizes the storage ratio is adopted as the critical duration. This study is applied to 18 urban drainage watersheds with pump station in Seoul, of which the range of watershed area is $0.24-12.70\textrm{km}^2.$ The result of simulation shows that the duration which maximizes storage ration is 30 and 60 minutes on the whole. It is shown also that the storage ration of 2nd- and 3rd-quartile pattern is larger than that of 1st- and 4th-quartile pattern of temporal distribution. A simplified empirical formula for Seoul area is suggested by using the regression analysis between the maximum storage ration and the peak ratio, and can be utilized for the preliminary design and planning of detention pond with pump station.

• Journal of Korea Water Resources Association
• /
• v.39 no.7 s.168
• /
• pp.593-603
• /
• 2006

The basin response to storm is regarded as nonlinearity inherently. In addition, the consistent nonlinearity of hydrologic system response to rainfall has been very tough and cumbersome to be treated analytically. The thing is that such nonlinear models have been avoided because of computational difficulties in identifying the model parameters from recorded data. The parameters of nonlinear system considered as dynamic effects in the conceptual model are optimized as the sum of errors between the observed and computed runoff is minimized. For obtaining the optimal parameters of functions, the historical data for the Bocheong watershed in the Geum river basin were tested by applying the numerical methods, such as quasi-linearization technique, Runge-Kutta procedure, and pattern-search method. The estimated runoff carried through from the storage function with dynamic effects was compared with the one of 1st-order differential equation model expressing just nonlinearity, and also done with Nash model. It was found that the 2nd-order model yields a better prediction of the hydrograph from each storm than the 1st-order model. However, the 2nd-order model was shown to be equivalent to Nash model when it comes to results. As a result, the parameters of nonlinear 2nd-order differential equation model performed from the present study provided not only a considerable physical meaning but also a applicability to Korean watersheds.

• Journal of Korea Water Resources Association
• /
• v.42 no.1
• /
• pp.21-31
• /
• 2009

The characteristics, intensity and direction of groundwater.surface water interactions are controlled by groundwater head gradients, hydraulic conductivity and by the riverbed geometry. As a result of the spatial heterogeneity of these factors and the subsequent variability of the impact of these interaction processes, the water balance is also characterized by highly variable spatial patterns and temporal dynamics. However, spatially detailed studies concerning the spatio-temporal variability of the extent and intensity of surface-groundwater interactions have been limited to the investigation of cross-sections or small stream reaches. Thus, the extensive study on the watershed based interaction between surface water and groundwater is to be analyzed. In this study, the intensity and the spatial extent of interactions along the stream were found by using integrated SWAT-MODFLOW model. This integrated modeling approach was applied to Anyangcheon watershed in Korea. The effluent stream characteristics were found in the watershed, namely, baseflow was annually discharged except heavy rainy periods. The intensity and the spatial extent of surface-groundwater interactions in different sub-watersheds were found on a daily basis. The influential extent of surface-groundwater interaction become larger as the watershed elevations are lower.

• Journal of Korea Water Resources Association
• /
• v.49 no.11
• /
• pp.951-959
• /
• 2016

This study was conducted to evaluate observed runoff data collected every 10 minutes at stream gauging stations in Jeju Island using a physically-based model, SWAT. The Hancheon watershed was selected as study area, and ephemeral stream algorithm suggested by previous research was incorporated into the model, which is able to simulate ephemeral runoff pattern of Jeju streams. Simulated runoff and runoff rates were compared to observations during 2008-2013, which showed 'very good' performance rating in Nash-Sutcliffe model efficiency (ME) and determination coefficient ($R^2$). Some observations had problems such that runoff rates were very high for some rainfall events with little amount of antecedent rainfall, and were very low or missing with much rainfall comparing to previous researches. Additionally, regression equation between precipitation and simulated runoff was generated with high degree of correlation. The equation can be utilized to simply predict reasonable runoff, or to investigate and complement the abnormal or missing data of observations on the assumption that modelling results were sufficiently reliable and satisfactory. As results, minimizing the error in calibrating the model by evaluation of observed data would be helpful to accurately model the rainfall-runoff characteristics and analyze the water balance components of watersheds in Jeju Island.

• Journal of Ecology and Environment
• /
• v.29 no.3
• /
• pp.305-321
• /
• 2006

Atmospheric Acid Deposition: Nitrogen Saturation of Forests: Volume weighted annual average wet deposition of nitroge at 33 sites in Korea during 1999-2004 ranged 7.28 to $21.05kgN{\cdot}ha^{-1}{\cdot}yr^{-1}$ with average $12.78kgN{\cdot}ha^{-1}{\cdot}yr^{-1}$, which values are similar level with nitrogen deposition of Europe and North America. The temperate forests that suffered long-term high atmospheric nitrogen deposition are gradually saturated with nitrogen. Such nitrogen saturated forest watersheds usually leach nitrate ion ($NO_3^-$) in stream water and soil solution. It may be likely that Korean forest ecosystems are saturated by much nitrogen deposition. In leaves with nitrogen saturation ratios of N/P, N/K and N/Mg are so enhanced that mineral nutrient system is disturbed, suffered easily frost damage and blight disease, reduced fine-root vitality and mycorrhizal activity. Consequently nitrogen saturated forests decrease primary productivity and finally become forest decline. Futhermore understory species are replaced the nitrophobous species by the nitrophilous one. In soil with nitrogen saturation uptake of methane ($CH_4$) is reduced and emission of nitrogen monoxide (NO) and nitrous oxide ($N_{2}O$) are increased, which gases are greenhouse gas accelerating global warming.