• 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.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.12 no.2
• /
• pp.82-98
• /
• 2009

This study applied SWAT model to analyze suspended sediment load that is influence on the high density turbid water in Donghyang and Cheoncheon basin, which are located in the upstream of Yongdam Dam. GIS data such as DEM, land cover map and soil map, and meteorological data were used as the input data of SWAT model. And the rating curve equation and Q-SS equation of Donghyang and Cheoncheon gauge station were applied as the measured values of them. As the result of flowout, the coefficient of determination ($R^2$) and the Nash-Sutcliffe coefficient of efficiency (EI) of model calibration showed high as 0.87 and 0.87 at Donghyang gauge station, and the $R^2$ and EI of model validation were high as 0.95 at Cheoncheon gauge station. Also, as the result of suspended sediment load, the $R^2$ and EI of model calibration were high as 0.77 and 0.76 at Donghyang gauge station, and the $R^2$ and EI of model validation marked high as 0.867 and 0.80 at Cheoncheon gauge station. It is considered that the suspended sediment load of 2003 showed the highest due to rainfall amounts and rainfall intensity in using SWAT model. The results of suspended sediment modeled in this study can be applied to the decision-making support data for the evaluation of soil erosion possibility and turbid water potential in the management of reservoir.

• The Journal of Engineering Geology
• /
• v.23 no.4
• /
• pp.381-387
• /
• 2013

In compiling flood hazard maps for the case of dam-failure, a scenario-based numerical modeling approach is commonly used, involving the modeling of important parameters that capture peak discharge, such as breach formation and progress. In this study, an earth-dam-break model is constructed assuming an identical mechanism and hydraulic process for all dam-break processes. A focus of the analysis is estimation of the hydrograph at the outlet as a function of time. The constructed hydrograph then serves as an upper boundary condition in running the flood routing model downstream, although flood routing is not considered here. Validation was performed using the record of the Tangjishan dam-break in China. The results were satisfactory, with a coefficient of determination of 0.974, Nash-Sutcliffe Coefficient of Efficiency (NSC) of 0.94, and Root Mean Square Error (RMSE) of $610m^3/sec$. The proposed model will contribute to assessments of potential flood hazards caused by dam-break.

• Journal of Biosystems Engineering
• /
• v.43 no.2
• /
• pp.148-159
• /
• 2018

Purpose: Unmanned air vehicle (UAV) remote sensing was applied to test various vegetation indices and make prediction models of protein content of rice for monitoring grain quality and proper management practice. Methods: Image acquisition was carried out by using NIR (Green, Red, NIR), RGB and RE (Blue, Green, Red-edge) camera mounted on UAV. Sampling was done synchronously at the geo-referenced points and GPS locations were recorded. Paddy samples were air-dried to 15% moisture content, and then dehulled and milled to 92% milling yield and measured the protein content by near-infrared spectroscopy. Results: Artificial neural network showed the better performance with $R^2$ (coefficient of determination) of 0.740, NSE (Nash-Sutcliffe model efficiency coefficient) of 0.733 and RMSE (root mean square error) of 0.187% considering all 54 samples than the models developed by PR (polynomial regression), SLR (simple linear regression), and PLSR (partial least square regression). PLSR calibration models showed almost similar result with PR as 0.663 ($R^2$) and 0.169% (RMSE) for cloud-free samples and 0.491 ($R^2$) and 0.217% (RMSE) for cloud-shadowed samples. However, the validation models performed poorly. This study revealed that there is a highly significant correlation between NDVI (normalized difference vegetation index) and protein content in rice. For the cloud-free samples, the SLR models showed $R^2=0.553$ and RMSE = 0.210%, and for cloud-shadowed samples showed 0.479 as $R^2$ and 0.225% as RMSE respectively. Conclusion: There is a significant correlation between spectral bands and grain protein content. Artificial neural networks have the strong advantages to fit the nonlinear problem when a sigmoid activation function is used in the hidden layer. Quantitatively, the neural network model obtained a higher precision result with a mean absolute relative error (MARE) of 2.18% and root mean square error (RMSE) of 0.187%.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.28 no.6B
• /
• pp.709-721
• /
• 2008

This paper is to test the applicability of ModKIMSTORM (Modified KIneMatic Wave STOrm Runoff Model) by applying it to Namgangdam watershed of $2,293km^2$. Model inputs (DEM, land use, soil related information) were prepared in 500 m spatial resolution. Using five typhoon events (Saomi in 2000, Rusa in 2002, Maemi in 2003, Megi in 2004 and Ewiniar in 2006) and two storm events (May of 2003 and July of 2004), the model was calibrated and verified by comparing the simulated streamflow with the observed one at the outlet of the watershed. The Pearson's coefficient of determination $R^2$, Nash and Sutcliffe model efficiency E, the deviation of runoff volumes $D_v$, relative error of the peak runoff rate $EQ_p$, and absolute error of the time to peak runoff $ET_p$ showed the average value of 0.984, 0.981, 3.63%, 0.003, and 0.48 hr for 4 storms calibration and 0.937, 0.895, 8.08%, 0.138, and 0.73 hr for 3 storms verification respectively. Among the model parameters, the stream Manning's roughness coefficient was the most sensitive for peak runoff and the initial soil moisture content was highly sensitive for runoff volume fitting. We could look into the behavior of hyrologic components from the spatial results during the storm periods and get some clue for the watershed management by storms.

• The Journal of Engineering Geology
• /
• v.33 no.4
• /
• pp.627-638
• /
• 2023

This study aimed to evaluate the spatial variability of downstream river flow resulting from the operation of the Gimcheon Buhang Dam in the Gamcheon watershed. The dam's effects on flood reduction during the flood season and on increasing streamflow during the dry season-two main functions of multipurpose dams-were quantitatively analyzed. Streamflow data from 2013 to 2021 for the study waterhsed were simulated on a daily basis using SWAT-K (Soil and Water Assessment Tool - Korea) model. Comparison of the simulated and observed values found goodness of fit values of 0.75 or higher for both the coefficient of determination and the Nash-Sutcliffe model efficiency coefficient. The spatial analysis of the dam's effect on flood reduction focused on the annual maximum flood: rates of flood reduction at the four stations ranged from 8.5% to 25.0%. The evaluation of streamflow increase during times of low flow focused on flow duration curves: in particular, compared to the case without an upstream dam, the average low flow at the four sites increased from 33% to 198%.

• Journal of Korea Water Resources Association
• /
• v.48 no.11
• /
• pp.925-936
• /
• 2015

The purpose of this study is to explore the effectiveness of dual-polarization radar rainfall by comapring with the flood inundation record map through KIMSTORM(Grid-based KIneMatic wave STOrm Runoff Model). For Namgang dam ($2,293km^2$) watershed, the Bisl dual-polarization radar data for 3 typhoons (Khanun, Bolaven, Sanba) and 1 heavy rain event in 2012 were prepared. For both 28 ground rainfall data and radar rainfall data, the model was calibrated using observed discharge data at 5 stations with $R^2$, Nash and Sutcliffe Model Efficiency (ME) and Volume Conservation Index (VCI). The calibration results of $R^2$, ME and VCI were 0.85, 0.78 and 1.09 for ground rainfall and 0.85, 0.79, and 1.04 for radar rainfall respectively. The flood inundation record areas (SY and MD/SG district) by typhoon Sanba were compared with the distributed modeling results. The spatial distribution by radar rainfall produced more surface runoff from the watershed and simulated higher stream discharge than the ground rainfall condition in both SY and MD/SG district. In case of MD/SG district, the stream water level by radar rainfall near the flood inundation area showed 0.72 m higher than the water level by ground rainfall.

• Journal of Korea Water Resources Association
• /
• v.48 no.10
• /
• pp.793-806
• /
• 2015

The simulation of natural streamflow at ungauged basins is one of the fundamental challenges in hydrology community. The key to runoff simulation in ungauged basins is generally involved with a reliable parameter estimation in a rainfall-runoff model. However, the parameter estimation of the rainfall-runoff model is a complex issue due to an insufficient hydrologic data. This study aims to regionalize the parameters of a continuous rainfall-runoff model in conjunction with a Bayesian statistical technique to consider uncertainty more precisely associated with the parameters. First, this study employed Bayesian Markov Chain Monte Carlo scheme for the estimation of the Sacramento rainfall-runoff model. The Sacramento model is calibrated against observed daily runoff data, and finally, the posterior density function of the parameters is derived. Second, we applied a multiple linear regression model to the set of the parameters with watershed characteristics, to obtain a functional relationship between pairs of variables. The proposed model was also validated with gauged watersheds in accordance with the efficiency criteria such as the Nash-Sutcliffe efficiency, index of agreement and the coefficient of correlation.

• Journal of Soil and Groundwater Environment
• /
• v.26 no.6
• /
• pp.106-117
• /
• 2021

The purpose of this study is prediction of landslide occurrence reflecting the subsurface flow characteristics within the soil layer in the future due to climate change in a large scale watershed. To do this, we considered the infinite slope stability theory to evaluate the landslide occurrence with predicted soil moisture content by SWAT model based on monitored data (rainfall-soil moisture-discharge). The correlation between the SWAT model and the monitoring data was performed using the coefficient of determination (R²) and the model's efficiency index (Nash and Sutcliffe model efficiency; NSE) and, an accuracy analysis of landslide prediction was performed using auROC (area under Receiver Operating Curve) analysis. In results comparing with the calculated discharge-soil moisture content by SWAT model vs. actual observation data, R² was 0.9 and NSE was 0.91 in discharge and, R² was 0.7 and NSE was 0.79 in soil moisture, respectively. As a result of performing infinite slope stability analysis in the area where landslides occurred in the past based on simulated data (SWAT analysis result of 0.7~0.8), AuROC showed 0.98, indicating that the suggested prediction method was resonable. Based on this, as a result of predicting the characteristics of landslide occurrence by 2050 using climate change scenario (RCP 8.5) data, it was calculated that four landslides could occur with a soil moisture content of more than 75% and rainfall over 250 mm/day during simulation. Although this study needs to be evaluated in various regions because of a case study, it was possible to determine the possibility of prediction through modeling of subsurface flow mechanism, one of the most important attributes in landslide occurrence.

• Journal of Korea Water Resources Association
• /
• v.42 no.1
• /
• pp.33-50
• /
• 2009

The impact on hydrologic components considering future potential climate, land use change and vegetation cover information was assessed using SLURP (Semi-distributed Land-Use Runoff Process) continuous hydrologic model. The model was calibrated (1999 - 2000) and validated (2001 - 2002) for the upstream watershed ($260.4\;km^2$) of Gyeongancheon water level gauging station with the coefficient of determination and Nash-Sutcliffe efficiency ranging from 0.77 to 0.60 and 0.79 to 0.60, respectively. Two GCMs (MIROC3.2hires, ECHAM5-OM) future weather data of high (A2), middle (A1B) and low (B1) emission scenarios of the IPCC (Intergovernmental Panel on Climate Change) were adopted and the data was corrected by 20C3M (20th Century Climate Coupled Model) and downscaled by Change Factor (CF) method using 30 years (1977 - 2006, baseline period) weather data. Three periods data of 2010 - 2039 (2020s), 2040 - 2069 (2050s), 2070 - 2099 (2080s) were prepared. To reduce the uncertainty of land surface conditions, future land use and vegetation canopy prediction were tried by CA-Markov technique and NOAA NDVI-Temperature relationship respectively. MIROC3.2 hires and ECHAM5-OM showed increase tendency in annual streamflow up to 21.4 % for 2080 A1B and 8.9 % for 2050 A1B scenario respectively. The portion of future predicted ET about precipitation increased up to 3 % in MIROC3.2 hires and 16 % in ECHAM5-OM respectively. The future soil moisture content slightly increased compared to 2002 soil moisture.