• Proceedings of the Korea Water Resources Association Conference
• /
• 2007.05a
• /
• pp.142-146
• /
• 2007

본 연구에서는 강우강도식의 매개변수를 보다 효율적으로 산정하기 위해서 유전자알고리즘을 적용한 매개변수 산정법을 제시하였으며, 지속기간의 장, 단기간에 따른 매개변수의 변화를 고려하기 위하여 다목적 유전자알고리즘을 적용하여 매개변수를 추정한 후 기존의 강우강도식에 의한 결과와 비교해 보았다. 매개변수는 지점빈도해석을 통해 산정된 확률강우량을 사용하여 추정하였고, 유전자 알고리즘의 목적함수로는 Nash & Sutcliffe Index, Root Mean Square Error(RMSE), Relative Root Mean Square Error(RRMSE), 결정계수, 평균들을 사용하여 가장 효율적인 형태의 목적함수를 구성하였다. 그 결과 기존의 매개변수 추정 방법들에 비해 유전자알고리즘을 이용한 경우에 더 정확한 강우량값을 산정할 수 있었고, 특히 다목적 유전자 알고리즘을 사용할 경우 장기간과 단기간에 걸쳐서 동시에 정확도를 향상시킬 수 있는 매개변수를 추정할 수 있는 것으로 나타났다.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.56 no.4
• /
• pp.9-19
• /
• 2014

The objective of this study was to evaluate of the Tank model in simulating runoff discharge from rural watershed in comparison to the SWAT (Soil and Water Assessment Tool) model. The model parameters of SWAT was calibrated by the shuffled complex evolution-university Arizona (SCE-UA) method while Tank model was calibrated by genetic algorithm (GA) and validated. Four dam watersheds were selected as the study areas. Hydrological data of the Water Management Information System (WAMIS) and geological data were used as an input data for the model simulation. Runoff data were used for the model calibration and validation. The determination coefficient ($R^2$), root mean square error (RMSE), Nash-Sutcliffe efficiency index (NSE) were used to evaluate the model performances. The result indicated that both SWAT model and Tank model simulated runoff reasonably during calibration and validation period. For annual runoff, the Tank model tended to overestimate, especially for small runoff (< 0.2 mm) whereas SWAT model underestimate runoff as compared to observed data. The statistics indicated that the Tank model simulated runoff more accurately than the SWAT model. Therefore the Tank model could be a good tool for runoff simulation considering its ease of use.

• Journal of Korea Water Resources Association
• /
• v.53 no.6
• /
• pp.395-408
• /
• 2020

This paper aims to develop and apply three different machine learning approaches (i.e., artificial neural networks (ANN), adaptive neuro-fuzzy inference systems (ANFIS), and wavelet-based neural networks (WNN)) combined with an evolutionary optimization algorithm and the k-fold cross validation for multi-step (days) streamflow forecasting at the catchment located in Algeria, North Africa. The ANN and ANFIS models yielded similar performances, based on four different statistical indices (i.e., root mean squared error (RMSE), Nash-Sutcliffe efficiency (NSE), correlation coefficient (R), and peak flow criteria (PFC)) for training and testing phases. The values of RMSE and PFC for the WNN model (e.g., RMSE = 8.590 ㎥/sec, PFC = 0.252 for (t+1) day, testing phase) were lower than those of ANN (e.g., RMSE = 19.120 ㎥/sec, PFC = 0.446 for (t+1) day, testing phase) and ANFIS (e.g., RMSE = 18.520 ㎥/sec, PFC = 0.444 for (t+1) day, testing phase) models, while the values of NSE and R for WNN model were higher than those of ANNs and ANFIS models. Therefore, the new approach can be a robust tool for multi-step (days) streamflow forecasting in the Seybous River, Algeria.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.57 no.4
• /
• pp.121-133
• /
• 2015

The TANK model has been widely used in rainfall-runoff modeling due to its simplicity of concept and computation while achieving forecast accuracy. A major barrier to the model application is to determine parameter values for ungauged watersheds, leading to the need of a method for the parameter estimation. The objective of this study was to develop regression equations for estimating the 3th TANK model parameters considering their variations for the ungauged watersheds. Thirty watersheds of dam sites and stream stations were selected for this study. A genetic algorithm was used to optimize TANK model parameters. Watershed characteristics used in this study include land use percent, watershed area, watershed length, and watershed average slope. Generalized equations were derived by correlating to the optimized parameters and the watershed characteristics. The results showed that the TANK model, with the parameters determined by the developed regression equations, performed reasonably with 0.60 to 0.85 of Nash-Sutcliffe efficiency for daily runoff. The developed regression equations for the TANK model can be applied for the runoff simulation particularly for the ungauged watersheds, which is common for upstream of agricultural reservoirs in Korea.

• Journal of Korean Society on Water Environment
• /
• v.29 no.2
• /
• pp.184-195
• /
• 2013

Urbanization from agricultural/forest areas has been causing increased runoff and pollutant loads from it. Thus, numerous models have been developed to estimate NPS loading from urban area and Long-Term Hydrologic Impact Analysis (L-THIA) model has been used to evaluate effects of landuse changes on runoff and pollutant loads. However, the L-THIA model could not consider rainfall intensity in runoff evaluation. Therefore, the L-THIA model, capable of simulating runoff using 10-minute rainfall data, was applied to the study areas for evaluation of estimated runoff and NPS. The estimated Nash-Sutcliffe coefficient (NSE) values were over 0.6 for runoff, BOD, TN, and TP for most sites and watershed. The calibrated model was further extended to other counties for pollutant load potential evaluation. Pollutant load potential maps were developed and target areas were identified. As shown in this study, the L-THIA 2012 can be used for evaluation runoff and pollutant loads with limited data sets and its estimation could be used in identifying pollutant load hot spot areas for implementation of site-specific Best Management Practices.