• 한국농공학회지
• /
• 제35권3호
• /
• pp.23-35
• /
• 1993

The objectives of this paper were to validate the proposed network flow model using field data and to demonstrate the model applicability for various purposes. The model was tested with data from the Banweol watershed, where an intentive streamflow gauging system has been established. Model parameters were not calibrated with field data so that it can be validated as ungaged conditions. Three different schemes were employed to represent the drainage system of the tested watershed : a single, complex, and detailed network. The single network assumed the watershed as a cell, while complex and detailed networks considered several cells. The results from different schemes were individually compared satisfactorily to the observed daily stages at the Banweol reservoir located at the outlet of the watershed. The results from three schemes were in close agreement with each other, Justifying that the model performs very well for different network schemes being used. Daily streamflow from three network schemes was compared for a selected reach within the watershed. The results were very close to each other regardless of network formulation. And the model was applied to simulate daily streamflow before and after the construction of a reservoir at a reach. The differences were discussed, which reflected the influences of the dam construction upon the downstream hydrology. Similar appliocations may be possible to identify the effects of hydraulic structures on streamflow.

• 한국환경과학회지
• /
• 제13권7호
• /
• pp.611-618
• /
• 2004

The Neural Network Models which mathematically interpret human thought processes were applied to resolve the uncertainty of model parameters and to increase the model's output for the streamflow forecast model. In order to test and verify the flood discharge forecast model eight flood events observed at Kumho station located on the midstream of Kumho river were chosen. Six events of them were used as test data and two events for verification. In order to make an analysis the Levengerg-Marquart method was used to estimate the best parameter for the Neural Network model. The structure of the model was composed of five types of models by varying the number of hidden layers and the number of nodes of hidden layers. Moreover, a logarithmic-sigmoid varying function was used in first and second hidden layers, and a linear function was used for the output. As a result of applying Neural Networks models for the five models, the N10-6model was considered suitable when there is one hidden layer, and the Nl0-9-5model when there are two hidden layers. In addition, when all the Neural Network models were reviewed, the Nl0-9-5model, which has two hidden layers, gave the most preferable results in an actual hydro-event.

• 물과 미래
• /
• 제23권1호
• /
• pp.109-118
• /
• 1990

본 연구에서는 한강 및 낙동강유역의 유량관측망에 대하여 개발된 지역화회귀모형을 이용하여 모형이 지니는 평균표본오차의 최소화를 목적함수로 하는 비선형 정수계획법에 의하여 5,6,10,15 및 20년의 계획년수를 갖는 각 운영계획에 따른 기존 관측망의 유효성을 평가하며, 경제적인 측면에서 관측망효율을 증가시킬 수 있도록 하천유량관측망을 조정·계획하였다.

• 한국습지학회지
• /
• 제21권3호
• /
• pp.236-243
• /
• 2019

현재까지 많은 수문 시계열은 전통적인 선형 모형을 이용하여 분석되고 예측되어 왔다. 하지만, 자연현상과 수문시계열의 패턴 및 변동과 관련하여 비선형적 구조의 증거가 발견되고 있다. 따라서 시계열 분석 및 예측을 위한 기존의 선형 모형은 비선형적 특성에 적합하지 않을 수 있다. 본 연구에서는 미국 플로리다 코코아 지역 인근에 있는 St.Johns 강의 일유량 자료에 대한 카오스 분석을 수행하였고, 그 결과 낮은 차원의 비선형 동역학적 특성을 가진 흥미로운 결과가 나타났지만 한국의 소양강댐 일유량 자료는 확률적 특성을 보여주었다. 카오스 특성을 토대로한 DVS(결정론적 vs 추계학적) 알고리즘을 이용해 두 시계열 시스템의 특성을 파악하였고 단기 예측을 수행하였다. 또한 본 연구에서는 일 유량 시계열 예측을 위해 인공신경망 방법을 사용하였고, DVS 알고리즘에 의한 예측을 비교 분석하였다. 분석 결과, 카오스 특성을 갖는 시계열 자료가 보다 정확한 예측성을 보였다.

• 한국농공학회지
• /
• 제43권2호
• /
• pp.69-77
• /
• 2001

An artificial neural network model was developed to analyze and forecast daily steamflow flow a small watershed. Error Back propagation neural networks (EBPN) of daily rainfall and runoff data were found to have a high performance in simulating stremflow. The model adopts a gradient descent method where the momentum and adaptive learning rate concepts were employed to minimize local minima value problems and speed up the convergence of EBP method. The number of hidden nodes was optimized using Bayesian information criterion. The resulting optimal EBPN model for forecasting daily streamflow consists of three rainfall and four runoff data (Model34), and the best number of the hidden nodes were found to be 13. The proposed model simulates the daily streamflow satisfactorily by comparison compared to the observed data at the HS#3 watershed of the Baran watershed project, which is 391.8 ha and has relatively steep topography and complex land use.

• 한국지리정보학회지
• /
• 제21권4호
• /
• pp.50-63
• /
• 2018

급격한 도시화를 겪으면서 자연적인 물순환 체계의 왜곡을 초래하였다. 이러한 물순환 구조의 변화는 기존 수자원 이용 경향을 변화시키며 하천 건천화 현상을 유발하고 있다. 이를 관리하기 위해 건천화 평가 및 예측이 가능한 하천 건천화 영향 평가 기술이 필요하다. 하천 건천화 영향평가 기술 수행을 위해서는 기초자료로써 GIS 기반의 공간자료 구축이 필수적이나, 관련 연구는 미흡한 실정이다. 따라서 본 연구에서는 하천 건천화 평가를 위한 GIS 기반의 시계열 공간자료 활용에 대한 연구를 수행하였다. 이에 6개 하천 건천화 영향요소(기상, 토심, 산림밀도 및 높이, 도로망, 지하수 이용량, 토지이용)을 대상으로, 과거 수십년 간의 변화과정을 전국 단위 GIS 자료로 구축하여 연속수문모형 운용에 대한 기초자료로 활용하였다. 이러한 영향요소를 대상으로 시계열에 따라 하천 건천화 원인을 분석하고 해석할 수 있는 분포형 연속수문모형 기반의 DrySAT을 활용하여 하천 건천화 영향요소별 연유출량 및 건천화 평가를 수행하였다. 그 결과, 다른 요소들은 고려하지 않고 주어진 기상 조건하에 연유출량은 기본값 977.9mm로 산출되었다. 반면, 토심 감소, 산림 높이 증가, 도로 개발 증가, 지하수이용량 증가, 토지이용 개발변화를 고려하였을 때의 연평균 유출량은 각각 1,003.5mm, 942.1mm, 961.9mm, 915.5mm, 1003.7mm로 산출되었다. 산출된 결과는 하천건천화의 주요 원인으로서 지표유출량을 증가시켜 하천유량을 감소시키는 토심의 감소, 지표유출량을 감소시키는 산림 밀도의 증가, 지표하유출량을 감소시키는 도로의 증가, 기저유출량을 감소시키는 무분별한 지하수 개발과 지하수이용량의 증가, 지표유출량을 증가시키는 불투수지역의 증가를 들 수 있다. 또한, 하천 건천화 정의 및 등급 범위를 통해서 건천화 등급에 따라 표준유역별로 나타내었으며, 기상, 토심 감소 고려, 산림 높이 증가, 도로 개발 증가, 지하수이용량 증가, 토지이용 개발변화를 고려하였을 때의 건천화 등급은 각각 2.1, 2.2, 2.5, 2.3, 2.8, 2.2로 나타났다. 기본값인 강우조건을 제외한 5개 하천 건천화 영향요소에 대한 건천화 영향순위는 지하수 이용량 변화에 대한 건천화 영향이 제일 컸으며, 산림 밀도 변화, 도로 건설 변화, 토지이용 변화 및 토심 변화 순으로 나타났다. 향후 전국 하천 건천화 평가시스템 개발을 통해 6개 하천 건천화 영향요소에 대한 미래 자료 변화 및 이에 대한 건천화의 진행전망 등 시스템에 의한 평가결과를 토대로 맞춤형 하천 건천 관리 및 방지 방안을 제공할 수 있을 것으로 판단된다.

• 한국농공학회지
• /
• 제35권1호
• /
• pp.40-49
• /
• 1993

The Objectives of this paper were to develop a modified tank model that is capable of simulating daily streamflow from a small watershed using daily watershed evapotranspiration and to test the applicability of the model to different watersheds. Tank model was restructured to consist of three series of tanks, each of which may mathematically reflect watershed runoff mechanisms from different components of surface runoff, interflow, and baseflow. And pan evaporation was correlated to potential evapotranspiration estimated from a combination method, and was multiplied by monthly crop and landuse coefficients, and watershed storage coefficient to estimate the watershed evapotranspiration losses. Ten watersheds were selected to calibrate model parameters that were defined using an optimization scheme, and the results were correlated with watershed parameters. Simulated daily runoff was compared to the observed ones from the tested watersheds. The simulating results were in good agreement with the observed values when optimal and calibrated parameters were used. Ungaged conditions were also applied to compare simulated values to the observed. And the results were in fair conditions for all the tested watersheds which differ considerably in their sizes, landuse types, and physiological features.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2021년도 학술발표회
• /
• pp.131-131
• /
• 2021

Deep learning models, especially those based on long short-term memory (LSTM), have presented their superiority in addressing time series data issues recently. This study aims to comprehensively evaluate the performance of deep learning models that belong to the supervised learning category in streamflow prediction. Therefore, six deep learning models-standard LSTM, standard gated recurrent unit (GRU), stacked LSTM, bidirectional LSTM (BiLSTM), feed-forward neural network (FFNN), and convolutional neural network (CNN) models-were of interest in this study. The Red River system, one of the largest river basins in Vietnam, was adopted as a case study. In addition, deep learning models were designed to forecast flowrate for one- and two-day ahead at Son Tay hydrological station on the Red River using a series of observed flowrate data at seven hydrological stations on three major river branches of the Red River system-Thao River, Da River, and Lo River-as the input data for training, validation, and testing. The comparison results have indicated that the four LSTM-based models exhibit significantly better performance and maintain stability than the FFNN and CNN models. Moreover, LSTM-based models may reach impressive predictions even in the presence of upstream reservoirs and dams. In the case of the stacked LSTM and BiLSTM models, the complexity of these models is not accompanied by performance improvement because their respective performance is not higher than the two standard models (LSTM and GRU). As a result, we realized that in the context of hydrological forecasting problems, simple architectural models such as LSTM and GRU (with one hidden layer) are sufficient to produce highly reliable forecasts while minimizing computation time because of the sequential data nature.

• 한국물환경학회지
• /
• 제26권5호
• /
• pp.860-870
• /
• 2010

Long term monthly monitoring data showed that the water quality of streams flowing into Lake Paldang has been improved by various strategy for water. However, the effect of quality on Lake Paldang is still insufficient because of nonpoint source from watershed. In order to evaluate quantifying methods for pollution source and make a suggestion on improvements, Storm Water Management Model (SWMM) was constructed by using data set from the water quality and streamflow monitoring network in the Kyoungan watershed for Total Maximum Daily Loads (TMDLs). Load duration curve (LDC) based on the result of the Kyoungan watershed SWMM indicated that the water quality criterion on $BOD_5$ was often exceeded in up-stream than down-stream. From flowrate-load correlation curve, SS load significantly increased as streamflow increases. 75.3% of streamflow and 62.1% of $BOD_5$ loads is discharged especially in the zone of high flows, but monitoring data set didn't provide proper information about the conditions and the patterns associated with storm events. Therefore, it is necessary to acquire representative data set for comparing hydrograph and pollutograph through monitoring experimental watershed and to establish methods for quantifying point and nonpoint source pollutant loads.

• Asian Journal of Innovation and Policy
• /
• 제10권2호
• /
• pp.212-235
• /
• 2021

In spite of the increasing popularity of the Ricardian model for the study of the impact of climate change on agriculture, there has been few attempts to examine the role of interregional spillovers in this framework and all of them rely on geographical proximity-based weighting schemes. We remedy to this gap by focusing on the spatial externalities of surface water flow used for irrigation purposes and demonstrate that farmland value, the usual dependent variable used in the Ricardian framework, is a function of the climate variables experienced locally and in the upstream locations. This novel approach is tested empirically on a spatial panel model estimated across the counties of the Southwest USA over 1997-2012. This region is one of the driest in the country, hence its agriculture relies heavily on irrigated surface water. The results highlight how the weather conditions in upstream counties significantly affect downstream agriculture, thus the actual impact of climate change on agriculture and subsequent adaptation policies cannot overlook the streamflow network anymore.