• 검색 결과가 없습니다.

IEG 환경지질연구정보센터

N/A
N/A
Protected

Academic year: 2021

Share "IEG 환경지질연구정보센터"

Copied!
1
0
0

로드 중.... (전체 텍스트 보기)

전체 글

(1)

수리지질학․10월 29일(토)

2005 대한지질학회 추계학술발표회 초록집

143

Forecasting Hydrogeologic Time Series Data Using Artificial Neural Networks

Yoon, Heesung*․Jun, Seong-Chun․ Bae, Kwang-Ok․ Lee, Kang-Kun Hydrogeology Laboratory, Seoul National University, [email protected]

An effective management of the groundwater requires the forecasting temporal varia- tions of hydrogeologic variables, such as the level of groundwater table, the concentration of contaminants in groundwater and so on. These hydrogeologic time series data tend to have a nonlinear relationship between input and output time series because the subsur- face medium is highly heterogeneous. The conventional time series forecasting models, however, are based on a linear relationship between inputs and outputs. Thus they have difficulties in forecasting the hydrogeologic time series data whose relationship between inputs and outputs is nonlinear.

Recently, the application of artificial neural networks (ANN) as an approach to fore- casting water resource variables is growing. The ANN is a flexible mathematical structure patterned after a biological nerval system and is considered the standard computational tool for nonlinear problems in a variety of fields.

In this study, the applicability of an ANN time series model to forecasting the level of groundwater table is investigated. For the model development feed forward network and recurrent neural network are used. The input time series are the amount of the precip- itation or the precipitation and the level of the tide in the case of the coastal site. The opti- mal model structure and parameters are examined and the effect of lag times between in- puts and outputs on model performance is also investigated. The forecasting results show that the ANN time series model can be an effective method to forecast the hydrogeologic time series data.

9-35 포스터

참조

관련 문서

Double pie chart showing the taxonomic composition at phylum and species level for groundwater (IYAW1), hyporheic water (IYHW1, IYHW2, IYHW3, and IYHW4) and stream water (IYSW1)

It can be summarized that the sea level rise derived from global warming and climate change is a big threat to groundwater resources in coastal areas in the world.. Thus,

This study was performed to evaluate the applicability of pump and treat technology as well as to identify the changes of groundwater level by continuous pumping at the

Fig. Variations in safety factors of a slope when considering both the downward velocity of the wetting front and the upward velocity of the groundwater level... 13은 Bishop의

In this study, kriging and conditional simulation were applied for the supplement of the missing groundwater-level data in alluvium and bedrock aquifer with peak type variations

A time-series analysis was performed on the daily groundwater level data from 34 monitor- ing wells in bedrock aquifer of Seoul (a capital of South Korea) for three years, to

요약 (Abstract): In this study, a semi-analytical model to address groundwater level fluctuations in response to precipitations and its infiltration is developed

Optimal method of radon analysis in groundwater was studied using ultra low-level liquid scintillation counter (ULL- LSC) which is well known as an analytical instrument for