• The Journal of Engineering Geology
• /
• v.23 no.2
• /
• pp.137-147
• /
• 2013

For the effective management of groundwater resources, it is necessary to predict groundwater level fluctuations in response to rainfall events. In the present study, time series models using artificial neural networks (ANNs) and support vector machines (SVMs) have been developed and applied to groundwater level data from the Gasan, Shingwang, and Cheongseong stations of the National Groundwater Monitoring Network. We designed four types of model according to input structure and compared their performances. The results show that the rainfall input model is not effective, especially for the prediction of groundwater recession behavior; however, the rainfall-groundwater input model is effective for the entire prediction stage, yielding a high model accuracy. Recursive prediction models were also effective, yielding correlation coefficients of 0.75-0.95 with observed values. The prediction errors were highest for Shingwang station, where the cross-correlation coefficient is lowest among the stations. Overall, the model performance of SVM models was slightly higher than that of ANN models for all cases. Assessment of the model parameter uncertainty of the recursive prediction models, using the ratio of errors in the validation stage to that in the calibration stage, showed that the range of the ratio is much narrower for the SVM models than for the ANN models, which implies that the SVM models are more stable and effective for the present case studies.

• Proceedings of the Korea Society of Environmental Toocicology Conference
• /
• 2003.10a
• /
• pp.27-27
• /
• 2003

Lake Texoma is located on the border of southern Oklahoma and northern Texas. It has 93,000 surface acres, and is a focus of the recreation, and farming industries in the region. There are potential stressors around the Lake Texoma watershed that may cause adverse ecological effects in the lake. System assimilative capacity (SAC) is the ability of abiotic and biotic processes to atteuniate the stressors. SAC Exceeded indicates potential of occuring adverse eco-effects. A number of representative chemical release sites and stressor sources in the surrounding watershed were characterized, and several impact sites having stressors sources, such as being near agriculture, landfills, housing areas, oil production fields and heavy use recreational activity, were selected for surface water, sediment, and groundwater monitoring. A paired reference site, having similar physical characteristics as its impact site, was also chosen based on its proximity to the impact site. Lake water samples were collected at locations identified as marina entrance, gasoline filling station, and boat dock at five marinas selected on Lake Texoma from September 1999 to December 2001. Paired water and sediment samples were also collected. Groundwater samples were collected at about 70 producing monitoring wells. Water quality parameters measured were inorganics (nitrate, nitrite, orthophosphate, ammonia, sulfate, and chloride), dissolved methane, total organic carbon (TOC) (or DOC), volatile organic compounds (VOCs) such as methyl tert-butyl ether (MTBE) and BTEX, and a suite of metals. Biotic communities were evaluated at impact and reference sites. Five basic components were measured; two terrestirial components (plants and bird comminitires) and three aquatic components (benthic inverbrates, litteral-zone fishes, ecosystem attribures). Potential impacts to these comminites were evaluated.

• Economic and Environmental Geology
• /
• v.49 no.5
• /
• pp.361-370
• /
• 2016

It is well known that the inflow of rainwater and the infiltration of groundwater to sewerage (I/I) increase the sewage and burden sewage treatment plants and lower their treatment efficiency. Therefore, it is important to estimate the amount of I/I. In this study, well groundwaters, public water supplies (PWSs), and sewage and rainwater channels were investigated to check whether oxygen and deuterium isotopes could be used as a tool for I/I estimation. This study shows that the isotopic composition of PWS in Jeonju area is very consistent over time and distinctly lighter than the circulating local rainwater (CLR) because it is supplied from Yongdam Dam, which is located about 40 km inland to the east in the mountainous area. Considering the fact that sewage mostly originates from the PWS, we could calculate the amounts of CLR in the sewerage from a monitoring station using unaffected rainwater and tap water as mixing end members. The calculation revealed that the CLR fraction ranged from 50% to 90% depending on observation time. This is well supported by the dilute natures of the sewages at the station. The fraction of PWS in investigated well waters were about 46%, indicating that leaking of PWS is very serious and is an important groundwater source in the study area. Since the infiltration of such groundwater may not alter the isotopic composition of sewage significantly, the actual I/I would be much greater than the calculated ones.

• Economic and Environmental Geology
• /
• v.41 no.6
• /
• pp.645-655
• /
• 2008

The Cr(VI) concentrations at the shallow aquifer well (MPH-0-1) of the Moonpyung groundwater monitoring station were in the range of 0.5 to 3.1 mg/L exceeding 10 to 62 times the guideline for drinking-water quality, indicating continuous contamination. However, Cr was not detected at the deep bedrock well and the other subsidiary monitoring wells except for MPH-1 and 6. Cross-correlation analyses were conducted for rainfall and groundwater level time series, resulting in the mean time of recharge after precipitation events to be 5.6 days. For rainy season, the water level was raised and the Cr(VI) concentration was several times lower than that during dry season at well MPH-0-1 well. Correlation of the Cr(VI) concentration with the groundwater-level showed that the Cr(VI) reduction was closely related with the groundwater-level rise in the well. However, the groundwater level rise during high water season induced the lateral migration of the Cr(VI)-contaminated groundwater at well MPH-4. We enriched and isolated a chromium reducing bacteria, Enterobacter aerogenes, from the Cr(VI)-contaminated groundwater in the wells MPH-0-1 and MPH-1. The bacteria may play an important role for immobilizing Cr(VI) in the Cr(VI)-contaminated groundwater. Therefore, the migration of the contaminant (Cr(VI) must has been restricted because of the natural attenuation by microbial reduction of Cr(VI) in the groundwater. This research suggests that the bioremediation of the Cr(VI)-contaminated groundwater by the indigenous bacteria may be feasible in the Cr(VI) contaminated groundwater.

• Journal of Soil and Groundwater Environment
• /
• v.8 no.4
• /
• pp.64-67
• /
• 2003

In order to identify the origin and nature of the spilled oil in the potential source, we analyzed the concentrations of specific fuel constituents in fuel standard and environmental samples. The ratios of pristane/phytane are virtually unaltered because these compounds have the same bolatility in environmental samples. These were useful to identify the source of the fuel oil and to assess the effect of microbial degradation and weathering of the fuel oil. We analyzed the ratios of pristane/phytane in neat white kerosene, boiler kerosene, JP-8 and diesel products from L and S gas station. The ratios of pristane/phytane in L-white kerosene and JP-8 was 3.10 $\pm$0.03 and 1.77 $\pm$ 0.01, respectively. Otherwise, the ratios of pristane/phytane in water phase after distribution of fuel oil and water was 2.97 $\pm$0.02 in case of white kerosene and 1.65 $\pm$ 0.02 in case of JP-8. It is apparent from the results that the ratios of pristane/phytane were as product-specific, especially between kerosene and JP-8, and therefore, can also be used for fuel type identification in free products and groundwater samples which were collected in monitoring wells.

• Journal of the Geological Society of Korea
• /
• v.54 no.5
• /
• pp.545-556
• /
• 2018

This study was conducted to estimate the deep percolation using numerical modeling and field observation data based on rainfall in Haean basin. Soil moisture sensors were installed to monitoring at 30, 60 and 90 cm depths in four sites (YHS1-4) and automatic weather station was installed to around YHS3. Soil moisture and meteorological data was observed from March 25, 2017 to March 25, 2018 and May 06, 2016 to May 06, 2018, respectively. Numerical analysis was performed from June to August, 2017 using the HYDRUS-1D. Average soil moisture contents were high to generally in YHS3 for 0.300 to $0.334m^3/m^3$ and lowest in YHS1 for 0.129 to $0.265m^3/m^3$ during the soil moisture monitoring period. The results of soil moisture flow modeling showed that field observations and modeling values were similar but the peak values were larger in the modeling result. Correlation analysis between observation and modeling data showed that r, $r^2$ and RMSE were 0.88, 0.77, and 0.0096, respectively. This show high correlation and low error rate. The total deep percolation was 744.2 mm during the period of modelling at 500 cm depth. This showed that 61.3% of the precipitation amount (1,214 mm) was recharged in 2017. Deep percolation amount was high in the study area. This study is expected to provide basic data for the estimation of groundwater recharge through unsaturated zone.