• Proceedings of the Korea Water Resources Association Conference
• /
• 2002.05a
• /
• pp.43-50
• /
• 2002

Accurate quantitative forecasting of rainfall for basins with a short response time is essential to predict streamflow and flash floods. Previously, neural networks were used to develop a Quantitative Precipitation Forecasting (QPF) model that highly improved forecasting skill at specific locations in Pennsylvania, using both Numerical Weather Prediction (NWP) output and rainfall and radiosonde data. The objective of this study was to improve an existing artificial neural network model and incorporate the evolving structure and frequency of intense weather systems in the mid-Atlantic region of the United States for improved flood forecasting. Besides using radiosonde and rainfall data, the model also used the satellite-derived characteristics of storm systems such as tropical cyclones, mesoscale convective complex systems and convective cloud clusters as input. The convective classification and tracking system (CCATS) was used to identify and quantify storm properties such as life time, area, eccentricity, and track. As in standard expert prediction systems, the fundamental structure of the neural network model was learned from the hydroclimatology of the relationships between weather system, rainfall production and streamflow response in the study area. The new Quantitative Flood Forecasting (QFF) model was applied to predict streamflow peaks with lead-times of 18 and 24 hours over a five year period in 4 watersheds on the leeward side of the Appalachian mountains in the mid-Atlantic region. Threat scores consistently above .6 and close to 0.8 ∼ 0.9 were obtained fur 18 hour lead-time forecasts, and skill scores of at least 4％ and up to 6％ were attained for the 24 hour lead-time forecasts. This work demonstrates that multisensor data cast into an expert information system such as neural networks, if built upon scientific understanding of regional hydrometeorology, can lead to significant gains in the forecast skill of extreme rainfall and associated floods. In particular, this study validates our hypothesis that accurate and extended flood forecast lead-times can be attained by taking into consideration the synoptic evolution of atmospheric conditions extracted from the analysis of large-area remotely sensed imagery While physically-based numerical weather prediction and river routing models cannot accurately depict complex natural non-linear processes, and thus have difficulty in simulating extreme events such as heavy rainfall and floods, data-driven approaches should be viewed as a strong alternative in operational hydrology. This is especially more pertinent at a time when the diversity of sensors in satellites and ground-based operational weather monitoring systems provide large volumes of data on a real-time basis.

• Journal of Korea Water Resources Association
• /
• v.54 no.12
• /
• pp.1255-1263
• /
• 2021

While it could become an alternative water resource, fog could undermine traffic safety and operational performance of infrastructures. To reduce such adverse impacts, it is necessary to have spatially continuous fog risk information. In this work, tree-based machine-learning models were developed in order to quantify fog risks with routine meteorological observations alone. The Extreme Gradient Boosting (XGB), Light Gradient Boosting (LGB), and Random Forests (RF) were chosen for the regional fog models using operational weather and visibility observations within the Jeollabuk-do province. Results showed that RF seemed to show the most robust performance to categorize between fog and non-fog situations during the training and evaluation period of 2017-2019. While the LGB performed better than in predicting fog occurrences than the others, its false alarm ratio was the highest (0.695) among the three models. The predictability of the three models considerably declined when applying them for an independent period of 2020, potentially due to the distinctively enhanced air quality in the year under the global lockdown. Nonetheless, even in 2020, the three models were all able to produce fog risk information consistent with the spatial variation of observed fog occurrences. This work suggests that the tree-based machine learning models could be used as tools to find locations with relatively high fog risks.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.17 no.4
• /
• pp.317-325
• /
• 2015

An adequate downscaling of synoptic forecasts is a prerequisite for improved agrometeorological service to rural areas in South Korea where complex terrains and small farms are common. In this study, geospatial schemes based on topoclimatology were used to scale down the Korea Meteorological Administration (KMA) temperature forecasts to the local scale (~30 m) across a rural catchment. Then, using these schemes, local temperatures were estimated at 14 validation sites at 0600 and 1500 LST in 2013/2014 and were compared with the observations. The estimation errors were substantially reduced for both 0600 and 1500 LST temperatures when compared against the uncorrected KMA products. The improvement was most notable at low lying locations for the 0600 temperature and at the locations on west- and south-facing slopes for the 1500 LST temperature. Using the downscaled real-time temperature data, a pilot service has started to provide the field-specific weather information tailored to meet the requirements of small-scale farms. For example, the service system makes a daily outlook on the phenology of crop species grown in a given field using the field-specific temperature data. When the temperature forecast is given for next morning, a frost risk index is calculated according to a known relationship of phenology and frost injury. If the calculated index is higher than a pre-defined threshold, a warning is issued and delivered to the grower's cellular phone with relevant countermeasures to help protect crops against frost damage.

• Journal of Korea Water Resources Association
• /
• v.46 no.2
• /
• pp.195-204
• /
• 2013

As drought is phenomenon of nature with unavoidability and repeated characteristic, it is necessary to plan to respond to it in advance and construct drought management system to minimize its damage. This study suggested standard for classification of drought, which is appropriate for our nation to respond to drought by assessing drought severity in the regions for this study. For data collection, 61 locations were selected - the locations keep precipitation data over 30 years of observation. And data for monthly precipitation for 37 years from 1973 were used. Based on this, this study classified unified drought interval into four levels using drought situation phases which are used in government. For standard for classification of drought severity fit to our nation, status of main drought was referred and these are classified based on accumulated probability of drought - 98~100% Exceptional Drought, 94~98% Extreme Drought, 90~94% Severe Drought, 86~90% Moderate Drought. Drought index (SPI, PDSI) was made in descending order and quantitative value of drought index fit to standard of classification for drought severity was calculated. To compare classification results of drought severity of SPI and PDSI with actual drought, comparison by year and month unit were analyzed. As a result, in comparison by year and comparison by month unit of SPI, drought index of each location was mostly identical each other between actual records and analyzed value. But in comparison by month unit of PDSI for same period, actual records did not correspond to analyzed values. This means that further study about mutual supplement for these indexes is necessary.

• Atmosphere
• /
• v.23 no.2
• /
• pp.131-141
• /
• 2013

Many studies have related the recent variations of wildfire regime such as the increasing number of occurrances, their patterns and timing changes, and the severity of their extreme cases with global warming. However, there are only a few numbers of wildfire studies to assess how the future wildfire regime will change in the interactions between land and atmosphere with climate change especially over East Asia. This study was performed to estimate the future changing aspect of wildfire danger with global warming, using Haines Index (HI). Calculated from atmospheric instability and dryness, HI is the potential of an existing fire to become a dangerous wildfire. Using the Weather Research and Forecasting (WRF) model, two separated 5-year simulations of current (1995~1999) and far future (2095~2099) were performed and analyzed. Community Climate System Model 3 (CCSM3) model outputs were utilized for the model inputs for the past and future over East Asia; future prediction was driven under the IPCC A1B scenario. The results indicate changes of the wildfire danger regime, showing overall decreasing the wildfire danger in the future but intensified regional deviations between north and south. The overall changes of the wildfire regime seems to stem from atmospheric dryness which is sensitive to soil moisture variation. In some locations, the future wildfire danger overall decreases in summer but increases in winter or fall when the actual fire occurrence are generally peaked especially in South China.

• Journal of the Society of Disaster Information
• /
• v.14 no.2
• /
• pp.154-164
• /
• 2018

Purpose: Recent climate change makes weather-related disasters such as summer storms, heavy rains, winter snowfall disasters, and extreme cold temperature increase in trend. Heavy snowfall disasters requires speedy response due to various effects to traffic flows, buildings, and infrastructure. Heavy snowfall disaster response of South Korea is insufficient, even though heavy snowfall disasters affect urban safety. There have been lack of policy studies for heavy snowfall disasters. Method: This research analyzes case studies and explores implications using Information system technologies to snow removal vehicles and equipments for speedy snow removal during the heavy snowfall disasters. Results: Information system technology attachment to snow removal equipments can identify locations of snow removal vehicles and equipments for emergency period to support snow removal of adjacent jurisdictions. Conclusion: Case studies of this research can be further used for efficient application of snow removal tools of local governments.

• Journal of Korean Society for Atmospheric Environment
• /
• v.25 no.3
• /
• pp.219-236
• /
• 2009

Particle-phase organic tracers (molecular markers) have been shown to be an effective method to assess and quantify the impact of sources of carbonaceous aerosols. These molecular markers have been used in chemical mass balance (CMB) models to apportion primary sources of organic aerosols in regions where the major organic aerosol source categories have been identified. As in the case of all CMB models, all important sources of the tracer compounds must be included in a Molecular Marker CMB (MM-CMB) model or the MMCMB model can be subject to biases. To this end, the application of the MM-CMB models to locations where reasonably accurate emissions inventory of organic aerosols are not available, should be performed with extreme caution. Of great concern is the potential presence of industrial point sources that emit carbonaceous aerosols and have not been well characterized or inventoried. The current study demonstrates that emissions from industrial point sources in the St. Louis, Missouri area can greatly bias molecular marker CMB models if their emissions are not correctly addressed. At a sampling site in the greater St. Louis Area, carbonaceous aerosols from industrial point sources were found to be important source of carbonaceous aerosols during specific time periods in addition to common urban sources (i.e. mobile sources, wood burning, and road dust). Since source profiles for these industrial sources have not been properly characterized, method to identify time periods when point sources are impacting a sampling site, needs to avoid obtaining biases source apportionment results. The use of real time air pollution measurements, along with molecular marker measurements, as a screening tool to identify when point sources are impacting a receptor site is presented.

• Journal of the Society of Naval Architects of Korea
• /
• v.41 no.6
• /
• pp.75-83
• /
• 2004

Ice rubble pieces broken by the bow impact load and side hull of an icebreaking vessel usually pass along the ship's bottom hull and may hit the propeller/rudder or other stern structures causing serious damage to ship's hull . Therefore it is important to estimate the size of broken ice pieces during the icebreaking process. The dynamic interaction process of icebreaker with infinite ice sheet is simplified as a wedge type beam of finite length supported by elastic foundation. The wedge type ice beam is leaded with vertical impact forces due to the inclined bow stem of icebreaking vessels. The numerical model provides locations of maximum dynamic bending moment where extreme tensile stress arises and also possible fracture occurs. The model can predict a failure length of broken ice sheet given design parameters. The results are compared to Nevel(1961)'s analytical solution for static load and observed pattern of ice sheet failure onboard an icebreaker. Also by comparing computed failure length with the characteristic length, the meaning of ice rubble sizes is discussed.

• Journal of Environmental Science International
• /
• v.16 no.12
• /
• pp.1451-1462
• /
• 2007

Heavy metal concentrations in the soil were investigated for the abandoned Samkwang metal mine, Cheongyang-Gun, Chungnam Province, Korea. The concentrations of heavy metal(As, Cd, Cu, Ni, Pb, Zn) were determined in mine soils collected at the abandoned mine sites to obtain a general classification and specification of the pollution in this highly polluted region. The results estimated with the normal test and basis statistic on the central tendency and variation showed that the distribution of heavy metal concentration had significantly different at the range of all locations. The range of spatial distribution on the relationship of heavy metal concentration and pH was $4.8{\sim}8.8$ and heavy metal concentration on the type of land use was highest in forest land, and also Ni and Zn in farm and rice field showed the high concentration. The distribution of heavy metal concentration on the depth of a soil showed that the metal concentrations in subsoil were higher than of those in surface soil, while the concentration of Cu and Ni had no significant difference on the depth of soil. Results from the correlation analysis using the data except the extreme and unusual data revel that Zn-Cd(r=0.867), Zn-As(r=0.797), Zn-Pb(r=0.764), Cu-Cd(r=0.673), Cu-As(r=0.614) and Zn-Ni(r=0.605) were the most important parameters in assessing variations of heavy metal in soil. To discriminate pattern differences and similarities among samples, principal factor analysis(PFA) and cluster analysis(CF) were performed using a correlation matrix. This study suggests that PFA and CF techniques are useful tools for identification of important heavy metal and parameters. This study presents the necessity and usefulness of multivariate statistical assessment of complex databases in order to get better information about the quality of soil and gives the basis information to clean up the abandoned mine sites.

• Journal of The Korean Astronomical Society
• /
• v.52 no.6
• /
• pp.235-244
• /
• 2019

We present the results of near-infrared imaging observations of the galaxy overdensity around the z = 1.44 radio-loud active galactic nucleus (AGN) 6CE1100+3505, which was carried out with the purpose of sampling the redshifted Hα emission from the actively star-forming galaxies that could constitute the overdensity. The existence of the structure around this AGN was spectroscopically confirmed by previous grism observations which are however limited to the central region. Using the CH4Off narrow/medium-band and H broad band filters in the Wide Infrared Camera (WIRCam) on the Canada-France-Hawaii Telescope (CFHT), we constructed a sample of objects that show a flux excess in the CH4Off band due to line emission. The emission line flux is ~ 4.9 × 10^-16 erg s^-1 cm^-2, corresponding to a star formation rate (SFR) of ~ 50 M_⊙ yr^-1 for galaxies at redshifts z ~ 1.4. None of the galaxies with medium-band flux excess is located within 1 Mpc from the central AGN, and there is no evidence that the selected galaxies are associated with the proposed cluster. Along with the star formation quenching near the center that was found from the previous grism observations, the lack of extreme starbursts in the structure suggests that at z ~ 1.4, overdense regions are no longer favorable locations for vigorous star formation.