• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.3
• /
• pp.242-247
• /
• 2015

In this study, we propose the design of Radial Basis Function Neural Network(RBFNN) classifier in order to classify between precipitation and non-precipitation echo. The characteristics of meteorological radar data is analyzed for classifying precipitation and non-precipitation echo. Input variables is selected as DZ, SDZ, VGZ, SPN, DZ_FR, VR by performing pre-processing of UF data based on the characteristics analysis and these are composed of training and test data. Finally, QC data being used in Korea Meteorological Administration is applied to compare with the performance results of proposed classifier.

• Atmosphere
• /
• v.29 no.4
• /
• pp.451-461
• /
• 2019

Medium-range forecast is highly dependent on ensemble forecast data. However, operational weather forecasters have not enough time to digest all of detailed features revealed in ensemble forecast data. To utilize the ensemble data effectively in medium-range forecasting, representative weather patterns in East Asia in this study are defined. The k-means clustering analysis is applied for the objectivity of weather patterns. Input data used daily Mean Sea Level Pressure (MSLP) anomaly of the ECMWF ReAnalysis-Interim (ERA-Interim) during 1981~2010 (30 years) provided by the European Centre for Medium-Range Weather Forecasts (ECMWF). Using the Explained Variance (EV), the optimal study area is defined by 20~60°N, 100~150°E. The number of clusters defined by Explained Cluster Variance (ECV) is thirty (k = 30). 30 representative weather patterns with their frequencies are summarized. Weather pattern #1 occurred all seasons, but it was about 56% in summer (June~September). The relatively rare occurrence of weather pattern (#30) occurred mainly in winter. Additionally, we investigate the relationship between weather patterns and extreme weather events such as heat wave, cold wave, and heavy rainfall as well as snowfall. The weather patterns associated with heavy rainfall exceeding 110 mm day^-1 were #1, #4, and #9 with days (%) of more than 10%. Heavy snowfall events exceeding 24 cm day^-1 mainly occurred in weather pattern #28 (4%) and #29 (6%). High and low temperature events (> 34℃ and < -14℃) were associated with weather pattern #1~4 (14~18%) and #28~29 (27~29%), respectively. These results suggest that the classification of various weather patterns will be used as a reference for grouping all ensemble forecast data, which will be useful for the scenario-based medium-range ensemble forecast in the future.

• Journal of Korea Water Resources Association
• /
• v.46 no.3
• /
• pp.287-299
• /
• 2013

The objective of this study is to derive and evaluate the drought threshold level based on hydro-meteorological data using historical drought events. After collecting the drought events during 1991 to 2009 year, the observed meteorological data and estimated hydrological component from LSM are used as input for the percentile analysis that is drought analysis data. The drought threshold level that precipitation and runoff of 3 month duration are less than 35%, soil moisture of 2 month duration is less than 35% and evapotranspiration of 3 month duration is more than 65% is derived using ROC analysis that are objective test method. ROC analysis with SPI (3) is performed to evaluate the applicability of threshold level in the domestic. As a result, it can be concluded that the derived drought threshold level show better performance to reflect the historical drought events than SPI (3) and it reasonably explain the spatial drought situation through the spatial analysis.

• Atmosphere
• /
• v.31 no.1
• /
• pp.73-83
• /
• 2021

Numerical weather prediction (NWP) models play an essential role in predicting weather factors, but using them is challenging due to various factors. To overcome the difficulties of NWP models, deep learning models have been deployed in weather forecasting by several recent studies. This study adapts long short-term memory (LSTM), which demonstrates remarkable performance in time-series prediction. The combination of LSTM model input of meteorological features and activation functions have a significant impact on the performance therefore, the results from 5 combinations of input features and 4 activation functions are analyzed in 9 Automated Surface Observing System (ASOS) stations corresponding to cities/islands/mountains. The optimized LSTM model produces better performance within eight forecast hours than Local Data Assimilation and Prediction System (LDAPS) operated by Korean meteorological administration. Therefore, this study illustrates that this LSTM model can be usefully applied to very short-term weather forecasting, and further studies about CNN-LSTM model with 2-D spatial convolution neural network (CNN) coupled in LSTM are required for improvement.

• Atmosphere
• /
• v.17 no.1
• /
• pp.87-99
• /
• 2007

Simulation results of Asian Dust Aerosol Model (ADAM) for the period of April 7-9, 2006 were analyzed, comparing with observed PM10 data. ADAM simulated around ten times lower than on-site PM10 concentration in the source regions: Zhurihe, Tongliao, Yushe, Dalian and Huimin. As the result of this low concentration, transported amounts of Asian Dust were under-estimated as well. In order to quantify a forecasting accuracy, Bias and RMSE were calculated. Even though remarkably negative Biases and high RMSEs were observed, ADAM simulation had followed well up the time of dust outbreak and a transported path. However, the emission process to generate dust from source regions requires a great enhancement. The PM10 concentration at the surface reached up to $2,300{\mu}gm^{-3}$ at Baeknyoungdo and Seoul (Mt. Gwanak), up to $1,750{\mu}gm^{-3}$ at KGAWO about 18:00 LST in April 8, respectively; however, ADAM did not simulate the same result on its second peak. It is considered that traveling Asian dust might have been lagged over the Korean peninsula by the blocking of surface high pressure. Moreover, the current RDAPS's 30 km grid resolution (which ADAM adopts as the meteorological input data) might not adequately represent small-scale atmospheric motions below planetary boundary layer.

• Journal of Korean Society of Water and Wastewater
• /
• v.10 no.2
• /
• pp.88-103
• /
• 1996

The objective of this study is to develop stochastic and deterministic models which could be used to synthesize water application time series. Adaptive models using mulitivariate ARIMA(Transfer Function Model) are developed for daily urban water use forecasting. The model considers several variables on which water demands is dependent. The dynamic response of water demands to several factors(e.g. weekday, average temperature, minimum temperature, maximum temperature, humidity, cloudiness, rainfall) are characterized in the model by transfer functions. Daily water use data of Kumi city in 1992 are employed for model parameter estimation. Meteorological data of Seonsan station are utilized to input variables because Kumi has no records about the meteorological factor data.To determine the main factors influencing water use, autocorrelogram and cross correlogram analysis are performed. Through the identification, parameter estimation, and diagnostic checking of tentative model, final transfer function models by each month are established. The simulation output by transfer function models are compared to a historical data and shows the good agreement.

• Journal of the Korean earth science society
• /
• v.36 no.6
• /
• pp.567-579
• /
• 2015

The accurate simulation of micro-scale weather phenomena such as fog using the mesoscale meteorological models is a very complex task. Especially, the uncertainty arisen from initial input data of the numerical models has a decisive effect on the accuracy of numerical models. The data assimilation is required to reduce the uncertainty of initial input data. In this study, the limitation of the mesoscale meteorological model was verified by WRF (Weather Research and Forecasting) model for a summer fog event around the Nakdong river in Korea. The sensitivity analyses of simulation accuracy from the numerical model were conducted using two different initial and boundary conditions: KLAPS (Korea Local Analysis and Prediction System) and LDAPS (Local Data Assimilation and Prediction System) data. In addition, the improvement of numerical model performance by FDDA (Four-Dimensional Data Assimilation) using the observational data from AWS (Automatic Weather System) was investigated. The result of sensitivity analysis showed that the accuracy of simulated air temperature, dew point temperature, and relative humidity with LDAPS data was higher than those of KLAPS, but the accuracy of the wind speed of LDAPS was lower than that of KLAPS. Significant difference was found in case of relative humidity where RMSE (Root Mean Square Error) for LDAPS and KLAPS was 15.7 and 35.6%, respectively. The RMSE for air temperature, wind speed, and relative humidity was improved by approximately $0.3^{\circ}C$, $0.2m\;s^{-1}$, and 2.2%, respectively after incorporating the FDDA.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.14 no.2
• /
• pp.307-316
• /
• 1994

The rainfall forecasting model of the short term is improved at the point where meterological data is not gaged. In this study, the adopted model is based on the assumptions for simulation model of rainfall process, meteorological homogeneousness, prediction and estimation of meteorological data. A Kalman Filter technique is used for rainfall forecasting. In the existing models, the equation of the model is non-linear type with regard to rainfall rate, because hydrometer size distribution (HSD) depends on rainfall intensity. The equation is linearized about rainfall rate as HSD is formulated by the function of the water storage in the cloud. And meteorological input variables are predicted by emprical model. It is applied to the storm events over Taech'ong Dam area. The results show that root mean square error between the forecasted and the observed rainfall intensity is varing from 0.3 to 1.01 mm/hr. It is suggested that the assumptions of this study be reasonable and our model is useful for the short term rainfall forecasting at the ungaged point of the meteorological data.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.15 no.1
• /
• pp.144-158
• /
• 2012

A higher spatial resolution is preferable to support the accuracy of detailed climate analysis in urban areas. Airborne LiDAR (Light Detection And Ranging) and satellite (KOMPSAT-2, Korea Multi-Purpose Satellite-2) images at 1 to 4 m resolution were utilized to produce digital elevation and building surface models as well as land cover maps at very high(5m) resolution. The Climate Analysis Seoul(CAS) was used to calculate the fractional coverage of land cover classes in built-up areas and thermal capacity of the buildings from their areal volumes. It then produced analyzed maps of local-scale temperature based on the old and new input data. For the verification of the accuracy improvement by the precise input data, the analyzed maps were compared to the surface temperature derived from the ASTER satellite image and to the ground observation at our detailed study region. After the enhancement, the ASTER temperature was highly correlated with the analyzed temperature at building (BS) areas (R=0.76) whereas there observed no correlation with the old input data. The difference of the air temperature deviation was reduced from 1.27 to 0.70K by the enhancement. The enhanced precision of the input data yielded reasonable and more accurate local-scale temperature analysis based on realistic surface models in built-up areas. The improved analysis tools can help urban planners evaluating their design scenarios to be prepared for the urban climate.

• Atmosphere
• /
• v.25 no.2
• /
• pp.221-233
• /
• 2015

Pollen is closely related to health issues such as allergenic rhinitis and asthma as well as intensifying atopic syndrome. Information on current and future spatio-temporal distribution of allergenic pollen is needed to address such issues. In this study, the Community Multiscale Air Quality Modeling (CMAQ) was utilized as a base modeling system to forecast pollen dispersal from oak trees. Pollen emission is one of the most important parts in the dispersal modeling system. Areal emission factor was determined from gridded areal fraction of oak trees, which was produced by the analysis of the tree type maps (1:5000) obtained from the Korea Forest Service. Daily total pollen production was estimated by a robust multiple regression model of weather conditions and pollen concentration. Hourly emission factor was determined from wind speed and friction velocity. Hourly pollen emission was then calculated by multiplying areal emission factor, daily total pollen production, and hourly emission factor. Forecast data from the KMA UM LDAPS (Korea Meteorological Administration Unified Model Local Data Assimilation and Prediction System) was utilized as input. For the verification of the model, daily observed pollen concentration from 12 sites in Korea during the pollen season of 2014. Although the model showed a tendency of over-estimation in terms of the seasonal and daily mean concentrations, overall concentration was similar to the observation. Comparison at the hourly output showed distinctive delay of the peak hours by the model at the 'Pocheon' site. It was speculated that the constant release of hourly number of pollen in the modeling framework caused the delay.